基于卫星遥感数据的吉林省西部地区积雪参数研究

发布时间：2018-12-16 23:55

【摘要】：积雪是地球表面覆盖物的重要组成部分,其对全球气候环境和人类生存条件都有着十分重要的影响,因此准确的监测和分析积雪的特征参数具有十分重要的意义。吉林省西部地区位于我国东北地区松嫩平原的中南部,当地冬季漫长,且具有积雪覆盖面积大、覆盖时间长的特点,积雪对当地的经济发展和人们的日常生活产生了较为明显的影响。此外,该地区土地盐碱化程度较高,形成了其特有的下垫面特征。本文主要选取风云三号B星微波成像仪(FY3B-MWRI)数据作为实验数据,结合光谱遥感数据,从积雪覆盖和积雪深度两个方面对吉林省西部地区的积雪情况进行研究,主要工作及研究成果如下:(1)基于MWRI被动微波遥感数据的吉林省西部地区积雪覆盖判识算法研究本文通过对现有的基于被动微波遥感数据的积雪覆盖判识算法进行分析比较,从中选取了具有代表性的Singh积雪决策树判识算法、李晓静积雪决策树判识算法和潘金梅积雪决策树判识算法,对2010年12月份和2012年至2016年每年1月份期间研究区域内的积雪覆盖情况进行判识,并将积雪覆盖判识结果与MOD10A1积雪产品进行对比。研究结果表明,在观测期间内,采用的三种积雪覆盖判识算法的判识精度均未能达到较高精度。通过对其误差来源进行分析,本文对原有判识算法的结构和参数进行了优化,提出了一种更加适用于吉林省西部地区的积雪决策树判识改进方法。实验结果表明,针对研究区域,本文提出的改进方法的积雪覆盖判识精度达到了95.4%,明显高于Singh积雪决策树判识算法的78.3%、潘金梅积雪决策树判识算法的76.7%和李晓静积雪决策树判识算法的89.6%。(2)基于MWRI被动微波遥感数据的吉林省西部地区雪深反演算法研究本文采用FY3B业务化雪深反演算法和Chang雪深反演经验算法,利用MWRI被动微波遥感数据,实现吉林省西部地区2010年12月份及2012年至2015年每年1月份的雪深反演,并结合土地分类数据对不同下垫面上的雪深均值进行了统计和对比分析。为了进一步提高雪深反演精度,本文将改进方法得到的雪盖判识结果与雪深反演算法相结合,并将雪盖判识结果为无雪区域的雪深值进行剔除,只统计雪盖判识结果为有雪区域内的雪深值,此时统计结果显示四种下垫面上的积雪深度均值明显增加。此外,研究结果表明:观测地区的积雪深度在空间分布上呈现出一种自东南向西北逐步递减的趋势。
[Abstract]:Snow cover is an important part of the earth surface cover, which has a very important impact on the global climate environment and human living conditions. Therefore, it is of great significance to accurately monitor and analyze the snow cover characteristics. The western region of Jilin Province is located in the central and southern part of the Songnen Plain in Northeast China. The local winter is long and has the characteristics of large snow cover area and long covering time. Snow has an obvious impact on local economic development and people's daily life. In addition, the salinization degree of the land in this area is relatively high, forming its unique underlying surface characteristics. In this paper, the data of Fengyun No. 3 B Star Microwave Imager (FY3B-MWRI) are selected as experimental data, combined with spectral remote sensing data, the snow cover and snow depth in western Jilin Province are studied from two aspects: snow cover and snow depth. The main work and research results are as follows: (1) based on MWRI passive microwave remote sensing data, snow cover recognition algorithm based on passive microwave remote sensing data in western Jilin Province is studied. In this paper, the existing snow cover recognition algorithm based on passive microwave remote sensing data is studied. Method for analysis and comparison, The typical Singh snow decision tree recognition algorithm, Li Xiaojing snow decision tree recognition algorithm and Pan Jinmei snow decision tree recognition algorithm are selected. The snow cover in the study area between December 2010 and January 2012 to 2016 was identified, and the results were compared with MOD10A1 snow cover products. The results show that the accuracy of the three snow cover recognition algorithms can not reach higher accuracy during the observation period. Based on the analysis of the error sources, this paper optimizes the structure and parameters of the original recognition algorithm, and puts forward an improved method of snow decision tree recognition, which is more suitable for the western region of Jilin Province. The experimental results show that the accuracy of snow cover recognition of the improved method proposed in this paper is 95.4, which is obviously higher than that of Singh snow decision tree. 76.7% of Pan Jinmei snow decision tree and 89.6% of Li Xiaojing snow decision tree recognition algorithm. (2) based on MWRI passive microwave remote sensing data, the snow depth inversion algorithm in western Jilin Province is studied in this paper. FY3B operational snow depth inversion algorithm and Chang snow depth inversion empirical algorithm, Using MWRI passive microwave remote sensing data, the snow depth inversion in December 2010 and from 2012 to 2015 is realized in western Jilin Province, and the mean value of snow depth on different underlying surfaces is statistically analyzed and compared with land classification data. In order to further improve the accuracy of snow depth inversion, this paper combines the result of snow cover recognition obtained by the improved method with the snow depth inversion algorithm, and the result of snow cover recognition is eliminated from the snow depth value of the region without snow. Only the result of snow cover recognition is the snow depth in the region with snow, and the statistical results show that the mean value of snow depth on the four kinds of underlying surfaces is obviously increased. In addition, the results show that the depth of snow in the observed area is decreasing gradually from southeast to northwest.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P407

【参考文献】