最大熵增蒸散模型在中国区域的应用及时间尺度对比分析

Application of Maximum Entropy Production Model for estimation of evapotranspiration in China

为解决传统蒸散发计算中模型参数、模型缺陷和不确定性等问题,基于最大熵增蒸散发(MEP)模型的模型假设,针对我国不同植被覆盖下的蒸散发(潜热)通量进行估算,并分析半小时、日、周、月等不同时间尺度下的模型表现,以期获得适合我国区域的蒸散发计算方法。结果表明:MEP模型具有较好的模型计算表现,与多个站点的涡度相关实测数据较为接近;MEP模型在半小时尺度上应用更优,所有站点的潜热通量平均决定性系数(R^2)为0.75,周、月尺度上潜热通量精度较显热通量高。对MEP模型的两个输入参数(表面温度和比湿)分析表明,表面温度对预测结果影响较大。MEP模型精度高且具有所需输入参数少的优势,研究结果可用于对各种植被覆盖地表的蒸散发估算。

To solve limitations of model parameters,model structure defects,and uncertainties in estimating ET by traditional models,this paper estimates ET flux under different vegetation-covered land surface in China based on assumptions of the Maximum Entropy Production(MEP)Model.Meanwhile,this study analyzed the MEP model performances at different temporal scales including half-hourly,daily,weekly,and monthly to obtain a suitable method for estimating ET in China.The results showed that:The MEP model had a good estimate performance and was consistent with the measurements of multiple eddy covariance stations;The MEP model was better applied on the half-hourly scale,the averaged coefficient of determination(R^2)of latent heat over all site was 0.75,and the accuracy of latent heat estimate was higher than that of sensible heat over the weekly and monthly scales;Analysis on the surface temperature and specific humidity of the MEP model’s inputs showed that the surface temperature had a greater influence on the predictions;Due to the high accuracy of the MEP model and advantage of fewer required input parameters,results can be used to estimate ET under various vegetation-covered land surfaces.