Abstract. Grasslands cover around two thirds of the land area of Great Britain (GB) and are important reservoirs of terrestrial biological carbon (C). Outside a few well-monitored sites the quantification of C dynamics in managed grasslands is made complex by the spatio-temporal variability of weather conditions combined with grazing and cutting patterns. Earth observation (EO) missions produce high-resolution frequently-retrieved proxy data on the state of grassland canopies but synergies between EO data and biogeochemical modelling to estimate grassland C dynamics are under-explored. Here, we show the potential of model-data fusion (MDF) to provide robust near-real time analyses of managed grasslands of GB (England, Wales andScotland). We combine EO data and process-based modelling to estimate grassland C balance and to examine the role of management. We implement a MDF algorithm to (1) infer grassland management from vegetation reduction data (Proba-V), (2) optimise model parameters by assimilating leaf area index (LAI) data (Sentinel-2) and (3) simulate livestock grazing, grass cutting, and C allocation and loss to the atmosphere. The MDF algorithm was applied for 2017 and 2018 at 1855 fields sampled from across GB. The algorithm was able to effectively assimilate the Sentinel-2 based LAI time series (overlap = 80 %, RMSE = 1 gCm⁻², bias = 0.35 gCm⁻²) and predict livestock densities per area that correspond with independent census-based data (r = 0.68). The mean total removed biomass across all simulated fields was 6 (±1.8) tDM ha⁻¹ y⁻¹. The simulated grassland ecosystems were on average C sinks in 2017 and 2018; the GB-average net ecosystem exchange (NEE) and net biome exchange (NBE) for 2017 was −232 ± 94 and for 2018 was −120 ± 103 gCm⁻² y⁻¹. The 2018 summer drought reduced C sinks, with a 9-fold increase in the number fields that were C sources (NBE > 0) in 2018 compared to 2017. We conclude that management in the form of sward condition and the timing, intensity and type of defoliation are key determinants of the C balance of managed grasslands. Nevertheless, extreme weather, such as prolonged droughts, can convert grassland C sinks to sources.

中文翻译：

---

受地球观测约束的英国管理草原的碳收支

摘要。草原约占英国 (GB) 陆地面积的三分之二，是重要的陆地生物碳 (C) 库。在一些监测良好的场地之外，由于天气条件的时空变化以及放牧和切割模式，管理草地中碳动态的量化变得复杂。地球观测 (EO) 任务产生了关于草地冠层状态的高分辨率、经常检索的代理数据，但 EO 数据与用于估计草地 C 动力学的生物地球化学模型之间的协同作用尚未得到充分探索。在这里，我们展示了模型数据融合 (MDF) 的潜力，可以对英国（英格兰、威尔士和苏格兰）的管理草地提供可靠的近实时分析。我们结合 EO 数据和基于过程的建模来估计草地碳平衡并检查管理的作用。我们实施 MDF 算法以 (1) 从植被减少数据 (Proba-V) 推断草地管理，(2) 通过同化叶面积指数 (LAI) 数据 (Sentinel-2) 和 (3) 模拟牲畜放牧，优化模型参数，割草，C分配和大气损失。MDF 算法应用于 2017 年和 2018 年的 1855 个来自 GB 的采样字段。该算法能够有效地同化基于 Sentinel-2 的 LAI 时间序列（重叠 = 80 %，RMSE = 1 gCm^-2，偏差 = 0.35 gCm ^-2 ) 并预测与独立的基于人口普查的数据相对应的每个区域的牲畜密度 ( r  = 0.68)。所有模拟田间的平均总去除生物量是 6 (±1.8) tDM ha ^-1  y ^-1。2017年和2018年模拟草地生态系统平均为C汇；2017 年 GB 平均净生态系统交换 (NEE) 和净生物群系交换 (NBE) 为 -232 ± 94，2018 年为 -120 ± 103 gCm ^-2  y ^-1. 2018 年夏季干旱减少了碳汇，2018 年碳源田数（NBE > 0）比 2017 年增加了 9 倍。落叶的数量是管理草地碳平衡的关键决定因素。然而，极端天气，如长期干旱，可以将草地碳汇转化为碳源。