University of Twente Student Theses


Estimation of Evapotranspiration from Satellite Remote Sensing and Meteorological Data over the Fogera Flood Plain - Ethiopia

Enku, Temesgen (2009) Estimation of Evapotranspiration from Satellite Remote Sensing and Meteorological Data over the Fogera Flood Plain - Ethiopia.

[img] PDF
Abstract:Evapotranspiration (ET) is not only an important component in the land surface energy balance, but also major component in the water cycle and an important parameter in water resources management. Conventional methods for ET estimation which use meteorological data can only be used to estimate ET at certain location, but not at a larger scale. This limitation can be overcome by using a combination of ground based and remote sensing data in a surface energy balance model. In this study, the annual cycle of ET in for the Fogera flood plain, in northwest Ethiopia is estimated, using MODIS images, weather station data and the surface energy balance model SEBS. The results were compared to conventional ground based methods. The Fogera flood plain is located east of Lake Tana, in northwest Ethiopia. The area is about 500 km2 and elevation of 1800 m, flooding occurs frequently during the rainy season. Rice is the major crop. A meteorological station was installed in the flood plain in June 2008, recording standard meteorological values at 5-min interval. In addition, eddy covariance instruments were installed for 3 days in September 2008. MODIS products were downloaded and processed with monthly intervals. Weather station data and the MODIS products were used as input for SEBS to calculate monthly images of ET of the whole floodplain. For months where no weather station data were available, daily data of the nearby weather station in Bahir Dar were used. Eddy covariance data of three days were used as validation. The annual actual ET from SEBS in the year 2008 was 1519 mm, while P-M reference ET was 1498 mm. The annual ET over the area was larger by about 17 % from the mean annual rainfall of the floodplain. The excess ET was from the flooding caused mainly by the rivers. The spatial and temporal variation of ET over the flood plain was also evaluated. The actual ET from SEBS was larger than the P-M ETo during the wet seasons and lower in drier periods. Spatial variation of ET was well pronounced in drier periods than wet seasons. The instantaneous albedo from the remote sensing method was less by 15% from the daily albedo computed from the eddy flux tower observation. The ground based conventional methods Penman-Monteith (P-M), Priestley-Taylor, Makkink, and Abtew simple equations were used for the estimation of reference ET. A comparison of these methods showed that simple radiation methods like Makkink and Abtew equations perform as well as P-M in the area. A sensitivity analysis of P-M for input variables showed that reference ET is most (and equally) sensitive to incoming solar radiation and air temperature, and least sensitive to wind speed. This explains why simple ET methods based on either radiation or temperature perform well. The temperature method that only uses maximum temperature was used in this area. This method was found to perform comparable with other conventional methods. With this method, it was possible to estimate ETo with less than 10 % error from P-M reference ET when daily estimations are compared and with in 2 % error when yearly estimations compared. Key words: Evapotranspiration, MODIS, SEBS, Penman-Monteith
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