Estimating spatial distribution of daily snow depth with kriging methods: combination of MODIS snow cover area data and ground-based observations
Estimating spatial distribution of daily snow depth with kriging methods: combination of MODIS snow cover area data and ground-based observations
复制标题
DOI：
10.5194/tcd-9-4997-2015
复制
发表时间：
2015-09
期刊：
影响因子：
--
通讯作者：
C. Huang;H. Wang;J. Hou
中科院分区：
文献类型：
--
作者：
C. Huang;H. Wang;J. Hou
抽象。准确测量积雪深度的空间分布是困难的，因为站点稀疏，特别是在中国西部。在这项研究中，我们开发了一种新的方案，产生一个合理的空间分布的日积雪深度使用克里格插值方法。这些方法结合了联合收割机的海拔与信息的影响，从中分辨率成像光谱仪（MODIS）的积雪面积（SCA）产品。该方案利用去除云层的中分辨率成像分光仪SCA图像中的无雪像素来确定虚拟台站，或雪深为零的地区，以补偿台站的稀缺和分布不均。四种类型的克里金方法进行了测试：普通克里金（OK），通用克里金（UK），普通协同克里金（OCK），通用协同克里金（UCK）。将这些方法应用于中国新疆北方50个气象站的逐日雪深观测。结果表明，这些克立格方法可以准确地重建积雪深度的空间分布。增加的虚拟站改善了雪深的分布，并减少了克里金过程的平滑效果。在积雪较浅的情况下，OK方法的性能最好，而在积雪较广的情况下，UCK方法的性能最好。
Abstract. Accurately measuring the spatial distribution of the snow depth is difficult because stations are sparse, particularly in western China. In this study, we develop a novel scheme that produces a reasonable spatial distribution of the daily snow depth using kriging interpolation methods. These methods combine the effects of elevation with information from Moderate Resolution Imaging Spectroradiometer (MODIS) snow cover area (SCA) products. The scheme uses snow-free pixels in MODIS SCA images with clouds removed to identify virtual stations, or areas with zero snow depth, to compensate for the scarcity and uneven distribution of stations. Four types of kriging methods are tested: ordinary kriging (OK), universal kriging (UK), ordinary co-kriging (OCK), and universal co-kriging (UCK). These methods are applied to daily snow depth observations at 50 meteorological stations in northern Xinjiang Province, China. The results show that the spatial distribution of snow depth can be accurately reconstructed using these kriging methods. The added virtual stations improve the distribution of the snow depth and reduce the smoothing effects of the kriging process. The best performance is achieved by the OK method in cases with shallow snow cover and by the UCK method when snow cover is widespread.