The Louisiana Department of Wildlife and Fisheries, Inland Fisheries Division supports and endorses the use of the following Index of Biotic Integrity system developed by the Aquatic Resource Workgroup for the Sabine River Authority Relicensing program as per the following:

Development of an Index of Biotic Integrity (IBI) for the Sabine River Downstream of the Toledo Bend Hydropower Project

An aquatic resource subgroup (Roy Kleinsasser, Texas Parks and Wildlife Department; Steven Arnold, HDR; Bobby Reed, Louisiana Department of Wildlife and Fisheries; and Robert Hansen and Mark Fisher, Texas Commission on Environmental Quality) was formed to develop a tool to assess biotic integrity and biodiversity relative to the Toledo Bend project.  The work was also done with the intent of developing a long-term monitoring tool that could subsequently be used to evaluate success of project operational alternatives and other basin efforts.  The tool, however, is not designed to replace the existing regional IBI relative to TCEQ’s overall water quality assessment program.

Given the time constraints, it was decided to use the regionalized IBI (ecoregions 33/35 aggregate) developed by Linam et al. (2001) as a foundation, re-scaling the existing metric scoring criteria to the larger Sabine system and determining if certain existing metrics were appropriate for evaluating a hydropower project. Scaling the metrics to the system was important given that the largest drainage area for any site used in developing the regional IBI was 865 km2 compared to the lower Sabine which has a drainage area approaching 20,000 km2.  Species richness metrics that were thought to be sensitive to drainage area included total number of species, number of cyprinid species, number of benthic invertivore species, and number of intolerant species.  The group also wanted to consider the appropriateness of existing metrics for evaluating project operations and whether alternative metrics would aid in characterizing the lower Sabine fish assemblages relative to Toledo Bend operations.

A database of 94 stream sites was assembled using fish assemblage collections from ecoregions 33 and 35, including the original streams used by Linam et al. (2002) to develop the regional IBI.  Other sources included Kleinsasser et al. (2004), which represents data from streams throughout ecoregions 33 and 35 and covers a range of stream sizes; Carver (1966), pre-impoundment data from Sabine River; Seidensticker and Parks (1994), data from upstream and downstream of Toledo Bend; Moring (2003) from the Neches River; and LDWF data from the Calcasieu River.  In general, sites were excluded if seining was not done in conjunction with another collection method (e.g., seining plus netting or seining plus electrofishing).

An important point is that this IBI is applied to individual collections.  For the basis of this project, that means a single electrofishing pass and 10 effective seine hauls.

Candidate metrics are listed below.  The first 12 were those employed by Linam et al. (2002), whereas the 13th was added similar to Bergstedt et al. (2004).  The large river faunal group was developed by Bobby Reed with input from the subgroup.  The application of trophic and tolerance metrics follows Linam et al. 2002 and Linam and Kleinsasser 2002.
1.	Total Number of Fish Species
2.	Number of Native Cyprinid Species
3.	Number of Benthic Invertivore Species (see Linam et al 2002)
4.	Number of Sunfish Species
5.	Number of Intolerant Species
6.	Percent of Individuals as Tolerant 
7.	Percent of Individuals as Omnivores 
8.	Percent of Individuals as Invertivores
9.	Percent of Individuals as Piscivores
10.	Catch (averaging CPUE from seining and electrofishing)
11.	Percent of individuals as state non-natives
12.	Percent of individuals with disease or other anomaly
13.	Number of species in Large River Faunal Group

Raw metric data was plotted against a log of drainage area and examined for trends.  Metrics were scored 5, 3, and 1, similar to Linam et al. (2002).  Metrics for which a relationship to drainage area were observed and consequently rescaled for the Sabine, include Total Number of Fish Species, Number of Native Cyprinid Species, Number of Sunfish Species, Percent of Individuals as Tolerant, and Catch, whereas a drainage area relationship was evident for the large river faunal group. The Number of Benthic Invertivore Species and Number of Intolerant Species had the same scoring criteria as the regional IBI.  No effort was made to adjust the the trophic scoring metrics since those had been set on a statewide basis.   Ultimately, the non-native species and disease metrics were discarded.  Few exotic species are found in the basin and the majority of raw values were zero.  The disease metric appears more appropriate for investigating sublethal pollutant inputs.  Generally, little data was available for assessing this metric.

To evaluate the ability of individual metrics to distinguish between higher and lower quality biotic communities in larger systems, we then filtered the data to include only sites greater than 1,000 km2.  We then subdivided the data into groups representing “better” quality and “lesser” quality, based on assemblage structure and knowledge of the sites.  The A and B data groups were then plotted as both raw metric values and metric scores to evaluate the metric’s utility in properly distinguishing between the higher and lower quality sites.  The metrics generally performed as expected, though it was noted that the omnivore metric was higher at “better” sites and appeared redundant with the invertivore metric.  We ultimately eliminated the omnivore metric. 

A proposed 10-metric index was then applied to both better and lesser sites, which were compared using a t-test.  Scores at the lesser sites were significantly lower (p<0.05) than the better sites, indicating that the index could distinguish between the two groups .

Two approaches were discussed to using the index for evaluating the project.  One was to consider total IBI score and component metric gradients downstream of the project, evaluating whether a dose-response signature was apparent.  Alternately, the applicant expressed a desire to establish a threshold (e.g., minimum score necessary for qualification as a high quality aquatic life use subcategory as required by the Texas wadeable stream procedures), but also evaluate gradients. 

Linam et al. (2002) used 50th percentile for high use when establishing a “high” aquatic life use from a database of least-disturbed reference streams.  Those sites represented the best conditions given the screening critieria used to select them.  Barbour et al. (1995) and Hughes (1995) suggested a higher percentile (e.g. 75th percentile) if using a mixed dataset such as the 94 sites employed in this effort.  After some discussion, the subgroup agreed to establish a threshold at a percentile of 62.5, resulting in an IBI score of 35.  This approach does not exclude using a gradient and examining all of the component metrics within the 10-metric IBI.