Water quality assessment and apportionment of pollution sources using APCS-MLR and PMF receptor modeling techniques in three major rivers of South Florida

Mohammad Haji Gholizadeh; Assefa M Melesse; Lakshmi Reddi

doi:10.1016/j.scitotenv.2016.06.046

Water quality assessment and apportionment of pollution sources using APCS-MLR and PMF receptor modeling techniques in three major rivers of South Florida

Sci Total Environ. 2016 Oct 1:566-567:1552-1567. doi: 10.1016/j.scitotenv.2016.06.046. Epub 2016 Jun 15.

Authors

Mohammad Haji Gholizadeh¹, Assefa M Melesse², Lakshmi Reddi³

Affiliations

¹ Department of Civil Engineering, Florida International University, 10555 W Flagler Street, EC3781, Miami, FL 33174, USA. Electronic address: mhaji002@fiu.edu.
² Department of Earth and Environment, Florida International University, AHC-5-390, 11200 SW 8th Street, Miami, FL 33199, USA. Electronic address: melessea@fiu.edu.
³ Department of Civil Engineering, Florida International University, PC 230, 11200 SW 8th St., Miami, FL 33199, USA. Electronic address: lreddi@fiu.edu.

PMID: 27317134
DOI: 10.1016/j.scitotenv.2016.06.046

Abstract

In this study, principal component analysis (PCA), factor analysis (FA), and the absolute principal component score-multiple linear regression (APCS-MLR) receptor modeling technique were used to assess the water quality and identify and quantify the potential pollution sources affecting the water quality of three major rivers of South Florida. For this purpose, 15years (2000-2014) dataset of 12 water quality variables covering 16 monitoring stations, and approximately 35,000 observations was used. The PCA/FA method identified five and four potential pollution sources in wet and dry seasons, respectively, and the effective mechanisms, rules and causes were explained. The APCS-MLR apportioned their contributions to each water quality variable. Results showed that the point source pollution discharges from anthropogenic factors due to the discharge of agriculture waste and domestic and industrial wastewater were the major sources of river water contamination. Also, the studied variables were categorized into three groups of nutrients (total kjeldahl nitrogen, total phosphorus, total phosphate, and ammonia-N), water murkiness conducive parameters (total suspended solids, turbidity, and chlorophyll-a), and salt ions (magnesium, chloride, and sodium), and average contributions of different potential pollution sources to these categories were considered separately. The data matrix was also subjected to PMF receptor model using the EPA PMF-5.0 program and the two-way model described was performed for the PMF analyses. Comparison of the obtained results of PMF and APCS-MLR models showed that there were some significant differences in estimated contribution for each potential pollution source, especially in the wet season. Eventually, it was concluded that the APCS-MLR receptor modeling approach appears to be more physically plausible for the current study. It is believed that the results of apportionment could be very useful to the local authorities for the control and management of pollution and better protection of important riverine water quality.

Keywords: APCS–MLR; PMF; Pollutants; Source apportionments; South Florida; Water quality.

MeSH terms

Environmental Monitoring / methods*
Factor Analysis, Statistical
Florida
Linear Models
Principal Component Analysis
Rivers / chemistry*
Water Pollution / analysis*
Water Quality