王丽霞 张珈玮 张双成 隋立春 杨耘 刘招

摘要  以陜西省作为研究区域，首先基于MODIS NDVI中国月合成数据计算研究区2000、2005、2010和2015年度的植被覆盖度并进行等级划分;而后利用CA-Markov模型，以植被覆盖度等级作为元胞类型，计算不同时期各植被覆盖度等级的转移矩阵，由此模拟2010和2015年的NDVI值;对比模拟NDVI结果和原始影像数据，评价模拟精度，并预测2020年NDVI的空间分布状况。结果表明，利用CA-Markov模型进行植被覆盖空间分布的模拟，得到2010和2015年模拟结果的Kappa系数分别为0.797 5、0.853 2，符合精度要求，可以用于植被覆盖空间分布的预测。陕西省植被覆盖存在明显的空间差异性，呈现出陕北—关中—陕南地区植被覆盖度逐渐递增的纬度地带性规律。2000—2020年各级植被覆盖区均有向更高一级植被覆盖区变化的趋势。总体上植被覆盖度上升趋势明显，尤以陕北地区北部变化最为显著，关中地区和陕南秦巴山区植被覆盖度增幅较小。

关键词  归一化植被指数;植被覆盖度;CA-Markov模型;模拟预测

中图分类号  P  237文献标识码  A

文章编号  0517-6611（2020）04-0053-04

doi：10.3969/j.issn.0517-6611.2020.04.016

开放科学（资源服务）标识码（OSID）：

Simulation and Prediction of Vegetation Coverage in Shaanxi Province Based on CA-Markov Model

WANG Li-xia1， ZHANG Jia-wei2， ZHANG Shuang-cheng1 et al  （1.School of Geology Engineering and Geomatics， Changan University， Xian， Shaanxi 710054;2. School of Earth Science and Resources， Changan University， Xian， Shaanxi 710054）

Abstract  Taking Shaanxi Province as the research area. Firstly， based on MODIS NDVI China monthly synthetic data， the vegetation coverage in the region of 2000， 2005， 2010 and 2015 was calculated and graded. Then using the CA-Markov model and taking the vegetation coverage level as the cell type， the transfer matrix of various vegetation coverage levels in different periods was calculated， and NDVI in 2010 and 2015 was simulated. Comparing simulated NDVI with the original image data， the accuracy of simulation was evaluated and the spatial distribution of NDVI in 2020 was predicted. The results indicated that using CA-Markov model to simulate the spatial distribution of vegetation coverage， the Kappa coefficient of the simulation results in 2010 was 0.797 5， and in 2015 was 0.853 2， which met the accuracy requirements and could be used to predict the spatial distribution of vegetation coverage.There were obvious spatial differences in vegetation cover of Shaanxi Province， showing the latitudinal zonal law of increasing vegetation coverage from the northern of Shaanxi to Guanzhong to the southern of Shaanxi area. Between 2000 and 2020， vegetation coverage areas at all levels had a tendency to change to higher-level vegetation coverage areas. In general， the vegetation coverage increased obviously， especially in the northern part of northern Shaanxi. The vegetation coverage in Guanzhong and Qinba Mountains in southern Shaanxi increased little.

Key words  Normalized difference vegetation index;Vegetation coverage;CA-Markov model;Simulation and prediction

植被是陆地生态系统的主体，起着连接土壤、大气和水分等要素，参与生态系统物质循环的重要作用。植被生长和覆盖状况存在明显的时空差异性，在一定程度上反映了气候变化的趋势，是生态环境研究的核心问题之一[1]。遥感技术以其信息量大、观测范围广、精度高、速度快以及实时性和动态性等特点为开展植被监测和评估提供了有效的手段[2]。目前研究主要体现为通过遥感影像提取各类植被指数，在不同空间尺度下对植被覆盖进行长时间序列的动态监测[3-8]。……