Science Service System
You are here : Home : Proposals_Summary

Summary of Proposal LAN1393

TitleSoil moisture variability mapping in the Malinda wetland using TerraSAR data
Investigator Kuria, David - University of Bonn, Geography
Team Member
Professor Misana, Salome - University of Dar-es-Salam, Geography
Professor Menz, Gunther - University of Bonn, Geography
SummaryIntroduction For the detection of wetland, techniques are required that have the capabilities to differentiate between dry and wet terrestrial areas as well as to detect variable degrees of soil moisture contents. Since the microwave region of the electromagnetic spectrum is primarily sensitive to dielectric properties of the ground targets and therefore sensitive to soil moisture, microwave sensing techniques provide data that can be seen as complementary and supplementary to that of the optical region of the spectrum. Another unique capability of microwave sensor like the synthetic aperture radar (SAR) is that a portion of the energy penetrates into the ground surface and hence provides sub-canopy information, which in particular, matters in the context of soil moisture detection. Besides this sensitivity, the detection of wetlands using microwave remote sensing is independent of weather conditions as clouds are penetrated at wavelengths > 2 cm and rainfall at wavelengths > 4 cm. Objectives • Determine the profile as well as the seasonal and spatial variations of soil moisture from radar satellite imagery and TDR-based ground-truthing • Adapt the semi-empirical backscatter model for soil moisture estimation from radar imagery. • Provide linakges to the potential and actual yields of crops and the biomass of the semi-natural or crop-associated weedy vegetation to the key production-limiting factors and their relative importance by modeling approaches and extrapolate the yield gap of major crops spatially to all initially inventoried 51 wetlands of the project areas by GIS (spatial modeling / up-scaling). Methods Several processing steps have to be applied to data sets to derive estimations of spatio-temporal soil moisture distribution by using a semi-empircal backscatter model (after Galarneau et al 2001) based on multi-frequency ALOS PALSAR (L-band, 6,25 m resp. 12,5 m spatial resolution and 46 days recurrence cycle), RADARSAT (C-band, 3 m spatial resolution and 24 days repeat cycle) and TerraSAR-X (X-band, 1 m pixel size and 11 days repetition rate). The image processing chain includes different aspects such as geo-coding and multi-looking, speckle filtering, texture analysis and decibel conversion, image segmentation, backscatter statistics, processing of the data sets like binary wetland classification, decision tree classification and the implementation of the semi-empirical backscatter model for the selected SWEA wetlands in Kenya and Tanzania. The calibration and validation of the regionalized soil moisture model will be based on regular soil moisture measurements along representative transects within the Malinda wetlands in Tanzania using a TDR probe. Data requirements Data obtained in Spotlight, StripMap and ScanSAR modes for diverse dates (dry period: Jan –March, Aug – Sept; wet periods: April – June, October – November) Data required from TerraSAR-X, ALOS-PALSAR and RADARSAT sensors. TDR probe observations for same duration Satellite derived soil moisture products Deliverables The expected outcome from this WP3 will cover two main aspects: 1. improved delineation and differentiation land cover and land use classification of the selected wetlands. For the definition of the natural und agricultural used areas will be performed in close collaboration the WP1, WP2 & WP 4. 2. spatio-temporal distribution assessment of soil moisture based on the high resolution semi-empirical backscatter model (mean spatial resolution: 10 m and average temporal resolution: 30 days). The output of the semi-empirical radar backscatter model will be quantitative soil moisture maps and serve as an input variable into the potential crop modeling and for the assessment of the production potential of natural vegetation.

Back to list of proposals

© DLR 2004-2016