University of Twente Student Theses


Validation of Surface Energy Balance System (SEBS) over forest land cover and sensitivity analysis of the model

Badola, Anupam (2009) Validation of Surface Energy Balance System (SEBS) over forest land cover and sensitivity analysis of the model.

[img] PDF
Abstract:In order to estimate evapotranspiration over regional scale, a Surface Energy Balance System (SEBS) has been developed recently. SEBS is a powerful tool to estimate actual evapotranspiration over local, regional and continental scale using remote sensing and standard meteorological data. It is based on the principle of surface energy balance closure, and variables such as net radiation, soil heat flux, etc. have been parameterized in the model. SEBS calculates evapotranspiration using optical and thermal remote sensing beside weather station data. In this study, SEBS was for the first time tested for forest land cover over the central part of the Netherlands using high resolution remote sensing data of ASTER. Besides, four methods for calculation of fractional cover and two methods for calculation of soil heat flux were validated using ground data. Results showed that the method proposed by Gutman and Ignatov is the best of the tested algorithms for retrieval of fractional vegetation cover from remote sensing data. The two methods validated for soil heat flux estimation showed that both of them were overestimating the soil heat flux. Comparison of the estimated evaporative fraction with validation data confirms that SEBS can estimate evaporative fraction with RMSE 0.47 [-]. However, there are huge differences in the estimation of sensible and latent heat flux RMSE >50 Wm-2. It has been found that image resolution (90m ASTER data) did not have any impact on the accuracy of results. It was assumed that it is due to large fetch of flux tower, (300-500m2) used for validation. Sensitivity analysis of the SEBS model was carried out on remote sensing inputs to SEBS model. Results show that amongst surface albedo, emissivity, land surface temperature, fractional vegetation cover and NDVI, the SEBS model is most sensitive to land surface temperature with sensible heat increasing with exponential rate beyond 350C surface temperature and decreasing exponentially below 200C. This exponential increase/decrease in sensible heat results in abrupt changes in evaporative fraction.
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