Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Master → Pays Bas → Analysing performance of optical and passive mecrowave data to infer soil moisture in the upper soil layers for the prominent crops in eastern parts of Rajasthan, India

International Institute for Geo-Information Science and Earth Observation (ITC) 2008

Analysing performance of optical and passive mecrowave data to infer soil moisture in the upper soil layers for the prominent crops in eastern parts of Rajasthan, India

Ambika Mukund

Titre : Analysing performance of optical and passive mecrowave data to infer soil moisture in the upper soil layers for the prominent crops in eastern parts of Rajasthan, India

Auteur : Ambika Mukund

Etablissement de soutenance : University of Twente International Institute for Geo-Information Science and Earth Observation (ITC)

Grade : Master of Science in Geo-Information Science and Earth Observation 2008

Résumé
Soil Moisture is a key state variable of the energy and water cycle on earth surface. For areas like Rajasthan in India, which remain under constant threat of drought, comprehensive information on soil moisture variability can help in early drought prediction, drought monitoring and evaluation of drought impact on agricultural production. Limitations of acquiring adequate in-situ measurements and the high demand of data exhibited by models make remote sensing a highly viable option for extraction of knowledge about soil moisture variation at the root-zone. With the advent of optical remote sensed data with high spatial resolution and passive microwave data with high temporal resolution, it is desirable to use them to obtain soil moisture measurements for agricultural drought assessment. This study assesses the ability of Vegetation Temperature condition Index, developed from interpretation of NDVI –Ts space using surface reflectance and land surface temperature provided by Terra MODIS on an 8 day basis, to estimate soil moisture status at a spatial resolution of 1km. VTCI was attempted to be computed at 250mts by downscaling LST, and based on the agriculture land cover alone. Further estimation was done by regressing VTCI and 31 in-situ volumetric soil moisture measurements for the surface and at depths of 15cm, 30 cm and 45cm for the prominent crops in Eastern Rajasthan. The ability of passive microwave parameters, horizontally polarised brightness temperature (T BH ) and Polarisation Difference (PD) provided by Aqua – AMSRE on a daily basis at a coarse spatial resolution of 25 km, to assess soil moisture condition was also explored in the study. Comparison of the assessments by both VTCI and the passive microwave parameters were done with soil moisture simulate d using a GIS based simple water balance model. The model adopts the equation for water balance suggested by Thornthwaite and Mather (1955). The simulation is done for 5 * 5 km grids on a weekly basis and the inputs required are effective rainfall, maximum crop evapotranspiration and maximum available water capacity. The results for VTCI at 1km and computed for th e whole image was found to be more representative of the average profile soil moisture condition than the other two versions. A strong relation of VTCI and the average profile soil moisture till 50cms depth beneath the surface was observed with a coefficient of determination of 0.63. The relations hip was more relevant at the surface and at 30cm depth than 15cm and 45cm. All relationships were significant at 0.01 level for 2-tailed analysis. Estimation of average profile soil moisture from VTCI was done for 2003. The estimated volumetric soil moisture content showed high variability with different climatic region. The mean measure for semi-arid region was found to be 13.45% and 26. 19% for the sub-humid region. Soil moisture estimated from VTCI and that simulated using the SWBM, was found to vary with an RMSE of 7.04% for the study area, where the mean value is 14.2%. 62% in the semi-arid area and only 44% in the sub- humid area were found to be correctly estimated. Th e poor relation may be due to underestimation in the simulated soil moisture by the SWBM owing to uncertainties in the conventional methods used to compute its inputs. The passive microwave parameters exhibited significant relation with simulated soil moisture at a lower frequency of 6.9GHz. The correlation was better for semi-arid region than sub- humid region due to the influence of vegetation dens ity. The coefficient of determination was 0.78 for PD and 0.75 for T BH for semi-arid region and 6.9 GHz. It is found that T BH and PD can well depict the temporal variation in soil moisture but fail to represent the spatial patterns owing to the high geographical dependence of microwave data. It h as been concluded from these relationships that VTCI, T BH and PD can be used to infer soil moisture beneath the surface but in semi-arid areas, or specifically in sparse or moderately vegetated region.

Mots clés : Root zone Soil Moisture, VTCI, Microwave Temperature Brightness, Polarisation Difference, Terra- MODIS, Aqua- AMSRE, Simple Water Balance Model

Version intégrale (ITC )

Page publiée le 8 février 2018