Deciphering the mechanism of groundwater temperature changes associated with longwall mining in a coalfield, China, using the extreme gradient boosting method
{"title":"Deciphering the mechanism of groundwater temperature changes associated with longwall mining in a coalfield, China, using the extreme gradient boosting method","authors":"Shen Qu, Guangcai Wang, Shouchuan Zhang, Zheming Shi, Xiangyang Liang, Ankun Luo","doi":"10.1007/s10040-024-02807-w","DOIUrl":null,"url":null,"abstract":"<p>Investigations focusing on the impacts of mining on groundwater systems typically provide a qualitative analysis of groundwater flow and chemistry, whereas relatively few studies quantitatively analyze groundwater temperature perturbations induced by mining. This study aims to identify the hydrogeological mechanism responsible for changes to groundwater temperature associated with longwall coal mining. Here, the extreme gradient boosting (XGBoost) method was used to construct three models at different phases of mining disturbance to identify the factors governing groundwater temperature dynamics: (1) a pre-disturbance model; (2) an in-disturbance model; and (3) a post-disturbance model. The feature relative importance (FRI) of input variables contributing to groundwater temperature dynamics was quantified for a long-term groundwater monitoring dataset collected from the Ningtiaota Coalfield, Ordos Basin, China. Pre-mining disturbance groundwater temperatures were stable, and the XGBoost model identified the groundwater level of the respective monitoring wells to be the greatest predictor for variation in groundwater temperature. During mining disturbance, proximal monitoring wells exhibited a decline in groundwater temperature, where the FRI of groundwater temperature in an upgradient monitoring well increased by 151–662% relative to the pre-mining disturbance model. The monitoring of aquifer properties and stable isotope composition of groundwaters provided additional evidence to suggest groundwater temperature decreases were associated with increased recharge contributions from surficial Quaternary aquifers. Post-mining disturbance, groundwater temperature and aquifer specific storage demonstrated recovered to pre-mining conditions. This study provides insights into mining-induced groundwater temperature dynamics as a result of changes to hydraulic connection between aquifers.</p>","PeriodicalId":13013,"journal":{"name":"Hydrogeology Journal","volume":"81 1","pages":""},"PeriodicalIF":2.4000,"publicationDate":"2024-06-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Hydrogeology Journal","FirstCategoryId":"89","ListUrlMain":"https://doi.org/10.1007/s10040-024-02807-w","RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"GEOSCIENCES, MULTIDISCIPLINARY","Score":null,"Total":0}
引用次数: 0
Abstract
Investigations focusing on the impacts of mining on groundwater systems typically provide a qualitative analysis of groundwater flow and chemistry, whereas relatively few studies quantitatively analyze groundwater temperature perturbations induced by mining. This study aims to identify the hydrogeological mechanism responsible for changes to groundwater temperature associated with longwall coal mining. Here, the extreme gradient boosting (XGBoost) method was used to construct three models at different phases of mining disturbance to identify the factors governing groundwater temperature dynamics: (1) a pre-disturbance model; (2) an in-disturbance model; and (3) a post-disturbance model. The feature relative importance (FRI) of input variables contributing to groundwater temperature dynamics was quantified for a long-term groundwater monitoring dataset collected from the Ningtiaota Coalfield, Ordos Basin, China. Pre-mining disturbance groundwater temperatures were stable, and the XGBoost model identified the groundwater level of the respective monitoring wells to be the greatest predictor for variation in groundwater temperature. During mining disturbance, proximal monitoring wells exhibited a decline in groundwater temperature, where the FRI of groundwater temperature in an upgradient monitoring well increased by 151–662% relative to the pre-mining disturbance model. The monitoring of aquifer properties and stable isotope composition of groundwaters provided additional evidence to suggest groundwater temperature decreases were associated with increased recharge contributions from surficial Quaternary aquifers. Post-mining disturbance, groundwater temperature and aquifer specific storage demonstrated recovered to pre-mining conditions. This study provides insights into mining-induced groundwater temperature dynamics as a result of changes to hydraulic connection between aquifers.
期刊介绍:
Hydrogeology Journal was founded in 1992 to foster understanding of hydrogeology; to describe worldwide progress in hydrogeology; and to provide an accessible forum for scientists, researchers, engineers, and practitioners in developing and industrialized countries.
Since then, the journal has earned a large worldwide readership. Its peer-reviewed research articles integrate subsurface hydrology and geology with supporting disciplines: geochemistry, geophysics, geomorphology, geobiology, surface-water hydrology, tectonics, numerical modeling, economics, and sociology.