A new perennial forage module coupled with the ECOSMOS terrestrial ecosystem model: Calibration and evaluation for Urochloa (syn. Brachiaria) brizantha

Abstract

Pastures represent the main land use type in Brazil and are used for livestock production and maintenance of land ownership. Most pasture areas exhibit signs of degradation related to inadequate management, which affects pasture biomass production. Terrestrial ecosystem models are effective tools for evaluating the economic and ecological returns associated with distinct management scenarios. In this paper, a new forage module implemented in the ECOSystem MOdel Simulator (ECOSMOS) is formulated, and ECOSMOS-Forage model calibration and evaluation results were obtained for estimating tropical forage growth under continuous and rotational stocking methods subject to grazing or cutting regimes. A set of equations to simulate the daily carbon allocation, growth, senescence, and morphological characteristics of Urochloa (syn. Brachiaria) brizantha cvs. Marandu and Piatã were proposed. Biophysical and physiological parameters, including the solar radiation balance, soil water content, evapotranspiration and carbon assimilation, were calibrated and evaluated with one micrometeorological experiment. Additionally, two sets of parameters were calibrated for Marandu grass growth under continuous and rotational stocking. Based on the Marandu parameterization, a third set of parameters, thereby adjusting a few parameters, was derived for the Piatã grass under rotational stocking. The ECOSMOS-Forage model could successfully capture the energy, carbon and water fluxes and biomass dynamics during growth cycles. The net radiation, evapotranspiration and gross primary production were simulated with agreement index (d) values of 0.97, 0.89, and 0.95, respectively, during the calibration phase and 0.98, 0.90, and 0.93, respectively, during the evaluation phase. Regarding model application to simulate forage growth under rotational stocking, the model could estimate aboveground, leaf and stem dry matter accumulation levels with d values ranging from 0.91 to 0.98. Although forage growth is relatively difficult to simulate, compared with annual crops, the results suggested that the model could provide high potential for simulating tropical perennial forage grasses.