Mar 4, 1991 - scheme (VHS) located in Jan Kempdrop, Northern Cape province of South ...... field derived spectra of salinised soils and vegetation as indicator of .... combination of salts and the type of soil surface, texture and organic matter content. ...... Remote Sensing and Image Interpretation (3rd. Ed.). John Wiley ...
Mapping and Modeling of Irrigation Induced Salinity of Vaal-Harts Irrigation Scheme in South Africa
By Olumuyiwa Idowu Ojo
Submitted in partial fulfillment of the requirements for the degree DOCTOR TECHNOLOGIAE In the Department of Civil Engineering FACULTY OF ENGINEERING AND THE BUILT ENVIRONMENT TSHWANE UNIVERSITY OF TECHNOLOGY
Supervisor: Prof F A O Otieno Co-Supervisor: Prof G M Ochieng’ Co-Supervisor: Dr B Mwaka
i
DECLARATION BY CANDIDATE I hereby declare that the thesis submitted for the degree Doctor Technologiae, at the Tshwane University of Technology, is my original work and has not previously been submitted to any other institution of higher education. I further declare that all the sources cited or quoted are indicated and acknowledged by means of a comprehensive list of references.
Olumuyiwa Idowu Ojo
Copyright: @Tshwane University of Technology 2013
ii
This study is dedicated to God Almighty who has been my all in all and to the following: 1. Engr. Remi Olu – my 1st degree (B. Eng) supervisor and mentor 2. Dr. Kola Ogedengbe – my 2nd degree (M.Sc. Eng.) supervisor and mentor 3. Prof. Fredrick Otieno – my 3rd degree (DTech) main supervisor and mentor
iii
ACKNOWLEDGEMENTS First and foremost, I would like to thank God, who made it possible for me to begin and complete this work successfully and also for his protection and favour in my life in the course of the programme.
Many thanks to my main Supervisor; Prof. Fredrick Otieno, for his fatherly role in my life and on the research work, co-supervisors; Prof George Ochieng’ cannot be forgotten for his wisdom and constant support and Dr Beason Mwaka for his supporting roles. All of them had patiently read and edited the manuscript and provided me valuable comments and pieces of advice throughout my thesis work. I would like to express my gratitude to Prof Julius Ndambuki whose mature touch on me always added the needed catalyst to make the work completed in time. The financial support from Tshwane University of Technology and Nigerian Educational Trust Fund (ETF) is thankfully acknowledged.
I am also thankful to Dr. J. Taiwo and Mr. A. Adeola, both of Department of Geography, University of Ibadan, Nigeria for assisting me greatly and constantly in the area of Remote Sensing and GIS Program training and analysis. The entire staff members of the department of Civil Engineering, Tshwane University of Technology, Pretoria are all wonderful; there is the need to mention; Ms Salome van der Merwe and Ms Daphney Ngoma. Also, the staff members of the Agricultural Research Council (ARC) and Department of Water & Environmental Affairs, Pretoria office, particularly Celeste, Eric, and Alloyous who provided me with various ancillary data and assistance during the course of the work are acknowledged. The staff members of Water Users’ Association in Vaal harts, Northern Cape were also very supportive during the field’s visits and questionnaire administration.
I have not enough words to express my feeling and gratitude to my family; my spouse (Gbemisola) who is like a wine that get sweeter with age and my wonderful children (Samuel, Shalom and Sharon) for their emotional and material support as well as the sacrifices and prayers from the beginning till now to enable me to complete successfully. Many thanks to my beloved mother (Elizabeth Oyinlola), my sister and her husband (Mr and Mrs Bademos) and my brothers (Babatunde and Taiwo) for their love, encouragement and support before, during and after the work. iv
I would like to thank my colleagues and friends: Dr Josiah Adeyemo, Dr Toyin Taiwo, Nkwonta Onyeka, Revd Amos Adeniyi and Fisayo Olufayo for sharing their wisdom and experiences as well as spending good time together with me during the course of the programme. My LAUTECH family members are also wonderful as they challenged and motivated me into this. They include; Prof. JO Ojediran (my mentor), Prof. SO Jekayinfa, Prof. WA Adekojo and Prof. Olajide. Others include Prof. Bola Fashina, Prof. Adegboola, Dr. Mrs. BA Adejumo, Engr. F Ola, Engr. K Oriola, Dr. A Adebayo, Engr. Mrs. M Adejumobi and all other staff members of the Department of Agricultural Engineering, Faculty of Engineering and Technology, to you all I say thank you.
Last but not the least; I would like to convey my special thanks to my God given family that are not blood related I appreciate; Dr. Josiah Adeyemo, Prof. Chris Enweremadu, Dr. Williams Kupolati , Dr. Mrs Patricia Popoola and Prof Rotimi Sadiku. Others are my South Africa based sister and brothers: Oluranti Agboola, Adesola Adegbola and David Adedokun.
v
ABBREVIATIONS ARC
Agricultural Research Council (of South Africa)
ASTER
Advanced Space Borne Thermal Emission and Reflection Radiometer
DEM
Digital Elevation Model
dS/m
Decisiemens per meter
ECe
Electrical conductivity of the saturation extract of soil
EIA
Environmental Impact Assessment
EOSAT
Earth Observation Satellite Company
ERDAS
A remote sensing application with raster graphics editor abilities for geospatial applications
ESP
Exchangeable Sodium Percentage
ETM+
Enhanced Thematic Mapper plus
FCC
False Colour Composition
GIS
Geographic Information System
GLCF
Global Land Cover Facility
GPS
Global Positioning System
IDRISI
Integrated geographic information system and remote sensing software
IWMI
International Water Management Institute
JPEG
Joint Photographic Experts Group
LAI
Leaf Area Index
MAGI
Maryland Geographic Information
MSS
Multi-Spectral Scanner
NDSI
Normalized Difference Salinity Index
NDVI
Normalized Difference Vegetation Index
NIR
Near Infrared
NOAA-AVHRR
The Advanced Very High Resolution Radiometer vi
PCA
Principal Component Analysis
pH
hydrogen ion concentration
RGB
Red, Green, Blue
SI
Salinity Index
SPOT
Système Pour l’Observation de la Terre
SRS
Satellite Remote Sensing
TCC
True Colour Composite
TDS
Total Dissolved Solids
TIR
Thermal Infrared
TM
Thematic Mapper
VHS
Vaal Harts Irrigation Scheme
vii
ABSTRACT Irrigation induced soil salinity is a critical environmental problem which has great impact on soil fertility and overall agricultural productivity. Satellite remote sensing (SRS) and GIS are modern tools that provide information on salinity variation over time essential for environmental monitoring and change detection. SRS and GIS can also help in the reduction of conventional time and labour expended by the expensive traditional method of field sampling techniques presently in use in South Africa.
The overall objective of this study was, therefore, the mapping and modeling of the spatiotemporal trend of soil surface salinity and land use / land cover pattern in Vaal Harts irrigation scheme (VHS) located in Jan Kempdrop, Northern Cape province of South Africa for sustainable land management and development that is eco-friendly.
The temperature, rainfall and relative humidity data of VHS were collected from South Africa Weather Service. Mean, variance, standard deviation and coefficient of variation including Spearman correlation coefficient and the autocorrelation coefficient were determined from the data. A Topographic/base map (scale: 1: 150 000), soil map and the flood line data, acquired in map form from the Agricultural Research Council and South Africa Department of Agriculture, were converted to JPEG in the form of point data in a GIS environment. Baseline surveys were conducted and structured questionnaires were administered among selected 40 farmers and a project manager in order to access other necessary information on VHS.
viii
In achieving the main study objective, Landsat (TM and ETM+) images data of VHS acquired for three Epochs; 1991, 2001 and 2005 totaling 15 years were obtained from the Global Land Cover Facility (GLCF) hosted by the University of Maryland, USA. Groundtruthing of the SRS data using a Garmin handheld GPS with a receiver accuracy of 2 m was used to obtain geographical coordinate of salt affected area. GIS assisted spatial analysis and modeling of the data collected (field and SRS) were carried out through data preprocessing of images clipped in ArcGIS 9.2 and ERDAS IMAGINE softwares, also image classifications were done using maximum likelihood method and Markov change prediction algorithm developed in IDRISI for the models.
There were 12 observations for each of three variables considered; temperature, rainfall and relative humidity. The average yearly temperature of the study area increases insignificantly by a constant of about 0.1168444. Coefficient of variation (CV) for temperature was found to be about 29.59 while the variance of this set of observations is 26.63. The involvement of non-zero values in the serial correlation indicated the significance of the deterministic component in the data. The Pearson correlation coefficient was used to test for the relationship between rainfall and temperature in order to determine their effect on the build up of soil salinity. Their relationship is significantly negative with P < 0.05.
The analyses of water samples from main, northern and western canals of VHS indicated corresponding average electrical conductivities of 0.009, 0.012 and 0.011 mmho/cm and mean SARs of 3.53, 3.00 and 3.89 respectively for the three canals. Electrical conductivity (EC) and total dissolved solids obtained are indicators of salinity. All the three water samples gave adjustable RNA of 0.12 to 0.17, thus indicating a medium to high salinity and sodium hazard tendencies in the scheme. ix
The static land use / land cover distribution pattern of VHS for the year 1991 to 2005 revealed a drastic change in the normal course of cultivated land. The year 2005 showed a remarkable increase of area of fallow land depicting a scenario of a likelihood of salinity problem. The overall percentage of fallow land has increased by 37.86 %; settlement or builtup area decreased by 18.97 %, water bodies decreased by 0.48 % and cultivated / irrigated land area also decreased by 18.21 %. The land use / land-cover map was classified with an overall accuracy of 93.14 %.
The temporal trend showed that there has been an increase in the area occupied by salt between 1991 and 2005. It was shown that the area covered by salt increased from 7,077.97 Km2 in 1991 to 7,117.61 Km2 in 2001 and to 12,651.52 Km2 in 2005. This means an increase of
39,642 Km2 were actually covered between 1991 and 2001 while this increased
significantly to 5,533.91 Km2 between 2001 and 2005. It was observed during the modeling, that there is the probability that much of the community area will show visible sign of salt (surface salinity) cover of 46% over the next fifteen years (2005 to 2020).
An overall conclusion drawn from this study is that the results obtained have shown the feasibility of using SRS (Landsat ET and ETM+) data and GIS to estimate soil salinity trend in VHS. Maps and model developed on salinity in this study are of great importance to planners, project engineers and managers in monitoring the consequences of land use change on the scheme. Based on this result, there is, therefore, an urgent need for management programme to be initiated in order to control the spread of salinity and reclaim the damaged land in order to make the scheme more economically viable.
x
TABLE OF CONTENT CHAPTER ONE ........................................................................................................................ 1 1.0 INTRODUCTION .............................................................................................................. 1 1.1 Environmental Effects ......................................................................................................... 3 1. 2 Irrigation problems ............................................................................................................. 4 1.2.1 Salinization and waterlogging .......................................................................................... 4 1.2.2 Pollution ............................................................................................................................ 4 1.3 Environmental issues and the use of remote sensing ........................................................... 5 1.4 Problem statement................................................................................................................ 6 1.4.1 Reason for the Study ......................................................................................................... 8 1.4.2 Aim and objectives ........................................................................................................... 9 1.4.3 Justification of the study ................................................................................................. 10 1.4.4 Scope of the research ...................................................................................................... 10 1.4.5 Relevance of the research ............................................................................................... 11 CHAPTER TWO ..................................................................................................................... 11 2.0 LITERATURE REVIEW .................................................................................................. 11 2.1 Environmental impacts of irrigation .................................................................................. 11 2.2 South Africa experience: Irrigation development and problems in Vaal harts .................. 20 2.3 Salinity assessment ............................................................................................................ 23 2.3.1 Direct method ................................................................................................................. 23 2.3.2 Indirect method: Geoinformatics and computational tools............................................. 24 2.3.2.1 Global Positioning System (GPS)................................................................................ 25 2.3.2.2 Satellite Remote Sensing (SRS) .................................................................................. 25 2.4 SRS for land use and land cover change ............................................................................ 26 2.4.1 SRS for mapping of salt-affected .................................................................................. 28 2.4.2 SRS applications in different countries .......................................................................... 31 xi
2.4.3 SRS Landsat and other Sensors ...................................................................................... 34 2.5 Visual interpretation using photo imagery......................................................................... 36 2.5.1 Digital analysis ............................................................................................................... 38 2.5.2 Digital analysis using surface vegetation index.............................................................. 39 2.5.3 Geographical Information System .................................................................................. 40 2.5.4 GIS in Soil Salinity Modeling ......................................................................................... 40 2.5.5 Soil Salinity analysis and modeling ................................................................................ 41 2.5.5.1 Multi-criteria decision evaluation ................................................................................ 41 2.5.5.2 Spectral Mixture modeling .......................................................................................... 42 2.5.5.3 Cover-radiance Relationships ...................................................................................... 42 2. 5. 5. 4 Vegetation Indices .................................................................................................... 42 2.5.5.5 Wireless sensor networks............................................................................................. 43 2.5.5.6 Hybrid Method: SRS data, field data and GIS tool ..................................................... 44 2.5.5.7 GIS tool: IDRISI Software .......................................................................................... 45 2.6 Modeling the Potential for Change................................. ..................................................44 2.6.1 Predicting changes modeling .......................................................................................... 46 2.6.2 Markov chains and transitional probabilities matrix ...................................................... 47 2.7 Statistical analysis .............................................................................................................. 47 2.7.1 Student’s T-test ............................................................................................................... 47 2.7.2 Pearson’s Chi-squared test .............................................................................................. 47 CHAPTER THREE ................................................................................................................. 48 3.0 STUDY AREA .................................................................................................................. 54 3.1 Location of the study area .................................................................................................. 54 3.2 Surface run-off, rainfall, temperature and evapotranspiration ........................................... 55 3.3 Geomorphology, Soil and Parent Material ........................................................................ 57 3.4 Infrastructure and irrigation methods................................................................................. 57 xii
3.5 Crop types .......................................................................................................................... 59 3.6 Irrigation development in VHS.......................................................................................... 59 CHAPTER FOUR....................................................................................................................64 4.0 MATERIAL AND METHODS ......................................................................................... 63 4.1 Selection of the study area..................................................................................................64 4.2 Analysis of climatic data .............................................................................................. 63 4.3 Field data collection ..................................................................................................... 64 4.4 Soil sampling and analytical methods ............................................................................... 66 4.4.1 Irrigation water sampling and analytical methods ......................................................... 66 4.5 Remote Sensing Data used ............................................................................................. 67 4.5.1 Preprocessing of the data ................................................................................................ 68 4.5.1.1 Radiometric correction of data .................................................................................... 70 4.5.1.2 Geometric correction of d ata ................................................................................. 70 4.5.2 Image classification methods .................................................................................... 77 4.5.3 Unsepervised classification............................................................................................. 77 4.5.4 Supervised classification ............................................................................................... 77 4.5.5 Pixel matching ............................................................................................................... 77 4.5.5.1 Training stage .......................................................................................................... 77 4.5.5.2 Classification stage ................................................................................................. 78 4.5.5.3 Maximum Likelihood Classifier ........................................................................... 79 4.5.5.4 Band selection .............................................................................................................. 79 4.6 Methods of satellite data analysis ...................................................................................... 80 4.6.1 Analyses for land use/ land cover ................................................................................... 80 4.6.2 Analyses for salinity ....................................................................................................... 81 4.6.3 Conversion of digital number to radiance....................................................................... 81
xiii
4.6.4 Conversion of radiance to radiance ................................................................................ 81 4.7 Modeling of land use and salinity ...................................................................................... 86 4.7.1 Overlay operation method .............................................................................................. 88 4.7.1.1 DEM map..................................................................................................................... 89 4.7.1.2 Land cover map ........................................................................................................... 89 4.7.1.3 Soil salinity map ......................................................................................................... 89 4.8 Statistical analysis .............................................................................................................. 92 4.8.1 Student’s test ................................................................................................................... 92 4.8.2 Pearson’s chi-squared test ............................................................................................... 92 4.9 Limitations in the Study..................................................................................................... 90 CHAPTER FIVE ..................................................................................................................... 94 5.0 RESULTS AND DISCUSSIONS ...................................................................................... 94 5.1 Results of weather data analysis ........................................................................................ 94 5.1.1 Temperature .................................................................................................................... 94 5.1.2 Rainfall.......................................................................................................................... 100 5.2 Irrigation water quality analyses ...................................................................................... 102 5.3 Base line survey ............................................................................................................... 106 5.4 Results from satellite image processing ........................................................................... 112 5.4.1Results from supervised classification ........................................................................... 112 5.4.2 Land use / land cover analysis ...................................................................................... 118 5.5 Change prediction using markov analysis ....................................................................... 121 5.6 Results from indices analysis ........................................................................................... 130 5.7 Distribution of the salt-affected areas in VHS.................................................................. 141 5.8 Validation of the method ................................................................................................. 151 5.9 Empirical models using soil salinity (EC) ........................................................................ 154 5.10 Findings from structured questionnaries ........................................................................ 158 xiv
CHAPTER SIX .....................................................................................................................169 6.0 CONCLUSIONS AND RECOMMENDATIONS...........................................................169 6.1 Conclusions ...................................................................................................................... 169 6.1.1 Introduction ................................................................................................................... 169 6.1.2 Research questions ........................................................................................................ 169 6.1.3 Land use/ land cover analysis........................................................................................170 6.1.4 Results from indices analysis ............................................................................................ 171 6.1.5 Salinity in VHS ............................................................................................................. 172 6.1.6 Modeling of Salinity ..................................................................................................... 173 6.1.7 General Conclusion ....................................................................................................... 173 6.2 Recommendations............................................................................................................174 REFERENCES ...................................................................................................................... 188 Appendix A ............................................................................................................................ 207 Appendix B ............................................................................................................................ 243 Appendix C ............................................................................................................................ 271 Appendix D ............................................................................................................................ 277 Appendix E ............................................................................................................................ 277
xv
LIST OF TABLES Table 2.1: Enviromental impact assessment of irrigation by basin ......................................... 15 Table 2.2: Irrigated land damaged by salinazation, top five irrigations and world estimate, mid-1980s Global estimate of secondary salinisation in the worlds irrigation lands ................................................................................................................................. 18 Table 2.3: Regional distribution of salt affected soils ............................................................. 18 Table 2.4: Global estimate of secondary salinity in the world’s irrigation lands .................... 19 Table 2.5: Irrigation development in South Africa .................................................................. 21 Table 2.6: Examples of various techniques in the detection of soil salinity ............................ 30 Table 3.1: Meterological data of Metehara town ..................................................................... 50 Table 4.1: Checklists for collation of field data ....................................................................... 59 Table 4.2: Landsat time-series used in the study ..................................................................... 62 Table 4.3: LMAX and LMIN values for landsat ..................................................................... 78 Table 4.4: Mean solar exoatmospheric irriadiansces ............................................................... 78 Table 5.1: Summary of the analysed temperature data ............................................................ 99 Table 5.2: Summary of the analysed temperature data ............................................................ 99 Table 5.3: Summary of the analysed weather data ................................................................ 102 Table 5.4: Pearson correlation coefficient for temperature and rainfall ................................ 102 Table 5.5: Results of irrigation water sample analysed ......................................................... 105 Table 5.6: Cover change between 1991 and 2005 ................................................................. 121 Table 5.7: Land-use/land-cover for the years with predicted for 2020.................................. 122 Table 5.8: Change in cover type in the study area (km2) ....................................................... 122 Table 5.9: Rate of change in cover type in the study area....................................................122 Table 5.10: General conversion pattern ................................................................................. 122 Table 5.11: Markov Transitional probability matrix ............................................................. 123 Table 5.12: Markov Transitional area matrix ........................................................................ 123 Table 5.13: Area of land-use: Farm land, fallow land, builtup and water body (1999) ........ 124 xvi
Table 5.14: Area of land-use: Farm land, fallow land, builtup and water body (2001) ........ 124 Table 5.15: Area of land-use: Farm land, fallow land, builtup and water body (2005) ........ 125 Table 5.16: Fertility rate vs irrigation practice ...................................................................... 160 Table 5.18: Fertility rate vs irrigation method......................................................................160 Table 5.19: Fertlity rate vs source of water ........................................................................... 161 Table 5.20a: Irrigation purpose vs irrigation practise ............................................................ 161 Table 5.20b: Reason for not practising irrigation vs irrigation method ................................. 162 Table 5.20c: Reason for not practising irrigation vs irrigation method ................................. 162 Table 5.21: Reason for not practising irrigation vs irrigation method ................................... 163 Table 5.22: Reason for not irrigation vs source of water ....................................................... 163 Table 5.23a: Crop yields vs irrigation practise.......................................................................164 Table 5.23b: Crop yields vs irrigation method ...................................................................... 164 Table 5.23c: Crop yield vs irrigationb method ...................................................................... 165 Table 5.24: Crop yield vs irrigation method .......................................................................... 165 Table 5.25: Yield vs source of water ..................................................................................... 166 Table 5.26a: Cultivable land area vs irrigation practise ........................................................ 166 Table 5.26b: Cultivable land area vs irrigation method ......................................................... 167 Table 5.26c: Cultivable land area vs irrigation method ......................................................... 167 Table 5.27: Cultivable land area vs irrigation method..........................................................167 Table 5.28: Cultivable land area vs source of water .............................................................. 168 Table 5.29a: Factors affecting yield vs irrigation practise ..................................................... 169 Table 5.29b: Factors affecting yield vs irrigation method ..................................................... 170 Table 5.30: Factors affecting yield vs irrigation method ....................................................... 171 Table 5.31: Factors affecting yield vs irrigation method ....................................................... 172 Table 5.32: Fators affecting yield vs sourcs of water ............................................................ 172 Table 5.33a: Effect of irrigation practise on the soil ............................................................. 173 xvii
Table 5.33b: Effect of irrigation practise on the soil vs irrigation method............................173 Table 5.34: Effect of irrigation practise on the soil vs irrigation method .............................. 174 Table 5.35: Effect of irrigation practise vs irrigation method ............................................... 175 Table 5.36a: Effect of irrigation practise of the soil vs source of water ................................ 175 Table 5.36b: Health problem vs irrigation practise ............................................................... 176 Table 5.36c: Health problem vs irrigation method ................................................................ 176 Table 5.37: Health problem vs irrigation method .................................................................. 177 Table 5.38a: Health problem vs irrigation method ................................................................ 178 Table 5.38b: Health problem vs source of water....................................................................179 Table 5.39a: Basic group statistics......................................................................................... 180 Table 5.39b: Indenpendent sample test .................................................................................. 180 Table 5.39c: Indenpendent sample test .................................................................................. 181 Table 5.39d: Indenpendent sample test..................................................................................181
xviii
LIST OF FIGURES Figure 2.1: An EM survey taken on a paddock basis at ground level ................................................... 23 Figure 2. 2: Global distribution of salt-affected soils ........................................................................... 47 Figure 2.3: Flow chart of remote sensing processing of satellite images.............................................49 Figure 3.1: Position of the study area (Vaal Harts) in relation to South Africa...................................50 Figure 3. 2: Canal detail.......................................................................................................54 Figure 3.3: Aerial Overview of VIS.......................................................................................................55 Figure 3.4: VHS main canal with low level of water.............................................................................55 Figure 3.5: VHS main canal ..................................................................................................................56 Figure 4.1: Mosaic Image of 1991.........................................................................................................74 Figure 4.2: Mosaic Image of 2001. ....................................................................................................... 74 Figure 4.3: Mosaic Image of 2005 ........................................................................................................ 75 Figure 4. 4: Flow chart for the processing of the SRS data .................................................................. 75 Figure 4.5: Processing of the SRS data in IDRISI window...................................................................76 Figure 4.6: Pre-processing of the SRS data for decision making to mode............................................ 77 Figure 4.7: Markov chain analysis processing of the SRS data ............................................................ 77 Figure 4.8: Classificationv of salt areas (Salinity areas) by satellite imagery. ..................................... 78 Figure 4.9: Classification of the salt affected areas based on geology land form/land cover and floodline level ..................................................................................................................... 79 Figure 5.1a: Autocorrelation graph of maximum temperature ............................................................. 96 Figure 5.1b: Autocorrelation graph of maximum temperature ............................................................. 96 Figure 5.1c: Autocorrelation graph of maximum temperature ............................................................. 96 Figure 5.2a: Average min. and max Temperature for the study area (1983-1992) ............................... 97 Figure 5.2b: Average min. and max Temperature for the study area (1993-2002)............................... 98 Figure 5.2c: Average min. and max Temperature for the study area (2003-2010) ............................... 98 Figure 5.3: Average yearly Temperature in VHS ................................................................................. 99
xviii
Figure 5.4: Average yearly rainfall for the study area....................................................................... .100 Figure 5.5: VHS main irrigation water supply canal..........................................................................103 Figure 5.6: VHS Northern irrigation water supply canal....................................................................104 Figure 5.7: VHS Western irrigation water supply canal................................................................104 Figure 5.8: Effect of Salinity in VHS................................................................................................. 106 Figure 5.9: Effect of Salinity in VHS................................................................................................. 107 Figure : 5.10: Effect of Salinity in VHS..............................................................................................107 Figure 5.11: Effect of Salinity in VHS in VHIS................................................................................. 108 Figure 5.12: Waterlogging in VHS in VHIS...................................................................................... 108 Figure 5.13: The contour map of the VHS developed from the GIS analysis .................................... 109 Figure 5.1 4: Digital Elevation Model (DEM) developed from GIS analysis ................................... 110 Figure 5.15: Digital Elevation Model (DEM): 3D view developed from GIS....................................111 Figure 5.16: Clipped Mosaic image of 1991....................................................................................... 114 Figure 5.17: Clipped Mosaic image of 2001....................................................................................... 115 Figure 5.18: Clipped Mosaic image of 1991....................................................................................... 116 Figure 5.19: Spectral signatures of VHS during 1991 ........................................................................ 117 Figure 5.20: Spectral signatures of VHS during 2001 ........................................................................ 117 Figure 5.21: Spectral signatures of VHS during 2005 ........................................................................ 118 Figure 5.22: Land cover map of VHS from 1991, 2001, 2005 and predicted 2019/2020...................120 Figure 5.23: Classified land use 1991 ................................................................................................. 126 Figure 5.24: Classified land use 2001.................................................................................................127 Figure 5.25: Classified land use 2005.................................................................................................128 Figure 5.26: Land-use: farm land, fallow land, built up, water body............................................129 Figure 5.27: Land-use: farm land, fallow land, built up, water body..................................................129 Figure 5.28: Land-use: farm land, fallow land, built up, water body..................................................130 Figure 5.29: Normalised difference vegetative index (NDVI) 1991...................................................132 Figure 5.30: Normalised difference vegetative index (NDVI) 2001,,,,,,.............................................133 xix
Figure 5.31: Normalised difference vegetative index (NDVI) 2005. ................................................. 134 Figure 5.32: 1991 Temperature distribution map of VHS as performed using GIS analysis ............. 135 Figure 5.33: 1991 Temperature distribution map of VHS as performed using GIS analysis ............ 136 Figure 5.34: 2001 Temperature distribution map of VHS as performed using GIS analysis..............137 Figure 5.35 2001 Temperature distribution map of VHS as performed using GIS analysis .............. 138 Figure 5.36: 2005 Temperature distribution map of VHS as performed using GIS analysis ............. 139 Figure 5.37: 2005 Temperature distribution map of VHS as performed using GIS analysis ............. 140 Figure 5.38: Salinity map indicating the salinity classification 1991 (low, medium, high) ............... 144 Figure 5.39: Salinity map indicating the salinity classification 2001 (low, medium, high) ............... 145 Figure 5.40: Salinity map indicating the salinity classification 2005 (low, medium, high) ............... 146 Figure 5.41: Salinity map indicating the salinity classification 2010 (low, medium, high)................147 Figure 5.42: Salinity map indicating the salinity classification 2020 (low, medium, high) ............... 148 Figure 5.43: Salinity chart indicating the salinity classification 1991.................................................149 Figure 5.44: Salinity chart indicating the salinity classification 2005................................................149 Figure 5.45: Projected salinity chart indicating the salinity classification.2010................................150 Figure 5.46: Projected salinity chart indicating the salinity classification.2015.................................150 Figure 5.47: Projected salinity chart indicating the salinity classification.2020.................................151 Figure : 5.48: Salinity graph indicating the salinity classification regression analysis .....................153 Figure 5.49: Projected salinity model chart ........................................................................................ 153 Figure 5.50: Projected salinity model chart ........................................................................................ 154 Figure 5.51: Electrical conductivity (EC) variation in the study area................................................156 Figure 5.52: NDVI and temperate plant giving 67% correlation ........................................................ 157
xx
xi
CHAPTER ONE 1.0 INTRODUCTION Irrigated agriculture is crucial to the economy, health and welfare of a very large part of the developing world and it is vital for world food security. It has contributed significantly to poverty alleviation, food security, and improving the quality of life for rural populations. However, irrigation project schemes are sometimes faced with the lack of sound objective assessments of their environmental and social implications. These schemes can only be adequately sustained by taking into consideration both environmental effects and the availability of funds for the maintenance of the implemented schemes. Irrigation can be used to accelerate economic growth with equity, self-reliance and to improve the health, income and living conditions of the poor majority. It can also be used to ensure equitable and sustainable use of the environment and natural resources for the benefit of present and future generations (MMSD-SA, 2001).
Three main systems must be considered in setting goals for sustainable development. These are the social system, the economic system and the natural system (MMSD-SA, 2001). State of the Environment reports (SOE), of which several hundreds have been published, addresses the following questions: What is happening in the environment? (Conditions and trends); Why is it happening? (Causes, links between human influences and natural processes); Why is it significant? (Ecological, economic and health effects); and what are we doing about it? (That is implications for planning and policy, Berger, 1997). Since sustainable development is defined as the integration of social, economic and environmental considerations, then an irrigation project that is developed and operated in an environmentally and socially acceptable manner could be seen as contributing to sustainable development. The benefits 1
derived from the irrigation system must develop an area in such a way that will survive long after the closure of the system (UNEP 2001). Despite these benefits, irrigated agriculture often radically changes land use, thus has major impacts on the environment especially soil degradation. These impacts may be on the natural and physical environments and also on the human environment (FAO, 2004). Communities lands can be affected by the ways in which environmental activities such as irrigation interacts with their physical environments. These activities include a large number of aspects; from disposal of fertigation / chemical salts, wastewater (drainage water) causing pollution of soil, water and air (Ashton, 1999).
The extent and degree of changes in the environment are influenced by the spatial variation of physical parameters in the form of soil types and land use. Therefore, environmental impacts of irrigation are the changes in quantity and quality of soil and water and the ensuing effects on natural and social conditions at the tail-end and downstream of the irrigation scheme. The impacts stem from the changed hydro- geological conditions owing to the installation and operation of the scheme.
An irrigation scheme often draws water from the river and distributes it over the irrigated area. As a result, it is found that, the downstream river discharge is reduced, the evaporation in the scheme is increased, the groundwater recharge in the scheme is increased, the level of the water table rises and the drainage flow is increased. These may be called direct effects. The effects thereof on soil and water quality are indirect and complex; water logging and soil sanilization are part of these, whereas the subsequent impact on natural, ecological and socioeconomic conditions is very intricate (ILRI, 1988).
2
1.1
Problems associated with irrigation
Irrigated agriculture is a stabilizing factor in food production scenario, but irrigation-related salinisation has resulted in a number of chemical-physical-biological problems jeopardizing sustainable irrigated agriculture. Soil deterioration due to increased salinity levels, rise in water table resulting into water logging, river or downstream water pollution is prevalent in most irrigation schemes worldwide. Ahmad and Singh (1991) reported that the problems of rising water tables and salinization in the areas of canal irrigation projects have affected large irrigation countries such as India, Pakistan, Afghanistan, Egypt, Iran, Sudan, Syria and the USSR. India has lost several million hectares of land and about 15 million hectares of irrigated land are already suffering from salinization and water logging in Pakistan. They also stated that about half of the world irrigated area has already been damaged to some degree by water logging and salinization and that much of the land expected to be irrigated in the future is highly vulnerable to similar damages, and on a global scale, at least 200 000 to 300 000 hectares of irrigated land are lost every year due to salinization and water logging.
It has been estimated that in South Africa more that 50% of the wetland ecosystem including irrigated land have been lost mainly through agricultural development and poor land management (Walmsley, 1988; DEAT, 1999). In South Africa, certain development activities including irrigation do have substantial detrimental effects on the environment, thus requiring a thorough Environmental Impact Assessment (EIA). It was reported that in South Africa, limited research has been conducted in the area of EIA (Sandham et al, 2008). De Villiers et al (2002) reported that irrigation schemes in South Africa are beset with problems and that only 37% of such schemes are commercially oriented. They envisaged that soil degradation, and notably salinization and water logging could in future, become a formidable problem jeopardizing the objectives concerning irrigated agriculture to create opportunities for 3
smallholder and empower poor farmers by improving the efficiency and therefore the competitiveness of this sector. They also found in the Nkomazi irrigation scheme for smallscale farmers that water logging, salinity and sodicity had reduced agricultural productivity within one year of irrigation to such an extent that reclamation was inevitable.
1.1.1 Salinisation and water logging The salinization of soils results in a number of chemical, physical and biological problems. According to the South African database (Barnard et al., 2002), an estimated 776 131 hectares of South African soils are strongly saline while soils with weak profile development usually occurring on flood plains that can be saline, comprise of 1 447 988 hectares. This is associated with waterlogging and pollution caused by irrigation, thus resulting into salinity.
Factors discovered to be involved in the salinization of irrigation schemes or soil suitability, poor design and methods of irrigation systems operation, and others are water quantity, quality and prevailing climate (Bowonder et al, 1996). According to a survey by the South African Department of Agriculture in 1990, out of 128, 000 hectares of cultivated land 54, 000 hectares are seriously alkaline and waterlogged and moderately affected. Bowonder et al (1996) stated that waterlogging and salinity are two major environmental problems that can arise in large irrigation projects.
1.1.2 Soil pollution Soil pollution is ever increasing since soil serves as the major medium to discharge the ever increasing volume of solid waste products and effluents produced by the growing population, industrialization and urbanization. According to Barnard et al (2002), the biggest contributor to the solid waste stream in South Africa is mining waste with 72%, pulverized fuel ash with 4
6.7%, urban waste with 4.5% and sewage sludge with 3.6%. The contribution of agriculture to the solid waste stream is estimated to be 0.02%. Water quality is in jeopardy as a result of pollution, as water is a limited and strategic resource.
Urbanization, industrialization, irrigation, inorganic fertilizers, herbicides and pesticides used by agriculture, polluted return flows and the recycling of water, largely contribute to water quality deterioration. Van der Merwe (1995) also noted the particular importance of an increasingly heavy mineral salt load as it renders water less suitable for urban and industrial use while it has a devastating effect on food crops. In the past decades, a consistent effort has been made in the field of agricultural research to improve the understanding of the physical processes involved in an irrigation system. More recently, the spread of modeling techniques using distributed parameters has largely encouraged the use of input data from remote sensing with the support of GIS for manipulating large data sets (Azzali et al., 2001).
1.2 Environmental issues and the use of remote sensing Remote sensing is a tool that can provide spatial and temporal information essential for environmental monitoring and change detection in mining areas (Schmidt et al., 1997; Schmidt and Glaesser, 1998; Serra et al., 2003; Manu et al., 2004). It also helps in the reduction of time and expensive field sampling methods (Limpitlaw and Woldai, 1998; Rathore and Wright, 1993; Haboudane et al., 2002). With the launching of the Landsat satellite in 1972, researchers began to use satellite data for monitoring environmental activities in different parts of the world (Coker, 1977). Sahin et al. (2006) used three Landsat geocover dataset from 1970 to 2000 to detect temporal changes in the Zonguldak coal test field. Brogaard and Prieler (1990) in the interim report submitted to the Institute for Applied Systems Analysis described how Landsat MSS can be used for the identification of broad 5
land cover changes of the Western part of Horqin steppe, Inner Mongolia Autonomous Region. Trinh et al. (2004) used Landsat images for studying land use dynamics and soil degradation in the Tamduong district of Vietnam. Eranani and Gabriels (2006) used Landsat data from 1976 to 2002 to detect changes in land cover in the Yazd-Ardakan basin, Iran. Latifovic and Fytas (2005) analyzed the land cover change of the Oil Sands Mining Development in Athabasca, Canada using information extraction method applied to two Landsat scenes.
New sensors with high spectral/temporal resolution such as the Advanced Space Borne Thermal Emission and Reflection Radiometer (ASTER) can allow even more precise land cover classification. Data collected by the MSS and TM will continue to be used as a historical global database (Stefanov et al., 2001). Dean et al. (2007) used processed satellite images to support environmental management, monitoring and sustainable development reporting of the Shell Canada and Albian Sands mine operations in northern Alberta. They used change detection methods to monitor the expansion of mining activity and its progressive reclamation on the environment. This has enabled more research activities in this regard with better results.
1.3 Problem statement According to Streutker et al. (1981), within 35 years of the existence of the Vaal Harts scheme (VHS), the fine sandy soils of this scheme were severely salinised. About 30 000 hectares of saline and saline-sodic soils (depth 0.3 m) at VHS resulted in about 1.4 – 2.1 million South African rand losses in gross income to the irrigation scheme farmers (Streutker et al., 1981). The installation of 218 drainage systems totalling 500 km of subsurface lateral drains at a cost of 2 million rand were undertaken between 1975 and 1977. Although the 6
subsurface drains were successful in keeping the groundwater table under 0.7m and leaching salts from recent salinised patches, there were still some 1 500 hectares of saline soils by the end of 1977 and an additional 1 000 hectares saline soils in 1980. A further 2 million rand then had to be invested to install sub-surface drains on the farm and to link these drains to the partially developed system of open storm water drains, in an attempt to lower the water table and to leach salts (Streutker et al., 1981; Du Plessis, 1986). To date more land area has been lost in the problem. Mapping of the land damaged due to salinization is a frantic task and requires a lot of manpower and time since it requires identification, sampling and classifying the land by conventional surveying methods and field sampling. Techniques like Remote Sensing (RS) and Geographical Information System (GIS) can do these tasks more efficiently.
The approach to the problem dealing with salt affected land using RS and GIS has been proven in many recent studies to be the most efficient (Rao et al., 1997; Schmidt et al, 1997; Serra et al, 2003, & Manu et al, 2004). Large coverage, good resolution in visible and near visible spectrum and repetitive passes are the advantages of RS Satellites and Advanced Analytical Techniques in GIS can also be useful in detection and intensity analysis of salt affected land (Moulders, 1987).
Lack of maps and models of land use / land cover and salinity trends in the Vaal Harts irrigation scheme affects the effective management of environmental problems in the scheme especially waterlogging and salinity problems thereby maximising their occurrence in the scheme over the years. Thus, this research work has mapped and modeled the irrigation induced salinity of the Vaal Harts irrigation scheme in Jan Kempdrop, Northern Cape, South Africa using Landsat SRS data and GIS tools. The result of this study provides a basic soil 7
salinity management guide for irrigation project managers, farmers and policy makers involved in irrigated agriculture particularly in the selected study area with a possibility of its usage in other irrigation schemes within the whole country.
1.4 Scope of the research The research was limited to Vaal-Harts irrigation scheme in the Northern Cape Province with particular focus on large scale surface salinity in the scheme. Although irrigation started in the 1930’s in the scheme, the study covered 15 years from 1991 to 2005; this is due to the date of the earliest satellite data collected.
1.5 Relevance of the research Since the global tendency is towards sustainability, a critical analysis and modeling of irrigation impacts on the environment will help determine whether the current land-use and socioeconomic situation is sustainable in the Vaal Harts scheme. Therefore, this research is important to assist in the sustenance of the irrigation scheme through the application of the SRS and GIS techniques which are safe and cost effective in the timely assessment of environmental impacts in terms of surface salinity. This view is similar to the one put forward by Jensen (1983) that environmental impact assessment should be carried out periodically in order to identity changes in the environment caused by irrigation development.
1.6 Research questions This study answered the following research questions: 1. What are the significant environmental impacts that have arisen due to irrigation and associated spatial developments in the Vaal Hart’s irrigation scheme? 8
2. Can land use changes be assessed with satellite imagery? 3. Where has salinity occurred in VHS and what are their extent? 4. What are the likely factors that influence land use change in the scheme? 5. And can these trends be modelled for forecasting purposes in the scheme?
1.6.1 Aims and objectives The aims of this study were to map and model the salinity trends in Vaal Harts irrigation schemes in Jan Kempdrop, Northern Cape Province of South Africa using Landsat SRS, GIS computational tools (IDRIS and ERDAS). The resulted maps and models will aid the effective management of salinity problem and thereby minimizing its occurrence in the future through planning and policy formulation, on the irrigation scheme within the context of environmental sustainability. The specific objectives of the study were: 1.
To identify spatio-temporal trends in soil salinity boundary (extent) in the study area by using empirical methods of satellite remote sensing (SRS) data for fifteen (15) years (1991 to 2005);
2.
To model and forecast the future trends in soil salinity based on historical patterns observed in the study area;
3.
To identify spatio-temporal trends of land use/land cover changes in the study area by using empirical methods of satellite remote sensing (SRS) data for fifteen (15) years (1991 to 2005);
4.
To analyze the relationship between soil salinity and temperature regime; and
5.
To analyze the perceived effect of irrigation induced salinity on soil fertility and crop yield using a structured questionnaire.
9
1.6.2 Justification of the study Monitoring environmental changes and time series analysis is quite difficult with traditional method of soil sampling and field surveying. Geoinformatics approach is a technique which involved integrating Satellite remote sensing (SRS) data and Geographical information system (GIS) tools. They both provide information on variation over time essential for environmental monitoring and change detection in land usage (Schmidt et al., 1997; Serra et al., 2003; Manu et al., 2004). They also help in the reduction of conventional time consuming and expensive field sampling method (Haboudane et al., 2002). Since soil salinity is a dynamic process, SRS and GIS can be used for its mapping and modeling. The maps and models developed will thus, serve as important tools for planners in monitoring the consequences of land use change in the socioeconomic environment of the study area.
10
CHAPTER TWO 2.0 LITERATURE REVIEW 2.1 Environmental impacts of irrigation Economic, social and environmental change is inherent to any development. The need to avoid adverse impacts and to ensure long term benefits led to the concept of sustainability. In order to predict environmental impacts of any development activity and to provide an opportunity to mitigate against negative impacts and enhance positive impacts, the environmental impact assessment (EIA) procedure was developed in the 1970s as a useful tool to identify potential consequence of unsustained development projects.
Environment and development are complementary and interdependent and thus, EIA is a technique for ensuring that the two are mutually reinforcing (Wathern, 1988). EIA was defined as a formal process to predict the environmental consequences of human development activities and to plan appropriate measures to eliminate or reduce adverse effects and to augment positive effects. EIA was also defined as an instrument to identify and evaluate the potential environmental consequences of a development action in order to support decision making and sound environmental management (Glasson et al., 1999; Wood, 2003).
The United Nations Economic Commission for Europe in 1991 put together a more succinct definition of EIA as: “an assessment of the impact of a project on the environment”. In essence, EIA is a systematic process that examines the environmental consequences of developmental actions in advance. The emphasis compared with many other mechanisms for environmental protection is on protection (Glasson et al, 1999). Project planners have 11
traditionally assessed the impacts not on the systematic, holistic and multidisciplinary way required by EIA. Internationally, EIA is one of the most successful and widely adopted environmental policy implementation instruments that have emerged over the past three decades (Sandham et al, 2008). Environmental assessment is appropriate for both site specific projects and wider programmes. EIA provides more opportunities to correct situations where the environment is adversely affected (Tiffen, 1989). Therefore, the three main functions of EIA are predicting the likely problems, find ways to avoid them and to enhance the positive effects of such projects.
Irrigated agriculture often radically changes land use and is a major consumer of fresh water. It thus has a major impact on the environment especially soil degradation. This degradation may extend both upstream and downstream of the irrigated area. The impacts may be both to the natural, physical environments and to the human environment (FAO, 2004).
Irrigation schemes are sometimes embarked upon in the absence of sound objective assessments of their environmental and social implications. These schemes can only be adequately sustained by taking into consideration both environmental effects and the availability of funds for the maintenance of the schemes. An irrigation system consists of the following segments (Sarma and Murthy, 2004): 1. Catchment area- from where water flows to the river and the dam; 2. Dam and water spread areas of the reservoir; 3. Water distribution system (main canal, branch canals if any, distributaries, outlets and field irrigation channels) and; 4. Command area-where crops are grown.
12
The positive impacts of irrigation in the expansion and intensification of agriculture also have the potential to cause: increased erosion; pollution of surface water and ground water from agriculture chemicals; deterioration of water quality; increased nutrient levels in the irrigation and drainage water resulting in algal blooms, proliferation of aquatic weeds and eutrophication in irrigation canals and downstream waterways. Also, poor water quality below an irrigation project may render the water unfit for other users, harm aquatic species and because of its high nutrient content, result in aquatic weed growth that obstructs waterways and has health, navigation and ecological consequences. Elimination of dry season die-back and the creation of a more humid micro climate may result in an increase of agricultural pests and plant diseases (Graham and Singh, 1997; Singh, 1995).
Generally, the major potential negative environmental impacts of most large irrigation projects for which literature is replete of are: water logging and salinization of soils, increased incidence of water – borne and water – related diseases, possible negative impacts of dams and reservoirs, problems of resettlement or changes in the lifestyle of local populations. The environmental impact assessment of irrigation by world basins according to FAO, 2004, is summarized in Table 2.1 totalling 42,504,00 million ha. Irrigated agriculture also has impacts on the water environment as related to abstractions, return flows, influences on water quality and modifications to the hydrological regime due to changed land use.
Afoz and Singh (1991) categorised the negative impacts as follow: 1. Impacts during construction stage and 2. Problems that may be created after its operation.
13
Environmental impacts of irrigation are the changes in quantity and quality of soil and water as a result of irrigation and the ensuing effects on natural and social conditions at the tail-end and downstream of the irrigation scheme. The development of large-scale irrigation projects, which involves diversions of rivers, construction of large reservoirs and the irrigation of large landscapes, causes large changes in the natural water and salt balances of entire hydro geologic systems. Large and increasing proportions of the world's irrigated land are deleteriously affected by water logging and excessive salinity (Afoz and Singh, 1991). Table 2.1 presents environmental impact hazards of irrigation of world basins in terms of salinity, health, forest, fishery and wildlife against the irrigation potential area in hectares.
14
Table 2.1: Environmental impact assessment of irrigation of world basins Basin
Irrigation potential
Environmental impact hazard
(100 ha)
Salinity
Health
Forest
Fishery
wildlife
Senegal river
420
+++
++
+
+
+
Niger river
2817
+++
++
+
++
++
Lake Chad
1163
+++
++
+
++
++
Nile river
8 000
+++
+
+
+
++
Rift Valley
844
+
++
+
+
+
Shebelli-Juba
351
+++
+
+
+
+
Congo/Zaire river
9 800
+
+
++
+
+
Zambezi river
3 160
++
++
+
+
+
Okavango
208
++
+
+
+
+++
Limpopo river
295
++
++
+
+
+
Orange river
390
++
+
+
+
+
South interior
54
+++
+
+
+
+
North interior
71
+++
+
+
+
+
Mediterranean Coast
850
+++
+
+
+
+
North West Coast
1 200
+++
+
+
+
+
West Coast
5 113
+
++
+
+
+
835
+
++
++
+
+
1 808
++
++
+
+
++
South Atlantic Coast
84
++
+
+
+
+
Indian Ocean Coast
1 500
+
+
+
+
+
East Central Coast
1 928
+
++
+
+
+
78
++
+
+
+
+
1 500
+
++
+
+
+
35
++
+
+
+
+
West Central Coast South West Coast
North east coast Madagascar Islands Total +++: Serious
42 504 ++: moderate
+: Low or nil
Source: FAO (2004)
15
The problems of water logging and secondary salinity prevalent in most irrigated lands has resulted from the excessive use of water for irrigation, inadequate and inappropriate drainage management and the discharge of drainage water into good-quality water supplies which are used elsewhere for crop production (Jensen et al., 1990; Biswas, 1990). The USDA Salinity Laboratory defines a saline soil as having an ECe of 4 dS/m or more. ECe is the electrical conductivity of the ‘saturated paste extract’, that is, of the solution extracted from a soil sample after being mixed with sufficient water to produce a saturated paste.
The moisture content of a drained soil at field capacity may be much lower than the water content of its saturated paste. Under dry land agriculture, the soil water content might drop to half of field capacity during the life of the crop. The actual salinity of a rain-fed field whose soil had an ECe of 4 dS/m could be 8-12 dS/m (USDA, 1995). Salinization is a natural process in soils and water, especially in areas of water deficit, but various human activities are increasing its extent and severity a process called accelerated or enhanced salinization (Szabolcs and Varallyay, 1979; Goudie, 2003).
Goudie and Viles (1997) stated that increasing salinity accelerates the weathering of buildings and engineering structures. It is estimated that approximately 25 % of the world's irrigated land is damaged by salinisation (Postel, 1989). The world’s total land area under affected soils at the present time is estimated at over 950 million hectares. Approximately 40 or 50 % of the irrigated land in the arid and semiarid regions in the world has some degree of soil salinity and /or sodicity problems (Szabolcs, 1987). Adams and Hughes, 1990 claimed that up to 50 % of the world's irrigated land may be affected by salt. According to Rhoades (1988) salt-related problems occur within the boundaries of at least seventy-five countries. Countries with notable salinity problems include Australia, China, Egypt, India, Iraq, 16
Mexico, and Pakistan. Others are the republics of the ex-Soviet Union, Syria, Turkey, and USA. According to Peck (1978) the area of irrigated land has increased from roughly 8 million hectares (20 million acres) at the end of the eighteenth century to 250 million hectares at the end of the twentieth. Increases in soil and water salinity are not restricted to irrigated areas. In certain parts of the world, such as Australia, salinisation has resulted from vegetation clearance. This is called “dryland salinity.” Similar problems exist in North America, notably in Manitoba, Alberta, Montana, and North Dakota (Bari and Schofield, 1992). It has been estimated that the area of salt-affected and waterlogged soils amounts to 50% of the irrigated area in Iraq, 23% in Pakistan, 50% in the Euphrates Valley of Syria, 30% in Egypt, and over 15% in Iran (Worthington, 1977).
Ahmad and Singh (1991) reported that the problems of rising water tables and salinisation in the areas of canal irrigation projects have affected large irrigation countries such as India, Pakistan, Afghanistan, Egypt, Iran, Sudan, Syria and the former USSR. India has lost several million hectares of land and about 15 million hectares of irrigated land are already suffering from salinization and water logging in Pakistan. Rozanov et al. (1991) stated that the global areas of irrigated land increased from 50,000 to 2,200,000 km2, from 1700 to 1984, while at the same time some 500,000 km2 were abandoned as a result of secondary salinisation. They believe that in the last three centuries total soil loss due to irrigation is 1 million km2 of land destroyed, plus 1 million km2 of land with diminished productivity due to salinisation. They also stated that about half of the world’s irrigated area has already been damaged to some degree by water logging and salinization and that much of the land expected to be irrigated in the future is highly vulnerable to similar damage, and, on a global scale, at least 200 000 to 300 000 hectares of irrigated land is lost every year due salinization and water logging. In Africa, less than 10 % of salt-affected soils are also affected by human action (Thomas and 17
Middleton, 1993). South Africa has total irrigatable land of 13 million ha with about 1 million ha under irrigation giving a total of 9% of the total area being cropped. Also the area of irrigated land that is salt affected is 0.1 million ha giving a total of about 10% of irrigated land area (Ghassemi et al, 1995). Tables 2.2, 2.3 and 2.4 present data on the extent of saltaffected lands worldwide.
Table 2.2: Irrigated land damaged by salinisation in world top five irrigators Country
Area damaged (million hectares)
The share of irrigated land damaged (%)
India
20.0
36
China
7.0
15
United States
5.2
27
Pakistan
3.2
20
ex-Soviet Union
2.5
12
Total
37.9
24
WORLD
60.2
24
Source: (Postel, 1989) Table 2.3: Regional distribution of salt-affected soils Regions
Total area Mha
Saline soils
Sodic soils
Mha
%
Mha
%
Africa
1,899
39
2.0
34
1.8
Asia, the Pacific and Australia
3,107
195
6.3
249
8.0
Europe
2,011
7
0.3
73
3.6
Latin America
2,039
61
3.0
51
2.5
Near East
1,802
92
5.1
14
0.8
North America
1,924
5
0.2
15
0.8
Total
12,781
397
3.1%
434
3.4%
Source: FAO Land and Plant Nutrition Management Service (2005)
18
Table 2.4: Global estimate of secondary salinisation in the world's irrigated lands
Country
Total land
Area irrigated
area cropped
Area of irrigated land that is saltaffected
Mha
%
Mha
%
Mha China
97
45
46
6.7
15
India
169
42
25
7.0
17
USSR
233
21
9
3.7
18
United States
190
18
10
4.2
23
Pakistan
21
16
78
4.2
26
Iran
15
6
39
1.7
30
Thailand
20
4
20
0.4
10
Egypt
3
3
100
0.9
33
Australia
47
2
4
0.2
9
Argentina
36
2
5
0.6
34
South Africa
13
1
9
0.1
9
Subtotal
843
159
19
29.6
20
World
1,474
227
15
45.4
20
Source: Ghassemi et al. (1995)
2.2 South Africa experience: Irrigation development and problems Irrigated agriculture is a stabilizing factor in South Africa’s food production scenario, responsible for 30% of total food production. According to Barnard et al. (2002), a total of 231, 362 hectares of farmland were under irrigation as at 1910. This increased to 736,932 hectares in 1965 and to 1, 300, 000 ha in 2000, Table 2.3 gives the details of the regional distribution of salt affected soils with Africa region having 39 Mha saline soils and 34 Mha sodic soils out of a total of 397 Mha saline soils and 434 Mha sodic soils respectively. An 19
estimated 100, 000 ha out of the 885, 000 ha of the land of South African state irrigation schemes were salt-affected by 1976 (Du Plessis, 1986). The cost of reclamation is justified only if maintenance of the water table does not require a considerable amount of overirrigation and does not cause pollution down-stream (Streutker et al., 1981) which requires a high level of irrigation and drainage management.
Irrigation related salinisation results in a number of chemical-physical-biological problems jeopardizing sustainable irrigated agriculture in South Africa. Water logging mostly associated with irrigation is an incessant countrywide problem while the salinisation of both soil and water is a threat to irrigated agriculture. Bowonder et al (1996) stated that water logging and salinity are two major environmental problems that are likely to occur in large irrigation projects. Van der Merwe (1995) also noted the particular importance of an increasingly heavy mineral salt load as salinity as it renders water less suitable for urban and industrial use while it has a devastating effect on food crops.
De Villiers et al (2002) reported that South African irrigation schemes are beset with salinity and waterlogging problems and that only 37% of the participants are commercially oriented. It is envisaged that soil degradation, and notably salinization and water logging could in future become a formidable problem jeopardizing the objectives of irrigated agriculture which is to create opportunities for smallholder and resource poor farmers by improving the efficiency and therefore the competitiveness of this sector. They also found in the Nkomazi irrigation scheme for small-scale farmers that water logging, salinity and sodicity had reduced agricultural productivity within one year of irrigation to such an extent that reclamation is inevitable.
20
The salinisation of soils results in a number of chemical, physical and biological problems. According to Barnard et al. (2002), an estimated 776, 131 ha of South African soils (arable lands) are strongly saline while soils with weak profile development usually occurring on flood plains that could be saline. They stated that salinity is most often associated with water logging and pollution caused by irrigation. From some literature, factors identified to be responsible for potential salination at irrigation schemes are soil suitability, poor design and methods of irrigation systems operation, water quantity and quality. In another survey by the South African Department of Agriculture in 1990, out of 1, 128, 000 ha of cultivated land, 154, 000 ha is seriously alkaline and waterlogged and moderately affected (Barnard et al., 2002).
This is alarming and calls for drastic actions and demanding researches to be
conducted to find ways of effectively reducing the rate of this menace. Table 2.5 presents information about irrigation development in South Africa detailing year and total area irrigated.
Table 2.5: Irrigation development in South Africa Year
Total Area Irrigated (ha)
1910
231 362
1924
318 767
1965
736 932
1996
1 290 132
2000
1 300 000
Source: Barnard et al., 2002
2.2.1 Irrigation development and problems in Vaal harts In 1938, Act No. 38 of 1934 was approved, providing permission to construct the Vaal Dam and develop the Vaal Harts Irrigation Scheme. The area stretches from Jan Kempdrop in the 21
south to Taung (the Dry Harts River) in the north. Salt deposited through irrigation water amounts to 4.65 t/ha per annum (Streutker et al., 1981). Flood irrigation commenced at the VHS established in 1940 covering more than 30,000 ha and soon after, the groundwater table rose to between 0.9 and 1.5 m over the whole area. In 1956 several salinisation cases were reported and in the 1960s, a number of soil profiles from all over the Scheme, contained more salts in the subsurface than measured during the initial soil survey, indicating a disturbing tendency, although not alarming at that stage (Streutker et al., 1981).
According to Streutker et al., (1981), within 35 years of VIS existence, the fine sandy soils of this scheme were severely salinised. They further stated that between the periods of 1974– 1976, the semi-arid Vaal Hart’s area (with 218 farms) received above average rainfall resulting in a critical level of soil salinisation. Streutker et al. (1981) reported that the reclamation of some 30, 000 ha saline or saline-sodic soils (depth 0.3 m) at the VHS resulted in 1.4 - 2.1 million South African rand losses in gross income for the irrigation scheme farmers. They reported a further re-investment of 2 million rand was used to install subsurface drains on the scheme to link to the partially developed system of open storm water drains, in an attempt to lower the water table and to leach salts using drains.
The installation of 218 drainage systems totalling 500 km of subsurface lateral drains at a cost of 2 million rand was undertaken between 1975 and 1977. Although the subsurface drains were successful in keeping the groundwater table below 0.7 m and leaching salts from recent salinised patches, there were still some 1 500 ha of saline soils by the end of 1977 and 1 000 ha in 1980. In VHS, the total dissolved salts (TDS) averaged 1 005 mg/l in 1976 and 1 350 mg/l in 2004. As at 2009 the TDS was 1 476 mg/l, representing an increase of 96 mg/l in
22
5 years; an average increase per annum of 19.25 mg/l. Irrigated salt deposits that are not drained out are build up in the soil at the rate of 0.8 t/ha per annum (Streutker et al., 1981).
2.3 Salinity assessment 2.3.1 Direct method Saline fields can be simply identified by the presence of spotty white patches of precipitated salts. Such precipitates usually occur in elevated or non vegetated areas, where the water evaporates and leaves salt behind. Soil salinity is measured as the salt concentration of the soil solution in terms of Electrical conductivity (EC) in dS/m. The standard for the determination of soil salinity is from an extract of a saturated paste of the soil and is expressed as ECe. The salinity can be measured in a 2:1 or 5:1 water: soil mixture (in terms of g water per g dry soil) (ILRI, 2003). Soils are considered saline when the ECe > 4. At 4 < ECe < 8, the soil is called slightly saline, at 8 < ECe < 16 it is moderately saline, and at ECe > 16 the soil is severely saline (Richards, 2010).
Electrical conductivity of irrigation water can also be measured by the handheld conductivity meter. Irrigation water quality is often expressed as total soluble salts; an international convention being that 1 dS/m is equivalent to 640 mg/L of mixed salts. Soil salinity on a large scale is mapped with an electromagnetic (EM) conductivity meter. An EM38 survey can be used to map the extent of subsoil salinity as well as discharge areas. Figure 2.1 illustrates the heterogeneity of soil salinity as shown by the EM meter.
23
Figure 2.1: An EM survey taken on a paddock basis at ground level (Source: Rampant, 2004, Department of Primary Industry, Bendigo, Australia).
2.3.2 Indirect method: Geoinformatics and computational tools Geoinformatics approach is a modern technique which involves the combination of Global Positioning System (GPS), Satellite remote sensing data (SRS) and Geographical information system (GIS) modeling tools. This is because salinity is a dynamic process and to assess the extent of salinity, modeling is often required. Geoinformatics involves the acquisition, processing, analysis and management of geographic or spatial information. Spatial information is concerned with knowing what (object) is where (space) and when (time). Data is collected using techniques such as GPS, remote sensing, orthography, total station and the use of the more traditional surveying equipment such as theodolites and level. Among these techniques the newer and most prominent of them are the GPS and remote sensing.
24
2.3.2.1 Global Positioning System (GPS) GPS is a satellite positioning and a navigating system technology that continues to gain wide usage and applications worldwide. Inputting of data directly on the field through the use of GPS, saves time and greatly facilitates subsequent processing. The GPS data logger permits direct interfacing with GIS which is the database Management System for spatial analysis (Adeleke and Aremu, 2002).
2.3.2.2
Satellite Remote Sensing (SRS)
According to Olorunfemi (1983), monitoring environmental changes and time series analysis is quite difficult with traditional method of surveying. In recent years, satellite remote sensing techniques have been developed, which have proved to be of immense value in preparing accurate land use, land cover maps and monitoring changes at regular intervals of time. In case of inaccessible region, this technique is perhaps the only method of obtaining the required data on a cost and time effective basis. A remote sensing device records response which is based on many characteristics of the land surface, including natural and artificial cover.
An interpreter uses the element of tone, texture, pattern, shape, size, shadow, site and association to derive information about land cover. The generation of remotely sensed images of various types of sensor flown aboard different platforms at varying heights above the terrain and at different times of the day and the year does not lead to a simple classification system. It is often believed that no single classification could be used with all types of imagery and all scales. Xiaomei and Rong Qing in 1999 noted that information about the change is necessary for updating land cover maps and the management of natural resources. 25
The information may be obtained by visiting sites on the ground and or extracting it from remotely sensed data. Singh (1989) defined change detection as the process of identifying differences in the state of an object or phenomenon by observing it at different times. Change detection is an important process in monitoring and managing natural resources and urban development because it provides quantitative analysis of the spatial distribution of the population of interest. Macleod and Congation (1998) listed four aspects of change detection which are important when monitoring natural resources. These are: 1. Detecting the changes that have occurred; 2. Identifying the nature of the change; 3. Measuring the areal extent of the change; and 4. Assessing the spatial pattern of the change
2.4 SRS for land use and land cover changes Often, land use and land cover change may result in environmental, social and economic impacts of greater damage than benefit to the area (Moshen, 1999). Therefore data on land use change are of great importance to planners in monitoring the consequences of changes in land use in the area. Such data is of value to resources management and agencies that plan and assess land use patterns and in modeling and predicting future changes.
In an investigation by Shosheng and Kutiel (1994), they discovered the advantages of remote sensing techniques in relation to field surveys in providing a regional description of vegetation cover. The results of their research were used to produce four vegetation cover maps that provided new information on spatial and temporal distributions of vegetation in this area and allowed regional quantitative assessment of the vegetation cover. Arvind and 26
Nathawat (2006) carried out a study on land use and land cover mapping of Panchkula, Ambala and Yamunanger districts, Hangana State in India. They observed that the heterogeneous climate and physiography conditions in these districts had resulted in the development of different land use/land cover in these districts. An evaluation by digital analysis of satellite data indicates that the majority of areas in these districts is used for agricultural purpose. It is inferred that land use and land cover pattern in the area are generally controlled by agro climatic conditions, ground water potential and a host of other factors. Generally, satellite imagery is able to provide a more frequent data collection on a regular basis unlike aerial photographs which although may provide more geometrically accurate maps, is limited in respect to its extent of coverage and is expensive.
In 1985, the U.S Geological Survey carried out a research program to produce 1:250,000 scale land cover maps for Alaska using Landsat MSS data (Fitz Patrick et al, 1987). The State of Maryland Health Resources Planning Commission also used Landsat TM data to create a land cover data set for inclusion in their Maryland Geographic Information (MAGI) database. All seven TM bands were used to produce a 21 class land cover map (EOSAT 1992). Also, in 1992, the Georgia Department of Natural Resources completed mapping the entire State of Georgia to identify and quantify wetlands and other land cover types using Landsat Thematic Mapper data (ERDAS, 1992). The State of South Carolina Lands Resources Conservation Commission developed a detailed land cover map composed of 19 classes of TM data (EOSAT, 1994). This mapping effort employed multi-temporal imagery as well as multispectral data during classification. Dimyati (1995) conducted an analysis of land use and land cover changes using the combination of MSS Landsat and land use map of Indonesia. The implication of this is that Landsat Thematic mapper data is suitable for developing land use and land cover pattern maps for irrigated land. 27
2.4.1 SRS for mapping of salt-affected areas Application of satellite remote sensing for surveying and mapping of salt-affected areas began with the use of black and white photography. The relatively bright appearance provides the information about salinity due to the efflorescence of salt crust. The effect of salinity on crops provides the information on salinity indirectly. The aerial photographs have been used to delineate units based on the combination of geomorphologic differences and differences in grey tones. Attempts were also made to relate the differences in the grey tones with the salt content (Coker, 1977). Indirect features like landscape may help to identify the problems of soil salinity. Relative elevation is one of the most evident landscape features in relation to salinity and moisture provided by saline and shallow groundwater table.
Satellite Remote sensing (SRS) data are modern tools that provide information on variation over time essential for environmental monitoring and change detection (Schmidt et al., 1997; Schmidt and Glaesser, 1998; Serra et al., 2003; Manu et al., 2004). They also help in the reduction of conventional, time consuming and expensive field sampling methods, which is used in the traditional method of monitoring pollution and salinity (Limpitlaw and Woldai, 1998; Rathore and Wright, 1993; Haboudane et al., 2002). Dehaan and Taylor (2002) used field derived spectra of salinised soils and vegetation as indicator of irrigation induced soil salinisation for identification of saline soil regions. Foopa et al. (2002) used spectral unmixing in snow cover estimation using NOAA-AVHRR data and examined the ability of real time snow cover estimation at subpixel level. Okin (2001) demonstrated the use of multiple end member spectral mixture analysis (MESMA) in retrieving information about soil. MESMA is capable of mapping soil surface types even when vegetation type cannot be reasonably retrieved. Kenneth et al. (2000) compared linear mixture model based on calibrated atmospherically corrected hyper spectral imagery t o show i t s r el at i v e ab i l i t y 28
t o measure small differences in percent green vegetation cover in the areas of sparse vegetation in arid environments. This can be used to measure NDVI for detecting and mapping land use changes of large areas.
Metternicht and Zink (1998) reported that multi-temporal optical and microwave remote sensing can significantly contribute to detecting temporal changes of salt related surface features. Bastiaanssen (1998), summarised the IWMI review of different RS applications for water resources management, bands in the near and middle infrared spectral bands give information on soil moisture and salinity. This was in agreement with findings of Mulders, (1987); Agbu, Fehrenbacher and Jansen, (1990). Salinised and cropped areas can be identified by a salinity index based on greenness and brightness that indicate leaf moisture influenced by salinity, with classical false-colour composites of separated bands or with a computer-assisted land-surface classification (Kauth and Thomas 1976; Hardisky, Klemas and Daiber, 1983; Steven et al., 1992; Vincent et al., 1996). The brightness index detects brightness appearing at high levels of salinity. Table 2.6 shows that TM bands 5 and 7 are frequently used to detect soil salinity or drainage anomalies (Mulders and Epema, 1986; Menenti, Lorkeers and Vissers, 1986; Zuluaga, 1990; Vincent et al., 1996).
The physiological condition of a crop is shown best at TM 5 and 7; TM bands 3 and 4 are better suited to describing the overall crop development. Most of the studies in Table 2.6 are based on multispectral scanner (MSS) and TM data, this is due to the fact that the Satellite pour l'Observation de la Terre (SPOT) and the Indian Remote Sensing Satellite (IRS) have no bands greater than 1.7 mm. Joshi and Sahai (1993), found that TM, which had an accuracy of 90 percent for soil salinity mapping, was better than MSS, which was 74 % accurate. Goossens et al. (1993) compared the accuracy of TM, MSS, and SPOT and found TM to be 29
the best multispectral radiometer for soil-salinity mapping. Table 2.6 gives a summary of examples in literature of various techniques for detecting soil salinity.
Table 2.6: Examples of various techniques in the detection of soil salinity Source Chaturvedi etal., 1983 Mulders and Epema, 1986 Menenti, Lorkeers and Vissers, 1986 Everitt et al., 1988 Sharma and Bhargava, 1988 Singh and Dwivedi, 1989 Timmerman, 1989
Study Area South Dakota, USA Tunisia
Sensor PMW TM 5, 7
Methodology Brightness/temperature Digital classification
Tunisia
TM 5, 6, 7, albedo
Digital classification
Texas, USA Uttar Pradesh, India
Video imagery NISS
False-colour composite Supervised classification
Uttar Pradesh, India
NISS
Supervised classification
Qattara Depression, Egypt Gujarat, India
TM 5, 6, albedo
Brightness/temperature
PMW, C-band
Supervised classification
Uttar Pradesh, India
TM
Digital classification
Mendoza, Argentina Uttar Pradesh, India USA
TM 4, 5, 7 TM 2, 3, 4 TM 4, 5
Wiegand, Everitt and Richardson, 1992 Joshi and Sahai, 1993 Goossens et al., 1993 Casas, 1995
USA
XS
Classification False-colour composite Near/mid infrared difference Multiple regression
Saurashtra Coast, India Western Delta, Egypt Tamaulipas, Mexico
MSS, TM MSS, XS, TM XS 3, TM 5
Brena, Sanvicente and Pulido, 1995 Mirabile et al., 1995 Vincent et al., 1995
El Carrizo, Sonora, Mexico Mendoza, Argentina Gharb Plain, Morocco
TM 2, 3, 4
Vincent et al.,1996
Punjab, Pakistan
XS
Dwivedi, 1996 Vidal et al., 1996
Uttar Pradesh, India Punjab, Pakistan
MSS 1, 2, 3, 4 XS
Singh and Srivastav, 1990 Saha, Kudrat and Bhan, 1990 Zuluaga, 1990 Rao et al., 1991 Steven et al., 1992
TM 3, 5 XS 1, 2, 3
Source: Tabet (1999)
30
False-colour composite Supervised classification Brightness, supervised classification Multiple regression Kauth-Thomas index Greenness, brightness, classification tree Greenness, brightness, classification tree Principal component Greenness, brightness, classification tree
2.4.2 SRS applications Johnston and Barson (1990) reviewed Satellite Remote Sensing applications in Australia. They found that identification of saline areas was most successful during peak vegetative growth. In other periods, low fractional vegetation cover in salinised areas could not be distinguished from areas that were bare because of overgrazing, erosion or ploughing. Siderius (1991) stated that salinity is best seen at the end of irrigation or the rainy season when the plots are bare. This was at variance with Johnston and Barson (1990) findings. Venkataratnam (1983) used MSS images of pre-monsoon, post-monsoon and harvest seasons to map soil salinity in the Punjab, India. He concluded that the spectral curves of highly and moderately saline soils change considerably during the annual cycle, which significantly complicates the time-compositing procedure.
Vincent et al. (1996) in a study in Pakistan using MSS data used a tree like classification procedure. The first treatment is to mask vegetation from non-vegetation using normalised difference vegetation index (NDVI). Then the brightness index was calculated to detect moisture and salinity on fallow land and abandoned fields. This approach is suitable for locating blocks that had malfunctioning drainage networks. Different classes of soil salinity can be mapped using this approach, but its accuracy needs to be determined and compared to other available methods.
Goossens et al. (1998) as cited in Salman (2000) analysed the beginning, middle and end of the growing season in the western Nile Delta and concluded that single image may be suitable for detecting severely salinised soils but more gradations can be determined using temporal images. IDNP (2002) reviewed studies conducted by different researchers on direct observations on bare soils and indirectly by vegetation cover. In the visible part of the 31
spectrum, the soil reflectance of salt cover areas was found to be prominent. Bands in the middle infrared gave information of moisture content, which was often associated with salt content differences and some information on type of salts. The lack of vegetation or scattered vegetation and highly salt-affected surface make it possible to directly detect salt on the surface. Ground observations and radiometric measurements indicated that the main factors affecting the reflectance are the quantity and mineralogy of salt, moisture, colour and roughness. The evaluation of soil surface remains under the influence of external factors as groundwater quality and variation of depth, wetting/drying cycles and wind.
Metternicht and Zinck (1996) through their studies related to ground observation and radiometric measurement in the visible and near infrared wavelengths stated that the main factors affecting the reflectance are the quantity and mineralogy of salts together with soil moisture, soil colour and terrain roughness, which in turn are controlled by different combination of salts and the type of soil surface, texture and organic matter content. According to them, the physiological status of the crop is best manifested at TM 5 and 7, while TM bands 3 and 4 are better suited to describe the overall crop development. The multiple regression analyses between SPOT spectral data and soil morphological, physical, and chemical properties showed that many surface and some subsurface soil properties were significantly correlated. Brightness index proved to be a more useful spectral parameter if surface soil properties are to be extracted from satellite data, but the ratio of the values in red and infrared band seems to be a better technique to employ when subsurface soil properties are of interest. In general, bands in the near and the middle infrared region give reasonable information on soil moisture and salinity (Moulders, 1987). Steven et al. (1992) confirmed this finding by showing that near to middle infrared index is a better indicator for chlorosis occurring in stressed crops (normalised difference for TM bands 4 and 5). This 32
new ratio is immune to colour variations and provides an indication of leaf water potential. The spectral behaviour of salt-affected soils as compared to normal cultivated soils showed relatively higher spectral response in visible and near-infrared regions. This indicator will better assist in detecting salinity by using the indexes.
Strongly saline-sodic soils were found to have a higher spectral response as compared to moderately saline-sodic soils; also that the vegetation cover modifies the overall spectral response pattern of salt-affected soils especially in the green and red spectral bands. Spatial resolution has a significant effect on enhancing the identification of salt affected soils and crops. Steven et al. (1992) stated that based on some past researches in comparing the accuracy of TM, MSS, and SPOT they found TM to be the superior multi-spectral radiometer for soil salinity mapping and that digital classification techniques can also help in improving the identification and mapping of salt-affected soils or crops.
According to Salman (2000), excess soil moisture can cause a change in soil colour and a change in soil reflectance properties being detected by remote sensing and accumulation of organic matter. Soil colour is generally darker in poorly drained areas than in well drained soils. The visible bands in Landsat-MSS data can be used to identify this colour. Baber (1982), as cited in IDNP (2002) pointed out that colour infrared photography could indicate drainage problems by soil moisture saturation or plant stress. Shallow water tables exhibit an increase in surface moisture, which can be detected from visible reflectance and microwave emissivity.
33
2.4.3 SRS Landsat and other Sensors With the launching of the Landsat Satellite in 1972, researchers began to use satellite data for monitoring mining activities in different parts of the world (Coker, 1977). In 1984, WAPDA conducted a study on the applicability of Landsat imagery for monitoring soil salinity trends in two areas in Punjab and Sindh, India. Black and white mosaics of Landsat-MSS band 5 scale 1: 250 000 taken in March, April and December of 1984 were visually interpreted and compared with 1:250 000 surface salinity maps. Sahin et al. (2006) used three Landsat geocover dataset from 1970 to 2000 to detect temporal changes in the Zonguldak coal test field. Trinh et al. (2004) used Landsat images for studying land use dynamics and soil degradation in the Tamduong district of Vietnam. Eranani and Gabriels (2006) used Landsat data from 1976 to 2002 to detect changes in land cover in the Yazd-Ardakan basin, Iran. Latifovic and Fytas (2005) analyzed the land cover change of the Oil Sands Mining Development in Athabasca, Canada using information extraction method applied to two Landsat scenes. For all the above studies, their findings showed that Landsat MSS can be used easily for the identification of broad land cover changes where there are bare and cultivated lands or used lands and also extended for salinity detection and mapping.
Tabet (1995), Vidal et al. (1998) and Tabet (1999) conducted studies on vegetation and brightness indices derived from SPOT-XS. The indices were used to classify salinity for vegetative and non-vegetative areas. The resulting classification allowed for the identification of highly saline and non-saline areas, but areas with low to medium salinity levels were difficult to distinguish. Cialella et al. (1997) used the combined GIS/RS approach for predicting soil drainage classes. They studied soil drainage by means of a classification-tree analysis using airborne NDVI data, digital elevation data and soil types. Though the study did not focus on soil salinity, it can be modified for salinity detection based on the use of NDVI. 34
Verma et al. (1994) combined the TM false-colour composite (FCC) bands 2, 3 and 4 with thermal data at 10.4 to 12.5 mm to solve the problem of spectral similarity where the dullwhite tone of salt-affected and sandy soils has been difficult to distinguish. They found the data between March and the first week of April significantly better because of maximum contrast. They classified salt-affected soils in Etah, Aligarh, Mainpuri and Mathura districts of India into S1: 75 %, using the integrated approach to image interpretation. This FCC approach can also be integrated with other earlier discussed for better analysis and classification.
Dwivedi and Rao (1992) adopted a quantitative approach to identify the most appropriate three-band combination of Landsat TM reflective-band data for identifying salt-affected soils. They used the standard deviation and correlation coefficient values of TM data to compute a statistical parameter called the optimum index factor (OIF), an indication of the variance of the data. Of the 20 possible 3-band combinations, the combination of 1, 3, and 5 was found to be the best in terms of information content. The validation of the results revealed a mixed relationship between rankings obtained from OIF values and estimated accurately. Dwivedi and Sreenivas (1998) demonstrated the potential of image transformations such as principalcomponent analysis (PCA), rationing, and image differencing to detect changes in the extent and distribution of salt-affected soils, using Landsat MSS data for 1975 and 1992 to study the alluvial plains of Uttar Pradesh. Results indicated that the third principal component, image differencing and rationing of the first two bands provided substantial information about behaviour of salt-affected soils over time in the two periods. Metternicht and Zinck (1997) applied a synergistic approach to map salt-affected surfaces, combining digital image classification with field observation of soil-degradation features and laboratory determinations. Landsat TM bands1, 2, 4, 5, 6 and 7 were combined to obtain the 35
highest separability between salt- and sodium-affected soils. Overall accuracy was 64 percent; for some soils 100 percent accuracy was obtained. The overall accuracy can be improved on if the TM bands were better selected based on supervised classification information obtained during the field survey. Data collected by the MSS and TM will continue to be used as a historical global database (Stefanov et al., 2001). Newer sensors with either greater spectral/temporal resolution such as the Advanced Space borne Thermal Emission and Reflection Radiometer (ASTER) will allow even more precise land cover classification.
2.4.4 Image pre-processing and corrections Landsat images have been widely used for land cover mapping and the creation of vegetation inventories at different spatial scale information on the Earth’s surface characteristics (Bossard et al., 2000; Cohen and Goward, 2004). Nevertheless, there exist limitations in using Landsat data for multitemporal studies because of problems in obtaining homogeneous time series. Efforts have been made in the past to reduce non-surface noise in Landsat images and also to calibrate the sensor to correct radiometric trends (Teillet et al., 2004; de Vries et al., 2007) and also reduce the influence of topography (Gu and Gillespie, 1998; Pons and Solé, 1994). According to Schroeder et al. (2006) other studies have shown that the application of accurate sensor calibrations and complex atmospheric corrections does not guarantee the multitemporal homogeneity of Landsat datasets because complete atmospheric properties are difficult to quantify and simplifications are commonly assumed. There are many protocols proposed in pre-processing multitemporal Landsat data sets according to Han et al., (2007), these protocols comprise of the following steps:
36
1. Geometric correction, 2. Calibration of the satellite signal to obtain “Top of the Atmosphere Radiance”, 3. Atmospheric correction to estimate surface reflectance, 4. Topographic correction, and 5. Relative radiometric normalization between images obtained on different dates.
2.4.4.1 Radiometric correction Radiometric processing is always recommended to be done prior to geometric processing since this resampling step generally smoothness the data set (Paolini et al., 2006). Some studies have analyzed the role of complete radiometric correction protocols in processing multitemporal Landsat data when a number of different vegetation processes are of interest. Their results found out that land classification and forest succession serve as a function of the radiometric correction applied (Norjamäki and Tokola, 2007; Schroeder et al., 2006). Radiometric correction and geometric correction processes are required to obtain accurate time series of Landsat imagery.
2.4.4.2 Geometric correction The objective of geometric corrections is to compensate for the distortions and degradations caused by the errors due to the variation in altitude, velocity of the sensor platform, variation in scan speed and in the sweep of the sensor's field of view, earth curvature and relief displacement.
37
2.5 Visual interpretations using photo imagery IDNP (2002), reviewed a study carried out for the entire Indian territory (329 million ha) using Landsat - MSS FCC of 1:1, 000, 000 scale and categorised wasteland as salt-affected, gullies, waterlogged or marshy, undulating upland with or without scrub, forest blank, sandy areas (coastal or desert), barren hill ridge or rock outcrops and snow covered/glacial areas. The interpretation technique was supported by intensive ground data and geographical knowledge of the area. An accuracy of 80 to 90 % was achieved in the identification and mapping of wastelands when compared with the ground survey. Rao and Venkataratnam (1991), used Landsat TM standard FCC and delineated strongly sodic soils as bright white patches with fine texture, and moderately sodic soils as a dull white to strong brown. This technique can effectively be used in the classification of salinity in various grades.
Aerial photographs and Landsat TM data could be used to monitor changes in the status of salt- affected soils. According to Kalra and Joshi (1997), Landsat (MSS and TM), SPOT and IRS (LISS-I & 11), FCC images during fallow period April-May, January/February crop and rain fed crop (October) were used and evaluated the capability of multi-sensor data for delineating salt-affected soils in arid Rajasthan. It was concluded that the moderately and severely salt-affected soils could be mapped from any season’s FCC of Landsat, SPOT and IRS.
2.5.1 Digital analysis Remote sensing investigation on soil salinity can be broadly divided into the delineation of salt-affected soils under the bare condition and cropped condition. Salinised and cropped areas can be identified with a salinity index based on greenness and brightness that describes 38
leaf moisture as influenced by salinity, with classical false colour composites of separated bands, or with a computer assisted land surface classification (Steven et al., 1992; Vincent et al., 1996). Basically, a brightness index is meant to detect high levels of brightness appearing at high levels of salinity. The unique patterns of geomorphologic shapes are thought to be helpful in discriminating the salinisation process from a physiographic perspective.
In a review by Salman (2000), he stated that the application of remote sensing in contextual classifier for soil salinity mapping with a built GIS to link the location of the irrigation feeders and drainage master canals in the western Nile Delta with digital elevation data and satellite classifications is possible. He stated that, soil salinity risks are considered to be proportional to the distance of the field from the main irrigation canals, as well as to the field elevation difference with the main irrigation canals. TM bands 2, 3, 4, 5, 6 and 7 were used to classify three different stages of water logging according to a simple supervised procedure.
2.5.2 Digital analysis using surface vegetation index Salinity can be detected through its impact on the vegetation. A vegetation index is a common spectral index that identifies the presence of chlorophyll. A number of vegetation indices have been proposed. In a study by Richardson et al. 1976, an inverse relationship was observed between reflectance and salinity, as salt content induces less plant cover (decreasing of density, LAI (leaf area index), and height) and sometimes slight salt deposition on surface associated with vegetation have a similar reflectance as that of the normal cropped area. Salt tolerant plants are good references of salinity level of salt marshes but require good calibration. Contrasted associations of vegetation and bare soils can be more useful for salinity detection than individual surface types. RS information can be improved when it integrates with other data, for which GIS is an appropriate tool. 39
2.5.3 Geographical Information System (GIS) A GIS is commonly defined as a computer system for the input, editing, storage, maintenance, management, retrieval, analysis, synthesis and output of geographically referenced or spatial information (ESRI, 1998). GIS is a tool to analyse and interpret the remotely sensed data. It can also be used for the analysis and interpretation of physically collected data. Arora and Goyal (2003) highlighted the use of geographical information system (GIS) in the development of the conceptual groundwater model. Various layers of information such as canal network recharge zones, subsurface geology and digital terrain model (DTM) of Hanumangarh and Sriganganagar districts of India were developed in GIS and were then transferred to the finite difference grid for developing a mathematical groundwater flow model of the area.
2.5.4 GIS in soil salinity modeling A model is used to represent a reality. Models can help understand, describe, or predict how things work in the real world. Generally, there are two types of models; those that represent the objects in the landscape (representation models); and those that attempt to simulate processes in the landscape (process models). Representation models try to describe the objects in the landscape, such as buildings, streams, or forest, while process models attempt to describe the interaction of the objects that are modelled in the representation model. Different types of process models include suitability modeling, distance modeling and hydrological modeling.
40
2.5.5 Soil Salinity analysis and modelling 2.5.5.1 Multi-criteria decision evaluation Decision Theory is concerned with the logic by which one arrives at a choice between alternatives (Saaty, 1977). The recommended alternatives vary from problem to problem. They might be alternative actions, alternative hypotheses about a phenomenon, alternative objects to include in a set and so on. Resource allocation decisions are also prime candidates for analysis with a GIS. Indeed, land evaluation and allocation are the most fundamental activities of resource development (FAO, 1976). To meet a specific objective, several criteria are to be evaluated. Such a procedure is called Multi-Criteria Evaluation (Voogd, 1983; Carver, 1991). Multi-criteria evaluation (MCE) with respect to salinity is most commonly achieved by one of two procedures: A pairwise comparison method has been used for the development of weights of the factors in the salt affected soil analysis. Here, breaking the information down into simple pairwise comparisons in which only two criteria are considered at a time can greatly facilitate the weighting process, and will likely produce a more robust set of criteria weights. The technique being implemented in IDRISI software is that of pairwise comparisons developed by Saaty (1997) in the context of a decision making process known as the Analytical Hierarchy Process (AHP). The first introduction of this technique to a GIS application was that of Rao et al. (1997), the procedure was developed outside the GIS software using a variety of analytical resources.
Purevdorj et al. (1998), listed the current available methods to assess and model land vegetation cover and biomass from remotely sensed data into three basic methods; spectral mixture models, calibrated cover-radiance relationships, and vegetation indices approaches. The combination of these methods will produce a better result. 41
2.5.5.2 Spectral mixture modeling Mixture model includes two types of modeling, linear and nonlinear (Schowengerdt, 1997). The linear mixture model assumes that each field within a ground pixel contributes an amount characteristic of the cover type in that field to the signal received at the satellite sensor and is proportional to the area of the cover type. The main problem in the usage of linear mixture modelling is with the location of pure end member for the green cover component, because vegetation density in grassland is relatively low (Purevdorj et al., 1998). According to Schowengerdt (1997), nonlinear mixing occurs when radiation transmission through one material and second reflectance occur from other materials, or there are multiple reflections within or between materials.
2.5.5.3 Cover-radiance relationships The Cover-radiance relationship approach investigates the relationship between field collected canopy cover data and radiance data. They are best suited to medium spatial resolution satellite sensor data, such as Landsat TM, MSS, and SPOT since they require accurate measurement of vegetation cover on the ground covering the same area. This method has been used in many earlier studies (Ferro 1998, Lewis 1994, Schmidt and Karnieli 2002, Todd and Hoffer 1998, Wang et al. 2002). Problems with comparing satellite data and ground measurement include the accuracy of estimating a large area and the efficiency of the model for describing the canopy condition (Purevdorj et al., 1998).
2. 5. 5. 4 Vegetation indices Soil and green vegetation have different modes of reflectance characteristics. The mixture of soil, green vegetation, and shade in the pixels make remote sensing of land cover a challenge 42
(Perry and Lautenschlager, 1984; Todd and Hoffer, 1998). According to Jenson (2006), vegetation indices can be used to minimize the impacts of soil background and biological aging materials. Red and near infrared have been found to be good at detecting green vegetation (Purevdorj et al. 1998). Therefore most vegetation indices make use of the red and near infrared portion of spectral reflectance. The selection and suitability of a vegetation index are generally determined by its sensitivity to the characteristics of interest (Gao et al., 2000; Perry and Lautenschlager, 1984; Schmidt and Karnieli, 2002; Weiser et al., 1986).
Many efforts have been made to optimize vegetation indices that make them insensitive to variations in sub-surface-sensor geometries, atmosphere, calibration, and canopy background (Gao, 2000). Frequently used vegetation indices include Simple Ratio (SR), Normalised Difference Salinity Index (NDSI), Normalised Difference Vegetation Index (NDVI), SoilAdjusted Vegetation Index (SAVI), Enhanced Vegetation Index (EVI), Green Vegetation Index (GVI), and Transformed Soil Adjusted Vegetation Index (TSAVI) (Chen, 1996; Todd and Hoffer, 1998). These results are not complete without the support of field collected information.
2.5.5.5 Wireless sensor networks The wireless sensor networks are a new technology for collecting data about the natural or built environment. They consist of low cost embedded sensory and computational devices (audio and video which detect bio-chemicals in novel ways) with wireless capability forming ad hoc networks. The network provides information on an unprecedented temporal and spatial scale with the aim to reduce the cost of data capture by a factor of 1 000 and increase the spatial and temporal resolution by a factor of 1 000 within ten years. The sensor can allow
43
greater understanding, accurate monitoring and modeling, better prediction, planning and control sustainable management (CSIRO, 2009).
Other mapping techniques to perform change detection with satellite imagery have become numerous as a result of increasing versatility in manipulating digital data and increasing computer power. A wide variety of digital change detection techniques have been developed over the last two decades. Singh (1989), and Coppin and Bauer (1996), summarize eleven different change detection algorithms that were found to be documented in the literature by 1995. These include: 1. Mono-temporal change delineation; 2. Delta or post classification comparisons; 3. Multidimensional temporal feature space analysis; 4. Composite analysis; 5. Image differencing; 6. Multi-temporal linear data transformation; 7. Change vector analysis; 8. Image regression; 9. Multi-temporal biomass index; 10. Background subtraction; and 11. Image ratioing’
2.5.5.6 Hybrid method: integration of SRS data, field data and GIS tool There are close relationships between field data and spectral data (Weiser et al., 1986; Lewis, 1994). It is reasonable to hypothesize that there are high correlations between variation of 44
field collected variables of vegetation canopy and spectral variation. There have been a few works that utilize remotely collected data and field data (Jorgensen and Nohr, 1996; Lauver, 1997; Gould, 2000; Zhang et al., 2005). The new generation of satellites now being launched carry instruments with much greater spatial resolution (as high as 0.6m to 1m); sufficient to see many more features of interest in mapping than it was before (Akinyede and Boroffice, 2004).
At present, the identification and mapping of saline soil are a combination of the following: 1. Visual interpretation of photographs; 2. Digital analysis of false colour composite (FCC); and, 3. Digital analysis of surface radiation and vegetation index. All methods require ground truth information for calibration and validation. The actual use depends on the specific aim of the survey, data availability, human skill and availability of time and money.
2.5.5.7 GIS tool: IDRISI software The IDRISI software is an integrated feature rich GIS and image processing software system for the analysis and display of spatial data. IDRISI (Taiga edition) was released in February, 2009 and has a range of tools for GIS analysis, natural resource monitoring and satellite image processing. Others include land change and time series analysis, modeling, baseline land resource mapping, decision support and uncertainty management. It also provides a full suite of mathematical and relational modeling tools that allow models to be entered as equations, with map layers as variables (Clark Labs, 2009).
45
IDRISI includes the largest suite of supervised and unsupervised classification techniques in the industry, based on scientifically proven algorithms and methods, for both multispectral and hyper spectral imagery. The IDRISI software includes all of the general purpose and advanced processing tools required to prepare satellite imagery at an extremely affordable cost. Although IDRISI provides an extensive suite of image processing tools, what makes the software critical for this research is that image processing data can be completely integrated with IDRISI's equally extensive set of raster GIS tools, saving effort, costs and resources. It provides Land Change Modeler, an automated application for the monitoring and prediction of land cover change (Clark Labs, 2009).
2.6 Modeling the potential for change Change is modelled empirically by using past changes to develop a mathematical model and GIS data layer expression of transitional potential. Transitions can be grouped into a set of sub-models and the potential power of explanatory variables can be explored. Variables can be added to the model as either static or dynamic components. Once model variables have been selected, each transition is modelled using either a Multilayer Perception Neural Network or Logistic Regression. The result for either model is a potential map for each transition an expression of time-specific potential for change.
2.6.1 Predicting changes / modeling According to Clark Lab (2009), after specifying the end date, the quantity of change in each transition can be modelled through a Markov Chain analysis model. Two basic models of change are provided. The soft prediction model yields a map of vulnerability to change for the selected set of transitions. The soft prediction model is generally preferred for habitat and 46
biodiversity assessment because it provides a comprehensive assessment of change potential. The hard prediction model is based on a multi-objective land competition model. The hard prediction yields only a single realisation out of many possible realisations.
Land Change Modeler allows for the input of dynamic variables as well as planning interventions in the change prediction set up. A validation tool is included to assess the quality of the prediction map in relation to a map of reality (the validation tool is only available in the LCM implementation within IDRISI) Wu (2006).
2.6.2 Markov chains and transition probability matrix Markov chains have been widely used to model land use changes including both urban and nonurban areas at large spatial scales (Jahan, 1986; Muller and Middleton, 1994; Wu, 2006). A time-stationary Markov chain is completely determined by the Markov transition matrix P. An aggregate, macroscopic, stochastic, modeling process, a technique for predicting change modeling, predictions of future change are based on changes that have occurred in the past. A transition matrix contains the probability that each land cover category will change to every other category. A transition area matrix contains the number of pixels that are expected to change from each land cover type to each other land cover type over the specified number of time units, while a set of conditional probability images report the probability that each land cover type would be found at each pixel after the specified number of time units (Muller and Middleton, 1994; Wu, 2006). The Markov Transitional Probability Matrix can be used to forecast the future trend in soil salinity and land use-land cover using Markov chain analysis as proved by the studies. The Markov transition probability matrix tends to show the probability that a given cover type will change to another cover type over the time period under consideration. 47
2.7 Statistical analysis 2.7.1 Student's t-test Student's t-test is any statistical hypothesis test in which the test statistic follows a Student's t distribution if the null hypothesis is supported (Fisher Box, 1987). When the scaling term is unknown and is replaced by an estimate based on the data, the test statistic (under certain conditions) follows a Student's t distribution (Fadem, 2008). The most commonly used t-tests are: 1.
A one-sample location test of whether the mean of a normally distributed population has a value specified in a null hypothesis.
2.
A two sample location test of the null hypothesis that the means of two normally distributed populations are equal. All such tests are usually called Student's t-tests, though strictly speaking that name should only be used if the variances of the two populations are also assumed to be equal; the form of the test used when this assumption is dropped is sometimes called Welch's t-test. These tests are often referred to as "unpaired" or "independent samples" ttests, as they are typically applied when the statistical units underlying the two samples being compared are non-overlapping (Fisher Box, 1987).
3.
A test of the null hypothesis that the difference between two responses measured at the same statistical unit has a mean value of zero. For example, suppose we measure the size of a cancer patient's tumor before and after a treatment. If the treatment is effective, we expect the tumor size for many of the patients to be smaller following the treatment. This is often referred to as the "paired" or "repeated measures" t-test (Fisher Box, 1987; Raju, 2005). 48
4.
A test of whether the slope of a regression line differs significantly from 0. Most t-test statistics have the form t = Z/s, where Z and s are functions of the data. Z is designed to be sensitive to the alternative hypothesis whereas s is a scaling parameter that allows the distribution of t to be determined.
As an example, in the one-sample t-test in equation 1: Z=
,
where
..............................................................1
is the sample mean of the data, is the sample size, and is the population standard deviation of the data;
s in the one-sample t-test is
,
where is the sample standard deviation. The assumptions underlying a t-test are that: 1.
Z follows a standard normal distribution under the null hypothesis
2.
s2 follows a χ2 distribution with p degrees of freedom under the null hypothesis, where p is a positive constant
3.
Z and s are independent.
Two-sample t-tests for a difference in mean involve independent samples, paired samples and overlapping samples. Paired t-tests are a form of blocking, and have greater power than unpaired tests when the paired units are similar with respect to "noise factors" that are independent of membership in the two groups being compared (Raju, 2005). In a different context, paired t-tests can be used to reduce the effects of confounding factors in an observational study. Others are independent samples and one-sample t-test.
49
The independent samples t-test is used when two separate sets of independent and identically distributed samples are obtained, one from each of the two populations being compared. In testing for one-sample t-test, the null hypothesis that the population mean is equal to a specified value μ0, one uses the statistic described in equation 2:
.......................................................................2 Where, is the sample mean, s is the sample standard deviation of the sample and n is the sample size. The degrees of freedom used in this test are on − 1. Dependent t-test for paired samples is used when the samples are dependent; that is, when there is only one sample that has been tested twice (repeated measures) or when there are two samples that have been matched or "paired". This is an example of a paired difference test also descibed in equation 3 (Fadem, 2008).
............................................................3 For equation 3, the differences between all pairs are calculated. The pairs are either one person's pre-test and post-test scores or between pairs of persons matched into meaningful groups (for instance drawn from the same family or age group: see table). The average (XD) and standard deviation (sD) of those differences are used in the equation. The constant μ0 is non-zero if you want to test whether the average of the difference is significantly different from μ0. The degree of freedom used is n − 1.
50
Two-Sample t-Test for Equal Means:
The two-sample t-test for unpaired data is defined by equations 4, 5 and 6: H0: Ha: Test Statistic:
...........................4 where N1 and N2 are the sample sizes, means, and
and
and
are the sample
are the sample variances.
If equal variances are assumed, then the formula reduces to:
...........................5 where
................6
2.7.2 Pearson's chi-squared test Pearson's chi-squared test (χ2) is used to assess two types of comparison: tests of goodness of fit and tests of independence (Greenwood and Nikulin, 1996). Pearson's chi-squared test is the best-known of several chi-squared tests such as Yates, likelihood ratio, portmanteau test in time series and others which are statistical procedures whose results are evaluated by reference to the chi-squared distribution. Its properties were first investigated by Karl Pearson
51
in 1900 (Plackett, 1983). Pearson's chi-squared test is used to assess two types of comparison: tests of goodness of fit and tests of independence. 1.
A test of goodness of fit establishes whether or not an observed frequency distribution differs from a theoretical distribution.
2.
A test of independence assesses whether paired observations on two variables, expressed in a contingency table, are independent of each other.
The first step is to calculate the chi-squared test statistic, X2, which resembles a normalized sum of squared deviations between observed and theoretical frequencies. The second step is to determine the degrees of freedom,
, of that statistic, which is essentially the number of
frequencies reduced by the number of parameters of the fitted distribution. In the third step, X2 is compared to the critical value of no significance from the
distribution, which in
many cases gives a good approximation of the distribution of X2. Calculating the teststatistic, the value of the test-statistic is described by equation 7:
.......................................................7 Where, = Pearson's cumulative test statistic, which asymptotically approaches a
distribution.
= an observed frequency; = an expected (theoretical) frequency, asserted by the null hypothesis; = the number of cells in the table.
The chi-squared statistic can then be used to calculate a p-value by comparing the value of the statistic to a chi-squared distribution. The number of degrees of freedom is equal to the number of cells, minus the reduction in degrees of freedom. The result about the number of 52
degrees of freedom is valid when the original data were multinomial and hence the estimated parameters are efficient for minimizing the chi-squared statistic. More generally however, when maximum likelihood estimation does not coincide with minimum chi-squared estimation, the distribution will lie somewhere between a chi-squared distribution with and
degrees of freedom (Chernoff and Lehmann, 1954).
For test of independence, an observation consists of the values of two outcomes and the null hypothesis is that the occurrence of these outcomes is statistically independent. Each observation is allocated to one cell of a two-dimensional array of cells (called a table) according to the values of the two outcomes. If there are r rows and c columns in the table, the "theoretical frequency" for a cell, given the hypothesis of independence is given by equations 8 and 9:
......................................8 where N is the total sample size (the sum of all cells in the table). The value of the teststatistic is
.......................................................9 Fitting the model of independence reduces the number of degrees of freedom by p = r + c − 1. The number of degrees of freedom is equal to the number of cells rc, minus the reduction in degrees of freedom, p, which reduces to (r − 1)(c − 1). For the test of independence, also known as the test of homogeneity, a chi-squared probability of less than or equal to 0.05 (or the chi-squared statistic being at or larger than the 0.05 critical points) is commonly interpreted by applying workers as justification for rejecting the null hypothesis that the row variable is independent of the column variable (Plackett, 1983). 53
CHAPTER THREE 3.0 STUDY AREA 3.1 Location of the study area Vaal Harts irrigation scheme (VHS) located in the east of the Fhaap Plateau in the Northern Cape and North West province border covering areas from Jan Kempdrop in the south to Taung (the Dry Harts River) in the north in South Africa, served as the study area. VHS covers about 36, 950 hectares of irrigated land and are one of the largest irrigated areas under irrigation in Africa. Water is provided to some 680 farmers. The scheme is supplied with water abstracted from the Vaal River at the Vaal Harts weir about 8 km upstream of Warrenton. A canal is used to convey the water to the scheme (Grove, 2006). Figures 3.1a and 3.1b presents the position of the study area (VHS) in relation to South Africa.
Figure 3.1a: Location of the study area (VHS) in relation to South Africa Source: adapted from van Vuuren, et al (2004). 54
Figure 3.1b: Location of the study area (VHS) in relation to South Africa Source: adopted from van Vuuren, et al (2004).
3.2 Surface run-off, rainfall, temperature and evapotranspiration VHS is located in a summer rainfall area of South Africa. This area experiences low, seasonal and irregular rainfall with an average rainfall of 442 mm per year (Jager, 1994). The average precipitation in the summer months, October to February varies between 9.1 and 9.6 mm/day while in July precipitation is only 3.6 mm/day. The rainy season in the area is usually from October to March. In the winter months, almost no rainfall occurs. The average annual rainfall in Jan Kempdrop and in Taung weather stations (close to the study area) is 477 mm and 450 mm respectively (AGIS, 2009). The average temperature of the spring and summer months is above 30°C and which is the highest in the month of February. The median annual
55
simulated runoff in the area is in the range of 20 to 41 mm, with the lowest 10-year recording at 4.8 to 9.3 mm (Schmidt et al., 1987).
Evapotranspiration due to the application of irrigation water, rainfall and plant growth in these summer months are high. The importance and relevance of evapotranspiration refer to the calculation of water use by a plant during a given season. According to Fritz (2009) the average evapotranspiration for the crops planted in the study area was 1, 030 mm. Taking the growth cycle of the 2 most planted crops (maize and wheat) into consideration, this figure can be reworked to 774 mm for the research period. The nearest meteorological station in the study area is at Jan kempdrop which is about 600 Km from Pretoria. Table 3.1 presents the monthly and annual meteorological data of Vaal Harts obtained from South Africa Weather Services.
Table 3.1: Meteorological data of Vaal Harts Total for the year 621.1
Elem Rainfall (mm) T* (°C)
Start
End
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
01\01\1997
25\10\2010
116.6
84.5
56.2
53.3
76.2
42.04
3.4
21.7
12.9
40.3
54.3
59.7
01\01\1997
25\10\2010
22.4
22.7
21.2
17.2
12.8
9.8
9.7
12.4
16.2
19.8
21.6
23.5
RH (%)
01\01\1999
25\10\2010
61.0
64.5
66.2
66.1
60.1
59.9
53.9
48.5
44.6
51.4
51.8
55.6
683.6
Tx (°C)
01\01\1997
25\10\2010
32.2
32.0
30.1
27.0
23.4
21.7
20.4
23.2
26.8
29.2
30.6
32.1
328.9
Tn (°C) RHx (%) RHn (%) ET0 (mm)
01\01\1997
25\10\2010
16.9
17.4
14.8
24.4
18.2
6.9
0.2
2.0
5.7
10.6
12.6
15.1
144.8 1051.7
01\01\1998
25\10\2010
90.9
92.5
92.9
92.7
90.0
87.9
85.4
81.3
77.9
84.8
86.9
88.7
01\01\1998
25\10\2010
29.4
30.1
33.0
43.2
37.1
30.3
22.7
20.0
18.3
22.5
22.5
26.2
13\06\1999
25\10\2010
5.4
4.5
4.2
3.3
3.0
2.2
2.4
3.2
4.2
4.8
5.6
5.8
335.1 48.6
Source: South African Weather Service (2009)
56
3.3 VHS Geomorphology and soil material The VHS area is situated in a glacial valley which is drained by the Harts River. The altitude ranges from 1 050 to 1 175 m a.m.s.l. (Above mean sea level), changing towards the west (AGIS, 2009). The VHS is averagely flat, as 70% of the area comprises slopes of less than 1%. The lithostratigraphy classification of the area was named as the Bothaville Formation, Rietgat Subformation. The Rietgat Formation in the Taung-Jan Kempdrop area is known as the Phokwane Formation of the Harts water Group. The Phokwane Formation mainly consists of porphyrite lava, volcanic tuffs, tuffaceous sediments and chert (Schutte, 1994).
According to Hough and Rudolph (2003), the soil in the area comprises of alluvial and Kalahari Sand. The soil layer is more than 3 m deep in the area. The soil types found in the study area are: Hutton, Kimberley, Hutton/Mispah, Dundee and Katspruit/Kroonstad (Barnard, 2008). On the average, soil mainly consists of 75 % sand, 10 % silt and 15 % clay.
3.4 Infrastructure and irrigation methods Irrigation water is relayed to the plots on the Vaal Harts and Taung Irrigation Schemes through an extensive network of open channels, siphons and pipes. The main canal is 18.4 km long. It splits into the northern canal, which is 82 km long and serves 33 400 ha, and the western canal, which is 22 km long, serving 4 800 ha. The water reaches the plots by means of a feeder (45 km) and tertiary (580 km) canals. There are 5 balancing dams on the scheme. Farmers also make use of overnight storages to enable them irrigate when the canal is dry and to assist with scheduling. The average size of an overnight storage is 3 600 m3, while the capacity of the canal in VHS is 4mm/day. The water quota for the North and West canal is 9, 140 m3 per ha/annum. The total water use charge is 8.77 cents per cubic meter of water 57
which includes a charge of 8.24 cents for irrigation water use, a catchment management charge of 0.5 cents per cubic meter and a water research charge of 0.03 cents per cubic meter of water (Grove, 2006). Figure 3.2 shows the canal details.
Figure: 3. 2: Canal detail. Source: van Vuuren, et al (2004).
58
3.5 Crop types According to Grove (2006), the most common cash crops grown in the area are barley/wheat, maize, cotton, grapes, potatoes, oats and groundnuts. Other tree crops grown are lucerne, pecan nuts, peanuts, grapes, citrus, and olives. A wide variety of fruits, nuts and crops are planted in the area throughout the year. Peanuts, citrus and olives as cash crops are exported to the USA, Europe and Japan. Some of the VHS farmers still make use of flood irrigation, while many farmers have changed to other practices like centre irrigation pivots and drip, due to the greater effectiveness and efficiency of these systems. Figures 3.3 to 3.5 present an aerial overview and main canal of VHS.
Figure: 3.3: Aerial overview of VHS (Source: Aurecongroup, 2010)
59
Figure 3.4: VHS main canal (Source: Field survey)
Figure: 3.5: VHS main canal (Source: Field survey) 60
3.6 Irrigation development in VHS In 1934, Act No. 38 of 1938 was approved, providing permission to construct the Vaal Dam and develop the Vaal harts Irrigation Scheme. The area of study stretches from Jan Kempdrop in the south to Taung (the Dry Harts River) in the north. The first farmers received their plots in 1938. Today there are 1, 200 plots that vary in size from 25 to 75 ha and cover a total area of 35, 302 ha, which include 31, 732 ha in the Northern Cape and 3, 570 ha in the North-West Province.
Water logging and salinisation problems have been experienced in the area. To remedy the problem, a main subsurface drainage system was installed in 1972; the feeder canals were also lined with concrete. In 2000 approximately 50% of the plots did not have proper discharge points for the drained water, although ± 80% do have internal subsurface drains (Van Niekerk, 2009). Several studies have been conducted in the Vaal harts Irrigation Scheme area to determine the influence of the irrigation on groundwater. The study by Herold and Bailey (1996) claimed that salts are accumulating in the groundwater sources below the area by leaching through the upper soils. The study also indicated the possibility of a salt sink, mainly due to a perched water table in the area. A study by Ellington et al. (2004), on the other hand, indicated that water levels do not differ more than a few centimetres in deep and shallow water systems. The water quality profiled in piezometers indicated no major stratification of groundwater, and the deep aquifer does not perform separately. If the net storage of the aquifer remains the same, the TDS increase will be in the order of 14 mg/l per annum.
The irrigation water is the greatest contributor to the groundwater system and contributes much more to the salt load than fertilisers. According to Streutker et al. (1981), flood 61
irrigation commenced at the Vaal harts Irrigation Scheme of 30, 000 ha established in 1940 and soon after, the groundwater table rose by between 0.9 and 1.5 m over the whole area. In 1956 several salinisation cases were reported and in the 1960s, a number of soil profiles from all over the Scheme, contained more salts in the subsurface than measured during the initial soil survey, indicating an increasing tendency, although not alarming at that stage. In the 1960s, after 35 years of VHS, the fine sandy soils of the scheme were severely salinised, which led to a reclamation project of some 30, 000 ha saline or saline-sodic soils (depth 0.3 m) between 1977 and 1980 (Streutker et al.,1981).
According to Du Plessis (1986) salt-affected soils resulted in 1.4 - 2.1 million South African rand losses in gross income for the irrigation scheme farmers. The installation of 218 drainage systems totalling 500 km of subsurface lateral drains at a cost of 2 million rand were undertaken between 1975 and 1977. Although the subsurface drains were successful in keeping the groundwater table below 0.7 m and leaching salts from recent salinised patches, there were still some 1, 500 ha of saline soils by the end of 1977 and additional 1 000 ha in 1980. Bearing in mind that an estimated 100, 000 ha of the 885, 000 ha of the land of State Irrigation Schemes were salt-affected by 1976 (Du Plessis, 1986), the cost of reclamation is justified only if maintenance of the water table does not require a considerable amount of over-irrigation and does not cause pollution down-stream which requires a high level of irrigation and drainage management. A further 2 million rand then had to be invested to install sub-surface drains on farms and to link these drains to the partially developed system of open storm water drains, in an attempt to lower the water table and to leach salts (Streutker et al., 1981).
62
CHAPTER FOUR 4.0 MATERIAL AND METHODS 4.1 Selection of study area The Vaal harts irrigation scheme (VHS) with headquarter located in Jan Kempdrop, Northern Cape province of South Africa w as selected for the study. VHS covers an area of about 37,000 hectares of irrigated land and the scheme is an area affected with m an y environm ent al problem s such as soil salinity as seen during the baseline survey. A baseline survey was conducted on the scheme in April, 2009. During the baseline survey the problem of irrigation induced salinity was overwhelming and thus much of the land has been left uncultivated due to this menace. At the time of this study an independent consultant was hired to study the problem of salinity in the scheme with a view to suggest ways of combating it. Therefore, this study will assist in combating the menace of irrigation induced salinity.
4.2 Analysis of climatic data Temperature is a measure of the quantity of heat energy possessed by a body or medium as a result of heat transfer. Air temperature, a determinant of agricultural production since it influences the rate of evapotranspiration which is a significant component of the hydrologic cycle. This analysis was carried out in order to understand the effect of air temperature as a key environmental factor.
The temperature, rainfall and relative humidity data collected for the year 1997 to 2010 were analysed. For each of these years; the values of the variables for each month of the year were computed. Basic statistics of mean, variance, standard deviation and coefficient of variation 63
including Pearson's correlation coefficient and the autocorrelation coefficient were determined from the data. Pearson's correlation coefficient is a measure of the correlation (linear dependence) between two variables X and Y, giving a value between +1 and −1 inclusive to give the covariance of the two variables divided by the product of their standard deviations. It is used to measure the strength of linear dependence between two variables (Rodgers and Nicewander, 1988; Stigler, 1989). For a sample of size n, the n raw scores Xi, Yi are converted to ranks xi, yi and ρ is computed from these:
i xi x yi y
i xi x yi y 2
2
...........................................................................................(1)
Where ρ is the Pearson's correlation coefficient,
xi and x are variable 1 and its mean, while yi and y are variable 2 and its mean respectively
4.3 Field data collection In VHS, salinity problem had developed over the years due to excessive irrigation and poor physical condition of the soil. Information about the soil samples collected and their locations were recorded using Garmin handheld GPS; the readings were later downloaded into ARCGIS for further analysis.
A topographic/base map (scale: 1: 150,000) of Vaal harts of Latitude E 24.7178°-27 2612 º S and Longtitude S 25. 2765°-28.0180° E (path 172, row 079) was obtained from the Department of Water Affairs was used. Soil map, soil data and climatic data for the area were obtained from the Agricultural Research Council (ARC) and the South African Weather Service, Pretoria office. The flood line data was acquired from Department of Agriculture (DOA) office in Jan Kempdrop in map form and was converted to JPEG and hence in GIS environment. Information about soil types, water and soil quality, agricultural practices, crop 64
yield, irrigation facilities and necessary ground truth data of the study area were obtained through baseline survey and structured questionnaires administered among 40 farmers selected randomly and a project manager. Table 4.1 served as a guide to the data collection.
Table 4.1: Checklists for collation of field data a. Location of project area - Locate nearest town - Latitude and Longitude - Locate the field office
b. Climatic data (past/present)
- Check on monthly/annual basis -Rainfall - Temperature (max, min) - Relative humidity etc.
c. Soil status (collects all available soil data from inception to date
-Check for past soil survey report - Data on soil classes on scheme - Number of groups - Map of the scheme - Chemical status of each group including pH, Ca2+, Na+ and Mg2+
d. Water regime i. Surface from the intake
- Sources of water for irrigation
ii. Drainage water
- Data on quality of each source pH, Na+ SAR Mg2+, Ca2+ - Water table levels (flood line) -Quality of some parameters if available Note if this has been monitored on a monthly basis 65
4.4 Soil sampling and analytical methods Soil samples were collected along the flood lines to determine the corresponding ECe values of the affected soil and waterlogged land. The flood line map was provided by the Northern Cape Department of Agriculture. Soil samples of surface soil (0-20 cm depth) were collected at each location with a soil auger. Soil sample for bulk density of surface soil was also collected at each location using a soil core sampler which is 100 cm3. At each location three samples were taken at an interval of 5-6 m and then samples were mixed to make a composite sample separate from surface and subsurface. In all, 120 soil samples were collected in the study area for analysis. They were oven dried and the samples were sieved using 2 mm sieve and used for analysis. The soil samples collected using soil core samplers for bulk density were kept in a hot air oven at 105o C for 24 hrs. The dry weights of soils were taken with electronic balance with accuracy of +0.1 gm.
4.4.1 Irrigation water sampling and analytical methods Water quality is increasing becoming a concern to irrigation both from a supply point of view and with respect to the environmental impacts of irrigation (Backeberg et al, 1996). To avoid problems when using poor quality water supplies, there must be sound planning to ensure that the quality of water available is put to the best use. Three irrigation canals: main, northern and western canals were used as sampling points. Water samples were taken at least 30 minutes after the water is released. At each sampling point, four 100ml samples were taken using clean plastic sampling bottles labeled for the purpose of water analysis identification.
The samples were analyzed for total dissolved solids (TDS), pH, electrical conductivity (Ec), carbonate (CO3²-), bicarbonate (HCO3-) sulphate (SO4²-), nitrate (NO3-), chloride (Cl-), 66
calcium (Ca²+), magnesium (Mg²+), potassium (K+) and sodium (Na+) according to the methods described by Chopra and Kanwar (1991). The TDS was determined by the evaporation and drying method. The pH and ECa were read on a pH-meter and conductivity meter, respectively. CO3²- and HCO3- were estimated by the volumetric titration method. The SO4²-, NO3- and Cl- were read on an LED photometer (LF200). Ca²+ and Mg²+ were determined by the EDTA titration method while Na+ and K+ by flame photometry. The adjusted sodium adsorption ratio of Suarez (1991) was calculated using the equation: Adj. RNa =
Na Cax Mg 2
………………………………………........................................(2)
Where, Cax = a modified Ca value taken from Ojo et al, 2006 modified from Ayers and Wescot (1985).
4.5 Remote sensing data used The main research work consists of remote sensing, GIS assisted spatial analysis and modeling of the data collected (field and SRS) through data pre-processing and image classification (using the maximum likelihood algorithm). For the study, Landsat (TM and ETM+) images data of VHS with Latitude E 24.7178°-27 2612 º S and Longtitude S 25. 2765°-28.0180° E (path 172, row 079) acquired for 1991\03\04, 2001\05\26 and 2005\03\02; representing 15 years (details in Table 4.2). They were obtained from the Global Land Cover Facility (GLCF) an Earth Science Data Interface hosted by the University of Maryland, USA. These were used to consider variation in land use/land cover and salinity in order to assess the environmental impact of the scheme. The Landsat images/data collected were mosaic for 1991, 2001 and 2005 (15 years) to cover the whole VHS area as shown in Figures 4.1, 4.2 and 4.3. 67
Basically, the underlisted software packages were used for this study: 1.
ArcGIS 9.2. This was used for displaying and subsequent processing and enhancement of the image;
2.
ERDAS IMAGINE 9.1. This was also used to process the data (mosaic);
3.
IDRISI Taiga. This was used for the development of land use/land cover classes and salinity subsequently.
Table 4.2: Landsat time-series used in the study Year 1991 2001 2005
Sensor Landsat TM Landsat ETM+ Landsat ETM+
Path 172 172 172
Row 079 079 079
Date 1991/03/04 2001/05/26 2005/03/02
4.5.1 Preprocessing of the data Since changes in vegetation activity commonly occur as a consequence of climate seasonally it is therefore important that the capture dates of the images are similar in different years. Images taken in different months and combined will have a strong influence on the temporal homogeneity of the dataset. All the used Landsat Thematic Mapper (TM) and Enhanced Thematic Mapper (ETM+) images were obtained from the archives of the Global Land Cover Facility (GLCF) an Earth Science Data Interface hosted by the University of Maryland, USA. Since most studies based on Landsat data consider ETM+ and TM radiometry to be comparable, TM data from 1991 was used and then switched to the ETM+ data when it became available in 2001 to 2005 due to the better calibration of this sensor (Teillet et al., 2001); subsequently this was switched back to TM data due to the ETM+ SLC failure in 2003. Months in summer pose fewer problems in terms of obtaining reliable clouds-free images; therefore, we selected the month of May to create a second time series. A total of 28 68
Landsat-TM and ETM+ images taken between 1991 and 2005 were acquired as shown in Table 4.2 lists the dates and types of the selected images.
4.5.1.1 Radiometric correction of data One common cause of radiometric error is sensor malfunction. Two types of sensor malfunction are line start error and banding (also called striping). The effects of this type of internal radiometric error can be easily reduced. All data used for this study were in saved in the z drive on the computer. The image files were loaded into ERDAS Imagine software package and then the bad lines were determined, as they appear white. The color of the inquire cursor was changed by selecting utility/inquire color to something other than white. The inquire cursor was moved over the bad lines. The coordinate system was then changed from “map” to “file.” the process repeated for the other bad lines, and record the line number for each. The image was fitted to the viewer and then the bad lines replaced by clicking interpreter. The image file was loaded, the band combination of this TM scene was changed, which has only the three visible bands included in order to view each band separately. This was done to carry out radiometric Enhancement/correction of the Landsat TM Data. The output file was saved in the working directory and the unstriped images displayed in the viewer.
4.5.1.2 Geometric correction of d ata The objective of geometric corrections is to compensate for the distortions and degradations caused by the errors due to the variation in altitude, velocity of the sensor platform, variation in scan speed and in the sweep of the sensor's field of view, earth 69
curvature and relief displacement. The random distortion was corrected through the selecting sufficient number of ground control points with correct coordinates and from maps and GPS points and then localized in the satellite image. The procedures are as shown in Figures 4.4 and 4.5.
To eliminate any discrepancies of mismatching during overlaying of the images, Georeferencing image is needed. Georeference is a process that establishes the relation between row / column numbers and real world coordinates. Two commonly used Georeference
approaches
was followed:
Georeference corners:
Specifying
the
coordinates of the lower left (as Xmin, Ymin) and upper right corner (as Xmax, Ymax) of the raster image and the actual pixel size. Georeference tie points: specifying reference points in an image so that specific row / column numbers to obtain a correct X, Y coordinates, this was done with the aid of topographic maps. To delineate and ground truth the data, a Garmin handheld GPS with a receiver accuracy of 2 m was used to obtain geographical coordinate of salt affected area. The geographical coordinates were transformed to the Hartebeesthoek 94 Datum and compared to ground coordinates. A second order polynomial transformation was used in an effort to keep the Root Mean Square Error (RMSE) lower than 1 pixel according to Mather (2004) recommendation. Figures 4.1 to 4.3 show the mosaic images used for the pre-processing; Figure 4.4 details the flow chart steps for the preprocessing and processing of the SRS data, while Figure 4.5 shows the actual process in IDRISI window.
70
Figure 4.1: Mosaic Image of 1991 (Source: GIS analysis)
71
Figure 4.2: Mosaic Image of 2001. (Source: GIS analysis) 72
Figure 4.3: Mosaic Image of 2005. (Source: GIS analysis) 73
Figure 4.4: Developed flow chart for the processing of the SRS data
Figure 4.5: Processing of the SRS data in IDRISI window. (Source: GIS analysis) 74
Figure 4.6: Processing of the SRS data for decision making to model. (Source: GIS analysis)
Figure 4.7: Markov Chain analysis of the SRS data. (Source: GIS analysis)
75
4.5.2 Image classification methods Image classification is done by assigning a specific part of the feature space corresponding to a specific class. Once the classes have been defined in the feature space, each image pixel can be compared to these classes and assigned to the corresponding class. Classes to be distinguished in an image classification need to have different spectral characteristics. Most of the procedures commonly applied to the classification of remote sensing images are based on the radiometric information contained in the image bands.
According to Price (1994), the pixels are expected to be more or less grouped in the multispectral space in clusters corresponding to different land cover types. This can be analyzed by comparing spectral reflectance curves. The principle of image classification is that a pixel is assigned to a class based on its feature vector, by comparing it to predefined clusters in the feature space. Doing this for all image pixels results in a classified image (Janssen and Gorte, 2001). In pixel based classification, two types of traditional classification methods: unsupervised classification and supervised classification are used.
4.5.3 Unsupervised classification For an unsupervised classification the computer software develops the spectral signatures that are used in the classification process. In addition, with an unsupervised classification, pixels are assigned to a class based on their spectral characteristics alone. During the unsupervised classification, the pixels will be clustered based on the natural spectral groupings present in the dataset. Pixels are assigned to classes based on their spectral distance from a class mean. This is an iterative process, with the class means shifting after each iteration. The process ends when either a maximum number of iterations have been 76
performed, or a maximum percentage of unchanged pixels has been reached between iterations. The various clusters created will then need to be compared to in situ data, in order to assign meaningful values to them.
4.5.4 Supervised classification In supervised classification, the image pixels’ categorization process was done by specifying to the computer algorithm, numerical descriptors of the various land cover types present in an image. Three basic stages involved in the supervised classification method are: training stage, classification stage and accuracy assessment stage. For supervised classification o f land use / land covers of the area, four classes were identified and training set was generated. The training set was generated with the aid of fieldwork and image interpretation of all images. Training sets layer was generated for the data set (Landsat TM and ETM+). The supervised classification technique using the maximum likelihood algorithm used in this study is described in 4.5.4.
4.5.5 Pixel matching 4.5.5.1 Training stage Training samples are those pixels in the image that represent the typical spectral information on land-cover classes and are selected by the user to train the classifier. To yield acceptable classification results, training data must be both representative and complete and must cover the spectral variance of the information classes (Lillesand and Kiefer, 2001). Irrespective of how training areas are delineated, when using any statistically based classifier, the theoretical lower limit of the number of pixels that must be contained in a training set is n+1, where n is the number of spectral bands. Equation 77
3 describes Markov change prediction algorithm developed in matrix form and used to perform the iterations. P = (Pij) =
P11
P12
-
-
P1n
P21
P22
-
-
P2n
P31
P32
-
-
P3n
-
-
-
-
-
Pn1
Pn2
-
-
Pnn
nj1 Pij 1 ..................................................................................................................(3) Where P (Pij) = number of pixels that must be contained in a training set is n+1, n is the number of spectral bands.
For this study training samples were obtained by outlining training areas using a reference cursor being controlled by any of several means. This was done according to the rule by Lillesand and Kiefer (2001) that all spectral classes constituting each information class must be adequately represented in the training set statistics t o b e used to classify the images. Training samples were selected according to the ground truth from the fieldwork. During fieldwork, coordinates for homogeneous land cover areas were recorded by GPS receiver. These homogeneous areas are identified in the image to form the training samples for all of the information classes.
4.5.5.2 Classification stage In this stage, classification is performed after specifying a set of training samples and a certain classification algorithm. The classification algorithms (classifiers) function this 78
way: pixels in the image are compared to each training sample numerically and are allocated to the land-cover classes according to certain algorithms. The classic classifiers used in pixel based image analysis are hard classifiers, which assign a membership of 1 or 0 to the objects, expressing whether an object belongs to a certain class, or not. Here the classifiers are called hard classifiers because they express the objects’ membership to a class only in a binary manner (yes or no). The commonly used classifiers are minimum distance to mean classifier, parallelepiped classifier, and maximum likelihood classifier. In this study, maximum likelihood classifier has been used. This classification w a s p e r f o r m e d a s described in Figure 4.5.
4.5.5.3 Maximum likelihood classifier The maximum likelihood decision rule is based on a normalized (Gaussian) estimate of the probability density function of each class as described by Pedroni (2003). The maximum likelihood classifier quantitatively evaluates both the variance and covariance of the category spectral response patterns when classifying an unknown pixel. After evaluating the probability in each category, the pixel was then assigned to the one with the highest probability value if the probability values are all below a threshold set by the analyst (Lillesand and Kiefer, 2001).
4.5.5.4 Band selection For this study, band selection was done through the aid of the reflectance properties of features, the correlation matrix of the bands and the spectral reflectance curve of known features (during the series of field visits) in all bands. Spectral profile was generated from the image using ERDAS IMAGINE 9.1. The different band combinations used for the analysis 79
were: 1.
Combination of band 4, band 3 and band 2;
2.
Combination of band 3, band 4 and band 2;
3.
Combination of band 3, band 2 and band 1;
4.
Combination of band 4, band 5 and band 3.
4.6 Methods for satellite data analysis Based on the above, five methods were adopted in this study for satellite data analysis namely: 1.
Calculation of the area in hectares of the resulting land use/land cover types and salinity area for the study years and subsequently comparing the results;
2.
Markov chain analysis was used for predicting change;
3.
Overlay Operations for the production of the maps;
4.
Image thinning; and
5.
Maximum likelihood classification.
4.6.1 Analyses of land use- land cover Based on the prior knowledge of the study area through a reconnaissance survey in addition to information obtained from literatures, the classification scheme was better developed giving a broad classification where the land use / land cover was identified by a single digit. The first three methods (area calculation, Markov chain analysis and overlay operation) above were used for identifying changes in the land use/land cover types. The land use/land cover were categorized as follows: 1.
Irrigated land;
2.
Built up land / settlement; 80
3.
Fallow land; and
4.
Water bodies.
The comparison of the land use / land cover statistics assisted in identifying the percentage change, trend and rate of change between 1991 and 2005. In achieving this, the first task was to develop a table showing the area in hectares and the percentage change for each year (1991, 2001 and 2005) measured against each land use land cover type. In obtaining an annual rate of change, the percentage change is divided by 100 and multiplied by the number of study year 1991 – 2005 (15years).
4.6.2 Analyses for salinity Contrast, brightness enhancements and true colour composite of the images were performed on the images using band 3, 4 and 5 for each scene. Mosaicing of the composite images was done in ERDAS Imagine 9.1 software to achieve a whole image (of four scenes) of the study area. The image was clipped in ArcGIS 9.2 and ERDAS IMAGINE software packages to extract the study area from the mosaic image see Figures 4.1, 4.2 and 4.3. It was ensured that co-registration of the images were within 0.5 pixels and 30m resolution. After the data w e r e pre processed, then the image classification using the maximum likelihood algorithm of classified images of Landsat were carried out for the years obtained.
Salt affected soils are usually characterized by areas of poorly developed vegetation and as such, the state of stress vegetation could be used as an indirect sign of the presence of salts in the soils. Soil salinity modeling is a methodology involving a set of analytical procedures that simulate real natural conditions within a GIS using their spatial relationships of geographic features to locate the problem of salinity geographic areas for a specific land-use. The procedure employed was by identification of the location of sampling point sources of 81
satellite data, this was done by following the flow charts in Figures 4.8 and 4.9. The corresponding reflectance of soil samples from different sampling zones were retrieved for different bands and indices. The following indices; Salinity Index (SI), Normalised Salinity Differential Index (NSDI) and Normalised Differential Vegetative Index (NDVI) proposed by Tripathi et al. (1997), was applied to give better results in the re-classification of saltaffected lands. Figures 4.4 to 4.8 give procedures of the classifications.
SI is the ratio of red band to near infrared (NIR) band while NSDI is the ratio of the difference of the red to NIR and divided by the summation of the two. This concept has emerged from the Red Edge concept for vegetation vigour mapping. In red edge concept, the spectral reflectance of the NIR is radioed with red band, which gives very high values for vegetation than other features on Earth. The SI, NSDI and NDVI indices used for salinity classification and quantification are computed as follows:
Band 3 SI = ...........................................................................................................4 Band 4
Band 3 Band 4 NDSI = ......................................................................................5 Band 3 Band 4 Band 4 Band 3 NDVI = .....................................................................................6 Band 3 Band 4
The actual calculation of NDVI is given in Equation 6 according to Chander et al (2009), but before the actual NDVI and NDBI were calculated, the Digital Number (DN) of the involving bands (3, 4 and 5) was converted to radiance and radiance to reflectance.
82
4.6.3 Conversion of Digital Number to Radiance In the analysis of multi-temporal images from sensors, the conversion of digital numbers to radiance values is an inevitable step. For this study, the conversion of digital numbers to radiance units was done to minimise bias in judgment of observed changes. The Digital Number (DN) of the involving bands (3, 4 and 5) was converted to radiance and radiance to reflectance as explained by Chander et al (2009) using equation 7: ..................................7 Where: = Spectral radiance at the sensor's aperture [W/(m2 sr μm)] = the spectral radiance that is scaled to QCALMAX [W/(m2 sr μm)] = the spectral radiance that is scaled to QCALMIN [W/(m2 sr μm)] Qcal = Quantized calibrated pixel value [DN] Qcalmax = the maximum quantized calibrated pixel value (corresponding to LMAXλ) in DN = 255 Qcalmin = the minimum quantized calibrated pixel value (corresponding to LMINλ) in DN = 1 for LPGS products = 1 for NLAPS products processed after 4/4/2004 = 0 for NLAPS products processed before 4/5/2004 LMAX and LMIN are obtained from the Meta data file available with the image and according to Gyanesh Chander et al. (2009) are given in the Table 4.3.
4.6.4 Conversion of Radiance to Reflectance The spectral radiances of the involving images as calculated with equation 7 was thereafter converted to planetary reflectance or abode which is a physical measurement to achieve 83
reduction in between-scene variability. The combined atmospheric and surface reflectance is therefore calculated using equation 8. .............................................................................................8 Where: = Unit less planetary reflectance = Spectral radiance at the sensor's aperture from equation 7 = Earth-Sun distance in astronomical units ESUN = Mean solar exoatmospheric irradiances (values from Table 4.4) = Solar zenith angle in degrees provided in the Meta file
Table 4.3: LMAX and LMIN values of Landsat data Band number
Satellite/Sensor
LMAX
LMIN
3
Landsat5/TM
264
-1.17
4
Landsat5/TM
221
-1.51
5
Landsat5/TM
30.2
-0.37
6
Landsat5/TM
15.3032
1.2378
3
Landsat7 /ETM+
152.90
-5.00
4
Landsat7 /ETM+
241.10
-5.10
5
Landsat7 /ETM+
31.06
-1.00
6.1
Landsat7 /ETM+ High
31.06
-1.00
6.2
Landsat7 /ETM+ low gain
12.65
3.20
84
Table 4.4: Mean solar exoatmospheric irradiances (E0) (w/(m2 * µm))[13] Sensor/Band 1
2
3
4
5
7
TM
1.983
1.796
1.536
1.031
220.0
83.44
ETM+
1.997
1.812
1.533
1.039
230.8
84.90
85
Secondary Data
Primary Data
Pre-field data organization and analysis
Study area boundary, topographic map and satellite image (ETM+)
Pre-processing Topographic map
Image processing Image enhancement
Supervised Classification
Post-processing NSDI Map of salt affected soil
Field verification Post supervised classification
Map of salt affected soil
Figure 4.8: Classification/Mapping of salt areas (salinity areas) by satellite imagery
86
Geological Formation
Reclassified
Flood line
Land sat
Level
ETM+
Interpolate
Supervised
(CRSI)
Classification
Reclassified
Land cover
Weighted Reclassified
DEM
Classification
Landform
Reclassified
Overlay Analysis
Salt affected soil map
Figure 4.9: Mapping of the salt affected areas based on land use/land cover classification
87
4.7 Modeling of land use / land cover and salinity Markov Chain Analysis is a convenient tool for modeling land use and salinity changes when changes and processes in the landscape are difficult to describe. A Markovian process is one in which the future state of a system can be modeled purely on the basis of the immediately preceding state. Markovian chain analysis was used to describe land use and salinity change from one period to another and also using this as the basis to project future changes (modeling). This is achieved by developing a transition probability matrix of land use change from time one to time too, which shows the nature of change while still serving as the basis for projecting to a later time period. The Markov Chain algorithm developed in IDRISI environment was used to perform 168 iterations for the change predictions. Figure 4.6, 4.7 and Tables 4.2 show the processes.
4.7.1 Overlay operation method The overlay operation method identifies the actual location and magnitude of change. This method was used to generate the different maps on the study. To assess the spatial soil salinity for the VHS, a model of salinity development was formulated upon the interaction of the factors. This was done by coupling together a GIS model relating the interaction of four thematic (factors) layers: geologic formation, flood line level, landform and land cover. Boolean logic was applied to the result through the re-classification module of IDRISI which assisted in mapping out separate areas of change whose magnitude was later computed for. A 30 m cell size was taken on the basis of the DEM (Digital Elevation Model) and Landsat ETM+ resolution. The processes were as shown in Figures 4.8 and 4.9.
88
4.7.1.1 DEM map Digital Elevation Model (DEM) maps of the VHS generated were clipped from SRS Landsat TM of 30 m resolution by using a masking layer of the VHS topography map. All the input parameters were re-sampled to 30 m cell size resolution.
4.7.1.2 Land cover map The land-cover map based on vegetation density was generated using Landsat ETM+ data for 1991, 2001 and 2005. The land-cover layer was reclassified to make the parameter compatible for GIS analysis in the salinity model with other model parameters. This is derived from DEM of the study area in a GIS platform using surface analysis in ArcGIS 9.2 software. The procedure used is illustrated in Figure 4.9.
4.7.1.3 Soil salinity map After the salt affected areas were identified, attempts were made to assess their spatial distribution in the form of maps. Thematic salinity maps were generated by integrating salinisation factors such as geomorphic, soil and elevation attributes in GIS and to calculate the areas affected in hectares. The output maps of elevation and soil physiographies map was crossed with a topographical map of the study area. The schematic diagram of the processes involved in generating the maps is as shown in Figures 4.5 to 4.9, depicting the step by step process to generate the land use / land cover and salinity maps.
89
4.8 Statistical analysis 4.8.1 Student's t-test Student's t-test is any statistical hypothesis test in which the test statistic follows a Student's t distribution if the null hypothesis is supported (Fisher Box, 1987). One-sample t-test for the data collected using the structured questionnaire was computed equation 9: Z= where
,
..............................................................9
is the sample mean of the data, is the sample size, and is the population standard deviation of the data;
s in the one-sample t-test is
,
where is the sample standard deviation. The assumptions underlying the t-test were that: 1.
Z follows a standard normal distribution under the null hypothesis
2.
s2 follows a χ2 distribution with p degrees of freedom under the null hypothesis, where p is a positive constant
3.
Z and s are independent.
In testing for one-sample t-test, the null hypothesis that the population mean is equal to a specified value μ0, one uses the statistic described in equation 10:
.......................................................................10 Where, is the sample mean, s is the sample standard deviation of the sample and n is the sample size. 90
The degrees of freedom used in this test were − 1. Dependent t-test for paired is also described in equation 11 (Fadem, 2008).
............................................................11 From equation 11, the differences between all pairs are calculated. The pairs are either one person's pre-test and post-test scores or between pairs of persons matched into meaningful groups (for instance drawn from the same family or age group: see table). The average (XD) and standard deviation (sD) of those differences were used in the equation. The constant μ0 is non-zero if you want to test whether the average of the difference is significantly different from μ0. The degree of freedom used is n − 1.
4.8.2 Pearson's chi-squared test Pearson's chi-squared test (χ2) was used to assess two types of comparison: tests of goodness of fit and tests of independence (Greenwood and Nikulin, 1996). The following steps were adopted. The first step was to calculate the chi-squared test statistic, X2, which resembles a normalized sum of squared deviations between observed and theoretical frequencies. The second step then determined the degrees of freedom, , of that statistic, which is essentially the number of frequencies reduced by the number of parameters of the fitted distribution. In the third step, X2 was compared to the critical value of no significance from the distribution, which in many cases gives a good approximation of the distribution of X2. For calculating the test-statistic, the value of the test-statistic is described by equation 12:
.......................................................12 Where, 91
= Pearson's cumulative test statistic, which asymptotically approaches a
distribution.
= an observed frequency; = an expected (theoretical) frequency, asserted by the null hypothesis; = the number of cells in the table.
The number of degrees of freedom is equal to the number of cells, minus the reduction in degrees of freedom. The result about the number of degrees of freedom is valid when the original data were multinomial and hence the estimated parameters are efficient for minimizing the chi-squared statistic. However, when maximum likelihood estimation does not coincide with minimum chi-squared estimation, the distribution will lie somewhere between a chi-squared distribution with
and
degrees of freedom
(Chernoff and Lehmann, 1954).
For test of independence, an observation consists of the values of two outcomes and the null hypothesis is that the occurrence of these outcomes is statistically independent. Each observation is allocated to one cell of a two-dimensional array of cells (called a table) according to the values of the two outcomes. If there are r rows and c columns in the table, the "theoretical frequency" for a cell, given the hypothesis of independence is given by equations 13 and 14:
......................................13 where N is the total sample size (the sum of all cells in the table). The value of the teststatistic is
92
.......................................................14 Fitting the model of independence reduces the number of degrees of freedom by p = r + c − 1. The number of degrees of freedom is equal to the number of cells rc, minus the reduction in degrees of freedom, p, which reduces to (r − 1)(c − 1). For the test of independence, also known as the test of homogeneity, a chi-squared probability of less than or equal to 0.05 (or the chi-squared statistic being at or larger than the 0.05 critical points) is commonly interpreted by applying workers as justification for rejecting the null hypothesis that the row variable is independent of the column variable (Plackett, 1983).
4.9 Limitations on the Study The weather data collected had some few point values missing, so data cleaning was carried out to fix some of the missing point values, thus this limited the use to only data in the year 1997 to 2010. Also, there was a limitation as a result of resolution difference. Landsat image of 1991 was acquired by the multi spectral scanner (MSS) which has a spatial resolution of 80 meters, whilst the images of 2001 and 2005 were acquired by the Thematic Mapper TM and Enhanced Thematic Mapper (ETM) respectively. Both TM and ETM have a spatial resolution of 30 meters. Although this limitation was corrected for through image thinning of the 1991, this still reduced its optimum use for projecting into the future so as to have a consistent result.
93
CHAPTER FIVE 5.0
RESULTS AND DISCUSSIONS
5.1
Results of weather data analysis
Three major variables, temperature, rainfall (precipitation) and relative humidity were considered as described below, detailed results of the analyses of these and other variables are given in Appendix B.
5.1.1 Temperature The study shows the average yearly minimum and maximum temperatures recorded as 9.720C and 23.520C respectively. There was a constant increase in temperature over the years; from 1983 to 2010. The marginal increases were observed between the years 1983 and 1987, 1998 and 2000, 2002 and 2004, 2005, 2008 and 2010 as shown in Figures 5.2a, 5.2b and 5.2c. These figures give an idea of the distribution pattern of temperature of the studied years in VHS. In temperature patterns, January was found to be the warmest month with maximum and minimum temperatures of 32.7 °C and 17.4 °C respectively. July was the coldest month with minimum day temperature of 2.4 °C. Common to VHS is the significant difference between the maximum and minimum temperatures as the seasons change. These results confirmed the finding of Arvind and Nathawat (2006) that land use and land cover pattern are generally influenced by agro climatic variables like temperature, ground water potential and a host of other factors.
The average yearly temperature of the study area increases significantly by a constant of about 0.1168444 (Pearson correlation coefficient, p = 0.163; 95 % confidence level: -0.054 – 0.288) as shown in Tables 5.1, 5.2 and 5.3. Coefficient of variation (CV) to describe the 94
dispersion of the variables in a way that it does not depend on the variable’s measurement unit. Here, the CV for temperature was found to be about 29.59. The variance is a measure of how far a set of numbers is spread out; and the variance of this set of observations is 26.625. The involvement of non-zero values in the serial correlation indicated the significance of the deterministic component in the data. Figures 5.1a, 5.1b and 5.1c show the spectral density function derived from autocorrelation of Temperature, rainfall and relative humidity respectively, while Figures 5.2a, 5.2c, 5.2c and 5.3 and Table 5.1 shows 12 temperature observations with a mean value of 17.45, standard deviation of 5.16 and average minimum and maximum valves of 9.720 and 23.520 respectively. Table 5.2 indicates a summary of statistical analyzed
temperature data with coefficients ranging from 0.117 to -216.84,
standard error from 0.079 to 157.43 and an average maximum of -0.54 to -559.84 and interval of 0.29 and 126.17 respectively and Table 5.3 show summary of the analyzed weather data. With a total of 4330 observations for maximum temperature and 4332 observations of minimum temperature giving a mean of 27.23 and 11.83 and standard deviation of 8.1 and -4.0 respectively. Other parameters such as rainfall and relatively humidity are discussed in 5.1.2 and 5.1.3.
95
Figures 5.1a: Autocorrelation graph of maximum temperature for VHS
Figures 5.1b: Autocorrelation graph of rainfall for VHS 96
Figures 5.1c: Autocorrelation graph of relative humidity for VHS
Figure 5.2a: Average min. & max. Temperature for the study area (1983-1992)
97
Figure 5.2b: Average min. & max. Temperature for the study area (1993-2002)
Figure 5.2c: Average min. & max. Temperature for the study area (2003-2010)
98
Figure 5.3: Average yearly temperature in the study area
Table 5.1: Summary of the analyzed basic temperature data Variable
Observation
Mean
Std. Dev
Ave. Min.
Ave. Max.
Temp ºC
12
17.438
5.160
9.720
23.520
Table 5.2: Summary of statistical analyzed temperature data Temp mean
Coefficient
Std. Error
T
P > |t|
Ave. Max.
Interval
Max
0.117
0.079
1.490
0.163
-0.544
0.288
Min
-216.839
157.427
-1.380
0.194
-559.843
126.165
99
5.1.2 Rainfall Rainfall is also an important climatic variable because of the critical role it plays in agricultural processes activating plant growth and indirectly water logging and salinity. It was observed that rainfall is maximum in the summer and minimum during winter. Rainfall data were considered from the year 1983 to 2010 similar to temperature data. Precipitation reached a peak of yearly average of 60 mm in 1988 and 44 mm in 1991, while the lowest with an average of 12 mm was recorded in 1992. Table 5.3 show summary of the analyzed rainfall data. With a total of 4802 observations which gives a mean of 1.77 and standard deviation of 10.37 with the minimum being zero (0) that is a period of no rain and maximum of 304 mm rainfall. The average yearly rainfall with normal weather conditions as well as years with favourable and severely unfavourable weather conditions are shown in Figure 5.4. VHS received the highest rainfall from the months of November to March. The rainfall is the lowest from the months of April to October. The details of the analysis are shown in Appendix B.
Figure 5.4: Average yearly rainfall for the study area. 100
5.1.3 Relative humidity High relative humidity is usually recorded during the months of May to September and low values in October to April which coincide with the peak of winter and summer seasons respectively. Table 5.3 show summary of the analyzed relative humidity data. With a total of 4356 observations which gives a mean of 87.55 and standard deviation of 13.05 with the minimum and maximum being -3.0 % and 100 % respectively. Details of the analysis are shown in Appendix B.
The Pearson correlation coefficient was used to test for correlation between rainfall and temperature parameters. The correlation between the two parameters was significantly negative (P= -0.0359; P < 0.05), meaning that increase in one resulted in a decrease in the other parameter. However, the strength of the correlation was weak. Also, the Pearson correlation coefficient was used to test for correlation between temperature and relative humidity. The correlation between the two parameters was also significantly negative (P= 0.0973; P < 0.05). The strength of the relationship was also very weak as summarised in Table 5.4. Detailed results of the analysis are in Appendix A.
101
Table 5.3: Summary of the analyzed weather data Variable
Obser.
Mean
Std. Dev.
Min
Max
Temp. max oC
4330
27.237
7.583
8.1
339.2
Temp. min oC
4332
11.833
24.010
-4.0
195.2
Rainfall (mm)
4802
1.766
10.370
0
304.0
Rel. Hum. (%)
4356
87.547
13.048
-3.0
100.0
rhn
4311
27.783
13.048
-17.7
174.9
4.048
1.677
0
25.74
Evapotranspiration 3943 (mm)
Table 5.4: Pearson correlation coefficient for temperature, rainfall and relative humidity Temperature
Rain
Relative Humidity
Temperature
1.000
1.000
1.000
Rain
-0.036*1.000
1.000
0.018 Relative Humidity1
-0.097*1.000
1.000
0.000
5.2 Irrigation Drainage water quality analyses The analyses of water samples from the main, northern and western canals indicated a mean pH of 7.5, 6.9 and 7.10 respectively. They all fell outside the range of 5.5 to 6.4, which is above the normal range recommended for irrigation water reuse. Table 5.5 shows the values of the electrical conductivity (EC) and total dissolved solids obtained which are indicators of 102
salinity. The samples have corresponding average electrical conductivities of 0.009, 0.012 and 0.011 mmho/cm and mean SARs of 3.53, 3.00 and 3.89 respectively. All the three canals (main, northern and western canals) gave water samples of low levels of Adj.RNA with a range of 0.12-0.17; these pose some hazards with respect to Na and salts build up in the soils. Most of the results are moderate when compared to the irrigation water standards and also fall within the USDA class Cl–Sl with low to medium salinity and sodium hazard tendencies. Generally, the quality of water from the sources appear suitable but are likely to cause soil deterioration in terms of increased salinity, toxicity or decreased permeability with continued usage as at the time of assessment. In agreement with the findings of Graham and Singh (1997) therefore, the continued use of such waters for irrigation might lead to nutrient imbalances and infiltration problems in soils as a result of salinity built up in the soil.
Figure 5.5: VHS Main Irrigation Water Supply Canal
103
Figure 5.6: VHS Northern Irrigation Water Supply Canal
Figure 5.7: VHS Western Irrigation Water Supply Canal 104
Table 5.5: Results of irrigation water sample analysis Sample point
1
2
3
Source: average values
MC
NC
WC
pH
7.50
6.90
7.10
PO-4(Mg/l)
0.01
N.D
0.02
NO-3(Mg/l)
2.70
3.40
2.01
K+ (Mg/l)
1.50
1.20
1.60
Na+(Mg/l)
1.00
1.20
1.20
Ca2+(Mg/l)
0.14
0.30
0.16
Mg2+(Mg/l)
0.02
0.02
0.03
SO42-(Mg/l)
10.40
15.40
9.50
Cl- (Mg/l)
8.00
6.80
1.40
CO3-(Mg/l)
0.02
N.D
0.01
HCO3- (Mg/l)
5.50
6.50
4.50
Fe3+(Mg/l)
0.12
0.12
0.10
solids 282
294
290
267
218
239
2.38
2.26
2.40
0.024
0.002
0.021
3.00
3.89
Total (Mg/l)
dissolved
Total suspended (Mg/l)
solids
Total hardness solids (Mg/l) Conductivity (ECW)
mmhos/cm 3.53
Key: MC = Main Canal NC= Northern Canal WC = Western Canal N.D = Not Detected 105
5.3 Baseline survey During the baseline survey and field visits, it was discovered visually that there is a high level of salinity in the study area as captured in Figures 5.8, 5.9, 5.10 and 5.11. The contour map in Figure 5.13 was used to develop the DEM and 3D DEM maps in Figures 5.14 and 5.15 in order to better understand the contribution of the land terrain in the formation of water logging and salinity in the area.
Figure 5.8: Effect of salinity in VHS. Source: Field survey on: 25 September, 2011
106
Figure 5.9: Effect of salinity in VHS. Source: Field survey on: 25 September, 2011
Figure 5.10: Effect of salinity in VHS. Source: Field survey on: 25 September, 2011
107
Figure 5.11: Effect of salinity in VHS. Source: Field survey on: 25 September, 2011
Figure 5.12: Water logging in VHS. Source: Field survey on: 25 September, 2011 108
Figure 5.13: The contour map of the VHS. Source: GIS analysis
109
Figure 5.14: Digital Elevation Model (DEM). Source: GIS analysis 110
Figure 5.15: Digital Elevation Model (DEM): 3D view. Source: GIS analysis 111
5.4
Results from satellite image processing
During the field checks, most locations in the scheme showed low to medium salinity effect demonstrated by little white patches on the soil surface. After field visits, areas which were classified as salt affected before were identified and marked out from the overall analysis. Informal discussion, interviews, questionnaire administration and field visits with the farmers and project managers in VHS were also used as inputs for pre and post-classification. The training areas given were based on the reflectance signature and combination of different bands.
Contrast, brightness enhancements and true colour composite of the images were performed on the images using band 3, 4 and 5 for each scene. Mosaicing of the composite images was done in ERDAS Imagine 9.1 software to achieve a whole image (of four scenes) of the study area. The images were clipped in ArcGIS 9.2 and ERDAS IMAGINE softwares to extract the study area from the mosaic image. It was ensured that co-registration of the images were within 0.5 pixels and 30m Resolution as seen in Figures 5.16, 5.17 and 5.18. After the data was pre-processed, the image classification using the maximum likelihood algorithm of Landsat image of was carried out for the years 1991, 2001 and 2005.
5.4.1 Results from supervised classification The results from supervised classification of ETM+ data are shown in Figure 5.23, 5.24 and 5.25. These results can be summarized as follows. The supervised classification algorithm was used in the identification of the areas with high salt contents. In order to be able to separate the areas where the salt is readily visible from other areas in the delineated portion of the study area, supervised classification algorithm was used to separate the four distinct land 112
use signature observed in the image. The land use / land covers are areas dominated by salt, water/swampy and vegetation. The supervised classification made use of 15 training sites per habitat identified on the image. As a rule there should be an adequate sample of pixels for each cover type for statistical characterization. A general rule of thumb is that the number of pixels in each training set (i.e., all the training sites for a single cover class) should not be less than ten times the number of bands. Hence, in this case three bands were used in the classification with 45 training sites with each cover type having fifteen each.
The distinctiveness of the training data was evaluated by displaying the signature graph against the three bands of the Landsat image used. The purpose of this exercise was to evaluate the distinctiveness and non-overlapping nature of the training site data obtained during the field data collection. Figure 5.23 shows that the three habitat classes observed are distinct as the line of each of the training data did not overlap or cross each other. The band 3 (red band) provides the best separation for vegetation, while band 4 (near infrared band) provide the best separation for the area with salt while water is distinct in all the bands. Band 4 provided a better separation of the area occupied by salt than band 5 (middle infrared band). Figures 5.23 to 5.25 also show that between band 4 and band 5, band 4 performed better when it comes to the spectral separation of areas dominated by salt but band 3 was extremely excellent in the vegetation discrimination. Figures 5.19, 5.20 and 5.21 depict the spectral signatures which aided in the classification processes. Salt area, community/built-up, vegetation and water swamp were clearly distinguished.
113
Figure 5.16: Clipped /mosaic image of VHS in 1991
114
Figure 5.17: Clipped / mosaic image of VHS in 2001 115
Figure 5.18: Clipped mosaic image of VHS in 2005 116
Figure 5.19: Spectral signatures of VHS during 1991
Figure 5.20: Spectral signatures of VHS in 1991 in point form 117
Figure 5.21: Spectral signatures of VHS in 2005
5.4.2 Land use / land cover analysis The static land use / land cover distribution for each study year is presented in Tables 5.6 to 5.12 and Figure 5.22. From Table 5.6 and Figure 5.22 in the year 1991 the settlement area occupied 28.78 % (10,453.26 Km2) of the total land mass with pockets of building around VHS, fallow land 41.4 % (15,038.31 Km2), water body 0.87 % (315.52 Km2) and irrigated land area 28.95 % (10,515.09 Km2). The study further revealed fallow land is the largest of the four classes. In 2001 from Table 5.6, the settlement area has decreased to 18 % (6,537.80 Km2) in the study area; water body and fallow land have increased to 1.0 % and 66. 48 % respectively, Irrigated / farming activities decreased to 14.52 %, this is represented in Figure 5.24. From Table 5.6, the year 2005 shows a further drop of 9.61 % in settlement areas, while there was significant reduction in water body (0.39 %), bare surface increased to 79.26 % and cultivated/irrigation land dropped to a low of 10.74 %. This change depicts that decrease in 118
the settlement area is proportional to decrease in cultivated land. The situation in 2005 is represented in Figure 5.25. The spatial variation in land use-land cover for the period 1991 to 2005 is as represented in Figure 5.23 and 5.25. The period 1991 to 2005 reveals a drastic change in the normal course of cultivated land. The year 2005 has shown a remarkable growth of fallow areas depicting a scenario of a likelihood of salinity problem. The process has led to elimination of crop, forest and scrub lands. The later period resulted in a rapid decline in the water bodies mainly because of the increase in the salinity problem, high temperature in the scheme over the years and topographical terrain of the study area are shown in Figures 5.38, 5.39 and 5.40. The overall projected change scenarios explained are shown in Figures 5.41 and 5.42. The land use/land-cover map was classified with an overall accuracy of 93.14 % details are as shown in the Figures in Appendix D.
119
24°30'0"E
25°0'0"E
25°30'0"E
24°30'0"E
26°0'0"E
25°0'0"E
27°30'0"S
27°30'0"S
27°30'0"S
27°30'0"S
26°0'0"E
Ipelegeng
Ipelegeng
Bloemhof
25°30'0"E Schweizer-Reneke
Schweizer-Reneke
Bloemhof Boitumelong DAM
DAM
Boitumelong
Tikwana
Tikwana
Warrenton DAM
Warrenton DAM
Windsorton
Windsorton
Seretse Kareehof
Seretse Kareehof
Kimdustria Bunn El Toro Park KimberleyBeaconsfield
29°0'0"S
29°0'0"S
29°0'0"S
Kimdustria Bunn El Toro Park KimberleyBeaconsfield
24°30'0"E
Barkley West
25°0'0"E
25°30'0"E
29°0'0"S
28°30'0"S
28°30'0"S
28°30'0"S
Barkley West
28°30'0"S
26°0'0"E
24°30'0"E
25°0'0"E
25°0'0"E
25°30'0"E
24°30'0"E
26°0'0"E
25°0'0"E
26°0'0"E
26°0'0"E
Ipelegeng 27°30'0"S
27°30'0"S
27°30'0"S
Ipelegeng
Bloemhof
25°30'0"E Schweizer-Reneke
Schweizer-Reneke
27°30'0"S
25°30'0"E
Fig B: 2001 Cover Map
Fig A: 1991 Cover Map 24°30'0"E
28°0'0"S
28°0'0"S
28°0'0"S
28°0'0"S
Hoopstad
Christiana
Hoopstad
Christiana
Bloemhof
DAM
DAM
Boitumelong
Boitumelong
Tikwana
Tikwana
Warrenton DAM
Warrenton DAM
Windsorton
Windsorton
Seretse Kareehof
28°30'0"S
28°30'0"S
25°30'0"E
29°0'0"S
29°0'0"S
29°0'0"S
25°0'0"E
26°0'0"E
0 12.525
50
24°30'0"E
25°0'0"E
25°30'0"E
26°0'0"E
Fig D: 2019 Predicted Cover Map
Fig C: 2005 Cover Map
Seretse Kareehof
Kimdustria Bunn El Toro Park KimberleyBeaconsfield
Kimdustria Bunn El Toro Park KimberleyBeaconsfield
24°30'0"E
Barkley West
29°0'0"S
28°30'0"S
28°30'0"S
Barkley West
28°0'0"S
28°0'0"S
28°0'0"S
28°0'0"S
Hoopstad
Christiana
Hoopstad
Christiana
Legend
75
100 Kilometers
Communities
2005 Cover Map
Salt Area
Class_Name
Communities
Forest
Water Swamp
Figure 5.22: Land cover maps of VHS from 1991, 2001, 2005 and predicted 2019/2020 120
5.5 Change prediction (modeling) using Markov analysis Markov Chain Analysis is a convenient tool for modeling land use and salinity changes where changes and processes in the landscape are difficult to describe. This process is one in which the future state of a system can be modeled purely on the basis of the immediately preceding state. Markov chain analysis was used in this study to describe land use and salinity change from one period to another and also as the basis to project future changes (modeling). This is achieved by developing a transition probability matrix of land use change from time one to time two, which shows the nature of change while still serving as the basis for projecting to a later time period. The Markov Chain algorithm developed in IDRISI environment was used to perform 168 iterations for the change predictions (Tables 5.10 and 5.11 for details). Each iteration performed during the processing lasted an average of 6 hours.
Table 5.6: Cover change between 1991 and 2005
Change Salt Area/ Salt Area Urban/Salt Area Vegetation /Salt Area Water Swamp/Salt Area Salt Area/Urban Urban/Urban Vegetation/Urban Water Swamp/Urban Salt Area/Vegetation Urban/Vegetation Vegetation/Vegetation Water Swamp/Vegetation Salt Area/Water Swamp Urban/Water Swamp Vegetation/Water Swamp Water Swamp/Water Swamp Total
Change (x 10-3Km2) Percentage Change 1991 & 2001 & 1991 & 1991 & 2001 & 2001 1991 & 2005 2005 2001 2005 2005 148383.70 247922.42 311649.12 3.91 6.53 8.21 46415.86 71233.35 175888.43 1.22 1.88 4.63 309726.36 592180.12 444302.21 8.16 15.60 11.71 207235.11 353815.94 333312.07 5.46 9.32 8.78 158144.26 299444.49 176577.79 4.17 7.89 4.65 109560.91 175374.85 262441.55 2.89 4.62 6.92 234174.84 346199.71 314033.96 6.17 9.12 8.27 171554.51 137061.91 205027.66 4.52 3.61 5.40 265707.85 93948.33 64369.84 7.00 2.48 1.70 74374.08 57034.49 112393.14 1.96 1.50 2.96 584474.55 229186.25 190153.74 15.40 6.04 5.01 401853.54 127910.45 141162.80 10.59 3.37 3.72 135561.03 66481.61 159164.29 3.57 1.75 4.19 108760.44 35468.68 122711.40 2.87 0.93 3.23 408027.78 368837.45 377920.11 10.75 9.72 9.96 431278.60 3795233.41
593133.38 3795233.41 121
404125.32 3795233.41
11.36 100
15.63 100
10.65 100
Table 5.7: Land-use / land-cover for the years with prediction for 2020 Cover Type 1 2 3 4
Salt Covered Urban Water/Swamp Vegetation Total
1991 Area (x 10-3 Km2) 707,796.84 339,111.28 1,536,403.52 1,211,921.75 3,795,233.4
2001 Area (x 10-3 Km2) 711,761.03 673,434.53 1,326,410.02 1,083,627.84 3,795,233.4
2005 Area (x 10-3 Km2) 1,265,151.83 958,080.96 1,063,921.11 508,079.52 3,795,233.4
2020 Predicted Area (x 10-3 Km2) 1,166,512.03 1,254,804.33 561,391.38 812,525.68 3,795,233.4
Table 5.8: Rate of change in cover type in the study area 1991-2001 3964.19 334,323.25 -209,993.50 -128,293.91
Salt Covered Urban Water/Swamp Vegetation
2001-2005 553,390.80 284,646.43 -262,488.91 -575,548.33
1991-2005 557,354.99 618,969.68 -472,482.41 -703,842.24
2001-2005 138,347.70 71,161.61 -65,622.23 -143,887.08
1991-2005 398,110.71 44,212.12 -337,487.44 -502,744.46
Table 5.9: Rate of change in cover type in the study area 1991-2001 396.42 33,432.33 -20,999.35 -12,829.391
Salt Covered Urban Water/Swamp Vegetation
Table 5.10: General conversion pattern
Change Salt Area/ Salt Area Urban/Salt Area Vegetation /Salt Area Water Swamp/Salt Area
1991 & 2001
Change (x 10-3 Km2) 1991 & 2001 & 2005 2005
Percentage Change 1991 & 1991 & 2001 & 2001 2005 2005
148,383.70 46,415.86
247,922.42 71,233.35
311,649.12 175,888.43
3.91 1.22
6.53 1.88
8.21 4.63
309,726.36
592,180.12
444,302.21
8.16
15.60
11.71
207,235.11
353,815.94
333,312.07
5.46
9.32
8.78
122
Table 5.10 shows the pattern of conversion of the four cover types observed in the delineated area. The result shows that generally, the area occupied by vegetation which incidentally entails part of the area currently devoted to agriculture are at risk of being converted than other cover types. The other threatened cover type is the swampy waterlogged areas which are particularly in close proximity to already salinised areas. The pattern is consistent across the different years studied.
Table 5:11: Markov transitional probability matrix
Salt Area
Urban
Vegetation
Water/Swamp
Salt Area
0.2977
0.4573
0.1435
0.1015
Urban
0.2438
0.4396
0.1952
0.1214
Vegetation
0.3956
0.2313
0.1268
0.2464
Water/Swamp
0.3339
0.1294
0.1207
0.416
Table 5.12: Markov transition area matrix Cells in (x 10-3 Km2)
Expected to Transit to:
Salt Area
Urban
Vegetation
Water/Swamp
Salt Area
4,637,448
7,122,503
2,234,628
1,581,313
Urban
2,875,801
5,185,092
2,302,574
1,431,928
Vegetation
2,474,368
1,446,562
793,130
1,541,152
Water/Swamp
4,373,866
1,694,356
1,581,226
5,448,995
123
Table 5.13: Land-use (1991) for the training site Land Use Type
Area (Km²)
Percentage (%)
Irrigated Land
10,515.091
28.95
Fallow Land
15,038.314
41.40
Settlement
10,453.264
28.78
Water Body
315.521
0.87
Total
36,325.000
100
Table 5.14: Area of land-use (2001) training site Land Use Type
Area (Km²)
Percentage (%)
Irrigated Land
5,272.961
14.52
Fallow Land
24,148.468
66.48
Settlement
6,537.797
18.00
Water Body
362.787
1.00
Total
36, 325.000
100
124
Table 5.15: Area of land-use (2005) training site Land Use Type
Area (Km²)
Percentage (%)
Irrigated Land
3,900.702
10.74
Fallow Land
28,790.348
79.26
Settlement
3,489.911
9.61
Water Body
141.229
0.39
Total
36, 325.000
100
125
Figure 5.23: Classified land-use classification 1991 126
Figure 5.24: Classified land-use classification 2001 127
Figure 5.25: Classified Land-use classification 2005 128
Figure 5.26: Land-use: farmland, fallow land, built up, Water body (1991)
Figure 5.27: Land-use: farmland, fallow land, built up and water body (2001) 129
Figure 5.28: Land-use: farmland, fallow land, built up and water body (2005)
5.6
Results from Indices Analysis
The NDVI maps developed are as shown in Figures 5.29, 5.30 and 5.31 below. The results of the vegetation index (NDVI) indicated the scarcity of vegetation as a whole in the highly saline area; confirming the ground truth. Minimum to lower threshold represented a negative change in NDVI, lower to upper threshold represented no change in NDVI, upper threshold to maximum value represented positive change in NDVI. Figures 5.32 to 5.37 show the temperature distribution over the area in map forms. A plot of NDVI and temperature gave a correlation of 67% to show the effect of temperature on the rate of vegetation/salinity level in the area. Figure 5.57 indicated that temperature is an important environmental factor in the build up of salinity in an area known to have a high temperature. Figures 5.38 to 5.42 show 130
occurrence pattern of salt-affected soils through supervised classification method and was grouped into salinity classes of low, medium and high. In year 1991, the salinity covered 4,978.7 km2, 15,675.8 km2 and 10,317.3 km2 (low, medium, high respectively) and for the year 2005, it covered 9,740.5 km2, 11,399.4 km2 and 9,831.9 km2 (low, medium, high respectively). This is an indication that in overall, there is an increase in the salinity trend over the years and this increase is still continuing.
From the result of NDSI, which is an inverse of NDVI it was observed that an area with high raster value or high reflectance was delineated as t he area affected by salinity problem. Once an area has high reflectance value was identified, based on their reflectance value the level of the salinity was determined. In the NDSI image the salt-affected areas, depicted in grey colour, could be roughly differentiated from those of non-salt-affected areas, water logged (in cyan colour) and vegetation (in green colour). Since this salinity problem is dynamic in nature and even varies with time, the image used which was captured in 2005 with ground truth information only gave slight indications of the presence of salinity. However, the extent of the soil salinity problem in the study area was found to be increasing. The salinity level classified in NDVI is given in Figures 5.29 to 5.31. The area-wise salinity distribution is given in the Table 5.6. Out of the total area, 6 % were classified as high salinity distribution throughout the study area with a moderate salinity area covering about 20 % of the total area.
131
Figure 5.29: Normalised Difference Vegetative Index (NDVI) 1991 132
Figure 5.30: Normalized Difference Vegetative Index (NDVI) 2001 133
Figure 5.31: Normalized Difference Vegetative Index (NDVI) 2005
134
Figure 5.32: 1991 Temperature distribution map of VHS. Source: GIS analysis 135
Figure 5.33: 1991 Temperature distribution map of VHS. Source: GIS analysis
136
Figure 5.34: 2001 Temperature distribution map of VHS. Source: GIS analysis 137
Figure 5.35: 2001 Temperature distribution map of VHS. Source: GIS analysis
138
Figure 5.36: 2005 Temperature distribution map of VHS. Source: GIS analysis
139
Figure 5.37: 2005Temperature distribution map of VHS. Source: GIS analysis 140
5.7 Distribution of the salt-affected areas in VHS The Visual interpretation, reflectance properties of salinity area coupled with the ground truth as well as other features were used to derive the resultant salinity classifications and maps. Based on these, the study area was categorized into three salinity levels: 1.
Low salinity: Low salinity is normally shown in pinkish white colour on the False Colour Composite. They normally show a little red or pink mottled textures,
2.
Medium salinity: Medium salt-affected areas are normally shown in light blue colour on the False Colour Composite. They also show fine texture with few mottled spots of other colours and,
3.
High salinity: High salinity mostly occurs where the relative reflectance is very high as compared to other features.
The findings of Tabet (1995), Vidal et al. (1998) Tabet (1999) and Latifovic and Fytas (2005) showed that Landsat MSS can be used easily for the identification of broad land cover changes where there are bare lands and cultivated or used lands as shown in the results obtained in this study. These indices were used to classify salinity for vegetative and nonvegetative areas, the resulting classifications allowed for the identification of highly saline and non-saline areas, but areas with low to medium salinity levels were difficult to distinguish.
The examination of the trend in accumulation of salt on the surface was accomplished using the Landsat images. The temporal trend shows that there has been an increase in the area occupied by salt between 1991 and 2005 as shown in Tables 5.7 to 5.12. Table 5.8 provides information on the trend in salt coverage and that the area covered by salt increased from 141
7,077.9 Km2 in 1991 to 7,117.6 Km2 in 2001 and to 12,651.5183 Km2 in 2005. Tables 5.8 and 5.9 show that 396.42 Km2 were actually covered between 1991 and 2001 while this increased to 5,533.90 Km2 between 2001 and 2005. The overall increase in the area of salt accumulation in soils between 1991 and 2005 was 5,573.549 Km2. Table 5.9 on the other hand shows the rate of increase in salt accumulation, and the rate of change in the seated area for 1991-2001, 2001-2005 and 1991-2005.
Figures 5.38 to 5.42 shows occurrence pattern of salt-affected soils through supervised classification method and was grouped into salinity classes of low, medium and high. In year 1991, the salinity covered 4,978.7 km2, 15,675.8 km2 and 10,317.3 km2 (low, medium, high respectively) and for the year 2005, it covered 9,740.5 km2, 11,399.4 km2 and 9,831.9 km2 (low, medium, high respectively). This means that low salinity class has increased by 4, 8618 km2, while medium and high salinity classes decreased by 4,296.4 km2 and 485.4 km2 respectively. This decrease is a result of continuous control measures in place at VHS which mainly focus on the high salinity area in the scheme. This is an indication that overall, there is an increasing trend in the salinity over the years.
To forecast the future trend in soil salinity based on historical patterns observed in the study area use was made of Markov chain analysis. The Markov transition probability matrix shows the probability that a given cover type will change to another cover type over the time period under consideration. Over the next fifteen years from 2005 to 2020, it was observed that the probability that much of the community area will show visible sign of salt cover is 0.46 (46%) shown in Table 5.11. Incidentally these areas are currently devoted to agriculture. The probability that area currently showing visible coverage will remain same is 0.2977. Based on Table 5.11 therefore, the most threatened cover type is the urban and the vegetation areas 142
while the area currently covered by water as well as the swamp area has the least probability of being converted. This matrix shows the numbers of cells that are likely to change to the other ones in the column. As an example, 7122503 of urban cells are likely to transit to an area covered by salt in the study area over the next 14years. Table 5.12 shows the result of the Markov analysis. This is in agreement with the findings of Muller and Middleton (1994); Wu (2006) that a set of conditional probability images give the probability that each land cover type would be found at each pixel after the specified number of times.
143
Figure 5.38: Salinity map indicating the salinity classifications (1991)
144
Figure 5.39: Salinity map indicating the salinity classifications (2001) 145
Figure 5.40: Salinity map indicating the salinity classifications (2005)
146
Figure 5.41: Salinity map indicating the salinity classifications (2015)
147
Figure 5.42: Salinity map indicating the salinity classifications projection (2020)
148
Figure 5.43: Salinity chart indicating the salinity classifications (1991)
Figure 5.44: Salinity chart indicating the salinity classifications (2005) 149
Figure 5.45: Projected salinity chart indicating the salinity classifications (2010)
Figure 5.46: Projected salinity chart indicating the salinity classifications (2015)
150
Figure 5.47: Projected salinity chart indicating the salinity classifications projected (2020)
5.8 Validation of the method Regression analysis is widely used for prediction and forecasting, it is also used to understand which among the independent variables relate to the dependent variable, and to explore the forms of these relationships. The performance of regression analysis methods in practice depends on the form of the data generating process and how it relates to the regression approach being used. For this study, years under consideration were used as dependent variables, while the area covered was used as independent variables. The relationship between the two variables for high salinity classification (R2) was 0.4864 approximately 0.5 depicting that high salinity classification is significant. However, the variations for low classification (R2 = 0.0675) and medium classification (R2 = 0.0229) are not explained because of their low values are approximately zero, which is as a result of how the data was
151
fitted due to non uniformity in the years interval (1991 to 2005, 2005 to 2010, 2010 to 2015 and 2015 to 2020).
Figures 5.48, 5.49 and 5.50 showed the regression analysis; therefore the salinity model can be used to predict soil salinity in the VHS more accurately. The area in square kilometres on the y-axis and years under consideration (1991 to 2020) on the x-axis represented by 1 to 6. From the validation and comparing the model results, it can be seen that, the salinity model is found to be an indicator of soil salinity. The salinity graphs plotted in Figure 5.48 and the correlation of ECe and salinity model gave the following model equations for each salinity class considered: 1. Low Salinity Class: y 329.68x 7158.9; R 2 0.0675 2. Medium Salinity Class: y 161.9 x 13869; R 2 0.0229 3. High Salinity Class: y 168.5x 9942.2; R 2 0.4864 4. Correlation of ECe & Salinity Model: y 25.04 x 22.2; R 2 0.4561
Figure 5.48: Salinity graph indicating the salinity classification regression analysis 152
Figure 5.49: Projected salinity model chart
Figure 5.50: Projected salinity model chart
153
5.9 Empirical Models of Soil Salinity Using ECe An empirical model for ECe vs. NDSI was prepared using regression analysis and it was observed that the model of ECe vs. NDVI offered a coefficient of determination of 77.5%. The ECe values were obtained from the flood lines digitalised map in Figure 5.51 and these were plotted against NDVI values. This shows a higher efficiency compared to NDVI vs. Temperature plot of 67% as shown in Figure 5.52. This model was extended for the whole image using modeler builder in ArcGIS and salinity levels were divided into three groups. The advantage of this model being generated through GIS was that it directly gives the salinity level at any point in the image (as seen in Figures 5.38 to 5.42, and 5.49 to 5.50). From the predicted salinity maps three ranges of salinity levels (classes) were generated; high, medium and low. The summaries of salinity level, the extent of the area in Km2 are given in the Figures 5.38 to 5.42 including the probability index. The spatial distribution of salt affected land derived from empirical model were checked by comparing it with salt affected area derived from N D V I - NDSI and it indicated the same areal extent in high salinity class but other classes did not show significant relation. Using the regression model derived, salinity maps were generated and the maximum and minimum EC values predicted were 29 and 0.75 dS/m, respectively. The use of this empirical model supported with less ground data helped give good results especially in assessing and modelling the surface soil salinity in VHS.
154
Figure 5.51: Electrical conductivity (EC) variation in the study area (sampling points)
155
Figure 5.52: NDVI and temperature plot
5. 10 Findings from structured questionnaires Tables 5.16 to 5.39 presented the findings obtained through the use of structured questionnaires administered among 40 farmers. Out of the 40 farmers randomly selected, approximately 41 % of the sampled class agreed that the irrigation practice better their crop production, while 25 % admitted that irrigation worsen their production and 21.9% believed that their production remain unchanged since the inception of the scheme (1930s till date). The chi-Square test gave a significance level of 0.4. The relationship between soil fertility and irrigation method is recorded at significance level of 0.3 as indicated by the chi-square test. This indicates that the sprinkler method improves the soil fertility which accounts for 34.4% of the sampled farmers. The farmers practicing furrow method indicated that the method does not significantly impact soil fertility and this represented 12.5% of the farmers.
156
The result of the soil fertility in relation to the sources of water used by the farmers indicated that 50% of the sampled farmers are of the opinion that water from the dam fairly improves soil fertility, 21.9% are of the opinion that water from dam worsen soil fertility while 25% of the farmers admitted that the soil fertility remains unchanged with the use of water from the dam while 3.1% of the farmers using water from central - pivot agreed that it worsens soil fertility. From the analysis it can be deduced that among the sampled farmers there are no significant changes in terms of irrigation method adopted in improving soil fertility and between soil fertility and water sources, as more than 70% indicated no changes among the factors considered. That is, there was almost no need in changing the irrigation method adopted as there was no accrued reason whether in relation to high income, higher demand for new crops or the viability of the method.
Tables 5.37 and 5.38 presented the results of the study that examined whether there is any significant difference between crops cultivated before and after the adoption of irrigation in the study area. The Chi square result shows a significance level of 0.03. The sample indicates that the dominant crops grown (wheat and groundnut) among the sampled farmers did not change before and after the adoption of irrigation in the study area. This is because of the economic importance of both, as they are cash crops and suitable for the area. The standard Deviation (SD) of 280.079 and 186.964 was derived from the number of land cultivated before and after irrigation respectively in the study area. The Student T-Test for equality of means shows a negative trend with -0.097 for Equal Variances Assumed and -0.-102 for Equal Variances not assumed. The Levene’s Test for Equality of Variances indicates a significance level of 0.7. This is as shown in Tables 5.39a to 5.39d, indicating less cultivated land area due to the salinity problem.
157
In conclusion, the sampling technique (random sampling) adopted proved to be best fitting as the results show a higher significance among the considered parameters. Among the parameters considered, water source and type of crops grown showed considerable effects on the accumulation of salinity in the soil in the study area.
Table 5.16: Fertility rate VS Irrigation practise Do you Irrigate Yes How would you rate the Fairly Better fertility Worse
No
40.6%
Total
9.4%
50.0%
25.0%
25.0%
About The Same
21.9%
3.1%
25.0%
Total
87.5%
12.5%
100.0%
Table 5.17a: Fertility rate VS Irrigation Method Irrigation Method
Sprinkler How would you rate the Fairly Better fertility Worse About The Same Total
34.4%
Furrow 3.1%
Flood Irrigation 3.1%
15.6%
6.3%
3.1%
12.5%
53.1%
15.6%
158
Sprinkler and Pivot
3.1%
3.1%
9.4%
Table 5.17b : Fertility rate VS Irrigation method Irrigation Method Furrow and Flood Irrigation How would you rate the Fairly Better fertility Worse
Furrow and Surface
6.3%
About the Same Total
6.3%
Total
3.1%
50.0%
3.1%
25.0%
6.3%
25.0%
12.5%
100.0%
Table 5.18: Fertility rate VS Source of Water Source of Water Dam How would you rate the Fairly Better fertility Worse
Pivot
50.0% 21.9%
About The Same
25.0%
Total
96.9%
159
Total 50.0%
3.1%
25.0% 25.0%
3.1%
100.0%
Table 5.19: Irrigation purpose VS Irrigation practise Do you Irrigate Yes If No Why did you change
Higher income
4.9%
New crops are in higher demand
4.9%
Low viability
2.4%
No
Total 4.9%
4.9%
9.8%
2.4%
No change
70.7%
12.2%
82.9%
Total
82.9%
17.1%
100.0%
Table 5.20a: Reason for not practising irrigation VS Irrigation method Irrigation Method Sprinkler If No why did you change
New crops are in higher demand
Furrow
Flood Irrigation
2.4%
2.4%
No Change
48.8%
9.8%
2.4%
Total
51.2%
12.2%
2.4%
160
Table 5.20b: Reason for not practising irrigation VS Irrigation method Irrigation Method
Sprinkler and Pivot If No Why did you change
Furrow and Flood Irrigation
Higher Income
2.4%
2.4%
New crops are in higher demand
2.4%
2.4%
Low Viability
Furrow and Surface
2.4%
No Change
2.4%
Total
7.3%
14.6% 7.3%
14.6%
Table 5. 20c: Reason for not practising irrigation VS Irrigation method Irrigation Method Basin and Surface If No Why did you change
Surface
Total
Higher Income
4.9%
New crops are in higher demand
9.8%
Low Viability
2.4%
No Change
2.4%
2.4%
82.9%
Total
2.4%
2.4%
100.0%
161
Table 5.21: Reason for not practising irrigation VS Source of Water Source of Water Dam If No Why did you change
Pivot
Total
Higher Income
4.9%
4.9%
New crops are in higher demand
9.8%
9.8%
Low Viability
2.4%
2.4%
No Change
80.5%
2.4%
82.9%
Total
97.6%
2.4%
100.0%
Table 5.22: Crop yield VS Irrigation practise Do you Irrigate Yes How Would You Assess Good your Yield Fair
No
Total
4.9%
2.4%
7.3%
19.5%
4.9%
24.4%
Poor
17.1%
4.9%
22.0%
No Difference
41.5%
4.9%
46.3%
Total
82.9%
17.1%
100.0%
162
Table 5.23a: Crop yield VS Irrigation method Irrigation Method
Sprinkler How Would You Assess Good your Yield Fair
Flood Irrigation Sprinkler and Pivot
Furrow
2.4% 12.2%
2.4% 7.3%
Poor
12.2%
No Difference
24.4%
4.9%
Total
51.2%
12.2%
2.4% 4.9%
2.4%
7.3%
Table 5.23b: Crop yield VS Irrigation method Irrigation Method Furrow and Flood Irrigation How Would You Assess Good your Yield Fair
Furrow and Surface
Basin and Surface
Surface
2.4% 2.4%
Poor
4.9%
No Difference
2.4%
9.8%
2.4%
2.4%
Total
7.3%
14.6%
2.4%
2.4%
163
Table 5.23c: Crop yield VS Irrigation method
Total How Would You Assess Good your Yield Fair
7.3% 24.4%
Poor
22.0%
No Difference
46.3%
Total
100.0%
Table 5.24: Yield VS Source of Water Source of Water Dam How Would You Assess Good your Yield Fair
Pivot
Total
7.3%
7.3%
24.4%
24.4%
Poor
19.5%
No Difference
46.3%
Total
97.6%
164
2.4%
22.0% 46.3%
2.4%
100.0%
Table 5.25: Cultivable land areas VS Irrigation practise Do you Irrigate Yes Are there Areas of Your Yes land that are not No Cultivable Total
No
Total
65.9%
2.4%
68.3%
17.1%
14.6%
31.7%
82.9%
17.1%
100.0%
Table 5.26a: Cultivable land area VS Irrigation method Irrigation Method
Sprinkler Are there Areas of Your Yes land that are not No Cultivable Total
Furrow
41.5%
Flood Irrigation
12.2%
9.8% 51.2%
12.2%
165
Sprinkler and Pivot 4.9%
2.4%
2.4%
2.4%
7.3%
Table 5.26b: Cultivable land areas VS Irrigation method Irrigation Method Furrow and Flood Irrigation Are there Areas of Your Yes land that are not No Cultivable Total
Furrow and Surface
Basin and Surface
4.9% 7.3%
9.8%
7.3%
14.6%
2.4%
2.4%
Table 5.26c: Cultivable land areas VS Irrigation method Irrigation method Surface Are there Areas of Your Yes land that are not No Cultivable
Total 2.4%
68.3% 31.7%
Total
2.4%
100.0%
Table 5.27: Cultivable land areas VS Source of water Source of Water Dam Are there Areas of Your Yes land that are not No Cultivable Total
Pivot
65.9%
Total 2.4%
31.7% 97.6%
166
68.3% 31.7%
2.4%
100.0%
Table 5.28: Factors affecting Yield VS Irrigation practise Do you Irrigate Yes Which of the Following Poor Soil Condition has affected Your Yield Poor soil and Loss of soil Due to Flood Everything
No
22.0% 17.1%
Total 22.0%
4.9%
2.4%
22.0%
2.4%
Loss of soil Due to Rain and Poor water from the Source
14.6%
2.4%
17.1%
Stagnant Water and Loss of Soil Due to Rain
12.2%
2.4%
14.6%
Poor Water From Source
14.6%
7.3%
22.0%
Total
82.9%
17.1%
100.0%
167
Table 5.29a: Factors affecting Yield VS Irrigation method Irrigation Method Sprinkler Furrow Which of the Following Poor Soil Condition has affected Your Yield Poor soil and Loss of soil Due to Flood
19.5%
2.4%
7.3%
7.3%
Everything
2.4%
Loss of soil Due to Rain and Poor water from the Source
7.3%
Stagnant Water and Loss of Soil Due to Rain
9.8%
Poor Water From Source
4.9%
2.4%
51.2%
12.2%
Total
Flood Irrigation
2.4%
2.4%
Irrigation Method
Sprinkler and Pivot Which of the Following Poor soil and Loss of has affected Your Yield soil Due to Flood Loss of soil Due to Rain and Poor water from the Source
2.4%
4.9%
Poor Water From Source Total
7.3% 168
Furrow and Flood Furrow and Irrigation Surface 2.4%
4.9%
2.4%
12.2%
7.3%
14.6%
Table 5.29b: Factors affecting Yield VS Irrigation method Irrigation Method Basin and Surface
Surface
Which of the Following Poor Soil Condition has affected Your Yield Poor soil and Loss of soil Due to Flood
Total 22.0% 22.0%
Everything
2.4%
Loss of soil Due to Rain and Poor water from the Source Stagnant Water and Loss of Soil Due to Rain
17.1%
2.4%
2.4%
Poor Water From Source
14.6%
22.0%
Total
2.4%
169
2.4%
100.0%
Table 5.30: Factors affecting Yield VS Source of water Source of Water Dam Which of the Following Poor Soil Condition has affected Your Yield Poor soil and Loss of soil Due to Flood Everything
Pivot
22.0% 19.5%
Total 22.0%
2.4%
22.0%
2.4%
2.4%
Loss of soil Due to Rain and Poor water from the Source
17.1%
17.1%
Stagnant Water and Loss of Soil Due to Rain
14.6%
14.6%
Poor Water From Source
22.0%
22.0%
Total
97.6%
170
2.4%
100.0%
Table 5.31: Effect of Irrigation practise on the soil Do you Irrigate Yes If Yes;
state the effect:
No
Total
4.9%
4.9%
Increase in Yellow nut edge
22.0%
2.4%
24.4%
White Patches
29.3%
4.9%
34.1%
Increased Salt
2.4%
2.4%
Worse
24.4%
9.8%
34.1%
Total
82.9%
17.1%
100.0%
Table 5.32: Effect of Irrigation practise on the soil VS Irrigation method Irrigation Method
Sprinkler If Yes;
State the effect:
Furrow
Flood Irrigation
Sprinkler and Pivot
4.9%
Increase in Yellow surface
12.2%
4.9%
White Patches
17.1%
7.3%
Worse
17.1%
Total
51.2%
171
12.2%
2.4%
4.9%
2.4%
2.4%
7.3%
Table 5.33a: Effect of Irrigation practise on the soil VS Irrigation method Irrigation Method Furrow and Flood Irrigation If Yes; the effect:
White Patches
Furrow and Surface
4.9%
Basin and Surface
Surface
2.4%
Increased Salt
2.4%
Worse
2.4%
9.8%
2.4%
2.4%
Total
7.3%
14.6%
2.4%
2.4%
Table 5.33b: Effect of Irrigation practise on the soil VS Irrigation method
Total If Yes;
State the effect;
4.9%
Increase in Yellow nut edge
24.4%
White Patches
34.1%
Increased Salt
2.4%
Worse
34.1%
Total
100.0%
172
Table 5.34: Effect of Irrigation practise on the soil VS Source of water Source of Water Dam If Yes; state
None
Pivot 4.9%
Total 4.9%
Increase in Yellow nut edge
22.0%
White Patches
34.1%
34.1%
Increased Salt
2.4%
2.4%
Worse
34.1%
34.1%
Total
97.6%
173
2.4%
2.4%
24.4%
100.0%
Table 5.35: Health problem VS Irrigation practise Do you Irrigate Yes Since The Scheme Worms Started, What Types Of Sinus Sickness
No
Total
19.5%
19.5%
2.4%
2.4%
Sinus and Worms
2.4%
Sinus and hearing
2.4%
2.4% 2.4%
Don't Know
58.5%
14.6%
73.2%
Total
82.9%
17.1%
100.0%
Table 5.36a: Health problem VS Irrigation method Irrigation Method
Sprinkler Since The Scheme Worms Started, What Types Of Sinus Sickness
Furrow
Flood Irrigation
Sprinkler and Pivot
19.5% 2.4%
Don't Know
31.7%
12.2%
2.4%
4.9%
Total
51.2%
12.2%
2.4%
7.3%
174
Table 5.36b: Health problem VS Irrigation method Irrigation Method Furrow and Flood Irrigation Since The Scheme Sinus and Worms Started, What Types Of Sinus and hearing Sickness
Furrow and Surface
Basin and Surface
2.4% 2.4%
Don't Know
2.4%
14.6%
2.4%
Total
7.3%
14.6%
2.4%
TTable 5.36c: Health problem VS Irrigation method Irrigation Method Surface
Total
Since The Scheme Worms Started, What Types Of Sinus Sickness
19.5% 2.4%
Sinus and Worms
2.4%
Sinus and hearing
2.4%
Don't Know
2.4%
73.2%
Total
2.4%
100.0%
175
Table 5.37: Health problem cases on the scheme VS Source of water Source of Water Dam Since The Scheme Worms Started, What Types Of Sinus Sickness
17.1%
Pivot 2.4%
Total 19.5%
2.4%
2.4%
Sinus and Worms
2.4%
2.4%
Sinus and hearing
2.4%
2.4%
Don't Know
73.2%
73.2%
Total
97.6%
176
2.4%
100.0%
Table 5.38a: Crop type grown presently VS crop type grown before Were these the Crops you Were Planting before Yes What Type of crops do Corn, Wheat, Pecan nut Count you Grow % of Total Corn, Wheat, Pecan Nuts, Barely, Lucerne
Count % of Total
Maize, Ground Nuts, Wheat, Barely, Pecan, Lucerne
Count
Maize and wheat
Count
% of Total
% of Total Wheat, Maize and Peanuts
Count % of Total
Maize, Wheat and Ground Nuts
Count % of Total
Corn, Wheat and Ground Nuts
Count % of Total
Wheat, Ground Nuts, Pecan and Alfalfa
Count % of Total
Total
Count % of Total
177
No 6
1
14.6%
2.4%
4
2
9.8%
4.9%
2
2
4.9%
4.9%
7
0
17.1%
.0%
4
5
9.8%
12.2%
1
5
2.4%
12.2%
0
1
.0%
2.4%
0
1
.0%
2.4%
24
17
58.5%
41.5%
Table 5.38b: Crop type grown presently VS Crop type grown before
Total What Type of crops do Corn, Wheat, Pecan nut Count you Grow % of Total Corn, Wheat, Pecan Nuts, Barely, Lucerne
Count % of Total
Maize, Ground Nuts, Wheat, Barely, Pecan, Lucerne
Count
Maize and wheat
Count
% of Total
% of Total Wheat, Maize and Peanuts
Count % of Total
Maize, Wheat and Ground Nuts
Count % of Total
Corn, Wheat and Ground Nuts
Count % of Total
Wheat, Ground Nuts, Pecan and Alfalfa
Count % of Total
Total
Count % of Total
178
7 17.1% 6 14.6% 4 9.8% 7 17.1% 9 22.0% 6 14.6% 1 2.4% 1 2.4% 41 100.0%
Table 5.39a: Basic group statistics Before and After Amount of land Cultivated
N
Before After
Mean
Std. Deviation
Std. Error Mean
10 1470.00
280.079
88.569
8 1481.25
186.964
66.102
Table 5.39b: Independent sample test t-test for Levene's Test for Equality of Equality Variances of Means
F Amount of land Cultivated
Equal variances assumed
Sig. .183
Equal variances not assumed
t .674
-.097
-.102
179
Table 5.39c: Independent sample test t-test for Equality of Means
Diff. Amount of land Cultivated
Equal variances assumed Equal variances not assumed
Sig. (2-tailed)
Mean Difference
16
.924
-11.250
15.597
.920
-11.250
Table 5.39d: Independent sample test t-test for Equality of Means 95% Confidence Interval of the Difference Std. Error Difference Amount of land Cultivated
Lower
Upper
Equal variances assumed
115.625
-256.363 233.863
Equal variances not assumed
110.516
-246.027 223.527
180
CHAPTER SIX 6.0 CONCLUSIONS AND RECOMMENDATIONS 6.1 Conclusions 6.1.1 Introduction Soil salinity is a critical environmental problem which has great impact on soil fertility and overall agricultural productivity. Since soil salinity processes are highly dynamic, the method of detection, mapping and modeling it should also be dynamic. The overall objective of this study was to map and model the spatio-temporal trend of soil surface salinity and land use / land cover of VHS using Landsat enhanced thematic mapper (ETM+) data along with other field data and topographical maps to show the spectral classes and salt-affected areas from 1991 to 2005 for sustainable land management and development that is eco-friendly.
It was also envisaged that the resulting maps and models should not only be simpler but also interpretable in meaningful management terms. This chapter presents the conclusions drawn from the results obtained in respect of mapping and modeling of the Vaal Harts Irrigation scheme.
6.1.2 Research questions The results obtained in Sections 5.1 to 5.10 and the discussions that followed have demonstrated the following: 1. That the significant environmental impacts that have arisen due to irrigation and associated spatial developments in the Vaal Hart Irrigation scheme are linked mainly to irrigation induced surface salinity;
181
2. That land use change in terms of salinity in the scheme can be assessed with satellite imagery, analysed and modelled; 3. That there are changes that have occurred in VHS due to the problem of salinity with varying areal extents in its trends / rate in the scheme over the years; and 4. That through the use of structured questionnaires and field survey, the factors influencing land use change and salinity problem in the scheme can be determined.
These are in line with similar research questions in a study of congestion (1998) on change detection and its importance to monitoring natural resources. The results obtained in this study are therefore linked to the identified problem of irrigation induced land use / land cover and salinity in VHS.
6.1.3 Land use / land cover analysis in VHS The results and discussions in Section 5.4.2 have demonstrated the ability of the use of remote sensing (Landsat TM and ETM +) and GIS to assess land use / land cover fairly well in a simple manner and to obtain a model that is interpretable in meaningful terms.
The static land use / land cover distribution for each study year is presented in section 5.4 (Tables 5.6 to 5.12 and Figure 5.22). For the year 1991 the settlement area occupied 28.78 % (10,453.26 Km2) of the total land mass with pockets of building around VHS, fallow land 41.4 % (15,038.31 Km2), water body 0.87 % (315.52 Km2) and irrigated land area 28.95 % (10,515.09 Km2). The study revealed fallow land as the largest of the four classes. The representation is shown in Figure 5.23 and Table 5.14. In 2001, the settlement area decreased to 18 % (6,537.80 Km2) in the study area; water body and fallow land increased to 1.0 % and 66. 48 % respectively, Irrigated / farming activities decreased to 14.52 %, this is represented 182
in Figure 5.24. In the year 2005 there was a further drop in the settlement area by 9.61%, while there was a significant reduction in water body by 0.39 %, bare surface increased to 79.26 % and cultivated/irrigation land dropped to a low of 10.74 %. This change depicts that the decrease in the settlement area is proportional to decrease in cultivated land; this is a reflection that more agricultural lands are being left uncultivated due to salinity problems. This situation in 2005 is represented in Figure 5.25. The spatial variation in land use / land cover for the period 1991 to 2005 is as represented in Figures 5.16, 5.17 and 5.18. The period 1991 to 2005 reveals a drastic change in the normal course of cultivated land. The year 2005 showed a remarkable growth of fallow areas depicting a scenario of a likelihood of salinity problem. The process has led to elimination of the crop, forest and scrub lands. The later period resulted in a rapid decline in the water bodies mainly because of the increase in the salinity problem, high temperature over the scheme with years and topographical terrain of the study area in Figures 5.26, 5.27 and 5.28. The percentage of fallow land increased by 37.86 %; settlement or built-up area decreased by 18.97 %, water bodies by 0.48 %, cultivated / irrigated land area also decreased by 18.21 %. The land use / land-cover map was classified with an overall accuracy of 93.14 %.
The implication of these results is that the Landsat Thematic mapper (TM and ETM +) data are very good and useful for developing salinity, land use / land cover pattern maps for irrigation land which is in agreement with findings of Dimyati (1995).
6.1.4 Results from Indices Analysis The results in Section 5.6 and the ensuing discussions have exhibited a further extension of the salinity index analysis using GIS. In this study, the results of the vegetation index (NDVI) indicated the scarcity of vegetation as a whole in the high salinity area, thus confirming the 183
ground truth findings during the field survey. Minimum to lower threshold represented a negative change in NDVI, lower to upper threshold represented no change in NDVI, upper threshold to maximum value represented positive change in NDVI. This was in agreement with findings of Mulders, 1987; Agbu, Fehrenbacher and Jansen, 1990; Zuluaga, 1990; Vincent et al., 1996; Dehaan and Taylor, 2002) that field derived spectra of salinised soils and vegetation indexes are good indicators of irrigation induced soil salinisation for identification of saline soil regions.
A plot of NDVI and temperature gave a correlation of 67% to show the effect of temperature on the rate of vegetation/salinity level in the area. This is an indication that temperature is an important environmental factor in the build up of salinity in areas known for high temperature (semi-arid climatic zone in which the study area fell into).
6.1.5 Salinity in VHS The findings of Latifovic and Fytas (2005) showed that Landsat MSS can be used easily for the identification of broad land cover changes where there are bare lands and cultivated or used lands. This can also be extended for salinity detection and mapping. The findings in this study confirm this. The occurrence pattern of salt-affected soils through supervised classification method was grouped into salinity classes of low, medium and high as shown in chapter 5 (Figures 5.38 to 5.42). In year 1991, the salinity covered 4,978.7 km2, 15,675.8 km2 and 10,317.3 km2 (low, medium, high respectively) and for the year 2005, it covered 9,740.5 km2, 11,399.4 km2 and 9,831.9 km2 (low, medium, high respectively). The temporal trend shows that there has been an increase in the area occupied by salt between 1991 and 2005. The overall trend in salt coverage showed that the area covered by salt increased from 7,077.9 Km2 in 1991 to 7,117.6 Km2 in 2001 and to 12,651.5183 Km2 in 2005. Results 184
(Table 5.8 and 5.9) show the rate of change in cover type in the study area to be 3,964.2 for salt cover between 1991 and 2001, while this increased to 5,533.90 between 2001 and 2005. The overall increase in salt accumulation in soils between 1991 and 2005 was 5,573.55. This is an indication that there is an overall increase in the salinity trend over the years. The areawise distribution of the salinity was given in Table 5.6. Out of the total area, 6 % were classified as high salinity which was distributed throughout the study area and medium salinity which covered approximately 20 % of the total area of VHS.
6.1.6 Modeling of salinity The future trend in soil salinity based on historical patterns observed in the study area was accomplished using Markov chain analysis. Over fifteen years from 2005 to 2020, it was observed that the probability that much of the community area will show visible sign of salt cover is 0.46 (46%) shown in Table 5.11. Incidentally these areas are currently devoted to agriculture. The probability that the area currently showing visible coverage will remain the same is 0.2977; therefore the most threatened cover type is the urban and cultivated areas while the area currently covered by water as well as the swamp area have the least probability of being converted as the salt is being leached into the groundwater.
6.1.7 General conclusion The results obtained have shown that the study has met its objectives. It has been shown that the application of RS and GIS for mapping and modelling of soil surface salinity can yield a simple model that represent the whole study area (VHS) well which can be interpreted in physically meaningful maps in accordance with the subject paradigms. The results obtained using TM and ETM + to map and model soil surface salinity and land use in VHS confirmed 185
the findings of Dimyati (1995), Lewis (1994) and Goossens et al (1993) that there are close relationships between field data and spectral data and that TM and ETM+ have a high detection accuracy compared with MSS and SPOT. The process of mapping and modeling is necessary for updating land cover maps and the management of natural resources as suggested by Xiaomei and Rong Qing in 1999 since land use and land cover change may result in environmental, social and economic impacts of greater damage than benefit to the area (Moshen, 1999). The maps and models developed in this study on salinity are of great importance to planners in monitoring the consequences of land use change on the area. These data are of value to resources management and government agencies that plan and assess land use patterns to model and predict future changes especially Department of Agriculture.
6.2 Recommendations The study has been successful but for a few shortfalls that could be looked at further with the availability of more time and financial resources. Therefore the following recommendations are suggested: 1. Despite the fact that the overlay salinity method is simple and easy to apply, the entire relationship changes with time and location. Therefore, requires some field visits to adjust the parameters being considered (both field salinity levels as well as the model parameters); 2. The Landsat sensor detects only the salinity of the surface of the soil and gives no detailed idea about the conditions below the surface; so subsurface information should be obtained indirectly through the use of groundwater salt level combined with overlay salinity methods using advanced sensors like SPOT. 3. Since, the spatial resolution of satellite images still limits the detection of small patches of salt affected area; high-resolution images from latest sensors (for 186
example hand held hyper spectrometer) should be applied to increase the accuracy of detection; 4. In order to assess the reliability of the Landsat (TM and ETM +) method applied, it is
recommended that similar studies on ground water incursion in similar agroecological conditions should be undertaken in order to know the effect of underground salt accumulation effect on the soil surface salinity problem of the scheme; 5. By incorporating the recommendations from 1 to 4 into the maps and model developed, this integrated approach will help to refine the methodology currently in use and determine what tools and technologies are best suited to improve the accuracy of remote sensing and GIS in mapping soil salinity (surface and underground) for the study area; and 6. Based on the results obtained in the study work showing low to medium surface salinity level with a probability of its increase, there is an urgent need for a management program to control the spread of the menace by re-grading the land surface (topography) and incorporating subsurface drainage system or selected appreciated drip irrigation system thereby reclaiming the degraded VHS farm land in order to make the scheme more economically viable.
187
REFERENCES
Adeleke, O.O. and Aremu, A.G. (2002). Application of GPS for mapping of Nigerian roads, Nigerian journal of technological development, Vol. 2 (1) Dec. 2002 pp 62-68. Afoz, A. and Singh, P. (1991). Environmental Impact Assessment for Sustainable Development: Chittaugarh on Project in Outer Himalayas. Ambio 20 (7): 298 – 302. Agbu, P. A., Fehrenbacher, D. I. and Jansen, I. J. (1990). Soil-property relationships with SPOT satellite digital data in East Central Illinois. Soil Sci. Soc. Am. J., 54: 807-812. AGIS (Agricultural Geo-Referenced Information System) (2009). URL: www.agis.agric.za. Accessed on May 24th 2009. Ahmad, A. and P. P. Singh (1991). Environmental Impact Assessment for sustainable development: Chittaurgarh Irrigation Project in Outer Himalayas. JSTOR: Ambio, Vol. 20, No. 7 pp. 298-302. Akinyede, J. O. and Boroffice, R. A. (2004). Geoinformation, space application and sustainable national development. Nigerian journal of surveying and Geoinformatics Vol. 1 No.1 pp17-40. Arora, A.N. and Goyal, Rohit (2002). Environmental and Socio-Economical Impacts of Water logging in Hanumangarh and Sriganganagar Districts. Nature Environment and Pollution Technology, 1 (3), pp. 307-316. Arora, A.N. and Goyal, Rohit (2003). Conceptual Groundwater Modeling using GIS. GIS India 2003, National Conference on GIS/GPS/RS/Digital Photogrammetry and CAD, Jaipur. Arvind, C. P. and. Nathawat, M. S. (2006). Land Use Land Cover Mapping Through Digital Image Processing of Satellite Data – A case study from Panchkula, Ambala and Yamunanagar Districts, Haryana State, India. 188
Ashton, P. (1999), Using environmental impact assessment to determine the consequences of mining activities and to highlight the costs of sustainable development, Proceeding, Conference of Environmental Management System Mining. Aurecongroup (2010). http://www.aurecongroup.com/en/Projects/Water/Vaalharts-IrrigationScheme-Study.aspx accessed on the 21st March, 2010. Azzali, S., Menenti, M., Meuwissen, I. J. M. and Visser, T. M. N. (1990). Mapping crop coefficients in an Argentinian irrigation scheme using remote sensing techniques. In M. Menenti, Ed, Remote sensing in evaluation and management of irrigation, p. 79101. Mendoza, Argentina. Baber, J. J. (1982). Detection of crop conditions with low-altitude aerial photography. In: Remote Sensing for Resource Management, pp. 407-412, C.J. Johannsen and J.L. Sanders, Iowa. Bari, M. A., and Schofield, N. J. (1992). Lowering of a Shallow, Saline Water Table by Extensive Eucalypt Reforestation. Journal of Hydrology 133.3–4 (1992), 273–291. Barnard, J. H. (2008). Personal communication; lecturing material. Department of Soil Crop and Climate Sciences, University Free State, Bloemfontein. South Africa. Barnard, R. O., Van der Merwe, A. J., Nell, J. P., De Villiers, M. C., Van der Merwe, G. M. E. and Mulibana, N. E., (2002). Technical country report/in-depth study on problem soils including degraded soils in South Africa: Extent, present use, management and rehabilitation (with emphasis on salt-affected soils). 4th Meeting of FAO Global Network Integrated Soil Management for Sustainable Use of Salt-Affected Soils. Valencia, Spain. May 2001. Bastiaanssen, W. G. M. (1998). Remote sensing in water resources management: the state of the art. Colombo, IWMI.
189
Berger, A. R. (1997). Assessing rapid environmental change using geo-indicators, Environmental Geology, 1, 36-44. Bossard, M., Feranec, J. and Otahel, J., (2000). CORINE Land Cover Technical Guide – Addendum
2000.
Technical
report
No
40.
Copenhagen
(EEA).
http://terrestrial.eionet.eu.int. Bowonder, B., Prassad, S. S. R and Reddy, R (1988). Project Sitting Countries, Project Appraisal 2 (1): 11-20 Brogaard, S. and Prieler S. (1998). Land cover in the Horqin grasslands, North China. Detecting changes between 1975 and 1990 by means of remote sensing. Interim report IR-98-044/July on work of the International Institute for Applied System Analysis. Carver, S. J. (1991). Integrating Multi-Criteria Evaluation with Geographical Information Systems. International Journal of Geographical Information Systems. 5:321-329. Chen, H. M., Arora, M. K., and Varshney, P. K. (2003). Mutual information-based image registration for remote sensing data. International Journal of Remote Sensing, 24, 3701-3706. Chernoff, H.; Lehmann E.L. (1954). "The use of maximum likelihood estimates in
tests
for goodness-of-fit". The Annals of Mathematical Statistics 25 (3): 579–586. doi:10.1214/aoms/1177728726. Choubey, V. K. (1996). Assessment of waterlogged area in IGNP Stage I by remotely sensed and field data. Hydrology Journal, Vol. XIX (2), pp. 81-93. Choudhri, M. B., Mian, M. A. and Rafiq, M. (1978). The nature and magnitude of salinity and drainage problems in relation to agricultural development in Pakistan. Pakistan Soils Bulletin, 8. Lahore, Pakistan.
190
Cialella, A. T., Dubayah, R., Lawrence, W. and Levine, E. (1997). Predicting soil-drainage class using remotely sensed and digital elevation data. Photogrammetric Engineering and Remote Sensing, 63(2): 171-178. Cohen, W. B., & Goward, S. N. (2004). Landsat's role in ecological applications of Remote Coker, A. E. (1997). The application of remote sensing technology to assess the effects of and monitor change in coal mining in Eastern Tennessee. Proceedings of the First Annual William Symposium, Falls Church, pp. 95-105. Coppin, P. and Bauer, M. (1996). Digital Change Detection in Forest Ecosystems with Remote Sensing Imagery. Remote Sensing Reviews. Vol. 13. p. 207-234. CSIRO (2009). Wireless sensor networks: a new instrument for observing our world. www.csiro.au/science/Sensors-and-network-technologies.html. Accessed on 26th May, 2009. De Villiers, M. C., Pretorius, D. J., Barnard, R. O., Van Zyl, A. J. and Le Clus, C. F. (2002). Land degradation assessment in dryland areas: South Africa. FAO Land Degradation Assessment in Dryland Project. Rome, Italy. de Vries, C., Danaher, T., Denham, R., Scarth, P., Phinn, S., and Miller, E. H. (2007). An operational radiometric calibration procedure for the Landsat sensors based on pseudo-invariant target sites. Remote Sensing of Environment, 7 (3), 414–429. Dean, A., Hughes, S., Gibbons, W., Syed, A., Renou, C., Dow, D., Mangin, A., and Boivin, T. (2007). Monitoring sustainable development in the oil sands region of Alberta, Canada,
using
SPOT-5
and
ENVISAT
ASAR
and
MERIS
http://www.eomd.esa.int/files/docs/131-176-149-30-2007101113598.pdf,
imagery, accessed
05-12-2009. Diamant, B.Z. (1980).
Environmental Repercussions on Irrigation Development in Hot
Climates. Environmental Conservation 7(1) 53 – 58. 191
Dimyati, E. (1995). An Analysis of Land Use/Land Cover Change Using the Combination of MSS Landsat and Land Use Map- A case study of Yogyakarta, Indonesia, International Journal of Remote Sensing 17(5): 931 – 944. Draeger, W.C. (1976). Monitoring irrigated level acreage using Landsat imagery: an application example. Open-file report No. 76-630. U.S. Geological Survey. Pp. 23. Du Plessis, H. M., (1991). Researching and applying methods to conserve natural irrigation resources. Proceedings of South Africa Irrigation Symposium 61-69. June 1991. Durban, South Africa. Dwivedi, R. S. (1996). Monitoring of salt affected soils of the Indo-Gangetic alluvial plains using
principal
component
analysis.
International
Journal
of
Remote
Sensing.17:1907- 1914. Dwivedi, R. S. and Sreenivas, K. (1998). Image transforms as a tool for the study of soil salinity and alkalinity dynamics. International Journal of Remote Sensing, 19(4), 605619. Dwivedi, R.S. (1996). Monitoring of salt-affected soils of the Indo-Gangetic alluvial plains using principal component analysis. International Journal of Remote Sensing, 17(10): 1 907-1 914. Dwivedi, R.S. and Rao, B.R.M. (1992). The selection of the best possible Landsat TM band combination for delineating salt-affected soils. International Journal of Remote Sensing, 13 (11), 2 051-2 058. Ellington R. G., Usher B. H. and Van-Tonder G. J. (2004). Quantification of the impact of irrigation on the groundwater resource in the Vaal harts Irrigation Scheme. WRC Report No. 1322/2/04. Water Research Commission, Pretoria. South Africa. EOSAT (1992). Landsat TM Classification International Georgia Wetlands in EOSAT Data User Notes, Vol. 7, No 1, EOSAT Company, Lanham, MD. 192
ERDAS (1992). ERDAS Production Services Map State for Georgia DNR in the Monitor, Vol. 4, No 1, ERDAS, Inc, Atlanta, GA. Ernani, M. Z. and Gabriels, D. (2006). Detection of land cover changes using Landsat MSS, TM, ETM+ sensors in YHazd-Ardakan basin, Iran, Proceedings of Agro Environ 2006. ESRI (1998). Introduction to ARC View GIS. Environmental system research institute. Inc. USA. Fadem, B. (2008). High-Yield Behavioral Science (High-Yield Series). Hagerstwon, MD: Lippincott Williams & Wilkins. ISBN 0-7817-8258-9. FAO (1976). A Framework for Land Evaluation, Soils Bulletin 32. Food and Agricultural Organization, Rome. FAO (2004). Environmental impact assessment of irrigation and drainage projects. Food and Agriculture Organization, Rome. FAQS (2011). http://www.faqs.org/periodicals/201104/2343235481.html#ixzz1PnmmdGUq, accessed on the 24th March, 2011. Fisher Box, J. (1987). "Guinness, Gosset, Fisher, and Small Samples". Statistical Science 2 (1): 45–52. doi:10.1214/ss/1177013437. JSTOR 2245613. Fritz, M. (2009). Personal communication; Agrometeorolgy material. Agricultural Research Council, Institute for Soil, Climate and Water, Potchefstroom. South Africa. Ghassemi, F., Jakeman A. J., Nix H. A. (1995). Salinisation of land and water resources: Human causes, extent, management and case studies. UNSW Press, Sydney, Australia, and CAB International, Wallingford, UK. Glasson, J., Therival, R., Chadwick, A., (1999). Introduction to Environmental Impact Assessment: Principle and Process. Practice and prospects. 2nd Ed. UCL, London. Pg. 3-24. 193
GLCF (2011). Global Land Cover Facility (GLCF) of the University of Maryland, Maryland, USA www.glcf.umiacs.umd.edu accessed on 23rd Sept., 2011. Goosens, R., Joshi, D.C and Ranst, E.V. (1998). The use of remote sensing to map gypsiferous soils in the Ismailia soil (Egypt). Geoderma, 87:47-56. Goossens, E. M. C., Goossens, R. E. A., Ghabour, T. K. and Gad, A. G. (1993). Detailed assessment of degradation processes in the agricultural land of a newly reclaimed area in Egypt, using multi-temporal high resolution satellite data. EARSeL Advances in Remote Sensing, 2(3): 91-101. Goossens, R. E. A., El Badawi, M., Ghabour, T.K. and de Dapper, M. (1993). A simulation model to monitor the soil salinity in irrigated arable based upon remote sensing and GIS. EARSeL Advances in Remote Sensing, 2(3): 165-171. Goudie, A. S. (2003). Enhanced Salinisation: In Water Resources Perspectives: Evaluation, Management, and Policy, edited by A. Alsharhan and W. W. Wood, pp. 287–294. Amsterdam, London, and Boston: Elsevier. Goudie, A. S., and Viles, H. A. (1997). Salt Weathering Hazards: An analysis of salt attack on buildings in a wide range of environments. Chichester, U.K., and New York: Wiley. Peck, A.J. “Salinization of Non-Irrigated Soils and Associated Streams: A Review.” Australian Journal of Soil Research 16 (1978), 157–168. DOI: 10.1071/SR9780157. Gould, W. (2000). Remote sensing of vegetation, plant species richness, and regional biodiversity hotspots. Ecological applications, 10 (6), pp. 1861-1870. Graham, W. R. and B. R. Singh (1997). Soil and water quality under large-scale irrigation in a semi-arid ecosystem. The Bakolori Irrigation Project area, Zamfara State. In: B.R. Singh (Ed.) Management of Marginal Lands in Nigeria, pp. 219-226.
194
Greenwood, P.E., Nikulin, M.S.(1996). A guide to chi-squared testing, J.Wiley, New York, ISBN 0-471-55779-X. Grove, B. (2006). Generalised Whole-farm stochastic dynamic programming model to optimise agricultural water use. Report to the Water Research Commission. Pretoria, Water Research Commission Gu, D. and Gillespie, A., 1998. Topographic normalization of Landsat TM images of a forest based on subpixel sun canopy- sensor geometry. Remote Sensing of Environment, 64 (2), pp. 166-175. Haboudane, D., Bonn, F., Royer, A., Sommer, S., and Mehl, W. (2002). Land degradation and erosion risk mapping by fusion of spectrally based information and digital geomorphometric attributes. International Journal of Remote Sensing, 23, 3795-3820. Han, T., Wulder, M., White, J., Coops, N., Alvarez, M., & Butson, C. 2007. An Efficient Protocol to Process Landsat Images for Change Detection With Tasseled Cap Transformation. Geoscience and Remote Sensing Letters, IEEE, 4 (1), 147-151. Hardisk, M. A., Klemas, V. and Daiber, F. C. (1983). Remote sensing salt marsh biomass and stress detection. Advances in Space Research, 2: 219-229. Herold, C. E. and Bailey, A. K. (1996). Long term salt balance of the Vaal harts Irrigation Scheme. WRC Report No. 420/1/96. Water Research Commission, Pretoria. South Africa. Hough, J. J. H. and Rudolph, D. C. (2003). Vaal harts Groundwater Protocol for on-site sanitation. Report by VKE Engineers, by GHT. GHT Consulting Report No.: RVN 331.1/482. South Africa. IDRISI (2009). 32 guides to GIS and Image processing, Volume 1, Release 2. Pp. 17. IGNP Status Report (2001). Monitoring of Water Table IGNP Command Stage I. CAD Groundwater Department, Government of Rajasthan. 195
ILRI (1988). Effectiveness and Social/Environmental Impacts of Irrigation Projects: a Review. In: Annual Report 1988, International Institute for Land Reclamation and Improvement (ILRI), Wageningen, The Netherlands, pp. 18 - 34 ILRI (2003). Lecture notes on salinity. International Course on Land Drainage International Institute for Land Reclamation and Improvement (ILRI), Wageningen, The Netherlands. www.waterlog.info. Accessed on 22nd March, 2012. Indo-Dutch Network Project (IDNP) (2002). A Methodology for Identification of Water logging and Soil Salinity Conditions Using Remote Sensing. CSSRI, Kamal and Alterra-ILRI, Wageningen. pp.78. Jager, J. M. (1994). Accuracy of vegetation evaporation formulae for estimating final wheat yield. Water SA, 20, 307-314. Jensen, M. E (1993). The Impacts of Irrigation and Drainage on the Environment ICID, 5 th N.D Gulhati Memorial Lecture, the Hague, Netherlands pp 26. Johnston, R. M. and Barson, M.M. (1993). Remote-sensing of Australian wetlands. An evaluation of Landsat TM data for inventory and classification. Australian Journal of Marine and Freshwater Research. 44:235−242. Johnstone, R. M. and Barson, M. M. (1990). An assessment of the use of remote sensing techniques in land degradation studies. Bulletin 5. Canberra, Australian Department of Primary Industries and Energy. 64 pp. Kalra, N. K. and Joshi, D. C. (1997). Evaluation of multi-sensor data for delineating saltaffected soils in arid-Rajasthan. Journal of the Indian Society of Remote Sensing. 25:79-91. Kauth, R. Y. and Thomas, G. S. (1976). The tasseled cap-A Graphic Description of the spectral- Temporal Development of Agricultural Crops as seen by Landsat. In: proceeding of the symposium on Machine processing of Remotely Sensed Data. 196
Purdure University of West Lafayette, Indiana, pp. 41-49. Kenneth, M. and Timothy, M. (2000). "Hyper spectral Mixture Modeling for Quantifying Sparse Vegetation Cover in Arid Environments." Remote Sensing of Environment 72: 360-374. Latifovic, R., Fytas, K., Chen, J. and Paraszczak, J. (2005). Assessing land cover change resulting from large surface mining development, International Journal of Applied Earth Observation and Geoinformation, 7, 29-48. Lillesand, T. M. and Kiefer, R. W. (1994). Remote Sensing and Image Interpretation (3rd Ed.). John Wiley & Sons, Inc., New York. Limpitlaw, D. and Woldai, T. (1998). Environmental impact assessment of mining by integration of remotely sensed data with historical data, Mine Environmental Engineering Conference 11, Brunel University, Ox Bridge, UK, 29-30 July. Macleod, S. and Congalton, K. (1998). A Quantitative Comparison of Change Detection Algorithms for Monitoring Eelgrass from Remotely Sensed Data. Photogrammetric Engineering and Remote Sensing. Vol. 64. No.: 3. p. 207 - 216. Manu, A., Twumasi, Y.A. and Coleman, T.L. (2004). Application of Remote Sensing and GIS technologies to assess the impact of surface mining at Tarkwa, Ghana. Proceedings of IGARSS’2004 Symposium, Honolulu, Hawaii, July 24-28. Matternicht, G. and Zinck, J. A. (2003). Remote sensing of soil salinity: potentials and constraints. Remote sensing of Environment, 5812:1-10. Menenti, M., Lorkeers, A. and Vissers, M. (1986). An application of thematic mapper data in Tunisia. ITC Journal, 1: 35-42. Metternicht, G. and Zinck, J. A. (1997). Spatial discrimination of salt and sodium-affected soil surfaces. Int. J. R.S., 18 (12): 2 571-2 586. Metternicht, G. and Zink, J. A. (1996). Spatial discrimination of salt and sodium affected soil 197
surfaces. International Journal of Remote Sensing. 18:2571-2576. MMSD (2001). Meeting report of the Finance, Mining and Sustainability Conference, Mining, Minerals and Sustainable Development (MMSD), workshop on Sustainable Development, Washington DC. MMSD-SA (2001). Locating the mining and minerals sector within the southern African vision for Sustainable development, Mining, Minerals and Sustainable DevelopmentSouthern Africa (MMSD-SA), workshop on Sustainable Development, Stellenbosch. Moshen, A. (1999). Environmental Land Use Change Detection and Assessment Using with Multi – temporal Satellite Imagery. Zanjan University. Mulders, M. A. (1987). Remote sensing in soil science. Developments in Soil Science 15, Elsevier Publication, Amsterdam. Mulders, M. A. and Epema, G. F. (1986). The thematic mapper: a new tool for soil mapping in arid areas. ITC Journal, 1: 24-29. Muller, E. and Mangold, S. (2002). State of the Environment Report, North West Province, South Africa, www.nmpg.gov.za/soer/FullReport/air.html, accessed 30/08/2009. Ojo, O. I., Ochieng, G. M. and Otieno, F. A. O. (2011). Assessment of water logging and salinity problems in South Africa: an overview of Vaal Harts irrigation scheme. WIT Transactions on Ecology and the Environment, Volume 153, 2011. WIT Press, ISSN 1743-3541. Ojo, O. I., Otieno, F. A. O. and Ochieng G. M., (2009). Irrigation Problems and Research needs in South Africa: A review of the Vaal Harts Irrigation Scheme. World Academy of Science, Engineering and Technology WCSET Conference, September 23-25, 2009. Amsterdam, the Netherlands ISSN 2070-3740. Okin, G. S., and Roberts, D. A. (2001). Practical limits on hyper spectral vegetation discrimination in arid and semiarid environments. Remote Sensing of Environment 198
77(2): 212-225. Olorunfemi, J. F. (1983). Monitoring urban land – use in Developed Countries – An aerial photographic approach, Environmental Int.9, 27 – 32. Plackett, R.L. (1983). "Karl Pearson and the Chi-Squared Test". International Statistical Review (International Statistical Institute (ISI)) 51 (1): 59–72. doi:10.2307/1402731. JSTOR 1402731. PONS, X. and SOLÉ-SUGRAÑES, L., 1994, A simple radiometric correction model to improve automatic mapping of vegetation from multispectral satellite data. Remote Sensing of Environment, 48, 191 -204. Price, J. C. (1994). "How unique are spectral signatures?" Remote Sensing of Environment 49: 181-186. Raju, TN (2005). "William Sealy Gosset and William A. Silverman: two "students" of science". Pediatrics 116 (3): 732–5. doi:10.1542/peds.2005-1134. PMID 16140715 Rao, B. R. and Venkataraman, L. (1991). Mapping the magnitude of sodicity in part of the Indo-Gangetic plains of Uttar Pradesh, northern India using Landsat data, International Journal of Remote Sensing. 12:419-425. Rao, B. R., Dwivedi, R. S., Venkataratnam, L., Ravishankar, T., Thammappa, S. S., Bhargava, G. P. and Singh, A. N. (1991). Mapping the magnitude of sodicity in part of the Indo-Gangetic plains of Uttar Pradesh, northern India using Landsat data. International Journal of Remote Sensing, 12: 419-425. Rao, M., Sastry, S.V.C., Yadar, P. D., Kharod, K., Pathan, S. K., Dhinwa, P. S., Majumdar, K. L., Sampat Kumar, D., Patkar, V. N. and Phatak, V. K. (1997). A Weighted Index Model for Urban Suitability Assessment-A GIS Approach. Bombay Metropolitan Regional Development Authority, Bombay.
199
Rathore, C. S. and Wright, R. (1993). Monitoring environmental impacts of surface coal mining, International Journal of Remote sensing, 14, 1021-1042. Rhoades, J. D. (1990). Soil Salinity: Causes and Controls: In Techniques for Desert Reclamation, edited by A. S. Goudie, pp. 109–134. Chichester, U.K., and New York: Wiley, 1990. Richards, L.A. (2010). Diagnosis and improvement of saline and alkali soils. USDA Agricultural
Handbook
60.
http://www.ars.usda.gov/Services/docs.htm?docid=10158&page=2. Accessed on 22nd March, 2012. Richardson, A. J., Gerbermann, A. H., Gausmann, H. W. and Cuellar, J. A. (1976). Detection of saline soils with Skylab multi-spectral scanner data. Photogrammetric Engineering and Remote Sensing. 42:679-684. Rodgers, J. L. and Nicewander W. A. (1988). Thirteen ways to look at the correlation coefficient, The American Statistician, 42 (1): 59–66, February 1988. Rozanov, B. G., V. Targulian, and D. S. Orlov. (1991). Soils: In the Earth as Transformed by Human Action, edited by B. Turner M. Pp. 203–214. Cambridge and New York: Cambridge University Press, 1991. Saaty, T. L. (1977). A Scaling Method for Priorities in Hierarchical Structures, J. Math. Psychology. 15:234-241. Sahin, H., Oruc, M. and Buyuksalih, G. (2006), Temporal analysis of multi epoch Landsat geocover images in Zonguldak test field, Proceedings of ISPRS Ankara Workshop, 7th session 16 February, 2006. Sahin, H., Oruc, M. and Buyuksalih, G. (2006). Temporal analysis of multi epoch Landsat geocover images in Zonguldak test field, Proceedings of ISPRS Ankara Workshop, 7th session 16 February, 2006. 200
Salman, A. (2000). Using the State-of-the-art Remote Sensing and GIS for monitoring water logging and Salinity, International Water-Management Institute, Lahore, Pakistan. Sandham, L. A., Moloto, M. J. and Retief, F. P. (2008). The quality of environmental impact reports for projects with the potentials of affecting wetlands in South Africa. Water SA Vol. 34 No.2 Pg 155-162. Sarma, P. V. and Murthy, K. L. N. (2004). Environment and Agriculture. A Case Study of Krishna Delta Region in Andhra Pradesh. Pp 95 – 115. Schmidt, H. and Glaesser, C. (1998), Multitemporal analysis of satellite data and their use in the monitoring of the environmental impacts of open cast lignite mining areas in Eastern Germany, International Journal of Remote Sensing, 24, no 16:3311-3340. Schmidt, H. and Karnieli, A., (2002). Remote sensing of the seasonal variability of vegetation in a semi-arid environment. Journal of Arid Environments, 45 (1), pp. 4360. Schroeder, T. A., Cohen, W. B., Song, C., Canty, M. J., & Yang, Z. (2006). Radiometric calibration of Landsat data for characterization of early successional forest patterns in western Oregon. Remote Sensing of Environment, 103, 16−26. Schutte, I. C. (1994). The Geology of the Christiana Area. Explanation of Sheet 2724 Scale 1:250 000. Council for Geosciences, Pretoria. South Africa. Sensing. Bioscience, 54 (6), 535−545. Serra, P., Pons, X. and Sauri, D. (2003). Post-classification change detection with data from different sensors: some accuracy considerations, International Journal of Remote Sensing, 19, 2245-2260. Shoshany, M, and Kutiel S. (1994). Monitoring Temporal Vegetation Cover Changes in Mediterranean and Arid Ecosystems Using a Remote Sensing Technique: case study 201
of the Judean Mountain and the Judean Desert. Journal of Arid Environments, 33: 9 – 21. Siderius, W. (1991). The use of remote sensing for irrigation management with emphasis on IIMI research concerning salinity, water logging and cropping patterns. Mission Report. Enschede, Netherlands, International Institute for Aerospace Survey and Earth Sciences. 97 pp. Singh, A. N. and Dwivedi, R. S. (1989). Delineation of salt-affected soils through digital analysis of Landsat MSS data. International Journal of Remote Sensing, 10: 83-92. Singh, B. R. (1995). Soil Management Strategies for the Semi-arid Ecosystem in Nigeria. The Case of Sokoto and Kebbi States. African Soils. 28: 317 – 320. Singh, R. P. and Srivastav, S. K. (1990). Mapping waterlogged and salt affected soils using microwave radiometers. International Journal of Remote Sensing, 11: 1 879-1 887. Stefanov, W. L., Ramsey, M. S. and Christensen, P. R. (2001), Monitoring urban land cover change: An expert system approach to land cover classification of semiarid to arid urban centers, Remote Sensing of Environment, 77, 173-185. Stefanov, W. L., Ramsey, M. S. and Christensen, P. R. (2001). Monitoring urban land cover change: An expert system approach to land cover classification of semiarid to arid urban centers, Remote Sensing of Environment, 77, 173-185. Steven, F. G, Zagrafos, C.and Oglethorpe, D. (1992). Multi-Criteria Analysis In Soil Salinity: Using Goal Programming To Explore Solutions. Current Issues in Soil Conservation, Retrieved
on
13.6.2009
from
http://www.multilingualmatters.net/cit/007/0020/cit05200350.pdf Steven, M. D., Malthus, T. J., Jaggard, F. M. and Andrieu, B. (1992). Monitoring responses of vegetation to stress. In A.P. Cracknell & R.A. Vaughan, Eds. Remote sensing from
202
research to operation: proceedings of the 18th annual conference of the Remote Sensing Society, p. 369-377. Nottingham, U.K., Stigler, S. M. (1989). Francis Galton's Account of the Invention of Correlation. Statistical Science 4 (2): 73–79. Doi: 10.1214/ss/1177012580. JSTOR 2245329. Szabolcs, I. (1979). Review of research on salt-affected soils. Paris, UNESCO. Szabolcs, I. (1987). Salt affected soils. Photographic and Videographic observations for determining and mapping the response of cotton to soil salinity, Remote Sensing of Environment, CRC Press, Boca Raton, FL, 49:212-213. Szabolcs, I. (1994). State and Perspectives on Soil Salinity in Europe. European Society for Soil Conservation Newsletter 3 (1994), 17–24. Tabet, D. (1995). Le suivi de la salinité dans les périmètres irrigués par télédétection et systèmes d'information géographique: application au périmèter de Fordwah, Pakistan. Montpellier, France, DEA National d'Hydrologie. 174 pp. Tabet, D. (1999). Intérêt d'une approche spatiale pour suivi de la salinité des sols dans les systèmes irrigués: cas de la subdivision de Chistian dans le Punjab, Pakistan. Ph.D. Thesis. Montpellier, France, Ecole Nationale du Génie rural, des Eaux et Forêts. Teillet, P. M., Helder, D. L., Ruggles, T. A., Landry, R., Ahern, F. J., Higgs, N. J., et al. (2004). A definitive calibration record for the Landsat-5 Thematic Mapper anchored to the Landsat-7 radiometric scale. Canadian Journal of Remote Sensing, 30(4), 631−643. Thomas, D. S. G., and N. J. Middleton. (1993). Salinization: New Perspectives on a Major Desertification Issue. Journal of Arid Environments 24.1 (1993), 95–105. Tripathi, N. K., B. K. Rai, and P. Dwivedi (1997). Spatial modelling of soil alkalinity in GIS environment using IRS data. Proceedings of 18th Asian Conference on Remote Sensing 20- 24 Oct.1997, Kualalampur. 203
UNEP (2002). Finance, Mining and Sustainability. United Nations Environment Programme (UNEP).
http:www.natural_resources.org/minerals/CD/docs/unep/finance_wshop-
report.pdf, accessed 05-10-09. USGS (1999). Land Subsidence in the United States. United State Geological Survey. Online: http://water.usgs.gov/ogw/pubs/fs00165/. Accessed 11/10/2011. USGS (1999). The Landsat Satellite System Link, U.S. Geological Survey on the World Wide Web. URL: http://landsat7.usgs.gov/landsat_sat.html. Accessed 11/10/2011. Van der Merwe, A. J., (1995). Wise land use: The basis for sustainable growth and development in South Africa, Proceeding of ARC-ISCW Wise Land Use Symp. 2-8, Pretoria, South Africa. Van Niekerk C. S. (2009). Personal communication. Northern Cape Provincial Department of Agriculture, Jan Kempdorp. South Africa. Van Trinh, M., Duong, N. D. and Van Keulen, H. (2004). Using Landsat images for studying land use dynamics and soil degradation: Case study in Tamduong District, Vinhphuc Province, Vietnam, International Symposium on Geoinformatics for Spatial Infrastructure Development in Earth and Allied Sciences 2004. Van Vuuren, E. (2004). Proceedings of the 2004 Water Institute of Southern Africa (WISA) Biennial Conference 2 .6 May 2004, ISBN: 1-920-01728-3 Cape Town, South Africa. Venkataratnam, L. (1983). Monitoring of soil salinity in the Indo-Gangetic plain of NW India using multidate Landsat data. In Proceedings of the 17th international symposium on remote sensing of the environment, Vol. 1, p. 369-377. Ann Arbor, Michigan, USA, Environmental Research Institute of Michigan. Verma, K. S., Saxena, R. K., Barthwal, A. K. and Deshmukh, S.N. (1994). Remote-sensing technique for mapping salt affected soils. International Journal of Remote Sensing, 15(9), 1 901-1 914. 204
Vidal, V., Tabet, D., Ahmad., M.D., Asif, S., Zimmer, D. and Strosser, P. (1998). Salinity assessment in irrigation systems using remote sensing and geographical information systems - application to Chistian subdivision, Pakistan. Report R-47. Lahore, Pakistan, IIMI. Vincent, B., Frejefond, E., Vidal, A. and Baqri, A. (1995). Drainage performance assessment using remote sensing in the Gharb plain, Morocco. In A. Vidal & J.A. Sagardoy, eds. Water report No. 4: use of remote sensing techniques in irrigation and drainage, p. 155-164. Rome, FAO. Vincent, B., Vidal, A., Tabbet, A.B. and Kuper, M. (1996). Use of satellite remote sensing for the assessment of water logging or salinity. In Evaluation of performance of subsurface drainage systems: 16th congress on Irrigation and Drainage, Cairo, Egypt, 15-22 September, 1996. Ed. Vincent, B, 203-216. New Delhi: International Commission on Irrigation and Drainage. Voogd, H. (1983). Multi-criteria Evaluation for Urban and Regional Planning. Pion, Ltd., London. WAPDA (1984). Applicability of Landsat imagery for monitoring soil salinity trends. Publ. 529. Lahore, Pakistan, Directorate of Reclamation and Investigation Organization. Worthington, E. B. (1977). Arid Land Irrigation in Developing Countries: Environmental Problems and Effects. Oxford and New York: Pergamon. Xiaomei, Y. and Ronqing, L. Q. Y. (1999). Change Detection Based on Remote Sensing Information Model and its Application to Coastal Line of Yellow River Delta – Earth Observation Center, NASDA, China. Zhang, C., Guo, X., Wilmshurst, J., and Sissons, R. (2006). Application of Radarsat imagery on grassland heterogeneity assessment. Canadian Journal of Remote Sensing, 32 (4), pp. 1-7. 205
Zimmerman, Donald W. (1997). "A Note on Interpretation of the Paired-Samples t Test". Journal of Educational and Behavioral Statistics 22 (3): 349–360. JSTOR 1165289. Zuluaga, J.M. (1990). Remote sensing applications in irrigation management in Mendoza, Argentina, In M. Menenti, Ed. Remote sensing in evaluation and management of irrigation, p. 37-58. Mendoza, Argentina, Instituto Nacional de Ciencia y Tecnic.
206
Appendix A: findings from the structured questionnaires Table 1a: Fertility rate VS Irrigation practise Do you Irrigate Yes How would you rate the Fairly Better fertility Worse
No
40.6%
9.4%
25.0%
Total 50.0% 25.0%
About The Same
21.9%
3.1%
25.0%
Total
87.5%
12.5%
100.0%
Table 1b: Chi-Square Test for Fertility rate VS Irrigation practise Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
1.714a
2
.424
Likelihood Ratio
2.642
2
.267
Linear-by-Linear Association
.403
1
.526
N of Valid Cases
32
207
Table 2a: Fertility rate VS Irrigation Method Irrigation Method
Sprinkler How would you rate the Fairly Better fertility Worse About The Same Total
Furrow
34.4%
Flood Irrigation
3.1%
Sprinkler and Pivot
3.1%
15.6%
6.3%
3.1%
12.5%
53.1%
15.6%
3.1%
3.1%
9.4%
Table 2a contd: Fertility rate VS Irrigation Method
Irrigation Method Furrow and Flood Irrigation How would you rate the Fairly Better fertility Worse
6.3%
About Tthe Same Total
6.3%
208
Furrow and Surface
Total
3.1%
50.0%
3.1%
25.0%
6.3%
25.0%
12.5%
100.0%
Table 2b: Chi-Square Tests Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
19.720
10
.032
Likelihood Ratio
21.920
10
.016
1.888
1
.169
Linear-by-Linear Association N of Valid Cases
32
Table 3a: Fertility rate VS Source of Water Source of Water Dam How would you rate the Fairly Better fertility Worse
Pivot
50.0% 21.9%
About The Same
25.0%
Total
96.9%
209
Total 50.0%
3.1%
25.0% 25.0%
3.1%
100.0%
Table 3b: Chi-Square Tests Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
3.097
2
.213
Likelihood Ratio
2.872
2
.238
Linear-by-Linear Association
.091
1
.763
N of Valid Cases
32
Table 4a: Purpose of irrigation VS Irrigate practise Do you Irrigate Yes If No Why did you change
Higher income
4.9%
New crops are in higher demand
4.9%
Low viability
2.4%
No
Total 4.9%
4.9%
9.8%
2.4%
No change
70.7%
12.2%
82.9%
Total
82.9%
17.1%
100.0%
210
Table 4b: Chi-Square Tests Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
3.815
3
.282
Likelihood Ratio
3.537
3
.316
Linear-by-Linear Association
.484
1
.487
N of Valid Cases
41
Table 5a: Reason for not practising irrigation VS Irrigation method Irrigation Method Sprinkler If No why did you change
New crops are in higher demand
Furrow
Flood Irrigation
2.4%
2.4%
No Change
48.8%
9.8%
2.4%
Total
51.2%
12.2%
2.4%
211
Table 5a contd: Reason for not practising irrigation VS Irrigation method
Irrigation Method
Sprinkler and Pivot If No Why did you change
Furrow and Flood Irrigation
Higher Income
2.4%
2.4%
New crops are in higher demand
2.4%
2.4%
Low Viability
Furrow and Surface
2.4%
No Change
2.4%
Total
7.3%
14.6% 7.3%
14.6%
Table 5a contd: Reason for not practising irrigation VS Irrigation method Irrigation Method Basin and Surface If No Why did you change
Surface
Total
Higher Income
4.9%
New crops are in higher demand
9.8%
Low Viability
2.4%
No Change
2.4%
2.4%
82.9%
Total
2.4%
2.4%
100.0%
212
Table 5b: Chi-Square Tests Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
33.788
21
.038
Likelihood Ratio
24.629
21
.264
Linear-by-Linear Association
1.561
1
.212
N of Valid Cases
41
Table 6a: Reason for not practising irrigation VS Source of Water Source of Water Dam If No Why did you change
Pivot
Total
Higher Income
4.9%
4.9%
New crops are in higher demand
9.8%
9.8%
Low Viability
2.4%
2.4%
No Change
80.5%
2.4%
82.9%
Total
97.6%
2.4%
100.0%
213
Table 6b: Chi-Square Tests Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
.211
3
.976
Likelihood Ratio
.380
3
.944
Linear-by-Linear Association
.186
1
.666
N of Valid Cases
41
Table 7a: Yield VS Irrigation Do you Irrigate Yes How Would You Assess Good your Yield Fair
No
Total
4.9%
2.4%
7.3%
19.5%
4.9%
24.4%
Poor
17.1%
4.9%
22.0%
No Difference
41.5%
4.9%
46.3%
Total
82.9%
17.1%
100.0%
214
Table 7b: Chi-Square Tests Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
1.364
3
.714
Likelihood Ratio
1.329
3
.722
Linear-by-Linear Association
1.066
1
.302
N of Valid Cases
41
215
Table 8a: Yield VS Irrigation method Irrigation Method
Sprinkler How Would You Assess Good your Yield Fair
Flood Irrigation Sprinkler and Pivot
Furrow
2.4% 12.2%
2.4% 7.3%
Poor
12.2%
No Difference
24.4%
4.9%
Total
51.2%
12.2%
2.4% 4.9%
2.4%
7.3%
Table 8a contd: Yield VS Irrigation method
Irrigation Method Furrow and Flood Irrigation How Would You Assess Good your Yield Fair
Furrow and Surface
Basin and Surface
Surface
2.4% 2.4%
Poor
4.9%
No Difference
2.4%
9.8%
2.4%
2.4%
Total
7.3%
14.6%
2.4%
2.4%
216
Table 8a contd: Yield VS Irrigation method
Total How Would You Assess Good your Yield Fair
7.3% 24.4%
Poor
22.0%
No Difference
46.3%
Total
100.0%
Table 8b: Chi-Square Tests Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
30.193
21
.088
Likelihood Ratio
26.022
21
.206
Linear-by-Linear Association
.697
1
.404
N of Valid Cases
41
217
Table 9a: Yield VS Source of Water Source of Water Dam How Would You Assess Good your Yield Fair
Pivot
Total
7.3%
7.3%
24.4%
24.4%
Poor
19.5%
No Difference
46.3%
Total
97.6%
2.4%
22.0% 46.3%
2.4%
100.0%
Table 9b: Chi-Square Tests Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
3.644
3
.303
Likelihood Ratio
3.124
3
.373
Linear-by-Linear Association
.005
1
.942
N of Valid Cases
41
218
Table 10a: Cultivable land areas VS Irrigate practise Do you Irrigate Yes Are there Areas of Your Yes land that are not No Cultivable Total
No
Total
65.9%
2.4%
68.3%
17.1%
14.6%
31.7%
82.9%
17.1%
100.0%
Table 10b: Chi-Square Tests
Value Pearson Chi-Square Continuity Correctionb Likelihood Ratio
Asymp. Sig. Exact Sig. (2- Exact Sig. (1(2-sided) sided) sided)
df
11.370
1
.001
8.562
1
.003
10.905
1
.001
Fisher's Exact Test
.002
Linear-by-Linear Association
11.093
N of Valid Cases
41
1
219
.001
.002
Table 11a: Cultivable land areas VS Irrigate practise method Irrigation Method
Sprinkler Are there Areas of Your Yes land that are not No Cultivable Total
Furrow
41.5%
Flood Irrigation
Sprinkler and Pivot
12.2%
4.9%
9.8% 51.2%
12.2%
2.4%
2.4%
2.4%
7.3%
Table 11a contd: Cultivable land areas VS Irrigate practise method
Irrigation Method Furrow and Flood Irrigation Are there Areas of Your Yes land that are not No Cultivable Total
Furrow and Surface 4.9%
7.3%
9.8%
7.3%
14.6%
220
Basin and Surface 2.4%
2.4%
Table 11a contd: Cultivable land areas VS Irrigate practise method
Irrigation method Surface Are there Areas of Your Yes land that are not No Cultivable
Total 2.4%
68.3% 31.7%
Total
2.4%
100.0%
Table 11b: Chi-Square Tests Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
16.810
7
.019
Likelihood Ratio
19.313
7
.007
Linear-by-Linear Association
4.824
1
.028
N of Valid Cases
41
221
Table 12a: Cultivable land areas VS Source of water Source of Water Dam Are there Areas of Your Yes land that are not No Cultivable
Pivot
65.9%
Total 2.4%
68.3%
31.7%
Total
97.6%
31.7% 2.4%
100.0%
Table 12b: Chi-Square Tests
Value
Asymp. Sig. Exact Sig. (2(2-sided) sided)
df
Pearson Chi-Square
.476
1
.490
Continuity Correctionb
.000
1
1.000
Likelihood Ratio
.774
1
.379
Fisher's Exact Test
1.000
Linear-by-Linear Association
.464
N of Valid Cases
41
1
222
.496
Exact Sig. (1-sided)
.683
Table 13a: Factors affecting Yield VS Irrigate practise Do you Irrigate Yes Which of the Following Poor Soil Condition has affected Your Yield Poor soil and Loss of soil Due to Flood Everything
No
22.0% 17.1%
Total 22.0%
4.9%
2.4%
22.0%
2.4%
Loss of soil Due to Rain and Poor water from Source
14.6%
2.4%
17.1%
Stagnant Water and Loss of Soil Due to Rain
12.2%
2.4%
14.6%
Poor Water From Source
14.6%
7.3%
22.0%
Total
82.9%
17.1%
100.0%
Table 13b: Chi-Square Tests
Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
3.947
5
.557
Likelihood Ratio
5.337
5
.376
Linear-by-Linear Association
2.122
1
.145
N of Valid Cases
41 223
Table 13b: Chi-Square Tests
Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
3.947
5
.557
Likelihood Ratio
5.337
5
.376
Linear-by-Linear Association
2.122
1
.145
N of Valid Cases
41
Table14a: Factors affecting Yield VS Irrigate method Irrigation Method Sprinkler Furrow Which of the Following Poor Soil Condition has affected Your Yield Poor soil and Loss of soil Due to Flood
19.5%
2.4%
7.3%
7.3%
Everything
2.4%
Loss of soil Due to Rain and Poor water from Source
7.3%
Stagnant Water and Loss of Soil Due to Rain
9.8%
Poor Water From Source
4.9%
2.4%
51.2%
12.2%
Total 224
Flood Irrigation
2.4%
2.4%
Table14a contd: Factors affecting Yield VS Irrigate method Irrigation Method
Sprinkler and Pivot Which of the Following Poor soil and Loss of has affected Your Yield soil Due to Flood Loss of soil Due to Rain and Poor water from Source
2.4%
4.9%
Poor Water From Source Total
7.3%
225
Furrow and Flood Furrow and Irrigation Surface 2.4%
4.9%
2.4%
12.2%
7.3%
14.6%
Table14a contd: Factors affecting Yield VS Irrigate method
Irrigation Method Basin and Surface
Surface
Which of the Following Poor Soil Condition has affected Your Yield Poor soil and Loss of soil Due to Flood
Total 22.0% 22.0%
Everything
2.4%
Loss of soil Due to Rain and Poor water from Source Stagnant Water and Loss of Soil Due to Rain
17.1%
2.4%
2.4%
Poor Water From Source
14.6%
22.0%
Total
2.4%
226
2.4%
100.0%
Table 14b: Chi-Square Tests Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
52.014
35
.032
Likelihood Ratio
47.019
35
.084
Linear-by-Linear Association
9.279
1
.002
N of Valid Cases
41
227
Table15a: Factors affecting Yield VS Source of water Source of Water Dam Which of the Following Poor Soil Condition has affected Your Yield Poor soil and Loss of soil Due to Flood
Pivot
Total
22.0% 19.5%
Everything
22.0% 2.4%
22.0%
2.4%
2.4%
Loss of soil Due to Rain and Poor water from Source
17.1%
17.1%
Stagnant Water and Loss of Soil Due to Rain
14.6%
14.6%
Poor Water From Source
22.0%
22.0%
Total
97.6%
2.4%
100.0%
Table15b: Chi-Square Tests
Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
3.644
5
.602
Likelihood Ratio
3.124
5
.681
Linear-by-Linear Association
.592
1
.441
N of Valid Cases
41 228
Table15b: Chi-Square Tests
Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
3.644
5
.602
Likelihood Ratio
3.124
5
.681
Linear-by-Linear Association
.592
1
.441
N of Valid Cases
41
Table 16a: Accumulated effect of Irrigation practise on the soil Do you Irrigate Yes If Yes State To ename
No
4.9%
Total 4.9%
Increase in Yellow nutadge
22.0%
2.4%
24.4%
White Patches
29.3%
4.9%
34.1%
Increased Salt
2.4%
2.4%
Worse
24.4%
9.8%
34.1%
Total
82.9%
17.1%
100.0%
229
Table16b: Chi-Square Tests
Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
2.355
4
.671
Likelihood Ratio
2.741
4
.602
Linear-by-Linear Association
1.953
1
.162
N of Valid Cases
41
Table 17a: Accumulated effect of Irrigation practise on the soil VS Irrigation Method Irrigation Method
Sprinkler If Yes State To ename
Furrow
Flood Irrigation
Sprinkler and Pivot
4.9%
Increase in Yellow nutadge
12.2%
4.9%
White Patches
17.1%
7.3%
Worse
17.1%
Total
51.2%
230
12.2%
2.4%
4.9%
2.4%
2.4%
7.3%
Table 17a contd: Accumulated effect of Irrigation practise on the soil VS Irrigation Method Irrigation Method Furrow and Flood Irrigation If Yes State White Patches
Furrow and Surface
4.9%
Basin and Surface
Surface
2.4%
Increased Salt
2.4%
Worse
2.4%
9.8%
2.4%
2.4%
Total
7.3%
14.6%
2.4%
2.4%
Table 17a contd: Accumulated effect of Irrigation practise on the soil VS Irrigation Method
Total If Yes State To ename
4.9%
Increase in Yellow nutadge
24.4%
White Patches
34.1%
Increased Salt
2.4%
Worse
34.1%
Total
100.0%
231
Table17b: Chi-Square Tests Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
26.416a
28
.550
Likelihood Ratio
28.605
28
.433
Linear-by-Linear Association
5.433
1
.020
N of Valid Cases
41
Table 18a: Accumulated effect of Irrigation practise on the soil VS Source of water Source of Water Dam If Yes State To ename
Pivot 4.9%
Total 4.9%
Increase in Yellow nutadge
22.0%
White Patches
34.1%
34.1%
Increased Salt
2.4%
2.4%
Worse
34.1%
34.1%
Total
97.6%
232
2.4%
2.4%
24.4%
100.0%
Table18b: Chi-Square Tests Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
3.177a
4
.529
Likelihood Ratio
2.901
4
.575
Linear-by-Linear Association
1.100
1
.294
N of Valid Cases
41
Table 19a: Health problem cases on the scheme VS Irrigation practise Do you Irrigate Yes Since The Scheme Worms Started, What Types Of Sinus Sickness
No
Total
19.5%
19.5%
2.4%
2.4%
Sinus and Worms
2.4%
Sinus and hearing
2.4%
2.4% 2.4%
Don't Know
58.5%
14.6%
73.2%
Total
82.9%
17.1%
100.0%
233
Table 19b: Chi-Square Tests Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
7.097a
4
.131
Likelihood Ratio
7.454
4
.114
Linear-by-Linear Association
1.300
1
.254
N of Valid Cases
41
Table 20a: Health problem cases on the scheme VS Irrigation method Irrigation Method
Sprinkler Since The Scheme Worms Started, What Types Of Sinus Sickness
Furrow
Flood Irrigation
Sprinkler and Pivot
19.5% 2.4%
Don't Know
31.7%
12.2%
2.4%
4.9%
Total
51.2%
12.2%
2.4%
7.3%
234
Table 20a: Health problem cases on the scheme VS Irrigation method Irrigation Method Furrow and Flood Irrigation Since The Scheme Sinus and Worms Started, What Types Of Sinus and hearing Sickness
Furrow and Surface
Basin and Surface
2.4% 2.4%
Don't Know
2.4%
14.6%
2.4%
Total
7.3%
14.6%
2.4%
Table 20a: Health problem cases on the scheme VS Irrigation method Irrigation Method Surface
Total
Since The Scheme Worms Started, What Types Of Sinus Sickness
19.5% 2.4%
Sinus and Worms
2.4%
Sinus and hearing
2.4%
Don't Know
2.4%
73.2%
Total
2.4%
100.0%
235
Table20b: Chi-Square Tests Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
48.029a
28
.011
Likelihood Ratio
28.849
28
.420
Linear-by-Linear Association
3.836
1
.050
N of Valid Cases
41
Table 21a: Health problem cases on the scheme VS Source of water Source of Water Dam Since The Scheme Worms Started, What Types Of Sinus Sickness
17.1%
Pivot 2.4%
Total 19.5%
2.4%
2.4%
Sinus and Worms
2.4%
2.4%
Sinus and hearing
2.4%
2.4%
Don't Know
73.2%
73.2%
Total
97.6%
236
2.4%
100.0%
Table 21b: Chi-Square Tests Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
4.228
4
.376
Likelihood Ratio
3.374
4
.497
Linear-by-Linear Association
3.626
1
.057
N of Valid Cases
41
237
Table 22a: Type grown presently VS Type grown before The Chi Square result
Were these the Crops you Were Planting Yes
What Type of crops do Corn, Wheat, Pecan nut Count you Grow % of Total Corn, Wheat, Pecan Nuts, Barely, Lucern
Count % of Total
Maize, Ground Nuts, Wheat, Barely, Pecan, Lucene
Count
Maize and wheat
Count
% of Total
% of Total Wheat, Maize and Peanuts
Count % of Total
Maize, Wheat and Ground Nuts
Count % of Total
Corn, Wheat and Ground Nuts
Count % of Total
Wheat, Ground Nuts, Pecon and Alfalfe
Count % of Total
Total
Count % of Total
238
No 6
1
14.6%
2.4%
4
2
9.8%
4.9%
2
2
4.9%
4.9%
7
0
17.1%
.0%
4
5
9.8%
12.2%
1
5
2.4%
12.2%
0
1
.0%
2.4%
0
1
.0%
2.4%
24
17
58.5%
41.5%
Table 22c: Type of crops grown presently VS Type grown before
Total What Type of crops do Corn, Wheat, Pecan nut Count you Grow % of Total Corn, Wheat, Pecan Nuts, Barely, Lucern
Count % of Total
Maize, Ground Nuts, Wheat, Barely, Pecon, Lucene
Count
Maize and wheat
Count
% of Total
% of Total Wheat, Maize and Peanuts
Count % of Total
Maize, Wheat and Ground Nuts
Count % of Total
Corn, Wheat and Ground Nuts
Count % of Total
Wheat, Ground Nuts, Pecan and Alfalfe
Count % of Total
Total
Count % of Total
239
7 17.1% 6 14.6% 4 9.8% 7 17.1% 9 22.0% 6 14.6% 1 2.4% 1 2.4% 41 100.0%
Table 22d: Chi-Square tests Value
df
Asymp. Sig. (2-sided)
Pearson Chi-Square
15.266a
7
.033
Likelihood Ratio
18.940
7
.008
Linear-by-Linear Association
7.654
1
.006
N of Valid Cases
41
T-Test was done to examine whether there is any significant difference between farm size presently and before now: Table 23a: Group statistics Before and After Amount of land Cultivated
Before After
N
Mean
Std. Deviation
Std. Error Mean
10 1470.00
280.079
88.569
8 1481.25
186.964
66.102
240
Table 23b: Independent sample test t-test for Levene's Test for Equality of Equality Variances of Means
F Amount of land Cultivated
Equal variances assumed
Sig. .183
t .674
-.097
Equal variances not assumed
-.102
Table 23c: Independent sample test t-test for Equality of Means
df Amount of land Cultivated
Equal variances assumed Equal variances not assumed
241
Sig. (2-tailed)
Mean Difference
16
.924
-11.250
15.597
.920
-11.250
Table 23d: Independent sample test t-test for Equality of Means 95% Confidence Interval of the Difference Std. Error Difference Amount of land Cultivated
Lower
Upper
Equal variances assumed
115.625
-256.363 233.863
Equal variances not assumed
110.516
-246.027 223.527
242
Appendix B: Results of analysed climate data Statistical analysis Variable | Obs Mean Std. Dev. Min Max -------------+-------------------------------------------------------TX | 4330 27.23646 7.583335 8.1 339.2 TN | 4332 11.83286 24.0106 -40 195.2 Rain | 4802 1.76586 10.37016 0 304 Rs | 4802 20.21344 6.834414 0 44.92 u2 | 4802 1.358803 .7346702 .05 12.06 -------------+-------------------------------------------------------Rhx | 4356 87.54715 13.04839 -.3 100 Rhn | 4311 27.78286 23.17768 -17.7 174.9 et0 | 3943 4.04804 1.676608 0 25.74 Hu | 4756 8.33545 4.545889 0 19.52 Cu | 4756 -8.546152 11.37729 -27 23.5 -------------+-------------------------------------------------------Dpcu | 4443 1.72271 3.551762 0 23.5 VP | 816 1.001985 .496615 .01 2.14 Svp | 4756 2.219281 .7518021 0 4.53 Vpd | 816 1.415895 .6399606 .15 3.49 Umax | 2319 8.299728 11.26632 0 257.2 -------------+-------------------------------------------------------Uhr | 2539 1238.453 509.0703 0 2359 Avet | 4330 19.5391 12.92468 1.35 167.1 Averh | 4356 57.52149 15.16374 -.15 137.45
For var rain rhx: pwcorr TX X, star (.05) sig -> Pwcorr TX rain, star (.05) sig | TX rain -------------+-----------------TX | 1.0000 | | Rain | -0.0359* 1.0000 | 0.0183
Pearson correlation coefficient was used to test for relationship between these two parameters. The relationship between rain and temp is significantly negative (P Pwcorr TX rhx, star (.05) sig | TX rhx -------------+-----------------TX | 1.0000 | | Rhx | -0.0973* 1.0000 | 0.0000 | . For var rain rhx: bys year: pwcorr TX X, star (.05) sig -> Bys year: pwcorr TX rain, star (.05) sig ----------------------------------------------------------------------------------------------------> Year = 1997 | TX rain -------------+-----------------TX | 1.0000 | | Rain | -0.4766 1.0000 | 0.1172 | ----------------------------------------------------------------------------------------------------> Year = 1998 | TX rain -------------+-----------------TX | 1.0000 | | Rain | -0.0038 1.0000 | 0.9443 | ----------------------------------------------------------------------------------------------------> Year = 1999 | TX rain -------------+-----------------TX | 1.0000 | | 244
Rain | -0.1127* 1.0000 | 0.0353 | ----------------------------------------------------------------------------------------------------> Year = 2000 | TX rain -------------+-----------------TX | 1.0000 | | Rain | 0.0743 1.0000 | 0.1560 | ----------------------------------------------------------------------------------------------------> Year = 2001 | TX rain -------------+-----------------TX | 1.0000 | | Rain | -0.0655 1.0000 | 0.2218 | ----------------------------------------------------------------------------------------------------> Year = 2002 | TX rain -------------+-----------------TX | 1.0000 | | Rain | -0.1298* 1.0000 | 0.0131 | ----------------------------------------------------------------------------------------------------> Year = 2003 | TX rain -------------+-----------------TX | 1.0000 | | Rain | 0.0450 1.0000 245
| 0.4547 | ----------------------------------------------------------------------------------------------------> Year = 2004 | TX rain -------------+-----------------TX | 1.0000 | | Rain | 0.0754 1.0000 | 0.1721 | ----------------------------------------------------------------------------------------------------> Year = 2005 | TX rain -------------+-----------------TX | 1.0000 | | Rain | 0.0358 1.0000 | 0.5080 | ----------------------------------------------------------------------------------------------------> Year = 2006 | TX rain -------------+-----------------TX | 1.0000 | | Rain | -0.0286 1.0000 | 0.6118 | ----------------------------------------------------------------------------------------------------> Year = 2007 | TX rain -------------+-----------------TX | 1.0000 | | Rain | -0.0328 1.0000 | 0.5344 246
| ----------------------------------------------------------------------------------------------------> Year = 2008 | TX rain -------------+-----------------TX | 1.0000 | | Rain | -0.0416 1.0000 | 0.4973 | ----------------------------------------------------------------------------------------------------> Year = 2009 | TX rain -------------+-----------------TX | 1.0000 | | Rain | -0.1532* 1.0000 | 0.0040 | ----------------------------------------------------------------------------------------------------> Year = 2010 | TX rain -------------+-----------------TX | 1.0000 | | Rain | 0.0112 1.0000 | 0.8473 | -> Bys year: pwcorr TX rhx, star (.05) sig ----------------------------------------------------------------------------------------------------> Year = 1997 | TX rhx -------------+-----------------TX | 1.0000 | | Rhx | . . 247
| |
.
----------------------------------------------------------------------------------------------------> Year = 1998 | TX rhx -------------+-----------------TX | 1.0000 | | Rhx | 0.1886* 1.0000 | 0.0005 | ----------------------------------------------------------------------------------------------------> Year = 1999 | TX rhx -------------+-----------------TX | 1.0000 | | Rhx | 0.0401 1.0000 | 0.4555 | ----------------------------------------------------------------------------------------------------> Year = 2000 | TX rah -------------+-----------------TX | 1.0000 | | Rhx | -0.0300 1.0000 | 0.5671 | ----------------------------------------------------------------------------------------------------> Year = 2001 | TX rhx -------------+-----------------TX | 1.0000 | | Rhx | -0.2725* 1.0000 | 0.0000 248
| ----------------------------------------------------------------------------------------------------> Year = 2002 | TX rhx -------------+-----------------TX | 1.0000 | | Rhx | -0.2004* 1.0000 | 0.0001 | ----------------------------------------------------------------------------------------------------> Year = 2003 | TX rhx -------------+-----------------TX | 1.0000 | | Rhx | -0.0263 1.0000 | 0.6614 | ----------------------------------------------------------------------------------------------------> Year = 2004 | TX rhx -------------+-----------------TX | 1.0000 | | Rhx | -0.1279* 1.0000 | 0.0202 | ----------------------------------------------------------------------------------------------------> Year = 2005 | TX rhx -------------+-----------------TX | 1.0000 | | Rhx | -0.1723* 1.0000 | 0.0013 | 249
----------------------------------------------------------------------------------------------------> Year = 2006 | TX rhx -------------+-----------------TX | 1.0000 | | Rhx | -0.1058 1.0000 | 0.0598 | ----------------------------------------------------------------------------------------------------> Year = 2007 | TX rah -------------+-----------------TX | 1.0000 | | Rhx | -0.0108 1.0000 | 0.8378 | ----------------------------------------------------------------------------------------------------> Year = 2008 | TX rhx -------------+-----------------TX | 1.0000 | | Rhx | -0.1016 1.0000 | 0.0971 | ----------------------------------------------------------------------------------------------------> Year = 2009 | TX rhx -------------+-----------------TX | 1.0000 | | Rhx | -0.2137* 1.0000 | 0.0000 | 250
----------------------------------------------------------------------------------------------------> Year = 2010 | TX rhx -------------+-----------------TX | 1.0000 | | Rhx | -0.0370 1.0000 | 0.5243 | . Log close Name: Log: C:\PROJECTS\PROJECTS_2012\Ojo\Ojo2.log Log type: text Closed on: 27 Mar 2012, 17:05:26
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
0.00
0.10
0.20 0.30 Frequency
0.40
-6.00 -4.00 -2.00 0.00 2.00 4.00 6.00
-6.00 -4.00 -2.00 Tx 0.00 2.00 4.00 6.00
Sample spectral density function
0.50
Evaluated at the natural frequencies
-6.00 -4.00 -2.00 0.00 2.00 4.00 6.00
Sample spectral density function
0.00 2.00 4.00 6.00
RHx
Figures 5a: Autocorrelation graphs of the study area (Max. Temp)
-6.00 -4.00 -2.00
Log Periodogram
Appendix C: Autocorrelation of climate variables
0.00
0.10
0.20 0.30 Frequency
Evaluated at the natural frequencies
Figures 5b: Autocorrelation graphs of the study area (Rel. Humidity)
271
0.40
0.50
-6.00 -4.00 -2.00 0.00 2.00 4.00 6.00
0.00 2.00 4.00 6.00
RHn
-6.00 -4.00 -2.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
Evaluated at the natural frequencies
Figures 5c: Autocorrelation graphs of the study area
0.00
0.10
0.20 0.30 Frequency
Evaluated at the natural frequencies
Figures 5d: Autocorrelation graphs of the study area
272
0.40
-6.00 -4.00 -2.00 0.00 2.00 4.00 6.00
Sample spectral density function
-6.00 -4.00 -2.00 Rs 0.00 2.00 4.00 6.00
Log Periodogram
Log Periodogram
Sample spectral density function
0.50
6.00 -6.00 -4.00 -2.00 0.00
2.00
4.00
6.00 4.00 2.00 U2 0.00 -6.00 -4.00 -2.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
Evaluated at the natural frequencies
Figures 5e: Autocorrelation graphs of the study area
0.00
0.10
0.20 0.30 Frequency
Evaluated at the natural frequencies
Figures 5f: Autocorrelation graphs of the study area 273
0.40
-6.00 -4.00 -2.00 0.00 2.00 4.00 6.00
Sample spectral density function
-6.00 -4.00 -2.00 HU 0.00 2.00 4.00 6.00
Log Periodogram
Log Periodogram
Sample spectral density function
0.50
0.00
0.10
0.20 0.30 Frequency
0.40
-6.00 -4.00 -2.00 0.00 2.00 4.00 6.00
-6.00 -4.00 -2.00 CU 0.00 2.00 4.00 6.00
Log Periodogram
0.50
Evaluated at the natural frequencies
Figures 5g: Autocorrelation graphs of the study area
-6.00 -4.00 -2.00 0.00 2.00 4.00 6.00
0.00 2.00 4.00 6.00
Sample spectral density function
-6.00 -4.00 -2.00
DPCU
Log Periodogram
Sample spectral density function
0.00
0.10
0.20 0.30 Frequency
Evaluated at the natural frequencies
Figures 5h: Autocorrelation graphs of the study area
274
0.40
0.50
-6.00 -4.00 -2.00 0.00 2.00 4.00 6.00
T 2.00 4.00 6.00 0.00 -6.00 -4.00 -2.00
Log Periodogram
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
Evaluated at the natural frequencies
Figures 5i: Autocorrelation graphs of the study area
-6.00 -4.00 -2.00 0.00 2.00 4.00 6.00
0.00 2.00 4.00 6.00
Sample spectral density function
-6.00 -4.00 -2.00
Rain
Log Periodogram
Sample spectral density function
0.00
0.10
0.20 0.30 Frequency
Evaluated at the natural frequencies
Figures 5j: Autocorrelation graphs of the study area
275
0.40
0.50
0.00
0.10
0.20 0.30 Frequency
Evaluated at the natural frequencies
Figures 5k: Autocorrelation graphs of the study area
276
0.40
-6.00 -4.00 -2.00 0.00 2.00 4.00 6.00
-6.00 -4.00 -2.00 Tn 0.00 2.00 4.00 6.00
Log Periodogram
Sample spectral density function
0.50
Appendix D: Guidelines for irrigation water quality Table D1: Guidelines for interpretation of water quality for irrigation use Degree of Restriction on Use Potential Irrigation Problem
Units None
Slight Moderate
to
Severe
Salinity(affects crop water availability) Electrical conductivity
dS/m < 0.7 0.7 – 3.0
> 3.0
mg/l < 450 450 – 2000
> 2000
(or) TDS Infiltration (affects infiltration rate of water into the soil. Evaluate using ECw and SAR together) SAR
=0–3
=
> 0.7 0.7 – 0.2
< 0.2
=3–6
=
> 1.2 1.2 – 0.3
< 0.3
= 6 – 12
=
> 1.9 1.9 – 0.5
< 0.5
= 12 – 20
=
> 2.9 2.9 – 1.3
< 1.3
= 20 – 40
=
> 5.0 5.0 – 2.9
< 2.9
And ECw
Specific Ion Toxicity (affects sensitive crops) Sodium (Na) Surface irrigation
SAR < 3
3–9
sprinkler irrigation
me/l < 3
>3
Surface irrigation
me/l < 4
4 – 10
sprinkler irrigation
me/l < 3
>3
Boron (B)
mg/l < 0.7 0.7 – 3.0
>9
Chloride (Cl)
Miscellaneous Effects (affects susceptible crops)
277
> 10
> 3.0
Nitrogen (NO3 - N)
mg/l < 5
5 – 30
> 30
Bicarbonate (HCO3) me/l < 1.5 1.5 – 8.5
(overhead sprinkling only)
Normal Range 6.5 – 8.4
PH
Source: Ojo, et al, 2006 adapted from Ayers and Westcot (1985)
278
> 8.5
Appendix E Questionnaire for Research Work: Modeling and mapping of irrigation induced salinity of Vaal-harts Irrigation Scheme, South Africa
Part A: Questionnaire for VH Water Manager 1.
Name of Organization:……………………………………………….
2.
Year of Establishment:………………………………………………
3.
Name of Officer:………………………………………………………
4.
Status/post:…………………………………………………………….
5.
Profession: (a) Engineering [ ] Discipline ……………………………. (b) Agriculture [ ] Areas of specialization…………………………… (c) Others specify…………...............................................
6.
Total Area cultivated (i)
At the beginning of scheme………………………………………..
(ii)
Currently:………………………………………………………….
7.
Year of commencement of irrigation scheme……………………………..
8.
Crops cultivated
irrigated or not
(i) ………………………………….
[ ] Yes [ ] No
(ii) …………………………………
[ ] Yes [ ] No
(iii)…………………………………..
[ ] Yes [ ] No 279
(iv)…………………………………...
[ ] Yes [ ] No
(v) ……………………………………. [ ] Yes [ ] No 9.
Total area irrigated: Ha
10. Average crop yield per hectare for the crops listed above Crop
Yield at inception current yield
(i) ………………………………………………………………………… (ii) ………………………………………………………………………… (iii) ……………………………………………………………………….. (iv) ………………………………………………………………………. (v) ………………………………………………………………………… 11.
Has there been any noticeable drop in yield? [ ] Yes [ ] No If yes, which year did it occur …………………………………………..
12.
What was the possible reason(s) for the drop in yield? ……………………………………………………………………………………………… ……………………………………………………………………………………………… ………………………………………
13.
What was the average yield that year? ….................................................
14.
What method(s) of irrigation is /are practiced on the scheme? (i)
Surface [ ]
Total areas irrigated ………………………. Ha please
specify: Basin [ ] Area covered …………………………………………ha Border [ ] Area covered ………………………………………... ha 280
Furrow [ ] Area covered ………………………………………. ha
15.
16.
(ii)
Sprinkler [ ] Area covered …………………………………..… ha
(iii)
Drip [ ] Area covered …………………………………………. ha
(iv)
Others specify……………………………………………………….
Source(s) of water for irrigation on scheme (i)
Dam/reservoir [ ]
(ii)
Stream
(iii)
Borehole
[ ]
(iv)
Well
[ ]
(v)
Others specify:………………………………………………………
[ ]
If source of water is Dam/reservoir (i)
What area of land was submerged?.............................................. ha
(ii)
How many families were displaced?............................................. ha
(iii)
How many families were resettled? ……………………………..ha
(iv)
Area of new land/forest opened up for resettled farers ……………………………………………..ha
(v)
Does the dam serve any other purpose? [ ] Yes
17.
[ ] No
if yes, please specify;
(a)
………………………………………………………………..
(b)
………………………………………………………………….
(c)
………………………………………………………………….
Are the following parameters monitored on the scheme? Yes
No
(i)
Water table depth
[ ]
[ ]
(ii)
Salt level in soil
[ ]
[ ]
281
18.
19.
(iii)
Top soil loss
[ ]
[ ]
(iv)
Surface water quality
[ ]
[ ]
(vii)
Pest in crops
[ ]
[ ]
Does water logging affect this irrigations scheme? [ ] Yes [ ] No (i)
When was it first noticed? ……………………………………
(ii)
How many hectares was water logged? ………………………ha
(iii)
Are these areas still cropped? Yes/No
Does salinity affect this irrigation scheme? [ ] Yes [ ] No (salinity is the accumulation of salt or whitish powdery material on soil surface) (i)
When was it first noticed? ……………………………………….
(ii)
How many hectares have been affected? …………………………ha
(iii)
Has there been any noticeable effect on yield? [ ] Yes [ ] No
(iv)
Please estimate the effect in yield (decrease or increase) ……………………………………. kg/ha increase ……………………………………... kg/ha decrease
20.
Does erosion affect this irrigation scheme? [ ] Yes [ ] No (i)
What is the average annual loss? ……………………….kg/ha
(ii)
What is the highest ever recorded?.................................kg/ha
(iii)
Under what condition(s) was this loss obtained? Rainfall
[ ]
Surface irrigation
[ ]
Sprinkler
[ ]
(iv)
Has erosion caused loss of cropping lands? [ ]Yes [ ] No
(v)
About how many hectares? …………………………… ha 282
21.
22.
23.
Is source of water (surface/stream water) polluted? [ ] Yes [ ] No (i)
Which of the following form(s) of pollution has been observed?
(a)
Increased suspended solid
[ ]
(b)
Silting of stream
[ ]
(c)
Salinity
[ ]
(d)
Muddy water
[ ]
(e)
Others specify:…………………………………………………….
(ii)
Has there been loss of fish population [ ] Yes [ ] No
What kind of ground water pollution is observed? (a)
Salinity/Alkalinity [ ]
(b)
Others specify:…………………………………………………
Does Weed infestation affect the irrigation scheme: [ ] Yes [ ] No (i)
Were the species of weeds existing in the area before irrigation started? Yes
(ii)
(iii)
24.
[ ] No
If no, what are the probable sources of the weeds? (a)
……………………………………………………………
(b)
…………………………………………………………....
(c)
…………………………………………………………….
list common species of weed found on scheme (a)
………………………………………………………………
(b)
……………………………………………………………..
(c)
…………………………………………………………….
Does pests’ infestation affect the irrigation scheme? [ ] Yes
283
[ ] No
[ ]
(i)
Were these species of pests existing in the area before the inception of the scheme? [ ] Yes [ ] No
(ii)
Please list common species of pest found on scheme (a)
………………………………………………………………
(b)
………………………………………………………………
(c)
………………………………………………………………
284
Questionnaire for Research Work Modeling and mapping of irrigation induced salinity of Vaal-Harts Irrigation Schemes, South Africa
Part B: Questionnaire for farmers 1.
Name of irrigation scheme ……………………………………………… Town:………………………………… Province:…………………………….
2.
Name of Farmer:…………………………………………………………..
3.
Occupation:
(i)
Full-time Farmer [ ]
(ii)
Part-time Farmer [ ]
If (2) what other business (es) are you involved in (a)
……………………………………………………………………..
(b)
…………………………………………………………………….
(c)
…………………………………………………………………….
4.
How long have you been Farming?................................................... Years
5.
Did you start on your present plot? Yes/No
6.
If No, How long have you been farming on this plot? …………………… Years
7.
Were you resettled here? [ ] Yes
[ ] No
If yes, (i)
What was the area of your former farmland?.................................ha
(ii)
What is the area of the land you now farm? ……………………..ha
(iii)
How would you assess the fertility of this land compared to your former land? (a)
Better
(b)
Much better
(c)
Fairly better
[ ] [ ] [ ] 285
(d)
About the same [ ]
(e)
Worse
(f)
Terribly worse [ ]
[ ]
8.
Was your former land submerged by the reservoir?
9.
What crop do young row? (i)
……………………………………………………………………..
(ii)
………………………………………………………………………
(iii)
………………………………………………………………………
(iv)
……………………………………………………………………..
(v)
…………………………………………………………………….
10.
Were these the crops you were growing before the scheme was established? Yes/No.
11.
If No, (a)
Why did you change?
(i)
The former crops will not do well on the new land [ ]
(ii)
The new crops are in higher demand [ ]
(iii)
You were advised to grow the new crops [ ]
(iv)
Other reasons……………………………………………………
(b)
Which crops were you growing before
(i)
………………………………………………………………………
(ii)
………………………………………………………………………
(iii)
………………………………………………………………………
(iv)
………………………………………………………………………
12.
Do you irrigate? Yes [ ] No [ ]
13.
If yes, when did you start? ………………………………………………… 286
Or
14.
(a)
Before schemed started
[ ]
(b)
When scheme started
[ ]
(c)
Years after scheme had started [ ]
What method of irrigation do you practice? (i)
Sprinkler
[ ]
(i)
Boarder
[ ]
(ii)
Basin
[ ]
(iii)
Furrow [ ]
(ii)
Surface
[ ]
(4) Others specify………………………………………………………… 15.
16.
17.
What is your source of water? (i)
From dam (reservoir -canal)
[ ]
(ii)
Borehole/well
[ ]
(iii)
River/Stream
[ ]
(iv)
Others specify………………………………………………………
Which crop do you irrigate? (i)
………………………………………………………………………
(ii)
………………………………………………………………………
(iii)
………………………………………………………………………
(v)
………………………………………………………………………
How would you assess you yield now, compared to those years before the scheme (or irrigation).
18.
Kindly estimate your yield in the last five years. Crop: (a)…………………… (b)…………………… (c)………………… 287
5 years ago ………………..
……………………..
……..……………
4 years ago ……………… ……………………….
…………………..
3 years ago ……………… ……………………….
…………………
2 years ago ……………… ……………………….
…………………
1 year ago ……………… ………………………. 19.
…………………
Are there areas of your land that are no longer cultivable? [ ] Yes [ ] No
20.
21.
If yes (a)
How large is this area? …………………………………. Ha
(b)
Why is it not cultivable? Because of
(i)
Stagnant water on land [ ]
(ii)
Crops don’t do well on it [ ]
(iii)
Some whitish powdery materials have become part of the soil. [ ]
(v)
Rain water (flood) has removed the top soil [ ]
Which of the following has affected your yield most? (i)
Stagnant water
[ ]
(ii)
Poor soil condition (whitish powder)
(iii)
Loss of soil due to rain (flood) [ ]
(iv)
Poor water from source
[ ]
[ ]
22.
Have you noticed strange weeds (grasses) on your farm? Yes/No
23.
Have these weeds caused any specific problem(s)? [ ] Yes
[ ] No 288
24.
If yes, please state:………………………………………………………
25.
Since the scheme started, what types of sickness have you observed as very common? (a)
Your fellow farmers
(i)
Malaria [ ]
(ii)
Bilharzia (blood in urine) [ ]
(iii)
Eye problems [ ]
(iv)
Others specify:…………………………….. [ ]
(b)
In Children
(i)
Malaria [ ]
(ii)
Bilharzia
(iii)
Eye problems (Oncho) [ ]
(iv)
Worms [ ]
[ ]
Thank you.
Questionnaire for Doctoral Research Work: Remote Sensing, GIS and Computational Tools for Environmental Impact Assessment, Analysis and Modeling of Vaal-harts Irrigation Schemes, South Africa
Questionnaire for farmers 1.
Name of irrigation scheme ……………………………………………… 289
Town:………………………………… Province:……………………………. 2.
Name of Farmer:…………………………………………………………..
3.
Occupation:
(i)
Full-time Farmer [ ]
(ii)
Part-time Farmer [ ]
If (2) what other business (es) are you involved in (d)
……………………………………………………………………..
(e)
…………………………………………………………………….
(f)
…………………………………………………………………….
4.
How long have you been Farming?................................................... Years
5.
Did you start on your present plot? Yes/No
6.
If No, How long have you been farming on this plot? …………………… Years
7.
Were you resettled here? [ ] Yes
[ ] No
If yes,
8.
(vi)
What was the area of your former farmland?.................................ha
(vii)
What is the area of the land you now farm? ……………………..ha
(viii)
How would you assess the fertility of this land compared to your former land? (a)
Better
[ ]
(b)
Much better
(c)
Fairly better
(d)
About the same [ ]
(e)
Worse
(f)
Terribly worse [ ]
[ ] [ ]
[ ]
Was your former land submerged by the reservoir? 290
9.
What crop do young row? (i)
……………………………………………………………………..
(ii)
………………………………………………………………………
(iii)
………………………………………………………………………
(ix)
……………………………………………………………………..
(x)
…………………………………………………………………….
10.
Were these the crops you were growing before the scheme was established? Yes/No.
11.
If No, (a)
Why did you change?
(i)
The former crops will not do well on the new land [ ]
(ii)
The new crops are in higher demand [ ]
(vi)
You were advised to grow the new crops [ ]
(vii)
Other reasons……………………………………………………
(c)
Which crops were you growing before
(vi)
………………………………………………………………………
(vii)
………………………………………………………………………
(viii)
………………………………………………………………………
(ix)
………………………………………………………………………
12.
Do you irrigate? Yes [ ] No [ ]
13.
If yes, when did you start? ………………………………………………… Or
14.
(a)
Before schemed started
[ ]
(b)
When scheme started
[ ]
(c)
Years after scheme had started [ ]
What method of irrigation do you practice? (i)
Sprinkler
[ ]
(ii) 291
Surface
[ ]
(i)
Boarder
[ ]
(ii)
Basin
[ ]
(iii)
Furrow [ ]
(4) Others specify………………………………………………………… 15.
16.
17.
What is your source of water? (i)
From dam (reservoir -canal)
[ ]
(ii)
Borehole/well
[ ]
(iii)
River/Stream
[ ]
(iv)
Others specify………………………………………………………
Which crop do you irrigate? (i)
………………………………………………………………………
(ii)
………………………………………………………………………
(iii)
………………………………………………………………………
(x)
………………………………………………………………………
How would you assess you yield now, compared to those years before the scheme (or irrigation).
18.
Kindly estimate your yield in the last five years. Crop: (a)…………………… (b)…………………… (c)………………… 5 years ago ………………..
……………………..
……..……………
4 years ago ……………… ……………………….
…………………..
3 years ago ……………… ……………………….
…………………
2 years ago ……………… ……………………….
…………………
1 year ago ……………… ………………………. 292
…………………
19.
Are there areas of your land that are no longer cultivable? [ ] Yes [ ] No
20.
21.
If yes (a)
How large is this area? …………………………………. Ha
(b)
Why is it not cultivable? Because of
(i)
Stagnant water on land [ ]
(ii)
Crops don’t do well on it [ ]
(iii)
Some whitish powdery materials have become part of the soil. [ ]
(viii)
Rain water (flood) has removed the top soil [ ]
Which of the following has affected your yield most? (i)
Stagnant water
[ ]
(ii)
Poor soil condition (whitish powder)
(iii)
Loss of soil due to rain (flood) [ ]
(iv)
Poor water from source
[ ]
[ ]
22.
Have you noticed strange weeds (grasses) on your farm? Yes/No
23.
Have these weeds caused any specific problem(s)? [ ] Yes
[ ] No
24.
If yes, please state:………………………………………………………
25.
Since the scheme started, what types of sickness have you observed as very common? (a)
Your fellow farmers
(i)
Malaria [ ]
(ii)
Bilharzia (blood in urine) [ ] 293
(iii)
Eye problems [ ]
(iv)
Others specify:…………………………….. [ ]
(c)
In Children
(i)
Malaria [ ]
(ii)
Bilharzia
(v)
Eye problems (Oncho) [ ]
(vi)
Worms [ ]
[ ]
Thank you.
294
