Estimation of Crop Coefficients Using Satellite Remote Sensing

Crop coefficient (K-c) based estimation of crop evapotranspiration (ETc) is one of the most commonly used methods for irrigation water management. The standardized FAO56 Penman-Monteith approach for estimating ETc from reference evapotranspiration and tabulated generalized K-c values has been widely adopted worldwide to estimate ETc. In this study, we presented a modified approach toward estimating K-c values from remotely sensed data. The surface energy balance algorithm for land model was used for estimating the spatial distribution of ETc for major agronomic crops during the 2005 growing season in southcentral Nebraska. The alfalfa-based reference evapotranspiration (ETr) was calculated using data from multiple automatic weather stations with geostatistical analysis. The K-c values were estimated based on ETc and ETr (i.e., K-c=ETc/ETr). A land use map was used for sampling and profiling the K-c values from the satellite overpass for the major crops grown in southcentral Nebraska. Finally, a regression model was developed to establish the relationship between the normalized difference vegetation index (NDVI) and the ETr-based crop coefficients (K-cr) for corn, soybeans, sorghum, and alfalfa. We found that the coefficients of variation (CV) for NDVI, as well as for K-cr of crops were lower during the midseason as compared to the early and late growing seasons. High CV values during the early growing season can be attributed to differences in planting dates between the fields, whereas high CVs during the late season can be attributed to differences in maturity dates of the crops, variety, and management practices. There was a good relationship between K-cr and NDVI for all the crops except alfalfa. Validation of the developed model for irrigated corn showed very promising results. There was a good correlation between the NDVI-estimated K-cr and the Bowen ratio energy balance system based K-cr with a R-2 of 0.74 and a low root mean square difference of 0.21. This approach can be a very useful tool for a large (watershed or regional) scale estimation of evapotranspiration using the crop coefficient and reference evapotranspiration approach.