Remote estimation of gross primary productivity in soybean and maize based on total crop chlorophyll content

Abstract

The synoptic quantification of crop gross primary productivity (GPP) is essential for studying carbon budgets in croplands and monitoring crop status. In this study, we applied a recently developed model, which relates crop GPP to a product of total crop chlorophyll content and incoming photosynthetically active radiation, for the remote estimation of GPP in two crop types (maize and soybean) with contrasting canopy architectures and leaf structures. The objective of this study was to evaluate performances of twelve vegetation indices used for detecting different vegetation biophysical characteristics, in estimating GPP of rainfed and irrigated crops over a period from 2001 through 2008. Indices tested in the model exhibited strong and significant relationships with widely variable GPP in each crop (GPP ranged from 0 to 19 gC/m 2/d for soybean and 0 to 35 gC/m 2/d for maize), however, they were species-specific. Only three indices, which use MERIS red edge and NIR spectral bands (i.e. red edge chlorophyll index, MERIS Terrestrial Chlorophyll Index and red edge NDVI), were found to be able to estimate GPP accurately in both crops combined, with root mean square errors (RMSE) below 3.2 gC/m 2/d. It was also shown that two indices, red edge chlorophyll index and red edge NDVI with a red edge band around 720 nm, were non-species-specific and yielded a very accurate estimation of GPP in maize and soybean combined, with RMSEs below 2.9 gC/m 2/d and coefficients of variation below 21%.