Arid
天山北坡三工河流域山地植被丰度变化遥感监测研究
其他题名Monitoring the Vegetation Density Change in San-gong River Watershed of Tianshan Mountains
陈耀亮
出版年2014
学位类型硕士
导师罗格平
学位授予单位中国科学院大学
中文摘要山地森林分布与地形、气候密切关联,但半个世纪以来人类活动对山地森林变化的贡献日益显著。中国西部干旱区具有典型的山盆结构,山地系统地表覆被主要由低山丘陵草原、山地森林和高山草地组成。半个多世纪来,在气候变化和人类活动共同影响下,干旱区山地森林的时空分布和内部结构发生了明显的变化,但是这种变化的量化研究还较为缺乏,尤其是在植被丰度的变化方面。植被丰度是一个用来反应植被分布密度和生长状况的抽象概念,植被丰度的高低可以通过植被覆盖度、植被指数、叶面积指数以及其它具有生物物理意义的遥感参数进行表征。传统的野外调查方法具有费时、费力、费财、监测面积小等缺陷,不能够很好地反映山地森林内部变化的时空差异特征。遥感方法因其具有丰富的光谱纹理信息和能够实现长时间的动态监测而为植被丰度变化的量化研究提供了新的思路。本文选择天山北坡典型的三工河流域为研究区,其山区主要分布有乔木(以雪岭云杉为主)、灌木(以欧亚圆柏为主)和草地三种植被类型。主要利用1990年、2001年和2010年3期Landsat 5 TM多光谱遥感数据,基于线性混合像元分解技术获得研究区植被分量(Green Vegetation Fraction,以下简称GV)、土壤分量(Soil Fraction, 以下简称Soil)、低反照率分量(Low Albedo Fraction,以下简称Low)、高反照率分量(High Albedo Fraction,以下简称High)和一个由GV与Low归一化的组合分量,即归一化植被-低反照率分量(Normalized Difference Index between Green Vegetation Fraction and Low Albedo Fraction,以下简称NDGLI)。首先利用基于分量的决策树分类方法确定了云杉、欧亚圆柏和草地的分布现状,2010年的分类后结果主要通过2010年的GoogleEarth影像和2006年Quickbird影像进行分层随机抽样验证,总体精度为0.86,Kappa系数为0.81;其次利用这些分量的变化及其与环境因子(海拔、坡度、坡向)的相关性探测云杉、欧亚圆柏丰度的时空分布和变化情况;最后利用研究区天池气象站点1990-2010年的温度和降水数据、阜康林场的1990-2010年的商业择伐与云杉抚育调查数据分析了这些变化的驱动机制。本文研究对推动干旱区山地植被变化的研究具有显著的意义。本项研究取得的主要结果如下: (1)云杉大部分分布在1900米-2700米的海拔高度。从海拔1800米开始,云杉分布面积迅速增加,在海拔2200米左右分布面积达到最高值,之后随着海拔的升高开始逐渐减少;云杉主要分布在10-40坡度之间,其中在25-32°之间分布范围最广;云杉主要分布在阴坡、半阴坡和半阳坡(292.5-360°,0-67.5°)。欧亚圆柏大部分分布在2500 -3000米的海拔高度。从海拔1900米开始,欧亚圆柏分布面积逐渐增加,在海拔2800米左右分布面积达到最高值,之后随着海拔的升高开始逐渐减少,但在海拔3300米左右仍有极少量分布;欧亚圆柏主要分布在15-45坡度之间,其中在25-38°之间分布范围最广;欧亚圆柏的分布主要分布在阳坡和半阳坡(30-60°, 180-270°)。 (2)云杉和欧亚圆柏的GV、NDGLI和归一化植被指数(Normalized Difference Vegetation Index,以下简称NDVI)呈现指数型增长的关系。但由于NDVI存在饱和现象,GV、Soil、Low、NDGLII等分量的变化范围比NDVI要广,因此可利用各分量取代NDVI进行植被丰度变化监测,这既可以避免光谱饱和现象又可以有效地识别其变化程度。 (3)1990-2001年期间,云杉的NDVI、GV、Low和NDGLI分量整体变化在空间上既存在增加的区域,也存在减少的区域,但是欧亚圆柏的 NDVI、GV和NDGLI分量绝大部分是增加的,Low分量绝大部分是减少的,表明欧亚圆柏的丰度在增加;2001-2010年期间,云杉的NDVI、GV和NDGLI整体上是增加的,而Low是减少的,变化幅度明显不同于1990-2001年期间。同样,欧亚圆柏也呈现NDVI、GV和NDGLI增加和Low减少的特征,显著不同于1990-2001年期间的变化。2001-2010年期间,云杉各分量的变化程度比欧亚圆柏各分量变化程度更为显著。 (4)混合线性分解各分量能够有效地反映研究区高山植被的变化。1990-2001年期间研究区种植云杉和择伐云杉的现象同时存在,这使郁闭度总体没有明显的变化,基本保持在0.3左右,这与此时期的云杉空间变化分布特征相吻合。2001-2010年由于“天然林保护工程”的实施,山区天然林全面禁伐,而云杉人工种植却未间断,并且这期间降水量和温度总体增加,致使云杉NDVI、GV和NDGLI的变化呈现增加的趋势,Low分量呈减少的态势。可以说20年间云杉DNVI、GV和NDGLI的变化是气候变化和人类活动共同作用的结果;1990-2010年期间欧亚圆柏基本上未受到人类活动影响,在气候变暖变湿的背景下,致使欧亚圆柏NDVI、GV和NDGLI在这20年期间呈现增加的趋势, Low呈减少的趋势。 本文主要利用Landsat 5 TM数据进行植被丰度变化的监测,但TM数据存在空间分辨率较低的缺陷。而且,由于缺乏历史时期的野外调查数据和高分遥感数据,使得变化监测的验证难以进行。高空间分辨率数据具有空间分辨率高、纹理信息丰富的优势,能够更加有效地监测植被丰度的变化,今后应该加强高空间分辨率数据在森林监测中的应用。另外,由于云杉和欧亚圆柏的生长状况直接与其生理参数和生长环境直接相关,还应该加强野外调查,建立长期的样地数据库,实现地-空一体化监测。
英文摘要The vegetation abundance change of arid forest is intensively related to global climate change and human activities. In the last past 20 years, what made arid forest change most greatly is not global warming but human activities. The vegetation in mountain systems of Central Asia arid region is composed of the hilly grasslands, mountainous forests and alpine meadows. In recent decades, as the joint impact of human activity and climate change, large changes in the distribution of forest and grassland took place, but few research was conducted to study them, especially in the aspect of vegetation abundance. Vegetation abundance is an abstract concept which can be used to indicate the vegetation density and growth condition. Vegetation index, vegetation cover, leaf area index and other biophysical values can be used to quantify the vegetation abundance. Traditional monitoring method like field survey can not well detect the spatio-temporal difference of vegetation abundance because it is time-consuming, expensive and can only monitor a small region. However, remote sensing can work this hot issue well due to its abundant spectrum and texture information and the ability to make a long term detection. Study area is located in the region of Sangonghe Basin, Xinjiang, China, where there are three kinds of vegetation types, P. Schrenkiana, J. Sabina and grassland. In this research, through analysing the change of three fractional values,green vegetaion fraction(GV), low albedo fraction(Low) and a new fraction combined with GV and Low, normalized difference index of GV and Low(NDGLI),which were derived from the linear spectral analysis in the whole region, a new method was established to detect the spatiotemporal distribution and abundance changes of P. Schrenkiana, J. Sabina in the last 20 years. The main conclusions include as follows: (1) P. Schrenkiana is mainly in 1900-2700 m a.s.l with rapidly increasing of area from 1800 m a.s.l and arriving its peak around 2200 m a.s.l while J. Sabina is mainly in 2500-3000m a.s.l with general area augment from 1900 m and reaching its highest point around 2800 m. As to the slope and aspect distribution, P. Schrenkiana is concentrated on medium slope(25-32 degree) and shady and semi-shady area(0-67.5 degree, 292.5-360 degree in aspect) while J. Sabina is mainly distributed on medium slope(25-38 degree) as well and sunny and semi-sunny area(30-120 degree, 180-270 degree in aspect). (2)GV, NDGLI of P. Schrenkiana showed perfectly exponential growth relationship with its normalized difference of vegetation index(NDVI), respectively, and so does the GV, NDGLI of J. Sabina. As the saturated phenomenon of NDVI is very apparent and the change extent of each fraction is fairly wider than that of NDVI, thus using the changed fraction to monitor vegetation density rather than the changed NDVI can not only avoid NDVI’s saturation but also can conduct a good division in the change degree. (3)In the period of 1990-2001, as to P. Schrenkiana, there were many increased pixels as well as decreased pixels of NDVI, GV, Low and NDGLI image while increased pixels of NDVI, GV and NDGLI image dominated the whole area of J. Sabina and so did the decreased pixels of Low image. In the period of 2001-2010, as to P. Schrenkiana, the change was totally different with the previous decade. The number of increased pixels of NDVI, GV and NDGLI image were far more than that of decreased pixels, and so did the amount of decreased pixels of Low image. In terms of J. Sabina, the basic trends of all the fraction change was as same as those of the previous decade but had a further deeper change. We also find these changes of both P. Schrenkiana and J. Sabina in this period are more deeper and bigger than corresponding changes in the period of 1990-2001. (4)The climate change can well explain the spatiotemporal change of J. Sabina. In the past 20 years, no human impacts was found to J. Sabina and with the slight increase of bot
中文关键词天山北坡 ; 遥感 ; 云杉 ; 欧亚圆柏 ; 线性混合像元分解 ; 植被丰度
英文关键词North Slope of Tianshan Mountain, Remote Sensing,Picea Schrenkiana,Juniperus Sabina,Linear Spectral Unmixing,Vegetation Abundance
语种中文
国家中国
来源学科分类地图学与地理信息系统
来源机构中国科学院新疆生态与地理研究所
资源类型学位论文
条目标识符http://119.78.100.177/qdio/handle/2XILL650/287450
推荐引用方式
GB/T 7714
陈耀亮. 天山北坡三工河流域山地植被丰度变化遥感监测研究[D]. 中国科学院大学,2014.
条目包含的文件
条目无相关文件。
个性服务
推荐该条目
保存到收藏夹
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[陈耀亮]的文章
百度学术
百度学术中相似的文章
[陈耀亮]的文章
必应学术
必应学术中相似的文章
[陈耀亮]的文章
相关权益政策
暂无数据
收藏/分享

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。