『Abstract
　Globally, it is possible that up to 25% of the world's population depends on karst water supplies. In karst areas, a high degree of groundwater and surface water linkage often results in the direct recharge of groundwater with polluted run-off following rainfall. In order to reveal the hydrochemical variations after rainfall, especially real-time variation of pollutants, high-resolution auto-monitoring techniques were used at the outlet of Qingmuguan subterranean stream (QSS), which is influenced by agricultural activities. In addition to rainfall, high-resolution measurements of pH, water level, electrical conductivity (Ec) and NO3^- concentration were recorded in the monsoon season and fertilizer application period using a data logger with time intervals of 15 min. In the six observed rainfall events, the pH value was mainly controlled by acidic rainfall inputs. The pH showed sharp decline after the rainfall event, and then increased. The Ec was impacted by the rainfall chemistry, dilution effect of rainfall and agricultural wastewater. NO3^- derived from agricultural activities was less impacted by rain chemistry; and its variations were mainly affected by the dilution effect of rainfall and agricultural wastewater. Under the influences of the R1 rainfall, the rapid changes of Ec and NO3^- were opposite in direction. As the rain continued, both the Ec and NO3^- rapidly changed in synchronization within the shortest period of 5 h and the longest of 27 h because of the impact of the agricultural wastewater. The groundwater quality changed due to the influx of agricultural wastewater over the entire monitoring period. According to the National Groundwater Quality Standard, People's Republic of China (GB/T 14848-9), the groundwater quality of the QSS moved through the following grades during the monitoring period: Grad III→Grade IV→Grade V→Grade IV→Grade V→Grade IV→Grade V→Grade IV→Grade III. Traditional sampling methods did not reveal accurate hydrochemistry changes of the QSS, and even generated misleading results. Consequently, the high-resolution auto-monitoring technique is necessary for the future protection and sustainable use of karst aquifer in Southwest China.

Keywords: High-resolution auto-monitoring; Subterranean stream; hydrochemistry; Nitrate real-time variations; Agriculture』

Introduction
Physical setting of the study site
Methods
Results and discussion
　Analyzing continuous monitoring data
　　The first rain event (R1)
　　The second rain event (R2)
　　The third rain event (R3)
　The remaining several rain events (R4, R5, R6)
　Contrast of auto-monitoring logger data and experimental result
　Variation of groundwater quality
Conclusion
Acknowledgments
References