A Study Of Groundwater Quality Spatio-temporal Characteristics And Genesis Analysis In A Costal City

In coastal cities,freshwater resources are less available,groundwater has been an important part of social development and residents' daily life.Studying comprehensively of groundwater spatio-temporal characteristics is the first step in management.On this basis,genesis analysis can improve the accuracy of protection measures.Based on longterm groundwater monitor data in a coastal city,this paper reveals the temporal evolution and spatial distribution characteristics of groundwater,and traces the causes of groundwater quality.This research result provides a more comprehensive analysis of groundwater resources,aiming to make up for the gaps in study area.Moreover,the results will provide a scientific basis for the groundwater resources management in study area and similar coastal cities.This paper is derived from "Research on the Sustainable Development of Groundwater Resources in a costal city".Building a groundwater quality prediction model based on Long Short-Term Memory networks.After testing,this model can achieve groundwater quality prediction and provide a basis for revealing the temporal evolution of groundwater quality.By using relevant methods of groundwater hydrochemistry,the hydrochemical characteristics and spatial distribution characteristics have been revealed.To show the distribution of groundwater components more intuitively,MATLAB is used in this section.Finally,combining geological conditions to determine the cause of groundwater quality.The main contents and results of this study are as follows:(1)Based on three long-term series of groundwater sample data and rainfall data in study area,a groundwater quality prediction model based on Long Short-Term Memory network is constructed.First,conduct descriptive statistical analysis on ion concentration data and rainfall of dry and wet periods,in order to clarify the ion concentration characteristics and determine model parameters.Then,the basic data is preprocessed,including filling vacancies and normalization.Randomly screen 90% from the data set as the training set,and the remaining 10% is the test set.By the training set,the model parameters optimize continuously.Finally,input the test set and perform error analysis on results.The results show that this model can effectively use long-term data to achieve higher-precision groundwater quality prediction.(2)Based on 42 short-time series of groundwater samples data in study area,the spatial distribution characteristics are studied.According to the statistical results,the groundwater quality in the northern inland area(AREA I)is better than that in the southern coastal area(AREA ?).The main anions and cations in AREA ? are similar to seawater.This phenomenon preliminarily proves that the groundwater in coastal areas may be affected by seawater.The ion concentration distribution images output by MATLAB can visually show the spatial distribution characteristics.Except for HCO3-,the rest indicators showed a similar distribution law: most points in AREA ? increased slowly,and the overall concentration was lower.By contrary,the concentration increased faster in AREA ?.By Shukarev classification and Piper diagram,the hydrochemical types are determined.It is verified again that the groundwater quality in AREA ? is controlled by water-rock interaction,while it is controlled by seawater in AREA ?.(3)Based on 42 short-time series of groundwater samples data in study area,the genesis analysis was carried out.By Pearson correlation method,a preliminary genesis analysis of the factors that are significantly related has been carried out.In the process of comparing similarities and differences,it is proved that different geographical locations,geological structures and media contribute to different groundwater quality characteristics.The Gibbs plots indicate that the groundwater quality is mainly controlled by rock weathering in AREA ? and AREA ?.Some samples are affected by seawater intrusion and human activities.Ion proportional coefficient method and Chlor-Alkali index are used to trace the source of specific factors.The results show that the main controlling factors in study area are rock dissolution,seawater intrusion and human factors.The positive cation exchange in AREA ? is the reason for the points that are out of Gibbs plots.