汾河流域景观破碎化时空演变特征
作者简介:付扬军(1995- ),男,山西中阳人,硕士,研究方向为土地资源评价与土地利用规划。E-mail: fuyangjun95@163.com
收稿日期: 2018-10-10
要求修回日期: 2019-05-20
网络出版日期: 2019-08-28
基金资助
国土资源部公益性行业科研专项经费课题(201411007)
The spatio-temporal evolution characteristics of landscape fragmentation in Fenhe River Basin
Received date: 2018-10-10
Request revised date: 2019-05-20
Online published: 2019-08-28
Copyright
以1990年、2000年、2015年遥感影像解译数据为基础,利用Fragstats 4.2软件分别采取标准法和移动窗口法获取最佳分析尺度;在此基础上探索各景观指数空间分异特征,揭示流域景观破碎化时空演变过程;采用空间自回归模型分析景观破碎化的影响因子。结果表明:随着粒度和幅度的改变,部分景观指数呈现一定规律性,90 m栅格大小、900 m方形窗口尺寸为研究区最佳分析尺度;斑块密度、蔓延度、有效网格大小、香农多样性指数在三个时期均存在空间集聚效应,且空间集聚类型分布形态具有较强的相似性;1990-2015年汾河流域景观破碎化空间分布格局发生明显变化,破碎化集聚区域呈现碎斑化;坡度、高程、距建设用地距离是影响三个时期流域景观破碎化的主要影响因素。
付扬军 , 师学义 , 和娟 . 汾河流域景观破碎化时空演变特征[J]. 自然资源学报, 2019 , 34(8) : 1606 -1619 . DOI: 10.31497/zrzyxb.20190804
There is fragmented land and fragile ecological environment in the loess hilly and gully region. With the rapid population growth and socio-economic development in recent years, the degree of fragmentation has been further aggravated. Therefore, based on the interpretation data of remote sensing images in 1990, 2000 and 2015, this paper uses Fragstats 4.2 to obtain the best analysis scale combined with the standard method and moving window method. On this basis, the spatial differentiation characteristics of each landscape index are explored; The spatial and temporal evolution process of watershed landscape fragmentation is revealed; The spatial autoregressive model is used to analyze the impact factors contributing to landscape fragmentation. The results show that with the change of grain size and extent, some landscape indices show a regularity to certain degree. The 90 m grid size and 900 m square window size are the best analysis scales in the Fenhe River Basin; patch density, contagion, effective grid size, and Shannon's diversity index have a spatial agglomeration effect in the three periods, and the spatial agglomeration type distribution has strong similarity. From 1990 to 2015, the spatial pattern of landscape fragmentation in the study area has obviously changed; the slope, elevation and distance from the construction land are the main factors affecting landscape fragmentation of river basin during the three periods. The research results can provide reference for landscape fragmentation analysis in Northwest China, and provide basis for future regional landscape pattern optimization and effective land management.
Key words: analysis scale; landscape fragmentation; driving force; Fenhe River Basin
Fig. 1 The location of the Fenhe River Basin图1 研究区地理位置 |
Fig. 2 Change curve of landscape index with spatial grain size图2 景观指数随空间粒度变化曲线 |
Fig. 3 Change curve of landscape index with spatial extent图3 景观指数随空间幅度变化曲线 |
Fig. 4 Moran scatter plot of landscape index图4 景观指数Moran散点分布 |
Fig. 5 LISA cluster graph of local spatial autocorrelation of landscape index图5 景观指数局部空间自相关LISA集群 |
Table 1 Principal component analysis results of landscape indices in three periods表1 3个时期景观指数主成分分析结果 |
年份 | KMO | 景观指数 | 公因子方差 | 成分 | 初始特征值 | 主成分 | |||
---|---|---|---|---|---|---|---|---|---|
初始 | 提取 | 特征值 | 方差贡献率/% | 累计方差贡献率/% | 1 | ||||
1990 | 0.770 | CONTAG | 1.00 | 0.819 | 1 | 3.380 | 84.505 | 84.505 | 0.905 |
MESH | 1.00 | 0.937 | 2 | 0.445 | 11.133 | 95.637 | 0.968 | ||
PD | 1.00 | 0.682 | 3 | 0.134 | 3.342 | 98.980 | -0.826 | ||
SHDI | 1.00 | 0.942 | 4 | 0.041 | 1.020 | 100.000 | -0.971 | ||
2000 | 0.778 | CONTAG | 1.00 | 0.835 | 1 | 3.416 | 85.399 | 85.399 | 0.914 |
MESH | 1.00 | 0.938 | 2 | 0.411 | 10.286 | 95.685 | 0.969 | ||
PD | 1.00 | 0.699 | 3 | 0.135 | 3.370 | 99.055 | -0.836 | ||
SHDI | 1.00 | 0.943 | 4 | 0.038 | 0.945 | 100.000 | -0.971 | ||
2015 | 0.758 | CONTAG | 1.00 | 0.794 | 1 | 3.317 | 82.937 | 82.937 | 0.891 |
MESH | 1.00 | 0.935 | 2 | 0.483 | 12.082 | 95.019 | 0.967 | ||
PD | 1.00 | 0.655 | 3 | 0.156 | 3.912 | 98.931 | -0.809 | ||
SHDI | 1.00 | 0.934 | 4 | 0.043 | 1.069 | 100.000 | -0.966 |
Fig. 6 Spatial distribution pattern of landscape fragmentation in three periods图6 景观破碎化程度空间分布格局 |
Table 2 Comparison of fitting results of spatial lag model and spatial error model of different impact factors表2 不同影响因子空间滞后模型和空间误差模型拟合结果对比 |
自回归模型 | R2 | LIK | AIC | SC | 驱动因子 | 回归系数 | Z值 | P值 |
---|---|---|---|---|---|---|---|---|
空间滞后模型 | 0.763 | -69708 | 139429 | 139491 | 高程 | -0.000095 | -6.003210 | 0.000 |
坡度 | 0.017500 | 16.859500 | 0.000 | |||||
坡向 | -0.000163 | -1.473480 | 0.141 | |||||
距建设用地距离 | 0.000013 | 7.766020 | 0.000 | |||||
距水域距离 | -0.000003 | -3.569110 | 0.000 | |||||
空间误差模型 | 0.784 | -68458 | 136928 | 136980 | 高程 | 0.002065 | 18.539400 | 0.000 |
坡度 | 0.058957 | 27.233100 | 0.000 | |||||
坡向 | -0.000140 | -1.060340 | 0.289 | |||||
距建设用地距离 | 0.004669 | 40.491100 | 0.000 | |||||
距水域距离 | -0.000087 | -6.252210 | 0.000 |
注:LIK:最大似然对数值;AIC:Akaike信息指标;SC:Schwartz指标。 |
[1] |
[
|
[2] |
[
|
[3] |
[
|
[4] |
|
[5] |
|
[6] |
[
|
[7] |
[
|
[8] |
[
|
[9] |
[
|
[10] |
[
|
[11] |
[
|
[12] |
[
|
[13] |
[
|
[14] |
[
|
[15] |
[
|
[16] |
[
|
[17] |
中国科学院. 地理空间数据云. (2016-12-01) [2017-08-09].
[Chinese Academy of Sciences. Geospacial Data Cloud. (2016-12-01) [2017-08-09].
|
[18] |
[
|
[19] |
|
[20] |
|
[21] |
|
[22] |
[
|
[23] |
[
|
[24] |
[
|
[25] |
[
|
/
〈 |
|
〉 |