论文部分内容阅读
青藏高原沼泽化草甸是土壤有机碳密度最高,对气候变化最敏感的高寒生态系统。对其生态系统总初级生产力(GPP)持续准确的量化是掌握站点到全球尺度的碳循环的关键。涡度相关技术(EC)是测量生态系统碳通量的最佳途径,而遥感模型可以实现从生态系统迹点(footprints)到区域乃至全球的尺度扩展。但是,大多遥感估算模型的适用性在这种高寒沼泽化草甸上并没有得到验证。本研究选取了四个近年来被广泛运用的遥感估算GPP的模型,即MODIS算法(MOD)、植被光合模型(VPM)、光合能力模型(PCM)和高寒植被模型(AVM)对青藏高原中部的一个典型高寒草甸生态系统的GPP进行了估算。结果显示:所有遥感模型对GPP的年内季节变异都有很好的解释(R2>0.89,P<0.0001),但很难解释其年际变化。与日均EC_GPP相比,VPM严重的低估了该生态系统的GPP,其估测值大约仅为EC观测值的54%。但是,其他三个模型可以较准确地进行GPP估算:相比之下,AVM可以反演94.5%的EC观测,相对于EC观测的均方根误差(RMSE)最小(1.47 g C m~(-2));PCM对EC_GPP有微小的高估(约12.0%的EC观测值),而MODR对EC_GPP有微弱的低估(约8.1%的EC观测值),但二者的偏差都不显著。本研究表明AVM对该高寒沼泽化草甸的GPP估算比其他较复杂的GPP估算模型更有优势。
The swampy meadow in the Qinghai-Tibet Plateau is the most alpine ecosystem with the highest density of soil organic carbon and the most sensitive to climate change. The continuous and accurate quantification of its total ecosystem primary productivity (GPP) is the key to mastering the carbon cycle from site to global scale. Eddy covariance (EC) is the best way to measure carbon fluxes in ecosystems, and remote sensing models can scale from ecosystem footprints to regions and globally. However, the applicability of most remote sensing estimation models has not been validated on this alpine swampy meadow. In this study, we selected four models of remote sensing GPP widely used in recent years, namely MODIS MOD, VPM, PCM and AVM, The GPP of a typical alpine meadow ecosystem was estimated. The results show that all the remote sensing models have a good explanation for the annual seasonal variation of GPP (R2> 0.89, P <0.0001), but it is difficult to explain the interannual variability. Compared to the average daily EC_GPP, the VPM severely underestimated the GPP of the ecosystem, with an estimated value of only about 54% of the EC observations. However, the other three models allow for more accurate GPP estimates: In contrast, AVM can retrieve 94.5% of the EC observations, with a minimum root mean square error (RMSE) of 1.47 g C m ~ (- 2)); PCM slightly over-estimated EC_GPP (about 12.0% of EC observations) while MODR showed a weak underestimation of EC_GPP (about 8.1% of EC observations), but both were not significantly different. This study shows that AVM estimates the GPP of this alpine swamp meadow more than any other more complex GPP estimation model.