Feb 18, 2018 - Abou-Khalifa (2012) evaluated rice varieties to different levels of nitrogen and found higher grain yield ...... Kader, M.A., Mia, M.S.R., Kafi, M.A., Hossain, M. S. and Islam, N. (2015). Improving Yield of ..... c) Poster f) Video show ...
Khulna University
Doctoral Dissertation
Farmers’ Knowledge Management Model for Increasing Rice Production in Southwest Bangladesh
Bidyuth Kumar Mahalder Roll No. PhD - 130805
A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy
Agrotechnology Discipline
February 2018
Declaration of Authorship I, the student of Agrotechnology Discipline of Khulna University, Bidyuth Kumar Mahalder, Roll No. PhD-130805, declare that this dissertation titled, ‘Farmers’ Knowledge Management Model for Increasing Rice Production in Southwest Bangladesh’ and the work presented in it are my own. I confirm that:
This work was done wholly while in candidature for Ph.D. degree at this University.
Any part of this dissertation has not previously been submitted for a degree or any other qualification at this University or any other institution.
Where I have consulted the published work of others, this is always clearly attributed.
Where I have quoted from the work of others, the source is always given. With the exception of such quotations, this dissertation is entirely my own work.
I have acknowledged all main sources of help.
Where the dissertation is based on work done by myself jointly with others, I have made clear exactly what was done by others and what I have contributed myself.
Signed: Date: February 18, 2018
--------------------------------------Counter signed by the Supervisor Name: Dr. Mohammad Bashir Ahmed Date: February 18, 2018 ii
Declaration of Acceptance This dissertation has been accepted by Agrotechnology Discipline under Khulna University in partial fulfillment of the requirements for the degree of Doctor of Philosophy
Doctoral Research Advisory Committee
Signature of Chairman of DRAL Professor Dr. Mohammad Bashir Ahmed
(Supervisor)
-fur^),9r_ Signature of Member Dr. Humnath Bhandari IRRI Representative for Bangladesh (Co-Supervisor)
Signature of Member Professor Khan Golam Quddus
Signature of Member Professor Dr. Md. Monirul Islam
Signature of Member Professor Dr. Sarder Safiqul Islam Date of Defense February 18,2018
l1l
Dedicated to
My Parents and Wife
iv
Acknowledgements I feel proud to express my deepest sense of gratitude to my Supervisor, Professor Dr. Mohammad Bashir Ahmed, Agrotechnology Discipline, Khulna University, for his constant encouragement, constructive comments, unending patience and friendly support, including effective guidance and valuable suggestion to accomplish the work. I also express gratefulness to my Co-supervisor Dr. Humnath Bhandari, Agricultural Economist, International Rice Research Institute (IRRI), for his continuous support, advice, guidance, review of analyzed data, results and editing the manuscript. Discussions with Dr. Moin Us Salam, Ex-Associate Professor, Department of Agronomy, Bangladesh Agricultural University, Mymensingh, Bangladesh provided me with clear guidelines on how should I proceed with data analysis, outline the analytical framework, illustrate the concept and implementation of knowledge management model. I thank him for his intellectual instructions and mental support as well. I am very grateful to the International Rice Research Institute (IRRI) for giving me the opportunity in conducting the collaborative research work between the Cereal Systems Initiative to South Asia in Bangladesh (CSISA-BD) project and Khulna University. I express my gratitude to Farzana Ahmad Julie, Unaiza Haris, Gotam Mondal, Nandidni Mondal, Debabrata Mahalder and other colleagues of IRRI and the USAID Agricultural Extension Support Activity (AESA) Project for their cordial support in collecting data from field and preparing the manuscript of the dissertation. Heartfelt gratitude is also extended to Dr. Subrata Chakraborty, Lecturer, School of Management and Enterprise, University of Southern Queensland, Australia for reviewing the draft manuscript, and all my respected teachers of the Agrotechnology Discipline of Khulna University including all the staff members for their exclusive support during my study. Specially, I deeply appreciate Dr. Bodrun Nesa, Senior Scientific Officer, Bangladesh Rice Research Institute (BRRI), Gazipur, for her valuable advice in synthesizing data and my wife Mrs. Sabita Rani Tapadar and daughters Dejani Mahalder and Delina Mahalder who provided their utmost support during my study and also for completion of the manuscript of the dissertation.
v
Abstract Quantifying knowledge on agriculture can have many benefits to stakeholders. While many knowledge-based systems exist in modern days for farmers’ decision support, specific models are lacking on how knowledge traits can impact on agricultural production systems. This study employed modelling technique, supported by field data, to provide a clear understanding and quantifying how knowledge management in production practices can contribute to rice productivity in the environmentally stressed southwest Bangladesh. This research accounted for ‘Boro’ rice as the target crop and ‘BRRI dhan28’ as the test variety. The ‘B-M Model’ was developed following the principle and procedure from published literature, ‘brainstorming’ and results from field survey. Three Knowledge Management Traits (KMT) were defined and quantified as the inputs of the model. Those are: Self-Experience and Observation (SEO), Extension Advisory Services (EAS) and Accessed Information Sources (AIS). The Yield Influencing Process (YIP), the intermediate state variable of the model, was deduced by accounting for the two dominant agronomic practices, seedling age for transplanting and Triple Superphosphate (TSP) application. ‘Knowledge drives farmers’ practice change which in turn influences yield’ was composed as the theoretical framework of the ‘B-M Model’. The model performed strongly against independently collected field data set. Across the 180 farmers’ data, the average Relative Rice Yield (RRY) predicted by the model (0.705) and observed in the field (0.716) was close (Root Mean Squired Deviation (RMSD) = 0.018). The difference between predicted and observed RRY was not statistically different (LSD = 0.03), indicating the model fully captured the field data. A regression of predicted and observed RRY explained 96% variance in observation, further proving the model’s strength in estimating RRY in wider range of farmers’ rice yields. In a normative analysis, the practicality and usefulness of the model to stakeholders were simulated for understanding of how much achievable yield could be expected by changing Farmers’ Knowledge Pool (the sum of three KMT) on rice production practices, and at what combination(s) of KMT to be considered at strategic hierarchy to materialize a targeted achievable yield. To best of the knowledge, a model quantifying rice yield in relation to Knowledge Management Trait does not exist in literature. Upon successful testing under diverse yield scenarios using multiple and sophisticated statistical tools that enhanced credibility of the model, it is concluded that the model has the potential to be used for identifying quantitative pathways of farmers’ knowledge acquisition for practice change leading to improved productivity of rice in the southwest region of Bangladesh. The Author Bidyuth Kumar Mahalder vi
Table of Contents Content
Page
Declaration of Authorship
ii
Deceleration of Acceptance
iii
Dedication
iv
Acknowledgements
v
Abstract
vi
Table of Contents
vii
List of Tables
x
List of Figures
xi
Abbreviations
xiii
Chapter 1
Introduction
1.1 Background
01
1.2 Statement of the Problem
02
1.3 Purpose of the Study
03
1.4 Contribution of the Study
03
1.5 Definition of Terms
04
1.6 Organization of the Dissertation
06
Chapter 2
Literature Review
2.1 Rice in the World
07
2.2 Agriculture and Rice in Bangladesh
07
2.3 Rice Crops in Bangladesh
08
2.4 Status of Rice Production in Bangladesh and Need for Future
08
2.5 The Challenges Meeting the Demand
09
2.6 The Way Forward
10
2.7 The Influence of Crop Management via Agronomic Practices in Addressing Yield Gap
10
2.8 Importance of Information, its Sources and Knowledge Gain on Rice Yield Improvement
14
2.9
Modelling the Change in Agricultural Production Systems
vii
17
Chapter 3
Methodology
3.1
Study Area
20
3.2
Target Crop and Variety
23
3.3
The Research Design
23
3.4
Model Development
23
3.4.1
Data collection
23
3.4.2
Data tabulation and analysis
26
3.4.3
The ‘B-M Model’
27
3.4.3.1 Quantification of knowledge management trait and farmers’ knowledge pool and their variability 28 3.4.3.2 Quantification of yield influencing process and its variability 30 3.4.3.3 Calculation of achievable rice yield
30
3.4.3.4 Presentation of the model output
31
3.5 Model Validation
31
3.5.1
Data collection
31
3.5.2
Data tabulation and analysis
32
3.5.3
Model testing
32
3.6 Potential Application of the ‘B-M Model’ 3.6.1
3.6.2 Chapter 4 4.1
33
Determination of achievable yield could be expected by changing farmers’ knowledge pool on rice production practices
33
The combination (s) of knowledge management trait to materialize a targeted achievable yield
34
Results and Discussion
Characteristics of Sampled Farmers
35
4.1.1 Demographic and socio-economic information
35
4.1.2 Resource base and capacity development of sampled farmers 37 4.2
Distribution of Rice Yield of Sampled Farmers Accounted for ‘B-M Model’ Development
38
4.3
Level of Salinity and Effect on Rice Yield
39
4.4
Rice Yield and Practice Changes of Sampled Farmers Accounted for ‘B-M Model’ Development 41 4.4.1 Difference in yield and practice change
41
4.4.2 The effect of practice change: relationship between seedling age and yield 42
viii
4.4.3 The effect of practice change: relationship between triple superphosphate use and yield 43 4.5
Development of ‘B-M Model’
44
4.5.1 Rationale and justification
44
4.5.2 Statement of the model
45
4.5.3 Blueprint of the ‘B-M Model’
45
4.5.4 Algorithms and parameter estimation
48
4.5.4.1 Knowledge management trait and farmers’ knowledge pool and their variability 48
4.6
4.5.4.2 Yield influencing process and its variability
49
4.5.4.3 Relationship between farmers’ knowledge pool and yield influencing process
51
4.5.4.4 Relationship between yield influencing process and relative rice yield
51
4.5.5 Model output
52
Validation of the ‘B-M Model’
53
4.6.1 Characteristics of sampled farmers for ‘B-M Model’ validation 53 4.6.1.1 Demographic and socio-economic information
53
4.6.1.2 Resource base and capacity development of sampled farmers 54
4.7
4.6.2 Distribution of rice yield of sampled farmers accounted for ‘B-M Model’ validation
55
4.6.3 Performance of the farmers for ‘B-M Model’
57
Potential Application of the ‘B-M Model’ – a Normative Analysis 58
Chapter 5
Conclusions and Recommendations
5.1
Summary of the Study
62
5.2
Recommendation
65
5.2.1 On policy implication
65
5.2.1 On future research
66
References
67
Appendices
75
ix
List of Tables Tables
Page
Table 3.1
Sampling for Data Collection from the ‘Boro’ Rice Farmers 24 in the Study Area
Table 3.2
Score Point of Knowledge Management Trait (KMT) 31 Designated for ‘B-M Model’ Run to Produce Output
Table 4.1
Demographic and Socioeconomic Information of Sampled 36 Farmers
Table 4.2
Resource base and Capacity Development of Sampled 37 Farmers
Table 4.3
Summary Statistics of Distribution of Farmers’ ‘Boro’ Rice Yield (variety, ‘BRRI dhan28’) in the Designated Period-2 38 in the Study Area
Table 4.4
Yield and Farmers’ Practice Change (PC) in Two Time Period for Cultivating ‘Boro’ Rice Yield (variety, ‘BRRI 41 dhan28’) in the Study Area
Table 4.5
Demographic and Socioeconomic Information of Sampled 54 Farmers in Relation to ‘B-M Model’ Validation
Table 4.6
Resource base and Capacity Development of Sampled 54 Farmers in Relation to ‘B-M Model’ Validation
Table 4.7
Summary Statistics of Distribution of Farmers’ ‘Boro’ Rice Yield (variety, ‘BRRI dhan28’) in the Study Area in Relation to ‘B-M Model’ Validation 56
x
List of Figures Figures
Page
Figure 3.1
Map of Study Area Showing Specific Study Locations in 22 Khulna and Satkhira Districts of Southwest Bangladesh
Figure 3.2
Fundamental Steps of Model Development (adapted from 28 Salam, 1992)
Figure 4.1
Status of Farmers’ ‘Boro’ Rice Yield (variety, ‘BRRI 38 dhan28’) in the Designated Period-2 in the Study Area
Figure 4.2
Histogram Showing the Distribution of Farmers’ ‘Boro’ Rice Yield (variety, ‘BRRI dhan28’) in the Designated Period-2 in the Study Area 39 Response of ‘Boro’ Rice Yield (variety, ‘BRRI dhan28’) to Level of Salinity in the Designated Period-2 in the Study Area during the Time of Transplanting 40 Response of ‘Boro’ Rice Yield (variety, ‘BRRI dhan28’) to Level of Salinity in the Designated Period-2 in the Study Area during the Time of Crop Maturity 40 Response of Seedling Age on ‘Boro’ Rice Yield (variety, ‘BRRI dhan28’) in the Designated Period-2 in the Study Area 43 Response of the Dose of Triple Superphosphate (TSP) on ‘Boro’ Rice Yield (variety, ‘BRRI dhan28’) in the Designated Period-2 in the Study Area 44 The Blueprint of the ‘B-M Model’ Showing the Flow of Inputs Translated into the Output. SEO is Self-Experience and Observation, EAS is Extension Advisory Services, AIS is Accessed Information Sources, SA is Seedling Age and 47 TSP is denoted for Triple Superphosphate
Figure 4.3
Figure 4.4
Figure 4.5
Figure 4.6
Figure 4.7
Figure 4.8
Measured Variability in three KMT (Knowledge Management Trait): (i) SEO (Self-Experience and Observation), (ii) EAS (Extension Advisory Services) and (iii) AIS (Accessed Information Sources), and Farmers’ Knowledge Pool (FKP) in Cultivating ‘Boro’ Rice (variety, ‘BRRI dhan28’) in the Study Area in Designated Period-2 accounted for the ‘B-M Model’ Development. Vertical 49 Lines Indicate the Ranges
Figure 4.9
Measured Variability in Yield Influencing Process (YIP) and Two Practice Changes (SA is Seedling Age and TSP is Triple Superphosphate) in Cultivating ‘Boro’ Rice (variety, ‘BRRI dhan28’) in the Study Area in Designated Period-2 Accounted for the ‘B-M Model’ Development. Vertical Lines Indicate the Ranges 50 xi
Figures Figure 4.10
Figure 4.11
Figure 4.12
Page Association between Farmers’ Knowledge Pool (FKP) and Yield Influencing Process (YIP) in Relation to ‘Boro’ Rice Yield (variety, ‘BRRI dhan28’) Cultivated in the Designated Period-2 in the Study Area
51
Association between Yield Influencing Process (YIP) and Relative Rice Yield (RRY) in relation to ‘Boro’ Rice Yield (variety, ‘BRRI dhan28’) Cultivated in the Designated Period-2 in the Study Area 52 Outputs of ‘B-M Model’ as Relative Rice Yield (RRY) of Achievable Yield, at Three Levels of Farmers’ Knowledge 53 Pool(FKP)
Figure 4.13
Status of Farmers’ ‘Boro’ Rice Yield (variety, ‘BRRI dhan28’) in the Study Area in Relation to Model Validation 55
Figure 4.14
Histogram Showing the Distribution of Farmers’ ‘Boro’ Rice Yield (variety, ‘BRRI dhan28’) in the Study Area in Relation to Model Validation 56 Comparison of Predicted and Observed Achievable Relative Rice Yield (RRY) in the study area. The 1:1 Line Shows no Significant Difference (P > 0.05) between Predicted and 57 Observed Values (R2 = 0.96)
Figure 4.15
Figure 4.16
Figure 4.17
Relationship between of Farmers’ Knowledge Pool (FKP) and Achievable Relative Rice Yield (RRY expressed as percentage) based on ‘B-M Model’ Run 59 Achievable Rice Yield in the Study Area in Combination of Extension Advisory Services (EAS) and Accessed Information Sources (AIS) on a Defined Self-Experience and Observation (SEO) Score Point of 10 at a Target of 60 80% Achievable Yield
Figure 4.18
Achievable Rice Yield in the Study Area in Combination of Extension Advisory Services (EAS) and Accessed Information Sources (AIS) on a Defined Self-Experience and Observation (SEO) Score Point of 20 at a Target of 61 80% Achievable Yield
Figure 4.19
Achievable Rice Yield in the Study Area in Combination of Extension Advisory Services (EAS) and Accessed Information Sources (AIS) on a Defined Self-Experience and Observation (SEO) Score Point of 30 at a Target of 61 80% Achievable Yield
xii
Abbreviations AESA
USAID Agricultural Extension Support Activity Project
AIS
Accessed Information Sources
AKIS
Agricultural Knowledge and Information System
AWD
Alternative Wetting and Drying
AYSF
Achieved Yield by Individual Sampled (Respondent) Farmer
BBS
Bangladesh Bureau of Statistics
BDT
Bangladesh Taka
B-M Model
A newly developed model that predicts ‘Boro’ Rice Yield
BRRI
Bangladesh Rice Research Institute
BTV
Bangladesh Television
CI
Confidence Interval
CV
Coefficient of Variance
DAE
Department of Agricultural Extension
dS
Deci Siemens
EAS
Extension Advisory Services
FAO
Food and Agriculture Organization of the United Nations
FGD
Focus Group Discussions
GO
Government Organization
ICT
Information and Communication Technology
IEC
Information, Education and Communication
IRRI
International Rice Research Institute
FKP
Farmers’ Knowledge Pool
Ha
Hectare
HYSF
Highest Yield Recorded in the Sampled Farmers
KM
Knowledge Management
KMT
Knowledge Management Trait
LASP
Local Agro Service Provider
LCS
Lack of Positive Correlation Weighted by Standard Deviation
LSD
Least Squared Difference
MoP
Muriate of Potash
MSD
Mean Squared Deviation
xiii
NGO
Non-Government Organization
NO
Number of Data Points in Observation
NP
Number of Data Points in Model Prediction
PC
Practice Change
PRA
Participatory Rural Appraisal
RMSD
Root Mean Squared Deviation
RRY
Relative Rice Yield
SA
Seedling Age
SB
Squared Bias
SD
Standard Deviation
SDP
Standard Deviation in Model Prediction
SDO
Standard Deviation in Observation
SDSD
Squared Difference between Predicted and Observed Standard Deviation
SED
Standard Error of the Difference
SEO
Self Experience and Observation
SPSS
Statistical Package for Social Science
TV
Television
TSP
Triple Superphosphate
YIP
Yield Influencing Process
xiv
CHAPTER 1 Introduction 1.1 Background Rice is the staple food of 165 million Bangladeshis. Rice production in the country has increased three-fold since 1971, the time of her independence. Positively, the country has transformed from so called “Bottomless Basket” to “Full of Food Basket” (Kabir et al., 2015) and the country has earned self-sufficiency in the crop sub-sector, and even briefly entered into the export regime (BER, 2015). LaFranchi (2015) has unreservedly narrated, “Bangladesh has emerged as a global model for combating hunger and obtained great success in becoming a country of food surplus from a country lagged with chronic food shortages”. The country will, however, need more and more food to feed the increasing population. A model-estimate, as presented by Kabir et al. (2015), shows that the current population (162.2 million) will reach to 170.8, 187.6, 203.0 and 215.4 million in 2021, 2031, 2041 and 2050, respectively. This will significantly affect the volume of the requirement of rice. For example, taking 2014 as baseline, the demand for clean rice in 2021, 2031, 2041 and 2050 will have to go up by 1%, 11%, 20%, and 27%, respectively. The supply of rice production in Bangladesh would be severely challenged by a number of constraints. These include decreasing land, scarcity of agricultural labour, deteriorating soil health, scarcity of water, and increasing climate vulnerability with the events of drought, salinity, flood, heat and cold. Climate change induced extreme events are likely to occur more frequently and become intensified in future. This will adversely affect rice production of the country. Such challenges in the coastal districts of southern Bangladesh will be felt more (Mainuddin et al., 2011). In this region, an estimated one million ha of area remains fallow during the dry (winter) season because of late planting and late harvesting of ‘Transplant Aman’ rice, soil and water salinity, and unavailability of 1
quality irrigation water. In order to bring more productivity in these regions and contribute to national food security, the government has prioritized the development and improvement of farming systems by growing ‘Boro’ rice in winter; this will also reduce pressure from the declining groundwater table in the northern region (CSISA, 2010).. ‘Boro’ rice production in southwest Bangladesh, like any other region, has two dimensions, horizontal and vertical. Horizontal dimension has two wings, cropping area and cropping intensity. The country as a whole has a limited scope for production increase from both wings of horizontal dimension. The net cropped area of the country is now standing as 7.81 million ha, which is likely going down to 6.87 million ha in 2050, if the current rate of decrease continues (Kabir et al., 2015). This means Bangladesh will be expecting less land for more production. On the other hand, cropping intensity which is currently standing as 194% can reach to a maximum of ~221 around 2050 (estimated by Kabir et al., 2015). Groundwater depletion in northwest Bangladesh is another issue affecting rice production. All these glim pictures point out that the required rice production increase will have to be realized vertically though yield increase or utilizing fallow area in the coastal zone.
1.2 Statement of the Problem Salam et al. (2016, 2017) puts forward that the classic equation of yield is ‘G’ by ‘E’, where ‘G’ is genotype, or variety of a crop, and ‘E’ is environment on which the variety is set to express its potential. In the recent years, the ‘E’ component has been segregated to ‘E’ by ‘M’, where ‘M’ is management. This segregation has been necessary because the whole atmosphere of the environment (E) is changed due to management (M); this change could be good or bad. Through good management, a farmer can achieve increased yield, while the yield could be poor due to poor management. A good management requires a good knowledge of the technology and its use. This explains the existence of yield gap between the farmers within a geographical location (Evenson et al., 1996). Kabir et al. (2015) have calculated the yield gap of clean rice in Bangladesh as 0.83 t ha -1, and 2
quantitatively shown how incrementally reducing this gap could immensely contribute to increased future rice production of this country (Alam and Hossain, 1998; Duwayri et al., 2000; Mondal, 2011). Salam et al. (2016) has stated that
management is ‘synonymous to agronomic practices’. Therefore, (agronomic) practice change can lead to changes in yield of ‘Boro’ rice in the southwest Bangladesh.
1.3 Purpose of the Study Knowledge is interpreted as a “sum of relationships of meaning that farmers create in their minds from available information, their experience, their feelings and their ideas” (Ferreira, 2002). Generated information through various sources becomes knowledge when farmers integrate those with what they already know (Dhewa, 2017). Innovative agronomic practices that either stem from scientific community, or farmers’ informal engagement through ‘trial and error’ method, or any other sources, can drive ‘Boro’ rice yield of a variety. The application of acquired ‘knowledge’ on those innovations can contribute to improve such yield under farmers’ circumstances. Velden (2002) argues that limited access of rice farmers to appropriate knowledge is a critical concern to achieve higher rice production. With the above background, this study aims to build a framework in order to understand which knowledge attributes and in what quantity of the attributes influence the rice yield to what degree. The research targeted the following specific objectives: (i)
To analyze temporal variation in ‘Boro’ rice yield in relation to farmers’ knowledge and agronomic practices in southwest Bangladesh;
(ii)
To develop a model to predict changes in ‘Boro’ rice yield based on farmers’ knowledge attributes;
(iii)
To validate the model with farmers’ yield changes in Boro’ rice under different knowledge attributes.
3
1.4 Contribution of the Study To the best of my knowledge, a model predicting rice yield variability through quantified knowledge attributes does not exist for Bangladesh. Potentially, the model can be used as a decision-making tool to guide various stakeholders to identify which knowledge attribute(s) of farmers and to what level those attributes are needed to reach a maximum rice yield in a locality.
1.5 Definitions of Terms Accessed Information Sources (AIS): One of the three knowledge traits of the ‘B-M Model’ which accounts for knowledge sharing networks, and electronic and printed media. B-M Model: A newly developed model that predicts relative achievable ‘Boro’ rice yield in a locality through translating farmers’ knowledge traits. The ‘B-M Model’ was named after its two innovators, Bidyuth Kumar Mahalder, the author and Moin Us Salam, a reputed agricultural scientist and modeller. Cropping intensity: The percentage of crops that can be grown in a piece of land in one year. Total Cropped Area % of CI (Cropping Intensity) =
X 100 Net Cropped Area
Extension Advisory Services (EAS): One of the three knowledge traits the ‘B-M Model’ accounts for. It includes agricultural knowledge gained through purposive interactions between farmers and public/private agricultural service extension agents. Farmers’ Knowledge Pool (FKP): Sum of the three knowledge trait (SEO, EAS and AIS), the ‘B-M Model’ accounts for.
4
Knowledge: Knowledge is a covert category of cultural elements and is understandable by interpreting verbal and non-verbal language. Therefore, knowledge is interpreted as a “sum of relationships of meanings that people create in their minds from available information, their experience, their feelings, and their ideas” (Ferreira, 2002). Knowledge Management: Refers to a range of practices used by an individual to identify, create, represent and distribute knowledge for reuse, awareness and learning. Knowledge Management Trait (KMT): The pathway of knowledge acquisition for farmers is defined as knowledge management trait. In the present study, three pathways i.e. KMT were considered. They are (i) Self-Experience and Observation (SEO), (ii) Extension Advisory Services (EAS) and (iii) Accessed Information Sources (AIS). Self-Experience and Observation (SEO): One of the three knowledge traits the ‘B-M Model’ accounts for. It refers to farmers’ agricultural knowledge gain through observation of practices within and outside his/her own household and his/her own experiences in farming over the years. Yield: The productivity of a crop in a unit land area. Yield gap: Difference between potential farm yield and actual farm yield. Yield Influencing Process (YIP): The combined effect of all agronomic practices to yield of a crop. Examples of such attributes are tillage operation, transplanting method, rice seedling age at transplanting, time of transplanting, type and time of weeding operations, fertilizer type and dose, type and time of insect-pest management and harvesting time.
5
1.6 Organization of the Dissertation This study is organized into five main chapters including the introductory chapter, which describes the background of the study, statement of the problem and purpose of the study and its objectives. Chapter 2 reviews the existing literature dealing with contemporary and relevant issues on rice production, especially ‘Boro’ rice production and rice farming system in south Asian countries, especially in Bangladesh. This chapter also focuses on farmers’ knowledge management practices in growing ‘Boro’ rice in southwest region of Bangladesh. Chapter 3 provides an account of the methodology employed in the present study. It describes the study design, study area and the newly developed farmers’ knowledge management model, data collection tools, sampling and collection of field data, both quantitative and qualitative, for construction and validation of the model. This chapter further states the data analysis methods and tools. Chapter 4 outlines the findings of the study with corresponding discussion pertaining to the results obtained from field survey as well as newly developed farmers’ knowledge management model. This chapter provides a more rigorous and detailed analysis in terms of calibration, validation and implications of farmers’ knowledge management model for increasing rice production in the study area. Finally, chapter 5 provides a summary and overview of the entire study and recommendations in relation to further expansion of the developed model, and its application at strategic level. It further points out the suggested policy recommendations relevant to the study.
6
CHAPTER 2 Literature Review 2.1 Rice in the World Rice (Oryza sativa L.) is grown in more than a hundred countries with total cultivated area of about 160 million hectares and occupies 11% of the world’s cultivated area with the production of more than 700 million tons (Alam et al., 2009; IRRI, 2010). It feeds more than half the people in the world (IRRI 2010). Rice is the staple crop of Asia where 90% of world rice is produced and consumed (Sarker at al., 2013).
2.2 Agriculture and Rice in Bangladesh Agriculture is the backbone of Bangladesh (BBS, 2010; Mondal, 2010). The sector contributes ~14% of the gross domestic product (GDP) of the country and provides 46% of the total employment (BBS, 2016). Crop production dominates Bangladesh agriculture, with the cropping intensity of 192%. Rice is the dominant crop, in terms of value adding, it accounts for more than 60% of total crop agriculture value. Almost 75% of the total cropped area is planted with rice, which accounts for over 90% of total cereal production (Assaduzzaman et al., 2010). The country is the fourth largest producer of rice in the world (Awal and Siddique, 2011). The importance of the crop in Bangladesh is manifold. Although infiltration of western food habits has somewhat diversified the traditional consumption pattern, rice still remains as a staple food of Bangladesh (Islam, 2012). It occupies a significant part of the nation’s social and cultural identity. A very common phrase ‘Macche Bhat-e Bangali’ (Fish and Rice is what defines a Bengali) is an embodiment of Bangladeshi identity. It is an integral part of festivals such as ‘Paus Praban’, ‘Durgapuja’ and ‘Annaprashan’ where rice consumption is symbolic religiously and culturally (Islam, 2012). Ahmed (2004) has discussed how rice has 7
been the most strategic commodity in the economy of Bangladesh. ‘Rice security’ is synonymous to ‘Food security’ in Bangladesh as in many other countries (Brolley, 2015). Rice security is not just an economic problem but also a social and political issue, as its insecurity is a key factor to create political instability in the country (Nath, 2015).
2.3 Rice Crops in Bangladesh Shelly et al. (2016) have presented rice growing season in Bangladesh. They are ‘Aus’, ‘Aman’ and ‘Boro’. ‘Aus’ is the pre-monsoon upland rice-growing season under rainfed conditions. This rice is direct or broadcast seeded during March and April after the pre-monsoon shower and harvested between July and August. The monsoon-season rainfed rice is the ‘Aman’, which is the most widespread, including along the coastal areas. ‘Aman’ is planted in two ways: direct seeding with ‘Aus’ in March and April and transplantation between July and August. Both types are harvested from November through December. ‘Boro’ is the dry-season irrigated rice planted from December to early February and harvested between April and June.
2.4 Status of Rice Production in Bangladesh and Need for Future Bangladesh once known as “Bottomless Basket”, which is now transformed into “Full of Food Basket” (Kabir et al., 2015). Since independence, there has been three-fold increase in rice production in the country, which jumped from nearly 11 million ton of milled rice in 1971-72 to about 35 million ton in 2014-15 (AIS, 2016). The country has earned self-sufficient in this crop sub-sector, and even briefly entered into the export regime (BER, 2015). More rice will be required in future because of increasing population. Kabir et al. (2015) estimate that the current population (162.2 million) will reach to 215.4 million in 2050. This will significantly affect the volume of the requirements of rice; the figure would be ~ 44.59 million ton (as milled rice). In a relative term, compared to 2014, the production demand in 2050 will go up by 27%. 8
2.5 The challenges Meeting the Demand Horizontal increase of rice production in Bangladesh is not a possibility (Sattar 2000). This is mainly because the total cultivable land is decreasing at a rate of 0.45% per year owing to the construction of industries, houses, roads and highways (Shelly et al., 2016). Jaim and Begum (2003) also highlights the issue of land availability for Bangladesh agriculture. On the other hand, increase in cropping intensity is a remote hope. The cropping intensity is currently standing as 194% (BBS, 2016). This can slowly raise maximum to ~221% towards 2050 according to a model estimation by Kabir et al. (2015). There are resource and climatic challenges as well. These are decreasing land resource, scarcity of agricultural labour, deteriorating soil health, scarcity of water, and increasing climate vulnerability with unpredictable events of drought (Biswas, 2014), salinity, flood, heat and cold (Shelly et al., 2016). Rice production in the southwest region of Bangladesh, which contributes about 16% of rice production, has become a challenge due to the increase in salinity. According to a report by the International Rice Research Institute (IRRI, 2007), farmers have faced a huge downfall in the saline-prone southwest region of Bangladesh. Tidal wave, shrimp culture and irrigation with saline water are the main reasons of causing salinity in cropland. The extensive shrimp farming system has long-term effect on soil salinization that negatively affects plant growth and crop production (Nahar and Hamid, 2016). A hiccup occurred in the rice sector in 2017 with two flood events, damaging ‘Boro” rice in the ‘Haor’ areas in the early part of the year, and delaying transplanting of ‘T. Aman’ rice in the middle part of the year. With such environmental shock, the country has again started importing rice.
9
2.6 The Way Forward Vertical increase of production remains the only option in Bangladesh to meet the demand of rice and maintain its sustainability. Kabir et al. (2015) has proved that three strategies - accelerating the genetic gain, reducing the yield gap and curtailing adoption lag – can be the keys for maintaining momentum of rice production in Bangladesh. The ‘genetic gain’ is a research issue, whereas ‘curtailing adoption lag’ is more of government’s policy and implementation issue. Therefore, addressing the ‘yield gap’ could the developmental strategy towards increasing rice production in Bangladesh.
2.7 Influence of Crop Management via Agronomic Practices in Addressing the Yield Gap Salam et al. (2016; 2017) have explained crop yield as a product of ‘G’, ‘E’ and ‘M’, where ‘G’, is genotype (or variety), ‘E’ is environment and ‘M’ is management. Awal and Siddique (2011) have emphasized that government of Bangladesh must focus on adopting high yielding varieties in order to increase the rice production. However, modern varieties almost entirely occupies for ‘Boro’ season rice cultivation in Bangladesh (BRRI, 2016); many of those varieties are specific to suited environments (BRRI, 2016). Therefore, management plays the vital role in variating yield under farmers’ field. The management is ‘synonymous to agronomic practices (Salam et al., 2016). Crop management practices can greatly alter rice yield (Ahmed et al., 2015). A survey by Sattar (2000) has shown the farmers’ management activities, such as use of poor quality seeds, improper use of fertilizers and other inputs, failure to control weeds during the critical competition period and ineffective control of pests and diseases, were responsible for the yield gap. According to Mainuddin and Kirby (2015), though Bangladesh has achieved selfsufficiency in rice, due to increased yields and the increased groundwater irrigated area in the dry season, but the rice yields are still below potential levels. The 10
continued development of higher yielding varieties and more productive management practices is anticipated for Bangladesh to maintain self-sufficiency in rice at least to 2050. Kabir et al. (2015) emphasized on alternate methods cultivation in order to bring technical efficiency of ‘Boro’ rice cultivation in Bangladesh. In assessing the effects of agronomic modification of traditional farming practices on productivity and sustainability of rice (wet season)–rice (dry season) system, Yadav et al. (2017) shows that modification of farmers’ practice with no-till and integrated nutrient management can increase rice yield. No-till with improved nutrient management in lowland rice cultivation has resulted in yield advantage (Das et al., 2014). Similar grain yield of rice under puddled and minimum or unpuddled conditions have also been reported (Ladha et al. 2009). The crop management practice of Alternate Wetting and Drying (AWD) is being promoted by IRRI and the national research and extension program in Bangladesh and other parts of the world as a water-saving irrigation practice that reduces the environmental impact of dry season rice production through decreased water usage, and potentially increases yield (Price et al., 2013). Bhuyan et al. (2014) examined fertigation in raised bed planting for transplanted ‘Boro’ rice as a new approach with higher yield and higher fertilizer and water use efficiency compared to the existing agronomic practice in Bangladesh. Research shows potential of greatly increasing rice yields in Bangladesh if nutrient management and other improved production practices are adopted (Haque et al., 1995). Proper management of crop nutrition is of huge importance as judicious and proper use of fertilizers makes remarkable improvement in the yield and quality of rice (Alam et al. 2009). Application of nitrogen fertilizer either in excess or less than the optimum level both affects yield and quality of rice to the significant extent (Manzoor et al., 2006).
11
Yoseftabar (2013) tested hybrid rice using three levels of nitrogen and reported maximum paddy yield by applying maximum level of nitrogen (300 kg ha -1). With ‘Basmati’ cultivars varieties, Sharma et al. (2012) recorded maximum grain yield when nitrogen was applied @ 90 kg ha-1. Abou-Khalifa (2012) evaluated rice varieties to different levels of nitrogen and found higher grain yield when nitrogen was applied at maximum level (220 kg ha-1). Awal et al. (2011) also reported maximum grain with rice variety, KSK-133 when nitrogen was applied at higher rate (156 kg ha-1). Pramanik and Bera (2013) recorded that grain yield increased gradually with increase in nitrogen up to 150 kg ha-1. Improved production practices that included younger seedlings (18 day-old. Compared to 27 day-old by farmers), mechanical transplanting (compared to manual by farmers), modified spacing (30 cm × 14 cm, compared to 25 cm × 15 cm by farmers), improved fertilizers doses (185, 120, 85 and 70, compared to 200, 120, 80, 45 kg ha -1 by farmers of urea, Triple Superphosphate (TSP), Muriate of Potash (MoP)) significantly improved ‘Boro’ rice yield (Kader et al., 2015). Phosphorus (P) influences flowering and ripening of rice (Diba et al., 2015). Acute P deficiency in soil caused a yield reduction in lowland rice by 50% or more (Saleque et al., 1995); Diba et al. (2015) recorded the optimum dose of P in ‘Boro’ rice variety ‘BRRI dhan28’ in Old Brahmaputra Floodplain as 29.11 kg ha -1 (~146 kg TSP ha-1, considering P content of TSP is 20%). This finding was similar to what had been reported by Mahajan et al. (1994). Alam et al. (2009) investigated the response of P on ‘Boro’ rice variety ‘BRRI dhan29’ rand found the highest yield response of P2O5 in as 72 kg ha-1 (~150 kg TSP ha-1, considering P2O5 content of TSP is 48%). On the other hand, Islam et al. (2010) observed that P responded to yield increase of ‘Boro’ rice up to 22 kg ha -1 (~110 kg TSP ha-1, considering P content of TSP is 20%).
12
The age of seedling may be important yield determinant for transplant rice because it has a notable influence on tiller production, grain formation and other yield contributing characters (Singh et al., 1998). It is a key factor which influences the tiller production, grain formation and other yield contributing parameters (Faruk et al., 2009). This attribute is the main factor for uniform stand establishment of rice (Ginigaddara and Ranamukhaarachchi, 2011) which controls its growth and yield (Faghani et al., 2011). The use of appropriate aged seedlings for transplantation and its timely planting are important non-cash inputs for attaining the higher yield of rice (Patra and Haque, 2011). Seedlings age at transplanting is an important factor for uniform stand of rice and regulating its growth and yield (Bassi et al., 1994). Tiller dynamics of the rice plant greatly depends on the age of seedlings at transplanting (Pasuquin et al., 2008). Tillering and growth of rice proceed normally when optimum aged seedlings are transplanted at the right time (Mobasser et al., 2007). Rahimpour et al. (2013) investigated the effect of seedling age on rice cultivars. They observed younger seedlings of 27 days age produced higher grain yield as compared to older seedlings of 35 days. Ali et al. (2013) tested ‘Boro’ rice of the variety “BRRI dhan28’ to different seedling ages, where the maximum grain yield resulted when young seedlings of 15 days were transplanted while minimum compared to the yield received from 30 day-old seedlings. Sarker at al. (2013) observed the yield of ‘Boro’ rice ‘BRRI dhan28’ increased progressively with the decrease in seedling age at transplanting from 70 to 30 days. A similar yield trend was also recorded in the variety ‘BRRI dhan29’. In both the varieties, the yield increased was attributed to the higher number of panicles m-2, filled grains panicle-1 and lower sterility (%). Brar et al. (2012) also reported significant effect of seedling age on paddy yield. According to their findings younger seedlings of 30 days of age produced more grain yield as compared to older seedlings of 60 days.
13
Sarkar et al. (2011) also recorded more grain yield from younger seedlings of 25 days of age while minimum was obtained from older seedlings of 35 days. In system of rice intensification rice variety ‘Ranjit’ was tested using seven levels of seedling age (6, 8, 10, 12, 14, 16 and18 days). It was reported that seedling age significantly affected grain yield and maximum grain yield was obtained by using 10 days old seedlings while minimum was given by 6 days old seedlings (review from Aslam et al., 2015). Pramanik and Bera (2013) recorded more grain yield (5,946 kg ha-1 when younger seedlings of 10 days were transplanted in case of hybrid rice. The highest grain yield with ‘BRRI dhan28”was obtained from intermittent irrigated plots where 15 days old seedlings were transplanted (Ali et al., 2013).
2.8 Importance of Information, its Sources and Knowledge Gain on Rice Yield Improvement Kabir et al. (2015) has emphasized that the ongoing momentum of rice production in Bangladesh can be achieved significantly by making information available to the farmers through the smart dissemination of information. They further emphasized that ‘smart technology and smart dissemination’ can help overcome the production barriers. The agricultural knowledge and information system integrates agricultural education, farmers, researchers, and extension workers to harness knowledge and information from various sources for better farming and improved livelihood (Kashem, 2013). Sattar (2000) had reflected the need for addressing the farmers’ knowledge management practice and the ways to improve it. He cited the incapability of the farmers in using the required amount of fertilizer is an indication that they need to be educated on that practice. He had further stated that lack of awareness of the farmers was responsible to increase yield gap.
14
Mainuddin and Kirby (2015) have indicated that the productive management not only requires in using modern technologies but also equipping the farmers to use those technologies in accordance with their own knowledge. Modern improved technologies must be locally accepted in order to carry out the activities appropriately and efficiently. To increase the production of rice through natural means, it is important to explore ways which will enable the Bangladeshi farmers to improve their knowledge management practices effectively and efficiently. A study undertaken by Iqbal (2013) in the saline prone regions of Bangladesh identified and crosschecked data on the highly tolerant crops to salinity and subsequently developed an approach for the management to enable farmers to produce more crops under the salinity prone southwest region of Bangladesh. In Bangladesh since ‘Boro’ rice often faces low temperature stress at the vegetative as well as reproductive stage, whereas, late ‘Aman’ faces low temperature stress at the reproductive stage subsequently causing a decrease in yield, farmers need to provide information on favourable conditions for growing the crops. Farmers also need to be trained in conditions like soil fertility, pest infestations etc. which are affecting the rice production in many ways (Shelly et al., 2016). According to Tiongco and Hossain (2015) farmers’ have choices in varietal diversity and it is important to understand their knowledge on seed preferences and typology in order to develop and implement support measures for improving the use of modern varieties. Nargis et al. (2009) in a study in the Ghatail upazila of Tangail district during the ‘Boro’ season, found that the farmers could not understand which insecticides should be used. In the most cases, they took suggestions of insecticide traders, neighbouring farmers, friends and relatives. Basak et al. (2012) confirmed that there was a significant amount of yield gap in the ‘Boro’ rice varieties and existing management practices that needed to be taken into account in order to secure a sustainable level of production in the future. 15
Hence, farmers needed to be well trained on management practices associated with the rice production. Omotayo (2015) stressed that knowledge management is a key driver of an entity’s performance. FAO (2016) has emphasized the necessity in identifying the ideas and insights that could be provided to enrich the farmers’ knowledge management practices. Janhanshiri and Walker (2015) discussed on the importance of forming a knowledge database rather than solely depending on huge documentations and publications that knowledge dissemination on improved methods and technologies is struggling in the agricultural arena as thorough research on underutilized and neglected crops in still in the process but the knowledge database that is being formed needs to be accepted locally at first. Ahmed (2012) explored how communication problems could cause significant resistance to adopting modern technologies and improved agricultural practices. In the study undertaken by him, it was also identified that farmers have their own interest to select the agricultural management practices as different farmers have different knowledge and experiences. Farmers’ knowledge in traditional cultivation, soil taxonomy and local varieties is an important aspect to consider as farmers often analyse through self-observation. Participatory approaches in development research should be included as being able to contribute motivates the farmers to engage in management practices more efficiently. This is reflected in the study carried out by Uddin et al. (2014) where the farmers’ adaptation strategies to environmental degradation and climate change effects are examined. Ahmed (2012) states that farmers have their own interest and own skill in selecting the kind of action or practice they want to adopt by using their own experiences in management.
16
Ghosh (2016) carried out a study in three villages of Bangladesh and used correlation co-efficient was used to analyze the data, where he found positive and significant correlation (at 5% level of probability) between knowledge and the farmers’ adoption of ‘BRRI dhan28’. One of the ways through which the farmers can improve their management practices is self-observation or self-learning, as Uddin et al. (2014) state.
2.9 Modelling the Change in Agricultural Production Systems Agricultural production systems is composed of many components. Conway (1985) defines a system as “an assemblage of components contained within a boundary such that the elements within the boundary have strong functional relationships with each other, but limited, week or non-existence relationships with elements in the other assemblages; the combined outcome of the strong functional relationships within the boundary is to produce a distinct behaviour of the assemblage such that it tends to respond to stimuli as a whole, even if the stimulus is only applied to one part”. In the system structure a boundary is drawn, although in reality it does not occur (Dent and Blackie, 1979). Within the boundary, the components exist and interact with each other. The system is run by inputs, both controllable and uncontrollable, which produce output (Salam, 1992). A system is represented by some form – this representation is termed a model. Models are imitations of the real systems (Moorby, 1985). Van Keulen et al. (1975) state that reality is always simplified in modelling, partly because our understanding of basic processes is limited, and partly because this enables us to handle the model. Jones et al. (2010) mention, “in classical nomenclature, models were placed within three categories according to their form: physical (also known as iconic), analogue and symbolic. Physical models provide tangible representations of a system; whereas, analogue models represent an equivalent but not identical system, and neither are applicable to agricultural systems generally. Symbolic models represent systems using symbols commonly joined in one or more mathematical 17
expressions”. Essentially, all models predicting agricultural production systems are symbolic and mathematical. Once a model on agricultural production systems is built, it needs an evaluation to sense how far it has mimicked the reality and how confidently it can be used for decision making (Dent and Blackie, 1979; Salam, 1992). Salam (1992) states that such evaluation is done to judge the accuracy of the model, where validation is a part of measuring the model accuracy (Jones et al., 1987). Validation compares model’s prediction with independent results; independent results mean data which have not been used previously in the development of the model (Spedding, 1975; Salam, 1992). Validation of models for agricultural production systems is a continuous process as these models are just working hypotheses; it is never possible to prove a hypothesis absolutely correct in science (Whisler et al., 1986). Knowledge is applied information and is composed of awareness, understanding, meaning, insight, creativity, ideas, intuition, judgment, and anticipating the outcome of actions. It only exists in the human mind. Generated information only becomes knowledge when people integrate it with what they already know (The Dhewa (2017). Kemp (1975) states knowledge is a system. Since every system grows, knowledge grows inevitably. Knowledge is a conceptual entity. Therefore, it cannot be visualized directly. It can be perceived through its effects or through the medium of documents (Subbaiah, 1985). Rogers (1995) states, “Knowledge occurs when an individual exposed to innovation’s existence and gains some understanding of how it functions”. Innovation is “an idea, practice, or object that is perceived as new by an individual or other unit of adoption” (Roger, 1983). Innovation starts with mobilizing existing knowledge (EU SCAR, 2015). Knowledge management is principally defined as organizational level ass the process of creating, sharing, using and managing the knowledge and information (Girard and Girard, 2015). In 1999, however, the term ‘personal knowledge
18
management’ was introduced which refers to management of knowledge at the individual level (Write, 2005). Brief history of agricultural system models has been provided by Jones et al. (2017a). Capalbo et al. (2017) have emphasized the need for knowledge products for both farm and policy decision makers. Antle et al. (2017) have asked for a new generation of agricultural systems models and knowledge products that can help accelerate the rate of agricultural innovation and meet the global need for food and fiber. Jones et al. (2017b) argue that the most important current limitation is data, both for on-farm decision support and for research investment and policy decision making. The knowledge-based systems such as the expert and decision support systems are useful tools for aiding the farmers in providing “if-then” question and answers, which are now available in the agriculture sector for variety of purposes (Yialouris and Sideridis, 1996; Khan et al., 2008). Soulignac et al. (2012) have presented a knowledge management system for exchanging and creating knowledge in organic farming, while Rahman (2015) has proposed an ‘innovationcycle framework’ of integrated agricultural knowledge system and innovation. Agricultural knowledge has a wide meaning to different players and sectors; farmers refer to it as experience; indigenous and tacit facts, extension and research organizations recognise it as proven good practices that maximises the crop yield, for example (Jahanshiri and Walker, 2015). Quantifying knowledge about agriculture and relate those to production systems can have many benefits to stakeholders. Van Den Ban (1993) states that the basic assumption of studying the Agricultural Knowledge and Information System (AKIS) is that information relevant for decision making is generated by actors and reaches farmers in many different ways. While many agricultural systems model are in existence, models are lacking on how knowledge traits, stem from different sources, can impact on agricultural production systems.
19
CHAPTER 3 Methodology 3.1 Study Area The total area of Bangladesh is 147,570 km2, where the coastal area covers about 20% of the country. Out of 2.85 million hectares of the coastal land, ~0.83 million hectares are arable land, which accounts for over 30% of the total cultivable lands of the country. This study represented southwest region of Bangladesh which is a part of the coastal region. It accounts for two administrative districts - Khulna and Satkhira (Fig. 3.1). These two districts cover an area of 8,253 km2, where 4.27 million people live in 1.02 million households (BBS, 2011).
The livelihood of the people of the study area primarily derives from agriculture, which includes crop cultivation, fish farming and livestock rearing. But rice is the predominant crop and farming system is largely rain fed. The study area is remarkably characterized by ‘ghers’. Land surrounded by large bunds mainly for shrimp and prawn cultivation is known as ‘gher’. Boro rice is mainly cultivated in ‘gher’ area where water depth is more or less one meter in the wet season is dominated by prawn or carp or mix of those culture followed by Boro rice (CSISA, 2014). Cropping intensity in the area is low due to soil salinity, unavailability of quality irrigation water in the dry season (Mainuddin et al., 2011), water logging, and the late harvest of the ‘aman’ crop, which prevents timely establishment of ‘rabi’ crops (Mondal et al., 2015). Peak salinity of the river water during the dry season exceeds 5 dS m-1 across about 59% of the arable land, while river water salinity during rainy season is less than 2 dS m-1 in about 39% of the study area (SRDI, 2012). 20
Other challenges to agricultural productivity include excessive flooding during the rainy season, severe cyclonic storms, and tidal surges throughout the year. The major cropping patterns in the study area are ‘aman-rabi’ and ‘aman’-fallow. Salinity of the topsoil (0–15 cm) varies from 4.0 to 6 dS m-1(electrical conductivity of the saturation extract) in the dry season and remains below 4.0 dS m -1 in the wet season (Mondal et al., 2006). Characteristically, the soils of Satkhira district experience soil salinity (2.32 to 5.95 dS m-1) at varying gradient. Single shrimp, integrated rice-fish/prawn and rice-rice are the major farming systems in the district. Average size of landholding per household in Satkhira district is 0.87 ha (Mainuddin et al., 2011). However, farmers have been cultivating ‘Boro’ rice predominantly using the varieties ‘BRRI dhan28’ and ‘BRRI dhan29’. About 60% farmers do not practice the improved agronomic practices like quality seed, younger seedling, timely planting of rice, row transplanting, skipped row, and improved fertilizer management for improved rice production. However, in 2015-2016 winter rice cropping season the farmers of Satkhira district cultivated high yielding variety of ‘Boro’ rice in 60,972 ha and total production was reported 228,337 tons. The average yield was 3.75 ton ha -1 (BBS, 2016). The soils of Khulna district experience varying gradient of soil salinity (3.5 to 8 dS m-1). A rice dominant cropping pattern is prevalent because of soil characteristics, weather and climate and availability of irrigation water. Single ‘T. Aman’ rice is the major cropping system in the district. Historically, the majority of the land remains fallow in dry season due to soil salinity and lack of fresh irrigation water. Average size of landholding per household in Khulna district is 0.60 ha (Mainuddin et al., 2011). In the last two decades, however, ‘Boro’ rice has been cultivating a part of the district predominantly using the varieties ‘BRRI dhan28’ and ‘BRRI dhan29’. In 2015-2016, the farmers of Khulna district cultivated high yielding variety of ‘Boro’ rice in 32,779 ha and total production was reported 109,752 tons. The farmers got average yield 3.35 ton ha-1 (BBS, 2016). 21
Fig. 3.1. Map of Study Area Showing Specific Study Locations in Khulna and Satkhira Districts of Southwest Bangladesh
3.2 Target Crop and Variety In this study, ‘Boro’ rice was taken as the target crop as it is getting interest to the farmers in the study area in the last two decades (BBS, 2015). The government of 22
Bangladesh has also prioritized the development and improvement of farming systems by growing ‘Boro’ rice in the region (CSISA, 2010). The chosen variety was ‘BRRI dhan28’. Country-wide, this variety is the dominant ‘Boro’ rice by area (BBS, 2015). Significant presence of the variety has been reported in the southwestern part of Bangladesh (Hossain et al., 2012). In a sample survey of 500 households in 344 villages in the region, Ahmed et al. (2016) reported that 39% farmers had been growing ‘BRRI dhan28’ in ‘Boro’ season.
3.3 The Research Design The research design encompassed three major components, described below, in relation to increasing rice production in the study area. (i) Development of a knowledge management model: This component accounts for developing a tool towards prediction of yield variability in the target rice variety based on acquired knowledge trait(s) of the farmers. (ii) Validation of the model: This component describes the procedures and techniques of measuring model-predicted yield variability in the target rice variety in relation to actual field scenarios. (iii) Potential application of the model: This component designs a platform for application of the developed and tested model by the stakeholder(s).
3.4 Model Development 3.4.1 Data collection Data for model development purposes were collected preliminary from 420 respondents, equally (210 each) from Khulna and Satkhira districts. In each district, three (3) upazila were purposively selected; in those upazilas farmers had been widely cultivating ‘Boro’ rice since early 2000s. Two (2) unions and one 23
village from each union were randomly selected from each sampled upazila. Finally, 35 ‘Boro’ rice farmers from each of the villages were randomly selected. Table 3.1 shows the distribution of sampled farmers specific to locations in the study area. Table 3.1. Sampling for Data Collection from the ‘Boro’ Rice Farmers in the Study Area
Daulatpur union
Agardari union
Nawapara union
Daulatpur village
Bashghata village Kamta village
35 35 35 35 35 35 35 35 35 35 Khulna District: 210 Satkhira District: 210 Total preliminary sample: 420
Magri village
Sokhipur union
Kaliganj union Rahimpur village
Jamira union Dhopkhola village
Vara Simla union
Bhanderpara union
Debhata Upazila
Vara Simla village
Batiaghata union
Aushkhali village Mikhali village
Sadar Upazila
Fultala union
Fultala Upazila
35
Kaliganj Upazila
Naodari village
Dumuria Upazila
Madhabkathi Atlia union village
Batiaghata Upazila
Gangarampur union
Satkhira District
Khalsibunia village
Khulna District
35
A semi-structured questionnaire was developed for data collection from sampled farmers (Appendix-1a). Data under following key variables were gathered through face-to-face of the principal decision maker in the household. i. Demography and socio-economic status of the respondents and the household; ii. Interviewee’s resources and capacity development pathways; iii. Agronomic practices impacting ‘Boro’ rice yield: variety, seedling age at transplanting, transplanting method, type and quantity of fertilize use, insect and disease management practices and yield; and iv. Acquisition of knowledge on agronomic practices – source and frequency.
24
The questionnaire was pre-tested with 25 farmer-respondents in April 2016, and edited as necessary. Interviews were then conducted using the revised questionnaire during June to August of 2016, which captured data related to the above stated key variables for current time period. This period was designated as ‘Period-2’. During interview process, recalled data were also gathered on agronomic practices impacting ‘Boro’ rice yield. Those recalled data represented the scenarios of ‘Boro’ rice yield and agronomic practices in the study area a decade ago. This period was designated as ‘Period-1’. These two time periods were considered to measure the changes in ‘Boro’ rice productivity and agronomic practices (henceforth designated as ‘practice change’) in the study area. Three pathways of knowledge acquisition, termed as ‘Knowledge Management Trait (KMT)’ were considered. They are (i) Self-Experience and Observation (SEO), (ii) Extension Advisory Services (EAS) and (iii) Accessed Information Sources (AIS). Self-Experience and Observation (SEO): This KMT accounts for agricultural knowledge gained by the farmers through self-observation of practices within and outside their own households. It also includes their own experiences in farming. The SEO may broadly be synonymous to farmers’ ‘indigenous knowledge’ on their production system (Rogers, 1995). Extension Advisory Services (EAS): This trait represents the agricultural knowledge gained from change agents; in this case,
agricultural extension
service workers, both public and private. Accessed Information Sources (AIS): This KMT includes sources of agricultural information through involving knowledge-sharing-networks, such as farmer groups, fairs, markets, relatives, friends,
neighbors, and social
networks such as ‘Krishi Bhabna’ (thorough Ministry of Agriculture) and ‘Krishi Kotha’ (Department of Agricultural Extension, DAE). AIS also accounts for information from media channels such as newspaper, radio and television, and Information, Education and Communication (IEC) materials from entities (e.g. government, private and development organizations). 25
Focus Group Discussion (FGD) technique (Chambers, 1994) was used to select and define the three (3) KMT. Six FGD sessions were conducted in six sampled upazilas. On each session, 20 - 30 ‘Boro’ rice farmers participated. Participants included 10 - 15% women. For compilation of opinions from FGD sessions, a checklist was used (Appendix-1b). In order to determine the level of salinity, samples of surface soils (0 - 15 cm depth) were collected randomly from 48 fields in the study area. The timing of these sample collection was at immediately after transplanting stage of the crop (January, 2016) and maturity stage (April, 2016). The soil salinity refers to the soluble plus readily dissolvable salts in the soil. Salinity is quantified in terms of the total concentration of such soluble salts, or more practically, in terms of the electrical conductivity of the solution. So, the electrical conductivity of the saturation extract was recommended as a general method for appraising soil salinity in relation to plant growth (SSSA, 1982) and (FAO, 1999). The soil samples were analyzed and soil salinity was determined by saturation extract method from the International Rice Research Institute (IRRI) established laboratory in Khulna. The corresponding yield of the ‘Boro’ rice of the 48 sampled fields were recorded through interviewing the farmers. Study area maps of Khulna and Satkhira districts locating soil salinity data points are shown in Appendix 4 (Figure A.4.1 & A.4.2). 3.4.2 Data tabulation and analysis Collected data were entered in MS-Excel and Statistical Package for Social Science (SPSS) software (IBM, 2012). Out of 420 interviewees, 156 farmers frequently changed varieties; therefore, those samples taken away from the analysis. The demography and socio-economic characteristics (Farmer’s age in years, farmer’s education in years, family size in number, male member in number, female member in number, farming experience in years, land size in ha, monthly income in BDT) and resources and capacity development pathways (Received training in days, farmers’ association meeting in number, extension agent’s visit in 26
number per year, received credit support in last two years in number and credit amount received in BDT) were summarized as range, average and standard deviation, as applicable, using SPSS software (IBM, 2012). Distribution of rice yield of the sampled farmers was presented in column graph in order to visualize the data. The ‘Data Analysis Tool Pack’ of MS-Excel was used for estimating ‘Descriptive Statistics’ of the rice yield data and developing ‘Histogram’ of the distribution. The change in yield and agronomic practices occurred between the two periods (Period-1 and Period-2) was calculated as average and 95% confidence interval using SPSS (IBM, 2012) and MS-Excel in-built functions. The attributes of practice change included seedling age at transplanting, and dose of urea, Triple Superphosphate (TSP) and Muriate of Potash (MP). These attributes were chosen based on farmers’ opinions during FGD sessions that those largely impacted the ‘Boro’ rice yield in the region. The relationships between seedling age and yield, and TSP and yield were determined through ‘best fit’ of observed data by regression analysis using ‘Data Analysis Tool Pack’ of MS-Excel (Nessa et al., 2015). In this analysis, seedling ages and corresponding yields were summarized as 5-day step of seedling age; whereas, the TSP and corresponding yields were summarized as 25 kg ha -1 of TSP dose. In both the cases, respective standard error (SE) was calculated for each summary data-point. 3.4.3 The ‘B-M Model’ The ‘B-M Model’ was named after its two innovators, Bidyuth Kumar Mahalder, the author and Moin Us Salam, a reputed agricultural scientist and modeller. The principle and procedure of the development of the ‘B-M Model’ followed according to Salam (1992) and Jones et al. (2010). The chronological steps are shown in Fig. 3.2.
27
Problem Identification and Objectives of Modelling
Data Acquisition and Analysis Model Construction (Blueprint, Algorithm, Calibration and Platform) Model Validation
Model Application
Fig. 3.2. Fundamental Steps of Model Development (adapted from Salam, 1992) The steps shown in Fig. 3.2 should not be viewed as a strict sequential process; rather the steps area repeated in an interactive fashion as information are gained (Jones et al., 1987; Salam et al., 1992). On the first step, the rationale and justification of the model were worked out through field experience of the author and supported by literature (Salam, 1992; PAB, 2007). The data acquisition technique and analysis for the model are described earlier (Sections 3.4.1 and 3.4.2). The ‘statement of the model’ was composed through ‘brain-storming’ in the line of the ‘rationale and justification of the model’. The ‘blueprint of the model’ was constructed following the principle of Dent and Blackie (1979), Salam (1992) and Jones et al. (2010). Its algorithms are described in Sections 3.4.3.1, 3.4.3.2 and 3.4.3.3. The ‘B-M Model’ implemented in MS-Excel platform. The validation and application processes are presented in Sections 3.5 and 3.6, respectively. 3.4.3.1 Quantification of knowledge management trait and farmers’ knowledge pool and their variability Three Knowledge Management Trait (KMT) were quantified as ‘Score Point’. This quantification was done through FGD exercise as described in Section 3.4.1. For this, farmers, firstly, were asked to make a list of all information sources available to them. Based on the list and characteristics of the information sources, they were 28
grouped into three (3) categories, which were named as: (i) Self-Experience and Observation (SEO), (ii) Extension Advisory Services (EAS) and (iii) Accessed Information Sources (AIS). The maximum ‘Score Point’ for each KMT was determined based on majority opinion of the FGD participants. The minimum ‘Score Point’ was also determined in the same way. The ‘Score Point’ for all three KMT were given equal weight based on discussion with participants. Similar scoring of quantifying farmers’ knowledge level on agricultural production and practices was also used by Sulaiman (1989), Bonny (1991), Shushma (1993), Jaganathan et al. (2012) and Sakib et al. (2014). Scoring of Knowledge Management Trait (KMT) was done as follows: (i) Self-Experience and Observation (SEO) a. If RS = 1, then MS = 10; b. If RS = 2, then MS = 20 c. If RS = 3, then MS = 30 d. If RS = 4, then MS = 40 (ii) Extension Advisory Services (EAS): a. If RS = 0, then MS = 0; b. If RS = 1, then MS = 10 c. If RS = 2, then MS = 20 d. If RS = 3, then MS = 30 (iii) Accessed Information Sources (AIS): a. If RS = 0, then MS = 0; b. If RS = 1, then MS = 10 c. If RS = 2, then MS = 20 d. If RS = 3, then MS = 30 ‘MS’ is the ‘Score Point’ used in the model which were determined through the FGD meetings. The each ‘MS’ was coded to facilitate data collection, expressed as ‘RS’. Farmers’ Knowledge Pool (FKP) = SEO + EAS + AIS Where, the range of the value is 10 to 100. 29
…. Eq. (I)
The variability in KMT and FKP in the sampled data was presented as range, mean, 50th percentile (mode), 25th percentile and 75th percentile. Those statistics were calculated using in-built functions of MS-Excel. 3.4.3.2 Quantification of yield influencing process and its variability The Yield Influencing Process (YIP) was calculated as follows: YIP = YIP (SA) + YIP (TSP)
… Eq. (II)
Where, YIP (SA) is the response of Seedling Age (SA) to achievable Relative Rice Yield (expressed as RRY (Section 3.4.3.3), and YIP (TSP) is the response Triple Superphosphate (TSP) to achievable Relative Rice Yield (expressed as RRY (section 3.4.3.3). The values of YIP (SA) and YIP (TSP) were derived through respective response curves using data as stated in Section 3.4.2 and presented in ‘Results and Discussion’ (Sections 4.4.2 and 4.4.3, respectively). In this modelling, other yield influencing factors, such as transplanting time, planting method, application of the dominant plant nutrient (urea) etc., were not considered as changes on those factors were not significant during the two time period (results provided in Section 4.4). The variability in YIP in the sampled data was presented as range, mean, 50 th percentile (mode), 25th percentile and 75th percentile. Those statistics were calculated using functions of MS-Excel. 3.4.3.3 Calculation of achievable rice yield The achievable Relative Rice Yield (RRY) was calculated as: RRY = AYISF / HYSF
… Eq. (III)
30
Where, AYISF is the yield achieved by individual respondentfarmer, and HYSF is the highest yield recorded in the sampled farmers. 3.4.3.3 Relationships between farmers’ knowledge pool and yield influencing process, and between yield influencing process and relative rice yield Both the relationships were determined through ‘best fit’ (Nessa et al., 2015) of summarized data (Section 3.4.2) by regression analysis using ‘Data Analysis Tool Pack’ of MS-Excel. 3.4.3.4 Presentation of the model output The ‘B-M Model’ was run using the inputs given in Table 3.2. The output of the model run was presented graphically. Table 3.2. Score Point of Knowledge Management Trait (KMT) Designated for ‘BM Model’ Run to Produce Output. Score Point of Farmers’ Knowledge Pool (FKP) 10 50 75
Knowledge Management Trait (KMT) SEO EAS AIS 10 25 30
0 15 20
0 10 25
3.5 Model validation 3.5.1 Data collection Data for ‘B-M Model’ validation purposes were collected from 180 respondents, equally (90 each) from Khulna and Satkhira districts. In each district, three (3) upazila were purposively selected; in those upazilas farmers had been widely cultivating ‘Boro’ rice since early 2000s. Two (2) unions and one village from each union were randomly selected from each sampled upazila. Finally, 15 ‘Boro’ rice farmers from each of the villages were randomly selected. The farmers 31
interviewed for data for model development and validation were different. For validation, data included only on rice yield and acquisition of knowledge on agronomic practices – source and frequency [as of section 3.4.1 (iv)]. 3.5.1 Data processing The analysis and presentation of data on demography and socio-economic status of the respondents and interviewees resources and capacity development pathways followed the procedure presented earlier (Section 3.4.1). Data processed for deriving achievable Relative Rice Yield (RRY) according to the procedure described in Section 3.4.3. Gathered data on Knowledge Management Trait (KMT) were processed as of Section 3.4.3.1. 3.5.2 Model testing Performance of the ‘B-M Model’ was analyzed statistically using three approaches: (i) correlation-regression approach (Kobayashi and Salam, 2000) and (Gauch et al., 2003, (ii) paired mean testing approach (predicted value versus observed value) (Mead et al., 2002) and (iii) a deviation approach (predicted value minus observed value) (Kobayashi and Salam, 2000). For correlation-regression approach (predicted value versus observed value), two regression statistics were used: (i) the coefficient of determination (R2) for the 1:1 (y = x) line and (ii) the slope (m) of the regression line which was forced through the origin (Asseng et al., 2000). The standard error of the slope, the level of significance (P) to test whether the slope was different from 1 and the number of points (n) included in the regression analysis were also used. For paired mean testing approach, the Standard Error of the Difference (SED) between two means was calculated as: SED = square root of [(SDP2 / NP) + (SDO2 / NO)]
32
… Eq. (IV)
Where, SDP and NP are the standard deviation and number of datapoints in model’s prediction and SDO and NO are the standard deviation and number of data-points in observation. The Least Significance Difference (LSD) was calculated using the SED and t-value at 5% level of significance and the means of model’s prediction and observation were compared. For the deviation approach, two deviation statistics were used. The first deviation statistic was the Root Mean Squared Deviation (RMSD), which is the average product of deviations for each ‘data-point pair’ in two datasets (Kobayashi and Salam, 2000). The second one was the Mean Squared Deviation (MSD). MSD has three components; Squared Bias (SB), squared difference between predicted and observed Standard Deviations (SD) and lack of positive correlation weighted by the standard deviations of predicted and observed values (LCS). MSD measures the total deviation between predicted and observed values. The lower the value of MSD, the closer the predicted value is to the observed value. SB indicates the agreement between the predicted and observed means, whereas SDSD and LCS together show how closely the model predicts variability around the mean. The two sources of this variability are the magnitude of fluctuations among the n observations and pattern of the fluctuations across n observations; SDSD and LCS quantify ability of the model to describe the magnitude and pattern of fluctuation, respectively.
3.6 Potential Application of the ‘B-M Model’ 3.6.1 Determination of achievable yield could be expected by changing farmers’ knowledge pool on rice production practices The ‘B-M Model’ was run for the range of Farmers’ Knowledge Pool (FKP) - 10 (lower bound) to 100 (upper bound) ‘Score Point’ with step 1 (one). The achievable rice yield (output of the model as percentage) was regressed over the ‘Score Point’ of FKP (input of the model) using the ‘Data Analysis Tool Pack’ of MS-Excel. 33
3.6.2 The combination(s) of knowledge management trait to materialize a targeted achievable yield A target of 80% achievable yield was set at three levels of SEOs – 10, 20 and 30. The combination(s) of the two KMT, Extension Advisory Services (EAS) and Accessed Information Sources (AIS) to reach a targeted achievable yield was investigated. For each level of SEO, the model was run in a combination of 6 (six) levels of EAS and AIS (both in the range of 5 to 30 at 5 steps). Altogether, there were 108 combination (3 [SEO] × 6 [EAS] × 6 [AIS]). Details analysis and results are shown in chapter 4, section 4.7.
34
CHAPTER 4 Results and Discussion Results of the study are presented and the findings have been discussed under seven headlines covering (i) demography of sampled farmers, their socioeconomic, resource base and capacity development attributes; (ii) distribution of farmers’ rice yield accounted for the model development; (iii) responses of the level of salinity in the fields on rice yield; (iv) relation of rice yield and farmers’ agronomic practice change over 10-years’ time period; (v) development of the knowledge management model with respect to its justification, statement, structure and algorithm; (vi) validation of model; and (vii) the model’s potential application. The details are presented below step by step.
4.1 Characteristics of Sampled Farmers 4.1.1 Demographic and socio-economic information The demographic and socio-economic information of farmers sampled in relation to ‘B-M Model’ development have been presented in Table 4.1. Out of 256 respondents, 81.25% were male and remaining 18.75% were female. The land size of the farmers varied from 0.01 to 4.47 ha, averaging 0.59±0.70 ha (± is SD). The data closely reflects the average cultivable land ownership of farmers in Bangladesh which is 0.50 ha per household (Thapa and Gaiha, 2011). The detailed analysis is shown in Appedix-2. The sampled farmers were as young as 21 year-old and as old as 82 year-old having a mean age of 42.55±12.60 years. This average closely resembled the mean farmers’ age in Bangladesh of 48 years (IRRI, 2016). In a separate study in the Satkhira region of Bangladesh, Ghosh (2016) observed the farmers’ age in the range of 25 to 71 years, averaging 47.05.
35
The respondents received, on average, 6.86±4.30 years formal education (Table 4.1); few (19.53%), however, did not attend school at all (Appendix-2, Table A2.4). The average education as observed by Ghosh (2016) in Satkhira region of Bangladesh was 4.48 years of schooling. Nationally, the adult literacy has been reported as 58.60%, whereas in Khulna division as 57.53% (BBS, 2015). This indicates literacy of the respondent farmers was much higher (80.47%) compared to national and regional figures. The family size varied from 1 to 12, averaging 4.80±1.80. The national average figure shows a very similar family size (4.35), where the rural population had, on average, 4.43 members per household (BBS, 2015). Ghosh (2016) in the Satkhira region found average family size of 5.58 in the range of 2 to 7. The sampled farmers had varying degree of farming experience (4 to 65 years), having a mean of 21.03±12.29 years. The study of Ghosh (2016) in Satkhira region observed the farming experience in the range of 1 to 50 years, on average 20.89. This figure is very close to the respondents in the study area. The average monthly income of the sampled farmers was BDT 7,762.50±3,682.34, which varied from BDT 2,500 to 30,000. In rural Bangladesh, the monthly average income was estimated as BDT 13,957 for 2014 (BBS, 2015). This indicate that farmers in the study area were less prosperous than the national average. However, the national average accounted for all rural population including non-farming professionals.
Table 4.1. Demographic and Socioeconomic Information of Sampled Farmers Items Farmer’s age (years) Farmer’s education (years) Family size (number) Male member (number) Female member (number) Farming experience (years) Land size (ha) Monthly income (in BDT)
Range 21 - 82 0 - 14 1 - 12 0-8 1-7 4 - 65 0.01 - 4.47 2,500 - 30,000
Note: Sample size was 256
36
Average 42.55 6.86 4.80 2.48 2.32 21.03 0.59 7,762.50
Standard deviation (±) 12.60 4.30 1.70 1.08 1.10 12.29 0.70 3,682.34
4.1.2 Resource Base and Capacity Development of Sampled Farmers Table 4.2 shows that a large number of respondents (70.70%) used their own resources to meet the expenses of agricultural production, whereas the others depended on credit as high as BDT 100,000 per year. The average annual credit was BDT 5,613.28±12,673.07; the high standard deviation indicates the amounts varied widely among the farmers. Access of formal credit to smallholder farmers in Bangladesh is limited. For example, in 2015, only little over nine percent of those group of farmers (owning less than 1 ha of land) received formal bank credit (Ahmed et al., 2017). In the microcredit sector, 8,547,214 borrowers received BDT 216,562 million in 2015 averaging BDT 25,337.14 per household (PKSF, 2015). Results shows that about 85.16% farmers had the opportunity of receiving training on rice production. The range of the training period was 1 to 7 days, on average 1.09±0.88. In the Satkhira region, Ghosh (2016) reported that 78.66% farmers received some form of training on rice cultivation; this figure is close to the present findings. Among the 256 respondents, 39.84% had some form of connections with farmers’ association(s). The involvement of farmers in association(s) may vary between regions. For example, in Parbatipur area of northwestern regions of Bangladesh only 15% farmers found involved with associations, whereas none in Sherpur area (Farid et al., 2015). Those farmers who had such connections, attended meetings as many as 14 times per year, but the average was low (1.02±2.93). Some respondents (41.41%,) never had contact with extension agents; those had, met little over once a year (average 1.41±1.52). Table 4.2. Resource Base and Capacity Development of Sampled Farmers Items Received training (days) Farmers’ association meeting (number) Extension agent’s visit (number / year) Received credit in two years (number) Credit amount received (BDT) Note: Sample size was 256
37
Range
Average
0-7 0 - 14 0-7 0-3 0 - 100,000
1.09 1.02 1.41 0.36 5,613.28
Standard deviation (±) 0.88 2.93 1.52 0.63 12,673.07
4.2 Distribution of Rice Yield of Sampled Farmers Accounted for B-M Model Development The rice yield of the sampled farmers spread in the range of 2,717 and 7,904 kg ha-1 (Fig. 4.1), where the mean was 5,674.25±835.37 kg ha -1 (Table 4.3). The median and mode of the yield distribution was similar (5,929 kg ha -1).
8000 7500
Rice Yield (kg ha-1)
7000 6500 6000 5500 5000 4500 4000 3500 3000 2500
Fig. 4.1. Status of Farmers’ ‘Boro’ Rice Yield (variety, ‘BRRI dhan28’) in the Designated Period-2 in the Study Area
Table 4.3. Summary Statistics of Distribution of Farmers’ ‘Boro’ Rice Yield (variety, ‘BRRI dhan28’) in the Designated Period-2 in the Study Area Statistical Parameters Mean Standard Error Median Mode Standard Deviation Sample Variance Kurtosis Skewness Range Minimum Maximum Sum Count Confidence Level (95%)
Value 5,674.25 52.21 5,928 5,928 835.37 697,849.32 0.12 -0.33 5,187 2,717 7,904 1,452,607 256 102.82
38
The histogram of the rice yield distribution is presented in Fig. 4.2. Around 45% of the yields (116 out of 256 samples) concentrated in the range of 5,500 to 6,000 kg ha-1. The histogram shows almost bell-shaped distribution of sampled yields (Fig. 4.2), which is slightly skewed (skewness = 0.33) on the left (Table 4.3). It has low kurtosis (0.12, Table 4.3), hence the curve has light tails with few outliers. Taken all together, distribution of rice yield of sampled farmers accounted for ‘B-M Model’ development fulfilled very closely with the criteria of normal distribution. 70 59
60
57 50
38
40 30
25
20 1
2
3000
3500
10
5
2
1
8000
16
7500
Frequency
50
7000
6500
6000
5500
5000
4500
4000
0 Achieved Yield (kg ha-1)
Fig. 4.2. Histogram Showing the Distribution of Farmers’ ‘Boro’ Rice Yield (variety, ‘BRRI dhan28’) in the Designated Period-2 in the Study Area
4.3 Level of Salinity and Effect on Rice Yield The relationship between soil salinity measured at transplanting and maturity stages of rice had insignificant relationship with yield. Fig. 4.3 shows the regression between the two variables (Y = 7100.6 - 531.26 X; R2 = 0.069; n = 48; P > 0.05) explained only ~7% variation in yield due to levels of salinity at transplanting time of rice. Results were similar when such relationship measured at the crop maturity time. In that situation, the regression between the two variables (Y = 6725.4 – 395.92 X; R2 = 0.048; n = 48; P > 0.05) explained only ~5% variation in yield due to levels of salinity (Fig. 4.4). Findings indicate that level of observed soil salinity did not affect rice yield in the study area. Mondal et al.
39
(2015) reported that ‘Boro’ rice variety, ‘BRRI dhan28’, has salt tolerant level about 4 dS m-1.
Rice Yield (kg ha-1)
8000 7000 6000 5000 4000 3000 2000 1.50 2.00 2.50 3.00 3.50 4.00 Soil Salinity at Transplanting Stage (dS m-1) Fig. 4.3. Response of ‘Boro’ Rice Yield (variety, ‘BRRI dhan28’) to Level of Salinity in the Designated Period-2 in the Study Area during the Time of Transplanting
Rice Yield (kg ha-1)
8000 7000 6000 5000 4000 3000 2000 1.50 2.00 2.50 3.00 3.50 4.00 -1 Soil Salinity at Maturity Stage (dS m ) Fig. 4.4. Response of ‘Boro’ Rice Yield (variety, ‘BRRI dhan28’) to Level of Salinity in the Designated Period-2 in the Study Area during the Time of Crop Maturity
40
4.4 Rice Yield and Practice Change of Sampled Farmers Accounted for B-M Model Development 4.4.1 Difference in yield and practice change The yield of rice the farmers have been achieving and the agronomic practices they are adopting in the recent years (Period-2) compared with the past, a decade ago (Period-1), are shown in Table 4.4. Farmers presently (Period-2) yielding 5,673.59±102.33 kg ha -1 of rice (± is 95% confidence interval or CI) compared to 4,944.94±64.54 kg ha -1 in the past (Period-1); statistically this difference was significant (P < 0.05). In a relative scale, this increase was 14.74%. The change in ‘Boro’ rice yield in the study area during the periods closely followed the national trend. During 2006 to 2015, the average yield of ‘Boro’ rice in Bangladesh increased by 15.34% (Appendix-3, Fig. A3.1). Table 4.4. Yield and Farmers’ Practice Change (PC) in Two Time Period for Cultivating ‘Boro’ Rice (variety, ‘BRRI dhan28’) in the Study Area Observed variables Yield (kg ha-1)
Seedling age (days)
Urea (kg ha-1)
TSP (kg ha-1)
MoP (kg ha-1)
Observation period Period 1 Period 2 P 20 – 30 > 30 – 40 > 40 – 50 > 50 – 60 > 60 Average Standard deviation (±)
Model development (n=256) (%) 0.00 21.09 30.86 22.27 14.06 11.72 42.55 12.57
Model validation (n=180) (%) 2.78 21.11 33.89 18.89 16.11 7.22 40.83 12.89
Table A2.4. Respondents’ Education Education range (level) Illiterate Primary Secondary Higher Secondary Graduate Masters Average (years) Standard deviation (±)
Model development (n=256) (%) 19.53 30.87 38.28 5.86 3.90 1.56 6.86 4.30
Model validation (n=180) (%) 15.56 38.34 37.22 5.56 1.66 1.66 5.73 3.97
Table A2.5. Respondents’ Household Size Household member (number) Only 1 2-6 7 - 11 >11 Average Standard deviation (±)
Model development (n=256) (%) 0.39 87.89 10.94 0.78 4.80 1.70
Model validation (n=180) (%) 2.22 86.66 10.56 0.56 4.71 1.75
Table A2.6. Respondents’ Land Resources Household land (hectare) Up to 0.05 > 0.05 – 0.50 > 0.50 – 1.00 > 1.00 – 2.50 > 2.50 Average Standard deviation (±)
Model development (n=256) (%) 10.16 50.39 21.09 12.89 5.47 0.59 0.70
84
Model validation (n=180) (%) 12.22 47.22 20.00 16.11 4.45 0.67 0.82
Table A2.7. Respondents’ Practices in Improved Rice Farming Types of Practices New Rice Variety Planting Date Line Spacing Seedling Age Fertilizer Application Irrigation Technology IPM Technology Soil Conservation
Model development (n=256) (%) 80.08 74.61 80.08 78.52 74.22 81.25 80.47 7.42
Model validation (n=180) (%) 88.89 96.67 87.22 97.78 90.00 62.78 75.56 13.89
Table A2.8. Respondents’ Accessibility to Credit Facility Types of Credit Support Organizations Received Credit from All Sources Bank Govt. Organization NGO / MFI Private Organization
Model development (n=256) (%)
Model validation (n=180) (%)
29.30 3.91 0.00 23.83 1.56
8.33 1.67 0.00 6.66 0.00
Table A2.9. Respondents’ Accessibility to Extension Agent Organization Provided Extension Services Accessed Extension Services / Agents Govt. Organization NGO Private Organization
Model development (n=256) (%)
Model validation (n=180) (%)
58.59 50.78 4.68 3.13
78.89 59.44 16.67 2.78
Table A2.10. Respondents’ Accessibility to Improved Rice Farming Training Organization Provided Training Received Training from All Sources Govt. Organization NGO Dev. Organization / Research Institute
Model development (n=256) (%)
Model validation (n=180) (%)
85.16 23.83 11.72
81.11 15.00 3.33
49.61
62.78
Table A2.11. Respondents’ Affiliation with Group / Association Types of Affiliated Association Affiliated with Group Farmers’ Field School IPM Club Water Users’ Group
Model development (n=256) (%) 39.84 24.61 9.77 5.46
85
Model validation (n=180) (%) 25.56 5.56 19.44 0.56
Table A2.12. Respondents’ Perception on Long Term Climate Change Types of Changes Perceived Rainy Season Dry Season Winter Season Drought Increased Salinity Increased Rainfall Decreased Temperature Increased Flood Frequency Increased
Model development (n=256) (%) 99.22 97.66 39.06 74.61 26.95 46.48 97.27
Model validation (n=180) (%) 95.56 90.00 78.33 67.22 47.22 62.78 97.78
26.56
16.67
Table A2.13. Respondents’ Sources in Acquiring Knowledge on Rice Farming Types of Sources for Knowledge Village Market /Fair Mass Media Neighbour / Relatives Local Learned Person NGO/Dev. Organization Inputs Retailers Govt. Organization Farmers’ Association
Model development (n=256) (%)* 71.09 46.09 95.70 25.00 63.67 39.06 18.75 17.19
Model validation (n=180) (%)* 37.22 53.33 97.22 7.22 78.33 24.44 47.78 28.89
* Multiple Responses
Table A2.14. Respondents’ Major Constraints in Practicing Improved Rice Farming Types of Constraints Lack of Financial Support Lack of Robust Technology Lack of Information and Communication Shortage of Agricultural Labour Lack of farm Machineries Lack of Improved Irrigation Facility Lack of Quality Inputs (e.g. Seeds, Fertilizer & Pesticides)
Model development (n=256) (%)* 75.00 78.91
* Multiple Responses
86
Model validation (n=180) (%)* 85.00 14.44
33.03 69.53 66.80
46.67 28.89 25.56
71.09
34.44
75.39
57.78
Appendix-3: National ‘Boro’ Rice Yield in Bangladesh (1972- 2016)
Rough Rice Yield (kg ha-1)
7000 6000 5000 4000 3000
1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
2000 Year Figure A3.1. National Boro’ rice Yield in Bangladesh during 1972 to 2016 (Arrows indicate the yield differences of the two time periods considered in the study)
87
Appendix -4: Soil Salinity Data Points Map
Figure A4.1. Soil Salinity Data Points Map of Khulna
88
Figure A4.2. Soil Salinity Data Points Map in Satkhira
89
