Department of Geography, Sussex University, Brighton, UK.
Department of Organismic and Evolutionary Biology, Harvard University, Cambridge, Massachusetts, USA.
Glob Chang Biol. 2022 Mar;28(5):1823-1852. doi: 10.1111/gcb.15995. Epub 2021 Dec 27.
Accurate descriptions of current ecosystem composition are essential for improving terrestrial biosphere model predictions of how ecosystems are responding to climate variability and change. This study investigates how imaging spectrometry-derived ecosystem composition can constrain and improve terrestrial biosphere model predictions of regional-scale carbon, water and energy fluxes. Incorporating imaging spectrometry-derived composition of five plant functional types (Grasses/Shrubs, Oaks/Western Hardwoods, Western Pines, Fir/Cedar and High-elevation Pines) into the Ecosystem Demography (ED2) terrestrial biosphere model improves predictions of net ecosystem productivity (NEP) and gross primary productivity (GPP) across four flux towers of the Southern Sierra Critical Zone Observatory (SSCZO) spanning a 2250 m elevational gradient in the western Sierra Nevada. NEP and GPP root-mean-square-errors were reduced by 23%-82% and 19%-89%, respectively, and water flux predictions improved at the mid-elevation pine (Soaproot), fir/cedar (P301) and high-elevation pine (Shorthair) flux tower sites, but not at the oak savanna (San Joaquin Experimental Range [SJER]) site. These improvements in carbon and water predictions are similar to those achieved with model initializations using ground-based inventory composition. The imaging spectrometry-constrained ED2 model was then used to predict carbon, water and energy fluxes and above-ground biomass (AGB) dynamics over a 737 km region to gain insight into the regional ecosystem impacts of the 2012-2015 Californian drought. The analysis indicates that the drought reduced regional NEP, GPP and transpiration by 83%, 40% and 33%, respectively, with the largest reductions occurring in the functionally diverse, high basal area mid-elevation forests. This was accompanied by a 54% decline in AGB growth in 2012, followed by a marked increase (823%) in AGB mortality in 2014, reflecting an approximately 10-fold increase in per capita tree mortality from ~55 trees km year in 2010-2011, to ~535 trees km year in 2014. These findings illustrate how imaging spectrometry estimates of ecosystem composition can constrain and improve terrestrial biosphere model predictions of regional carbon, water, and energy fluxes, and biomass dynamics.
准确描述当前的生态系统组成对于提高陆地生物圈模型预测生态系统对气候变异性和变化的响应至关重要。本研究探讨了成像光谱衍生的生态系统组成如何约束和改进陆地生物圈模型对区域尺度碳、水和能量通量的预测。将五种植物功能类型（Grasses/Shrubs、Oaks/Western Hardwoods、Western Pines、Fir/Cedar 和 High-elevation Pines）的成像光谱衍生组成纳入生态系统动态学（ED2）陆地生物圈模型，提高了在跨越西内华达山脉 2250 米海拔梯度的南塞拉 Critical Zone 观测站（SSCZO）的四个通量塔的净生态系统生产力（NEP）和总初级生产力（GPP）的预测。NEP 和 GPP 的均方根误差分别减少了 23%至 82%和 19%至 89%，并且在中海拔松树（Soaproot）、冷杉/雪松（P301）和高海拔松树（Shorthair）通量塔站点的水通量预测得到改善，但在橡树草原（圣华金实验牧场 [SJER]）站点则没有改善。这些在碳和水预测方面的改进与使用基于地面的清查组成进行模型初始化所取得的改进相似。然后，使用成像光谱约束的 ED2 模型来预测 737 公里区域的碳、水和能量通量以及地上生物量（AGB）动态，以深入了解 2012-2015 年加利福尼亚干旱对区域生态系统的影响。分析表明，干旱使区域 NEP、GPP 和蒸腾分别减少了 83%、40%和 33%，其中功能多样、高基底面积的中海拔森林减少幅度最大。这伴随着 2012 年 AGB 生长下降 54%，随后在 2014 年 AGB 死亡率显著增加（823%），反映出人均树木死亡率从 2010-2011 年的约 55 棵树·公里-2 增加到 2014 年的约 535 棵树·公里-2，增加了约 10 倍。这些发现说明了成像光谱估计的生态系统组成如何约束和改进陆地生物圈模型对区域碳、水和能量通量以及生物量动态的预测。
