【作者】 江振蓝；

【导师】 吴嘉平；

【作者基本信息】 浙江大学 ， 农业遥感与信息技术， 2013， 博士

【摘要】 严重的水土流失是生态恶化的集中反映,是我国生态环境最突出的问题之一,已成为生态安全的重要威胁因素。尽管水土流失是生态安全的关键因子,但目前鲜见水土流失与生态安全两者之间的系统研究。本文以浙江省安吉县为例,从水土流失的角度出发,基于水土流失动态监测、生态安全动态分析,以及水土流失对生态安全影响研究这条主线对水土流失时空过程及其生态安全效应进行了较为系统深入的研究。首先,利用修正通用水土流失方程(Revised Universal Soil Loss Equation, RUSLE),综合遥感(Remote Sensing, RS)和地理信息系统(Geographic Information System, GIS),构建基于背景因子和动态因子的水土流失的遥感定量化监测技术,对研究区1985-2008年间的水土流失进行时空动态监测,并利用探索性空间数据分析(Explotory Spatial Data Analysis, ESDA)方法对水土流失的时空分异规律进行分析；其次,以突变理论为基础,构建基于景观尺度的生态安全评价模型,并对研究区1985-2008年的生态安全进行了动态监测；第三,基于前两部分的研究结果,运用地理权重回归(Geographically Weighted Regression, GWR)模型对水土流失和生态安全的关系进行了定量的研究。主要研究结论如下：1、水土流失的遥感定量化监测：基于修正通用水土流失方程,将方程中相关的因子划分为背景因子和动态因子,对研究区1985、1994、2003、2008年的水土流失进行监测。结果表明,安吉县水土流失恶化与恢复过程并存。但是,由于水土流失的恶化速度大于恢复速度,且水土流失恶化区面积始终大于恢复区面积,总体上呈现恶化趋势。2、水土流失的时空分异规律：利用全局Moran’s I和局部空间关联指数Local Indicators of Spatial Association (LISA),从定量的统计学和可视化角度研究水土流失的时空变异规律。结果再次表明,水土流失呈现恶化趋势。水土流失的热点区域,包括严重的水土流失区域和水土流失恶化区不断扩大,且呈现较为集中的态势,研究区受水土流失影响的范围和影响的程度越来越大,水土流失越来越成为制约该区生态环境的主导因素。3、生态安全的时空分异规律：基于压力-状态-响应的理论框架,构建适合景观尺度的生态安全评价指标体系,并利用突变级数法进行研究区生态安全评价。结果表明,在1985-2008年间研究区生态安全以比较安全为主(占全区的55%以上),生态状况总体上较为良好。然而,随着不安全和很不安全区域比重的上升,生态安全性呈现持续的下降趋势。4、水土流失的生态安全效应：在500m×500m格网尺度下,利用地理权重回归(GWR)模型进行1985、1994、2003、2008年水土流失对生态安全及2000-2012年水土流失变化对生态安全变化的影响研究。结果表明,研究区水土流失的变化决定着生态环境的变化。随着时间的推移,研究区水土流失的持续恶化使得生态状况变得越来越不安全,水土流失问题已越来越成为该区生态环境最主要的问题。本研究主要创新点：依据RUSLE以及区分背景因子和动态因子,构建了适合区域长期监测的水土流失遥感定量化监测方法,并将探索性空间分析技术引入到水土流失的时空分异规律研究,兼顾了定量的统计学和可视化方法研究水土流失的时空分异规律；利用突变理论进行生态安全评价,最大可能地克服了以往模型权重确定的主观性,使评价结果更具可比性；将地理权重回归模型引入水土流失与生态安全关系研究,为充分体现两者关系的空间异质性和尽可能消除两者关系的空间自相关性提供了可能。受数据和能力的限制,仍有一些理论及技术问题需要进一步的探讨。本文所选择的时空尺度较为单一,致使对水土流失演变及水土流失的生态安全效应的时间和空间特征的分析有待进一步深入,水土流失时空过程及生态安全效应研究的尺度问题,也有待在今后的研究中做进一步的探讨；在基于景观尺度的生态安全评价中,受所获数据的限制,评价指标的选取在很大程度上依赖数据的可获取性,而这些指标是否完善、是否合理,如何与行政尺度或流域尺度评价指标体系进行匹配,有待在今后的研究中进一步深入；GWR模型采用的仅是线性拟合,存在一定的局限性,在今后的研究中,需对其模型进行改进。更多还原

【Abstract】 Severe soil erosion, indicative of the deterioration of eco-environment, has become one of the most critical ecological problems in China and a major threat to the nation’s eco-security. Despite the fact that soil erosion is a key factor of eco-security, there are few systematic reports focusing on uncovering their relationship. With a case study in Anji County, Zhejiang Province, south-eastern China, this work monitored the dynamics of soil erosion, analyzed the varions of eco-security, and assessed the impact of soil erosion on eco-security. Thus, a relatively systematic study was conducted into the spatiotemporal dynamics of soil erosion and their responses to eco-security.Firstly, by using Revised Universal Soil Loss Equation (RUSLE) and by integrating remote sensing (RS) and geographic information system (GIS), a quantitative method that separates background factors and dynamic factors in RUSLE was developed to estimate the spatiotemporal dynamics of soil erosion in the study area during1985-2008. The Exploratory Spatial Data Analysis (ESDA) was used to analyze the soil erosion dynamics; Then, a catastrophe theory was employed to generate a grid scale-based model for evaluating eco-security, which was then used to monitor the eco-security dynamics of the study area during1985-2008; At last, based on above results, a Geographically Weighted Regression (GWR) model was utilized to quantify the relationship between soil erosion and eco-security.The major findings of this study are as follows:1. Quantitative monitoring of soil erosion:On the basis of RUSLE, background and dynamic factors were employed to monitor soil erosion in the study area in1985,1994,2003and2008. The results revealed the co-existence of the deterioration and mitigation of soil erosion. There is, however, an overall trend of erosion deterioration, with both the scale and the range of deterioration exceeding that of mitigation.2. Spatiotemporal dynamics of soil erosion:The Moran’s I and Local Indicators of Spatial Association (LISA) were used to identify the spatial and temporal dynamics of soil erosion from the perspectives of quantitative statistics and visualization. The results revealed a trend of erosion deterioration. The hot spots of soil erosion, including areas with serious erosion and deteriorated erosion, were found expanding and clustering. The scale and degree of soil erosion in the area kept going up and erosion became an increasingly critical factor affecting the local eco-environment.3. Spatiotemporal dynamics of eco-security:A system of eco-security assessment indicators, adapted to the grid scale, was established on the basis of the pressure-state-response model. Then catastrophe theory was employed to evaluate the eco-security of the study area. The results showed that the eco-environment in major parts (accounting for over55%) of the study area was relatively secure, indicative of a relatively good eco-environment. However, with the increase of areas featuring insecure and very insecure eco-environment, the overall eco-security was on a trend of deterioration.4. Eco-security responsive to soil erosion:The Geographically Weighted Regression (GWR) model was used to quantify the relationship between soil erosion and eco-security at500m×500m grid scale. The results showed that the change of erosion regulated the change of eco-environment. Continuous erosion deterioration rendered the eco-environment increasingly less safe in the study area, and erosion gradually became the most important threat to the local eco-environment.New methods developed in this study are as follows:1) by separating background factors from dynamic factors, a quantitative method was developed to monitor soil erosion on a long-term basis. And the Exploratory Spatial Data Analysis (ESDA) was introduced to the study of spatial and temporal dynamics of erosion by integrating quantitative statistics and visualization;2) The use of catastrophe theory in eco-security assessment helped avoid the subjectivity in deciding the weight of different factors in a model and made the assessment results more comparable;3) The introduction of GWR to the analysis of the relationship between soil erosion and eco-security made it possible to fully reveal the spatial heterogeneity of the relationship and to get rid of their spatial autocorrelation at its best.With the limitation of data, further research is still required into some methods and theories:The selection of limited number of temporal/spatial scales caused an inadequate analysis of soil erosion change and the temporal/spatial features of eco-security responsive to soil erosion. Further research is to be conducted about the scales of study on the temporal/spatial dynamics of erosion and the responses of eco-security; the selection of indicators for landscape grid-based assessment of eco-security is mainly subject to data availability. A further study could be done on the reasonableness of the indicators and the methods of matching them up with those used in administrative or watershed scale assessment; As GWR model was established mainly with linear interpolation in this study, a further research could be performed to improve the GWR model.更多还原