Table 5.

Evidence organized by consideration for characterizing current background specific conductivity (SC) in the new area (Ohio Ecoregion 70) and for assessing its similarity to the original area (West Virginia Ecoregion 70)^a.

	Types of evidence	Results
A.	Regional properties
1	Similar climate across the region.	Climate is similar across the ecoregion. (+)
2	Similar geology across the region.	Geological strata are similar across the ecoregion. (++)
3	Similar physiography across the region.	The entire ecoregion has a hilly physiography. (+)
4	Geophysical modeling estimates low variance across the region.	Base-flow SC estimated by a geophysical model for the new area (195μS/cm) was 10.8% higher than modeled in the original area (176μS/cm). The model estimated that 76.2% of the new area has base-flow SC less than the background SC level in the original area (210 μS/cm). (++)
B.	Empirically estimated background
5	Lowest observed SC in the new area is approximately as low as in the original area.	The minimum SC was 40 and 49μS/cm in the original and new areas, respectively. (+)
6	Similar SC 10th or 25th centiles of the full data sets in both areas.	The 10th centile for the new area (163 μS/cm) is 26% higher than for the original area (120μS/cm). The 25th centile for the new area (244 μS/cm) is 31% higher than for the original area (169μS/cm). (−)
7	The proportion of measured sites less than the background limit from the original area (210μS/cm) is not negligible.	From a mixed data set, 19% of unique sites have measured SC ≤210μS/cm. (+)
8	The median of reference sites in the new area is less than the background level for the original area.	NE. Confirmed reference sites not available from Ohio. Sites identified in the Ohio data sets as reference sites often had poor habitat scores. (0)
9	In mixed data sets, the lower quartile of the new area should overlap the lower quartile of the original area in most months.	Both the lowest quartiles and the interquartile ranges of SC values overlap during months when more than 30 unique sites were sampled. (+)
C.	Distribution of low SC sites
10	Low SC sites (<210 μS/cm) are spatially dispersed.	Sites <210 μS/cm are widely distributed across Ohio Ecoregion 70, suggesting that they represent the regional background (Fig. 2c). (+)
11	Sites with SC less than a SC benchmark are spatially dispersed.	Sites ≤ 340 μS/cm are dispersed throughout the new area and represent 43.6% of unique sites in the new area (Fig. 2b). (+)
D.	Biological indicators of background SC
12	Salt-intolerant genera (XC95<340μS/cm) are found in the historical record.	Not examined. (0)
13	Salt-intolerant genera (XC95<340μS/cm) are found in recent records.	Seven genera occur in the new area with XC95 values <340μS/cm. (++)
14	Regional biota is similar.	Of the 139 genera found in the original area, 90.6% are represented in the data set for the new area. (+)
E.	Relevance and reliability of evidence
15	Larger data sets are more reliable.	The number of samples (4452) and sites (809) in the new area is sufficient to estimate background. (++)
16	Quality assured instream chemical measurements are more reliable.	Meter calibration and other good field and laboratory practices are required by Ohio EPA standard operating procedures. (+)
17	Appropriate reporting units are required.	Units are reported as SC. Other units and detection limits are as designated. (+)
18	Comparisons are appropriate only if ionic mixtures are similar.	Included samples in both areas are dominated by salts of HCO3‒ plus SO42‒ (Cl‒ represents less than 50% of anions by weight). (+)
19	Comparisons are more reliable if collection methods and sampling windows are similar.	The new area data set was collected throughout the year but was much more temporally concentrated and was not spatially distributed by a probability design. (−)
20	Validation increases reliability.	Other data sets are available for the new area, but they are very small (<20samples), so no validation was attempted. (0)
21	Inclusion of likely background sites makes identification of background SC more probable.	Small catchments, which tend to be in higher elevation and less disturbed, are included in all data sets, but the proportion of such sites is small. The proportion of land cover in native vegetation is not known. (0)
22	Abrupt changes in SC values at political boundaries make it likely that differences are not natural.	There are no abrupt changes in SC measurements at the state borders that divide the new area from the rest of the ecoregion. (+)
23−27	Not applicable, because reference sites for the new area are not useable.	No results. (0)

SC = specific conductivity, NE = no evidence