几种典型地表类型反照率时序变化特征及其参数化研究

Time series variation characteristics and parameterization of land surface albedo in several typical land cover types

地表反照率对地表能量平衡、气候模式和全球变化研究具有重要的作用.受季节和下垫面影响,地表反照率呈现一定的时序变化特征,刻画这种变化特征可以为地表反照率估算提供背景信息,有利于提高反照率反演精度.目前已有许多研究致力于分析地表反照率时序变化特征和影响因素,但分析仅限于某一特定区域的一种或几种地表类型,同时对反照率与叶面积指数(leaf area index,LAI)、归一化植被指数(normalized difference vegetation index,NDVI)等影响因子的相关性还缺少参数化研究.本文使用美国通量观测网络AmeriFlux观测数据资料,利用该网络站点分布范围广、地表类型丰富的特点,选取其中代表性较强的站点观测数据,分析了落叶阔叶林、常绿针叶林、草地、农田4种典型植被类型地表反照率的时序变化特征.发现各植被地表类型反照率呈现"U"形年内变化特征,在植被生长季,落叶阔叶林、农田、草地3种地表类型反照率均先减小再增大后减小,而常绿针叶林地表反照率变化不明显.利用统计方法分别建立了地表反照率与LAI和NDVI的参数化模型,结果显示:2种估算地表反照率的参数化模型均能较好反映植被生长过程中地表反照率的变化特征,与地面实测数据相比,计算得到的地表反照率具有较高的精度;2种参数化模型估算得到的反照率一致性好,相关系数为0.721 9.该参数化方案可为地表反照率的进一步估算提供背景参考模型.

Land surface albedo has an important role in surface energy balance,climate model and global change research.Affected by seasons and underlying surfaces,the surface albedo exhibits certain temporal variations.Characterizing this variation can provide background information for surface albedo estimation,which helps to improve albedo inversion accuracy.Many studies have devoted to analysis of surface albedo temporal variation and parameterization schemes,but the analysis is limited to one or several surface types in a specific area,and cannot express the variation law of surface albedo in different land types.In this study,AmeriFlux observations were used to analyze timing variations of land surface albedo including deciduous board-leaved forest,evergreen coniferous forest,grassland and farmland.Albedo parameterization models were established from site data in different land cover types with LAI and NDVI.It was found that during the growing season,evergreen coniferous forest and grassland albedo have identical relationship with LAI and NDVI:decreases first,then increases and finally maintains a stable state with the growth of vegetation.Albedo of deciduous broad-leaved forest increases first,then decreases slightly and then becomes stable with increase of LAI.For broad-leaved forests,albedo increases linearly with increased NDVI when NDVI is larger than 0.4.Albedo of farmland surface is approximately positively linear with LAI and NDVI.Regional scale land surface albedo was then estimated using established parameterization scheme.Parameterization models can well capture the time series variation pattern and spatial characteristics of the land surface albedo.This method can estimate land surface albedo in different land cover types,and provide a high accuracy albedo background reference for data assimilation model.