Abstract Grasslands offer many environmental and economic advantages that put them at the heart of future sustainable ruminant production systems. This study aimed to quantify and map the dry matter yield (DMY) and nitrogen yield (NY) of French grasslands resulting from cutting and grazing practices, based on the existing diversity of grassland vegetation, management, soil and climate conditions, using a research version of the STICS crop model called PâturSTICS. This model simulates daily dry matter (DM), nitrogen (N) and water fluxes involved in the functioning of grasslands and crops in response to management and environmental conditions. It was improved to represent deposition of animal waste on grassland soils during grazing and to simulate DM production and N content of grasses and legumes more accurately. Simulations were performed for locations across France on a high-resolution grid composed of pedoclimatic units (PCU) obtained by combining the spatial resolutions of climate and soil. The main grassland types and associated management types were determined for each PCU and then simulated over 30 years (1984–2013). Using the simulated values, predictive metamodels of annual grassland DMY and NY were developed from easily accessible explanatory variables using a random forest approach. Annual model predictions were aggregated and averaged at the PCU scale, then compared to regional observations. Predicted DMY agreed with available observations, except in semi-mountainous and mountainous regions, where PâturSTICS tended to overpredict DMY, probably because it ignores effects of snow, frost and slope, and due to how it represents effects of temperature and water stress on plant growth. According to results, three-quarters of French grasslands produce and export at least 7.6 t DM ha−1 yr−1 and 172 kg N ha−1 yr−1, respectively. One-quarter of French grasslands produce and export at least 10.7 t DM ha−1 yr−1 and 254 kg N ha−1 yr−1, respectively. The latter are located mainly in north-western France, the north-western Massif Central, the French Alps and the western Pyrenees, all of which have environmental conditions favourable for grass growth. The metamodels developed are interesting proxies for PâturSTICS’ predictions of grassland DMY and NY. Our results provided valuable knowledge that promotes better use of the potential forage production of French and European grasslands to improve protein self-sufficiency and N fertilisation management in ruminant livestock systems.

中文翻译：

---

法国草原干物质和氮产量的高分辨率评估

摘要 草原提供许多环境和经济优势，使其成为未来可持续反刍动物生产系统的核心。本研究旨在根据草原植被、管理、土壤和气候条件的现有多样性，使用研究版本对法国草原的干物质产量 (DMY) 和氮产量 (NY) 进行量化和绘制地图称为 PâturSTICS 的 STICS 作物模型。该模型模拟了草地和作物在应对管理和环境条件时的功能所涉及的每日干物质 (DM)、氮 (N) 和水通量。它被改进以表示放牧期间动物粪便在草地土壤上的沉积，并更准确地模拟草和豆类的 DM 产量和 N 含量。在由气候和土壤空间分辨率相结合获得的气候单位 (PCU) 组成的高分辨率网格上对法国各地的位置进行了模拟。确定每个 PCU 的主要草地类型和相关管理类型，然后模拟 30 年（1984-2013 年）。使用模拟值，使用随机森林方法从易于访问的解释变量中开发了年度草原 DMY 和 NY 的预测元模型。年度模型预测在 PCU 尺度上汇总和平均，然后与区域观察结果进行比较。预测的 DMY 与现有的观察结果一致，除了在半山地和山区，PâturSTICS 倾向于高估 DMY，可能是因为它忽略了雪、霜和坡度的影响，以及由于它如何代表温度和水分胁迫对植物生长的影响。根据结果​​，四分之三的法国草原分别生产和出口至少 7.6 t DM ha-1 yr-1 和 172 kg N ha-1 yr-1。四分之一的法国草原分别生产和出口至少 10.7 吨 DM ha-1 yr-1 和 254 kg N ha-1 yr-1。后者主要分布在法国西北部、西北部中央高原、法国阿尔卑斯山和比利牛斯山脉西部，均具有有利于草类生长的环境条件。开发的元模型是 PâturSTICS 对草地 DMY 和 NY 预测的有趣代理。