| With the increasingly prominent contradiction between groundwater supply and demand,the ecological environment,agricultural production,and socio-economic development of various countries and regions around the world have been constrained to varying degrees.Large-scale groundwater overexploitation has led to the depletion of over 50% of the world’s aquifers at an alarming rate,which has seriously threatened the healthy development of human society.Global warming has become an indisputable fact,and extreme weather is becoming increasingly frequent.Some experts and scholars have focused on groundwater issues and proposed the term "groundwater drought",closely combining climate issues with groundwater shortage issues,striving to provide a new research approach for the development and utilization of groundwater resources and promoting sustainable development of groundwater in the context of climate change.The North China Plain,as one of the regions with the most serious hydrogeological problems in the world,is the most important grain production base in China.The groundwater overdraft problem caused by farmland irrigation has led to a huge groundwater depression funnel in this region,causing a series of geological environmental problems,which has caused serious negative impact on the local social and economic development.Therefore,the study of groundwater drought in the North China Plain and the establishment of a groundwater drought research framework that integrates assessment,identification,response and prediction will help comprehensively and systematically reveal the characteristics of groundwater drought in the North China Plain,deeply understand the phenomenon of groundwater drought,and have important practical significance and theoretical value for the sustainable development of groundwater in the North China Plain,guiding local industrial and agricultural production,and promoting the harmonious coexistence between human beings and nature.Based on the data of gravity retrieval and climate experiment satellite GRACE and the global land data assimilation system GLDAS,this paper constructs groundwater drought evaluation indicators to quantitatively evaluate groundwater drought in the North China Plain,The gravity center transfer model of groundwater drought was established,the influencing factors and their response relationships of groundwater drought were discussed,and three different algorithms based on LSTM were optimized to predict the groundwater drought.Through comparative analysis,the most suitable model for groundwater drought prediction in the North China Plain was obtained.The main research contents and achievements are as follows:(1)The groundwater drought in the North China Plain was assessed,and its spatial-temporal evolution characteristics were revealed.The reliability of the GRACE data results was confirmed by validating the GRACE data with various types of data from GLDAS inverse performances;the groundwater drought index GGDI was constructed to assess the groundwater drought in the North China Plain,and the main oscillation period of groundwater drought in the study area was determined to be 3.3months by using the pole-symmetric modal decomposition method ESMD and the fast Fourier transform FFT,and the linear tilt rates were-0.012 and-0.013 months for the whole area and three geomorphic zones,respectively.and 5.4 months for Geomorphic Zone II,and2.8 months for Geomorphic Zone III,and the linear tilt rates for the whole region and the three geomorphic subzones were-0.012/month,-0.013/month,-0.012/month,and-0.011/month,respectively,and the groundwater drought showed a trend of continuous aggravation.From the spatial distribution of drought frequencies of different levels at monthly,seasonal and annual scales,it can be seen that in the monthly scale GGDI,light and moderate droughts dominate in the whole region of the North China Plain and administrative subdivisions;in the seasonal scale GGDI,spring droughts dominate in the whole region of the North China Plain;in the annual scale GGDI,light and moderate droughts dominate in the whole region of the North China Plain and in Hebei,Tianjin and Shandong are dominated by light and severe drought,while Hebei and Beijing are dominated by light and moderate drought.(2)A groundwater drought gravity center migration model was constructedA spatio-temporal coupled groundwater drought event identification method is proposed to identify groundwater drought events in the North China Plain from 2003 to 2020.49 groundwater drought events are identified,and 11 drought events exceeding three months are screened according to the principles of groundwater drought center of gravity migration model construction to construct a groundwater drought center of gravity migration model,and the results show that the migration direction is mostly related to the topographic tilt It was found that the migration direction was mostly related to the direction of topographic inclination,i.e.,migration in the southwest-northeast direction,and the drought center of gravity was mostly concentrated in the geomorphic zone II(central alluvial lacustrine plain).In addition,the drought intensities of 11 drought events ranged from 0.88 to 38.17.Visual mapping of the occurrence process of the three drought events with the highest drought intensities revealed that the occurrence of larger drought intensities did not concentrate at a fixed stage of the events,and there was no significant correlation between the drought intensities and the monthly maximum drought grid number.(3)The influencing factors and response relationships of groundwater drought are analyzedThe response relationships between the four influencing factors of meteorological drought index SPI,normalized vegetation index NDVI,and human activity and groundwater drought GGDI were studied using sliding correlation and Pearson correlation analysis,respectively,and the results showed that:(1)The regions with significant correlation between SPI and GGDI in North China Plain during 2003-2020 accounted for 93.12%,with the response time ranging from 8 to 10 months,among which the correlation between Beijing and Hebei region in the southern part of Yanshan Mountain is poor.Among the nine sliding windows,the area of the correlation between SPI and GGDI in 2011-2020 is the largest,93.57%,and the response time is mostly between 8 and 12 months;the area of the correlation in 2007-2016 is the smallest,29.36%,and the response time is mostly 11 months.(2)The area showing significant correlation between NDVI and GGDI at monthly scale accounted for18.55% of the whole region,and the area showing significant correlation between NDVI and GGDI at seasonal scale ranged from 9.95% to 40.27%,among which the correlation area was the most in spring and the least in autumn,and it mostly showed significant negative correlation in spring and summer when plants were growing,while it mostly showed significant positive correlation in autumn and winter when plants were not growing.The correlation between NDVI and GGDI at the annual scale was the lowest in autumn.The annual scale NDVI and GGDI showed significant correlation in 33.34% of the region.The areas with positive correlation between NDVI and GGDI in monthly and annual scales are mainly in the border area between Hebei and Shandong,and the areas with negative correlation are mainly in northern Hebei,Beijing and Tianjin.(3)The areas with significant correlations between groundwater level variability and GGDI in monthly,seasonal and annual scales account for 80.42%,45.45%～55.94% and 65.03% of the total grid with observation well data,respectively,and mostly show positive correlations,with positive correlations mostly concentrated in Beijing,Henan and Hebei’s pre-mountain plain areas,and negative correlations mostly concentrated in Hebei and Shandong’s The areas with positive correlation are mostly in the pre-mountain plain areas of Beijing,Henan and Hebei,and the areas with negative correlation are mostly in the border area between Hebei and Shandong and nearby areas.(4)The significant reduction of groundwater extraction can weaken the aggravation trend of groundwater drought,but with the increase of cumulative groundwater extraction,the trend of significant aggravation of groundwater drought will occur again.(4)Established a groundwater drought prediction modelUsing the monthly series data of SPI,NDVI and Groundwater level variation as the input data and GGDI as the output data,the prediction models of LSTM and PSO-LSTM,SAA-LSTM and WOA-LSTM with different parameter optimization algorithms were constructed for the whole region and each geomorphic zone of the North China Plain,and the results of the four prediction models were evaluated by three indexes of RMSE,MAE and R2 The results show that each region has different adaptability to the four prediction models,and the optimal prediction models are PSO-LSTM for Geomorphic Zone 1 and Geomorphic Zone3,and SSA-LSTM for the whole region and Geomorphic Zone 2.In addition,after the comparative analysis,the prediction accuracy of the algorithm-optimized LSTM prediction models for the whole region and the three geomorphic zones of the North China Plain is greater than that of the LSTM prediction models without In addition,the prediction accuracy of the algorithm-optimized LSTM model for the whole North China Plain and the three geomorphic regions is greater than that of the non-algorithm-optimized model. |
PDF Full Text Request