摘要
采用 RS和 GIS技术对科尔沁沙地景观斑块结构对沙漠化的影响进行了研究。通过选择固定沙丘斑块、农田斑块和景观中总斑块的数量、平均面积和平均斑块形状等九个因子 ,分别采用单因子关联分析和多因子 PCA分析 ,揭示了科尔沁沙地斑块结构对沙漠化过程的影响。单因子的分析表明 ,固定沙地斑块数 (SN )和平均斑块面积 (SA)、农田斑块数 (CN)和总斑块数 (TN)与沙漠化程度 (DG)的关联程度极为显著 (在 0 .0 1水平上 ) ,关联系数 >0 .92 1。各因子与沙漠化程度之间存在回归关系 ,其中上述 4个因子与沙漠化程度的关系比较密切 ,呈线性回归关系 ,相关系数变化在 0 .90 2～ 0 .991之间 ,相互之间有较高的置信度。PCA的分析表明 ,不同类型斑块数量对沙漠化景观的影响最大 ,其所代表的影响方向 (PC1轴 )占所有分析因子贡献率的5 3.3% ;不同类型斑块形状对沙漠化景观的影响次之 ,其所代表的影响方向 (PC2轴 )占所有分析因子贡献率的 2 7.5 %。从不同斑块类型对沙漠化的影响特点来看 ,固定沙丘斑块对沙漠化的发生、发展起正向的影响作用 ,斑块的各种指数 (SN、SA和 SF)与沙漠化的关系都为正关联。农田斑块对沙漠化的影响作用则较复杂 ,其中斑块数量 (CN )与沙漠化的关系为正关联 ,而平均斑块面积 (CA)
This study analyzes the effect of landscape patch structure influence upon desertification process by using RS and GIS techniques in Kerqin Sandy Land. Basic image adopt landsat TM image obtained on Sep. 6, 2000. RS and GIS software, such as erdas 8.4, Arcview 3.2, and Arcinfo 7.1, was used as an interpretation tools. Landscape is divided to eight types by means of land use and cover classification method. Analysis data come from samples that are same size and locate at different place on interpreted map. The factors of patch number, average patch area, and average patch shape of fixed sandy land, cropland, and total patch have been employed to judge the effect of patch structure influence upon the desertification process. Due to association and regression analysis of single factor between above mentioned factors and desertification degree (DG) which is quantitative indicator of desertification process, results showed that: (1) there were significant correlation (at 0.01 level) among the patch numbers of fixed sandy land (SN), the average area of total patch (SA), the patch numbers of cropland (CN), the numbers of total patch (TN) and the DG. Their association coefficients were more than 0.921. (2) There were linear regression relations between the DG and all of the factors. (3) However, there were not highest correlations between the factors and the DG only, but had highest correlations among inter-factors. These highlighted that there were complex relations among the DG and the factors that could be directly or indirectly influence upon DG. According to above findings we can confirm that the patch structure obviously affects the DG in a synthetic way, so principal component analysis (PCA) becomes a necessary method for best understanding the relation between the patch structure and the DG because it is a useful tool for analyzing multi-factors impact. The PCA results showed that all patch structure effects could be integrated to five vectors for the patch structure influence upon the DG, and total cumulative proportion was 99%. Among of these, the cumulative proportion of the first and the second vectors reached to 80.8% and were main comprehensive factors (vectors). The numbers of different kinds of patches were the major impact factor of desertification process, and its cumulative proportion reached to 53.3% at PC1 axis. The patch shapes were the secondary impact factor of desertification process and its cumulative proportion reached to 27.5% at PC2 axis. In terms of the characteristics of different patch types affecting on the desertification process, the fixed sandy land patches played a positive role for the DG and had a positive correlation between all indices and the DG. The cropland patches played a different role in different ways. First, it brought about a positive correlation between the patch numbers of cropland and the DG. Second, it result in a negative relationship between the average patch areas and the average patch shapes with the DG.
出处
《生态学报》
CAS
CSCD
北大核心
2004年第6期1237-1242,共6页
Acta Ecologica Sinica
基金
国家重点基础研究发展规划资助项目 ( G2 0 0 0 0 1860 7)
国家自然科学基金资助项目 ( 4 960 10 2 0 )~~
关键词
景观
斑块结构
沙漠化
关联分析
PCA分析
landscape
patch structure
desertification
association analysis
principal component analysis (PCA).
