Geoffrey M. Edwards, Environmental Specialist

Peter A. Tennant, P.E., Deputy Executive Director

John T. Lyons, P.E., Water Pollution Control Programs Manager

Ohio River Valley Water Sanitation Commission (ORSANCO)

5735 Kellogg Ave., Cincinnati, OH 45228

This paper presents findings of biological studies associated with a two-year, $2 million national demonstration project conducted on a segment of the Ohio River, which includes the Cincinnati/Northern Kentucky metropolitan area. Physical and chemical monitoring were also included in the project, which identified fecal coliform bacteria as the parameter with the most frequent and excessive violations of water quality standards. Biological community monitoring was undertaken to determine if direct impacts on aquatic life could be identified and attributed to wet weather sources.

Key Terms: Biological community monitoring, wet weather pollution sources, combined sewer overflows (CSOs), electrofishing, Hester-Dendy sampler unit

The primary objective of the study’s biological component was to utilize existing methods of biological sampling to determine the effects of wet weather pollution on the biological community of a large river. To achieve this goal both fish and macroinvertebrate populations were sampled.

Surveys were conducted in a segment of the Markland Pool- Ohio River Mile Points (ORMP) 462 to 492. Figure 1 displays the study area in relationship to the Ohio River Basin.

The scope of work for this study included a fish population assessment and the collection of macroinvertebrate samples on the Ohio River at designated locations in Markland Pool. The fish population surveys were conducted in three rounds of sampling at 21 sites. The macroinvertebrate samples were collected in four rounds of sampling, each round consisting of an eight-week colonization period. Objectives of the biological monitoring were to: (1) provide upstream background data, (2) examine the effects of major identifiable pollutant inputs (clusters of combined sewer overflows, sanitary sewer overflows, wastewater treatment plants, and tributaries) within the Greater Cincinnati/Northern Kentucky urban area, and (3) investigate the level of downstream recovery relative to upstream conditions within the confines of the study area.

The fish population assessment conducted for this component consisted of sampling 21 sites during Rounds One and Three, and 23 sites during Round Two (Table 1). Sites were chosen to produce optimum coverage for the study area. The surveys focused on sampling similar habitat areas (mud/gravel substrate) to reduce environmental variability as much as possible.

The fish population assessment was conducted in cooperation with the Ohio Environmental Protection Agency (Ohio EPA). Sampling was conducted from September 18 - 26, 1995 (Round One), August 13 - 29, 1996 (Round Two), and October 7 - November 6, 1996 (Round Three). Sites were approximately 500 meters in length, and were sampled at night to optimize catch abundance and diversity (Ohio EPA, 1987). Fish collected were counted, measured and identified to the lowest practical taxonomical level on site. All minnows and questionable identifications were preserved on site and later identified by Ohio EPA staff.

Macroinvertebrate samples were obtained using Hester-Dendy artificial substrate multi-plate sampler units. Sampler units consisted of five individual samplers/sampler unit. Each unit of five was anchored to a cement block at the sampling site to stabilize and submerge the unit. Macroinvertebrate sampling consisted of three distinct phases:

Phase One established 16 macroinvertebrate "longitudinal" sampling stations consisting of one Hester-Dendy sampling unit per station within the study area (Table 2). Sites between ORMP 462 and 492 were sampled at regular intervals (approximately two miles) for four rounds, each round lasting eight weeks.

Phase Two established three macroinvertebrate "cluster" sampling stations consisting of five Hester-Dendy sampling units per station (Table 2). Four rounds of sampling were also conducted at these sites, each round lasting eight weeks. The objective was to identify the extent of natural variability in macroinvertebrate populations within the study area.

Phase Three isolated the near-field effects of individual CSOs within the study area. In the first year of sampling (1995) individual sampling units were placed above and below each of three CSO discharges. In the second year the number of sites was expanded to six. Dye tests were used to determine the location of the sampling sites (Table 2). Each CSO was monitored for overflow frequency and duration. Sites were sampled for a total of three rounds, each round lasting eight weeks. The objective was to determine if overflows produce a measurable near-field impact on macroinvertebrate populations below the outfall.

The colonization period for macroinvertebrate samples was eight weeks. Sampling was conducted July 12 - September 5, 1995 (Round One), August 31 - October 26, 1995 (Round Two), July 9 - August 29, 1996 (Round Three), and August 27 - October 16, 1996 (Round Four). Recovery rates of the sampler units were as follows: Round One - 27 of 34 (79.4%), Round Two - 20 of 38 (52.6%), Round Three - 37 of 44 (84.1%), and Round Four - 41 of 44 (93.2%). Table 2 lists the specific units recovered for each round. Once retrieved, the individual plates from each sampler unit were processed in the field and the resulting composite of organisms stored in a preservative for shipment. Composites were sent to an independent laboratory, where they were counted and identified to the lowest practical taxonomical level.

Field measurable water quality parameters were collected at each site at the time of placement and retrieval of sampler units. Temperature, pH, dissolved oxygen and conductivity were recorded using a Hydrolab Model H-20 instrument. The Hydrolab instrument was pre- and post-calibrated to ensure the accuracy of the data collected.

Total precipitation was measured at gauges throughout the study area during Rounds One through Four, and is expressed below as an average of all the gauges within the study area (Table 4).

Total flow from the CSOs where sampler units were placed was monitored during Rounds Two through Four (CSO samplers were not placed during Round One). An example of the information collected at each of the sampling locations is expressed in Table 5.

Sites were evaluated with the help of Ohio EPA personnel who calculated an Index of Biotic Integrity (IBI) and a Modified Index of well-being (MIwb). It should be noted that both indices were designed to evaluate fish populations in inland streams and waterways. Since an index designed and calibrated specifically to evaluate fish populations for a large river like the Ohio River has not been developed, the IBI and MIwb were utilized in their present form.

The IBI is a multi-metric approach to evaluating fish populations and was originally described by Karr (1981) for use in Illinois streams. Ohio EPA uses a modified version of the IBI which takes into account regional differences between the fish populations of Ohio and Illinois. It consists of twelve metrics which are compared to the value expected at a reference site and then rated either a 5 (value approximates), 3 (deviates somewhat from) or 1 (strongly deviates from the value expected). The maximum IBI score is 60 and the minimum score is 12. Metrics used by Ohio EPA are: total number of species, sunfish species, sucker species, intolerant species, round body suckers, simple lithophils, tolerant fishes, omnivores, top carnivores, insectivores, DELT anomalies (Deformities, Eroded fins, Lesions, and Tumors), and relative number minus tolerant species (Ohio EPA, 1987).

IBI values expected at a reference site for the study area (Interior Plateau Ecoregion) for an Ohio inland stream would have a mean value of 43, a standard deviation of 1.1, and a range of 32 - 52 (Ohio EPA, 1987). Results from the Ohio River samples collected in 1995 (Round One) and those collected in 1996 (Rounds Two & Three) are displayed below (Table 6).

The MIwb is also a multi-metric approach to evaluating fish populations and was originally developed as the Index of well-being (Iwb) by Gammon (1976) for use on the Wabash River in Indiana. The Iwb consists of four measures of fish communities: numbers of individuals, biomass, Shannon Diversity based on numbers, and Shannon Diversity based on weight. Ohio EPA modified the Iwb by eliminating any of 13 highly tolerant species, hybrids, or exotic species from the numbers and biomass components of the Iwb, but not from the Shannon components (Ohio EPA, 1987). A minimum MIwb score is 0 and the maximum is 12.

MIwb values expected at a reference site for the study area (Interior Plateau Ecoregion) for an Ohio inland stream would have a mean value of 9.2, a standard deviation of 0.1, and a range of 8.5 - 10.2 (Ohio EPA, 1987). Results from the Ohio River samples collected in 1995 (Round One) and those collected in 1996 (Rounds Two & Three) are displayed below (Table 7).

A total of 125 composite samples was collected over four rounds of sampling (Table 2) and was evaluated with the help of Ohio EPA staff. The following indices were calculated for each composite sample: total number of organisms, taxa richness, percent Dominant taxa, percent Chironomids, Chironomid richness, percent EPT (Ephemeroptera, Plecoptera, and Trichoptera), EPT richness, EPT/Chironomid ratio, Modified Hilsenhoff’s Biotic Index, Community Loss Index, Jaccard Coefficient, Ohio EPA’s Invertebrate Community Index (ICI) and various analyses associated with the Zebra Mussel population. It should be noted that these indices were designed to evaluate macroinvertebrate populations in inland streams and waterways and were used in the absence of an index specifically designed and calibrated to evaluate macroinvertebrate populations for a large river like the Ohio. Sites were compared statistically based on a mean value at each cluster location and standard deviation at 95 percent confidence level. The expectations prior to the initiation of sampling was that the indices would reflect a higher biological integrity at the upstream sites as compared to the sites within the urban area. In addition, it was expected that the downstream sites would display a recovery in the composition of the macroinvertebrate community which would more closely represent the upstream conditions. Similarly, the expectation was that indices would reflect higher biological integrity upstream of the CSO sites than immediately downstream. CSO site O-3 offered two years of results which are used below as representative of all of the CSO locations studied.

This index is simply a count of the organisms found in each macroinvertebrate sample. The expectation prior to sampling was that the total number of organisms would be highest at the upstream sites, show a decrease through the urban area and a recovery at the downstream sites. At the CSO sites, the expectation prior to sampling was that the total number of organisms upstream of the overflow would be higher than the number downstream. Results are summarized as follows:

• Cluster and Longitudinal site sampling resulted in the expected trend (Figure 2).
• The number of organisms above CSO O-3 was statistically the same as the number below the outfall during Rounds Two through Four. This does not confirm the expectation prior to sampling.

Taxa richness is simply a count of the taxa found in each macroinvertebrate sample. In this case, all roundworm taxa were counted as only one taxa per Ohio EPA protocol. The expectation prior to sampling was that the total number of taxa would be highest at the upstream sites, show a decrease through the urban area, and a recovery at the downstream sites. At the CSO sites the expectation prior to sampling was that the total number of taxa upstream of the overflow would be higher than the number downstream. Results are summarized as follows:

• The number of taxa at Cluster and Longitudinal sites show a steady, increasing trend throughout the study area which does not conform to expectation.
• The site O-3 CSO samplers recovered in Rounds Two through Four either show an increase in number of taxa from upstream to downstream of individual outfalls or remain statistically the same - opposite of the expectation. However, of the additional taxa below the outfall, a majority are of the family Chironomidae which are generally considered to be pollution tolerant organisms.

This index is a measure of the percentage of the Family Chironomidae (Midges) within the community found at each site. These organisms are generally tolerant of pollution and their numbers tend to increase in degraded conditions. The expectation prior to sampling was that the percentage of chironomids would be lowest at the upstream sites, show an increase through the urban area and decrease at the downstream sites. At the CSO sites, the expectation prior to sampling was that the percentage of chironomids upstream of the overflow would be lower than the percentage downstream. Results are summarized as follows:

• The Cluster site samples conformed to expectation, however the Longitudinal site samples do not meet the expectation (Figure 3).
• Of the site O-3 CSO samplers recovered in Rounds Two through Four, the mean Percent Chironomids above the CSO was 36.84 and the percentage below the outfall was 51.66 (statistically significant) indicating the expected performance. This may suggest that a CSO can have a quantifiable effect upon near-field macroinvertebrate communities even on large rivers where tremendous dilution can occur.

This index is a measure of the percentage of the Orders Ephemeroptera (Mayflies), Plecoptera (Stoneflies), and Trichoptera (Caddisflies) within the community found at each site. These organisms are generally considered to be pollution-sensitive species. The presence of EPT organisms at a site is generally an indicator of good water quality, since their sensitivity precludes them from inhabiting degraded areas. The expectation prior to sampling was that the percentage of EPT will be highest at the upstream sites, show a decrease through the urban area and a recovery at the downstream sites. At the CSO sites the expectation prior to sampling was that the percentage of EPT upstream of the overflow would be higher than the percentage downstream. Results are summarized as follows:

• The number of taxa at Cluster and Longitudinal sites show a steady, increasing trend through the study area which does not conform to expectation.
• Of the Site O-3 CSO samplers recovered in Rounds Two through Four, mean percentage of EPT above the CSO was 23.42 and the percentage below was 12.35. Based on this information, the difference in the percentage of EPT was statistically significant and indicated the expected performance.

This family level index was originally developed by Hilsenhoff (1977) for use in Wisconsin streams. The original tolerance classifications were based on a numerical range of 0 to 5 and later modified by Hilsenhoff (1987) to use a 0 to 10 scale. However, similar results can be obtained using an index value of either 0 to 5 or 0 to 10, and adequate information is not available for several species that would allow use of the more definitive 0 to 10 tolerance range (U.S. EPA, 1990). Therefore, a 0 to 5 scale was chosen as modified by the Maryland Department of the Environment (1992). Higher index values indicate a more pollution tolerant macroinvertebrate community, and generally a lesser degree of water quality. A score of:

0 to 1.75 = excellent water quality

1.76 to 2.50 = good water quality

2.51 to 3.75 = fair water quality

3.76 to 4.00 = poor water quality

> 4.00 = serious water quality problems

The expectation prior to sampling was that the HBI score would be lowest at the upstream sites, show an increase through the urban area and a decrease at the downstream sites. At the CSO sites the expectation prior to sampling was that the HBI score upstream of the overflow would be lower than the score downstream. Results are summarized as follows:

• Cluster and Longitudinal site sampling resulted in the expected trend.
• Of the Site O-3 CSO samplers recovered in Rounds Two through Four, mean HBI score above the CSO was 3.33 and the mean below was 3.66 (statistically significant) indicating the expected performance.

This index is a measure of the percentage of the Zebra Mussels (Dreissena polymorpha) and Asiatic Bivalves (Corbicula fulminea) within the community found at each site. Since the mussels were the largest contributors of organisms to the population in many samples, this index may also be considered the percentage of dominant taxa for those samples. These two mussel taxa are generally tolerant of pollution and their numbers tend to increase in degraded conditions. The expectation prior to sampling was that the percentage of mussels would be lowest at the upstream sites, show an increase through the urban area and decrease at the downstream sites. At the CSO sites, the expectation prior to sampling was that the percentage of mussels upstream of the overflow would be lower that the percentage downstream. Results are summarized as follows:

• The Cluster and Longitudinal sites displayed the opposite of the expected trend with high numbers upstream, lower numbers through the urban area, and increasing numbers at downstream sites.
• At the Site O-3 CSO samplers recovered in Rounds Two through Four, the percentage of mussels above the CSO was 18.63 and the percentage below the outfall was 0.86. This seems to confirm that these mussels may be more sensitive to urban influences than originally expected.

The indices used to evaluate the fish population assessment and the macroinvertebrate collections conducted in 1995 and 1996 were designed to evaluate inland streams as opposed to a large river like the Ohio River. Given that basis, the results of these analyses must be viewed with a certain amount of caution. ORSANCO is well aware that any attempt to evaluate water quality conditions using biological populations on the Ohio River must be conducted with new indices designed for, or existing indices calibrated for, the special conditions which exist on large rivers (i.e., large amounts of flow, transient sediments, etc.). However, biological populations have been valuable assessment tools for smaller streams, and quite probably will prove to be of similar value on large rivers in the future. In the interim, biological results from this project did show some interesting results using available methods of evaluation.

In general, for the fish population assessed during this study, neither of the two indices was able to demonstrate any consistent, statistically reasonable difference between the upstream sites, the urban sites, and the downstream sites. It is important to note that the standard deviation for both the IBI and MIwb was rather high. As sampling efforts continue both river-wide and in the study area, and as the sample size becomes more robust, the standard deviation should be compressed for both of the indices.

In Rounds One through Four, several indices performed as expected. In particular, total number of organisms, percent chironomids, Hilsenhoff’s Biotic Index, and percent mussels showed clear statistically significant results over the study area. As with the fish population indices, it is clear that a larger, more robust sampling size is important to compress the standard deviations for many indices.

Of the CSO samples recovered in Rounds Two through Four, several indices performed as expected. In particular, percent Chironomids, percent EPT, EPT/Chironomid ratio, Hilsenhoff’s Biotic Index and the Invertebrate Community Index showed clear statistically significant differences in the makeup of the macroinvertebrate populations above and directly below particular CSOs. Future sampling efforts should focus on sampling at a variety of outfalls and at different seasons. There is evidence that, at least at these sites, CSOs have a quantifiable impact on the near-field populations of organisms irrespective of the unique qualities that large rivers, like the Ohio River, possess. Once these impacts are defined, efforts could be focused on examining the length of the impact in terms of distance downstream.

The authors would like to acknowledge the efforts of a number of individuals from Ohio EPA particularly, Chris Yoder and Jeff DeShon, who provided a great deal of assistance toward the completion of the biological component of this study. In addition to actually participating in the electrofishing surveys, representatives from OEPA also conducted specific evaluations of both the fish and macroinvertebrate data which proved extremely valuable in efforts to interpret the information collected. Special thanks to Bob Ovies of ORSANCO for assistance with graphics and HTML formatting.

We would also like to recognize the tremendous amount of in-kind and monetary resources dedicated to this study from the US EPA, the Metropolitan Sewer District of Greater Cincinnati, and the Sanitation District No. 1 of Northern Kentucky.