This report describes Southern California Edison’s (SCE) study of aquatic benthic macroinvertebrates in the Kern River for relicensing of the Borel Hydroelectric Project (FERC Project No. 382). The study documents the recent conditions of the community of benthic macroinvertebrates in the portion of the Kern River affected by the Borel Project. It evaluates the effects of the project on the macroinvertebrate community and, more generally, on the biological integrity of the river’s ecosystem. The study also assesses the availability of benthic macroinvertebrate food resources for the river’s fish populations.

An initial plan for this study was prepared at the request of the Sequoia National Forest (SQF), in their comments on the First Stage Consultation Package for the Borel Project relicensing. The final study plan was developed in consultation with the State Water Resources Control Board (SWRCB), the California Department of Fish and Game (CDFG), the SQF and other interested parties and conforms to the sampling and analysis protocols of CDFG’s California Stream Bioassessment (CSBA) program. The SWRCB recently has encouraged use of CSBA procedures for their Section 401 Water Quality Certification program for hydropower projects.

This report discusses the objectives and methods of the study, and presents the results and discusses their significance. A detailed description of the Borel Project facilities, and an overview of the environmental setting, can be found in SCE’s First Stage Consultation Package. This is available on request, by calling Candace Irelan at 909-394-8714, or it can be downloaded from www.resourceinsights.com.

The objectives of the benthic macroinvertebrate study were:

• to describe the macroinvertebrate community in the portion of the Kern River affected by the Borel Project,
• to assess the biological integrity of the river ecosystem using CSBA community metrics,
• and to evaluate the macroinvertebrate food resources for fish in the river.

The results of the study provide the CDFG an opportunity to expand their CSBA database, and enhance the information used by SWRCB and others to evaluate effects of hydropower projects in California.

The results for the sampling conducted near the Borel Powerhouse may also be useful as an indicator of possible effects of algicide treatments in the Borel Canal. This will be considered in a separate report.

The field data collection was carried our over a five-day period from February 19 through 23, 2001. The data collection and analytical methods used for the study conform to CDFG’s CSBA protocols.

General locations for seven reaches of the Kern River were selected for sampling in consultation with staff of the SQF, CDFG, SWRCB, interested parties, and members of the general public. The specific locations of these reaches were determined during onsite inspection of available habitat. The station reach locations and their lengths are given in Table 1.

The locations for Stations 1 through 6 are shown in Figure 1. Station 7 is not included on the map because it is distant from the other stations, located in the North Fork Kern River above Lake Isabella. The station elevations varied from 2,270 feet msl for Station 1 to 2,640 feet msl for Station 7.

The sampling stations can be grouped according to potential Project influence. Group 2 stations are in the diverted reach of the Kern River and include Stations 3 through 6. The benthic macroinvertebrates at these stations are subject to potential effects of diverting flow from the river and receiving flow from the lower depths of Lake Isabella. Group 1 includes Stations 1 and 2, which are downstream of the Powerhouse where discharge from the Canal generally augments the river’s flow. The Powerhouse tailrace discharges directly into Station 2. It should be noted that neither the diverted reach nor the reach below the Powerhouse has a natural flow regime as flows in both reaches are strongly affected by the COE’s operations of the reservoir. Group 3 has no Project effects. It includes Station 7 only, which is upstream of Lake Isabella and downstream of SCE’s Kern No. 3 Project. This location is relatively unaffected by human impacts on flow. This station was included in the study as a relatively undisturbed site for purposes of comparison.

The specific locations sampled within each of the sample station reaches were selected following the CSBA protocol for "non-point source sampling". This protocol is used to assess conditions within an entire stream section. The protocol calls for selecting stream locations with productive riffle habitats. Riffle habitat is defined as low to moderately low gradient habitat with cobble substrate. These locations were the sampling station reaches identified in Table 1. Four or five riffles were identified along the stream channel in each station reach and three of the riffles were randomly selected for sampling. A sampling transect was selected in each riffle by pacing off the distance along the bank of the entire riffle and randomly selecting a meter interval along the upstream third of the riffle. Modifications to this protocol were necessary for Stations 1, 2 and 6 because suitable riffle habitat was limited at these locations. Stations 1 and 6 each had three closely spaced riffles. One transect was located in each of these riffles. Station 2 had only one small riffle, which was sampled. The other sampling for this station was conducted in areas of gravel substrate in and directly across from the tailrace.

Table 1

Figure 1 (Coming from Mary Engbring – She should be delivering it to us Friday)

Three sites were sampled along each sampling transect of the stations (or, for Station 2, in each of the areas of gravel substrate) and the material collected from the three sites was combined into one composite sample. Where the substrate and structure of the stream channel along the transect was fairly uniform, the three sampled sites were located near each side and the center of the stream. Where the substrate and channel structure were not uniform, the three sampled sites were selected to represent the variation present.

A kick sample of 2 to 3 minutes duration was collected at each sample site along each transect. A D-framed kick net (12-inch x 10-inch with a 0.5 mm mesh) was used for most of the stations. However, because flow in the North Fork Kern River upstream of Lake Isabella (Station 7) was relatively high, this station was sampled with a 40-inch x 40-inch Zo Seine with handles (0.5-mm mesh). During sampling with either net, one biologist held the net immediately downstream of a 1-foot x 2-foot patch of riffle substrate. A 17-inch wrecking bar (crow bar) was used by the second biologist to uproot cobbles so they could be hand-rubbed to wash invertebrates into the current. The curved end of the wrecking bar was also used to rake the smaller substrate materials. Unmovable substrates were rubbed by hand to wash invertebrates free of the surface. The three samples from each sampling transect were composited to form one sample per transect. The lid of each sample container was labeled with the date, stream name, and station code. Each sample was then drained of water and filled with a 95 percent ethanol solution. A total of 21 samples (three composite samples per sampling station) were collected.

Water temperature, dissolved oxygen, pH, electrical conductivity and standard physical habitat characteristics were recorded for each station reach. Information on the physical habitat characteristics of each station was recorded using two CSBA worksheets: the "Physical Habitat Quality" evaluation worksheet and the "California Bioassessment Worksheet" for habitat descriptions. Each sampling transect or area was photographed.

A laboratory specializing in stream macroinvertebrates processed the 21 samples collected. The lab identified the taxa in the samples to genus or the lowest practical taxonomic level and enumerated the organisms in each taxon. Subsampling was used for samples with more than 300 organisms. The laboratory used CSBA quality assurance (QA) procedures during sample processing. Five of the samples were randomly selected and sent to the CDFG for validation of the lab’s taxonomic identifications.

The results from the laboratory were used to compute 19 CSBA community metrics (Table 2). Eleven of the metrics that provided the most meaningful results were statistically analyzed for differences among the stations. For each of these metrics, the Analysis of Variance (ANOVA) procedure (Sokal and Rohlf 1981) was used to test for any difference among the station means. Where a significant difference was found, Tukey’s multiple comparison method was used to compare the individual means (Analytical Software 1996). As indicated in Table 2, many of the metrics are computed

Table 2

as percentages. Percentage data are generally not normally distributed, which is an important assumption of ANOVA. The arcsine transformation is the most commonly used method for obtaining a normal distribution from percentage data (Sokal and Rohlf 1981), so the percentage metrics were arcsine transformed prior to being analyzed.

The California Bioassessment Worksheet habitat characteristics of the seven station reaches and their sampling transects are provided in Table 3. Note that for all the stations, transect a is the most downstream sampling location. In Station 2, transect c was located in the Powerhouse tailrace, while transects a and b were downstream where water from the tailrace and the Kern River were mixing.

The recorded habitat conditions do not show any consistent differences among stations in Groups 1 (downstream of Powerhouse) and 2 (diverted reach of Kern River). However, Station 7 (Group 3) was deeper and, even though it was sampled in the late afternoon, its water temperature was lower. Riparian cover was absent or low at all stations. Group 1 and 2 stations had relatively little riffle habitat and pools were the predominant habitat, but riffle habitat was more available in Station 7.

Table 4 lists the numbers of each taxon collected at each sampling transect of each of the seven stations. Midge (Dipteran: Chironomidae) and blackfly (Diptera: Simuliidae) larvae, Baetid mayfly nymphs (Ephemeroptera: Baetidae) and aquatic annelid worms (Oligochaeta: Naididae) were abundant at all stations. Hydropsychid caddisfly larvae (Trichoptera: Hydropsychidae) were abundant except at Station 2, the station immediately below the Powerhouse tailrace. Stone fly nymphs (Plecoptera) were rare or absent except in Station 7, perhaps because summer water temperatures are too warm for stoneflies in the downstream stations. Water bears (Tardigrada) and zooplanktonic cladocerans and copepods were abundant at Stations 1 and 2. These taxa are generally found in quiet water habitats. Stations 1 and 2 are below the Powerhouse and receive water from Lake Isabella via the Borel Canal. Copepods were also abundant at Station 6, which is the first station downstream of the Main Dam. Planarid flatworms (Turbellaria: Planariidae) were common in all the stations downstream of the reservoir but were relatively rare in Station 7, while Corbicula clams (Bivalvia: Corbiculidae) were present at all the downstream stations but were entirely absent from Station 7.

The CSBA sampling methods are not designed to generate information on the quantity of food provided by a stream’s macroinvertebrates, but the information on taxa and on numbers of organisms collected provides some insight into feeding conditions for the fish. Fish species that inhabit the Kern River in and downstream of the Borel reach include Sacramento suckers, Sacramento pikeminnows, hardhead, smallmouth bass, rainbow trout, and riffle sculpin (see Fish Population Survey Technical Report). Benthic macroinvertebrates are major prey organisms for all these species.

Table 3

Table 4

Table 4 (cont.)

Table 4 (cont.)

The macroinvertebrate community of all the stations included many taxa that are important prey of the fish species in the river. These taxa include all the true flies (Diptera), mayflies (Ephemeroptera), and caddisflies (Trichoptera). The stone flies (Plecoptera), which were abundant only in Station 7, are important prey of trout.

The densities of macroinvertebrates at the stations in the diverted reach of the Kern River (Group 2 stations) were high relative to the densities at the Groups 1 and 3 stations and were also high relative to densities of macroinvertebrates reported for other studies in the Sierras. Densities of organisms (number per square meter) were estimated from the number of individuals in the samples and the surface area of riffle substrate sampled (six square feet per transect). Table 5 gives the densities estimates for macroinvertebrates at each transect in the seven stations of this study. The mean densities at the Group 2 stations were mostly at least twice those at the other three stations. If the zooplankton species, which have very little food value to the fish in the river, are excluded, the differences are even greater.

The macroinvertebrate densities in the diverted reach were compared with those from reaches of two other streams that are roughly similar in size and elevation: the Horseshoe Bend reach of the upper San Joaquin River (Fresno County) and the South Fork American River downstream of the El Dorado Diversion Dam (El Dorado County). Both of these stream reaches are diverted for hydropower projects (Big Creek No. 4 on the San Joaquin River and the El Dorado Project on the South Fork American River). Three riffles in the Horseshoe Bend reach of the San Joaquin River were sampled during November 1995 (SCE 1997) and one riffle in the diverted reach of the South Fork American River was sampled in October 1999 and again in August 2000 (Unger 2001). The mean density of benthic macroinvertebrates in the Borel Project diverted reach was 17,706 individuals per square meter as compared to a mean density for the San Joaquin River stations of 8,391individuals per square meter and mean densities for the South Fork American River station of 11,104 individuals per square meter in 1999 and 7,386 in 2000.

A couple of difficulties in using the density estimates should be noted. The San Joaquin River study did not follow the CSBA protocols and sampling methods for this study were different from those of the other two studies. However, these differences probably do not greatly affect the density estimates. It should also be noted that the density estimates are less than ideal indicators of food supply because they provide information for one type of habitat only (riffles) and because they provide no information about the biomass or productivity of prey species. Nevertheless, the density estimates suggest that the macroinvertebrate food supply for fish in the Borel Project diverted reach is fairly high relative to that in the other stream reaches.

Table 5 gives the CSBA metrics for the Kern River samples. Note that Table 2 provides descriptions of the metrics. Table 2 also shows how the metric values reflect the degree of disturbance to which the macroinvertebrate community is exposed. For instance, as the ecosystem becomes increasingly impaired, Taxa Richness generally decreases and mean Tolerance Value increases. All the metrics vary consistently in relationship to ecosystem disturbance except Percent Grazers and Percent Predators.

Table 5

As indicated earlier, eleven of the metrics that give the most meaningful results were analyzed statistically. The metrics not analyzed include those for individual taxa (Ephemeroptera, Plecoptera, Trichoptera, Hydropsychidae and Baetidae) and those for the functional groups, scrapers, predators and shredders. Shredders were absent in almost all samples, and, as noted above, predators and scrapers show no consistent relationship to ecosystem impairment (see Table 2).

Figures 2a through 2k plot the station means for the eleven selected metrics. To highlight the relationship between metric value and level of impairment, the means are ordered by magnitude, with the mean representing greatest impairment on the left side of the graph and those representing decreasing impairment ordered to the right. Most of the metrics indicate that the Group 3 station, Station 7, has the least impaired ecosystem and the Group 1 stations, Stations 1 and 2, are the most impaired. As noted earlier, Station 7 is upstream of Lake Isabella, but all other stations are below the reservoir. The mean for Station 7 shows the least impairment for all but two of the eleven selected metrics, whereas the mean for either Stations 1 or 2, which are below the Powerhouse, showed the most impairment for eight of the metrics. Station 6, which is the closest station in the diverted reach to the Main Dam, also was among the most impaired stations for most of the metrics.

The results of the multiple comparison statistical tests indicate that the Station 7 means are significantly different from most or all of the other means for most of the selected metrics. The multiple comparison tests group together means that are not significantly different, as indicated by the dashed lines in Figures 2a through 2k. Station 7 is in a group by itself for six of the metrics, including Taxa Richness, EPT Taxa and Index, Sensitive EPT Taxa, Tolerance Value, and Percent Intolerant Value (Figures 2a through 2d, 2f and 2g). These results provide evidence that, as expected, the biotic integrity of the ecosystem at Station 7 is less impaired than that at the other stations. However, it must be noted that because Station 7 is some distance from the other stations, natural environmental factors may contribute to the differences in the metrics.

The results of the multiple comparison tests for the other stations are less consistent than those for Station 7. Nonetheless, the tests for three of the metrics, EPT Index (Figure 2c), Percent Tolerant Taxa (Figure 2h), and Percent Filterers (Figure 2k) indicate that the means of the Group 1 stations, Stations 1 and 2, are significantly different than those of the other stations. These results, combined with the high "impairment scores" of these stations for most of the metrics, suggest that the river’s ecosystem at the Group 1 stations is more impaired than that at the other stations.

Differences in sampling conditions may have influenced the results from some stations, but these differences were unlikely to have caused the differences in the mean metrics for most of the stations. As noted in the methods sections, Stations 1, 2 and 6 did not have a sufficient number of riffles to allow random selection of a sampling riffle, as specified by the standard CSBA protocol. Also, these stations had shorter stream reaches than the other stations downstream of Lake Isabella (Table 1). Station 2, the station immediately below the Powerhouse, had only one small riffle and one of its sampling sites was in the tailrace, so sampling conditions may have had some effect on the results from this station. However, the Station 1 and 6 riffles were substantial and appeared generally similar to

Figures 2a – 2k

those in the other stations. Therefore, it seems unlikely that sampling conditions affected the results from these two stations. It should be noted that the high "impairment scores" of Station 6 for many of the metrics may be related to its proximity to the Main Dam. This station shares important macroinvertebrate community attributes with Stations 1 and 2, including high numbers of zooplanktonic organisms and annelid worms (Table 4) and a high Percent Tolerant Taxa (Figure 2h). However, in other respects, such as a high total number of organisms (Table 4) and high percentages of Hydropsychidae and Baetidae (Table 5), this station is more similar to the other stations in the diverted reach.

The benthic macroinvertebrate community in the Kern River shows clear differences among the three station Groups: stations in the diverted reach of the Borel Project, stations in the reach downstream of the Borel Powerhouse and the station in the reach upstream of Lake Isabella. The macroinvertebrate community of the station upstream of the reservoir, which has been relatively unaffected by human influences on flow, is the most characteristic of undisturbed sites. All of the stations downstream of the reservoir, including those in the diverted reach and those downstream of the Powerhouse that receive the diverted flows, show more impairment than the upstream station, which suggests that the reservoir has an overriding adverse effect on the biotic integrity of the river’s ecosystem. Of the stations downstream of the reservoir, those in the diverted reach show somewhat less ecosystem impairment and have a higher density of organisms than those downstream of the Powerhouse.

Copyright © 2002 by Southern California Edison Company, Rights Reserved