X img

官方微信

img

群号:冰川冻土交流群

QQ群:218834310

高级检索
作者投稿 专家审稿 编辑办公 编委办公 主编办公

冰川冻土 ›› 2013, Vol. 35 ›› Issue (1): 40-47.doi: 10.7522/j.issn.1000-0240.2013.0005

• 冰冻圈与全球变化 • 上一篇    下一篇

基于AMSR-E的北疆地区积雪深度反演

卢新玉1, 王秀琴3, 崔彩霞1, 谢国辉2   

  1. 1. 新疆气象台, 新疆 乌鲁木齐 830002;
    2. 新疆气象局, 新疆 乌鲁木齐 830002;
    3. 昌吉州气象局, 新疆 昌吉 831100
  • 收稿日期:2012-08-08 修回日期:2012-11-15 出版日期:2013-02-25 发布日期:2013-07-22
  • 作者简介:卢新玉(1979-),男,新疆乌鲁木齐人,2011年在新疆师范大学获硕士学位,现主要从事天气预报技术开发方面的研究.E-mail:luxy31810964@sina.com
  • 基金资助:

    公益性行业(气象)科研专项(GYHY201106007); 国家科技支撑项目(2012BAC23B01); 国家自然科学基金项目(41271098)资助

Snow Depth Retrieval Based on AMSR-E Data in Northern Xinjiang Region, China

LU Xin-yu1, WANG Xiu-qin3, CUI Cai-xia1, XIE Guo-hui2   

  1. 1. Xinjiang Meteorological Observatory, Vrümqi Xinjiang 830002, China;
    2. Xinjiang Meteorological Bureau, Vrümqi Xinjiang 830002, China;
    3. Changji Meteorological Bureau, Changji Xinjiang 831100, China
  • Received:2012-08-08 Revised:2012-11-15 Online:2013-02-25 Published:2013-07-22

摘要:

利用北疆地区2007/2008-2009/2010年度积雪季(12月至次年2月)的AMSR-E降轨19 GHz与37 GHz波段的水平极化亮温数据, 结合北疆地区45个气象台站的实测雪深数据, 建立了北疆地区基于AMSR-E亮度温度数据的雪深反演模型, 并对模型的精度进行评价. 结果显示: 雪深在3~10 cm时, 模型反演的雪深值负向平均误差为-5.1 cm, RMSE值为6.1 cm; 雪深在11~30 cm时, 模型反演雪深值的平均误差仅为2.6 cm, RMSE、 正向平均误差、 绝对平均误差均较小; 雪深大于30 cm时, 模型反演的各项误差较大. 用合成方法反演北疆地区2006/2007-2010/2011年度5个积雪季的平均雪深分布和最大雪深分布, 结果显示北疆地区积雪主要分布于北部阿尔泰山和南部天山一带, 其中阿勒泰地区所占比重最大, 中部的准噶尔盆地腹地、 克拉玛依地区雪层较浅.

关键词: 被动微波遥感, AMSR-E, 积雪深度, 雪深反演, 北疆地区, 验证

Abstract:

Combining brightness temperature data at 19 GHz and 37 GHz vertical and horizontal polarization channel of AMSR-E data and measured snow depth data from 45 meteorological stations in Northern Xinjiang Region in 2007-2009 snow season (December-February), a snow depth retrieval model was established based on AMSR-E brightness temperature data. The accuracy of the model was evaluated. The result shows the negative average error of retrieval snow depth is -5.1 cm and RMSE is only 6.1 cm for snow depth ranging from 3 to 10 cm. When snow depth is between 11 and 30 cm, the average error of retrieval snow depth is 2.6 cm only, and its RMSE, positive average error and absolute average error are all less. When snow depth is larger than 30 cm, the errors of retrieval snow depth are larger. Comparing the retrieval snow depth with the observed data, it is found that the simulations show a great agreement with the observations at meteorological stations in Northern Xinjiang Region. This basically reflects the snow depth distribution in Northern Xinjiang Region. Comparing Chang algorithm with the retrieval model in this study reveals that the latter is superior to the former and can describe snow depth variation characteristics well in Northern Xinjiang Region. Meanwhile, an average snow depth distribution map and a most heavy snow depth map are drawn by using synthetic method from five snow seasons (2006-2010). The maps show that snow cover is mainly distributed in the Altay Mountains in the north and the Tianshan Mountains in the south, with a maximum in the Altay Mountains. The snow cover is thin in the hinterland of Junggar basin and Karamay area.

Key words: passive microwave remote sensing, AMSR-E, snow depth, snow depth retrieval, Northern Xinjiang Region, validation

中图分类号: 

  • P426.63+5