基于CA-Markov的渭河流域NDVI时空变化模拟及预测
王丽霞1, 张珈玮2, 孟妮娜1, 隋立春1,4, 张双成1, 刘 招3

(1.长安大学 地质工程与测绘学院, 西安 710054; 2.长安大学 地球科学与资源学院, 西安 710054; 3.长安大学 环境科学与工程学院, 西安 710054; 4.地理国情监测国家测绘地理信息局 工程技术研究中心, 西安 710054)

归一化植被指数; CA-Markov模型; 渭河流域

Simulation and Prediction of Temporal and Spatial Changes of NDVI in Weihe River Basin Based on CA-Markov
WANG Lixia1, ZHANG Jiawei2, MENG Nina1, SUI Lichun1,4, ZHANG Shuangcheng1, LIU Zhao3

(1.School of Geology Engineering and Geomatics, Chang'an University, Xi'an 710054, China; 2.School of Earth Science and Resources, Chang'an University, Xi'an 710054, China; 3.School of Environmental Science and Engineering, Chang'an University, Xi'an 710054, China; 4.Engineering Technology Research Center, National Geographical Monitoring National Surveying and Mapping Geographic Information Bureau, Xi'an 710054, China)

normalized difference vegetation index; CA-Markov model; Weihe River Basin

备注

为综合分析流域植被覆盖的时空变化特征,进而为生态环境保护提供科学参考,以渭河流域作为研究区域,首先基于MODIS NDVI中国月合成数据计算研究区年际NDVI值,并进行等级划分; 而后利用CA-Markov模型,以NDVI等级作为元胞类型,计算了不同时期各等级的转移矩阵,由此模拟了2015年NDVI的空间分布; 对比模拟NDVI结果和原始影像数据,评价模拟精度,并预测了2020年和2025年NDVI的空间分布状况。结果 表明:(1)利用CA-Markov模型模拟渭河流域NDVI的空间变化,得到2015年模拟结果的Kappa系数为0.785 0,叠置分析的准确度为73.4%,符合精度要求,可以用于NDVI空间分布的预测。(2)渭河流域植被覆盖存在明显的空间差异性,低度植被覆盖区、较低植被覆盖区主要分布于陇中和陕北黄土高原地区; 中度植被覆盖区主要分布在泾河和北洛河河谷地区,较高植被覆盖区主要分布在关中平原、子午岭等地,高度植被覆盖区主要分布在秦岭、六盘山等山区。(3)2015—2025年,预测流域植被覆盖状况将进一步改善。其中,低度植被覆盖区、较低植被覆盖区和中度植被覆盖区面积将有所减少,空间上整体向北迁移; 较高植被覆盖区和高度植被覆盖区面积进一步增加,空间上向北扩张。

Dynamic monitoring of vegetation plays an important role in indicating changes in climate and ecological environment. The Weihe River Basin is used as the research area. Firstly, the interannual NDVI value of the study area is calculated based on the MODIS NDVI China monthly synthesis data, and the grade is divided. Then based on the CA-Markov model, the NDVI grade is used as the cell type to calculate the transfer matrix of grades in different periods, in order to simulate the spatial distribution of NDVI in 2015; the simulated NDVI results with raw image data were compared to evaluate simulation accuracy and predict the spatial distribution of NDVI in 2020 and 2025. The results showed that:(1)the CA-Markov model was used to simulate the spatial variation of NDVI in the Weihe River Basin, the Kappa coefficient of the simulated results in 2015 was 0.785 0, the accuracy of the overlay analysis was 73.4%, and meets the accuracy requirements, the CA-Markov model can be used for the prediction of NDVI spatial distribution;(2)there are obvious spatial differences in vegetation cover in the Weihe River Basin; low vegetation cover areas and relatively lower vegetation cover areas mainly distributed in northern Shaanxi Province and the central of Gansu Province in the Loess Plateau, the medium vegetation cover areas mainly distributed in the Weihe River and Beiluohe River; in the valley area, the higher vegetation coverage areas mainly distributed in the Guanzhong Plain and Ziwuling Mountains, and the high vegetation coverage areas mainly distributed in the Qinling and Liupan Mountains;(3)during the period 2015—2025, the predicted vegetation cover in the basin will be further improved; the low vegetation coverage area, lower vegetation coverage area and medium vegetation coverage area will reduce, and the whole space will migrate northward; the area of higher vegetation coverage area and high vegetation coverage area will further increase, and the space will expand northward.