Application Of Combination Model Of Markov Chain In Prediction Of Precipitation

Due to the complexity,diversity and variability of climatic conditions,there are a lot of vagueness and uncertainty in the process of weather forecasting,which make the prediction of medium and long term rainfall as a difficult problem in computer prediction and meteorological science.Although China is large in territory and complex in terrain,there is some regularity in its spatial distribution of rainfall.Both people's lives and economy development are closely related to the weather.Based on satellite observations,weather forecast can accurately predict wind,humidity,temperature and other weather conditions in the coming days,but how to predict the rainfall changes in coming months or even coming years?This is a question that the paper focused on.Markov chain is suitable for the prediction of stochastic volatility,but it requires no state effect and the prediction of extreme value is not ideal.However,the combination model can optimize its defects by combining other algorithms.Therefore,this paper forecasts the rainfall based on the combination model of Markov chain and puts forward two corresponding forecasting methods.Next,the urban rainfall distribution system is designed and implemented by using the Arcgis components.The main research work and achievements are as follows:(1)This paper proposes a prediction algorithm based on Markov chain.and fuzzy set combination model.The clustering of rainfall is sorted first and then we calculate the state transition matrix.Next,the concept of membership degree is introduced and the influence of each state on the other states is obtained.Afterwards,the predicted state interval is calculated by weighted method.Finally,the specific rainfall forecast value is calculated by fuzzy set formula.The algorithm eliminates the more general classification of traditionally used,and then introduces the membership degree of membership to describe the sequence of random variables of more value which belong to all states.It's better to reduce the error brought by the original"either this or that" thought.(2)This paper also proposes a prediction algorithm based on Markov chain and grey theory combined model.In this method,firstly the sliding GM(1,1)model is introduced and we get the prediction function by the least squares method.Then the first prediction value and its error with the actual value are obtained by substituting the specific time variables.Next,the error is classified into different states.At last,we use the weighted Markov model to complete the prediction.The algorithm mainly adopts the second prediction method,which reduces the influence of accidental factors,simplifies the modeling steps and improves the accuracy and speed of prediction.(3)In this paper,we study and use the research techniques of rainfall distribution map and rainfall forecast at home and abroad,and utilize the interpolation algorithm and Arc Toolbox to construct the equivalent model.Then through the ArcGIS server,we publish geographic information services and succeed in making an urban rainfall distribution map.Finally,we achieve annual rainfall forecasting and mapping function through the weighted Markov algorithm,which provide convenience for people's production and life as well as disaster prevention and mitigation in relevant departments.Through the above studies,we optimize the Markov combination forecasting algorithm,improve the speed and precision of the prediction and lay a solid foundation for the prediction of long-term rainfall.