基于MODIS的青藏高原逐日无云积雪产品算法

青藏高原积雪对高亚洲地区水和能量循环起着重要的反馈和调节作用，其变化影响着融雪性河流流量，对下游水资源和经济活动具有重要影响。中分辨率成像光谱仪（MODIS）具有较高的时空分辨率，被广泛应用于积雪遥感动态监测，然而光学遥感积雪受云层影响严重，且青藏高原地区水汽分布不均，局地对流活跃，积雪的赋存时间变化快，这给高原地区逐日积雪监测及其气候学制图带来挑战。在考虑青藏高原地形和积雪分布特征情况下，结合现有的云覆盖下积雪判别算法，采用8个不同方法的组合，逐步实现MODIS逐日无云积雪算法。选取2009年10月1日-2011年4月30日两个积雪季为研究期，并采用145个地面台站观测雪深数据对去云算法各步骤过程开展精度验证，结果表明：当积雪深度>3 cm时，逐日无云积雪产品总分类精度达到96.6%，积雪分类精度达83%，积雪判对概率（召回率）达到89.0%，算法可实现青藏高原地区逐日无云积雪动态监测和积雪覆盖气候学数据重建，对高亚洲地区的水、生态和灾害等全球环境变化影响研究具有重要的意义。

Cloud removing algorithm for the daily cloud free MODIS-based snow cover product over the Tibetan Plateau

Snow cover on the Tibetan Plateau plays an important role in feedback and regulating to water and energy cycles of the High Asia. Snow melt supplies river runoff, which has an important influence on the social and economic development for the downstream areas. Medium resolution imaging spectrometer (MODIS) is widely used in dynamic remote sensing to monitor snow cover variation, because of its high temporal and spatial resolution. However, optical remote sensing snow cover is seriously affected by cloudy weather. Especially, water vapor over the plateau is low and inhomogeneous, with active convection, resulting in snow cover rapid changing. All the problems bring a challenge to the daily cloud-free snow monitoring and snow climatology mapping in the plateau. Considering the local terrain and snow cover characteristics in the Tibetan Plateau, in this study, based on the existing cloud removal algorithms, gradually controlling cloudage, the daily cloud-free MODIS snow cover algorithm has been carried out. The snow depth records from 145 ground stations for two snow seasons, from 1^st October, 2009 through 30th April, 2011, are used to qualify the accuracies of various algorithm processes. The qualifying show that when snow depth is above 3 cm, the total classification accuracy of daily cloud-free snow products is 96.6%, snow classification accuracy is 83.0% and the recall rate is 89%. The new cloud removal process is proved having high precision and is suitable for daily dynamic monitoring snow cover. It provides a new way for reconstructing snow climatology data over the Tibetan Plateau. It is useful to study global environmental change, such as water, ecology, disaster and their impacts on the High Asia.