地球科学进展  2018 , 33 (6): 590-605 https://doi.org/10.11867/j.issn.1001-8166.2018.06.0590

综述与评述

基于被动微波遥感的积雪深度和雪水当量反演研究进展

肖雄新, 张廷军*

兰州大学资源环境学院,西部环境教育部重点实验室,甘肃 兰州 730000

Passive Microwave Remote Sensing of Snow Depth and Snow Water Equivalent: Overview

Xiao Xiongxin, Zhang Tingjun*

Key Laboratory of Western China’s Environmental Systems (Ministry of Education), College of Earth Environmental Sciences, Lanzhou University, Lanzhou 730000,China

中图分类号:  P426.635

文献标识码:  A

文章编号:  1001-8166(2018)06-0590-16

通讯作者:  *通信作者:张廷军(1957-),男,甘肃庆阳人,教授,主要从事冰冻圈科学研究.E-mail:tjzhang@lzu.edu.cn

收稿日期: 2017-12-11

修回日期:  2018-05-15

网络出版日期:  2018-06-20

版权声明:  2018 地球科学进展 编辑部 

基金资助:  *国家重大科学研究计划项目“复杂地形积雪遥感及其多尺度积雪变化研究”(编号:2013CBA01802)国家自然科学基金重大研究计划项目“黑河流域上游多年冻土地表水、地下水过程及其效应研究”(编号: 91325202)资助.

作者简介:

First author:Xiao Xiongxin (1991-), male, Fuping County, Shaanxi Province, Master student. Research areas include remote sensing of snow. E-mail:xiaoxiongxin5118@126.com

作者简介:肖雄新 (1991-),男,陕西富平人,硕士研究生,主要从事积雪遥感研究.E-mail:xiaoxiongxin5118@126.com

展开

摘要

积雪是冰冻圈重要组成要素之一,也是对天气和气候响应最为敏感的自然要素。被动微波能够穿透云层、积雪和大气进行全天候、全天时地工作,在估算积雪深度、雪水当量等积雪参数上有很大优势。综述了国内外基于被动微波遥感的积雪参数反演研究的进展,首先介绍了被动微波遥感监测积雪的基本理论,以及被动微波遥感数据;然后将当前的积雪深度和雪水当量反演算法总结为4类:①基于统计的线性反演算法;②基于微波积雪模型的反演算法;③基于先验知识的非线性反演算法;④数据融合与数据同化。随后介绍了常用的7种积雪数据产品,并讨论了影响积雪深度和雪水当量反演精度的几个因素,最后对未来积雪参数反演研究方向做出了展望。

关键词: 被动微波遥感 ; 积雪深度 ; 雪水当量 ; 积雪产品

Abstract

Snow cover is an informative indicator of climate change because it affects local and regional surface energy and water balance, hydrological processes and climate. Passive Microwave (PM) works all weather and round the clock and penetrates clouds and snow. Passive microwave remote sensing data have been widely applied to retrieving snow depth and snow water equivalent in the past few decades. Recently, the snow depth retrieval study has rapidly developed. This paper reviewed the research progress of snow depth and snow water equivalent inversion algorithm using PM data at home and abroad. Firstly, the basic theory of passive microwave remote sensing snow monitoring and passive microwave remote sensing data were introduced. Then, the current snow depth and snow water equivalent inversion algorithm were summarized into four categories: ① A statistically based linear inversion algorithm; ② An inversion algorithm based on microwave transmission snow model; ③ A nonlinear inversion algorithm based on prior knowledge; ④ Data fusion and data assimilation. Afterwards, the commonly used seven kinds of snow data products were introduced, and several factors affecting the snow depth and the snow water inversion accuracy were discussed. Finally, the possible direction of future snow parameter inversion research was prospected.

Keywords: Passive microwave remote sensing ; Snow depth ; Snow water equivalent ; Snow products.

0

PDF (1346KB) 元数据 多维度评价 相关文章 收藏文章

本文引用格式 导出 EndNote Ris Bibtex

肖雄新, 张廷军. 基于被动微波遥感的积雪深度和雪水当量反演研究进展[J]. 地球科学进展, 2018, 33(6): 590-605 https://doi.org/10.11867/j.issn.1001-8166.2018.06.0590

Xiao Xiongxin, Zhang Tingjun. Passive Microwave Remote Sensing of Snow Depth and Snow Water Equivalent: Overview[J]. Advances in Earth Science, 2018, 33(6): 590-605 https://doi.org/10.11867/j.issn.1001-8166.2018.06.0590

1 引言

积雪是冰冻圈的重要组成部分,是地球气候系统内分布广泛,季节性明显,对天气和气候响应最为敏感的冰冻圈要素。全球大概有5%的降水是以雪的形式降落到地面上[1],南北极地区这一比例可达到50%~90%[2]。积雪主要分布在北半球,北半球冬季(1月)最大积雪面积可以达到4 600万km2,约占陆地面积的50%[3]。融雪作为河流湖泊重要的水资源,蕴藏着体量丰富的淡水资源,并对水循环、水文、气象以及水资源管理起着至关重要的作用。此外,积雪对地表—大气之间的能量交换具有重要的影响,对其下覆土壤具有保温和保湿效应[4],直接影响冬季土壤中CO2的排放和土壤中C和N等矿物元素的存在[5,6]。积雪对于人类生产生活会产生重要的影响,例如,为人类生活提供饮用水和生活用水;带来经济效益,如冬季滑雪娱乐项目等;带来洪水灾害,对铁路、公路、民航等交通运输带来破坏性影响。

积雪范围广泛,但是空间分布具有强的异质性,使得少量的站点很难充分显示大空间尺度上积雪的时空变化特征。卫星观测技术可以提供持续长时间、大范围积雪监测数据,克服了传统积雪实测方法中站点少的局限。到目前为止,积雪的遥感观测已经为水文、气象、防灾减灾和气候变化等领域的研究带来了巨大的便利[7,8,9,10,11]。现在利用光学遥感传感器主要进行积雪面积、积雪反射率等的遥感监测,但是不能有效地估算积雪深度、雪水当量等参数,而且光学遥感受到天气的影响,云在其中起了很大的作用。但是,微波可以穿透云层和积雪层,并且可以探测到积雪下覆地表的信息,具有全天候、大范围监测积雪的特点[11]。基于主动、被动微波遥感成为反演积雪深度和雪水当量的有效方式,具有光学遥感无可比拟的优势[11,12,13,14,15,16,17]

自20世纪70年代以来,国内外学者已经对积雪深度和雪水当量的研究做了大量工作,发展了多种反演方法。本文主要探讨基于被动微波遥感的积雪深度和雪水当量反演研究进展。首先较为详细地介绍利用被动微波遥感监测积雪的理论基础以及目前可用的被动微波遥感数据;然后对前人的积雪深度和雪水当量反演算法总结归纳,并对每种类型的积雪深度和雪水当量反演算法(基于统计的线性算法、基于微波积雪模型的算法、基于先验知识的非线性算法、数据融合与数据同化算法)的研究情况做简要介绍;在此基础上介绍了几种积雪数据产品,以及每种产品的特点和生成的算法;最后探讨了增加反演复杂性以及影响积雪深度和雪水当量反演精度的几个因素,为积雪参数反演研究提供依据。

2 数据和理论基础

2.1 被动微波遥感监测积雪的理论原理

微波的波长一般在0.1~100 cm (300~0.3 GHz)范围内。根据瑞利散射理论,散射波强度与其波长的4次成反比,水粒子组成的云、雾以及沙尘形成的气溶胶等的波长一般比微波的小,所以微波可以穿透云、雾、沙尘等组成的气溶胶层[18,19]。此外,大气对微波的散射、吸收作用也较小,一般情况下大气对微波的影响忽略不计,因此微波在通过大气时具有较高的透射率。这些优点使微波传感器几乎可以在全天时、全天候条件下监测地表所发射的微波信号,获取地表地物的特征数据,最后通过对微波传感器接收到的信息进行分析来估算地物的参数[15,20~22]。微波信号对积雪层具有较高的穿透力,能够穿透一定厚度的积雪,使被动微波遥感应用于积雪深度监测成为可能。在积雪覆盖区,卫星搭载的微波辐射计最终以亮温的形式记录下来自地面的微波辐射能量(亮温Tb)[23,24],其来自于4个部分,可用公式(1)表示(图1):

Tb=Tb1+Tb2+Tb3+Tb4, (1)

式中:Tb1是积雪层下覆地表发射的能量,Tb2是积雪层发射的能量,Tb3是经积雪—地表界面反射的天空辐射,Tb4是经积雪—空气界面反射的天空辐射。

积雪向上微波辐射的能量主要来源于积雪层下覆地表(Tb1)和积雪层内部(Tb2)[24,25]。在积雪微波辐射中,雪颗粒的体散射起着重要作用,会衰减和散射部分积雪向外辐射的能量。积雪的微波辐射特性会随着积雪厚度、积雪粒径[26]、积雪密度、雪温、水分含量[27]、积雪结构以及下垫面介质的变化而改变。积雪深度越大,意味着积雪层内的散射微波信号的粒子越多,就会有越多的能量在穿透积雪层的过程中被吸收或者散射,由此会引起卫星传感器接受的到能量越少,亮温越低[25,28,29]。当积雪融化时,由于积雪层内液态水的存在,积雪的微波辐射能量很大一部分会被吸收而无法穿透积雪层。积雪在能量辐射时,散射作用占主导,其中干雪表现为强散射体,其会随微波频率的增加而散射作用增强[30]

微波辐射理论的提出已有100多年的历史,其应用也日益成熟。研究者发现可以利用积雪对不同频率的敏感性不同来探测地表积雪雪深和雪水当量。Chang等[31]提出积雪粒径的微波辐射向上辐射的能量积雪深度、积雪粒径和积雪密度等的影响,此研究结果为后来的运用微波遥感遥感监测积雪提供了理论基础。

图1   积雪覆盖地表微波辐射

Fig.1   Microwave radiation of the surface with snow cover

2.2 被动微波遥感数据介绍

微波传感器主要接收来自于积雪和其下垫面的辐射能量,传感器以亮温值表示接收到的能量,而积雪的属性信息(如深度、雪水当量)与亮温呈现一定的函数关系。在多年的积雪参数反演研究中,主要用到了国际上已有的星载被动微波传感器包括Nimbus-7 上的SMMR(Scanning Multichannel Microwave Radiometer),DMSP (Defense Meteorological Satellites Program)系列卫星上的SSM/I(Special Sensor Microwave Imager)和SSMIS (Special Sensor Microwave Imager/Sounder),Aqua卫星搭载的AMSR-E(Advanced Microwave Scanning Radiometer-EOS)传感器,GCOM-W1(Global Change Observation Mission-Water)卫星上的AMSR2(Advanced Microwave Scanning Radiometer 2)传感器,以及FY-3系列卫星上的MWRI (Microwave Radiation Imager) 传感器(表1)。

表1   常用被动微波传感器特征参数

Table 1   Parameters summary of passive microwave sensors

传感器卫星平台运行时间频率/GHz极化方式(H,V)瞬间视场/(km×km)
SMMRNimbus 71978.10-1987.086.6H,V136×89
10.7H,V87×57
18.0H,V54×35
21.0H,V47×30
37.0H,V47×30
SSM/IDMSP1987.07-2009.1119.35H,V70×45
22.24V60×40
37.0H,V38×30
85.5H,V16×14
SSMISDMSP2006.12-19.35H,V70×45
22.24V60×40
37.0H,V38×30
91.66H,V38×30
AMSR-EAqua2002.06-2011.106.93H,V75×43
10.65H,V51×39
18.7H,V27×16
23.8H,V32×18
36.5H,V14×8
89.0H,V6×3
AMSR2GCOM-W12012.05-6.93H,V62×35
7.3H,V62×35
10.65H,V42×24
18.7H,V22×14
23.8H,V26×15
36.5H,V12×7
89.0H,V5×3
MWRIFY-3B/3C2010.11-10.65H,V51×85
18.70H,V50×30
23.80H,V27×45
36.50H,V18×30
89.00H,V9×15

新窗口打开

随着1978年搭载于Nimbus卫星和DMSP系列卫星上的被动微波传感器发射,基于被动微波遥感的积雪研究得到了快速发展。SMMR,SSM/I以及SSMIS的仪器特征参数已在表1中列出。相较于SSM/I和SSMIS,2002年发射的Aqua卫星上搭载的AMSR-E传感器具有更高的空间分辨率和更多的频段,其微波频率位于6.9~89 GHz。现在所使用的这4种被动微波传感器数据(SMMR,SSM/I,SSMIS和AMSR-E)都来自美国国家冰雪数据中心(The National Snow and Ice Data Center,NSIDC),为了保证数据最大一致性,对这4种传感器的数据全部采样,以25 km×25 km的等面积可扩充地球格网(Equal-Area Scalable Earth Grid,EASE-Grid)数据格式存储[32],在南北极地区采用的是等面积极地方位投影,全球视角采用的等面积圆柱投影。

我国星载被动微波传感器的发射较晚,2008年发射的风云三号A星(FY-3A)搭载有被动微波成像仪,它是我国首个可以获取积雪参数的业务化卫星。由于仪器自身的问题,FY-3A卫星搭载的MWRI并没有获取连续的观测数据[33]。自2010年FY-3B发射成功后,FY-3B MWRI在轨平稳运行,仪器各方面的性能都优于FY-3A星,最新的MWRI传感器已于2015年搭载于FY-3C卫星发射成功,其现在已经可以提供中国区域乃至全球的积雪参数产品[33,34,35,36,37]。中国气象局已经将FY-3系列的被动微波遥感数据采样成以HDF5格式存储的10 km×10 km的格网[38]。我国FY-3系列卫星搭载的MWRI传感器为现有积雪参数反演研究增加了新的数据源,可作为其他国际通用的被动微波数据的补充。

3 积雪深度和雪水当量反演算法

针对积雪参数(积雪深度和雪水当量)的反演研究已经有大量的成果。积雪深度和雪水当量是2个不同的积雪物理参数,一般情况下当得到其中一个物理量时,就会很容易得到另外一个物理参数。积雪深度和雪水当量之间的转化一般采用如下公式[38,39]:

SWE=SD×ρsnow,(2)

式中:ρsnow为积雪密度,SD为积雪深度,SWE为雪水当量。积雪密度的取值通常有2种方式,一种是使用固定值[29,40,41],即采用一个固定的积雪密度值用于积雪深度和雪水当量的关系转换,如0.24 g/cm3;由于积雪密度在大的空间尺度上具有较强的异质性,不同地区和不同时间内热积雪密度会受风、地形、下垫面植被、时间、气温等的影响而不同,随之就出现了大量针对积雪密度演化的研究工作,进而就有另外一种方法——采用动态积雪密度[42,43,44]进行积雪深度和雪水当量之间的转化,即积雪密度值非固定值,其会随着时间而变化,且/或随积雪类型的变化而出现不同的积雪密度值。

本文将学者们所提出的积雪深度和雪水当量反演算法大致归结为以下几类:基于统计的线性反演算法、基于微波积雪模型的反演算法、基于先验知识的非线性反演算法、数据融合与数据同化算法。

3.1 基于统计的线性反演算法

3.1.1 静态反演算法

积雪深度静态反演算法是国际上应用最为广泛的一种反演算法,最早是由Chang等[24]提出,其假设积雪粒径为0.3 mm,积雪密度为0.3 g/cm3,采用辐射传输模型模拟结果与地面站点观测数据做线性回归分析,最终得到积雪深度与18和37 GHz水平极化亮温差之间的线性关系式(公式(3))。Aschbacher[45]提出SPD(Spectral Polarization Difference)算法,它是基于亮温梯度所提出的积雪深度反演算法,即以19和37 GHz的水平和垂直极化下的亮温差作为回归算法的基础(公式(4)和(5)),其他研究者对SPD算法也做了相关的研究工作[46,47,48]。由于中国区域内的积雪密度相对偏低,而且深霜层较为发育,与北美及其他地区的积雪属性有很大的差异,前期车涛[49]在中国区域内对Chang算法做了初步修订,分别针对SMMR和SMM/I传感器提出了适合中国区域积雪深度反演算法(公式(6)和(7)),此处简称车涛1算法,具体公式见表2

表2   基于亮温梯度的积雪深度静态反演算法

Table 2   The static snow depth retrieval algorithm based on brightness temperature gradient

算法名称算法公式参考文献公式编号
Chang算法SD=1.59×(Tb18h-Tb37h)[24](3)
SPD算法SPD=(Tb19v-Tb37v)+Tb19v-Tb19h)
SD=A0×SPD-A1
[45](4)
(5)
车涛1算法SD=0.78×(Tb18h-Tb37h)+b
SD=0.66×(Tb19h-Tb37h)+b
[49](6)
(7)

注:SD为积雪深度,Tb18h和Tb37h分别为SMMR传感器中18和37 GHz水平极化亮温;Tb19v,Tb19h,Tb37v和Tb37v为SSM/I传感器中19和37 GHz水平垂直极化亮温(下面的公式字母含义与此相同,后面不再赘述),A0=0.68,A1=-0.67;b为补偿值,随月份变化

新窗口打开

在提出仅使用亮温数据的积雪深度反演算法后,由于森林、水体、高程、地形等对积雪属性的变化产生影响,其后又有众多研究者对该算法做了很多的验证和改进工作,形成了基于森林覆盖率、下垫面类型等参数修订的积雪深度静态反演算法。

由于植被会衰减来自积雪的上行微波辐射,植被冠层的存在会增加积雪深度反演的难度。基于上述植被覆盖带来的积雪深度反演问题,Foster等[50]对Chang算法做了修订,认为在不同的地区雪深反演算法不应是一样的,故引入植被覆盖度参数来修订反演算法提高积雪深度的反演精度(公式(8)),随后Foster等[51]又将积雪分类数据和地表覆盖类型数据引入到积雪反演中,并建立新的雪水当量反演方法。针对AMSR-E被动微波数据,有研究者在Chang算法基础上提出了改进的积雪深度反演算法(公式(9)),该算法考虑了森林覆盖度及湿雪对积雪估算的影响[41,52]。我国已有很多研究人员利用微波遥感反演和估算积雪深度或雪水当量,曹梅盛等[53]利用数字地形数据将中国西部地区分成了5个地貌单元(高山、高原、盆地、丘陵和低山),通过对Chang算法的修订得出了利用SMMR微波亮温反演积雪深度的修订算法,并分析了中国西部地区的积雪时空分布特征。车涛等[54]结合森林覆盖度,并剔除湿雪、冻土、降水和寒漠等像元,对公式(5)和(6)做了进一步的修正,得到了新的积雪深度反演算法(公式(10)和(11)),使之在中国区域内具有较小的估算误差。蒋玲梅等[37]利用不同频率的亮温对积雪深度的敏感性不同,使用AMSR-E微波亮温和地面台站数据建立中国区域不同地表覆盖类型下的积雪深度半经验统计反演算法(公式(12)),最后将此算法应用于FY-3B被动微波遥感数据估算中国区域积雪深度,研究结果显示在估算纯像元区域站点实测积雪深度时均方根误差(Root-Mean-Square Error,RMSE)在2~6 cm,而混合像元区域积雪深度反演的残差值分布在-5~+5 cm之内(表3)。

此类使用亮温差与站点实测通过线性回归得到的积雪深度反演公式,适用于对研究区域的积雪属性(积雪粒径、积雪密度等)不了解的情况下,当在反演公式里对森林、地形等影响因子参数化后,在一定程度上也能提高反演的精度。但是此类算法的普适性较差,不同地区的系数会大不相同,因而不能准确为其他地区提供高精度的积雪参数估算结果,而且此类算法并没有考虑积雪物理属性特征的变化。

表3   基于森林覆盖率修订的积雪深度静态反演算法

Table 3   The static snow depth retrieval algorithm based on forest fraction

算法名称算法公式参考文献公式编号
Forest 算法SD=0.78×(Tb18h-Tb37h)/(1-f)[50](8)
AMSR-E 算法SD=f×SDf+(1-f)×SD0[41,52](9)
车涛2算法SD=(0.78×(Tb18h-Tb37h)+b)/(1-f)
SD=(0.66×(Tb19h-Tb37h)+b)/(1-f)
[54](10)
(11)
蒋玲梅算法SD=fgrass×SDgrass+fbarren×SDbarren
+fforest×SDforest+ffarmland×SDfarmland
[37](12)

注:SD为积雪深度,b为补偿值,随月份变化;f为森林覆盖率;SDf根据表示森林区域积雪深度反演公式计算得到的积雪深度;SD0为根据非森林区域的积雪深度反演公式得到的积雪深度。详细的算法信息在参考文献[46,52]中有详细的说明;fgrass,fbarren,fforestffarmland分别为草地、裸地、森林和农田在像元内覆盖度;SDgrass,SDbarren,SDforest和SDfarmland分别为在纯像元内建立对应地表类型下的积雪深度反演算法,公式的详细信息在参考文献[37]有详细介绍

新窗口打开

3.1.2 动态反演算法

动态反演算法是在静态积雪深度反演算法的基础上,考虑了积雪物理属性(积雪粒径和积雪密度等)的变化对微波亮温的影响而做的修正。Josberger等[8]利用温度梯度指数(Temperature Gradient Index, TGI)来描述在积雪上温度梯度的累积效应,用积雪底层和表层的温度差表示温度梯度,它表示了积雪层内积雪粒径增长的一个累积变化指标(公式(13))。研究发现温度梯度指数与亮温差存在一个近似的线性关系,然后就可以利用温度梯度计算积雪深度(公式(14)),一般会假定地表温度为零度。研究结果发现公式(14)在北美平原地区的积雪深度反演精度较高,但在极地地区的精度并不高。Kelly等[55]认为植被的存在会对积雪的微波微波辐射信号产生影响,以及积雪的物理属性(积雪密度和积雪粒径)也会随着时间而产生变化,该研究者利用致密介质模型(Dense Media RadiativeTransfer Model,DMRT)模拟积雪粒径和密度的变化,使用SSM/I被动微波亮温差和实测积雪深度发展了雪水当量的反演模型,将积雪属性的变化对亮温的影响融入到积雪参数的反演过程中。也有研究者结合静态和动态算法各自的优势,形成了动、静结合的积雪深度反演算法[56,57]

TGI=1CTground-TairD(t)dt,(13)

D(t)=-αTairCddt(ΔTb19h-37h),(14)

式中:C为比例常数,20 ℃/m;Tground为积雪的底层温度;Tair为空气温度;D(t)为积雪深度;ΔTb19h-37h为19和37 GHz的水平极化亮温差;α为经验系数。

相较于静态算法,动态算法考虑积雪属性的变化,在某些地区的积雪深度或雪水当量的估算中精度有所提高,但是普适性受到了限制,在不同的地区得到的参数也会不同,主要是因为植被、大气和地形异质性的影响。

3.2 基于微波积雪模型的反演算法

为了可以更准确地描述微波在积雪内传输特征,目前建立了多种微波积雪模型。HUT(The Helsinki University of Technology)模型和积雪辐射传输模型[58]是一种半经验模型,该模型的基本假设是积雪中热辐射的散射以前向散射为主,然后对散射到前向的辐射能量比例以一个固定的经验系数q=0.96度量。此外HUT模型中考虑了森林覆盖度和大气的影响。在Roy等[59]的研究中发现HUT模型雪水当量的估算结果高于基于亮温差的线性回归算法,反演误差小于300 mm。Tsang等[60]应用致密介质辐射传输理论模型(Dense Media Radiative Transfer, DMRT)计算积雪中电磁波的有效传播常数, 并据此修正积雪的消光系数、散射系数和反照率,以此来准确反映积雪的微波辐射特性。Chen等[61]借助DMRT 模型和其他数据提高反演积雪深度和雪水当量的精度,并很好地模拟了在不同积雪粒径情况下亮温差和积雪深度的关系。多层积雪的微波辐射传输(Microwave Emission Model of Layered Snowpacks,MEMLS)模型[62]是一种采用六流技术描述积雪层内多次散射和吸收过程,并考虑积雪层之间的界面散射,研究发现该模型较好地模拟了积雪的微波辐射亮温,也可以模拟出亮温差在积雪深度达到50 cm左右时的饱和现象。

积雪模型一般可以较好地模拟出微波在积雪层内的辐射传输过程,以及对积雪各个参数的变化所引起的辐射亮温的变化。目前所提出的积雪模型一定程度上对积雪的辐射传输做了简化处理,而且模型中的参数有时很难通过实测来获取,从而限制了模型在大范围内使用。

3.3 基于先验知识的非线性反演算法

大量的研究表明,微波亮温与积雪参数之间是非线性函数关系,是多对多的复杂函数关系。因此发展一个显式的积雪参数反演关系式是不现实的,需要引入非线性的方法,如:神经网络、支持向量机、决策树、贝叶斯方法和遗传算法等机器学习算法来提高积雪参数的估算精度[63]。到目前为止,这些非线性的方法已经较为成功地应用于积雪深度和雪水当量的反演估算研究中。

3.3.1 神经网络

神经网络方法已经应用到了各个领域:控制和优化、模式识别和图像处理、预测预报等,目前神经网络也被应用于地表参数(包括土壤湿度、积雪深度、雪水当量等)估算研究中。Davis等[64]使用神经网络训练DMRT并反演积雪的4个参数(即积雪粒径、积雪深度、积雪密度和积雪温度),其中使用了SSM/I的5个观测值(19和37 GHz的水平和垂直极化以及22 GHz垂直极化亮温),估算结果显示积雪密度误差小于10%,雪水当量误差为9%~57%。Tedesco等[65]利用神经网络技术反演了积雪深度和雪水当量,并与其他4种积雪反演算法的估算结果作对比,研究结果显示神经网络估算积雪深度和雪水当量精度最高,Tabari 等[66]在伊朗Samsami盆地使用神经网络估算的积雪深度和雪水当量得到了同样的结论。此外,Forman等[67]在北美积雪覆盖条件下使用神经网络模拟估算被动微波不同频率和极化方式下的亮温,并对比分析了AMSR-E观测结果与神经网络模拟结果,模拟结果产生了接近于零的偏差,而且通过在积雪积累期和积雪融化期所做的神经网络模拟实验,研究结果证明神经网络可以作为大陆尺度数据同化的高效的计算方法。

多项研究表明微波亮温差随着积雪深度的增加而增加,但是当积雪深度超过一定值时,积雪深度的反演结果会有很大的偏差[41,68,69],单一的线性回归算法已经无法反演深雪。神经网络可以克服大尺度中存在的各种复杂问题,如通过学习和归纳大量数据知识进行非线性建模、分类以及关联,并且不需要在建模时对物理过程有过多的先验知识或了解[63,70]。神经网络尽管有这些优点,但是也有自身不可回避的问题:首先,对输入参数的要求必须是相关性较小,否则输出结果会存在很大的误差;其次,神经网络的结构对于高精度的输出结果也很重要;最后,神经网络对于训练样本的依赖都将限制神经网络的进一步应用。

3.3.2 支持向量机

支持向量机(Support Vector Machine, SVM)在图像分类中应用较为广泛,现在很多研究者已经将该方法应用于地学中,解决地学中的非线性问题。Liang等[47]集成了被动微波遥感亮温和反射率数据,然后使用SVM回归算法反演新疆地区的积雪深度。通过与其他积雪深度反演算法的对比分析得出基于SVM方法的积雪深度反演算法可以以较高精度估算积雪深度。Xiao等[48]使用被动微波亮温数据和气象台站资料以及其他辅助数据,并运用支持向量回归(Support Vector Regression, SVR)算法建立了基于不同植被类型和不同积雪期情况下的积雪深度反演模型,实验结果显示,该研究所提出的算法与其他现有的4种算法(Chang算法,SPD算法、神经网络算法和线性回归算法)相比,可以提高积雪深度反演结果精度,并在一定程度上减少“积雪饱和效应”。此外Forman等[71]使用SVM模拟估算北美地区积雪覆盖条件下AMSR-E不同频率和不同极化方式下(10.65,18.7,36.5 GHz水平和垂直极化)的亮温数据,其后又与基于神经网络算法估算的被动微波亮温作对比,研究发现基于SVM的估算结果优于神经网络的估算结果,因此,SVM是一种替代神经网络的更好的估算方法。再者,Xue等[72]在北美积雪覆盖条件下使用被动微波遥感数据对比分析了人工神经网络(Artificial Neural Network,ANN)和SVM估算积雪参数时的预测敏感度,研究结果发现不论深浅雪区域,有无森林覆盖地区,以及积雪积累和融化期,基于SVM算法的结果都很敏感。这个主要是由于SVM的模型结构相比于人工神经网络具有优势,对积雪属性的变化更敏感,可以做出更加准确的回馈效果。

3.3.3 其他机器学习算法

相对于神经网络,相关的研究表明决策树的优势在于有清楚的规则并可以训练得更快,且决策树的规则简单容易理解[63,70]。Balk等[73]使用二元决策树方法和地理统计技术对山区积雪空间分布特征做了建模分析,结果显示该方法可以提升山区雪水当量的预测。Davis等[64]在其研究中使用SMMR被动微波数据和贝叶斯迭代方法反演积雪参数,研究结果显示该方法的反演结果优于线性亮温梯度算法的估算结果。武黎黎等[74]使用改进的HUT积雪辐射传输模型(The Improved Helsinki University of Technology,IMPHUT)模拟18.7和36.5 GHz水平极化亮温,然后利用遗传算法反演积雪深度,反演结果显示基于IMPHUT模型的反演雪深结果精度优于HUT模型和Chang算法,误差、RMSE、平均误差比HUT模型的3个指标均减少一半左右,且反演雪深与实测雪深一致性较高。

基于先验知识的非线性反演算法较好地描述了微波亮温和积雪参数之间的非线性关系,它克服了线性算法在不同地区应用时的局限性。虽然该类型算法的使用范围广、反演精度高,但反演过程中缺少详实积雪物理模型过程参与[35],而且机器学习算法对算法自身结构的依赖度很高[72]

3.4 数据融合与数据同化

由于在积雪反演过程中存在各种影响反演精度的因素,例如,被动微波的空间分布率低,微波探测的深雪饱和效应,植被、水分等,使得基于被动微波遥感数据的反演结果并不能如实反映积雪属性的地面积雪实际信息[75]。站点实测可以直接获得积雪属性的“真值”,但是限于点观测,难以表达大空间范围内积雪变化特征;遥感观测具有多时相、宏观、全天候等特点,但是在不连续的时间尺度上只能获取某个瞬间的地面积雪信息;模型模拟具有明确的物理机制,其可以对过去、现在及未来的积雪参数进行模拟,但这是限定在特定时空尺度下对积雪参数进行的模拟,而且模型模拟大都是对实际积雪属性的变化做了一定简化处理,因此影响了模型模拟的精度[76]。卫星资料、地面站点资料及模型模拟都具有各自优势和局限性,根据不同来源的积雪资料进行时空分布和变化分析时可能存在着差异,会得出不同的结论[75]。为了综合不同来源观测资料的优势,实现各种资料源的优势互补,提高积雪数据的质量,众多研究人员提出不同数据源之间的融合和同化方法[77],例如,观测资料与模型模拟之间,观测资料与遥感资料之间,遥感资料与模型模拟之间,观测资料、遥感资料和模型模拟三者之间。

3.4.1 观测资料与模型模拟

Liston等[78]采用直接将积雪深度转化后的雪水当量值同化到区域大气模型系统(Regional Atmospheric Modeling System)模拟中,即把积雪的观测值直接代入到耦合的陆地—大气模型模拟结果中,改进雪水当量的估算结果,研究结果显示该方法可以显著提高积雪模型的积雪过程模拟。加拿大气象中心使用积雪深度实时实测数据和简化的积雪融化与积累期积雪模型,生成了全球逐日积雪深度再分析数据集[79],与北半球积雪深度实测值相比每个月估算值的RMSE在10~20 cm,偏差一般小于5 cm。而且在北美陆地数据同化系统(North American Land Data Assimilation System,NLDAS)中,也是将地面站点实测积雪数据同化到陆面过程模式中[80],结果显示在所有区域内模型估算都会低估最大雪水当量值,估算值与实测值的差异最大可以达到1 000 mm,东部地区一般会在100 mm以内。

3.4.2 观测资料与遥感资料

Takala等[40]提出了一种数据同化算法,同化地面实测积雪深度数据和被动微波辐射数据,并计算北半球30年长时间序列季节性的雪水当量数据,该数据集使用前苏联、芬兰以及加拿大的雪水当量观测值作为验证数据集。研究结果显示该算法具有较高的可行性,可以提高估算雪水当量的精度,在芬兰地区雪水当量估计值的RMSE均小于40 mm;欧亚大陆区域的RMSE和偏差会随着时间而变化,当雪水当量小于150 mm时RMSE和偏差分别在30~40 mm和-3~9 mm变化;在加拿大地区RMSE大概为40 mm。赵亮等[81]提出了一种融合被动微波遥感数据和地面测站数据动态反演积雪深度的方法,该反演方法可以在不同时空条件下根据测站观测雪深,灵活调整积雪深度反演系数,而且该算法在观测站稀少的西部高山地区的反演结果具有较高空间分辨率,与站点实测相比、判识有无雪的准确率高于80%。

3.4.3 遥感资料与模型模拟

Andreadis等[82]运用集合卡尔曼滤波(Ensemble Kalman Filter,EnKF)将遥感观测数据同化到水文变量渗透能力模型(Variable Infiltration Capacity hydrological model,VIC)中,使用1999—2003年4个冬季的MODIS积雪覆盖范围产品更新VIC模拟的雪水当量。在此过程中应用简单的积雪消融模型反演雪水当量—积雪面积,研究发现在深雪部分估算的雪水当量误差更大,对低海拔高度积雪融化期地区的雪水当量精度提高幅度大,对高海拔积雪积累期精度提高幅度不是很大。此外,为了更新积雪的状态信息,Durand等[83]将被动微波传感器SSM/I和AMSR-E多个波段的亮温以及宽波段albedo观测数据使用EnKF同时同化至陆面过程模式中估算雪水当量,其中该研究中使用集成陆面过程模式SSiB3中简单积雪—大气—土壤传输模型(Simple Snow-Atmosphere-Soil,SAST),研究结果显示,该方法对研究积雪属性和观测值之间的复杂关系具有很好的效果,实验结果显示经同化后的结果对积雪面积识别率可达到90%以上,显著提高了雪水当量估算的精度,雪水当量估算值的RMSE可以低至32 mm。车涛等[84]使用EnkF将被动微波亮温数据同化至积雪过程模型中,使用积雪深度地面实测数据作为同化反演结果的验证数据,通过对实验区7个站点的数据分析,结果显示该同化算法在积雪积累期可以提高估算积雪深度的精度,同化后结果的偏差和RMSE分别是-1.13 cm和4.60 cm,但是在积雪融化期并不能提高积雪深度估算精度,森林、降水以及模型所需的驱动数据都会影响同化结果的精度。

3.4.4 观测资料、遥感资料和模型模拟

Pulliainen[85]同化卫星微波辐射数据和地面观测数据来估算积雪深度和雪水当量。该同化算法基于HUT模型,并运用贝叶斯方法对卫星数据与地面实测数据赋权重。实验结果显示通过数据同化技术并使用SSM/I和AMSR-E数据估算北欧和芬兰地区积雪深度和雪水当量时,与实测插值后结果相比数据同化可以提高积雪深度和雪水当量的反演精度,但是由于该算法对积雪粒径很敏感,所以该模型在长时间序列或大空间尺度上应用时存有一定的不足。

4 被动微波卫星遥感积雪产品

目前,使用被动微波遥感数据反演得到的积雪产品主要应用于:①作为大尺度气候模型的输入和验证数据;②天气预报及中尺度模型;③数值天气预报;④气候变化的监控与监测;⑤水文监测和模拟融雪径流等;⑥农业方面:霜冻旱涝灾害预报与预警;⑦冰盖融化的判识;⑧湖面冰盖冻结与融化[86]。基于不同积雪深度和雪水当量反演算法可以得到不同的积雪数据产品,现在已经有很多种积雪深度数据产品,如不同时间尺度、不同空间尺度和不同数据格式,但是本文中主要介绍3种空间尺度范围下的积雪深度数据产品:中国区域、北半球尺度和全球尺度(表4)。

表4   积雪产品数据集

Table 4   Snow product dataset

序号数据名称空间范围空间分辨率时间范围时间
分辨率
获取地址
1中国雪深长时间序列
数据集
中国25 km(或0.25°)1978.10.24-2016.12.31逐日http:∥westdc.westgis.ac.cn
2FY3-MWRI雪深雪水
当量产品
全球25 km2011.7.15-逐日http:∥www.nsmc.org.cn
3AMSR-E积雪产品全球25 km2002.6.19-2011.11.3逐日http:∥nsidc.org/
4SMMR、SSM/I 和
SSMIS积雪产品
全球25 km1978.11.11-逐日http:∥nsidc.org/
5GlobSnow-2积雪产品北半球25 km1979—2012年逐日http:∥www.globsnow.info/
6加拿大气象中心逐日
积雪深度分析数据
全球24 km1998.08.01-2016.12.31逐日http:∥nsidc.org/
7兰德公司月平均积雪
深度数据集
全球(不包含
非洲和南美)
4°×5°1950—1976年逐月http:∥nsidc.org/

新窗口打开

4.1 积雪产品生成的算法

中国雪深长时间序列数据集使用的是车涛等[54]中国区域对Chang算法修正后的反演算法。中国地区积雪深度反演公式如公式(10)和(11):SMMR传感器下的被动微波数据使用公式(10)反演积雪深度,SSM/I和SSMIS下的被动微波数据使用公式(11)反演积雪深度。需要注意的是,使用车涛的雪深反演算法前,会基于Dai等[87]的研究将3种传感器下的微波亮温数据交叉定标。

FY3-MWRI雪深雪水当量数据产品使用的积雪深度反演算法有2种:一种是全球积雪深度反演算法,该算法是半经验型算法,与AMSR-E积雪深度反演算法相同(公式(9));另一种是区域积雪深度反演算法,该算法也是一种半经验型算法,但是该算法只适用于中国区域(公式(12))。

AMSR-E的积雪产品可以提供全球逐日、5天和逐月3种类型的积雪产品,该数据集的积雪反演算法是在Chang算法的基础上发展起来的,数据处理时首先会剔除湿雪像元,然后使用MODIS数据将非积雪像元掩膜掉。因为算法考虑了森林的影响,所以会基于森林和非森林区域分别建立反演算法。最后使用公式(9)估算积雪深度。

SMMR积雪产品空间覆盖南、北纬85°之间区域,SMMR积雪产品是在时空上做了插值处理,以极地方位立体投影形式显示,数据限制在0.5°×0.5°经纬格网内,数据中只有陆地区域,其中海洋和海湾地区数据被掩膜。SMMR,SSM/I和SSMIS积雪产品是美国NSIDC依据NASA算法[50,88]使用SMMR,SSM/I 和SSMIS被动微波传感器亮温数据生成的长时间序列(1978年至今)全球积雪数据产品。

GlobSnow-2数据集是欧洲太空局(European Space Agency,ESA)提供的积雪产品,包括积雪面积和雪水当量,雪水当量反演算法——数据同化方法[85]使用到了NSIDC的被动微波亮温数据(SMMR,SSM/I,AMSR-E)辅助以时间序列融化探测算法,整体算法在文献[40]中有详细介绍。

加拿大气象中心逐日积雪深度数据来自于实时地面的积雪深度实测值,使用最优插值法和一个简单的积雪融化和积累期的积雪模型,其中积雪模型使用的是来自于加拿大预测模型(Canadian Forecast Model)的气温分析数据和降水预测数据[79]。19世纪80年代初,兰德公司将所有可以获取到的气象站点的每月积雪深度集成到一个全球数字数据集——兰德公司月平均积雪深度数据集,然后以4°纬度×5°经度进行了网格划分。

4.2 积雪产品的精度

戴礼云等[38,68]在中国北方地区,用站点实测对比了积雪产品1(即中国雪深长时间序列数据集)、3(即AMSR-E)和5(即GlobSnow-2积雪产品),发现积雪产品3的误差较大即高估了雪水当量,主要是由于该产品的反演算法中积雪粒径和雪密度的信息不是非常准确;因为产品5使用了数据同化算法,同化了地面台站的实测数据,很明显产品5的精度高于产品3;相较于产品3,产品1在该研究区域的精度是相对较高,而且研究发现森林地区的积雪产品精度都相对不高,但是李小兰等[75]在研究中发现,产品1和中国区域的地面台站实测资料相比,在不同的区域,误差存在很大的差异。产品4(即SMMR,SSM/I 和SSMIS积雪产品)低估了前苏联地区积雪深度,而在中国的内蒙古地区对积雪深度和积雪面积存在明显的高估[89,90],这主要是由于我国的积雪密度偏低,深霜层较为发育。Liu等[91]在全球尺度下对比分析了积雪产品3(AMSR-E积雪产品)、4(SMMR,SSM/I和SSMIS积雪产品)和5(GlobSnow-2积雪产品)的雪水当量精度,结果显示产品5的雪水当量估算精度比产品3和4的都高,产品3和4更易受到积雪饱和效应的影响,在雪水当量大于120 mm时会显著低估。

5 影响积雪深度反演精度的因素

5.1 积雪的变质作用

由于受多种因素的影响,在运用被动微波遥感进行积雪监测时存在很多局限性。一个很大的影响因素是积雪内的液态水,它抑制了微波传感器探测雪的能力[92]。随着积雪中液体水含量的增加(意味着积雪介电常数的增大),微波辐射的吸收将逐渐超过微波散射而在衰减作用中占主导地位,导致积雪的散射特征减弱。在一定程度上,湿雪的存在会使积雪深度或雪水当量与微波亮温的变化没有关系[93,94]

由于积雪粒径的大小和微波波长很接近,可以有效地散射从地球表面发出的微波辐射。而积雪粒径越大,散射作用就越强。例如,以亮温差为主导的积雪深度线性反演算法中,积雪粒径偏大会导致亮温差越大,积雪深度估计值偏大[24]。而在同一层积雪的粒径会随着时间在积雪的变质作用下逐渐增大;在垂直剖面上,每个积雪层的粒径也会不一样;底层的积雪层积雪粒径最大,即深霜层是整个积雪剖面上粒径最大的。当积雪粒径发生改变后,其会改变积雪的微波辐射特性[51]

了解积雪粒径和密度对准确反演积雪深度和雪水当量很重要,而积雪密度在积雪变质的过程中也会发生变化。根据现有的研究发现欧亚大陆地区的积雪密度普遍比北美地区的积雪密度低,原苏联地区的积雪密度为0.21~0.31 g/cm3,北美地区的积雪密度为0.24~0.31 g/cm3[25],而中国地区的积雪密度更低,为0.16 g/cm3[95,96] ,深霜层也较为发育。在积雪的垂直方向上,顶层雪为新雪,其密度最小;从垂直剖面自上而下,积雪密度逐渐增大,积雪密度可以从0.12 g/cm3增大到0.22 g/cm3[68],因为下层积雪正经历着积雪压实,积雪粒径增大。

5.2 积雪下垫面

由于被动微波遥感数据的空间分辨率较低,在一个像元内会存在多种地物类型,如裸地、森林、草地等,植被又对积雪的微波辐射信号具有衰减作用,这就使得像元内地物的异质性增加了微波遥感反演积雪深度或雪水当量的复杂度[29,50,54,97]。Vander等[98]在其研究中指出植被的存在使微波信号对于积雪深度的敏感度减少23%~63%。对于不同地表覆盖类型,其被动微波辐射的衰减效应不同,在不同的植被覆盖度下的微波辐射衰减也不同。有研究者已经基于不同的地表覆盖类型或植被类型建立了不同的积雪深度或雪水当量的反演估算模型[37,48,51,90,99,100],或在积雪深度或雪水当量的反演方程中引入森林覆盖度提高积雪参数的反演精度[50,51,54]。为了能够准确对积雪进行定量或者建模研究,需要对每种地表覆盖类型或不同植被覆盖度下的积雪做研究。

对混合像元的研究,周胜男等[101]评估了混合像元内优势地物类型下AMSR-E积雪深度的反演精度,研究结果显示当植被由高到低再到无时(森林、灌木、草原和裸地),积雪深度的高估程度不断减少。现有的研究表明采用混合像元分解会使积雪的反演精度得到一定程度的提高[37]

5.3 其他影响因素

冻土、沙漠等地物的微波辐射信号对识别积雪微波辐射信号存在干扰。积雪的反演是积雪越厚,微波散射越强,不同频率之间的亮温差越大,但是由于积雪的微波辐射信号特征与降雨、冻土以及沙漠的具有相似的信号,会高估积雪深度或雪深当量[54,102]

在青藏高原或其他高海拔地区,由于大气稀薄从地表传输到卫星传感器的微波信号衰减可能较小,会高估这些高海拔地区的积雪深度或雪水当量[103,104]。之前的研究采用被动微波亮温数据反演积雪深度或雪水当量时,一般都默认反演过程中大气作用的存在,然而在高海拔地区(如青藏高原),空气稀薄,Savoie等[103]将青藏高原地区调整到低海拔地区,相当于增加了大气对于微波辐射信号的影响。也有研究者利用简化的辐射传输方程和从几个台站收集的大气探空资料对微波亮温数据进行了大气校正,研究发现经大气校正后的微波亮温可以提高积雪面积判识的精度[105]

复杂地形会影响积雪微波散射信号的分布,使得在25 km像元内的微波亮温并不能如实反映地面积雪信息,大大限制了卫星传感器接受到的积雪散射信号。在积雪深度或雪水当量反演时,复杂地形会影响积雪深度的反演精度[41,106~108]。例如,青藏高原地区地形差异很大,山体阴坡和阳坡同类地物的微波亮温差异都很大[15]。因此复杂地形地区的积雪研究成为了需要解决的问题,但有研究者在做积雪反演研究时对山区做出剔除处理。

此外,微波信号的穿透深度也会影响深雪的反演精度,微波信号不是无限穿透所有深度的积雪,蒋玲梅[86]在研究中发现,积雪粒径、含水量、积雪密度、积雪温度等都会影响微波在积雪内的穿透深度,也会随着微波频率的变化发生变化。其研究发现随着频率的增加,积雪的穿透深度不断减小。当使用被动微波亮温数据反演积雪深度时,就会存在饱和效应,即当积雪深度达到某个阈值时,微波亮温不会发生大的变化[25,69]

6 结 语

以上综述了基于被动微波遥感的积雪深度和雪水当量的反演研究进展。简述了被动微波遥感反演积雪的理论基础,并简要介绍了用于反演研究所用的被动微波遥感数据,随后较详细地介绍了4种类型的积雪深度和雪水当量反演算法(基于统计的线性反演算法,基于微波积雪模型的反演算法,基于先验知识的非线性反演算法,及数据融合与数据同化)。同时也介绍了目前的几种积雪数据产品,包括中国雪深长时间序列数据集,FY3-MWRI雪深雪水当量产品,AMSR-E积雪产品,SMMR,SSM/I和SSMIS积雪产品,GlobSnow-2积雪产品,加拿大气象中心逐日积雪深度分析数据,兰德公司月平均积雪深度数据集。最后对影响积雪深度的几个主要影响因素做了简要分析。现阶段被动微波遥感技术仍是我们认识积雪(冰冻圈中最为活跃也最为重要的要素之一)的主要手段,但是现有的基于被动微波遥感获取积雪参数的方式方法仍然有很多不确定性。存在这些不确定性的原因是:①被动微波遥感空间分辨率太粗;②不同区域积雪自身物理属性(密度、粒径、含水量、雪层结构等)的异质性对积雪的微波辐射和散射信号造成影响。而要解决反演中的不确定性问题,需要加强以下几个方面的研究:

(1)微波传感器接收到的亮温对积雪粒径、含水量、积雪密度、植被、地形等变化敏感,浓密的植被大大减弱了向上微波辐射信号,地形的高低起伏增大了卫星传感器接受到的积雪辐射信号的复杂性,而目前所使用的积雪反演方法在某种程度上经常会低估实际值,因而利用被动微波积雪监测中如何减小植被、地形等地表属性以及积雪的物理属性的影响有待进一步研究。

(2)因被动微波遥感数据的粗分辨率而带来的混合像元问题,导致了基于被动微波遥感反演积雪参数时并不能准确描述地表真实积雪信息。所以将其他高分辨率遥感影像数据或产品,如MODIS,Landsat,Lidar和SAR等数据,与被动微波遥感数据融合处理,使得融合处理后的数据更好地表达被动微波遥感数据粗分辨率像元内的异质性信息以及提高多源数据的利用率,实现优势互补,弥补各自传感器数据的缺陷。使用多源遥感数据融合可以更加准确地了解积雪属性(密度、粒径、深霜层、雪层结构)的特征,经数据融合处理生成的高精度积雪遥感产品可用于小区域尺度的高精度的水资源评估、雪灾监测与气象预报等领域。

(3)现有主流的积雪深度和雪水当量反演算法采用的微波频率主要是18 GHz 或19 GHz以及37 GHz,而现有研究结果显示目前的反演算法存在一个反演积雪深度阈值的瓶颈,即当雪深达到50 cm左右时亮温差逐渐趋于饱和,需要尝试探索除以上常用频率波段以外频率下的数据,如可以采用穿透深雪能力较强的10 GHz波段以及利用监测浅雪较好的89 GHz 或更高频波段的亮温数据。

(4)数据同化可以破解积雪物理过程不明确、被动微波数据分辨率低以及无法长时间持续监测积雪变化等问题。微波辐射传输模型在积雪参数反演估算过程中的使用和发展,更加有利于了解积雪物理特性和提高积雪参数反演的精度。在数据同化方法体系下同化积雪过程模型、站点观测数据、卫星遥感观测(主被动,光学,雷达等遥感观测)等多源数据,可为提高积雪深度和雪水当量的估算精度提供新的方法。

The authors have declared that no competing interests exist.


参考文献

[1] Hoinkes H.

Glaciology in the international hydrological decade[R]∥IAHS Commission of Snow and Ice, Reports and Discussion.

IUGG General Assembly,Bern, 1967, 79: 7-16.

[本文引用: 1]     

[2] Winther J G, Hall D K.

Satellite-derived snow coverage related to hydropower production in Norway: Present and future

[J]. International Journal of Remote Sensing, 1999, 20(16): 2 991-3 008.

DOI      URL      [本文引用: 1]      摘要

Hydropower derived from snow-melt runoff is a major source of electricity in Norway. Therefore, amount of snow-melt runoff is key to the prediction of available water. The prediction of water quantity may be accomplished through the use of hydrological models. These models, which may be run for individual basins, use satellite-derived snow-covered area in combination with snow-cover depletion curves. While it is known that snow albedo information would increase the accuracy of the models, large-scale albedo measurements have not yet been obtained from satellites on a regular basis. This paper presents Landsat-5 Thematic Mapper (TM) reflectances recorded in May 1989 from a mountainous catchment at Kvikne, Norway. Satellite-derived albedo values are analysed, and compared with simultaneously measured in situ albedo. The satellite-derived shortwave snow albedo is comparable with bare ground albedo and values as low as 0.19 were found in areas where the snow was highly metamorphosed and heavily blackened by organic material. To map snow-covered areas, the contrast between snow and snow-free areas can be improved by using a normalized TM Band 2-5 difference image. While TM Band 2 alone shows varying degrees of snow surface contamination within the study area, the normalized difference snow index (NDSI) is not affected by impurities. This paper also discusses the use of NASA''s EOS (Earth Observing System) Moderate Resolution Imaging Spectroradiometer (MODIS) sensor, which is planned to be launched in the summer of 1999 for mapping of large-scale geophysical parameters including snow-cover. MODIS will enable snow cover and albedo to be mapped in Norway on a daily basis, and should enhance our ability to estimate snow coverage and thus manage hydropower production.
[3] Frei A, Robinson D A.

Northern Hemisphere snow extent: Regional variability 1972-1994

[J]. International Journal of Climatology, 1999, 19(14): 1 535-1 560.

DOI      URL      [本文引用: 1]     

[4] Zhang T.

Influence of the seasonal snow cover on the ground thermal regime: An overview

[J]. Reviews of Geophysics, 2005, 43(4): 589-590.

DOI      URL      [本文引用: 1]      摘要

[1] The presence of seasonal snow cover during the cold season of the annual air temperature cycle has significant influence on the ground thermal regime in cold regions. Snow has high albedo and emissivity that cool the snow surface, high absorptivity that tends to warm the snow surface, low thermal conductivity so that a snow layer acts as an insulator, and high latent heat due to snowmelt that is a heat sink. The overall impact of snow cover on the ground thermal regime depends on the timing, duration, accumulation, and melting processes of seasonal snow cover; density, structure, and thickness of seasonal snow cover; and interactions of snow cover with micrometeorological conditions, local microrelief, vegetation, and the geographical locations. Over different timescales either the cooling or warming impact of seasonal snow cover may dominate. In the continuous permafrost regions, impact of seasonal snow cover can result in an increase of the mean annual ground and permafrost surface temperature by several degrees, whereas in discontinuous and sporadic permafrost regions the absence of seasonal snow cover may be a key factor for permafrost development. In seasonally frozen ground regions, snow cover can substantially reduce the seasonal freezing depth. However, the influence of seasonal snow cover on seasonally frozen ground has received relatively little attention, and further study is needed. Ground surface temperatures, reconstructed from deep borehole temperature gradients, have increased by up to 4脗掳C in the past centuries and have been widely used as evidence of paleoclimate change. However, changes in air temperature alone cannot account for the changes in ground temperatures. Changes in seasonal snow conditions might have significantly contributed to the ground surface temperature increase. The influence of seasonal snow cover on soil temperature, soil freezing and thawing processes, and permafrost has considerable impact on carbon exchange between the atmosphere and the ground and on the hydrological cycle in cold regions/cold seasons.
[5] Schimel J P, Bilbrough C, Welker J M.

Increased snow depth affects microbial activity and nitrogen mineralization in two Arctic tundra communities

[J]. Soil Biology and Biochemistry, 2004, 36(2): 217-227.

DOI      URL      [本文引用: 1]     

[6] Tape K, Sturm M, Racine C.

The evidence for shrub expansion in northern Alaska and the Pan-Arctic

[J]. Global Change Biology, 2006, 12(4): 686-702.

DOI      URL      [本文引用: 1]      摘要

Abstract One expected response to climate warming in the Arctic is an increase in the abundance and extent of shrubs in tundra areas. Repeat photography shows that there has been an increase in shrub cover over the past 50 years in northern Alaska. Using 202 pairs of old and new oblique aerial photographs, we have found that across this region spanning 620km east to west and 350km north to south, alder, willow, and dwarf birch have been increasing, with the change most easily detected on hill slopes and valley bottoms. Plot and remote sensing studies from the same region using the normalized difference vegetation index are consistent with the photographic results and indicate that the smaller shrubs between valleys are also increasing. In Canada, Scandinavia, and parts of Russia, there is both plot and remote sensing evidence for shrub expansion. Combined with the Alaskan results, the evidence suggests that a pan-Arctic vegetation transition is underway. If continued, this transition will alter the fundamental architecture and function of this ecosystem with important ramifications for the climate, the biota, and humans.
[7] Goodison B, Walker A.

Use of snow cover derived from satellite passive microwave data as an indicator of climate change

[J]. Annals of Glaciology, 1993, 17(1): 137-142.

DOI      URL      [本文引用: 1]     

[8] Josberger E G, Mognard N M.

A passive microwave snow depth algorithm with a proxy for snow metamorphism

[J]. Hydrological Processes, 2002, 16(8): 1 557-1 568.

DOI      URL      [本文引用: 2]      摘要

Abstract Passive microwave brightness temperatures of snowpacks depend not only on the snow depth, but also on the internal snowpack properties, particularly the grain size, which changes through the winter. Algorithms that assume a constant grain size can yield erroneous estimates of snow depth or water equivalent. For snowpacks that are subject to temperatures well below freezing, the bulk temperature gradient through the snowpack controls the metamorphosis of the snow grains. This study used National Weather Service (NWS) station measurements of snow depth and air temperature from the Northern US Great Plains to determine temporal and spatial variability of the snow depth and bulk snowpack temperature gradient. This region is well suited for this study because it consists primarily of open farmland or prairie, has little relief, is subject to very cold temperatures, and has more than 280 reporting stations. A geostatistical technique called Kriging was used to grid the randomly spaced snow depth measurements. The resulting snow depth maps were then compared with the passive microwave observations from the Special Sensor Microwave Imager (SSM/I). Two snow seasons were examined: 1988–89, a typical snow year, and 1996–97, a record year for snow that was responsible for extensive flooding in the Red River Basin. Inspection of the time series of snow depth and microwave spectral gradient (the difference between the 19 and 37 GHz bands) showed that while the snowpack was constant, the spectral gradient continued to increase. However, there was a strong correlation (0·6 < R 2 < 0·9) between the spectral gradient and the cumulative bulk temperature gradient through the snowpack (TGI). Hence, TGI is an index of grain size metamorphism that has occurred within the snowpack. TGI time series from 21 representative sites across the region and the corresponding SSM/I observations were used to develop an algorithm for snow depth that requires daily air temperatures. Copyright 08 2002 John Wiley & Sons, Ltd.
[9] Gao Feng, Li Xin, Armstrong R L, et al.

Preliminary application of passive microwave data to operational snow monitoring in Tibetan Plateau

[J]. Remote Sensing Technology and Application, 2003, 18(6): 360-363.

Magsci      [本文引用: 1]     

[高峰,李新,ARMSTRONG R L,.

被动微波遥感在青藏高原积雪业务监测中的初步应用

[J].遥感技术与应用,2003,18(6):360-363.]

DOI      URL      Magsci      [本文引用: 1]      摘要

<p>积雪范围、积雪深度和雪水当量等参数的遥感监测与反演对气候模式的建立以及积雪灾害的评估具有重要意义。被动微波遥感在这些参数的反演方面具有明显优势,但目前尚未应用到青藏高原地区的积雪遥感业务监测上来。2001年10月至2002年4月,利用SSM/I数据对青藏高原地区的积雪范围和积雪深度进行了实时监测,为西藏、青海遥感应用部门提供逐日的雪深分布图。对这次监测的总效果进行了分析和评价,并对发生在青海省内一次较大的降雪过程进行了遥感分析,结果表明:SSM/I反演的积雪范围变化趋势与MODIS结果总体上较为一致;SSM/I的雪深监测结果为当地遥感部门对大于10 cm的雪深做出正确判断提供了重要信息,是对雪灾定位的重要信息源。</p>
[10] Tedesco M, Miller J.

Observations and statistical analysis of combined active-passive microwave space-borne data and snow depth at large spatial scales

[J]. Remote Sensing of Environment, 2007, 111(2): 382-397.

DOI      URL      [本文引用: 1]     

[11] Shi Jiancheng, Xiong Chuan, Jiang Lingmei.

Review of snow water equivalent microwave remote sensing

[J]. Science in China (Series D), 2016,46(4):529-543.

[本文引用: 3]     

[施建成, 熊川, 蒋玲梅.

雪水当量主被动微波遥感研究进展

[J]. 中国科学:D辑, 2016, 46(4): 529-543.]

[本文引用: 3]     

[12] Sun Shaobo, Che Tao.

A review of the research on snow cover monitored with Sythetic Aperture Radar (SAR)

[J]. Journal of Glaciology and Geocryology, 2013, 35(3): 636-647.

Magsci      [本文引用: 1]     

[孙少波, 车涛.

基于合成孔径雷达 (SAR) 的积雪监测研究进展

[J]. 冰川冻土, 2013, 35(3): 636-647.]

DOI      Magsci      [本文引用: 1]      摘要

<p>积雪是冰冻圈中最重要的组成要素之一, 积雪研究对于气候变化、 水文循环等科学研究和农业灌溉、 减灾防灾等生产活动都具有重要意义.合成孔径雷达(SAR)不仅具有穿云透雾, 全天候观测地表的能力, 而且可穿透地表覆盖一定深度获取地表覆盖物内部特征信息.近年来SAR技术在冰冻圈科学研究中已广泛应用. 综述了SAR积雪监测研究的国内外进展, 对当前主要的SAR积雪遥感模型进行了总结分析, 着重介绍了当前主要的SAR和SAR干涉测量技术(InSAR)积雪面积制图方法、 雪水当量(SWE)反演算法、 积雪密度和雪深提取方法, 并对未来可能的研究方向进行了展望.</p>
[13] Hou Huishu, Yang Hongye.

A general introduction to MODIS snow products and its researching application

[J]. Remote Sensing Technology and Application, 2009, 24(2): 252-256.

Magsci      [本文引用: 1]     

[侯慧姝, 杨宏业.

MODIS 积雪产品及研究应用概述

[J]. 遥感技术与应用, 2009, 24(2): 252-256.]

URL      Magsci      [本文引用: 1]      摘要

<p>MODIS是新一代图谱合一的光谱成像仪,适合进行雪情监测。概要介绍了MODIS积雪产品及NDSI算法在积雪制图方面的应用,也介绍了MODIS积雪产品在国内外研究应用的现状和今后的发展趋势。并且给出了应用MODIS数据制作内蒙古雪盖图的实例。</p>
[14] Li Jinya, Yang Xiuchun, Xu Bin, et al.

Snow monitoring using MODIS and AMSR-E in six main pastoral areas of China

[J]. Scientia Geographica Sinica,2011, 31(9): 1 097-1 104.

[本文引用: 1]     

[李金亚, 杨秀春, 徐斌, . 基于

MODIS 与 AMSR-E 数据的中国 6 大牧区草原积雪遥感监测研究

[J]. 地理科学, 2011, 31(9): 1 097-1 104.]

[本文引用: 1]     

[15] Che Tao, Li Xin.

The development and prospect of estimating snow water equivalent using passive microwave remote sensing data

[J]. Advances in Earth Science, 2004, 19(2): 204-210.

Magsci      [本文引用: 3]     

[车涛, 李新.

被动微波遥感估算雪水当量研究进展与展望

[J]. 地球科学进展, 2004, 19(2): 204-210.]

Magsci      [本文引用: 3]      摘要

<p>被动微波遥感可以透过云层,全天候地提供地表一定深度的信息。星载被动微波遥感传感器的时间分辨率很高,在冰冻圈动态研究中有着重要的地位。在最近的二三十年中,大量被动微波遥感的应用都是在美国、加拿大、欧洲等地,而我国在这方面的研究相对较少。首先介绍了被动微波遥感数据在监测积雪方面的国内外研究进展,对现存的雪水当量(SWE)估算算法(和模型)的适用性进行讨论。然后,详细讨论了我国西部的青藏高原地区雪水当量的估算,阐明了利用SSM/I数据估算青藏高原地区雪水当量的复杂性,并指出了其复杂性产生的原因,提出了解决问题的方法,为该地区积雪动态的进一步研究提供了理论依据。</p>
[16] Li Xin, Che Tao.

A review on passive microwave remote sensing of snow cover

[J] Journal of Glaciology and Geocryology, 2007, 29(3): 487-496.

Magsci      [本文引用: 1]     

[李新, 车涛.

积雪被动微波遥感研究进展

[J]. 冰川冻土, 2007, 29(3):487-496.]

Magsci      [本文引用: 1]      摘要

积雪是冰冻圈中最活跃的要素之一,被动微波遥感具有高时间分辨率且能够迅速覆盖全球,在积雪时空变化监测中作用突出.总结分析了积雪被动微波遥感的主要模型,并对其方法、特点和适用性进行了较详细评述,重点介绍了NASA算法在雪深和雪水当量反演中的应用、反演结果的不确定性以及对它的改进.讨论新兴的积雪数据同化方法,介绍了同化被动微波观测以改进雪深和雪水当量反演精度的研究案例.评述了我国积雪被动微波遥感的进展,并且对未来可能的研究方向做出展望.
[17] Wu Yang, Zhang Jiahua, Xu Haiming, et al.

Advances in study of snow-cover from remote sensing data

[J]. Meteorological Monthly,2007, 33(6): 3-10.

[本文引用: 1]     

[吴杨, 张佳华, 徐海明, .

卫星反演积雪信息的研究进展

[J]. 气象, 2007, 33(6): 3-10.]

DOI      URL      [本文引用: 1]      摘要

综合分析了积雪信息反演的主要遥感信息源和提取方法。在光学遥感方面,应用较广的主要是改进型甚高分辨率扫描辐射仪(AVHRR)资料和中分辨率成像光谱仪(MODIS)资料;提取积雪信息大多是根据积雪在可见光波段的高反射率和近红外波段的低反射率,并通过建立回归模型反演积雪面积和深度。由于传感器的改进,MODIS卫星资料在空间分辨率、积雪反演算法等方面明显优于AVHRR资料。光学仪器受云层和大气的影响很大,由于云和积雪在可见光和近红外波段上都具有高反射率。并且由于云层的遮挡,云下的地表信息不能被光学遥感仪器所接收到。微波遥感方面,被动微波遥感仪如微波辐射计成像仪(SSM/I)、高级微波扫描辐射计(AMS
[18] Zhao Yingshi.Principle and Method of Remote Sensing Application Analysis (Second Edition)[M].Beijing: Science Press, 2013.

[本文引用: 1]     

[赵英时. 遥感应用分析原理与方法 (第二版)[M]. 北京:科学出版社, 2013.]

[本文引用: 1]     

[19] Sun Zhiwen,Shi Jiancheng,Jiang Lingmei, et al.

Development of snow depth and snow water equivalent algorithm in western China using passive microwave remote sensing data

[J]. Advances in Earth Science,2006,21(12):1 363-1 369.

[本文引用: 1]     

[孙之文,施建成,蒋玲梅,.

被动微波遥感反演中国西部地区雪深、雪水当量算法初步研究

[J].地球科学进展,2006,21(12):1 363-1 369.]

DOI      URL      [本文引用: 1]      摘要

雪深、雪水当量是积雪研究中重要参数,其在流域水量平衡和融雪径流预报以及雪灾监测与评价中起着重要作用。Chang等(1987)以辐射传输理论和米氏散射为理论基础,假定积雪密度和颗粒大小为常数,利用实测雪深数据和SMMR的亮温数据,通过统计回归方法,建立了雪深与18GHz和37GHz水平极化的亮温梯度之间的关系,发展了SMMR半经验的反演雪深的算法。后在此基础上又发展了针对SSM/I的半经验反演雪深算法。2002年发射的装载于Aqua卫星上的AM-SR-E是新一代的被动微波辐射计,性能较以往星载被动微波辐射计有较大提高,采用了改进后的SSM/I的半经验算法作为其估算全球雪水当量的反演算法。将AMS
[20] Jin Rui, Li Xin.

A review on the algorithm of frozen/thaw boundary detection by using passive microwave remote sensing

[J]. Remote Sensing Technology and Application, 2002, 17(6):370-375.

Magsci      [本文引用: 1]     

[晋锐, 李新.

被动微波遥感监测土壤冻融界限的研究综述

[J]. 遥感技术与应用, 2002, 17(6): 370-375.]

DOI      URL      Magsci      [本文引用: 1]      摘要

<p>主要介绍了被动微波遥感数据在研究土壤冻融状态方面的国内外研究进展。目前常用的而且比较成熟的冻融界限判别指标是37GHz亮度温度及负亮温谱梯度,重点对这两个冻融边界判别指标的发展进行了详细讨论,并对冻融判别和积雪判别的关系及其相互影响进行了分析和讨论。发展可靠实用的微波遥感土壤冻融状态的判别算法,以提供区域和全球尺度上的土壤状态信息,对水文学、气象学以及农业科学、工程地质研究与应用都有重要意义。</p>
[21] Zhang Tingjun, Jin Rui, Gao Feng.

Overview of the satellite remote sensing of frozen ground: Passive microwave sensors

[J]. AdvanceS in Earth Science, 2009,24(10):1 073-1 083.

[张廷军, 晋锐, 高峰.

冻土遥感研究进展:被动微波遥感

[J]. 地球科学进展, 2009, 24(10): 1 073-1 083.]

DOI      URL      摘要

多年冻土和季节冻土分别占北半球裸露地表的24%和55%。近地表土壤冻融的范围、冻结起始 日期、持续时间及冻融深度对寒季/寒区植物生长、大气与土壤间能量、水分及温室气体交换都具有极其重要的影响。自20世纪70年代以来,应用卫星遥感结合 地面观测资料研究局地到区域尺度的季节冻土和多年冻土已取得诸多成果,而遥感在冻土研究中的最直接应用是利用微波探测近地表土壤冻融状态。相对于主动 SAR,星载被动微波传感器具有多通道观测且重访周期较高,空间分辨率很低的特点。重点评述了近几十年来被动微波辐射计在冻土研究中的算法发展及其应用前 景,主要包括双指标算法、时间序列变化检测算法及判别树算法3类,其核心均是基于冻土的低温特征和“体散射变暗”效应。发展可靠实用的微波遥感土壤冻融状 态判别算法,提供区域和全球尺度上的土壤冻融状态信息,对水文学、气象学以及农业科学、工程地质研究与应用都具有重要意义。
[22] Zhang Tingjun, Jin Rui, Gao Feng.

Overview of the satellite remote sensing of frozen ground: Visile thermal infrared and radar sensor

[J]. Advances in Earth Science, 2009,24(9):963-972.

Magsci      [本文引用: 1]     

[张廷军, 晋锐, 高峰.

冻土遥感研究进展——可见光、红外及主动微波卫星遥感方法

[J]. 地球科学进展, 2009, 24(9): 963-972.]

DOI      URL      Magsci      [本文引用: 1]      摘要

<p>多年冻土和季节冻土分别占北半球裸露地表的24% 和55%。近地表土壤冻融的范围、冻结起始日期、持续时间及冻融深度对寒季/寒区的植被生长、大气与土壤间能量、水分及温室气体交换都具有极其重要的影响。卫星遥感结合地面观测资料研究局地到区域尺度的季节冻土和多年冻土已取得诸多成果。综述了近几十年来卫星遥感技术在冻土研究中的应用。监测多年冻土和地表冻融循环通常需要综合利用可见光、红外、被动微波及主动微波(包括合成孔径雷达SAR和散射计)遥感数据,任何单一波段的传感器都无法满足研究需求。SAR图像能提供空间分辨率较高的寒季/寒区近地表土壤冻融状态的起始日期、持续时间和区域演变等信息,但目前在轨SAR的重访周期相对于春秋季的土壤冻融循环变化过长;星载被动微波传感器具有多通道观测且重访周期较高,但空间分辨率很低的特点;光学和热红外传感器的时空分辨率介于SAR和被动微波遥感之间,但应用于冻土研究时需要具备多年冻土分布和冻融深度与环境因子相关关系的先验信息。总体而言,微波遥感是探测无雪覆盖近地表土壤冻融循环的有效技术手段,而利用热红外传感器反演的地表温度研究土壤冻融过程具有极大潜力。应用卫星遥感反演的积雪范围、雪深、融雪、地表类型、归一化差值植被指数、地表反照率和土壤水分等参数对研究局地、区域乃至全球尺度的冻土都大有裨益。</p>
[23] Armstrong R L, Chang A, Rango A, et al.

Snow depths and grain-size relationships with relevance for passive microwave studies

[J]. Annals of Glaciology, 1993, 17(1): 171-176.

DOI      URL      [本文引用: 1]     

[24] Chang A, Foster J, Hall D.

Nimbus-7 SMMR derived global snow cover parameters

[J]. Annals of Glaciology, 1987, 9(9): 39-44.

DOI      URL      [本文引用: 4]      摘要

Snow covers about 40 million km2 of the land area of the Northern Hemisphere during the winter season. The accumulation and depletion of snow is dynamically coupled with global hydrological and climatological processes. Snow covered area and snow water equivalent are two essential measurements. Snow cover maps are produced routinely by the National Environmental Satellite Data and Information Service of the National Oceanic and Atmospheric Administration (NOAA/NESDIS) and by the US Air Force Global Weather Center (USAFGWC). The snow covered area reported by these two groups sometimes differs by several million km2, Preliminary analysis is performed to evaluate the accuracy of these products. Microwave radiation penetrating through clouds and snowpacks could provide depth and water equivalent information about snow fields. Based on theoretical calculations, snow covered area and snow water equivalent retrieval algorithms have been developed. Snow cover maps for the Northern Hemisphere have been derived from Nimbus-7 SMMR data for a period of six years (1978芒聙聯1984). Intercomparisons of SMMR, NOAA/NESDIS and USAFGWC snow maps have been conducted to evaluate and assess the accuracy of SMMR derived snow maps. The total snow covered area derived from SMMR is usually about 10% less than the other two products. This is because passive microwave sensors cannot detect shallow, dry snow which is less than 5 cm in depth. The major geographic regions in which the differences among these three products are the greatest are in central Asia and western China. Future study is required to determine the absolute accuracy of each product. Preliminary snow water equivalent maps have also been produced. Comparisons are made between retrieved snow water equivalent over large area and available snow depth measurements. The results of the comparisons are good for uniform snow covered areas, such as the Canadian high plains and the Russian steppes. Heavily forested and mountainous areas tend to mask out the microwave snow signatures and thus comparisons with measured water equivalent are poorer in those areas.
[25] Cao Meisheng, Li Xin, Chen Xianzhang, et al.Remote Sensing of Cryosphere[M]. Beijing: Science Press, 2006.

[本文引用: 4]     

[曹梅盛, 李新, 陈贤章, . 冰冻圈遥感[M]. 北京: 科学出版社, 2006.]

[本文引用: 4]     

[26] Hofer R, Mätzler C.

Investigations on snow parameters by radiometry in the 3-to 60-mm wavelength region

[J]. Journal of Geophysical Research: Oceans, 1980, 85(C1): 453-460.

DOI      URL      [本文引用: 1]      摘要

We report on a 2-year period of monitoring parameters of a natural snowpack by ground-based microwave radiometry on a high-altitude Alpine test site. The microwave brightness temperatures are compared to a large set of ground-truth data. Three stages in the seasonal development of the snow cover are easily distinguishable which allow the prediction of the beginning of the snow melting. The moisture content of the melting surface layer is estimated by the aid of the typical daily variations of microwave brightness temperatures in spring. The test site was composed of two snow fields. The first one was lying on slightly reflecting soil, and the second one was lying on a completely reflecting metal foil. By measuring on both fields some microwave snow parameters can be determined. The damping coefficients for microwaves between 5 and 100 GHz were estimated by comparing the results of two extreme theories. Both theories gave results from less than 1 dB/m to more than 30 dB/m depending on the snow state, especially its liquid water content.
[27] Hallikainen M T, Ulaby F T,Van Deventer T E.

Extinction behavior of dry snow in the 18-to 90-GHz range

[J]. IEEE Transactions on Geoscience and Remote Sensing, 1987,(6): 737-745.

DOI      URL      [本文引用: 1]      摘要

The extinction properties of several dry snow types were examined in the 18-to 90-GHz range. The snow types ranged from newly fallen snow to refrozen snow, and the density and mean grain size varied from 0.17 to 0.39 g/cm3 and from 0.2 to 1.6 mm, respectively. From measurements of the transmission loss as a function of sample thickness at a temperature of -1500°C, the extinction coefficient and the surface scattering loss (due to surface roughness at the front and back surfaces of the snow slab) were determined for each snow type. The experimental values were compared against theoretical results computed according to the strong fluctuation theory. In general, good agreement with the experimental data was obtained at 18, 35, and 60 GHz when the grain size used in the theoretical calculations was chosen to be slightly smaller than the observed snow-particle size. However, the extinction coefficient of large-grained refrozen snow as predicted by the strong fluctuation theory is much larger at 90 GHz than the values determined experimentally. The attenuation in snow was observed to increase only slightly with increasing temperature in the -35 to -100°C range.
[28] Chang A, Foster J, Hall D, et al.

Snow parameters derived from microwave measurements during the BOREAS winter field campaign

[J]. Journal of Geophysical Research: Atmospheres, 1997, 102(D24): 29 663-29 671.

DOI      URL      [本文引用: 1]      摘要

Passive microwave data have been used to infer the snow-covered area and snow water equivalent (SWE) over forested areas, but the accuracy of these retrieved snow parameters cannot be easily validated for heterogeneous vegetated regions. The Boreal Ecosystem-Atmosphere Study Winter Field Campaign provided the opportunity to study the effect of boreal forests on snow parameter retrieval in detail. Microwave radiometers (18, 37, and 92 GHz) were flown on board the Canadian National Aeronautical Establishment's Twin Otter. Flight lines covered both the southern study area near Prince Albert and the northern study area near Thompson, Canada. During the 1994 winter campaign, extensive ground-based snow cover information, including depth, density, and grain size, was collected along most of the flight lines, jointly by U.S. and Canadian investigators. Satellite data collected by the special sensor microwave imager are also used for comparison. Preliminary results reconfirmed the relationship between microwave brightness temperature and SWE. However, the effect of forest cover observed by the aircraft sensors is different from that of the satellite observations. This is probably due to the difference in footprint averaging. There were also several flight lines flown over Candle Lake and Waskesiu Lake to assess lake ice signatures. Preliminary results show the thickness of the lake ice may be inferred from the airborne microwave observations. The microwave signature relationship between lake ice and snow matches the results from radiative transfer calculations.
[29] Che T, Dai L, Zheng X, et al.

Estimation of snow depth from passive microwave brightness temperature data in forest regions of northeast China

[J]. Remote Sensing of Environment, 2016, 183:334-349.

DOI      URL      [本文引用: 3]     

[30] Ulaby F T, Moore R K, Fung A K. Microwave Remote Sensing, Active and Passive, Microwave Remote Sensing Fundamentals and Radiometry[M]. US: Addison Wesley Publishing Company, 1981.

[本文引用: 1]     

[31] Chang T, Gloersen P, Schmugge T, et al.

Microwave emission from snow and glacier ice

[J]. Journal of Glaciology, 1976, 16(74): 23-39.

DOI      URL      [本文引用: 1]      摘要

The microwave emission from a model snow field, consisting of randomly spaced ice spheres which scatter independently, is calculated. Mie scattering and radiative transfer theory are applied in a manner similar to that used in calculating microwave and optical properties of clouds. The extinction coefficient is computed as a function of both microwave wavelength and ice-particle radius. Volume scattering by the individual ice particles in the snow field significantly decreases the computed emission for particle radii greater than a few hundredths of the microwave wavelength. Since the mean annual temperature and the accumulation rate of dry polar firn mainly determine the grain sizes upon which the microwave emission depends, these two parameters account for the main features of the 1.55 cm emission observed from Greenland and Antarctica with the Nimbus-5 scanning radiometer. For snow particle sizes normally encountered, most of the calculated radiation emanates from a layer on the order of 10 m in thickness at a wavelength of 2.8 cm, and less at shorter wavelengths. A marked increase in emission from wet versus dry snow is predicted, a result which is consistent with observations. The model results indicate that the characteristic grain sizes in the radiating layers, dry-firn accumulation rales, areas of summer melting, and physical temperatures, can be determined from multispectral microwave observations.
[32] Armstrong R, Brodzik M.

An Earth-gridded SSM/I data set for cryospheric studies and global change monitoring

[J]. Advances in Space Research, 1995, 16(10): 155-163.

DOI      URL      [本文引用: 1]      摘要

The National Snow and Ice Data Center (NSIDC) has distributed DMSP Special Sensor Microwave Imager (SSM/I) brightness temperature grids for the Polar Regions on CD-ROM since 1987. In order to expand this product to include all potential snow covered regions, the area of coverage is now global. The format for the global SSM/I data set is the Equal Area SSM/I Earth Grid (EASE-Grid) developed at NSIDC. The EASE-Grid has been selected as the format for the NASA/NOAA Pathfinder Program Level 3 Products which include both SSM/I and SMMR (Scanning Multichannel Microwave Radiometer) data (1978鈥1987). Providing both data sets in the EASE-Grid will result in a 15 year time-series of satellite passive microwave data in a common format. The extent and variability of seasonal snow cover is recognized to be an important parameter in climate and hydrologic systems and trends in snow cover serve as an indicator of global climatic changes. Passive microwave data from satellites afford the possibility to monitor temporal and spatial variations in snow cover on the global scale, avoiding the problems of cloud cover and darkness. NSIDC is developing the capability to produce daily snow products from the DMSP-SSM/I satellite with a spatial resolution of 25 km. In order to provide a standard environment in which to validate SSM/I algorithm output, it is necessary to assemble baseline data sets using other, more direct, methods of measurement. NSIDC has compiled a validation data set of surface station measurements for the northern hemisphere with specific focus on the United States, Canada, and the former Soviet Union. Digital image substraction is applied to compare the surface station and satellite measurements.
[33] Yang Hu, Li Xiaoqing, You Ran, et al.

Environmental data records from Fengyun-3B microwave radition imager

[J]. Advances in Meteorological Science and Technology, 2013,3(4):138-145.

[本文引用: 2]     

[杨虎, 李小青, 游然, .

风云三号微波成像仪定标精度评价及业务产品介绍

[J]. 气象科技进展, 2013, (4): 138-145.]

DOI      URL      [本文引用: 2]      摘要

风云三号卫星(FY-3)为极轨系列卫星,目前为止分别于2008年5月和2010年10月发射了上午轨道(A)和下午轨道(B)两颗卫星。微波成像仪(MicroWave Radiometer Imager,MWRI)是装载于FY-3上的重要遥感仪器。该仪器为10通道双极化微波成像仪器,中心观测频率设置为10.65,18.7,23.8,36.5和89.0GHz,每个频点有垂直(V)和水平(H)两个探测通道。获取的对地观测亮温数据可用于定量获取大气降水、水汽、海面风速、海温、海冰分布、土壤湿度和陆表温度等地球物理参数信息。目前微波成像仪运行状态稳定,每天获取两次全球覆盖数据。主要介绍微波成像仪定标状况和主要业务产品算法。
[34] Sun Zhiwen.

Estimating Snow Depth and Snow Water Equivalent Algorithm for Fy-3 MWRI and Development of System[D]

. Beijing: Beijing Normal University, 2007.

[本文引用: 1]     

[孙知文.

风云三号微波成像仪(FY-3 MWRI)积雪参数反演算法研究与系统开发[D]

. 北京:北京师范大学, 2007.]

[本文引用: 1]     

[35] Sun Zhiwen, Yu Pengshan, Xia Lang, et al.

Progress in study of snow parameter inversion by passive remote sensing

[J]. Remote Sensing for Land and Resources, 2015,27(1):9-15.

Magsci      [本文引用: 2]     

[孙知文, 于鹏珊, 夏浪, .

被动微波遥感积雪参数反演方法进展

[J]. 国土资源遥感, 2015, 27(1): 9-15.]

DOI      URL      Magsci      [本文引用: 2]      摘要

雪深(snow depth,SD)和雪水当量(snow water equivalent,SWE)是气候水文研究中的重要参数,在雪灾监测中尤为重要。首先,简要介绍了被动微波遥感SD和SWE反演算法的物理基础——积雪微波辐射传输模型,分析了不同微波频段、不同特点的积雪微波辐射和散射特性。然后,根据前人的研究从数学角度将反演算法分为线性亮温梯度法和基于先验知识法,总结了2类算法的优势和局限性: 线性亮温梯度法相对简单、速度快,一般只适用于特定的研究区; 先验知识法需要获取研究区的样本数据,并反复训练才能达到较好的精度,但对样本的独立性及其均值差异显著性的要求较高。最后,重点介绍了我国风云三号微波成像仪(FY-3 MWRI)的全球SD和SWE反演算法和针对中国区域的改进算法,并对未来的研究热点进行了展望。
[36] Wang Gongxue, Jiang Lingmei, Wu Shengli, et al.

Intercalibration FY-3B and FY-3C/MWRI for synergistic implementing to snow depth retrieval algorithm

[J]. Remote Sensing Technology and Application, 2017, 32(1): 49-56.

[本文引用: 1]     

[王功雪, 蒋玲梅, 武胜利, .

FY-3B与FY-3C/MWRI交叉定标及雪深算法应用

[J]. 遥感技术与应用, 2017, 32(1): 49-56.]

[本文引用: 1]     

[37] Jiang Lingmei, Wang Pei, Zhang Lixin, et al.

Improvement of snow depth retrieval for FY3B-MWRI in China

[J]. Science in China (Series D),2014, 44(3):531-547.

[本文引用: 5]     

[蒋玲梅, 王培, 张立新, .

FY3B-MWRI 中国区域雪深反演算法改进

[J]. 中国科学: D辑, 2014, 44(3):531-547.]

[本文引用: 5]     

[38] Dai Liyun.

Study on Passive Microwave Remote Sensing of Snow in northern China[D]

. Beijing:University of Chinese Academy of Sciences, 2013.

[本文引用: 3]     

[戴礼云.

我国北方积雪被动微波遥感反演研究[D]

.北京: 中国科学院大学, 2013.]

[本文引用: 3]     

[39] Che Tao, Li Xin, Gao Feng.

Estimation of snow water equivalent in the Tibetan Plateau using passive microwave remote sensing data (SSM/I)

[J]. Journal of Glaciology and Geocryology, 2004, 26(3): 363-368.

Magsci      [本文引用: 1]     

[车涛, 李新, 高峰.

青藏高原积雪深度和雪水当量的被动微波遥感反演

[J]. 冰川冻土, 2004, 26(3): 363-368.]

DOI      URL      Magsci      [本文引用: 1]      摘要

利用1993年1月份的SSM/I亮度温度数据反演了青藏高原的雪水当量,首先使用被动微波SSM/I数据19和37GHz的水平极化数据来反演雪深,根据积雪时间的函数来计算实时的雪密度,由雪的深度和密度计算出雪水当量.最后,利用SSM/I数据的19和37GHz的垂直极化亮度温度梯度对计算出的雪水当量进行回归分析,得到了利用SSM/I数据直接反演雪水当量的算法.
[40] Takala M, Luojus K, Pulliainen J, et al.

Estimating northern hemisphere snow water equivalent for climate research through assimilation of space-borne radiometer data and ground-based measurements

[J]. Remote Sensing of Environment, 2011, 115(12): 3 517-3 529.

DOI      URL      [本文引用: 3]     

[41] Tedesco M, Narvekar P S.

Assessment of the NASA AMSR-E SWE Product

[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2010, 3(1): 141-159.

DOI      URL      [本文引用: 4]      摘要

Since the launch of the Scanning Multichannel Microwave Radiometer (SMMR) in 1978, several studies have demonstrated the capability of spaceborne passive microwave sensors for mapping global snow water equivalent (SWE). Currently, SWE values are estimated operationally from microwave brightness temperatures measured by the Advanced Microwave Scanning Radiometer-Earth Observing System (AMSR-E) and distributed through the National Snow and Ice Data Center (NSIDC). In this study, we report results regarding the comparison between AMSR-E SWE and SWE/snow depth values distributed by the Snow Data Assimilation System (SNODAS) product of the NOAA's National Operational Hydrologic Remote Sensing Center and snow depth measured by automatic weather stations of the World Meteorological Organization. Generally, we found poor correlation between the AMSR-E and SNODAS SWE/snow depth values. The algorithm performance improves when considering WMO data, though the number of samples used for the analysis might play a role in this sense. We discuss algorithm-related sources of error and uncertainties, such as vegetation and grain size. Moreover, we report results aimed at evaluating whether replacing the linear approach with a nonlinear one and not using the brightness temperatures and ancillary data sets combined as in the current approach but taken separately as inputs to the algorithm might improve the performance of the algorithm.
[42] Sturm M, Taras B, Liston G E, et al.

Estimating snow water equivalent using snow depth data and climate classes

[J]. Journal of Hydrometeorology, 2010, 11(6): 1 380-1 394.

DOI      URL      [本文引用: 1]      摘要

In many practical applications snow depth is known, but snow water equivalent (SWE) is needed as well. Measuring SWE takes 09030409030420 times as long as measuring depth, which in part is why depth measurements outnumber SWE measurements worldwide. Here a method of estimating snow bulk density is presented and then used to convert snow depth to SWE. The method is grounded in the fact that depth varies over a range that is many times greater than that of bulk density. Consequently, estimates derived from measured depths and modeled densities generally fall close to measured values of SWE. Knowledge of snow climate classes is used to improve the accuracy of the estimation procedure. A statistical model based on a Bayesian analysis of a set of 25 688 depth--density--SWE data collected in the United States, Canada, and Switzerland takes snow depth, day of the year, and the climate class of snow at a selected location from which it produces a local bulk density estimate. When converted to SWE and tested against two continental-scale datasets, 90%% of the computed SWE values fell within 00±00±8 cm of the measured values, with most estimates falling much closer.
[43] Mccreight J L, Small E E.

Modeling bulk density and snow water equivalent using daily snow depth observations

[J]. Cryosphere, 2014, 8(2): 5 007-5 049.

[本文引用: 1]     

[44] Martinec J.

Expected snow loads on structures from incomplete hydrological data

[J]. Journal of Glaciology, 1977, 19(81): 185-195.

DOI      URL      [本文引用: 1]      摘要

An assessment of snow loads in Switzerland was required for a revision of the building code. Settling curves of snow are used to compute water equivalents of snow if direct measurements are not available. Based on a frequency analysis, relations between the snow load and the altitude are given for various return periods. Problems of regional effects and of converting the snow-cover data to roof loads are outlined.
[45] Aschbacher J.

Land Surface Studies and Atmospheric Effects by Satellite Microwave Radiometry[D]

. Innsbruck: University of Innsbruck, 1989.

[本文引用: 1]     

[46] Tedesco M, Kelly R, Foster J, et al.

AMSR-E/Aqua Daily L3 Global Snow Water Equivalent EASE-Grids V002

[M]. Boulder, CO:National Snow and Ice Data Center, 2004.

[本文引用: 1]     

[47] Liang J, Liu X, Huang K, et al.

Improved snow depth retrieval by integrating microwave brightness temperature and visible/infrared reflectance

[J]. Remote Sensing of Environment, 2015, 156:500-509.

DOI      URL      [本文引用: 2]     

[48] Xiao X, Zhang T, Zhong X, et al.

Support vector regression snow-depth retrieval algorithm using passive microwave remote sensing data

[J]. Remote Sensing of Environment, 2018, 210:48-64.

DOI      URL      [本文引用: 3]     

[49] Che Tao.

Study on Passive Microwave Remote Sensing of Snow and Snow Data Assimilation Method

[D]. Lanzhou:Cold and Arid Regions Environmental and Engineering Research Institute,CAS, 2006.

[本文引用: 1]     

[车涛.

积雪被动微波遥感反演与积雪数据同化方法研究[D]

. 兰州:中国科学院寒区旱区环境与工程研究所, 2006.]

[本文引用: 1]     

[50] Foster J, Chang A, Hall D.

Comparison of snow mass estimates from a prototype passive microwave snow algorithm, a revised algorithm and a snow depth climatology

[J]. Remote Sensing of Environment, 1997, 62(2): 132-142.

DOI      URL      [本文引用: 4]     

[51] Foster J L, Sun C, Walker J P, et al.

Quantifying the uncertainty in passive microwave snow water equivalent observations

[J]. Remote Sensing of Environment, 2005, 94(2): 187-203.

DOI      URL      [本文引用: 4]      摘要

This study uses a novel approach to quantify these errors by taking into account various factors that impact passive microwave responses from snow in various climatic/geographic regions. Among these factors are vegetation cover (particularly forest cover), snow morphology (crystal size), and errors related to brightness temperature calibration. A time-evolving retrieval algorithm that considers the evolution of snow crystals is formulated. An error model is developed based on the standard error estimation theory. This new algorithm and error estimation method is applied to the passive microwave data from Special Sensor Microwave/Imager (SSM/I) during the 1990鈥1991 snow season to produce annotated error maps for North America. The algorithm has been validated for seven snow seasons (from 1988 to 1995) in taiga, tundra, alpine, prairie, and maritime regions of Canada using in situ SWE data from the Meteorological Service of Canada (MSC) and satellite passive microwave observations. An ongoing study is applying this methodology to passive microwave measurements from Scanning Multichannel Microwave Radiometer (SMMR); future study will further refine and extend the analysis globally, and produce an improved SWE dataset of more than 25 years in length by combining SSMR and SSM/I measurements.
[52] Kelly R.

The AMSR-E snow depth algorithm: Description and initial results

[J]. Journal of The Remote Sensing Society of Japan, 2009, 29(1): 307-317.

DOI      URL      [本文引用: 1]      摘要

Abstract This paper describes the development of the current version of the Advanced Microwave Scanning Radiometer for the Earth Observing System (AMSR-E) algorithm that is used to estimate snow depth for the Japan Aerospace Exploration Agency. The algorithm uses native resolution brightness temperature observations, except for the 89GHz channel which is resampled to the 36GHz footprint, with brightness temperature corrections made on the native measurements rather than using the aggregated brightness temperature observations. A shallow snow detector is developed using the 89GHz channels to detect shallow snow. Furthermore, algorithm retrievals are comprised of the sum of a forested component and a non-forested component with a dynamic estimation of snow depth related to snowpack evolution from selected polarization differences. When compared with up to 254 individual ground station measurements of snow depth, tests show that the new algorithm performs better than previous static parameterized versions both in overall terms and in terms of low to moderate fractional forest cover. For dense forest cover, the algorithm is similar in performance to the previous version. Bias improvements are also very encouraging, but further work is still required to improve the new algorithm's performance in overall error terms and for different fractional landcover mixtures.
[53] Cao Meisheng, Li Peiji, Robinson D, et al.

Evaluation and primary application of microwave remote sensing SMMR derived snow coer in Western China

[J]. Remote Sensing of Environment, 1993, 8(3): 260-269.

[本文引用: 1]     

[曹梅盛, 李培基, Robinson D, .

中国西部积雪 SMMR 微波遥感的评价与初步应用

[J]. 环境遥感, 1993, 8(3): 260-269.]

[本文引用: 1]     

[54] Che T, Xin L, Jin R, et al.

Snow depth derived from passive microwave remote-sensing data in China

[J]. Annals of Glaciology, 2008, 49(1): 145-154.

DOI      URL      [本文引用: 5]      摘要

In this study, we report on the spatial and temporal distribution of seasonal snow depth derived from passive microwave satellite remote-sensing data (e.g. SMMR from 1978 to 1987 and SMM/I from 1987 to 2006) in China. We first modified the Chang algorithm and then validated it using meteorological observation data, considering the influences from vegetation, wet snow, precipitation, cold desert and frozen ground. Furthermore, the modified algorithm is dynamically adjusted based on the seasonal variation of grain size and snow density. Snow-depth distribution is indirectly validated by MODIS snow-cover products by comparing the snow-extent area from this work. The final snow-depth datasets from 1978 to 2006 show that the interannual snow-depth variation is very significant. The spatial and temporal distribution of snow depth is illustrated and discussed, including the steady snow-cover regions in China and snow-mass trend in these regions. Though the areal extent of seasonal snow cover in the Northern Hemisphere indicates a weak decrease over a long period, there is no clear trend in change of snow-cover area extent in China. However, snow mass over the Qinghai-Tibetan Plateau and northwestern China has increased, while it has weakly decreased in northeastern China. Overall, snow depth in China during the past three decades shows significant interannual variation, with a weak increasing trend.
[55] Kelly R E, Chang A T, Tsang L, et al.

A prototype AMSR-E global snow area and snow depth algorithm

[J]. IEEE Transactions on Geoscience and Remote Sensing, 2003, 41(2): 230-242.

DOI      URL      [本文引用: 1]      摘要

A methodologically simple approach to estimate snow depth from spaceborne microwave instruments is described. The scattering signal observed in multifrequency passive microwave data is used to detect snow cover. Wet snow, frozen ground, precipitation, and other anomalous scattering signals are screened using established methods. The results from two different approaches (a simple time and continentwide static approach and a space and time dynamic approach) to estimating snow depth were compared. The static approach, based on radiative transfer calculations, assumes a temporally constant grain size and density. The dynamic approach assumes that snowpack properties are spatially and temporally dynamic and requires two simple empirical models of density and snowpack grain radius evolution, plus a dense media radiative transfer model based on the quasicrystalline approximation and sticky particle theory. To test the approaches, a four-year record of daily snow depth measurements at 71 meteorological stations plus passive microwave data from the Special Sensor Microwave Imager, land cover data and a digital elevation model were used. In addition, testing was performed for a global dataset of over 1000 World Meteorological Organization meteorological stations recording snow depth during the 2000-2001 winter season. When compared with the snow depth data, the new algorithm had an average error of 23 cm for the one-year dataset and 21 cm for the four-year dataset (131% and 94% relative error, respectively). More importantly, the dynamic algorithm tended to underestimate the snow depth less than the static algorithm. This approach will be developed further and implemented for use with the Advanced Microwave Scanning Radiometer-Earth Observing System aboard Aqua.
[56] Biancamaria S, Mognard N M, Boone A, et al.

A satellite snow depth multi-year average derived from SSM/I for the high latitude regions

[J]. Remote Sensing of Environment, 2008, 112(5): 2 557-2 568.

DOI      URL      [本文引用: 1]      摘要

The hydrological cycle for high latitude regions is inherently linked with the seasonal snowpack. Thus, accurately monitoring the snow depth and the associated aerial coverage are critical issues for monitoring the global climate system. Passive microwave satellite measurements provide an optimal means to monitor the snowpack over the arctic region. While the temporal evolution of snow extent can be observed globally from microwave radiometers, the determination of the corresponding snow depth is more difficult. A dynamic algorithm that accounts for the dependence of the microwave scattering on the snow grain size has been developed to estimate snow depth from Special Sensor Microwave/Imager (SSM/I) brightness temperatures and was validated over the U.S. Great Plains and Western Siberia. The purpose of this study is to assess the dynamic algorithm performance over the entire high latitude (land) region by computing a snow depth multi-year field for the time period 1987鈥1995. This multi-year average is compared to the Global Soil Wetness Project-Phase2 (GSWP2) snow depth computed from several state-of-the-art land surface schemes and averaged over the same time period. The multi-year average obtained by the dynamic algorithm is in good agreement with the GSWP2 snow depth field (the correlation coefficient for January is 0.55). The static algorithm, which assumes a constant snow grain size in space and time does not correlate with the GSWP2 snow depth field (the correlation coefficient with GSWP2 data for January is 鈭0.03), but exhibits a very high anti-correlation with the NCEP average January air temperature field (correlation coefficient 鈭0.77), the deepest satellite snow pack being located in the coldest regions, where the snow grain size may be significantly larger than the average value used in the static algorithm. The dynamic algorithm performs better over Eurasia (with a correlation coefficient with GSWP2 snow depth equal to 0.65) than over North America (where the correlation coefficient decreases to 0.29).
[57] Grippa M, Mognard N, Le Toan T, et al.

Siberia snow depth climatology derived from SSM/I data using a combined dynamic and static algorithm

[J]. Remote Sensing of Environment, 2004, 93(1): 30-41.

DOI      URL      [本文引用: 1]     

[58] Pulliainen J T, Grandell J, Hallikainen M T.

HUT snow emission model and its applicability to snow water equivalent retrieval

[J]. IEEE Transactions on Geoscience and Remote Sensing, 1999, 37(3): 1 378-1 390.

DOI      URL      [本文引用: 1]     

[59] Roy V, Goita K, Royer R, et al.

Snow water equivalent retrieval in a Canadian boreal environment from microwave measurements using the HUT snow emission model

[J]. IEEE Transactions on Geoscience & Remote Sensing, 2004, 42(9): 1 850-1 859.

[本文引用: 1]     

[60] Tsang L, Chen C T, Chang A T, et al.

Dense media radiative transfer theory based on quasicrystalline approximation with applications to passive microwave remote sensing of snow

[J]. Radio Science, 2000, 35(3): 731-49.

DOI      URL      [本文引用: 1]     

[61] Chen C T, Nijssen B, Guo J, et al.

Passive microwave remote sensing of snow constrained by hydrological simulations

[J]. IEEE Transactions on Geoscience & Remote Sensing, 2001, 39(8): 1 744-1 756.

DOI      URL      [本文引用: 1]      摘要

This paper describes a snow parameter retrieval algorithm from passive microwave remote sensing measurements. The three components of the retrieval algorithm include a dense media radiative transfer (DMRT) model, which is based on the quasicrystalline approximation (QCA) with the sticky particle assumption, a physically-based snow hydrology model (SHM) that incorporates meteorological and topographical data, and a neural network (NN) for computational efficient inversions. The DMRT model relates physical snow parameters to brightness temperatures. The SHM simulates the mass and heat balance and provides initial guesses for the neural network. The NN is used to speed up the inversion of parameters. The retrieval algorithm can provide speedy parameter retrievals for desired temporal and spatial resolutions, Four channels of brightness temperature measurements: 19V, 19H, 37V, and 37H are used. The algorithm was applied to stations in the northern hemisphere. Two sets of results are shown. For these cases, the authors use ground-truth precipitation data, and estimates of snow water equivalent (SWE) from SHM give good results. For the second set, a weather forecast model is used to provide precipitation inputs for SHM. Additional constraints in grain size and density are used. They show that inversion results compare favorably with ground truth observations
[62] Wiesmann A, Mätzler C.

Microwave emission model of layered snowpacks

[J]. Remote Sensing of Environment, 1999, 70(3): 307-316.

DOI      URL      [本文引用: 1]      摘要

A thermal microwave emission model of layered snowpacks (MEMLS) was developed for the frequency range 5鈥100 GHz. It is based on radiative transfer, using six-flux theory to describe multiple volume scattering and absorption, including radiation trapping due to total reflection and a combination of coherent and incoherent superpositions of reflections between layer interfaces. The scattering coefficient is determined empirically from measured snow samples, whereas the absorption coefficient, the effective permittivity, refraction, and reflection at layer interfaces are based on physical models and on measured ice dielectric properties. The number of layers is only limited by computer time and memory. A limitation of the empirical fits and thus of MEMLS is in the range of observed frequencies and correlation lengths (a measure of grain size). First model validation for dry winter snow was successful. An extension to larger grains is given in a companion article (M盲tzler and Wiesmann, 1999) . The objective of the present article is to describe and illustrate the model and to pave the way for further improvements. MEMLS has been coded in MATLAB. It forms part of a combined land-surface-atmosphere microwave emission model for radiometry from satellites (Pulliainen et al., 1998) .
[63] Gharaei-Manesh S, Fathzadeh A, Taghizadeh-Mehrjardi R.

Comparison of artificial neural network and decision tree models in estimating spatial distribution of snow depth in a semi-arid region of Iran

[J]. Cold Regions Science & Technology, 2016, 122:26-35.

[本文引用: 3]     

[64] Davis D T, Chen Z, Tsang L, et al.

Retrieval of snow parameters by iterative inversion of a neural network

[J]. IEEE Transactions on Geoscience and Remote Sensing, 1993, 31(4): 842-852.

DOI      URL      [本文引用: 2]      摘要

The inversion of snow parameters from passive microwave remote sensing measurements is performed, using an iterative inversion of a neural network (NN) trained with a dense-media multiple-scattering model. Inversion of four parameters is performed based on five brightness temperatures. The four parameters are mean grain size of ice particles in snow, snow density, snow temperature, and snow depth. Iterative inversion of a data-driven forward NN model is justified on a theoretical and methodological basis. An error analysis is performed, comparing iterative inversion of a forward model with the use of an explicit inverse for the retrieval of independent snow parameters from their corresponding measurements. The NN iterative inversion algorithm is further illustrated by reconstructing a synthetic terrain of snow parameters from their corresponding measurements, inverting all four parameters simultaneously. The reconstructed parameter contours are in good agreement with the original synthetic parameter contours
[65] Tedesco M, Pulliainen J, Takala M, et al.

Artificial neural network-based techniques for the retrieval of SWE and snow depth from SSM/I data

[J]. Remote Sensing of Environment, 2004, 90(1): 76-85.

DOI      URL      [本文引用: 1]     

[66] Tabari H, Marofi S, Abyaneh H Z, et al.

Comparison of artificial neural network and combined models in estimating spatial distribution of snow depth and snow water equivalent in Samsami Basin of Iran

[J]. Neural Computing & Applications, 2010, 19(4): 625-635.

[本文引用: 1]     

[67] Forman B A, Reichle R H, Derksen C.

Estimating passive microwave brightness temperature over snow-covered land in North America using a land surface model and an artificial neural network

[J]. IEEE Transactions on Geoscience & Remote Sensing, 2013, 52(1): 235-248.

[本文引用: 1]     

[68] Dai L, Che T, Wang J, et al.

Snow depth and snow water equivalent estimation from AMSR-E data based on a priori snow characteristics in Xinjiang, China

[J]. Remote Sensing of Environment, 2012, 127:14-29.

DOI      URL      [本文引用: 3]     

[69] Zhang Xianfeng, Bao Huiyi, Liu Yu, et al.

Snow parameter estimation from microwave remote sensing data

[J]. Mountain Research, 2014, 32(3): 307-313.

[本文引用: 2]     

[张显峰, 包慧漪, 刘羽, .

基于微波遥感数据的雪情参数反演方法

[J]. 山地学报, 2014, 32(3): 307-313.]

DOI      URL      [本文引用: 2]      摘要

微波遥感传感器在36.5 GHz通道会因雪深超过其穿透深度而出现信号饱和,从而导致雪深被低估.针对该问题,首先建立了18.7 GHz与36.5 GHz通道亮温差和10.7 GHz与18.7 GHz通道亮温差相结合的积雪深度分层反演新方法,然后利用GCOM-W1星上搭载的AMSR2传感器数据估算了2012年12月至2013年2月新疆每 日积雪深度,结合同期的气象站点观测数据与野外实测数据对遥感反演结果进行了评价.结果表明,所建立模型能够很好识别新疆地区积雪的空间分布状况,雪深的 估算结果明显优于常用的Chang模型.
[70] Etemad-Shahidi A, Mahjoobi J.

Comparison between M5' model tree and neural networks for prediction of significant wave height in Lake Superior

[J]. Ocean Engineering, 2009, 36(15/16): 1 175-1 181.

DOI      URL      [本文引用: 2]     

[71] Forman B A, Reichle R H.

Using a support vector machine and a land surface model to estimate large-scale passive microwave brightness temperatures over snow-covered land in North America

[J]. IEEE Journal of Selected Topics in Applied Earth Observations & Remote Sensing, 2015, 8(9): 4 431-4 441.

[本文引用: 1]     

[72] Xue Y, Forman B A.

Comparison of passive microwave brightness temperature prediction sensitivities over snow-covered land in North America using machine learning algorithms and the Advanced Microwave Scanning Radiometer

[J]. Remote Sensing of Environment, 2015, 170:153-165.

DOI      URL      [本文引用: 2]     

[73] Balk B, Elder K.

Combining binary decision tree and geostatistical methods to estimate snow distribution in a mountain watershed

[J]. Water Resources Research, 2000, 36(1): 13-26.

DOI      URL      [本文引用: 1]     

[74] Wu Lili, Li Xiaofeng, Chen Yueqing, et al.

The improvement of HUT model and its application in snow depth inversion

[J]. Geomatics and Information Science of Wuhan University, 2017, 42(7): 904-910.

[本文引用: 1]     

[武黎黎, 李晓峰, 陈月庆, .

HUT模型的改进及其雪深反演

[J]. 武汉大学学报:信息科学版, 2017, 42(7): 904-910.]

DOI      URL      [本文引用: 1]      摘要

基于风云-3B(FY-3B)卫星的微波成像仪(MWRI)数据对HUT模型(Helsinki university of technology snow emission model)进行验证,结果表明,无论是18.7GHz还是36.5GHz水平极化亮温,HUT模型模拟亮温都与MWRI亮温存在较大的偏差。因此,本文对消光系数进行了本地化改进,得到了改进的HUT模型(IMPHUT模型)。IMPHUT模型在18.7GHz水平极化和36.5GHz水平极化时的模拟亮温偏差分别为-0.91K和-4.19K,较原始的HUT模型模拟精度(偏差分别为14.03K和-16.33K)有很大提高。最后,利用遗传算法进行雪深反演,基于IMPHUT模型的雪深反演(偏差为-6.79cm)优于HUT模型和Chang算法,反演与实测雪深具有较好的一致性。
[75] Li Xiaolan, Zhang Feimin, Wang Chenghai.

Comparison and analysis of snow depth over China, observed and derived from remote sensing

[J]. Journal of Glaciology and Geocryology, 2012, 34(4):755-764.

Magsci      [本文引用: 3]     

[李小兰, 张飞民, 王澄海.

中国地区地面观测积雪深度和遥感雪深资料的对比分析

[J]. 冰川冻土, 2012, 34(4): 755-764.]

URL      Magsci      [本文引用: 3]      摘要

比较了气象台站观测和卫星遥感(SMMR、 SSM/I、 AMSR-E)的积雪深度两种资料在空间分布、 年际变化及其与中国夏季降水之间关系的异同性.结果表明: 两种资料在积雪稳定区的分布比较一致, 积雪深度的大值区位于东北地区、 新疆北部和青藏高原地区; 对于季节性积雪区且积雪深度不大的区域而言, 二者之间存在着较大的差异, 尤其在江淮流域及长江中下游地区, 台站观测的积雪深度大于遥感得到的积雪深度; 平均而言, 两种资料获得的积雪深度在各地区基本一致.在新疆北部和高原南部, 二种资料的年际变化存在着差异, 在新疆北部, 台站观测大于遥感得到的积雪深度, 而在高原东南部遥感大于台站观测积雪.近30 a来, 两种资料获得的积雪深度在新疆北部和青藏高原的年际变化趋势基本一致, 新疆北部为增加趋势, 青藏高原有减少的趋势.值得注意的是, 在东北地区, 近30 a来两种类型资料的年际变化趋势呈相反变化.两种资料在新疆北部的相关最强; 东北、 青藏高原其次; 而高原东南部最差, 在使用时应加注意.青藏高原地区的两种积雪资料与中国夏季降水的相关"信号"基本一致.青藏高原地区积雪与东北西部地区和长江中下游夏季降水之间的相关最为显著.资料间的差异性并不影响高原地区积雪对中国夏季降水"信号"的应用.
[76] Zhao Yingshi.Principle and Method of Remote Sensing Application Analysis[M]. Beijing:Science Press,2003.

[本文引用: 1]     

[赵英时. 遥感应用分析原理与方法[M]. 北京:科学出版社, 2003.]

[本文引用: 1]     

[77] Reichle R H.

Data assimilation methods in the Earth sciences

[J]. Advances in Water Resources, 2008, 31(11): 1 411-1 418.

DOI      URL      [本文引用: 1]      摘要

Although remote sensing data are often plentiful, they do not usually satisfy the users鈥 needs directly. Data assimilation is required to extract information about geophysical fields of interest from the remote sensing observations and to make the data more accessible to users. Remote sensing may provide, for example, measurements of surface soil moisture, snow water equivalent, snow cover, or land surface (skin) temperature. Data assimilation can then be used to estimate variables that are not directly observed from space but are needed for applications, for instance root zone soil moisture or land surface fluxes. The paper provides a brief introduction to modern data assimilation methods in the Earth sciences, their applications, and pertinent research questions. Our general overview is readily accessible to hydrologic remote sensing scientists. Within the general context of Earth science data assimilation, we point to examples of the assimilation of remotely sensed observations in land surface hydrology.
[78] Liston G E, Pielke R A, Greene E M.

Improving first-order snow-related deficiencies in a regional climate model

[J]. Journal of Geophysical Research Atmospheres, 1999, 104(D16): 19 559-19 567.

DOI      URL      [本文引用: 1]      摘要

A climate version of the Regional Atmospheric Modeling System (RAMS) is used to simulate snow-related land-atmosphere interactions in the Great Plains and Rocky Mountain regions of the United States. The availability of observed snow-distribution products allow snow-water-equivalent distribution data to be assimilated directly into the RAMS simulations. By performing two kinds of model integrations, one with and one without assimilating the snow-distribution observations, the differences between the model runs are used to highlight model deficiencies and limitations and thus identify areas of possible improvement in the atmospheric model. The need to simulate subgrid snow distributions is identified and addressed by implementing a snow submodel that accounts for subgrid variations in air temperature and precipitation. This subgrid snow model is found to significantly improve the model's simulation of snow-related processes.
[79] Brasnett B.

A global analysis of snow depth for numerical weather prediction

[J]. Journal of Applied Meteorology, 1999, 38(6): 726-740.

DOI      URL      [本文引用: 2]     

[80] Sheffield J,Pan M,Wood E F, et al.

Snow process modeling in the North American Land Data Assimilation System (NLDAS):1. Evaluation of model-simulated snow cover extent

[J]. Journal of Geophysical Research Atmospheres,2003,108(D22):2 101-2 110.

[本文引用: 1]     

[81] Zhao Liang, Zhu Yuxiang, Yang Hong, et al.

A dynamic approach to retrieving snow depth based on the technology of integrating satellite remote sensing and in situ data

[J]. Acta Meterologica Sinica, 2013, 71(4): 769-782.

Magsci      [本文引用: 1]     

[赵亮, 朱玉祥, 杨弘, .

一种基于卫星遥感与地面测站数据融合技术的雪深动态反演方法

[J]. 气象学报, 2013, 71(4): 769-782.]

DOI      Magsci      [本文引用: 1]      摘要

提出了一种新的基于被动微波遥感和地面测站数据融合技术的雪深动态反演方法.这种新方法不再依赖单一的地面测站数据或卫星遥感数据,而是利用它们联合建立雪盖可信度指数,共同确定雪盖分布;然后在此基础上采用时空距离权重法设定反演系数动态参数化方案,反演雪深.这种雪深反演方法具有以下特点:针对不同时空条件下反演系数的动态差异问题,提出利用实时测站观测雪深,灵活调整雪深反演系数的解决方案,使反演系数具备随时空动态调整的能力,这是与静态反演方法最大的区别;充分利用了被动微波遥感数据时空连续性好的优势,能够在测站稀少的西部高山地区反演出空间分辨率相对较高的雪深数据,这是地面观测无法做到的.初步检验结果显示,该方法较明显地提高了中国西部高原地区和东部雪盖南缘区的反演精度,并克服了原有融合方法在中国西部雪盖面积偏小的问题,有效避免了静态反演方法在高山地区严重高估而平原地区低估雪深的问题,实现了被动微波遥感和地面观测数据的有效融合,扩大了雪深监测的有效范围
[82] Andreadis K M, Lettenmaier D P.

Assimilating remotely sensed snow observations into a macroscale hydrology model

[J]. Advances in Water Resources, 2006, 29(6): 872-886.

DOI      URL      [本文引用: 1]      摘要

Accurate forecasting of snow properties is important for effective water resources management, especially in mountainous areas like the western United States. Current model-based forecasting approaches are limited by model biases and input data uncertainties. Remote sensing offers an opportunity for observation of snow properties, like areal extent and water equivalent, over larger areas. Data assimilation provides a framework for optimally merging information from remotely sensed observations and hydrologic model predictions. An ensemble Kalman filter (EnKF) was used to assimilate remotely sensed snow observations into the variable infiltration capacity (VIC) macroscale hydrologic model over the Snake River basin. The snow cover extent (SCE) product from the moderate resolution imaging spectroradiometer (MODIS) flown on the NASA Terra satellite was used to update VIC snow water equivalent (SWE), for a period of four consecutive winters (1999鈥2003). A simple snow depletion curve model was used for the necessary SWE鈥揝CE inversion. The results showed that the EnKF is an effective and operationally feasible solution; the filter successfully updated model SCE predictions to better agree with the MODIS observations and ground surface measurements. Comparisons of the VIC SWE estimates following updating with surface SWE observations (from the NRCS SNOTEL network) indicated that the filter performance was a modest improvement over the open-loop (un-updated) simulations. This improvement was more evident for lower to middle elevations, and during snowmelt, while during accumulation the filter and open-loop estimates were very close on average. Subsequently, a preliminary assessment of the potential for assimilating the SWE product from the advanced microwave scanning radiometer (AMSR-E, flown on board the NASA Aqua satellite) was conducted. The results were not encouraging, and appeared to reflect large errors in the AMSR-E SWE product, which were also apparent in comparisons with SNOTEL data.
[83] Durand M, Margulis S A.

Feasibility test of multifrequency radiometric data assimilation to estimate snow water equivalent

[J]. Journal of Hydrometeorology, 2006, 7(3): 443-457.

DOI      URL      [本文引用: 1]     

[84] Che Tao, Li Xin, Jin Rui, et al.

Assimilating passive microwave remote sensing data into a land surface model to improve the estimation of snow depth

[J]. Remote Sensing of Environment, 2014, 143(54/63): 54-63.

DOI      URL      [本文引用: 1]     

[85] Pulliainen J.

Mapping of snow water equivalent and snow depth in boreal and sub-arctic zones by assimilating space-borne microwave radiometer data and ground-based observations

[J]. Remote Sensing of Environment, 2006, 101(2): 257-269.

DOI      URL      [本文引用: 2]     

[86] Jiang Lingmei.

Passive Microwave Remote Sensing of Snow Water Equivalence Study[D]

.Beijing:Beijing Normal University, 2005.

[本文引用: 2]     

[蒋玲梅.

被动微波雪水当量研究[D]

.北京:北京师范大学, 2005.]

[本文引用: 2]     

[87] Dai L, Che T, Ding Y.

Inter-calibrating SMMR, SSM/I and SSMI/S data to improve the consistency of snow-depth products in China

[J]. Remote Sensing, 2015, 7(6): 7 212-7 230.

DOI      URL      [本文引用: 1]     

[88] Armstrong R L, Brodzik M J.

Recent northern hemisphere snow extent: A comparison of data derived from visible and microwave satellite sensors

[J]. Geophysical Research Letters, 2001, 28(19): 3 673-3 676.

DOI      URL      [本文引用: 1]     

[89] Cao Meisheng, Li Peiji.

Microwave remote sensing monitoring of snow cover in western China

[J]. Mountain Reseach, 1994, 12(4): 230-234.

[本文引用: 1]     

[曹梅盛, 李培基.

中国西部积雪微波遥感监测

[J]. 山地研究, 1994, 12(4): 230-234.]

URL      [本文引用: 1]      摘要

用1978─1987年多通过微波扫描辐射计(SMMR)所获取的地表微波亮温及亮温-雪深区域订正反演算式,计算了100°E以西中国境内年与季的平均雪量和雪盖率,以及它们的年际变化,阐明了积雪时空的变化。所取得的高原及高山低山积雪监测结果,为当地积雪资源的开发利用提供了可靠依据。
[90] Zheng Lei, Zhang Tingjun, Che Tao, et al.

Evaluation of snow depth products derived from passive microwave satellite remote sensing data using ground—Based snow measurements

[J]. Remote Sensing Technology and Application, 2015, 30(3):413-423.

Magsci      [本文引用: 2]     

[郑雷, 张廷军, 车涛, .

利用实测资料评估被动微波遥感雪深算法

[J]. 遥感技术与应用, 2015, 30(3): 413-423.]

DOI      URL      Magsci      [本文引用: 2]      摘要

<p>利用SSM/I微波亮温数据,结合地面站点实测资料,比较Chang算法和Che算法在前苏联、中国及蒙古境内6种不同积雪类型的反演精度,结果表明:被广泛应用于全球雪深反演的Chang算法低估了前苏联境内雪深7.6 cm,相对误差为-24.3%,而分别高估中国及蒙古境内雪深9.2 cm与11.4 cm,相对误差分别为108.8%和180.9%,区域反演效果很差;针对中国境内积雪的Che算法严重低估前苏联境内雪深,整体低估21.3 cm,相对误差为-68.6%,RMSE为31.4 cm;在中国及蒙古境内反演效果有所改善。6个积雪类型中,植被较单一,地形较平坦的苔原型积雪和草原型积雪雪深的反演效果较好。 随着纬度和积雪深度的增加被动微波雪深反演有由高估变为低估的趋势。Che算法反演的雪深大体以40&deg;N为界,以北表现为低估,以南表现为高估,另一方面,整体上该算法在雪深低于6.7 cm时表现为低高估,高于6.7 cm表现为低估;因此,全球算法应用到局部地区需要进行修正,不同下垫面性质以和气候条件下形成的积雪的被动微波反演应区别对待。</p>
[91] Liu J, Li Z, Huang L, et al.

Hemispheric-scale comparison of monthly passive microwave snow water equivalent products

[J]. Journal of Applied Remote Sensing, 2014, 8(1). DOI:10.1117/1.JRS.8.084688.

URL      [本文引用: 1]      摘要

The snow water equivalent (SWE) products from passive microwave remote sensing are useful in global climate change studies due to the long-time and all-weather imaging capabilities of passive microwave radiometry at the hemisphere scale. Northern Hemisphere SWE products, including products from the National Snow and Ice Data Center (NSIDC) and GlobSnow from the European Space Agency (ESA), have been providing long-time series information since 1979. However, the different algorithms used to produce the NSIDC and GlobSnow products lead to discrepancies in the data. To determine which product might be superior, this paper assesses their hemisphere-scale quality for the time period 1979-2010. By comparing the data with historical snow depth measurements obtained from 7388 meteorological stations in the Northern Hemisphere, the accuracies of the different SWE products are analyzed for the period and for different snow types. The results show that for SWEs above 30 mm but below 200 mm, GlobSnow estimates maintain a better linear relation with the ground measurements. NSIDC products are more influenced by microwave "saturation," producing obvious underestimations for SWEs over 120 mm. However, for shallow snow (SWE less than 30 mm), the slight overestimate produced by GlobSnow is more obvious than that of the other NSIDC products.
[92] Frei A, Tedesco M, Lee S, et al.

A review of global satellite-derived snow products

[J]. Advances in Space Research, 2012, 50(8): 1 007-1 029.

DOI      URL      [本文引用: 1]      摘要

Snow cover over the Northern Hemisphere plays a crucial role in the Earth鈥檚 hydrology and surface energy balance, and modulates feedbacks that control variations of global climate. While many of these variations are associated with exchanges of energy and mass between the land surface and the atmosphere, other expected changes are likely to propagate downstream and affect oceanic processes in coastal zones. For example, a large component of the freshwater flux into the Arctic Ocean comes from snow melt. The timing and magnitude of this flux affects biological and thermodynamic processes in the Arctic Ocean, and potentially across the globe through their impact on North Atlantic Deep Water formation. Several recent global remotely sensed products provide information at unprecedented temporal, spatial, and spectral resolutions. In this article we review the theoretical underpinnings and characteristics of three key products. We also demonstrate the seasonal and spatial patterns of agreement and disagreement amongst them, and discuss current and future directions in their application and development. Though there is general agreement amongst these products, there can be disagreement over certain geographic regions and under conditions of ephemeral, patchy and melting snow.
[93] Singh P R, Gan T Y.

Retrieval of snow water equivalent using passive microwave brightness temperature data

[J]. Remote Sensing of Environment, 2000, 74(2): 275-286.

DOI      URL      [本文引用: 1]      摘要

Existing algorithms for retrieving snow water equivalent (SWE) from the Special Sensor Microwave/Imager (SSM/I) passive microwave brightness temperature data were assessed and new algorithms that include physiographic and atmospheric data were developed for the Red River basin of North Dakota and Minnesota. The frequencies of SSM/I data used are 19 GHz and 37 GHz in both horizontal and vertical polarization. Encouraging calibration results are obtained for the algorithms using multivariate regression technique and dry snow cases of the 1989 and 1988 SSM/I data (from DMSP-F8). Similarly, validation results for data not used in calibration [e.g., 1988 (1989 as calibration data), 1989 (1988 as calibration data), and 1997 (from DMSP-F10 and F13)] are also encouraging. The nonparametric, Projection Pursuit Regression (PPR) technique also gave good results in both stages. However, for the validation stage, adding a shift parameter to all retrieval algorithms was always necessary, possibly because of different scatter-induced darkening (caused by scattering albedo), which could arise even for snowpacks of the same thickness because snowpacks undergo different metamorphism in different winter years. Screening criteria are also proposed to eliminate SSM/I footprints affected by large water bodies and depth-hoar鈥攁nother key step toward reliable SWE estimation from passive microwave data.
[94] Hall D K.

Influence of depth hoar on microwave emission from snow in northern Alaska

[J]. Cold Regions Science and Technology, 1987, 13(3): 225-231.

DOI      URL      [本文引用: 1]      摘要

More than 80% of the total land area of Eurasia and North America can be covered by snow during the winter. Snow is highly reflective and the amount of time that it remains on the ground is important for regional and global energy balance since 90% of the solar energy available to heat the earth can be reflected away by snow. In previous work, global snow depth maps have been prepared based on Scanning Multichannel Microwave Radiometer (SMMR) passive microwave satellite data. A two-layer radiative transfer model is used to generate an algorithm by which global snow depth is calculated from the SMMR brightness temperatures (T B). Analysis of these SMMR snow maps has shown that the snowpack of the Arctic Coastal Plain of Alaska displays values of microwave T B that are lower than would be expected for the shallow snow which is known to occur there. In addition, observed values of T B in northern Alaska decrease as the winter progresses. The snowpack has a structure that consists of a low density depth hoar (lower) layer and an upper layer which consists of dense, windpacked snow. In this paper a two layer radiative transfer model is employed to calculate microwave T B for model snowpacks that have a depth hoar layer, in order to simulate snowpack conditions in northern Alaska. The observational data consist of a time series of 37 GHz horizontally and vertically polarized SMMR data of the Arctic Coastal Plain of Alaska during the period from January through March 1980 and snow depth and air temperature measurements from Umiat, Alaska from the same time period. In the simulations, the thickness of the depth hoar layer was 10 cm in early January and increased by 0.50 cm per week, throughout a 3 month study period, to simulate a reported increasing thickness of the depth hoar layer as the winter progresses. When simulated and observed T B were correlated for 15 data points, a coefficient of correlation of R = 0.84 was obtained. Results show that the presence and variability of the depth hoar layer in northern Alaska can have a significant effect on the microwave emission from a snowpack.
[95] Wei Yue, Chen Shujiang, Chen Xia.

Analysis on the seasonal snow density change characteristics of north Xinjiang

[J]. Journal of Glaciology and Geocryology, 2010, 32(3): 519-523.

Magsci      [本文引用: 1]     

[魏玥, 陈蜀江, 陈霞.

新疆北部地区季节性积雪密度变化特征分析

[J]. 冰川冻土, 2010, 32(3): 519-523.]

DOI      URL      Magsci      [本文引用: 1]      摘要

<FONT face=Verdana>选取新疆北部地区季节性积雪期的定点站和典型区域,应用北疆20个气象站点观测资料和使用便携式测雪仪(snow fork),在不同地域、不同雪层和不同时间进行观测与测量,并且在积雪稳定期中的一次降雪过程对新雪密度变化过程中影响它的诸多因子进行观测,对新疆北部地区冬季季节性积雪密度变化特征进行的观测和分析. 结果表明:雪面辐射热量和雪层内温度梯度对积雪密度起主要作用,变化主要是通过雪层内深霜和粗粒雪层的温度减小而实现的;在隆冬期全层积雪密度最大的为深霜层,入春2月下旬回暖期以后,由于雪层含水率的增加,季节性积雪密度最大层则为粒雪层. </FONT>
[96] Li Peiji.

Dynamic characteristic of snow cover in Western China

[J]. Acta Geographic Sinica, 1993, 48(6): 505-515.

[本文引用: 1]     

[李培基.

中国西部积雪变化特征

[J]. 地理学报, 1993, 48(6): 505-515.]

DOI      URL      [本文引用: 1]     

[97] Foster J L, Hall D K, Chang A T, et al.

Effects of snow crystal shape on the scattering of passive microwave radiation

[J]. IEEE Transactions on Geoscience and Remote Sensing, 1999, 37(2): 1 165-1 168.

DOI      URL      [本文引用: 1]      摘要

A discrete dipole scattering model is used to measure the passive microwave radiation scattered by snow particles having different shapes and sizes. The model results demonstrate that the shape of the snow crystal is insignificant in scattering microwave energy in the 37-GHz region of the spectrum
[98] Vander Jagt B J, Durand M T, Margulis S A, et al.

The effect of spatial variability on the sensitivity of passive microwave measurements to snow water equivalent

[J]. Remote Sensing of Environment, 2013, 136:163-179.

DOI      URL      [本文引用: 1]     

[99] Goïta K, Walker A E, Goodison B E.

Algorithm development for the estimation of snow water equivalent in the boreal forest using passive microwave data

[J]. International Journal of Remote Sensing, 2003, 24(5): 1 097-1 102.

DOI      URL      [本文引用: 1]     

[100] Romanov P, Tarpley D.

Enhanced algorithm for estimating snow depth from geostationary satellites

[J]. Remote Sensing of Environment, 2007, 108(1): 97-110.

DOI      URL      [本文引用: 1]      摘要

Observations in the visible and infrared spectral bands from the Imager instrument onboard Geostationary Operational Environmental Satellite (GOES) have been used to derive snow depth. The technique makes use of correlation between depth of the snow pack and satellite-derived subpixel fractional snow cover. Previous efforts to infer snow depth from satellite data with this technique were focused on grasslands and croplands, where the snow depth/snow fraction relationship is most pronounced. In this paper we improve the retrieval algorithm to extend snow depth estimates to forested areas. The enhanced algorithm accounts for the tree cover fraction and for the type of forest, deciduous or coniferous. The developed technique was used to derive maps of snow depth over mid-latitude areas of North America during winter seasons of 2003–2004 and 2004–2005. Satellite-based snow depth maps were produced daily at 402km spatial resolution. To validate the retrievals we compared them with surface observations of snow depth and with the snow depth analysis prepared at the NOAA National Operational Hydrological Remote Sensing Center (NOHRSC). The estimated retrieval error was about 30% for snow depths below 3002cm and increased to 50% for snow depths ranging from 30 to 5002cm. Snow depth retrievals were limited to scenes with less than 80% deciduous forest cover fraction and less than 50% needle leaf forest cover.
[101] Zhou Shengnan, Che Tao, Dai Liyun.

Based on the type of ground site representative of snow remote sensing products precision evaluation

[J]. Remote Sensing Technology and Application, 2017, 32(2): 228-237.

[本文引用: 1]     

[周胜男, 车涛, 戴礼云.

基于地面站点类型代表性的积雪遥感产品精度评价

[J]. 遥感技术与应用, 2017, 32(2): 228-237.]

[本文引用: 1]     

[102] Grody N C, Basist A N.

Global identification of snowcover using SSM/I measurements

[J]. IEEE Transactions on Geoscience and Remote Sensing, 1996, 34(1): 237-249.

DOI      URL      [本文引用: 1]      摘要

Visible satellite sensors have monitored snowcover throughout the Northern Hemisphere for almost thirty years. These sensors can detect snowcover during daylight, cloud-free conditions. The operational procedure developed by NOAA/NESDIS requires an analyst to manually view the images in order to subjectively distinguish between clouds and snowcover. Because this procedure is manually intensive, it is only performed weekly. Since microwave sensors see through nonprecipitating clouds, snowcover can be determined objectively without the intervention of an analyst. Furthermore, microwave sensors can provide daily analysis of snowcover in real-time, which is essential for operational forecast models and regional hydrologic monitoring. Snowcover measurements are obtained form the Special Sensor Microwave Imager (SSM/I), flown aboard the DMSP satellites. A decision tree, containing various filters, is used to separate the scattering signature of snowcover from other scattering signatures. Problem areas are discussed and when possible, a filter is developed to eliminate biases. The finalized decision tree is an objective algorithm to monitor the global distribution of snowcover. Comparisons are made between the SSM/I snowcover product and the NOAA/NESDIS subjectively analyzed weekly product.
[103] Savoie M H, Armstrong R L, Brodzik M J, et al.

Atmospheric corrections for improved satellite passive microwave snow cover retrievals over the Tibet Plateau

[J]. Remote Sensing of Environment, 2009, 113(12): 2 661-2 669.

DOI      URL      [本文引用: 2]     

[104] Qiu Y, Shi J, Lemmetyinen J, et al.

The atmosphere influence to AMSR-E measurements over snow-covered areas: Simulation and experiments

[C]//Geoscience and Remote Sensing Symposium,2009 IEEE International,IGARSS, 2009. DOI:10.1109/IGARSS.2009.5418158.

[本文引用: 1]     

[105] Tedesco M, Wang J R.

Atmospheric correction of AMSR-E brightness temperatures for dry snow cover mapping

[J]. IEEE Geoscience and Remote Sensing Letters, 2006, 3(3): 320-324.

DOI      URL      [本文引用: 1]      摘要

Differences between the brightness temperatures (spectral gradient) collected by the Advanced Microwave Scanning Radiometer for EOS (AMSR-E) at 18.7 and 36.5 GHz are used to map the snow-covered area (SCA) over a region including the western U.S. The brightness temperatures are corrected to take into account for atmospheric effects by means of a simplified radiative transfer equation whose parameters are stratified using rawinsonde data collected from a few stations. The surface emissivity is estimated from the model, and the brightness temperatures at the surface are computed as the product of the surface temperature and the computed emissivity. The SCA derived from microwave data is compared with that obtained from the Moderate Resolution Imaging Spectroradiometer for both cases of corrected and noncorrected brightness temperatures. The improvement to the SCA retrievals based on the corrected brightness temperatures shows an average value around 7%
[106] Mizukami N, Perica S.

Towards improved snow water equivalent retrieval algorithms for satellite passive microwave data over the mountainous basins of western USA

[J]. Hydrological Processes, 2012, 26(13): 1 991-2 002.

DOI      URL      [本文引用: 1]     

[107] Smith T, Bookhagen B.

Assessing uncertainty and sensor biases in passive microwave data across high mountain asia

[J]. Remote Sensing of Environment, 2016, 181:174-185.

DOI      URL      摘要

While forest cover and elevation have been integrated into many SWE algorithms, wind speed and long-term maximal snow depth have not. Our results show that wind redistribution of snow can have impacts on SWE, especially over large, flat, areas. Using our regression results, we have developed an understanding of sensor-specific SWE uncertainties and their spatial patterns. The uncertainty maps developed in this study provide a first-order approximation of SWE-estimate reliability for much of HMA, and imply that high-fidelity SWE estimates can be produced for many high-elevation areas.
[108] Dai L, Che T, Ding Y, et al.

Evaluation of snow cover and snow depth on the Qinghai-Tibetan Plateau derived from passive microwave remote sensing

[J]. Cryosphere Discussions, 2017, 11(4): 1-31.

DOI      URL      [本文引用: 1]      摘要

Palsas and peat plateaus are permafrost landforms occurring in subarctic mires which constitute sensitive ecosystems with strong significance for vegetation, wildlife, hydrology and carbon cycle. Firstly, we have systematically mapped the occurrence of palsas and peat plateaus in the northernmost county of Norway (Finnmark, 藴鈥50 000 km) by manual interpretation of aerial images from 2005 to 2014 at a spatial resolution of 250 m. At this resolution, mires and wetlands with palsas or peat plateaus occur in about 850 kmof Finnmark, with the actual palsas and peat plateaus underlain by permafrost covering a surface area of approximately 110 km. Secondly, we have quantified the lateral changes of the extent of palsas and peat plateaus for four study areas located along a NW-SE transect through Finnmark by utilizing repeat aerial imagery from the 1950s to the 2010s. The results of the lateral changes reveal a total decrease of 33-71 % in the areal extent of palsas and peat plateaus during the study period, with the largest lateral change rates observed in the last decade. However, the results indicate that degradation of palsas and peat plateaus in northern Norway has been a consistent process during the second half of the 20th century and possibly even earlier. Significant rates of areal change are observed in all investigated time periods since the 1950s, and thermokarst landforms observed on aerial images from the 1950s suggest that lateral degradation was already an ongoing process at this time. The results of this study show that lateral erosion of palsas and peat plateaus is an important pathway for permafrost degradation in the sporadic permafrost zone in northern Scandinavia. While the environmental factors governing the rate of erosion are not yet fully understood, we note a moderate increase in air temperature, precipitation and snow depth during the last few decades in the region.

/