Evaluating a satellite-based seasonal evapotranspiration product and identifying its relationship with other satellite-derived products and crop yield: A case study for Ethiopia

• Published in: International journal of applied earth observation and geoinformation 2015-08, Vol.40, p.39-54
• Main Authors: Tadesse, Tsegaye, Senay, Gabriel B., Berhan, Getachew, Regassa, Teshome, Beyene, Shimelis
• Format: Article
• Language: English
• Subjects: Crop yield; Drought; Early warning; Evapotranspiration; Food security; Risk management
• ISSN: 1569-8432; 1872-826X
• DOI: 10.1016/j.jag.2015.03.006

[Display omitted] •Developed a process to evaluate satellite-derived actual evapotranspiration (ETa) products.•Compared and evaluated ETa products with other satellite-derived data and crop yield.•Identified crop growing zones in Ethiopia that could potentially use the ETa products for monitoring crop yields during the long rainy season to support risk management and food security. Satellite-derived evapotranspiration anomalies and normalized difference vegetation index (NDVI) products from Moderate Resolution Imaging Spectroradiometer (MODIS) data are currently used for African agricultural drought monitoring and food security status assessment. In this study, a process to evaluate satellite-derived evapotranspiration (ETa) products with a geospatial statistical exploratory technique that uses NDVI, satellite-derived rainfall estimate (RFE), and crop yield data has been developed. The main goal of this study was to evaluate the ETa using the NDVI and RFE, and identify a relationship between the ETa and Ethiopia’s cereal crop (i.e., teff, sorghum, corn/maize, barley, and wheat) yields during the main rainy season. Since crop production is one of the main factors affecting food security, the evaluation of remote sensing-based seasonal ETa was done to identify the appropriateness of this tool as a proxy for monitoring vegetation condition in drought vulnerable and food insecure areas to support decision makers. The results of this study showed that the comparison between seasonal ETa and RFE produced strong correlation (R2>0.99) for all 41 crop growing zones in Ethiopia. The results of the spatial regression analyses of seasonal ETa and NDVI using Ordinary Least Squares and Geographically Weighted Regression showed relatively weak yearly spatial relationships (R2