Relations between evaporation coefficients and vegetation indices studied by model simulations

Abstract

Calculations using a heat balance and a radiative transfer model have been done to study relations among evaporation coefficients and vegetation indices. The evaporation coefficients are the crop coefficient (defined as the ratio of total evaporation and reference crop evaporation) and the transpiration coefficient (defined as the ratio of unstressed transpiration and reference crop evaporation), while the vegetation indices considered in this study are the normalized difference, soil adjusted vegetation index, and transformed soil adjusted vegetation index. The reference crop evaporation has been calculated using the Priestley-Taylor equation. The observed variations of crop (wheat) height, leaf area index, and weather conditions for 30 days at Phoenix (Arizona), together with the reflectances of different types of soil in wet and dry states, are used in the simulation. The total evaporation calculated from the model compared well with lysimeter observations. Variations in soil evaporation can introduce considerable scatter in the relation between the crop coefficient and leaf area index, while this scatter is much less for the relation between transpiration coefficient and leaf area index. The simulation results for 30 days of crop and weather data and reflectances of 19 soil types in wet and dry conditions gave significant linear correlations between the transpiration coefficient and the vegetation indices, the explained variance (r 2) being highest for the soil adjusted vegetation index ( r 2 = 0.88) and lowest for the normalized difference ( r 2= 0.81). A clump model is used to address the effect of spatial heterogeneity on the relationship between the transpiration coefficient and soil adjusted vegetation index. These simulated relationships between transpiration coefficient and vegetation indices for wheat are discussed in the context of the relationships derived from observations for several crops and grasses. The present analysis provides a theoretical basis for estimating transpiration from remotely sensed data.