南方红壤典型水土流失区植被覆盖度变化及其生态效应评估—

引用本文

张灿, 徐涵秋, 张好, 唐菲, 林中立. .南方红壤典型水土流失区植被覆盖度变化及其生态效应评估——以福建省长汀县为例.自然资源学报, 2015,30(6): 917-928
ZHANG Can, XU Han-qiu, ZHANG Hao, TANG Fei, LIN Zhong-li. .Fractional Vegetation Cover Change and Its Ecological Effect Assessment in a Typical Reddish Soil Region of Southeastern China: Changting County, Fujian Province.Journal of Natural Resources,2015,30(6): 917-928 复制到剪切板

Permissions

《自然资源学报》编辑部所有

南方红壤典型水土流失区植被覆盖度变化及其生态效应评估——以福建省长汀县为例

张灿, 徐涵秋^*, 张好, 唐菲, 林中立

福州大学环境与资源学院,福州大学遥感信息工程研究所,福建省水土流失遥感监测评估与灾难防治重点实验室,福州 350108

第一作者简介：张灿（1990- ）,女,河南邓州人,硕士研究生,研究方向为环境与资源遥感。E-mail: 1721374724@qq.com

*通信作者简介：徐涵秋（1955- ）,男,江苏盐城人,博士（爱尔兰国立大学）,教授,博导,主要从事环境与资源遥感研究。E-mail: hxu@fzu.edu.cn

收稿日期:2014-04-17 网络出版日期:2015-06-20

基金:国家科技支撑计划项目（2013BAC0801-05）

摘要

南方红壤典型水土流失区——福建省长汀县曾因生态破坏导致严重的水土流失。经过多年以植树为主的生态修复,该县生态面貌有了明显的改观。论文首先采用线性光谱混合分析模型计算植被覆盖度,并在此原始模型的基础上提出了对地形阴影进行修正的方法来获取植被覆盖度。精度验证表明,在线性光谱混合分析模型中加入山地指数（NDMVI）波段能够削弱地形阴影问题,提高植被覆盖度反演精度。在此基础上利用多时相遥感影像分析了长汀2001—2013年植被覆盖度的时空变化,并利用遥感生态指数（ RSEI）定量评价了长汀水土流失生态修复的效果。结果表明,经过13 a的水土流失治理,长汀的植被覆盖度有了明显的升高,从2001年的75.1%上升到2013年的86.5%。 RSEI生态指数值也随之上升,生态等级为优良的面积比例从85.83%增加到90.59%,反映了长汀县生态质量整体有了明显的提高。植被的生态效应定量研究表明,长汀县的植被覆盖度每增加10%, RSEI生态指数值至少提高10%,植被覆盖度的生态提升效应显著。

关键词: 遥感; 植被覆盖度; 遥感生态指数（ RSEI）; 线性光谱混合分析; 长汀

中图分类号:TP79:Q948 文献标志码:A 文章编号:1000-3037(2015)06-0917-12

Fractional Vegetation Cover Change and Its Ecological Effect Assessment in a Typical Reddish Soil Region of Southeastern China: Changting County, Fujian Province

ZHANG Can, XU Han-qiu, ZHANG Hao, TANG Fei, LIN Zhong-li

College of Environment and Resources, Institute of Remote Sensing Information Engineering, Fujian Provincial Key Laboratory of Remote Sensing of Soil Erosion and Disaster Protection, Fuzhou University, Fuzhou 350108, China

Abstract

Changting County in Fujian Province is a typical reddish soil erosion region in southeastern China. The county has been called the flame mountains area due to severe soil erosion on barren terrains. After years of ecological restoration, the local ecosystem has been improved remarkably and the barren lands are now covered with forests. This paper used remote sensing techniques to study the fractional vegetation cover ( FVC) changes and quantitatively evaluated the effects of ecological restoration in Changting during the period from 2001 to 2013 by using the remote sensing ecological index ( RSEI). Two remote-sensing based models for estimating FVC have been compared in order to select a suitable model for the retrieval of the FVC in the study area. One is the commonly-used Linear Spectral Mixture Analysis (LSMA) model, and the other is the LSMA_M model. The LSMA_M is based on the LSMA but with a Normalized Difference Mountain Vegetation Index (NDMVI) derived band added into the model. The comparative analysis confirms that the LSMA_M model has higher accuracy than the LSMA model, indicated by its lower root mean square error and higher correlation with referenced FVC data because the addition of the NDMVI band into the model could eliminate shadow problem caused by topographic relief in the mountainous areas. Therefore, the LSMA_Mmodel was selected to retrieve the FVC in this study. The result indicates that the 13-year effort for treatment of the soil loss in the county has led to a notable increase in county's FVC from 75.1% (2001) to 86.5% (2013). Meanwhile, the RSEI-based analysis also indicates a significant improvement of the county's ecological quality during the same period, because of the increase in RSEI value from 0.750 to 0.787, along with an increase in high- RSEI-level area from 85.83% to 90.59%. Regression analysis between FVC and RSEI suggests that each 10% increment in FVC could raise the RSEI by 10%. This clearly indicates a significant effect of FVC on county's ecological quality.

Keyword: remote sensing; fractional vegetation cover; RSEI; Linear Spectral Mixture Analysis; Changting

Show Figures

福建省长汀县是我国南方红壤地区的典型水土流失区, 其河田镇一带的严重水土流失历来备受关注。在20世纪40年代初, 河田镇便与甘肃天水、陕西长安一起建立了水土保持试验站, 并同时被列为当时全国的3个重点水土保持试验区, 但治理成效甚微。新中国成立以来, 特别是从20世纪80年代开始, 各级政府逐步加大长汀水土流失治理的力度, 制订了一系列激励措施。福建省政府更是在2000年将长汀水土流失治理列入为民办实事项目之一。经过20多年的整治, 长汀县水土保持工作取得了明显的成就, 生态面貌有了明显的改观, 并成为我国水土流失治理的典范。

大面积地表裸露、植被稀疏是长汀水土流失的典型特征。遍布红土、不见绿被的山体曾有“ 红色沙漠” 之称。因此, 长汀县的水土流失治理多以植被恢复为主, 通过造林绿化来改善区域生态系统。随着长汀县水土流失治理力度的加大, 对其所产生的生态效应进行研究和评估就显得愈为重要。由于植被是生态系统的重要组成成分, 植被覆盖度的变化是地球内、外部作用与人类改造相互作用的结果^[1], 对区域生态系统具有重要的影响。因此围绕植被覆盖度及其生态效应的研究长期以来一直为科学界所重视, 例如植被覆盖度与土壤湿度、地表温度和气候变化^{[2, 3, 4]}等关系的研究。除此之外, 植被覆盖度还是气候模型、水文模型、水土流失土壤侵蚀模型以及生态系统模型的重要输入参数^{[5, 6, 7, 8, 9]}。特别是植被覆盖度作为直观描述陆地植被生长状况的综合量化指标, 在生态环境综合评价、植被变化监测、水土保持以及干旱、沙尘暴^{[10, 11, 12, 13]}等自然灾害监测中得到了广泛的应用。

本文在研究长汀县2001— 2013年植被覆盖度变化的基础上, 重点对其产生的生态效应进行评估, 并研究植被覆盖度与生态变化之间的定量关系, 这对长汀县进一步治理水土流失和改善生态质量具有重要的现实指导意义。

1 研究方法

1.1 研究区概况

长汀县位于福建省西南部, 其地理位置为东经116° 00′ ~116° 40′ , 北纬25° 18′ ~26° 02′ , 土地总面积3 099 km²。该县以低山、丘陵地带为主, 成土母岩类型主要为砂、泥质岩和酸性火成岩等, 质脆而面蚀强烈, 风化后的土壤类型主要为红壤和砂壤土。长汀属亚热带季风气候, 年平均气温18.3 ℃, 年均降雨量达1 700 mm, 独特的自然条件决定了该区生态的脆弱性。根据中国植被区划, 长汀县属于亚热带常绿阔叶林区。但由于长期不合理的开发利用, 原生植被已遭毁灭性破坏, 代之的是大量的马尾松林。长汀县在地形上呈四周隆起、中部下凹的盆地特征, 这一盆地即为水土流失严重的河田盆地区^[14]。

1.2 遥感影像数据及其预处理

本文采用Landsat系列遥感影像为数据源, 所用影像的获取时间为：2001年10月12日（ETM+）、2004年10月12日（TM）和2013年10月5日（OLI）。其中2013年影像为最新发射的Landsat 8遥感影像。遥感数据的预处理包括几何配准和辐射校正。首先, 以2001年影像为基准, 采用二次多项式和最邻近像元法分别对2004和2013年影像进行配准, 配准的均方根误差小于0.5个像元。为减少不同时间的影像在大气、光照等各方面的差异与影响, 采用Chander等^{[15, 16]}的模型和Chavez^[17]的大气校正参数, 将Landsat TM/ETM+原始影像的DN值转换为传感器处反射率来进行辐射校正, 公式如下：

$\begin{matrix} ρ_{λ} = \frac{π [(Q_{λ} - h_{λ}) gai n_{λ}] d^{2}}{ESU N_{λ} \cos θ_{s}} \end{matrix}$ （1）

式中： $\begin{matrix} ρ_{λ} \end{matrix}$ 为传感器处的反射率, λ 为波段值; Q_λ为波段λ 的像元DN值; h_λ为波段λ 最暗像元的DN值; gain_λ为波段λ 的增益值[W/(m² · sr· µ m)]; ESUN_λ为大气顶部的平均太阳辐照度[W/(m²· µ m)]; d为日地天文单位距离; θ _s为太阳天顶角。

对于新发射的Landsat 8遥感影像, 依据美国USGS（http://glovis.usgs.gov/）提供的定标公式将影像的DN值转换为表观反射率：

$\begin{matrix} ρ_{λ} = \frac{M_{ρ} Q_{cal} + A_{ρ}}{\cos θ_{SZ}} \end{matrix}$ （2）

式中：ρ _λ为表观反射率; θ _SZ为太阳天顶角; Q_cal为像元DN值; M_ρ、A_ρ分别为相应波段的调整系数和偏差值, 均可在影像头文件获得。

1.3 植被覆盖度模型的确定

当前, 植被覆盖度（Fractional Vegetation Cover, FVC）的估算以线性光谱混合分析模型和基于NDVI植被指数的像元二分模型最为常用^[12]。考虑到本文将研究植被覆盖度与遥感生态指数（RSEI）为代表的生态效应关系, 而RSEI指数的计算基础是NDVI, 因此为避免二者互为相关, 本文采用独立于NDVI的线性光谱混合分析方法来计算研究区的植被覆盖度。

1.3.1 线性光谱混合分析

在混合像元分解法中, 线性光谱混合分析（Linear Spectral Mixture Analysis, LSMA）以其物理意义明显且相对简单而得到广泛应用^{[18, 19, 20]}, 它将像元在某一光谱波段的反射率视为由构成像元的各终端单元（Endmember）的反射率（亮度值）以其所占像元面积比例为权重系数的线性组合, 公式及其约束条件如下^[21]：

$\begin{matrix} R_{b} = \overset{N}{\sum_{i = 1}} f_{i} R_{ib} + e {}_{b}, \overset{N}{\sum_{i = 1}} f_{i} = 1, f_{i} \geq 0 \end{matrix}$ （3）

式中：R_b为波段b的反射率; N为端元的数目; f_i为端元i的权重, 代表各个光谱端元所占的比率; R_ib为端元i在波段b的反射率; e_b为残差。模型的精度可通过计算每个波段的残差（e_b）的均方根来检验：

$\begin{matrix} RMS = {(\frac{\overset{M}{\sum_{b = 1}} e_{b}^{2}}{M})}^{\frac{1}{2}} \end{matrix}$ （4）

式中：RMS为残差e_b的均方根; M是影像中的波段数。RMS越小, 则模型的误差越小。

在线性光谱分析的端元确定中, 以V-L-H-S（植被-低反照地物-高反照地物-裸露土壤）的4端元模型^[20]最为常用。但由于该模型中未包含水体, 因此需先对影像进行水体掩膜, 然后对原始的反射率影像进行最小噪声分离变换（Minimum Noise Fraction, MNF）, 实现数据降维和噪音分离以消除波段间的相关性。在统计MNF变换后各分量对应的特征值发现, 原始影像的前3个主分量特征值在原始影像信息量中所占的百分比为85.87%, 可以有效地表示数据的基本维度。因此可在MNF变换的前3个分量所形成的散点图中, 圈出不同端元的样本区域, 得到各端元的光谱均值（图1）。在理想的线性光谱混合分析模型中, 各端元应该分布在三角特征空间的顶点区域, 每个顶点区域的散点代表一种光谱端元, 越靠近边缘, 像元越纯净^[21]。最后利用线性光谱混合分析进行波谱分离, 得到4个端元的丰度图像和残余误差图像, 其中植被端元的丰度图像即为植被覆盖度影像。

	Figure Option View Download New Window
	图1 原始影像MNF前3个分量特征空间散点图及纯像元位置Fig. 1 Scatterplots of the first three MNF components showing the purity of pixels

由于研究区的植被主要分布在山地上, 地形起伏的阴影会对阴坡的植被覆盖度估算带来误差^{[22, 23]}。对此, 在原始影像的6个多光谱波段中引入能够削弱阴影的山地指数NDMVI（Normalized Difference Mountain Vegetation Index）^[24]作为辅助波段, 以削弱阴坡对线性光谱分解结果的影响, 提高分解结果的精度。为了描述方便, 将引入山地指数波段的LSMA模型称为LSMA_M, 以与原始的LSMA区别。

1.3.2 精度验证

为了更准确地估算长汀的植被覆盖度, 以研究区2004年10月12日影像为例, 对LSMA和LSMA_M两种模型反演得到的植被覆盖度进行精度验证, 通过对比两种模型的精度, 选取最适合于研究区的植被覆盖度反演模型, 然后依此模型反演2001和2013年影像的植被覆盖度。

精度验证采用基于高分辨率影像的亚像元对比法^[21], 该方法不仅能检验精度, 而且还可以将由模型求出的植被覆盖度和高分辨率影像所代表的实际植被覆盖度关联起来。利用3 $\begin{matrix} \times \end{matrix}$ 3窗口随机选取了150个样点（每个样点包括9个TM像元）, 并在近同期（2004-10-05）的高分辨率（5 m）SPOT 5的数据上, 对每个样点上所对应的324个SPOT像元直接进行人工数字化, 计算出实际植被覆盖度, 然后与模型求出的植被覆盖度进行对比回归分析, 模型精度通过计算均方根误差（RMSE）来判断, RMSE的计算公式为：

$\begin{matrix} RMSE = {[\frac{{\overset{n}{\sum_{i = 1}} (X_{i} - Y_{i})}^{2}}{n}]}^{\frac{1}{2}} \end{matrix}$ （5）

式中：X_i是模型求出的植被覆盖度估计值; Y_i是实际的植被覆盖度值; n为样本数。系统误差RMSE反映的是样本的总体精度, RMSE值越小, 则表明模型精度越高。

1.4 生态指数计算

为了考察植被变化所产生的生态效应, 本文选用徐涵秋^[25]新近提出的遥感生态指数（Remote Sensing Ecological Index, RSEI）来对长汀县的生态进行评价。该指数是一综合多种要素的生态综合评价指数, 它不仅考虑了植被所代表的绿度变化, 还结合了湿度、热度、干度等多种因素, 因此可以较全面地考察区域的生态变化, 其结果与国家环保部的生态状况指数（EI）^[26]也有较好的可比性^[27]。RSEI可表示为上述各因子的函数, 即：

$\begin{matrix} RSEI = f (Green, Wet, Thermal, Dry) \end{matrix}$ （6）

式中：Green、Wet、Thermal和Dry分别为绿度、湿度、热度和干度, 这4个指标可分别由遥感的对应指数或参量获得。以Landsat TM影像为例, 它们的计算公式分别为：

绿度指标以植被指数NDVI^[28]来代表：

$\begin{matrix} NDVI = \frac{ρ_{TM 4} - ρ_{TM 3}}{ρ_{TM 4} + ρ_{TM 3}} \end{matrix}$ （7）

式中：ρ _TM₄和ρ _TM₃分别代表TM4和TM3波段的反射率。

湿度指标由缨帽变换中的湿度分量^[29]代表：

$\begin{matrix} Wet = 0.031 5 ρ_{1} + 0.202 1 ρ_{2} + 0.310 2 ρ_{3} + 0.159 4 ρ_{4} - 0.680 6 ρ_{5} - 0.610 9 ρ_{7} \end{matrix}$ （8）

式中： $\begin{matrix} ρ_{1}, \dots, ρ_{5}, ρ_{7} \end{matrix}$ 表示TM各对应波段的反射率。

热度指标由经过比辐射率校正的亮温来代表。亮温BT的计算参照Landsat用户手册的模型^[30], 再对其进行比辐射率校正^{[31, 32]}：

$\begin{matrix} T = \frac{BT}{1 + (\frac{λT}{ρ}) lnε} - 273.15 \end{matrix}$ （9）

式中：T为经过比辐射率校正后的亮温（℃）; λ 为第6波段的中心波长（11.45 μ m）; ρ =1.438× 10^-2 mK; ε 为地表比辐射率, 其取值见文献[33]。

干度指标采用由裸土指数SI^[34]和建筑指数IBI^[35]合成得到的干度指数NDSI来代表：

$\begin{matrix} NDSI = \frac{1}{2} (SI + IBI) \end{matrix}$ （10）

由于以上4个指标的量纲不统一, 为消除不同数据间的差异, 必须对其归一化, 使其数值范围统一于[0, 1]之间, 归一化公式为：

$\begin{matrix} N I_{i} = \frac{Indicato r_{i} - Indicato r_{\min}}{Indicato r_{\max} - Indicato r_{\min}} \end{matrix}$ （11）

式中：NI_i为归一化后的某一指标值, Indicator_i为该指标在像元i的值, Indicator_max和Indicator_min分别为该指标的最大值和最小值。

对经过归一化后的4个指标计算第一主成分（PC1）, 使好的生态条件由PC1值大的来代表, 再用1减去计算出的PC1, 得到初始生态指数RSEI₀：

$\begin{matrix} RSE I_{0} = 1 - \{PC 1 [f (NDVI, Wet, T, NDSI)]\} \end{matrix}$ （12）

同理, 利用公式（11）对RSEI₀进行归一化, 即得到RSEI, 经过归一化后的RSEI值介于[0, 1]之间, RSEI值越接近1, 生态越好。

2 结果

2.1 植被覆盖度反演结果的分析比较

根据1.3.2所述的精度验证方法, 分别对LSMA和LSMA_M模型反演的植被覆盖度进行精度验证（图2）。

	Figure Option View Download New Window
	图2 LSMA和LSMA_M模型反演的植被覆盖度（FVC）的精度比较注：方程通过0.1%的显著性检验。Fig. 2 Accuracy assessment of the LSMA and LSMA_M models

由图2可知, LSMA_M模型在各方面都有明显的优势, 与LSMA模型相较, 其相关系数R值增加了0.173, RMSE值降低近1倍, 这说明加入山地指数波段的LSMA_M模型比未加该波段的LSMA模型的精度有了大幅提高。显然, 在原LSMA模型中加入山地指数波段可以有效提升阴坡植被的反射率, 从而降低了阴坡植被和低反照地物的混淆度, 提高了植被覆盖度的反演精度。综上, 本文选用引入山地指数的LSMA_M模型来反演各研究年份的植被覆盖度。

2.2 植被覆盖度变化

图3(a) (b) (c) 和表1为利用LSMA_M模型反演的各研究年份的植被覆盖度影像图及其统计值。

由表1可知, 长汀县的植被覆盖度在这13 a间呈总体上升趋势, 其均值从2001年的75.1%上升到2013年的86.5%。但从不同的时间段来看, 其植被覆盖度经历了先下降后上升的趋势。在2001— 2004年间的植被覆盖度有所下降, 其均值从2001年的75.1%下降到2004年的72.1%, 下降了3个百分点; 但在2004— 2013年间植被覆盖度则有大幅度增加, 由2004年的72.1%增加到86.5%, 增加了14.4个百分点, 说明13 a来长汀加大水土流失治理的成效显著。从标准差来看, 其值也从13.2%减少到12.3%, 这说明长汀县的植被覆盖度在增加的同时, 空间差异性也在逐渐减小, 植被覆盖度趋于统一。

	Figure Option View Download New Window
	图3 长汀县各年份植被覆盖度（FVC）影像和遥感生态指数（RSEI）影像Fig. 3 FVC and RSEI images of each study year of Changting

从图3 (a) (b) (c) 可以看出, 这13 a来长汀县植被覆盖度在各地区均有不同程度的增加, 表现为2013年植被覆盖度影像整体较2001和2004年颜色偏红。且从空间分布来看, 植被覆盖度增加的区域主要集中在中部的河田盆地区, 该区中代表植被覆盖度提高的红色图斑在2013年有了明显的增加。与此同时, 盆地周围边缘地区的红色较前两个时相进一步加深, 表明了这一区域内的植被覆盖度也有所提升。

表1 研究区各年份植被覆盖度统计 Table 1 Statistics of fractional vegetation cover of each study year

2.3 生态质量变化

利用1.4的公式分别计算各研究年份的生态指数, 得到研究区各时相的遥感生态指数影像[图3 (d) (e) (f) ], 并对其进行统计（表2）。

表2 长汀县各年份4个生态分指标和遥感生态指数RSEI的统计值 Table 2 Statistics of four indicators and RSEI in Changting

在表2各指标第一主成分（PC1）的载荷中, 湿度和绿度指标的值与干度和热度指标的值符号相反, 说明它们对生态起不同的作用。由各指标均值变化情况可知, 2001— 2013年间, 代表生态条件差的干度指标和热度指标的均值表现为总体下降的趋势, 而对生态有利的绿度指标和湿度指标的均值表现为总体上升。4个指标的各自表现说明了长汀县的生态质量整体趋于变好和提高, 因此遥感生态指数RSEI整体也呈现为上升趋势, 其均值从2001年的0.750上升到2013年的0.787。但2004年的RSEI（0.705）比2001年的RSEI（0.750）有所下降, 虽然代表生态条件变好的湿度和绿度在该年份均表现为上升, 但是代表生态条件变差的热度和干度的均值也有明显增加, 特别是干度指标比2001年大幅提高了0.165, 因此湿度和绿度的上升不足以抵消热度和干度的提高所带来的负面影响, 从而导致了生态指数相应降低。同样, 2013年的湿度虽较2004年有所下降, 但其他指数均值皆往生态好的方向转变, 因此生态指数得以整体升高。

为了进一步了解长汀的生态质量分布, 以0.2为间隔将RSEI指数分为5个生态等级, 分别为差、较差、中等、良和优, 并统计各生态等级的面积和所占的比例（表3）。

从表3统计结果可以看出, 在研究期间, 长汀县的生态质量呈整体上升趋势, 具体表现在生态级别为优良等级（4、5级）所占的面积比例从85.83%上升到90.59%, 生态等级为差和较差（1、2级）所占的面积比例从3.09%下降到1.95%。但在2001— 2004年期间, 生态等级为优良（4、5级）所占的面积比例从85.83%下降到80.84%, 而差到中等的生态等级所占面积比例也相应上升, 说明生态质量出现下降; 在2004— 2013年期间生态质量大有好转, 生态等级为优良（4、5级）所占的面积比例显著增加, 从80.84%又上升到90.59%, 提高了约10%, 等级为较差到中等的面积比例相应大幅下降。

表3 长汀县生态等级面积和比例 Table 3 Area and percent changes of graded RSEI from 2001 to 2013 in Changting

2.4 空间分布变化

为了进一步直观了解长汀县的植被覆盖度和生态质量的空间变化, 利用遥感变化检测技术获得研究区植被覆盖度和生态质量等级的变化图（图4）。遥感变化检测技术采用三原色混色技术, 即植被覆盖度或生态质量上升的地区由绿色代表, 植被覆盖度或生态质量下降的地区由红色代表, 植被覆盖度或生态质量不变区则由黄色（红绿的混色）来代表。

	Figure Option View Download New Window
	图4 长汀县2001— 2013年植被覆盖度和生态质量变化检测Fig. 4 Change detection images of FVC (a) and RSEI (b) between 2001 and 2013

从植被覆盖度的变化[图4 (a)]来看, 盆地周围的地区13 a里基本未发生变化, 呈黄色, 这是由于盆地周边是地势较高、受人类活动干扰较少的山区, 因此植被覆盖度历年来都较好, 没有发生明显的变化。而中部盆地内及其边缘的低矮山地主要呈绿色, 说明这13 a来植被覆盖度增加的地区主要在盆地内, 长汀县针对这一水土流失重点区的治理已取得明显成效。而代表植被覆盖度降低的少部分红色斑块主要零散地分布于盆地中, 这主要是盆地中新增的建筑用地导致了植被覆盖度的下降。还有些零星的红色斑块则是由于火烧山导致的地表裸露而造成的植被覆盖度降低。

从遥感生态指数RSEI的变化[图4 (b)]可以直观看出, 生态指数的变化与植被覆盖度的变化在空间位置上高度吻合。凡是植被覆盖度增加的位置, 生态指数也增加, 同样表现为绿色; 而植被覆盖度下降的地方, 生态指数也同样表现为下降的红色。可见植被覆盖度与生态指数有密不可分的关系, 其变化直接影响着区域的生态质量。

从变化检测的统计结果来看（表4）, 2001— 2013年间, 长汀县生态质量等级下降地区的面积为108.31 km² , 仅占总面积的3.5%, 而生态质量上升地区的面积则达1 849.06 km² , 占到了59.74%, 表明长汀县生态质量有了较明显的上升。

表4 生态变化检测（2001— 2013年） Table 4 Change detection result of graded RSEI between 2001 and 2013

2.5 植被覆盖度与生态变化关系分析

为进一步验证植被覆盖度与生态的关系, 定量分析植被覆盖度的生态改善效应, 分别将各年份的植被覆盖度影像和生态指数影像做回归分析。在影像上按11× 11的网格随机采样, 每个年份各采28 000个样本, 大量样本可以保证统计结果的客观性和代表性。由于RSEI与FVC数据的表示形式不一致, 因此首先将RSEI数据转换为百分比形式再与FVC进行回归分析（图5）。

	Figure Option View Download New Window
	图5 植被覆盖度（FVC）与生态指数关系的回归分析注：方程通过0.1%的显著性检验。Fig. 5 Regression analysis of the relationship between FVC and RSEI

由以上各年份回归模型来看, 植被覆盖度FVC和生态指数RSEI呈显著的线性正相关关系, 相关系数皆大于0.9, 说明植被覆盖度的升高会明显地提升区域的生态质量。从回归方程来看, 在长汀县, 植被覆盖度每提升10%, 至少可以使生态质量相应地提升10%。因此可以看出植被覆盖度具有重要的生态提升效应, 在改善区域生态质量上具有举足轻重的作用。

3 结论与讨论

1）通过对比LSMA和LSMA_M两种模型反演植被覆盖度的精度, 发现LSMA_M模型的精度更高, 其相关系数R值比LSMA模型高出0.173, 均方根误差减小了近50%。这是因为基于增加山地指数波段的LSMA_M模型能够有效削弱阴阳坡植被光谱的差异, 并可以降低阴坡植被与低反照地物的混淆度, 从而有效地提高了植被覆盖度的反演精度, 因此更适合于山地的植被覆盖度反演。

2）RSEI可以有效地揭示长汀水土流失区的生态状况, 但由于RSEI的4个生态因子中含有NDVI, 因此在研究RSEI与植被覆盖度FVC的关系时, 不应再采用基于NDVI的植被覆盖度模型来反演FVC, 以避免二者互为相关。本次研究使用独立于NDVI的LSMA_M模型来计算FVC, 使得揭示的RSEI与FVC的关系更为客观。

3）长汀县水土流失治理工作在所研究的13 a里已取得明显成效, 植被覆盖度有了显著的上升, 从2001年的75.1%上升到2013年的86.5%。植被覆盖度增加的区域主要集中在河田盆地内。长汀县植被覆盖度的提高也产生了明显的生态效应, 其遥感生态指数RSEI从2001年的0.750提高到2013年的0.787。定量回归分析表明, 在长汀植被覆盖度与生态指数呈线性正相关关系, 植被覆盖度每增加10%, 生态指数最少可增加10%, 在改善区域生态质量上具有明显的效应。

The authors have declared that no competing interests exist.

参考文献

View Option

[1]	孙红雨, 王长耀. 中国地表植被覆盖变化及其与气候因子关系[J]. 遥感学报, 1998, 2(3): 204-210. [SUN Hong-yu, WANG Chang-yao. Analysis of the vegetation cover change and the relationship between NDVI and environmental factors by using NOAA time series data. Journal of Remote Sensing, 1998, 2(3): 204-210. ] [本文引用:1] [CJCR: 1.077]
[2]	Cosh M H, Brutsaert W. Microscale structural aspects of vegetation density variability[J]. Journal of Hydrology, 2003, 276(1): 128-136. [本文引用:1] [JCR: 2.693]
[3]	徐涵秋, 何慧, 黄绍霖. 福建省长汀县河田水土流失区植被覆盖度变化及其热环境效应[J]. 生态学报, 2013, 33(10): 2954-2963. [XU Han-qiu, HE Hui, HUANG Shao-lin. Analysis of fractional vegetation cover change and its impact on thermal environment in the Hetian basinal area of County Changting, Fujian Province, China. Acta Ecologica Sinica, 2013, 33(10): 2954-2963. ] [本文引用:1]
[4]	Ringrose S, Matheson W, Wolski P, et al. Vegetation cover trends along the Botswana Kalahari transect[J]. Journal of Arid Environments, 2003, 54(2): 297-317. [本文引用:1] [JCR: 1.822]
[5]	Marticorena B, Bergametti G, Gillette D, et al. Factors controlling threshold friction velocity in semiarid and arid areas of the United States[J]. Journal of Geophysical Research: Atmospheres, 1997, 102(D19): 23277-23287. [本文引用:1]
[6]	De Roo A P J, Wesseling C G, Ritsema C J. LISEM: A single-event physically based hydrological and soil erosion model for drainage basins. I: Theory, input and output[J]. Hydrological Processes, 1996, 10(8): 1107-1117. [本文引用:1]
[7]	Wang G, Wente S, Gertner G Z, et al. Improvement in mapping vegetation cover factor for the universal soil loss equation by geostatistical methods with Land sat thematic mapper images[J]. International Journal of Remote Sensing, 2002, 23(18): 3649-3667. [本文引用:1]
[8]	詹小国, 王平. 基于RS和GIS的三峡库区水土流失动态监测研究[J]. 长江科学院院报, 2001, 18(2): 41-44. [ZHAN xiao-guo, WANG Ping. Research on dynamical supervision design of water and soil loss in Three Gorges reservoir area based on RS and GIS. Journal of Yangtze River Scientific Research Institute, 2001, 18(2): 41-44. ] [本文引用:1] [CJCR: 0.515]
[9]	Sellers P J, Rand all D A, Collatz G J, et al. A revised land surface parameterization (SiB2) for atmospheric GCMs. Part I: Model formulation[J]. Journal of Climate, 1996, 9(4): 676-705. [本文引用:1] [JCR: 4.904]
[10]	王智, 师庆三, 王涛, 等. 1982—2006 年新疆山地-绿洲-荒漠系统植被覆盖变化时空特征[J]. 自然资源学报, 2011, 26(4): 609-618. [WANG Zhi, SHI Qing-san, WANG Tao, et al. Spatial-temporal characteristics of vegetation cover change in mountain-oasis-desert system of Xinjiang from 1982 to 2006. Journal of Natural Resources, 2011, 26(4): 609-618. ] [本文引用:1] [CJCR: 1.841]
[11]	江洪, 王钦敏, 汪小钦. 福建省长汀县植被覆盖度遥感动态监测研究[J]. 自然资源学报, 2006, 21(1): 126-132. [JIANG Hong, WANG Qin-min, WANG Xiao-qin. Dynamic monitoring of vegetation fraction by remote sensing in Changting County of Fujian Province. Journal of Natural Resources, 2006, 21(1): 126-132. ] [本文引用:1] [CJCR: 1.841]
[12]	张清春, 刘宝元, 翟刚. 植被与水土流失研究综述[J]. 水土保持研究, 2002, 9(4): 96-101. [ZHANG Qing-chun, LIU Bao-yuan, ZHAI Gang. Review on relationship between vegetation and soil and water loss. Research of Soil and Water Conservation, 2002, 9(4): 96-101. ] [本文引用:2] [CJCR: 1.173]
[13]	范一大, 史培军, 刘三超. 我国北方沙尘暴与植被覆盖度关系研究[J]. 自然灾害学报, 2010, 19(6): 1-7. [FAN Yi-da, SHI Pei-jun, LIU San-chao. Relationship between dust storm disaster and vegetation fraction in northern China. Journal of Natural Disasters, 2010, 19(6): 1-7. ] [本文引用:1] [CJCR: 1.081]
[14]	长汀县地方志编撰委员会. 长汀县志[M]. 北京: 生活图书新知三联书店, 1993. [Compilation Committee of Changting Chorography. County annals of Changting. Beijing: Life-Studying-Information Triple Bookstore, 1993. ] [本文引用:1]
[15]	Chand er G, Markham B. Revised Land sat-5 TM radiometric calibration procedures and postcalibration dynamic ranges[J]. IEEE Transactions on Geoscience and Remote Sensing, 2003, 41(11): 2674-2677. [本文引用:1] [JCR: 2.933]
[16]	Chand er G, Markham B L, Helder D L. Summary of current radiometric calibration coefficients for Land sat MSS, TM, ETM+, and EO-1 ALI sensors[J]. Remote Sensing of Environment, 2009, 113(5): 893-903. [本文引用:1] [JCR: 4.769]
[17]	Charvz Jr P S. Image-based atmospheric corrections—Revisited and revised[J]. Photogrammetric Engineering and Remote Sensing, 1996, 62(9): 1025-1036. [本文引用:1] [JCR: 2.071]
[18]	Small C, Lu J W T. Estimation and vicarious validation of urban vegetation abundance by spectral mixture analysis[J]. Remote Sensing of Environment, 2006, 100(4): 441-456. [本文引用:1] [JCR: 4.769]
[19]	Small C. Estimation of urban vegetation abundance by spectral mixture analysis[J]. International Journal of Remote Sensing, 2001, 22(7): 1305-1334. [本文引用:1]
[20]	Lee S, Lathrop R G. Sub-pixel estimation of urban land cover components with linear mixture model analysis and Land sat Thematic Mapper imagery[J]. International Journal of Remote Sensing, 2005, 26(22): 4885-4905. [本文引用:2]
[21]	Wu C, Murray A T. Estimating impervious surface distribution by spectral mixture analysis[J]. Remote sensing of Environment, 2003, 84(4): 493-505. [本文引用:3] [JCR: 4.769]
[22]	段四波, 阎广建. 山区遥感图像地形校正模型研究综述[J]. 北京师范大学学报: 自然科学版, 2007, 43(3): 362-366. [DUAN Si-bo, YAN Guang-jian. A Review of models for topographic correction of remotely sensed images in mountainous area. Journal of Beijing Normal University: Natural Science, 2007, 43(3): 362-366. ] [本文引用:1]
[23]	闻建光, 柳钦火, 肖青, 等. 复杂山区光学遥感反射率计算模型[J]. 中国科学D辑, 2008, 38(11): 1419-1427. [YAN Jan-guang, LIU Qin-huo, XIAO Qing, et al. Optical remote sensing reflectance calculation model of complicated mountain area. Science in China Series D: Earth Sciences, 2008, 38(11): 1419-1427. ] [本文引用:1]
[24]	吴志杰, 徐涵秋. 卫星影像数据构建山地植被指数与应用分析[J]. 地球信息科学学报, 2011, 13(5): 656-664. [WU Zhi-jie, XU Han-qiu. A new index for vegetation enhancements of mountainous regions based on satellite image data. Journal of Geo-Information Science, 2011, 13(5): 656-664. ] [本文引用:1]
[25]	徐涵秋. 水土流失区生态变化的遥感评估[J]. 农业工程学报, 2013, 29(7): 91-97. [XU Han-qiu. Assessment of ecological change in soil loss area using remote sensing technology. Transactions of the CSAE, 2013, 29(7): 91-97. ] [本文引用:1] [CJCR: 2.121]
[26]	国家环保总局. 中华人民共和国环境保护行业标准(试行) HJ/T192—2006 [S]. 北京: 中国环境科学出版社, 2006. [State Environmental Protection Administration. The environmental protection industry stand ards of the People's Republic of China (Trial) HJ/T192-2006. Beijing: China Environmental Science Press, 2006. ] [本文引用:1]
[27]	徐涵秋. 区域生态环境变化的遥感评价指数[J]. 中国环境科学, 2013(5): 889-897. [XU Han-qiu. A remote sensing index for assessment of regional ecological changes. China Environmental Science, 2013(5): 889-897. ] [本文引用:1]
[28]	Goward S N, Xue Y, Czajkowski K P. Evaluating land surface moisture conditions from the remotely sensed temperature/vegetation index measurements: an exploration with the simplified simple biosphere model[J]. Remote Sensing of Environment, 2002, 79(2): 225-242. [本文引用:1] [JCR: 4.769]
[29]	Crist E P. A TM tasseled cap equivalent transformation for reflectance factor data[J]. Remote Sensing of Environment, 1985, 17(3): 301-306. [本文引用:1] [JCR: 4.769]
[30]	NASA. Land sat 7 Science Data Users Hand book [EB/OL]. http://landsathandbook.gsfc.nasa.gov,2012-09-05. [本文引用:1]
[31]	Chand er G, Markham B L, Helder D L. Summary of current radiometric calibration coefficients for Land sat MSS, TM, ETM+, and EO-1 ALI sensors[J]. Remote Sensing of Environment, 2009, 113(5): 893-903. [本文引用:1] [JCR: 4.769]
[32]	Nichol J. Remote sensing of urban heat island s by day and night[J]. Photogrammetric Engineering and Remote Sensing, 2005, 71(5): 613-621. [本文引用:1] [JCR: 2.071]
[33]	Sobrino J A, Jiménez-Muñoz J C, Paolini L. Land surface temperature retrieval from Land sat TM 5[J]. Remote Sensing of Environment, 2004, 90(4): 434-440. [本文引用:1] [JCR: 4.769]
[34]	Rikimaru A, Roy P S, Miyatake S. Tropical forest cover density mapping[J]. Tropical Ecology, 2002, 43(1): 39-47. [本文引用:1] [JCR: 0.551]
[35]	Xu H Q. A new index for delineating built-up land features in satellite imagery[J]. International Journal of Remote Sensing, 2008, 29(14): 4269-4276. [本文引用:1]

1998

0.0

1.077

孙红雨, 王长耀. 中国地表植被覆盖变化及其与气候因子关系[J]. 遥感学报, 1998, 2(3): 204-210.

[SUN

Hong-yu

, WANG

Chang-yao.

Analysis of the vegetation cover change and the relationship between NDVI and environmental factors by using NOAA time series data.

Journal of Remote Sensing, 1998, 2(3): 204-210. ]

By using NOAA temporal series data during the period 1985一1990 (69 months),this study examines the Vegetation-Cover change with time and space in China. Furthermore, the ana1ysis about correlation . between Vegetation-Cover change and environmental factors is also accomplished by combining simu1taneous average temperature and precipitation .data in each month. The methods employedinclude Sample Transect Ana1ysis, Partia1 Correlation Ana1ysis, and GIS Spatia1 Calcu1ation.From the above ana1ysis, we can draw the following conclusions:1. The vegetation index generally represents the situation of the superficial vegetation cover. Heat and water condition determine the whole spatia1 distribution pattem of vegetation index and build the regional attribute in longitude and latitude. The partia1 fluctuation of the vegetation index is essentia11y affected by landscape and soil conditions. This defines its regiona1 distribution pattem.2. Lying on the same longitude, the vegetation index obviously. corresponds to the seasonal movement law, and span of the movement in time and space is much greater than that along the same latitude. The type and ways of the green wave movement synchronize with the seasonal movement of the continenta1 main controlling air mass and are influenced by it.3.The vegetation index increases and decreases with the warming and cooling of the climate.So it has typically seasona1 change characteristics. Along the same longitude, the higher the latitude is,the greater the scope of seasonal change will be. In summer there is no evident difference of vegetationindex between southem and northem region. But in winter，this difference is quite obvious.4. In a large range, the vegetation index change is driven by water and heat conditions. In eastem moist monsoon zone of China,along. the same longitude,the temperature is the most important influential parameter.The correlation between the vegetation index change and monthly average temperature is obvious.So the heat is key natural limited factor for the growth of the vegetation. Yet,in northem China, along the same latitude, from east to west，it sequentially covers the humid monsoon zone, semi-arid and arid desert zone. The relationship between the vegetation index change and monthly average precipitation is well correlated positively. It proves that in a dry area, water source is the essential limited condition for the growth of the vegetation.

本文利用1985-1990年连续69个月的NOAA时间序列数据, 进行中国植被覆盖变化的空间, 以及时间序列分析, 并且结合同期的月平均气温、降水数据, 进行植被覆盖变化与气候因子相关性分析。该文证实了在中国植被覆盖随时间的推移规律, 空间分布规律, 以及植被覆盖变化与气温、降水的定量关系。

... 由于植被是生态系统的重要组成成分,植被覆盖度的变化是地球内、外部作用与人类改造相互作用的结果^[1],对区域生态系统具有重要的影响 ...

2003

2.693

0.0

... 因此围绕植被覆盖度及其生态效应的研究长期以来一直为科学界所重视,例如植被覆盖度与土壤湿度、地表温度和气候变化^[2,3,4]等关系的研究 ...

2013

0.0

徐涵秋, 何慧, 黄绍霖. 福建省长汀县河田水土流失区植被覆盖度变化及其热环境效应[J]. 生态学报, 2013, 33(10): 2954-2963.

[XU

Han-qiu

, HE

Hui

, HUANG

Shao-lin.

Analysis of fractional vegetation cover change and its impact on thermal environment in the Hetian basinal area of County Changting, Fujian Province, China.

Acta Ecologica Sinica, 2013, 33(10): 2954-2963. ]

Soil erosion has caused severe ecological problems in China. Being a typical reddish soil erosion region in southtern China, the hilly areas of the Hetian Basin in County Changting, Fujian province, have been called flame mountainous area, due to lack of vegetation covers, extensively exposed bare soils, and hot weathers. After more than 20 years of soil-loss treatment, the local ecosystem and environment have been greatly improved. This study used remote sensing techniques and four scenes of Landsat Thematic Mapper (TM) images, enquired in 1988, 1998, 2004, and 2010, to study the fractional vegetation cover changes and accompanied land surface temperature variations in the Hetian basinal area during the period from 1988 to 2010. Two remote-sensing based models for estimating fractional vegetation cover have been compared in order to select a suitable model for the retrieval of the fractional vegetation cover in the study area. The two models compared are the Carlson-Ripley and Gutman-Ignatov models. The comparative analysis confirmed that the Carlson-Ripley model had a higher accuracy than the Gutman-Ignatov model as it has a lower systematic error value. Therefore, the Carlson-Ripley model was used to retrieve the fractional vegetation cover for the study area. While, the land surface temperature change during the study period was evaluated through the temperature difference between the areas inside and outside the basin in each study year. The results indicate that the 22-year effort for the treatment of the soil loss in the Hetian basinal area has resulted in the notably increase in fractional vegetation cover in the area, averagely from 48.83% in 1988 to 59.78% in 2010. Meanwhile, the difference in land surface temperature between the areas inside and outside the basin has also been reduced from 3.39℃ in 1988 to 2.48℃ in 2010, almost 1℃ declined. This suggests a lowering of the temperature in the area inside the basin. Obviously, the increase in fractional vegetation cover was accompanied with the lowering of regional temperature. Spatial and temporal analysis indicates that the changes in both fractional vegetation cover and land surface temperature mainly took place in the areas inside the Hetian Basin, while the areas outside the basin were rather consistent either in fractional vegetation cover or in land surface temperature. Quantitatively analysis suggests that each 10% increment in fractional vegetation cover would result in the lowering of local land surface temperature approximately by 0.6 to 1℃ based on regression models of the four study years.

南方红壤典型水土流失区——福建省长汀县河田盆地区的大片山地曾由于植被稀少、山岭裸露、气候炎热,而被称之为"火焰山"。通过20多年的水土流失治理,该区的生态环境有了很大的改善。利用遥感技术,计算了长汀县河田盆地区在1988-2010年间的植被覆盖度变化及其伴随的地表温度变化。结果表明,通过近22a的水土流失治理,该区的平均植被覆盖度有了明显的升高,从1988年的48.83%上升到2010年的59.78%;而同期盆地内外的温度差异也随之下降,盆地裸露区和周边地区的温差降低了近1 ℃。定量研究表明,河田盆地区的植被覆盖度每增加10%,可降低地表温度0.6-1℃。

... 因此围绕植被覆盖度及其生态效应的研究长期以来一直为科学界所重视,例如植被覆盖度与土壤湿度、地表温度和气候变化^[2,3,4]等关系的研究 ...

2003

1.822

0.0

... 因此围绕植被覆盖度及其生态效应的研究长期以来一直为科学界所重视,例如植被覆盖度与土壤湿度、地表温度和气候变化^[2,3,4]等关系的研究 ...

1997

0.0

... 除此之外,植被覆盖度还是气候模型、水文模型、水土流失土壤侵蚀模型以及生态系统模型的重要输入参数^[5,6,7,8,9] ...

1996

0.0

... 除此之外,植被覆盖度还是气候模型、水文模型、水土流失土壤侵蚀模型以及生态系统模型的重要输入参数^[5,6,7,8,9] ...

2002

0.0

... 除此之外,植被覆盖度还是气候模型、水文模型、水土流失土壤侵蚀模型以及生态系统模型的重要输入参数^[5,6,7,8,9] ...

2001

0.0

0.515

詹小国, 王平. 基于RS和GIS的三峡库区水土流失动态监测研究[J]. 长江科学院院报, 2001, 18(2): 41-44.

[ZHAN

xiao-guo

, WANG

Ping

. Research on dynamical supervision design of water and soil loss in Three Gorges reservoir area based on RS and GIS. Journal of Yangtze River Scientific Research Institute, 2001, 18(2): 41-44. ]

The process and method on dynamic monitoring of water and soil loss for the Thre ee Gorges reservoir area using RS and GIS are introduced.Firstly,the classifica tion of land and the calculation of available coverage of plant cover are perfor med by remote sensing image handling software ERDAS,and then,the space analyses of varieties of water and soil loss are done by the ARCGRID of geographical i nformation system software ARC/INFD,and finally,the intensities and the areas of water and soil loss for every counties or small valleys for different intensitie s are estimated respectively.

介绍了利用遥感和地理信息系统栅格技术进行三峡库区水土流失动态监测的过程和方法。首先利用遥感图像处理软件ERDAS进行了土地利用的分类和植被覆盖度的计算，然后利用地理信息系统软件ARC/INFO中GRID工具进行各类水土流失因素图的空间分析，估算各县或小流域的水土流失强度，并计算各种强度下的水土流失面积。　

... 除此之外,植被覆盖度还是气候模型、水文模型、水土流失土壤侵蚀模型以及生态系统模型的重要输入参数^[5,6,7,8,9] ...

1996

4.904

0.0

... 除此之外,植被覆盖度还是气候模型、水文模型、水土流失土壤侵蚀模型以及生态系统模型的重要输入参数^[5,6,7,8,9] ...

2011

0.0

1.841

王智, 师庆三, 王涛, 等. 1982—2006 年新疆山地-绿洲-荒漠系统植被覆盖变化时空特征[J]. 自然资源学报, 2011, 26(4): 609-618.

[WANG

Zhi

, SHI

Qing-san

, WANG

Tao

, et al.

Spatial-temporal characteristics of vegetation cover change in mountain-oasis-desert system of Xinjiang from 1982 to 2006.

Journal of Natural Resources, 2011, 26(4): 609-618. ]

Dynamics of vegetation cover plays an important role in the coexistence and balance evolution mechanism of mountain-oasis-desert system (MODS) in Xinjiang. Based on the satellite-derived NDVI data, produced by GIMMS (Global Inventory Monitoring and Modeling Studies) group, derived from the NOAA/AVHRR (National Oceanic and Atmospheric Administration/the Advanced Very High Resolution Radiometer) land dataset, the fractional vegetation cover ( fvc ) was calculated at a spatial resolution of 8 km×8 km and at a 15-day interval, for the period January 1982 to December 2006 and then the annual maximum fvc was obtained by the maximum value composites (MVC) method. Under the basic pattern of MODS in Xinjiang, the fvc spatial-temporal dynamics characteristics of trends, rate, amplitude, and variability of 25 years in Xinjiang were analyzed by dividing the whole region into four subregions in terms of regional differentiation, and mountain, oasis, desert system in every subregion by altitude and landuse. The results showed that the fvc dynamics increased significantly in general, but fluctuated remarkably in the period from 1982 to 1995 and relatively stable from 1996 to 2006, varied significantly with different subregions, and subsystems of mountain, oasis and desert, in which the vegetation cover of oasis increased mostly rapidly, but desert decreased most obviously. On the amplitude and variability, the eastern and southern subregions exceed the northern and Yili subregions, desert > mountain > oasis, and the lower fvc is larger than the higher fvc . The significant improved subregions are located in oasis, its surrounding areas and part of mountains, but the degenerated subregions are mostly situated in desert of eastern and southern Xinjiang. There are different change characteristics in different fvc subregions, that is, the vegetation degenerated in lower fvc subregions and improved in higher fvc subregions.

1. Ministry of Education Key Laboratory of Oasis Ecology of Xinjiang, Urumqi 830046, China; 2. a. Institute of Arid Eco-environmental Sciences, b. College of Resources and Environmental Sciences, c. College of Geology and Exploration Engineering, Xinjiang University, Urumqi 830046, China

植被覆盖变化在新疆山地-绿洲-荒漠系统(MODS)共存、平衡演变机制中扮演着重要角色。基于1982—2006年GIMMS/NDVI数据,提取新疆地区年最大植被覆盖指数( fvc ),针对新疆MODS的基本格局,通过先地理分区再因海拔而异讨论,划分出山地、绿洲、荒漠子系统,对比分析不同分区不同子系统下 fvc 变化趋势、变率、变幅、变异性等时空动态特征。结果表明:新疆植被覆盖总体呈显著上升趋势,可分为1982—1995年的剧烈波动期和1996—2006年的相对平稳期,不同分区和子系统差异显著;在变幅和变异性上,东、南疆地区>北疆、伊犁地区,荒漠>山地>绿洲,植被覆盖度低的地区>植被覆盖度高的地区;显著改善的地区位于绿洲内部、绿洲的周边地区及部分山地,显著退化的地区大多位于东、南疆的荒漠区;不同植被覆盖度下植被覆盖变化具有不同特征,植被覆盖度低的地区,植被退化趋势明显,植被覆盖度高的地区,植被改善趋势明显。

... 特别是植被覆盖度作为直观描述陆地植被生长状况的综合量化指标,在生态环境综合评价、植被变化监测、水土保持以及干旱、沙尘暴^{[10,11,12,13]}等自然灾害监测中得到了广泛的应用 ...

2006

0.0

1.841

江洪, 王钦敏, 汪小钦. 福建省长汀县植被覆盖度遥感动态监测研究[J]. 自然资源学报, 2006, 21(1): 126-132.

[JIANG

Hong

, WANG

Qin-min

, WANG

Xiao-qin.

Dynamic monitoring of vegetation fraction by remote sensing in Changting County of Fujian Province.

Journal of Natural Resources, 2006, 21(1): 126-132. ]

Vegetation index(VI)and the transformation method of VI are two key factors in estimating vegetation fraction by Remote Sensing.Based on the common VI and the FCD(Forest Canopy Density)Model principle put forward by ITTO(International Tropical Timber Organization),VBSI vegetation index was proposed in this paper that can decrease the interferences from mountain shadows,soil background,rocks and building.It is expressed as VBSI=f(VI,BI,SI)=VI+n·BI+SI which is suitable for estimating vegetation coverage in mountainous areas.Furthermore,mixed pixel model is one of the good transformation methods from VI to vegetation fraction,which can apply to images from different phases.The expression is:fc=(S-Snon)/(Sveg-Snon).Vegetation fraction estimation in 1994-2003 was processed with ERDAS software system aimed at Chang-ting County,Fujian Province,China.The results show that(1)With VBSI,the error of vegetation fraction resulted from the area of image shadow can be decreased,and the error is 50% of that from Normalized Difference Vegetation Index(NDVI).The overall accuracy from the vegetation fraction estimation based on VBSI and mixed pixel model is more than 80%,which meets the requirements for vegetation fraction dynamic monitoring in region.(2)The coverage of high vegetation in the study area increased by 150.47km 2 ,or an increase of 4.9% of the total land area from 1994 to 2003.This is correlative with the work of soil and water conservation in recent years in Changting County.

Key Laboratory of Data Mining & Information Sharing of Ministry of Education,Spatial Information Research Center of Fujian Province,Fuzhou University,Fuzhou 350002,China

遥感估算植被覆盖度的关键在于:一是植被指数的选择,二是植被指数转换方法。针对中国南方多山的特点,论文提出了能较好地削弱影像中山体阴影、土壤背景、岩石、建筑用地等地物对植被覆盖度信息干扰的复合植被指数V BSI;在植被指数转换方法上,采用混合像元法,对福建省长汀县1994年和2003年的遥感影像进行了植被覆盖度的估算。研究结果表明:采用VBSI进行植被覆盖度的估算,影像阴影信息的干扰作用可以被削减为NDV I的50%,基于VBSI的混合像元法估算植被覆盖度的总体精度达到80%以上;动态监测表明,从1994至2003年研究区高植被覆盖面积增加了150.47km 2 ,占国土面积的比例提高了4.9个百分点,据调查,这与近年来长汀县加大水土流失治理力度有着重要关系。

2002

0.0

1.173

... 3 植被覆盖度模型的确定当前,植被覆盖度（Fractional Vegetation Cover, FVC）的估算以线性光谱混合分析模型和基于NDVI植被指数的像元二分模型最为常用^[12] ...

2010

0.0

1.081

1993

0.0

... 长汀县在地形上呈四周隆起、中部下凹的盆地特征,这一盆地即为水土流失严重的河田盆地区^[14] ...

2003

2.933

0.0

... 为减少不同时间的影像在大气、光照等各方面的差异与影响,采用Chander等^[15,16]的模型和Chavez^[17]的大气校正参数,将Landsat TM/ETM+原始影像的DN值转换为传感器处反射率来进行辐射校正,公式如下： ...

2009

4.769

0.0

1996

2.071

0.0

2006

4.769

0.0

... 在混合像元分解法中,线性光谱混合分析（Linear Spectral Mixture Analysis, LSMA）以其物理意义明显且相对简单而得到广泛应用^[18,19,20],它将像元在某一光谱波段的反射率视为由构成像元的各终端单元（Endmember）的反射率（亮度值）以其所占像元面积比例为权重系数的线性组合,公式及其约束条件如下^[21]： ...

2001

0.0

2005

0.0

... 在线性光谱分析的端元确定中,以V-L-H-S（植被-低反照地物-高反照地物-裸露土壤）的4端元模型^[20]最为常用 ...

2003

4.769

0.0

... 在理想的线性光谱混合分析模型中,各端元应该分布在三角特征空间的顶点区域,每个顶点区域的散点代表一种光谱端元,越靠近边缘,像元越纯净^[21] ...

... 精度验证采用基于高分辨率影像的亚像元对比法^[21],该方法不仅能检验精度,而且还可以将由模型求出的植被覆盖度和高分辨率影像所代表的实际植被覆盖度关联起来 ...

2007

0.0

... 由于研究区的植被主要分布在山地上,地形起伏的阴影会对阴坡的植被覆盖度估算带来误差^[22,23] ...

2008

0.0

闻建光, 柳钦火, 肖青, 等. 复杂山区光学遥感反射率计算模型[J]. 中国科学D辑, 2008, 38(11): 1419-1427.

[YAN

Jan-guang

, LIU

Qin-huo

, XIAO

Qing

, et al.

Optical remote sensing reflectance calculation model of complicated mountain area. Science in China Series D:

Earth Sciences, 2008, 38(11): 1419-1427. ]

提出一种复杂山区光学遥感反射率计算模型, 以有效消除地形影响, 计算地表反射率. 通过推导地形影响下的方向反射, 结合辐射传输方程, 考虑地表反射目标接收的太阳直射辐射为方向-方向二向反射, 反射目标接收的大气漫散射和周围地形背景反射辐射为半球-方向二向反射的综合过程, 发展了基于方向反射的地形影响消除和地表反射率计算模型. 利用GOMS模型模拟数据及真实光学遥感数据对模型应用能力进行了评价, 结果表明此模型具有有效的地形消除能力以及大气校正能力, 模型的建立为后续山区定量遥感的应用提供技术支持.

... 由于研究区的植被主要分布在山地上,地形起伏的阴影会对阴坡的植被覆盖度估算带来误差^[22,23] ...

2011

0.0

吴志杰, 徐涵秋. 卫星影像数据构建山地植被指数与应用分析[J]. 地球信息科学学报, 2011, 13(5): 656-664.

[WU

Zhi-jie

, XU

Han-qiu.

A new index for vegetation enhancements of mountainous regions based on satellite image data.

Journal of Geo-Information Science, 2011, 13(5): 656-664. ]

In this study, a normalized difference mountain vegetation index (NDMVI) is proposed using Landsat TM/ETM+ images. The new index aims to correct terrain influence on the vegetation measurements of the NDVI (Normalized Difference Vegetation Index). Through analyzing the different spectral reflectance characteristics of vegetation between sunny and shaded slopes in the complex mountainous region, we developed a new index to introduce two parameters to adjust values of the near-infrared and red bands of the Landsat TM/ETM+ image. The newly developed index has advantages over the NDVI in correcting terrain impact and over the other similar terrain correction vegetation indices in that it does not need other supplement data such as DEM for the model. The study shows that the NDMVI has nearly no significant correlation with cosine value of solar incidence angle and is not sensitive to the terrain changes. Therefore, this new index can more effectively suppress the influence to vegetation observations due to terrain difference better than does the NDVI. In addition, the NDMVI has a wider dynamic range than the NDVI does, and thus, could distinguish different land cove types more effectively than the NDVI. We concluded that, compared with the C-correction model, the new index can achieve a better performance because it has avoided the over-correction bias.

1. Department of Resources Engineering, Longyan University, Longyan 364012, China; 2. College of Environment and Resources, Fuzhou University, Fuzhou 350108, China

本研究以Landsat影像为数据源,在分析复杂地形山地植被在阳坡和阴坡反射率差异特征的基础上,提出一种归一化差值山地植被指数NDMVI (Normalized Difference Mountain Vegetation Index)。该指数模型无需辅助数据(如DEM)的支持,通过同时降低近红外波段(TM4)和红光波段(TM3)反射率的方法来消除或抑制地形的影响,具有较强的可操作性。研究表明:NDMVI与太阳入射角余弦值(cos i)的相关性相当小,对地形起伏变化表现不敏感,可有效消除或抑制地形的影响;比NDVI值动态变化范围更宽,对地物有更强的遥感识别能力;该模型抑制地形影响的效果比用C校正模型的效果更佳,不会出现过度校正的问题。

... 对此,在原始影像的6个多光谱波段中引入能够削弱阴影的山地指数NDMVI（Normalized Difference Mountain Vegetation Index）^[24]作为辅助波段,以削弱阴坡对线性光谱分解结果的影响,提高分解结果的精度 ...

2013

0.0

2.121

徐涵秋. 水土流失区生态变化的遥感评估[J]. 农业工程学报, 2013, 29(7): 91-97.

[XU

Han-qiu.

Assessment of ecological change in soil loss area using remote sensing technology.

Transactions of the CSAE, 2013, 29(7): 91-97. ]

Abstract: Severe soil loss has caused ecological degradation for the global ecosystem, thus it is a major problem facing the world today. Timely and fast monitoring ecological changes in soil loss regions has become an increasing concern. This paper develops a remote sensing assessment method of soil erosion-induced changes in regional ecological quality based ecological index (RSEI). The proposed index combines four indicators from existing remote-sensing indices/components to represent greenness, dryness, wetness and heat, which are the important ecological indicators frequently used in assessing regional ecological status. The four remote-sensing indices/components are the normalized difference vegetation index (NDVI), soil index (SI), wetness component of the tasseled cap transformation (Wet), and land surface temperature (LST). The principal component analysis (PCA) was utilized to compress the four indicators into one index - RSEI, in order to assess overall ecological status. The new index, RSEI, was thus constructed using the first component as it was proved to have effectively combined the most information of the four indicators. The application of the RSEI in Hetian basin area in Changting county of Fujian province, one of the most serious reddish soil erosion areas in southern China, showed that the RSEI can quantitatively assess the ecological effects of soil loss treatment in the area and easily detect spatial and temporal changes of the ecological quality through a time period from 1988 to 2010. The application utilized three Landsat TM images of 1988, 2004 and 2010. The four indicators (NDVI, SI, Wet and LST) of each year were retrieved from the images and then combined through the PCA transform to form the RSEIs for the study years. The RSEI-based analysis indicated that after a more than 20 years fight for soil loss in the area by the local people and government, the ecological quality of the area has been significantly improved. This is suggested by an increase in the mean RSEI value from 0.5 in 1988 to 0.59 in 2010, accompanied by a decrease in low level RSEI area from 66.1% to 47.7%, and an increase of high level RSEI area from 33.9% to 52.3% in this duration. Quantitative analysis reveals that the greenness indicator represented by NDVI contributes most to the RSEI change among the four indicators used for generating the index. This suggests that the biological restore of soil erosion areas by planting tree and grass is an effective way to soil-erosion treatment for Hetian basin.

水土流失是世界面临的一个严峻问题，它已给全球的生态造成了严重的威胁，及时快速地监测水土流失区的生态变化已显得尤为重要。为此，该文提出遥感生态指数(RSEI)来监测水土流失区的生态变化。该指数选取了绿度、湿度、热度和干度作为四大生态指标，并分别以遥感植被指数、湿度分量、地表温度和土壤-建筑指数为代表。同时采用主成分变换技术来集成各个指标，使各指标的权重是由数据本身的性质来决定，而不是人为的设定。将RSEI应用于福建长汀水土流失区的结果表明，RSEI指数可以定量地评价水土流失区生态修复的效果。数据显示，该区经过20多年的水土流失治理，RSEI生态指数值上升了17%，生态为优良的等级所占的面积比例从33.9%上升到52.3%，总体反映了该区的生态质量有了较明显的提高。

... 4 生态指数计算为了考察植被变化所产生的生态效应,本文选用徐涵秋^[25]新近提出的遥感生态指数（Remote Sensing Ecological Index, RSEI）来对长汀县的生态进行评价 ...

2006

0.0

... 该指数是一综合多种要素的生态综合评价指数,它不仅考虑了植被所代表的绿度变化,还结合了湿度、热度、干度等多种因素,因此可以较全面地考察区域的生态变化,其结果与国家环保部的生态状况指数（EI）^[26]也有较好的可比性^[27] ...

0.0

2002

4.769

0.0

... 绿度指标以植被指数NDVI ^[28]来代表： ...

1985

4.769

0.0

... 湿度指标由缨帽变换中的湿度分量^[29]代表： ...

0.0

... 亮温BT的计算参照Landsat用户手册的模型^[30],再对其进行比辐射率校正^[31,32]： ...

2009

4.769

0.0

... 亮温BT的计算参照Landsat用户手册的模型^[30],再对其进行比辐射率校正^[31,32]： ...

2005

2.071

0.0

... 亮温BT的计算参照Landsat用户手册的模型^[30],再对其进行比辐射率校正^[31,32]： ...

2004

4.769

0.0

2002

0.551

0.0

... 干度指标采用由裸土指数SI ^[34]和建筑指数IBI ^[35]合成得到的干度指数NDSI来代表： ...

2008

0.0

... 干度指标采用由裸土指数SI ^[34]和建筑指数IBI ^[35]合成得到的干度指数NDSI来代表： ...