加权马尔可夫链模型对上海地区年降水量的预测

Prediction of Annual Precipitation in Shanghai by Weighted Markov Chain Model

摘要: 采用均值–标准差分级法，以上海地区1970~2020年的年降水量数据为样本降水序列，根据上海地区降水量特点，确定了样本降水序列的分级标准和状态。根据马尔可夫理论和统计学原理，验证了样本降水序列满足马尔可夫性(马氏性)，进而以规范化的各阶自相关系数为权重，建立了适用于该地区降水量的加权马尔可夫链的预测模型。以此模型预测了上海地区2021年和2022年的年降雨量，预测结果比较精确。

Abstract: Using mean-standard deviation classification method, taking annual precipitation data of Shanghai from 1970 to 2020 as sample precipitation series, according to the characteristics of precipitation in Shanghai, the classification standard and status of sample precipitation series were determined. According to Markov theory and statistical principle, it is verified that the sample precipitation series satisfies Markov property. Then, the weighted Markov chain prediction model suitable for the precipitation in this region is established by taking the normalized autocorrelation coefficients of each order as the weight. This model predicts the annual rainfall of Shanghai in 2021 and 2022, and the prediction results are relatively accurate.