动态植被模型对青藏高原植被的模拟检验

Evaluation of Vegetation Characteristics over Qinghai- XizangPlateau Simulated by A Vegetation Dynamic Model

动态植被模型是研究植被变化对气候反馈和影响的重要模型工具。本文对耦合了动态植被(Dynamic Vegetation,DV)和碳氮(Carbon and Nitrogen,CN)模型的NCAR陆面过程模式CLM4.5(Community Land Model version 4.5)对青藏高原(以下简称高原)植被的模拟性能进行了评估,获得了定量化的偏差信息,并对高原植被和气候变化因子的关系进行了初步探讨。结果表明:模型能大致再现叶面积指数(Leaf area index,LAI)在历史时期的季节循环、长期变化趋势和空间分布,但空间变率较遥感资料大。模拟的乔木覆盖度偏大,草地覆盖度偏小,因此严重高估了植被高原南部和东部的LAI。与遥感观测相比,模拟的LAI呈现了1~2个月的滞后,这与模式本身的植被动力机制不完善和模式的降水驱动偏差有关。高原植被变化趋势的时空分布与表层土壤水和降水等气象因子的趋势变化显示出较好的一致性,表明在该研究时段,地表水循环的变化(主要是降水和土壤水含量)对高原植被生长可能起主导作用。

In this study,performances of NCAR Community Land Surface Model CLM(version 4.5)coupled with dynamic vegetation and carbon-nitrogen model(CN-DV)in representing the observed vegetation characteristics of satellite over the Qinghai-Xizang Plateau(QXP)during historical period 1981—2000 were evaluated by comparing with remote sensing datasets of leaf area index(LAI)and coverage fraction.It was shown CLM4.5-CNDV can generally reproduce the spatial distribution of LAI,coverage fractions and their linear trends shown in the satellite data,but LAI was seriously overestimated due to the significantly overestimated tree coverage,and underestimated grass coverage as well.The 1~2 month lag shown both in LAI and precipitation may suggest that the simulated LAI bias could also be caused by the bias of atmospheric forcing which come from 18 model of CMIP5.The spatial distribution of LAI linear trend had a great agreement with precipitation and 10-cm soil moisture,implying that the surface hydrological variations are the key factors to QXP vegetation change.