色林错、纳木错湖区被动微波雪深反演算法评估

Evaluation of Passive Microwave Snow-Depth Retrieval Algorithm in Selin Co and Nam Co

被动微波雪深反演算法是当前大范围获取青藏高原地表雪深信息的重要途径,但由于缺乏地面雪深观测资料,导致对算法在高原中西部区域的表现认识不足。为了评估当前被动微波雪深反演算法在青藏高原色林错、纳木错地区的适用性,利用AMSR2亮温数据和地面站点雪深数据,以相关系数、偏差和均方根误差作为评价指标,评估了Chang2算法、Che算法、SPD算法、AMSR2算法和Jiang算法等5种算法。结果显示,Jiang算法综合表现最好,在纳木错站R值最高为0.68;Che算法对浅雪反演效果较好,其在班戈站Bias为-0.66 cm;Chang2算法对纳木错站、色林错站深雪反演效果较好,在两地R值分别为0.63、0.50;SPD算法的反演效果最不理想,对雪深高估明显,其中浅雪高估近20 cm;AMSR2算法在区域间的表现差异较大,在纳木错站的反演结果比色林错站、班戈站好。除SPD算法外,其余算法均低估了研究区雪深,与以往研究结果一致。

The passive microwave snow-depth retrieval algorithm is an important method to obtain the surface snow depth information of the Tibetan Plateau on a large scale.In order to evaluate the applicability of the current passive microwave snow-depth retrieval algorithms in the Selin Co and Nam Co regions of the Tibetan Plateau,AMSR2 brightness temperature data and snow depth data of ground stations are used,while R,Bias and RMSE are used as evaluation indicators.Five algorithms including Chang2 algorithm,Che algorithm,SPD algorithm,AMSR2 algorithm and Jiang algorithm are chosen.The results show that the Jiang algorithm has the best overall performance,with the highest R value of 0.68 at Nam Co station.The Che algorithm has a good retrieval effect on shallow snow,and its Bias at Bangor Station is-0.66 cm.The Chang2 algorithm performed well for the deep snow of Nam Co station and Selin Co station,with R values of 0.63 and 0.50 in the two places respectively.The retrieval effect of SPD algorithm is the most unsatisfactory,and the snow depth is overestimated obviously,among which shallow snow is overestimated by nearly 20 cm.The performance of AMSR2 algorithm differs greatly between regions,and the retrieved results at Namco Station are better than those at Selin Co Station and Bangor Station.Except for the SPD algorithm,all other algorithms underestimate snow depth in the study area,which is consistent with previous research results.