Precipitation Characteristics And Stochastic Simulation In Southern Ningxia

Nowadays,the population is increasing too fast,the economy is growing too fast,and the environment in which we live is severely damaged.The increasingly scarce water resources are becoming more and more a concern for us.The problems caused by this are becoming more and more prominent.In-depth research,the more aware that water resources have become the subject we need to solve today.The main source of fresh water on the surface is atmospheric precipitation.The amount of fresh water that is particularly sinilar to humans is small,and fresh water becomes a renewable resource because of a series of processes such as evaporation and precipitation.However,in many areas,the interannual precipitation and the uneven distribution of time and space during the year often lead to serious flood disasters everywhere.Therefore,in-depth study of the temporal and spatial variation of precipitation has far-reaching significance in all aspects of development.In this paper,we use the precipitation of the original 60 years in the original area of Guyuan City as the research object.Firstly,we use the sliding average method,the morlet wavelet analysis method,the Mann-Kendall mutation test method conference and the periodic mean value superposition to calculate the original and explore the original.Precipitation trends and precipitation distribution characteristics in the state.Then,using the Monte Carlo method,the distribution model test is carried out on the original state area of Guyuan City,and the most suitable mathematical model distribution in the original state area is obtained.Finally,in order to further carry out random simulation and find the most suitable Markov state partitioning method for the region,compare the conventional hydrological specification classification,mean-variance classification method,K-means clustering method and several hierarchical clustering methods.The obtained state division is combined with the sliding average-weighted Markov model to carry out random simulation and prediction of precipitation in the original state to achieve the final research goal.Through the analysis of The precipitation characteristics of the Wonju District,it is found that the precipitation in the Yuanzhou District of Guyuan City is unevenly distributed during the year,with the characteristics of rainy summer and less rain in winter.The annual average precipitation is 442.0mm,the annual variation is small,and the annual precipitation shows a significant downward trend.The precipitation tendency rate is-0.474mm/10a.The annual precipitation has three obvious mutation points in 1963,1992 and 2010.In terms of annual precipitation cycle changes,the main time scale of 60 years of precipitation in the original state is 6a,with a short period of 3a and a partial period of 29a.The precipitation is abundant and dry,showing a clear turbulent relationship.Through Monte Carlo simulation simulation analysis,after determining the parameters based on the original data,the computer software is used to generate the random values that match the corresponding distribution.Using the K-TEST distribution test method,the annual precipitation distribution in the original state has passed a.The gamma distribution,the lognormal distribution,and the normal distribution of=0.01 indicate that the three distributions are more significant.The data simulated by the three models fluctuated greatly,and the trend was basically consistent with the original data.In order to find the optimal model and judge by standard error,the annual precipitation in the original state area is more consistent with the lognormal distribution,and the standard error is 90.65.The relative error of the mean is 1.7%,the relative error of the standard deviation is 4.9%,and the maximum and minimum values of the simulated precipitation are 786.3mm and 235.6mm,respectively,which is not much different from the original data.Through further optimization of the Markov model,and considering the unique precipitation characteristics of the original state,the state is divided by hierarchical clustering based on Euclidean distance,and the original data is verified by this clustering method.The method is feasible.In addition,the average precipitation precipitation data in the original state was in a state of abundance from 2010 to 2014,and the predicted precipitation was 525.3mm,533.2mm,539.3mm,543.9mm and 545.72mm.