Modeling sugarcane development and growth within ECOSMOS biophysical model

Abstract

Sugarcane plays an important role in electricity and sugar production and is a viable biofuel. Developing and optimizing a mechanism that can predict crop growth and yield at different spatiotemporal scales can promote the understanding of the effects of cultivation on the ecosystem, while providing options for optimizing management measures and improving the operational procedures of sugarcane growers. The main objective of this study is to integrate the sugarcane module into the ECOSystem MOdel Simulator (ECOSMOS) model and calibrate a parameter set for sugarcane genotypes groups (using different datasets); the model supports datasets that vary in complexity (from flux tower experiments to operational plots), while accounting for high genotype-by-environment-by-management (GxExM) variability. First, we calibrated the ECOSMOS biophysical and physiological parameters for the sugarcane module using two micrometeorological experimental sites, based on eddy-covariance and biomass measurements. Second, sugarcane genotypes located in different regions of contrasting climate conditions were split into two groups based on their period of harvest, i.e., early or mid-to-late harvest season, and two parameter sets were proposed. The sugarcane module was used to estimate the yield of numerous plots, using two different parameter sets, namely, the general and regionally-specific parameter sets. The model could successfully simulate the biophysical and physiological processes of the biomass of stalks and leaves, energy and carbon fluxes, and soil-water dynamics; for Experimental Site 2, the Nash-Sutcliffe efficiency (NSE) was 0.14–0.86 and the relative root mean square error (RRMSE) was 13–112. However, the generic parameter set did not perform well in all production environments, and the difference between the observed and simulated yields ranged from 0.9 to 14.5 (Mg ha-1). Hence, a novel calibration approach adopted in this study improved the module's accuracy, while improving the performances for all five production environments, with the difference between the observed and simulate yields being 0.3–2.2 (Mg ha-1). Although the two parameter sets can be used as a reference for sugarcane plantations in Brazil, we recommend recalibrating the model (for ensuring higher accuracy) before operational applications. Notably, the ECOSMOS-sugarcane model is emerging as a complex ecosystem model that can support the quantifications and evaluations of the effects of sugarcane plantations on the carbon and water balances in different environmental conditions, particularly in tropical regions.