APSIM model performance in simulating Piatã palisade grass growth and soil water in different positions of a silvopastoral system with eucalyptus

Abstract

CONTEXTSilvopastoral systems aim to be more sustainable than conventional systems, increasing total productivity, diversifying agricultural production, and improving resource use efficiency. However, very few studies were performed to adapt crop models to simulate these systems. OBJECTIVEThis study investigates the current capacity of the APSIM Next Generation model to estimate Piatã palisadegrass growth and soil water in different positions in a silvopastoral system with eucalyptus by looking at the impact of modelling belowground and aboveground competition separately. METHODSA field experiment was conducted from December 2014 to January 2016 in a silvopastoral system with eucalyptus trees arranged in single rows, in East-West orientation, with 15 m between rows and 2 m between trees in the rows. This experiment was conducted under grazing management and rainfed conditions, during 11 growth cycles, with pasture variables, soil moisture and microclimatic conditions assessed at four distances from the eucalyptus north row (0.00 m, 3.75 m, 7.50 m and 11.25 m). APSIM testing was performed in two ways: focusing on belowground competition for soil water and nitrogen, using a multi-root-zone approach, and focusing on competition for solar radiation, using the measured data of solar radiation at each position. RESULTS AND CONCLUSIONSThe APSIM-Tropical Pasture model was effective in simulating pasture growth when only competition for solar radiation was considered (R2 from 0.75 to 0.86, Agreement index (d) from 0.88 to 0.96, and Nash-Sutcliffe efficiency (NSE) between 0.32 and 0.85). These simulations are promising taking into account the difficulties to simulate pasture growth in silvopastoral systems with grazing; however, in systems with strong belowground competition for soil water and nutrients, this approach has strong limitations. Simulations of pasture growth were not as effective when using the multi-root-zone approach (R2 between 0.57 and 0.87, d between 0.53 and 0.74, and NSE from −5.32 to −1.20), despite reasonable simulations of soil water for the positions between the tree rows (R2 between 0.70 and 0.80, d from 0.91 to 0.94, and NSE between 0.60 and 0.77). In view of these results, improvements should be performed in APSIM to better simulate silvopastoral systems, mainly for solar radiation transmission and tree roots growth and distribution. SIGNIFICANCEThis study allowed identifying the key drivers of competition within the studied environment and to inform where the main efforts should be targeted in building a more effective model for silvopastoral systems. This is important to help model development teams in organising and optimising their efforts.