【目的】青藏高原作为中国最大,世界海拔最高的高原,地表温度垂直递减率(Land Surface Temperature Lapse Rate,LTLR)的时空分布特征对气候变化、生态系统以及水文过程研究具有重要意义。已有研究无法准确表达复杂地形条件下的山区近地...【目的】青藏高原作为中国最大,世界海拔最高的高原,地表温度垂直递减率(Land Surface Temperature Lapse Rate,LTLR)的时空分布特征对气候变化、生态系统以及水文过程研究具有重要意义。已有研究无法准确表达复杂地形条件下的山区近地表气温直减率在时空分布与变化上的精细特征。因此本研究利用地表温度日变化模型估算出青藏高原逐小时地表温度,进而计算出逐小时月均LTLR,以获得青藏高原地区高时空分辨率的LTLR分布。【方法】本研究基于2022年中国西部逐日1 km全天候地表温度数据集TRIMS,利用地表温度日变化模型对青藏高原逐小时地表温度进行估算,进而采用滑动窗口法计算逐小时月均LTLR,分析了研究区LTLR在季节尺度上的时空分布与差异特征。弥补了青藏高原地区缺少高时空分辨率LTLR研究的现状。【结果】(1)4个季节平均LTLR分别为-6.12、-7.63、-5.89和-3.23℃/km,春夏季节整体高于秋冬季节,但横断山脉区域相反,冬季平均LTLR较夏季高出约0.57℃/km;(2)春夏季最大LTLR分别为-14.45℃/km、-13.92℃/km,相对于秋、冬季最大LTLR的-13.60℃/km、-11.61℃/km,更高,因高海拔和干旱晴朗天气影响,羌塘高原区不同季节的最大LTLR差异显著,其中冬季最大LTLR最小,为-13.67℃/km;(3)夏季最小LTLR最为大,高出其他季节约3.05℃/km,其中横断山脉四季最小LTLR均较大,其中春季最小LTLR为-1.16℃/km,比其他3个季节更高,最小的秋季最小LTLR为0.03℃/km,而羌塘高原区四季最小LTLR最小;(4)日变化曲线显示,春秋冬3个季节的LTLR在11:00—14:00最大,春季最小LTLR出现在20:00—23:00,秋季最小LTLR出现时间较春季提前了约1 h,而夏季一天中出现2次最大LTLR,分别在4:00—7:00和15:00—18:00,在21:00—23:00呈现出日最小LTLR特征。【结论】本研究对深入揭示青藏高原地表温度垂直递减率在季节尺度上的时空变化特征与相关影响机制有重要作用。展开更多
With the increasing depth and intensity of coal mining operations,high-energy mine tremors have become a major trigger for rockburst disasters,posing severe threats to mine safety.Conventional rockburst risk assessmen...With the increasing depth and intensity of coal mining operations,high-energy mine tremors have become a major trigger for rockburst disasters,posing severe threats to mine safety.Conventional rockburst risk assessment methods either lack real-time adaptability or rely heavily on qualitative microseismic data analysis,limiting their effectiveness in dynamic early warning.To address these limitations,this study proposed a predictive framework for rockburst risk assessment by integrating ensemble learning algorithms with Bayesian optimization.A dataset was constructed using a sliding time window approach,linking the highest MS energy in the subsequent days with predefined risk levels.Both undersampling and oversampling strategies were employed to mitigate class imbalance,and their performance was evaluated.Three ensemble models,i.e.CatBoost,Random Forest,and LightGBM,were developed,and their hyperparameters were optimized using Bayesian techniques to enhance predictive performance.The models were validated using MS data from the 6303 and 6306 working faces at the Dongtan Coal Mine.All three ensemble models outperformed conventional classification methods,particularly in accurately predicting high-risk categories.Among them,the CatBoost model exhibited the best performance,with an accuracy of 89.47%and an F1¯-score of 90.62%.Furthermore,SHapley Additive exPlanations analysis was used to enhance model interpretability,identifying key MS indicators influencing rockburst risk predictions.This study provides a systematic approach for leveraging MS data and machine learning to improve an early warning system for rockburst hazards,offering valuable insights for underground mining safety management.展开更多
文摘【目的】青藏高原作为中国最大,世界海拔最高的高原,地表温度垂直递减率(Land Surface Temperature Lapse Rate,LTLR)的时空分布特征对气候变化、生态系统以及水文过程研究具有重要意义。已有研究无法准确表达复杂地形条件下的山区近地表气温直减率在时空分布与变化上的精细特征。因此本研究利用地表温度日变化模型估算出青藏高原逐小时地表温度,进而计算出逐小时月均LTLR,以获得青藏高原地区高时空分辨率的LTLR分布。【方法】本研究基于2022年中国西部逐日1 km全天候地表温度数据集TRIMS,利用地表温度日变化模型对青藏高原逐小时地表温度进行估算,进而采用滑动窗口法计算逐小时月均LTLR,分析了研究区LTLR在季节尺度上的时空分布与差异特征。弥补了青藏高原地区缺少高时空分辨率LTLR研究的现状。【结果】(1)4个季节平均LTLR分别为-6.12、-7.63、-5.89和-3.23℃/km,春夏季节整体高于秋冬季节,但横断山脉区域相反,冬季平均LTLR较夏季高出约0.57℃/km;(2)春夏季最大LTLR分别为-14.45℃/km、-13.92℃/km,相对于秋、冬季最大LTLR的-13.60℃/km、-11.61℃/km,更高,因高海拔和干旱晴朗天气影响,羌塘高原区不同季节的最大LTLR差异显著,其中冬季最大LTLR最小,为-13.67℃/km;(3)夏季最小LTLR最为大,高出其他季节约3.05℃/km,其中横断山脉四季最小LTLR均较大,其中春季最小LTLR为-1.16℃/km,比其他3个季节更高,最小的秋季最小LTLR为0.03℃/km,而羌塘高原区四季最小LTLR最小;(4)日变化曲线显示,春秋冬3个季节的LTLR在11:00—14:00最大,春季最小LTLR出现在20:00—23:00,秋季最小LTLR出现时间较春季提前了约1 h,而夏季一天中出现2次最大LTLR,分别在4:00—7:00和15:00—18:00,在21:00—23:00呈现出日最小LTLR特征。【结论】本研究对深入揭示青藏高原地表温度垂直递减率在季节尺度上的时空变化特征与相关影响机制有重要作用。
基金funded by the National Natural Science Foundation of China(Grant No.42477208)Natural Science Foundation of Hubei Province,China(Grant No.2024AFA072)Open Research Fund of State Key Laboratory of Geomechanics and Geotechnical Engineering Safety(Grant No.SKLGME-JBGS2402).
文摘With the increasing depth and intensity of coal mining operations,high-energy mine tremors have become a major trigger for rockburst disasters,posing severe threats to mine safety.Conventional rockburst risk assessment methods either lack real-time adaptability or rely heavily on qualitative microseismic data analysis,limiting their effectiveness in dynamic early warning.To address these limitations,this study proposed a predictive framework for rockburst risk assessment by integrating ensemble learning algorithms with Bayesian optimization.A dataset was constructed using a sliding time window approach,linking the highest MS energy in the subsequent days with predefined risk levels.Both undersampling and oversampling strategies were employed to mitigate class imbalance,and their performance was evaluated.Three ensemble models,i.e.CatBoost,Random Forest,and LightGBM,were developed,and their hyperparameters were optimized using Bayesian techniques to enhance predictive performance.The models were validated using MS data from the 6303 and 6306 working faces at the Dongtan Coal Mine.All three ensemble models outperformed conventional classification methods,particularly in accurately predicting high-risk categories.Among them,the CatBoost model exhibited the best performance,with an accuracy of 89.47%and an F1¯-score of 90.62%.Furthermore,SHapley Additive exPlanations analysis was used to enhance model interpretability,identifying key MS indicators influencing rockburst risk predictions.This study provides a systematic approach for leveraging MS data and machine learning to improve an early warning system for rockburst hazards,offering valuable insights for underground mining safety management.
