基于MODIS数据的青藏高原积雪日数提取与时空变化分析

Extraction and Spatiotemporal Analysis of Snow Covered Days over Tibetan Plateau Based on MODIS Data

以青藏高原为研究对象,首先采用基于三次样条函数的去云算法对2001—2011年逐日MODIS积雪面积比例产品进行了去云处理,并对去云结果进行了精度验证。然后根据去云后的逐日无云MODIS积雪面积比例产品,提取了研究区近11年的积雪日数,并对积雪日数的时空分布特征进行了分析。结果表明:1.本文的去云算法能有效的获取云覆盖像元的积雪面积信息,总体平均绝对误差值为0.092。去云后MODIS积雪产品提取的积雪日数与地面观测值具有较高的一致性(87.03%),平均绝对误差3.8 d;2.青藏高原积雪日数的分布极不均匀,四周山区(特别是西部和南部山区)积雪分布广泛且积雪日数高,而高原腹地积雪日数低,年均稳定性积雪面积占27.24%;3.青藏高原积雪日数的年际波动较大,积雪日数在34.14%的地区呈减少的趋势,在24.75%的地区呈增加的趋势,其中显著减少和增加的地区分别为5.59%和3.9%。

Based on Moderate Resolution Imaging Spectroradiometer（ MODIS） daily fractional snow cover（ FSC）data,in this paper the spatiotemporal distribution of snow covered days（ SCD） over Tibetan Plateau（ TP） was examined from 2001 to 2011. Firstly,we described a cloud removal methodology using cubic spline interpolation algorithm to fill in data gaps caused by clouds. Then the yearly SCD images were extracted using the cloud removed MODIS FSC data,and the spatiotemporal distribution of the SCD in the study area was analyzed. The results were as follows：（ 1） The cloud removal methodology was efficient in cloud reduction and retrieve the FSC information of these cloud covered pixels（ overall mean absolute error of the retrieved FSC data is 0. 092）. There was a high consistency between MODIS-derived SCD and in-situ observed SCD,and the mean absolute error was 3. 82 days.（ 2） The spatial distribution of SCD over the TP was far from uniformity. In the peripheral mountainous area（ especially the mountains in the southern and western of TP）,snow cover distributes extensively had a long duration（ with a high SCD）. But in the vast interior area,snow cover was rare and with a very low SCD. About 27. 24% of the TP are stable snow cover areas.（ 3） A very high interannual variability of SCD was found in the 11 years. About 34. 14%（ 5. 56% with a significant decline） and 24. 75%（ 3. 9% with a significant increase） of the study area showed declining and increasing trend in SCD,respectively.