[Influence of spatial difference on water quality in Jiuxiang River watershed, Nanjing].

Exploring the quantitative relationship between landscape characteristics and surface water quality indicators can provide important information to urban landscape planning and water environment protection. Jiuxiang River watershed in Nanjing city, being as a typical case study area, its landscape classification maps was conducted in 2009 by remote sensing digital images interpretation. Based on the remote sensing images and real-time monitoring data from October 2009 to September 2010, geospatial analysis and statistical analysis were integrated to explore the relationship between landscape composition, landscape pattern (landscape-level and class-level) and river water quality at the subwatershed scale. Results showed that most of landscape compositions influenced river water quality. Percentage of built-up land and unused land was positively related to total nitrogen (TN), total phosphorus (TP), ammonia nitrogen (NH4(+) -N) and permanganate index, while percentage of forestland showed a negative relationship. At the landscape level, the water quality was good when the size of patch was big. But at class-level, the aggregated distribution of the built-up land, unused land and arable land might cause TN, TP, permanganate index and NH4(+) -N concentration increased. But the aggregated distribution of the forestland had the opposite effect on water quality indicators.