Magnitude, drivers, and patterns of gross primary productivity of rice in Arkansas using a calibrated vegetation photosynthesis model

Abstract

An estimate of the gross primary productivity (GPP) of rice fields is instrumental for understanding both their harvest yield and landscape greenhouse gas dynamics. Rice contributes $1.7 billion annually to Arkansas’s economy, however, there is a lack of understanding of the spatial variation of rice’s GPP and its predictors. We employ the satellite-based vegetation photosynthesis model (VPM) to estimate GPP for Arkansas rice cropland and evaluate our model findings against 16 site-seasons in-situ data (eddy covariance; EC). At the site scale (individual ∼10–30 ha rice fields), the results evaluated against 16 site-seasons revealed that the VPM with site-measured meteorological information (R² = 0.6, mean absolute error (MAE) = 3.68 g C m⁻² day^-1, and bias = -0.33 g C m⁻² day^-1) only slightly outperforms the VPM based on gridded meteorological information (R² = 0.57, MAE = 3.75 g C m⁻² day⁻¹, and bias = -0.22 g C m⁻² day⁻¹). Across the state’s rice fields (2008–2020), the mean photosynthetic carbon uptake of Arkansas rice fields was modeled as 1801 ± 288 g C m⁻² year⁻¹. Across rice production ecological zones, significant differences in the enhanced vegetation index and land surface water index between the Grand Prairie and Middle Delta regions, despite similar temperature and photosynthetic active radiation, suggest that regional agronomic and soil differences are more influential on GPP variation than climatic factors. The modeled annual cumulative GPP was positively correlated (R² = 0.16) with the county-scale yield. Our results demonstrated the potential importance of agronomic practices in the Grand Prairie region and enhanced vegetation index for understanding the spatial variation of GPP.