黄土高原景观生态格局分析的遥感影像空间分辨率适用性研究

发布时间：2018-04-29 01:35

本文选题：景观生态 + 尺度效应　；参考：《西北大学》2015年硕士论文

【摘要】：地理信息科学中,我们在进行一般地理对象的分析和具体决策实施时,需要各种空间数据作为支撑条件进行综合辅助,要从动态的联结机制中把握各尺度空间信息数据,由此,可以认为尺度问题在地理新信息科学的研究中地位重要。在遥感数据应用于景观生态格局分析方面,很多方面的认知都有不足之处,哪些遥感影像对于哪些类型及地区的景观生态格局的研究最为合适,哪种尺度最佳,就目前而言,这些问题研究相对较少,需要我们进行更深的探讨。本文通过利用landsat8.HJ-1A,MODIS等不同空间分辨率遥感影像数据进行分析、预处理,并对其进行景观生态格局提取,研究不同分辨率下景观生态格局变化规律,试图找出随着分辨率变化下的敏感性因子。本研究选取了陕北黄土高原上的五大实验样区进行研究,分别是彬县长武样区、延川延长样区、米脂绥德样区、洛川黄龙样区、定边靖边样区,对于每个样区来说,分别分析了他们不同空间尺度下的景观格局变化情况,得到了如下几点结论：(1)景观聚集度的高低与景观破碎度没有直接相关性,二者没有明显的此消彼长的联系。因为,在空间分辨率增加同时,斑块密度、边界密度、最大斑块指数都出现了明显的下滑态势,而相应的景观聚集度在空间分辨率发生变化的情况下,也出现了下降。(2)在不同空间分辨率所对应的景观生态格局发面,由不同分辨率遥感影像所产生的景观生态格局具有显著差异,分析结果中也可以清晰地看到种现象。景观生态类型中,受空间分辨率影像相对较大的要属线景观要素,他们都是廊道的典型代表类型,这其中道路和河流等最为明显。本文算选的研究区大多是自然生态景观,线性景观在总体景观中所占比重不大,所以,空间分辨率在这方面尺度效应的影像不是很大。(3)景观格局各指数的尺度效应很明显,尺度效应的差异会随着空间分辨率的变化而发生变化,但是,有个特例情况,Shannon多样性指数,它不存在明显的尺度效应,只与研究区景观分类体系存在相关性。对于MODIS数据来讲选择250-500m的尺度可以尽可能的降低避免道路尺度效应对分析结果影响。与MODIS数据相比,Landsat8数据因为分辨率较高,所以线状景观较为明显,所占比重大。尺度效应明显,所以尺度最好选择在60m-80m左右。
[Abstract]:In geographic information science, we need a variety of spatial data as a support condition to comprehensively assist in the analysis of general geographical objects and the implementation of specific decisions. We should grasp the spatial information data of various scales from the dynamic connection mechanism, and thus, It can be considered that the scale problem plays an important role in the study of new geographic information science. In the application of remote sensing data to landscape ecological pattern analysis, many aspects of cognition have shortcomings, which remote sensing image is the most suitable for which types and areas of landscape ecological pattern research, which scale is the best, for the present, The research of these problems is relatively few, need us to carry on the deeper discussion. In this paper, using Landsat 8.HJ-1AMODIS and other spatial resolution remote sensing image data to analyze, preprocess and extract the landscape ecological pattern, and study the change law of landscape ecological pattern under different resolution. Try to find out the sensitivity factor with the change of resolution. This study selected five experimental sample areas on the Loess Plateau of Northern Shaanxi, namely Changwu sample, Yanchuan Yanchang, Mizhi Suide, Luochuan Huanglong, Dingbian Jingbian, respectively, for each sample area. The changes of landscape pattern at different spatial scales were analyzed, and the following conclusions were obtained: there was no direct correlation between the degree of landscape aggregation and the degree of landscape fragmentation, and there was no obvious relationship between them. As the spatial resolution increases, the patch density, boundary density and maximum patch index all show a significant downward trend, and the corresponding landscape aggregation degree changes in the spatial resolution. The landscape ecological pattern corresponding to different spatial resolution also appeared. The landscape ecological pattern produced by remote sensing image with different resolution had significant difference, and the species phenomenon could be seen clearly in the analysis results. Among the landscape ecological types, the relatively large spatial resolution images belong to the line landscape elements, which are typical representative types of corridors, such as roads and rivers are the most obvious. Most of the study areas selected in this paper are natural ecological landscape, and linear landscape occupies a small proportion in the overall landscape. Therefore, the scale effect of spatial resolution in the scale effect image is not very large. 3) the scale effect of each index of landscape pattern is very obvious. The difference of scale effect will change with the change of spatial resolution. However, there is a special case, Shannon diversity index, which has no obvious scale effect and is only related to the landscape classification system in the study area. For the MODIS data, choosing 250-500m scale can reduce the influence of road scale effect on the analysis results as much as possible. Compared with MODIS data, Landsat8 data has a higher resolution, so the linear landscape is more obvious and occupies a large proportion. The scale effect is obvious, so it is best to choose the scale around 60m-80m.
【学位授予单位】：西北大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P901;P237

【参考文献】