Analysis Of Temporal And Spatial Characteristics Of Precipitation In Jiangxi Province In The Past 20 Years And Research On Prediction Models

Precipitation can intuitively describe the climate change trend within the study area,and it is also an essential natural resource in human life.Changes in precipitation are often accompanied by changes in human life,the natural environment and the national economy.Combining with the precipitation data over the years in the region,the variation law of precipitation on time scale and the distribution characteristics on spatial scale are studied.Accurate prediction of precipitation is of great significance to hydrological analysis and management in Jiangxi Province.Precipitation is a time series data with non-linear and unstable characteristics.The precision of precipitation forecast is usually determined by how to deal with the precipitation data and how to construct the forecast model.According to the monthly precipitation data of 97 meteorological stations in Jiangxi Province and its surroundings from 2000 to 2020,through GIS spatial interpolation method,Mann-kendall mutation analysis and Morlet wavelet analysis,the temporal and spatial characteristics of precipitation in Jiangxi Province are excavated and studied,and Jiangxi is obtained by analysis.The change law,fluctuation period and spatial distribution of precipitation in the past 20 years in the province,and the one-dimensional ARIMA model,two-dimensional ARIMA model,the optimized two-dimensional ARIMA model and the LSTM neural network model were used to predict the precipitation in Jiangxi Province,and the following conclusions were obtained:(1)The average annual precipitation in Jiangxi Province has shown a gradual increase trend,the inter-annual precipitation differences have become more obvious and the periodic fluctuations have changed significantly.The difference between seasonality and seasonality has been enhanced accordingly,and the distribution of precipitation in the four seasons is uneven.Jiangxi Province has a pattern of heavy precipitation years and low precipitation years.Extremely heavy precipitation is basically concentrated in June,and the maximum monthly precipitation shows an increasing trend.The extremely low precipitation in Jiangxi Province is usually concentrated in winter from October to December,but it does not last for more than two months.(2)The results of wavelet analysis show that the precipitation of Jiangxi province in recent 20 years is mainly affected by the periodic fluctuation of 4-8 a and 10-16 a,and there is a periodic oscillation with the first main period being 8 a in Jiangxi province.(3)The comparison results of precipitation interpolation analysis in Jiangxi Province show that the co-kriging interpolation method has the highest accuracy among the annual average precipitation interpolation methods,and the interpolation results can better describe the spatial distribution of precipitation in Jiangxi Province.Precipitation is concentrated in the east-north region,the southern precipitation is increasing,and the northern precipitation is gradually decreasing.(4)Using one-dimensional ARIMA model,two-dimensional ARIMA model and optimized ARIMA model to forecast precipitation in Jiangxi province.Secondly,LSTM neural network model is used to forecast and fit the precipitation in Jiangxi province.The analysis of forecast error shows that the optimized Arima model is superior to the other three models.