Science Service System
You are here : Home : Proposals_Summary

Summary of Proposal LAN3317

TitleCrop phenology and its relation to soil moisture; a case study of Ifakara, Tanzania and Nai Swamp, Kenya.
Investigator Kirimi, Fridah - University of Bonn, Geography, Remote Sensing Research Group
Team Member
Mr. Kuria, Bartholomew - university of bonn, geography
Dr Thonfeld, Frank - university of bonn, geography
Prof Menz, Gunter - university of bonn, geography
SummaryWetlands form important ecosystemsas they have diverse functions including agricultural, hydrological andecological. These areas are vital in enhancing food security throughagricultural production as they have sufficient water throughout the seasons.However, there is need to control and manage the activities within the wetlandsto ensure maximum production is attained while negating any negative effectscaused by the activities. This is implemented through the decisions made by theland managers within the wetlands. Equipping the users of the wetland withscientific knowledge acts a support tool from which they can base theirdecisions on to ensure sustainable use of the wetlands. It is with this respectthat this research will be carried out. Microwave imagery has all weatherability and is independent from solar irradiation. The microwave region is usedin soil moisture estimation due to its unique response to dielectric constantwhich is related to presence of soil moisture. The objectives of the researchinclude delineating and differentiating the Ifakara and Nai swamp wetlands usingmulti sensor datasets and retrieving the soil moisture. The proposed data sources to be usedare TerraSAR-X images and Sentinel 2 images. SAR and optical image fusion willbe carried out to improve the visual appearance of the SAR image. Thetechniques that will be used in delineation and differentiation of the wetlandswill include the use of NDVI and texture analysis from the SAR images. SupportVector Machine classification method will be used due to its ability togenerate hyper planes which enables separation of classes including thoseclasses that are linearly non separable. The backscatter recorded by thesensor is a function of the soil moisture, vegetation and surface roughness.The backscatter due to vegetation will be obtained from vegetation indiceswhereas that due to surface roughness will be obtained from inversion ofphysical models such as the Integral Equation Model. Field data collection ofsoil moisture and roughness will also be carried out concurrently as thesatellite overpass. The roughness and vegetation data will then be used in asoil moisture retrieval algorithm.It isalso proposed that the radar derived soil moisture be used to calibrate andtest a Soil Vegetation Atmosphere Transfer (SVAT) model to study the accuracyof the SVAT model and the probability of using that model in retrieval ofmoisture at higher temporal resolutions over the study site. The deliverables include classifiedland use maps and soil moisture maps of the Ifakara and Nai swamp and thespatial and temporal soil moisture variation analysis. A multi temporal dataset is necessary in assessing the land cover dynamics since radar signalbehavior will vary over time depending on the vegetation phenology stages andtillage practices. The results obtained can further be used as input inhydrological models as well as disseminated to farmers and stakeholders to aidin implementation of production techniques which will ensure maximum yieldswhile minimizing wastage and unsustainable use of the natural resources. Inaddition, it will provide a basis upon which continued monitoring andevaluation of the performance and health of the wetland will be carried out inthe future.

Back to list of proposals

DLR 2004-2016