聚集指数和最大羧化速率对基于遥感产品的植被生产力估算的影响

Effects of Clumping Index and Maximum Carboxylation Rate on Vegetation Productivity Estimation Based on Remote Sensing Data

【目的】明确聚集指数和最大羧化速率遥感产品对BEPS模型估算植被生产力的影响。【方法】利用中国陆地通量站点观测数据,分析BEPS模型中聚集指数(CI)和最大羧化速率(V_(cmax))的敏感性程度,并比较聚集指数和最大羧化速率遥感产品对植被生产力估算的精度提升作用。在此基础上估算2012年中国陆地生态系统植被生产力,通过与参数缺省值估算结果对比,研究CI和V_(cmax)的时空变化对模型估算结果的影响。【结果】1) CI和V_(cmax)均为BEPS模型中较为敏感的参数,两者均与植被生产力呈正相关关系,且不同植被类型下V_(cmax)敏感性均高于CI。2)聚集指数和最大羧化速率遥感产品同时使用情况下,模拟结果的误差最小,精度最高,总初级生产力(GPP)均方根误差从665.60 g·m^(-2)a^(-1)降至584.71 g·m^(-2)a^(-1),平均误差和相对平均误差均为4种模拟情况最低值。3)2012年中国陆地生态系统GPP和净初级生产力(NPP)总量分别为5.21和2.49 Pg·a^(-1),受CI遥感产品(NDHD-CI)和V_(cmax)遥感产品(SIF-V_(cmax))的时空变化影响,GPP和NPP估算分别较模型缺省值偏低3.06%和4.72%。【结论】NDHD-CI和SIF-V_(cmax)能够提升BEPS模型估算植被生产力的精度,未来可对其他高敏感度参数和模型机理进行优化改进。受CI和V_(cmax)时空变化影响,植被生产力估算结果略低于缺省情况。V_(cmax)对植被生产力估算影响高于CI。

【Objective】This study aims to investigate the effect of clumping index(CI)and maximum carboxylation rate(V_(cmax))from remote sensing products on estimation of vegetation productivity with the boreal ecosystem productivity simulator(BEPS)model.【Method】The FLUXNET and ChinaFLUX data were used to analyze the sensitivities of CI and V_(cmax) in BEPS model,and compare the effects of CI and V_(cmax) on Gross Primary Productivity(GPP)estimation.On this basis,the vegetation productivity of terrestrial ecosystems in China in 2012 was estimated.By comparing with the estimated results of the default value,we determined the impact of the spatio-temporal changes of CI and V_(cmax) on the model performance.【Result】1)The results showed that CI and V_(cmax) had high sensitivities in the BEPS model.They were positively correlated with vegetation productivity,and the sensitivity of V_(cmax) was higher than that of CI in different vegetation types.2)When CI and V_(cmax) remote sensing products(NDHDCI and SIF-V_(cmax))were used simultaneously,the simulation results had the smallest error and the highest accuracy.The root mean square error(RMSE)of GPP decreased from 665.60 g·m^(-2)a^(-1) to 584.71 g·m^(-2)a^(-1),and the mean absolute error(MAE)and mean relative error(MRE)were the lowest in the four simulation cases.3)In 2012,the total GPP and Net Primary Productivity(NPP)of terrestrial ecosystems in China were 5.21 Pg·a^(-1) and 2.49 Pg·a^(-1),respectively.Affected by the spatio-temporal dynamics in the CI and V_(cmax),the GPP and NPP estimates were 3.06%and 4.72%lower than the default results of the model,respectively.【Conclusion】Our results have demonstrated that NDHD-CI and SIF-V_(cmax) can improve the accuracy of BEPS models in estimating vegetation productivity,and other high-sensitivity parameters and model mechanisms can be optimized and improved in the future.Affected by the temporal and spatial changes of CI and V_(cmax),the estimation results of vegetation productivity are slightly lower than the default situation.The effect of V_(cmax) on vegetation productivity estimation is higher than that of CI.