Identifying key environmental factors influencing spatial variation of water quality in upper Shitoukoumen Reservoir Basin in Jilin Province, China

Based on the observed data in monitored drainage areas and GIS spatial analysis tools, watershed basic database of Shitoukoumen Reservoir Basin was built. The multivariate analysis and redundancy analysis (RDA) were used to analyze the spatial and temporal variations of water quality, identify the key environmental factors and their patterns influencing the spatial variation of water quality, and determine the main types and forms of the non-point source (NPS) pollutant export controlled by the key environmental factors. The results show that different patterns of environmental factors lead to great changes in water quality at spatial and seasonal scales. All selected environmental factors explain 64.5% and 68.2% of the spatial variation of water quality over dry season and rainy season, respectively, which shows clear seasonal difference. Over dry season, residential land is the most important environmental factor, which possesses 35.4% of the spatial variation, and drainage area is the second key environmental factor, which possesses 17.0% of spatial variation in the total variance. Over rainy season, slope length and drainage area are the key environmental factors, which possess 29.3% of the spatial variation together. Residential land influences nitrogen export by changing NH4 (+)-N and particulate organic nitrogen (PON) discharge over dry season, and drainage area controls phosphorus export by regulating dissolved phosphorus (DP) drainage over dry season and phosphorus associated particulate (PAP) loss over rainy season, respectively. Although slope length is an important environmental factor, it does not influence NPS pollutant export. It is interesting that soil organic matter, as a minor environmental factor, highly determines phosphorus and nitrogen export by enhancing the DP, PAP and PON loss.