site as well as the values of its upstream and downstream stations. Data quality assessment was done at multiple stages. I inspected the distribution of the original data using descriptive statistics and graphical representations (histogram,
box-whisker plot). Because the distributions of BOD,COD, SS, TP, and TN are positively skewed, the original data were log10(x+1) transformed. Any unusual or suspicious outliers were detected and then either reestimated (e.g.,decimal points are off) or removed from regression analysis.The data quality assessment follows the US EPA guidance (US Environmental Protection Agency (USEPA), 2007).