For multivariate analysis, data were z-score standardized and Euclidean distance matrices produced for each

parameter group. Permutational Multivariate Analysis of Variance (MANOVA) was used with GC# and site location as factors to determine if each category differed by stream and up and downstream of golf course facilities. Significant multivariate interactions were examined by trajectory analysis where the magnitude and direction of change for each stream and site location pair was explored ( Collyer and Adams, 2007). When interactions between stream and site location were not significant, multivariate post hoc tests LBH589 manufacturer were run to determine which streams differed. Multivariate categories for each sampling location were visualized with principle components analysis as biplots of components 1 and 2. Mantel and partial mantel tests and two block partial least squares were used to examine multivariate correlation between parameter groups. All statistical analyses were carried out in R 2.14.1 with the assistance of vegan and geomoph packages. Watershed area ranged for each sampling point from 10 to 93 km2. Anthropogenic land use (e.g., agriculture, development, tree plantations, etc.) ranged 48–78% among stream riparian zones (Table

1). The multivariate landscape group was PCI-32765 similar up and downstream of golf course facilities (Pillai’s Trace = 0.2, p = 0.914; Table 1; Fig. 2A). The landscape group significantly differed by stream (Pillai’s T = 16.9, p = 0.001). Post hoc comparison indicated that GC1 was only similar

to GC2 and GC5. The landscape of GC6 was Reverse transcriptase significantly different from GC2. The landscapes of GC2, GC3, and GC4 were similar ( Fig. 2A). Water quality among streams ranged from oligotrophic to eutrophic (Table 2). DOC ranged from 1.3 to 16.9 mg-C l−1 and was significantly lower downstream of golf courses (Wilcoxon’s paired test, p = 0.002; Fig. 3). SpCond, TDN, BACT, and BP were variable among sites but did not differ up and downstream of golf course facilities. TDP ranged from 4.1 to 44.1 μg-P l−1 and was significantly higher downstream of golf course facilities (Wilcoxon’s paired test, p = 0.023; Fig. 3). All together, the water quality group up and downstream of golf course facilities was similar (Pillai’s T = 0.2, p = 0.913), but significantly differed in water quality among streams (Pillai’s T = 14.3, p = 0.001; Fig. 2B). Post hoc comparison indicated that GC1 and GC2 were similar but significantly differed from the other streams, except between GC1 and GC5 which did not differ (p = 0.064). GC3, GC4, GC5, and GC6 had similar water quality. DOM ranged from strongly humic-like with features of terrestrial inputs (e.g., higher aromaticity (SUVA) and contributions of C2 and C3) to humic-like with features of microbial inputs (e.g.