基于特征跟踪方法的喜马拉雅山冬季冰川流速估算

Estimation of the Himalayas' winter glacier velocity based on feature tracking

基于2009年冬季的4景先进对地观测卫星(ALOS)相控阵型L波段合成孔径雷达(PALSAR)数据,结合数字高程模型(DEM),并用特征跟踪方法对喜马拉雅山中段30条冰川的三维流速进行了估算。结果表明,冰川流速在0～268m·a^(-1)范围内,面积较大、冰舌较长的冰川,如绒布、康雄、Kazhenpu和Khumbu等冰川整体流速较大。从空间范围来看,东坡、东南坡和西南坡冰川流速明显高于北坡冰川流速。这主要由于受印度洋吹来的暖湿气流的影响,三个坡向冰川补给较多,冰川运动较快,而北坡为干旱的大陆性气候,冰川补给少,流速慢。除此之外,形状较直的冰川,所受摩擦阻力较小,流速较大,如康雄、Kazhenpu和West Barun等冰川。其次,面积较大、冰舌较长的冰川,流速随高程的降低而减小;面积较小、冰舌较短的冰川,流速随高程的降低波动变化,或者增加,这主要是由于坡度引起的冰川流速变化。冰川流速与冰川面积有一定的正相关关系,即面积越大,流速越大,面积越小,流速越小。基于特征跟踪方法的冰川流速估算能够克服差分干涉测量对数据时间基线太长而造成的失相干,是比较有效的冰川流速估算方法,并且此结果为喜马拉雅山冰川冬季流速提供了补充,可以为后续的研究提供参考价值。

The three-dimensional velocity of 30 glaciers in the middle section of the Himalayas mountain is estimated by using feature tracking method basing on the section ＇s four scenes of winter advanced land observation satellite（ ALOS） and phased array-type L-band synthetic aperture radar（ PALSAR） data of 2009 and the DEM. The results show that the glacier velocity ranges from 0 to 268 m·a^（-1）. The glaciers with large area and long tongue have the large velocity,such as the Rongbuk,Kangxiong,Kazhenpu and Khumbu and so on. As far as the spatial range is concerned,the glacier velocity in the north aspect is obviously lower than that in others aspects. This is only because of the impact of warm air blowing from Indian Ocean： glaciers of the three aspects have more glacial supply,while the north aspect has a dry continental climate with little glacial supply and slow velocity. Besides,straight glaciers have small friction resistance,which causes larger velocity,such as Kangxiong,Kazhenpu and West Barun. Secondly,the velocity of the glaciers with large areas and long tongues decreases with altitude decreasing,and if glaciers have small areas and short tongues,the velocity may fluctuate along altitude,so the changing of the glacier velocity is mainly due to slope. There is a positive correlation between the glacier velocity and the glacier area,that is,the larger the area,the greater the velocity,and the smaller the area,the smaller the velocity. The glacier velocity estimation based on feature tracking can overcome the decorrelation caused by the data long baseline of time of differential interferometry,so it is an effective method to estimate glacier velocity. These results provide a supplement for the winter glacier velocity of the Himalayas,and can be used as a reference for further study.