Canadian Journal of Remote Sensing ( IF 2.6 ) Pub Date : 2021-03-22 , DOI: 10.1080/07038992.2021.1898938 Aaron Thompson 1 , Richard Kelly 1
Abstract
Mid-latitude snow is understudied compared to snow in the northern high latitudes despite its importance as a source of freshwater to this economically significant region. The mid-latitudes provide opportunity to understand the influence on SWE retrievals of ice and vegetation, conditions which may encroach the north with a warming climate. Successful retrieval of SWE and snow depth was demonstrated using the Microwave Emission Model of Layered Snowpacks, adapted for backscattering (MEMLS3&a) with multi-angle Ku-band UWScat snow observations (VV polarization), of agricultural fields in Maryhill and Englehart, Ontario. A single-layer parameterization provided best results for Maryhill (RMSE of 21.9 mm SWE) while a dual-layer parameterization provided best results for Englehart (RMSE of 24.6 mm). Accounting for soil effects using a soil-subtraction method improved RMSE by up to 6.3 mm SWE. Retrievals were repeated with in situ snow depths improving accuracy at both sites (RMSE = 12.0 and 10.9 mm SWE for Maryhill and Englehart, respectively). Observations containing ice lenses and partially buried vegetation yielded a retrieval accuracy of RMSE = 60.4 mm which improved with in situ snow depths (RMSE = 21.1 mm SWE). These results compared favorably with the often-cited accuracy requirement of RMSE = 30 mm SWE.
中文翻译:
中纬度安大略省农业用地的Ku波段雷达雪水当量检索的注意事项
摘要
与北部高纬度地区的降雪相比,中纬度地区的降雪量被研究不足,尽管该地区作为该重要经济地区的淡水来源非常重要。中纬度提供了机会,以了解冰和植被对SWE取回的影响,这些条件可能会随着气候变暖侵蚀北部地区。使用多层雪堆的微波发射模型成功地取得了SWE和积雪深度,该模型适用于安大略省Maryhill和Englehart的农田的多角度Ku波段UWScat雪观测(VV极化)的反向散射(MEMLS3&a)。单层参数化为Maryhill(RMSE为21.9毫米SWE)提供了最佳结果,而双层参数化为Englehart(RMSE为24.6毫米)提供了最佳结果。使用土壤减法计算土壤影响后,RMSE最多提高了6.3 mm SWE。重复检索现场降雪深度可提高两个站点的精度(Maryhill和Englehart的均方根误差(RMSE)分别为12.0和10.9 mm SWE)。包含冰晶和部分被掩埋的植被的观测结果的取回精度为RMSE = 60.4 mm,随原位积雪深度的改善(RMSE = 21.1 mm SWE)。这些结果与经常提到的RMSE = 30 mm SWE的精度要求相比具有优势。