Assessment of surface water quality using self-organizing map and Hasse diagram technique
基于自组织映射与哈斯图方法的地表水水质评价研究

Surface water quality is a very important environmental issue for the management and protection of water resources. The primary purpose of this study was to apply self-organizing map (SOM) and Hasse Diagram Technique (HDT) to classify, model, interpret and evaluate the monitoring data for surface water quality. During the period between June 2010 and May 2011, 600 monthly samples were collected from 50 sites in West Tiaoxi River Basin, China. Based on SOM analysis, all the samples were divided into 10 clusters which were related with the water quality status of wet, flat, and dry seasons in the area. pH and CODMn had little change during the period. DO showed relatively high values in the dry season, indicating that the water had a certain degree of self-purification capacity. The values of NH4+-N, TN and TP were relatively small in wet season. The average of TN:TP (mass ratio) in West Tiaoxi River Basin was 46, with a range from 2 to 482 during monitoring period. About 66.7% of water samples were in the state of the phosphorus limitation, with only 3.2% in nitrogen limitation status in the period, indicating the occurrence and alternation of phosphorus limitation and nitrogen limitation states. The entire watershed generally appeared to be in phosphorus limitation state. By selecting CODMn, NH4+-N, TN and TP for HDT comprehensive analysis, 50 monitoring sites in West Tiaoxi River Basin were divided into eight levels. Water quality in the upstream was better, but relatively poor in midstream and downstream. Sensitivity analysis showed that CODMn had the highest sensitivity value (235), followed by TN (156), NH4+-N (61), and TP (34), indicating that among the selected surface water quality indicators, CODMn and TN were the two important parameters with high environmental risk. Based on these results, we suggested to consider developing techniques and relevant policies to reduce the pollution of CODMn and TN in the basin.