Contamination of drinking water with human and animal faecal matter is the leading cause ofwaterborne disease worldwide. Populations in developing regions are at highest risk, oftenlacking an improved source of water and relying directly on catchment water resources.Water resources in karst geological terrain, which provides approximately 25% of the world’sdrinking water, are particularly susceptible to contamination. The southwest China karstregion contains some of the poorest provinces in China, making this a high priority area toadvance the knowledge base underpinning water resource management. Land-use, climatic,and hydrological variables are known to be key controls on faecal contaminationlevels. Faecal contamination is commonly measured via faecal indicators such asE. coli; however, there is inadequate availability of data to assess the relationshipbetween the suite of potential predictor variables and E. coli across the range ofland-use and geological scenarios relevant to contamination of water resources. Mixedland-use catchments present a particularly complex environment in which to quantifythe controls on E. coli dynamics due to the mosaic of land-use types and range ofpotential faecal sources. The objective of this study was to investigate how land-use,antecedent climate and meteorological conditions, and hydrology influence E.coli concentrations in receiving waters in a mixed land-use karst catchment in southwestChina. E. coli concentration was used as a measure of recent faecal contamination levels.Samples were collected at 30 sites at a bi-weekly interval between April and October 2018providing a long-term dataset over seasonal and agricultural cycles. Field parameters werealso collected for all samples, informing on residence times and water sources, andmeteorological data was available from a weather station within the catchment. E. coliconcentration ranged from below detection at epikarst springs on forested hillslopes to 8log10CFU/100 mL (colony forming units/100 mL) at a contaminated site below a sewagetreatment plant. The variation in E. coli concentration between sites was greaterthan the temporal variation during the monitoring period; the range of mean valuesbetween sites when all time points were compiled for each site was 4.29 log10CFU/100 mL, compared with 0.8 log10 CFU/100 mL when values from all sites werecompiled for each time point. Multi-variate regression modelling indicates karsthydrology and land-use explain the majority of variation in E. coli levels betweensites, while meteorological variables explain some of the temporal variation (workongoing). The results of this study are important because they suggest that land-use andhydrology are the most important controls on faecal contamination levels in thisenvironment, but that climatic variation can also cause periods of increased risk. This hasoptimistic implications for improving management across the extensive karst farmingregion of China, and potentially other karst regions of the world. E. coli samplingprogrammes combined with hydrological, land-use, and meteorological datasetscould be used to identify where mitigation strategies are required and the ability topredict periods of increased risk of high E. coli concentrations within catchments. [ABSTRACT FROM AUTHOR]