Landscape units of Tunduru district: digital elevation models as ... - Core

0 downloads 0 Views 322KB Size Report
not available for the adjacent district of Tunduru, landscape units were mapped .... square = 3.2; d.f. = 1) when testing the relationship between landscape units.
The Land (2003) 6(3): 185-192

185

Digital elevation models for mapping landscape units: a case study of Tunduru district, Tanzania S. Dondeyne1/2, L.B. Emmanuel1, S. Mugogo1 & J. Deckers2 1

Naliendele Agricultural Research Institute, PO Box 509 Mtwara, Tanzania 2

Laboratory for Soil and Water Management Catholic University of Leuven

Vital Decosterstraat 102, B-3000 Leuven, Belgium email [email protected]

Summary In the late seventies, landscape units of the regions of Mtwara and Lindi in South Eastern Tanzania were mapped using aerial photographs and field surveys. These maps have proven useful for assessing land suitability for cashew, the major cash crop in the area, and for assessing the potential environmental impact of sulphur used as a fungicide in cashew groves. As maps of similar quality were not available for the adjacent district of Tunduru, landscape units were mapped using digital elevation models (DEM). The DEM were generated from contour lines, digitised at a 100 m height interval from the 1:50,000 scale topographical maps. The derived landscape units are comparable with those obtained for the neighbouring regions of Mtwara and Lindi by more laborious methods. Overlaying the landscape units with data on soil texture points to a relationship between landscape units and soil properties (P = 0.07). The map can now serve as basis for stratifying field surveys. This case study shows that GIS software can simplify the mapping of landscape units.

186

Dondeyne et al.

1. Introduction Cashew nuts are one of Tanzania’s most important foreign exchange earning crops, representing 9% of the world total (FAO, 2000). About 75% of the national cashew nut production comes from the South Eastern part of the country (Topper et al., 1998) of which 66% in Mtwara region, 20% in Lindi region and 14% in Tunduru district (Fig. 1). In the late seventies, Bennett et al. (1979) carried out extensive surveys on the physical environment of Mtwara region (17,500 km²) and Lindi region (66,000 km²). Their maps and reports are the most comprehensive information about the area and have been entered into a geographical database. The digital maps have proven useful for assessing land suitability for cashew (Dondeyne et al., 2001) and for assessing potential environmental impact of sulphur used as a fungicide for controlling powdery mildew disease (Oidium anacardii Noack) in cashew groves (Ngatunga et al., 2003; Ngatunga et al., 2001).

Figure 1 Location of Tunduru district and the regions of Mtwara and Lindi in South Eastern Tanzania

According to the agro-ecological zones map of Tanzania (De Pauw, 1984), the landscapes of Tunduru consist of dissected plains in the west; gently rolling plateaux in the central part developed on Karoo sandstone and Neogene sandy

The Land (2003) 6(3): 185-192

187

sediments; and flat to rolling plains in the west with or without inselbergs, formed on Precambrian basement rocks. Apart from the agro-ecological zones map, information comparable to the one provided by Bennett et al. (1979) is not available for the district of Tunduru (19,000 km²). However, a poor match was obtained when overlaying the agroecological zones map with the contour lines and drainage network from the 1:50,000 scale topographic maps. As the means to carry out studies as Bennett et al. (1979) did, were not available, it was decided to derive landscape units from the topographic maps using digital elevation models. Digital elevation models have since long been proposed for landform analysis by many authors (see for example Dikau, 1989). The aim of this paper is to present the methodology and to discuss the resulting landscape classification and its scope for applications.

2. Materials and methods From the topographical maps covering Tunduru district at a 1:50,000 scale (sheets 289-291, 301-304, 313-316, 320-322), the contour lines at a 100 m height interval and the rivers were digitised and coded as ArcInfo coverages (ESRI, 1994). The contour lines of the separate sheets were all joined into one line coverage. From this, a point coverage was derived which reduced data redundancy and which was used as input for generating the digital elevation models with PC-SEM (ESRI, 1993). PC-SEM is a software product, which uses triangular irregular network models for storing, managing and performing analysis of three-dimensional surfaces. With PC-SEM three dimensional views, from two different angles, were generated on which rivers and a grid of the geographical co-ordinates were draped. Landform units such as ridges, flat plateau tops, hills, escarpments, alluvial plain, inselbergs and plains, were identified and boundaries drawn on the model, similarly as one would do when working with a stereoscopic pair of aerial photographs. Subsequently, the landform units were transposed onto a flat

188

Dondeyne et al.

projection as illustrated in Fig. 2. These were then generalised into landscape units. The distribution of textural soil classes per landscape unit was calculated to test if there would be any relationship between landscape units and soil properties. A chi-square test was used to verify the statistical level of significance of the relationship. The soil textural classes were derived from a map of the ItaloTanzanian Cashew Research Programme (1981) on which only the location of the soil profiles and the soil texture of subsequent horizons is indicated. Any other information as thickness of the horizons, depth of the profile or profile development is lacking. The soil profiles were grouped into sandy profiles, when the texture of all horizons is sand, loamy sand or sandy loam, and clayey when any or all of the layers have a more heavy soil texture.

Figure 2 Landform units identified on a digital elevation model and their transposition on a flat projection

The Land (2003) 6(3): 185-192

189

3. Results and discussion The landscape units identified for Tunduru district consist of inselberg plains, plains, plateaux, dissected plateaux and alluvial plains (Fig.3). Altitude of the plains ranges between 250 and 500 m a.s.l. The plains are extended into the neighbouring regions of Mtwara and Lindi. The eastern part of the plains are dominated by inselbergs, of which the highest reaches 800 m a.s.l. Dissected plateaux dominate the central part of the district, while higher plateaux are found in the western part. The plateaux level between 800 and 900 m a.s.l. and are separated from the dissected plateaux by a steep escarpment with its base at 700 m a.s.l. A residual hill, piercing through the plateaux, reaches 1100 m a.s.l., is the highest point of South Eastern Tanzania. Just as is the case with the plateaux in Mtwara and Lindi regions, alluvial plains are associated to the plateaux. This can be explained by a restriction of the drainage due to the plateaux but is also linked to the perennial water sources found at the base of the plateaux.

Figure 3 Landscape units of Tunduru district derived from DEM models

Dondeyne et al.

190

As shown in Fig. 4 soils of the inselberg plains and plains are mostly clayey. On the dissected plateaux they are mostly sandy, while on the top of the plateaux both sandy and clayey soils seem to occur. Two profiles in the alluvial plains were sandy while one was clayey. The chi-square test yielded a P value of 0.07 (chisquare = 3.2; d.f. = 1) when testing the relationship between landscape units (plains or plateaux) with soil texture (sandy or clayey soils). Because of the small number of observations per landscape unit, the observations of the “inselberg plains” had to be pooled with those of the “plains” to perform the chi-square test; similarly those of the “dissected plateaux” had to be pooled with the “plateaux” and the three observations of the “alluvial plains” were disregarded. Though the P value is slightly higher than the critical 0.05, it seems to indicate that the landscape units would provide a good framework for stratifying further field surveys.

Landscape unit

17%

Inselberg plain (6) 83%

38%

Sandy soils

Plain (8) 63%

Clayey soils

67%

Alluvial plain (3) 33%

64%

Dissected plateau (14)

36%

50%

Plateau (8) 50%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

Distribution of soil profiles

Figure 4 Relationship between soil texture and landscape units; the number of observation per landscape units is given in between brackets (total n = 39)

The Land (2003) 6(3): 185-192

191

With the general availability of personal computers and GIS software it is now possible to classify and map landscape units using DEM. In this particular case study, we have shown that this method yields landscape units similar to those obtained by more laborious methods using aerial photographs as done by Bennett et al. (1979).

Acknowledgement This study was funded by the Belgian Agency for International Co-operation in partnership with the Ministry of Agriculture and Co-operatives, Tanzania.

References Bennett, J.G., Brown, L.C., Geddes, A.M.W., Hendy, C.R.C., Lavelle, A.M., Sewell, L.G. & Rose Innes, R. (1979) - Mtwara/Lindi Regional Integrated Development Programme, Report of the zonal survey team in phase 2, Vol. 1, The physical environment. Land Resources Development Centre, Surrey. De Pauw, E. (1984) - Soils, physiography and agro-ecological zones of Tanzania. Ministry of Agriculture, Dar es Salaam/FAO, Rome. Dikau, R. (1989) - The application of a digital relief model to landform analysis in geomorphology. In: Raper, J. (Ed.) Three dimensional applications in geographical information systems, Taylor and Francis, London, pp 51-77. Dondeyne S., E.L. Ngatunga, S. Mugogo, N. Cools & J. Deckers (2001) Landscapes and soils of South Eastern Tanzania: their suitability for cashew. In: Proceedings of the 19th Conference of the Soil Science Society of East Africa, Moshi, Tanzania, 2-7 December 2001. ESRI (1993) - PC SEM - Structured Elevation Model, Version 1.3.0. SEM Users guide, Environmental Systems Research Institute Germany, Kranzberg. ESRI (1994) - PC Arc/Info - the world’s leading desktop geographical information system. GIS by ESRI - User guides, Environmental Systems Research Institute, Inc., Redlands. FAO (2000) - FAO Statistical database, domain Agricultural production. URL http://apps.fao.org/default.htm. Update of 5 April 2000. FAO, Rome. Italo-Tanzanian Cashew Research Programme (1981) - Reconnaissance soil map (first approach), Ruvuma region, Tunduru district. Ministero degli affari

Dondeyne et al.

192

esteri, Dipartimento per la cooperazione allo sviluppo. Istituto Agronomico per l’Otremare, Firenze. Ngatunga E., N. Cools, S. Dondeyne, J.A. Deckers & R. Merckx (2001) Buffering capacity of cashew soils in South Eastern Tanzania. Soil Use and Management, 17(3): 155-162. Ngatunga E., S. Dondeyne & J.A. Deckers (2003) Is sulphur acidifying cashew soils of South Eastern Tanzania? Agriculture, Ecosystems and Environment, 95(1): 179-184. Topper, C.P., Martin, P.J., Katinila, N, Kikoka, L.P., Lamboll, R., Masawe, P.A.L. and Shomari, S.H. 1998 The historical and institutional background of the Tanzanian cashew industry. In: Topper C.P. et al. (Eds.), Proceedings of the International Cashew and Coconut Conference: Trees for life, the key to development, BioHybrids International, Reading, pp 76-83.