University of Twente Student Theses


Soil moisture temporal stability and its application in remote sensing products validation Using the example of Maqu (China) and Twente (the Netherlands)

Wu, Jiexia (2010) Soil moisture temporal stability and its application in remote sensing products validation Using the example of Maqu (China) and Twente (the Netherlands).

[img] PDF
Abstract:Surface soil moisture plays a key role in hydrology, climate and agriculture; however the conventional in-situ point measurements are not appropriate to represent areal soil moisture in long term, due to high spatial and temporal variability. Remote sensing for example Advanced Microwave Scanning Radiometer (AMSR-E) can give soil moisture information of the surface layer at large scales throughout a long period, however the validation of these products is very critical, because of scale “jump” problem between in situ measurements and remote sensing products. In this study, soil moisture measurements from two networks, one in Twente (the Netherlands) consisting of 17 stations with available data during study period (from 1st April to 15th November), and the other in Maqu (China) consisting of 20 stations from 30th June 2008 to 12th May 2009 were used to do upscaling for AMSR-E soil moisture products validation. First, the representativity of stations installed under grassland in the Twente region with heterogeneous land cover types was tested by a 5 days’ fieldwork with intensive sampling at 8 stations during September and October 2009.All but 2(station st11 and station st17) stations were proved to be representative. At footprint scale, soil moisture based on stations in this region can represent this region. Geostatistical interpolation and temporal stability analysis were both to estimate soil moisture spatial distribution. Geostatistical interpolation cannot used in this study due to short soil moisture spatial correlation lengths (<250m) obtained by variogram analysis and big errors caused by interpolation While after applying temporal stability analysis for long term and for different seasons. Representative stations nst13, cst2 and st03, st15 in Maqu and Twente were identified. Correlations between different stations support the results. Point measurements based on representative station were upscaled to areal soil moisture, by using absolute differences, linear regression and higher order regression, the results was evaluated by RMSE, BIAS, R. The best upscaling algorithms were applied to validate AMSR-E soil moisture products by using R and RMSE for evaluation. In the Maqu region, at descending time high correlation of 0.73 and low RMSE 4.21%(vol /vol) were observed while quite low correlation and big RMSE 0.003 and 12.53%(vol /vol) for ascending. In the Twente region, correlation at descending time was high about 0.7, while RMSE is 10.18 % (vol/vol), for ascending mode , correlation was also high 0.67, but RMSE was big about 16.32% (vol /vol).
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page