土地利用/景观生态学研究中的马尔可夫链统计性质分析

马尔可夫链在土地利用和景观生态学研究中得到了广泛应用，而应用中通常假设土地利用变化为满足马尔可夫性的一阶时齐马尔可夫链，对马尔可夫链的统计性质是否成立却很少进行检验.本文以北京市土地利用变化监测数据为算例，提出了马尔可夫链统计性质的皮尔逊χ² 拟合优度检验方法.检验结果表明，土地利用研究中通常假设的时齐性和马尔可夫性(一阶性)在统计学上并不成立，即北京土地利用演变过程为非时齐的高阶马尔可夫链.相对于马尔可夫统计性质的似然比检验中转移概率大于零的要求，皮尔逊χ²检验对转移概率的要求相对宽松，允许转移概率为零，所以应用的范围较似然比检验更为广泛.

Statistical properties of Markov chain in land use and landscape study

Markov chain has been widely applied in the study of land use and landscape changes, but its statistical properties were less tested. Based on the land use change data monitored in Beijing, and with Pearson X^2 goodness of-fit test, this paper examined the time stability and time independence of Markov chain of land use change. The results indicated that the hypothesis of time stationary and Markov property is not tenable, which meant that the land use change in Beijing was an un-stationary and highly order Markov chain. Pearson X^2 test was not as restricted as the likelihood ratio test, its transitional probabilities being allowed to be greater than zero, and thus, could be more useful in testing the assumption of homogeneity and independence.