Improving remote estimation of winter crops gross ecosystem production by inclusion of leaf area index in a spectral model

Abstract

The hysteresis of the seasonal relationships between vegetation indices (VIs) and gross ecosystem production (GEP) results in differences between these relationships during vegetative and reproductive phases of plant development cycle and may limit their applicability for estimation of croplands productivity over the entire season. To mitigate this problem and to increase the accuracy of remote sensing-based models for GEP estimation we developed a simple empirical model where greenness-related VIs are multiplied by the leaf area index (LAI). The product of this multiplication has the same seasonality as GEP, and specifically for vegetative periods of winter crops, it allowed the accuracy of GEP estimations to increase and resulted in a significant reduction of the hysteresis of VIs vs. GEP. Our objective was to test the multiyear relationships between VIs and daily GEP in order to develop more general models maintaining reliable performance when applied to years characterized by different climatic conditions. The general model parametrized with NDVI and LAI product allowed to estimate daily GEP of winter and spring crops with an error smaller than 14%, and the rate of GEP over- (for spring barley) or underestimation (for winter crops and potato) was smaller than 25%. The proposed approach may increase the accuracy of crop productivity estimation when greenness VIs are saturating early in the growing season.