Abstract Tropical grasses are used as forage, to produce energy from biomass, for land restoration and carbon sequestration, among other applications. Many modelling approaches have been employed to simulate tropical grasses growth, but these have several limitations that must be solved by adapting them or creating new models. This study aimed to develop a tropical pasture model in the Agricultural Production Systems Simulator (APSIM) modelling framework, and to parameterise it to simulate Brachiaria brizantha ‘BRS Piata’ growth, under grazing and cut-and-carry management. For this, three field experiments were conducted in the South-east of Brazil where pasture growth was measured in a cut-and-carry system, with irrigated and rainfed treatments, and in a rainfed grazing system. Model evaluation was performed through precision and accuracy indices and simulation errors. Under cut-and-carry management, forage productivity was estimated with R2 values from 0.89 to 0.94, Willmott agreement indices between 0.97 and 0.98, and Nash-Sutcliffe Efficiency values of 0.88 to up to 0.92. This demonstrates the capacity of the APSIM-Tropical Pasture model to simulate tropical pastures. Simulation of phenology, early growth after sowing, partitioning and senescence during flowering, and reallocation and retranslocation of plant dry matter and nitrogen were important aspects for this capacity. Then, APSIM-Tropical Pasture can be used to simulate tropical pastures, but several requirements for further improvements have been identified, such as to improve the simulations of flowering for palisade grass, N effects on pasture yield, reallocation and retranslocation processes. Under grazing management, forage productivity was estimated with R2 = 0.80, Willmott agreement index of 0.91, and Nash-Sutcliffe Efficiency value of 0.62. Despite these reasonable results, simulations presented problems, since does not take into account the effect of grazing animals on pastures. This indicates that, in its current form, APSIM-Tropical Pasture is not able to simulate pastures under grazing effectively. However, this simulation of grazed system was important to identify main modelling constraints and direct future research to improve knowledge of processes and interactions needed for pasture model development.

中文翻译：

---

APSIM-热带牧场：模拟多年生热带草生长及其参数化栅栏草（Brachiaria brizantha）的模型

摘要 热带草被用作饲料，从生物质中生产能量，用于土地恢复和碳封存等应用。许多建模方法已被用于模拟热带草的生长，但这些方法有几个限制，必须通过调整它们或创建新模型来解决。本研究旨在在农业生产系统模拟器 (APSIM) 建模框架中开发热带牧场模型，并将其参数化以模拟 Brachiaria brizantha 'BRS Piata' 在放牧和扦插管理下的生长。为此，在巴西东南部进行了三项田间试验，在这些试验中，采用扦插系统、灌溉和雨育处理以及雨育放牧系统测量了牧场的生长情况。模型评估通过精度和准确度指标以及模拟误差进行。在随收随运管理下，估计牧草生产力的 R2 值为 0.89 至 0.94，威尔莫特一致性指数介于 0.97 至 0.98 之间，纳什-萨特克利夫效率值介于 0.88 至 0.92 之间。这证明了 APSIM-Tropical Pasture 模型模拟热带牧场的能力。物候模拟、播种后的早期生长、开花期间的分区和衰老以及植物干物质和氮的重新分配和再分配是这种能力的重要方面。然后，APSIM-Tropical Pasture 可用于模拟热带牧场，但已经确定了进一步改进的几个要求，例如改进栅栏草开花的模拟，氮对牧场产量的影响，再分配和再分配过程。在放牧管理下，估计牧草生产力为 R2 = 0.80，Willmott 一致性指数为 0.91，Nash-Sutcliffe 效率值为 0.62。尽管有这些合理的结果，模拟仍然存在问题，因为没有考虑放牧动物对牧场的影响。这表明，以目前的形式，APSIM-Tropical Pasture 无法有效模拟放牧下的牧场。然而，这种放牧系统模拟对于确定主要建模约束和指导未来的研究非常重要，以提高牧场模型开发所需的过程和相互作用的知识。和 Nash-Sutcliffe 效率值为 0.62。尽管有这些合理的结果，模拟仍然存在问题，因为没有考虑放牧动物对牧场的影响。这表明，以目前的形式，APSIM-Tropical Pasture 无法有效模拟放牧下的牧场。然而，这种放牧系统模拟对于确定主要建模约束和指导未来的研究非常重要，以提高牧场模型开发所需的过程和相互作用的知识。和 Nash-Sutcliffe 效率值为 0.62。尽管有这些合理的结果，模拟仍然存在问题，因为没有考虑放牧动物对牧场的影响。这表明，以目前的形式，APSIM-Tropical Pasture 无法有效模拟放牧下的牧场。然而，这种放牧系统模拟对于确定主要建模约束和指导未来的研究非常重要，以提高牧场模型开发所需的过程和相互作用的知识。