Assessment of crop productivity over intensively managed agriculture regions in India and Australia using solar-induced fluorescence remote sensing data

Abstract:

Satellite based estimation of solar-induced terrestrial fluorescence (SIF) is considered to be a direct measure of photosynthetic functional status of the vegetation. Prior studies have shown SIF to more accurately retrieve the productivity of intensively managed croplands, as in the U.S. corn belt. In this study, we assessed and compared agricultural productivity over two intensive crop production regions in Australia and India using SIF data, traditional spectral measures, and crop yield data. Regional level wheat yield data were obtained for the Indo-Gangetic Plains (IGP) in India and the Murray Darling Basin (MDB) in Australia for analyses with GOME-2 SIF satellite and MODIS VI measurements, and gross primary productivity from flux towers. We investigated the importance of integrating traditional meteorological parameters and ground based data with time-series vegetation indices for scaling of SIF to obtain robust yield prediction models for application across years and continents. This study further explored the relationship of inter annual variations in crop phenology metrics through SIF retrievals and its relationship with crop yields.

The IGP study region showed systematic cycles of double cropping. MDB region on the other hand showed cycles of pronounced winter cropping and a weaker and variable second cropping over the analysis period. For various winter wheat crop seasons in IGP, from 2007 to 2012, SIF explained and accounted between 48 to 74 per cent of the variations in regional wheat yields. Similar results were obtained in the case of MDB also, however, the relationship between SIF and yield estimates was weaker (R² = 0.44). SIF measurements, as a surrogate of crop productivity, were considerably higher over the highly productive IGP region in almost all the years considered. The SIF data shows immense potential for modelling agricultural productivity, particularly as the resolution of SIF retrievals continues to improve.