对数马尔可夫模型在乌鲁木齐市降水预测中的应用

Application of Logarithmic Markov Model to Precipitation Forecast in the City of Urumchi

中长期降水量的预测是气象科学的一个难点问题,也是水文学中的一个重要问题。建立对数马尔可夫模型预测降水量,弥补了传统的马尔可夫模型降水预测中峰值的不准确性,提高了预测精度,并用乌鲁木齐市气象站43年降水资料进行了验证。结果表明,模型预测精度较高,为干旱半干旱区中长期降水量预报提供了一条简便可行的途径。

The medium- and long-term precipitation forecast is a difficult problem in the meteorological science and an important problem in the hydrology as well. The establishment of logarithmic Markov Model for forecasting of precipitation helps improve forecast accuracy of precipitation as compared with the conventional Markov Model precipitation forecast, which is unable to provide an accurate peak value. This has been verified by the precipitation data obtained in a period of 43 years at Ununchi Meteorological Station. The result of verification shows that the accuracy of the logarithmic Markov Model precipitation forecast is rather high. It provides a simple and practicable way for medium- and long- term precipitation forecast in dry and semi- dry re,iota.