Table 3.

Standardized mean differences between one group and the other two combined, using the entire dataset

Variable	Grouped by log TN (unadjusted)			Adjusted by GPS
	[1.17, 1.78]	[1.78, 2.69]	[2.69, 4.08]	[1.17, 1.78]	[1.78, 2.69]	[2.69, 4.08]
ELEV	− 1.33	6.88	6.68	0.32	2.40	1.57
Longitude	− 7.78	− 7.07	− 15.73	− 1.94	0.51	− 1.79
Log(PRECIP)	12.06	2.98	13.20	0.79	− 0.07	2.61
Log(AREA)	− 4.73	− 7.02	− 12.46	− 1.85	− 0.74	− 0.76
Log(CL)	− 9.14	− 5.40	− 14.43	− 1.88	1.47	− 2.41
Log(HCO3)	− 9.28	− 4.77	− 13.59	− 0.94	− 1.52	− 3.24
Log(SO4)	− 7.74	− 8.00	− 17.23	− 0.76	− 0.63	− 4.16
SED	− 6.67	− 8.77	− 17.24	− 1.76	− 0.36	− 2.33
STRMTEMP	− 6.67	− 5.15	− 11.72	−0.64	0.26	− 1.32
Percent.AGT	− 3.71	− 12.21	− 18.96	− 1.15	− 4.67	− 7.54
Percent.URB	− 2.86	− 2.91	− 5.58	0.89	0.61	− 0.55
Percent.Canopy	− 5.72	− 5.40	− 11.19	− 1.72	0.56	− 1.14
Riparian.Disturb	− 8.04	− 3.88	− 11.15	1.11	0.60	− 1.85

The “Grouped by log TN” columns compare confounding factors in the three subsets divided by three continuous intervals of log TN (in brackets), each with about 1/3 of the total samples. The “Adjusted by GPS” columns compare the confounding factors based on GPS in two steps: (1) data were divided based on GPS into six groups and (2) data within each GPS group were further divided into three groups based on the same log TN brackets, and mean differences were calculated. The sample size-weighted averages of these mean differences are shown