基于改进PT-hybrid算法的地表潜热通量变化分析

潜热通量(LE)影响陆地生态系统的水量平衡和能量收支。【目的】提高不同植被类型对应的地表潜热通量的反演精度对于探究生态水文过程具有重要的现实意义。【方法】本研究采用动态植被模型--LPJ-GUESS模型,引入表征植被占比的参数fv对PT-hybrid潜热通量算法加以改进,从而提高陆面蒸散发的反演能力。【结果】根据全球49个通量站点的地面观测LE验证改进后的算法,改进算法所得LE值与实测值的决定系数R²从0.63增加到0.74,均方根误差RMSE从17.1 W/m²降低到11.7 W/m²。进一步分析1982-2014年全球陆地LE变化特征,发现低纬度地区的LE均高于高纬度地区,并且LE在同一纬度上也呈现出不同的分布趋势。【结论】改进后的模型对地表潜热通量模拟精度显著提高,有助于精细刻画地表LE的时空分布特征。

Calculating Latent Heat Flux from Soil Surface Using the Improved PT-hybrid Algorithm

【Background】Latent heat flux(LE)impacts carbon dynamics and energy budget in the terrestrial ecosystem.Its accurate estimation is essential to understanding energy budget and surface-atmosphere interactions on the Earth surface.Satellite imageries have been used to calculate LE at global scale,but most of them overlooked the impact of vegetation diversity at pixel scale.【Objective】Considering that carbon flow and heat flux in ecosystems are closely related through photosynthesis and that plants are diverse,calculating LE using sub-pixel data should improve its accuracy,especially for complex ecosystems.【Method】Here,a dynamic vegetation model,LPJ-GUESS,was proposed to improve the satellite-based PT-hybrid algorithm,by using a parameter fv to represent the proportions of different PFTs in a pixel.【Result】Validation against ground-true data obtained from 49 flux tower sites in the world showed that the proposed model increased the square of correlation coefficient from 0.63 to 0.74,and reduced the root mean square error from 17.1 W/m² to 11.7 W/m².Estimation of the global LE from 1982 to 2014 revealed the proposed method was more effective for low latitude areas than for other areas,and that for areas on the same latitude,the LE showed complicated spatial distribution.【Conclusion】Sub-pixel method was coupled with the LE algorithm for the first time to calculate surface latent heat flux.It significantly improved the accuracy and can be used to estimate spatiotemporal distribution of LE at different scales.