REFERENCE CROP EVAPOTRANSPIRATION ESTIMATION USING REMOTE SENSING TECHNIQUE

Abstract

Reference crop evapotranspiration (ETo) is an important element for irrigation water management and its estimation is crucial. Many empirical and physical based approaches have been developed over the years for its estimation. However, lack of conventional ground weather station data, required by these approaches is a big challenge in a developing country like South Sudan, Eastern Africa, whose population depend mainly on rain-fed agriculture whose production diminished tremendously in recent years. Irrigation is being taken up as a remedy for diminished food production but due to non-availability of data for reference crop evapotranspiration, irrigation planning and management is seriously affected. Thus, this study proposes simple remote sensing technique for estimating monthly reference crop evapotranspiration without ground weather station data, using high spatial resolution remote sensing data, in dry season of South Sudan. The study evaluated the use of land surface temperature retrieved from Landsat 8 data as an alternative input for seven commonly used temperature based models (Blaney-criddle, Thornthwaite, Hargreaves (1985), Trajkovic (2007) modified Hargreaves, Droogers et al. (2002) modified Hargreaves, Allen (1993) modified Hargreaves and Kharrufa models) in Juba county of South Sudan. The proposed methodology has also been compared with analogous procedure proposed by Maeda et al. (2011) that use moderate resolution imaging spectroradiometer (MODIS) land surface temperature. Further, the study also proposes the automatic satellite image processing algorithms for each of the temperature based ETo methods for simplicity of images processing and calculations involve. For broader analysis, the proposed methodology was also tested in Roorkee region, India. The evaluation of the modelled results with FAO Penman Montieth results (using station data) as reference, shows that the models parameterized with Landsat 8 LST performed better than MODIS based, with low RMSE and MAE that ranges from 0.1064 to 0.1165mm/day and 0.0163 to 0.0997 mm/day respectively and high coefficients of determination (R2) of above 0.9. Hargreaves (1985) method was the best of all, in the study area (Juba County), with overall RMSE of 0.1064 mm/day, MAE of 0.0163mm/day and coefficient of determination (R2) of 0.93. ETo maps of study area (Juba County) were prepared using selected method (Hargreaves (1985). So considering the lack of ground station data for ETo estimation in the study area, the proposed methodology in this study may be used for monthly ETo estimation and will be of great help in planning, design and management of irrigation systems as well as other water management activities