Identification of River Water Pollution Characteristics Based on Projection Pursuit and Factor Analysis

Hong Huang,Jun Lu
DOI: https://doi.org/10.1007/s12665-014-3247-0
IF: 2.8
2014-01-01
Environmental Earth Sciences
Abstract:Human activities contribute different pollutants to receiving waters, often with significant variations in time and space. Therefore, integrating multiple parameters of water quality and their spatiotemporal variations is necessary to identify the pollution characteristics. Based on the water quality monitoring data with 12 parameters for 2 years at 22 sampling sites in the Cao-E River system, eastern China, the projection pursuit method was used to project all parameters and their temporal variations into a one-dimensional vector through two projections. Accordingly, we could easily assess the comprehensive water quality in different sampling sites and then classify their water pollution features. Factor analysis was then used to identify the pollution characteristics and potential sources. Results showed that all sampling sites for the river system could be classified into four groups: headwater sites (HS), agricultural nonpoint sources pollution sites (ANPS), point sources pollution sites (PSPS), and mixed sources pollution sites. Water quality in HS was good, containing only a few nutrients from the woodland runoff and soil erosion. For ANPS, the main pollutants were dissolved phosphorus, total P, and nitrate nitrogen (NO3 −-N), mainly from farming land. For PSPS, ammonium nitrogen (NH4 +-N) and organic pollutants originated from industrial and municipal sewage. In HS and ANPS, NO3 −-N was the main form of nitrogen, and a high ratio of NO3 −-N/NH4 +-N was a remarkable characteristic, whereas NH4 +-N was the main form of nitrogen in PSPS. Except in HS, water quality in the other groups could not meet the local water quality control standard. Finally, suggestions were proposed for water pollution control for the different groups.
What problem does this paper attempt to address?