基于FVC指数的中国西北干旱区植被覆盖变化Markov过程

基于1982-2000年NOAA/AVHRR影像的FVC数据，对中国西北干旱区采用先分区再因海拔而异的分类方法进行植被覆盖的遥感分类，并在8 km×8 km空间分辨率下，对研究区植被覆盖变化的任意两年、连续平均和间隔平均转移概率矩阵下Markov过程进行分析与检验，探讨了研究区植被覆盖变化的Markov过程及其指示意义.结果表明：研究区植被覆盖变化受随机过程的控制和长期稳定的驱动因子影响，其转移变化是多重的Markov过程；仅使用两期的植被覆盖变化不能准确预测植被覆盖变化的发展趋势，无论这两期的时间是连续还是有一定时间间隔；对中国西北干旱区而言，连续10年以上的数据变化信息基本可以反映大部分影响该区植被覆盖的因素，采用长期平均转移概率矩阵可以得到较稳定的模拟与预测；植被覆盖变化是长期的动态平衡，平衡一旦被打破，建立新平衡是一个很长的时间过程.

Markov process of vegetation cover change in arid area of Northwest China based on FVC index

Based on the fractional vegetation cover (FVC) data of 1982-2000 NOAA/AVHRR (National Oceanic and Atmospheric Administration / the Advanced Very High Resolution Radiometer) images, the whole arid area of Northwest China was divided into three sub-areas, and then, the vegetation cover in each sub-area was classified by altitude. Furthermore, the Markov process of vegetation cover change was analyzed and tested through calculating the limit probability of any two years and the continuous and interval mean transition matrixes of vegetation cover change with 8 km× 8 km spatial resolution. By this method, the Markov process of vegetation cover change and its indicative significance were approached. The results showed that the vegetation cover change in the study area was controlled by some random processes and affected by long-term stable driving factors, and the transitional change of vegetation cover was a multiple Markov process. Therefore, only using two term image data, no matter they were successive or intervallic, Markov process could not accurately estimate the trend of vegetation cover change. As for the arid area of Northwest China, more than 10 years successive data could basically reflect all the factors affecting regional vegetation cover change, and using long term average transition matrix data could reliably simulate and predict the vegetation cover change. Vegetation cover change was a long term dynamic balance. Once the balance was broken down, it should be a long time process to establish a new balance.