A Study On Drought Risk Prediction Assessment In Guanzhong Region Based On Feature Analysis

Drought is a universal natural disaster and one of the most serious natural disasters in China.Since ancient times,drought has been the main factor leading to grain yield reduction,water resource shortage and ecological environment deterioration.Risk assessment studies combined with drought change patterns and predicted drought development trends can intervene between droughts without causing serious damage.Therefore,this paper firstly analyzes the drought change pattern in Guanzhong region;secondly,multiple models are simulated for drought prediction and a drought prediction model suitable for Guanzhong region is screened;Finally,drought hazard indicators are obtained based on drought characterization,future drought severity indicators are obtained based on drought forecasts,and drought vulnerability indicators are combined to evaluate the drought risk of the Guanzhong region.The results of risk evaluation provide a scientific basis for the formulation of drought prevention and drought relief policies.The main research results of the paper are as follows:(1)The study of the spatial and temporal evolution of drought in the Guanzhong region is based on the standardized precipitation index SPI.The analysis of temporal variation patterns in Guanzhong region is based on the SPI12 drought index and uses the M-K trend test method as well as the linear trend line method.The results show that the drought index of five cities in Guanzhong has a decreasing trend during 1961-2020;the patterns of temporal,intensity,peak,and frequency changes in different regions of Guanzhong are extracted from the drought characteristic variables by means of travel theory..(2)Multivariate Copula-based drought recurrence analysis.Based on the drought characteristics identified by the tour theory,a Markov chain Monte Carlo simulation algorithm under the Bayesian framework was used for Copula parameter estimation,and the optimal Copula function was screened based on the evaluation index.The multi-characteristic variable joint analysis of drought ephemeris and drought intensity,drought intensity and drought peak,and drought ephemeris and drought peak was carried out based on the obtained Copula function to obtain the distribution of recurrence periods in Guanzhong region.The analysis results show that the distribution of drought recurrence periods in the mild and moderate cases are relatively close,and the short-duration recurrence periods in the severe drought case are mainly distributed in the southeastern part of Guanzhong.(3)A study on multi-scale drought prediction based on multi-model optimization.A total of six models were used for 1 month lead time,3 month lead times,and 12 month lead times drought forecasting in the Guanzhong region,including three independent models SARIMA,online LSTM,and offline LSTM,respectively,three combined models based on EEMD decomposition EEMD-SARIMA model,EEMD-online LSTM model,and EEMD-offline LSTM model.The results show that the prediction results of the combined models are improved compared with the single model prediction results,in which the optimal prediction models for all three forecasting periods are EEMD-offline LSTM models.Therefore,the EEMD-offline LSTM model was used for future drought index prediction,and the overall drought situation in the coming year is light,and drought may occur in some areas in summer,but the drought situation in Weinan City is more serious and needs to be focused on.(4)Analysis of Drought Impact Factors and Risk Assessment in Guanzhong Region.Drought risk is influenced by a combination of natural environment,human activities,economic development and other factors.Therefore,in this paper,five cities in Guanzhong region are taken as the research objects by administrative districts,and in order to conduct drought risk evaluation,The Drought Risk Assessment Composite Index DRI1 is constructed from drought hazard and drought vulnerability,and the future drought risk projection composite index DRI2 is constructed from the results of drought vulnerability projections and future drought severity.Then,drought severity range are constructed Arc GIS 10.1 software was used to map the drought risk partition of Guanzhong region.Comparing the results of the two indices,it can be seen that the inclusion of the future drought severity index will lead to some changes in the risk level.But the risk is lower in Tongchuan and continues to be higher in Weinan.The factor contribution degree model was used to carry out the factor contribution degree calculation of the three index layers to obtain the main impact factors of drought risk.Based on the main impact factors and risk evaluation results,disaster prevention and mitigation recommendations are proposed for reference.