B-IBI Calculation Procedures Used by Streamkeepers of Clallam County—July 2007

Streamkeepers currently uses the 10-metric genus-level Benthic Index of Biological Integrity for the Puget Sound Lowlands to analyze its benthic macroinvertebrate samples.  This index is based on three replicate samples taken at the same time at the same site, and metrics are based on total and average counts for those replicates (see http://www.clallam.net/streamkeepers/html/biological_monitoring.html).  These metrics are based on a target sample count of 500 ±10%, and 250 specimens has been set as the minimum count for adequate statistical analysis (see http://www.clallam.net/streamkeepers/html/Leska_Fore_s_Level_III_analysis_study.htm).  However, the B-IBI sample-collection protocol is governed by spatial and temporal criteria rather than specimen counts, and therefore replicates can undershoot or overshoot the target of 250-550 specimens.  Here are the procedures Streamkeepers follows in those nonstandard cases.

Case I.  One or more replicates in a sample exceeded the maximum target value of 550 specimens.  In this case, the taxonomist subsamples those replicates using a widely-accepted procedure:

a.       The taxonomist lays the replicate onto a grid of squares and randomly counts squares one at a time until the total count at the end of a given square reaches or exceeds 500; specimens identified in the subsample are tallied as normal.

b.      With the uncounted grid squares, the taxonomist performs a so-called “Large and Rare” search (Streamkeepers prefers to call it an “Additional Taxa Search”) in which the taxonomist looks for any specimens of taxa not already included in the count for the subsample.  For any such specimens, a tally of “1” is made in a separate column.  (The taxonomist does not count beyond 1 for these additional taxa.)  The reasoning behind this additional count is that the replicate was taken from a defined area within a defined time, and that unique specimens collected under those limitations should be included within the replicate counts (otherwise the replicate gets “penalized” for having too many organisms in it).

c.       For calculation of metrics, the two columns are added together first.  The totals don’t reflect the numbers of individuals of the taxa found in the “Large and Rare” search, and thus the metrics will be skewed somewhat from what they would have been had the replicate been counted completely, but two mitigating factors tend to minimize the extent of this skewing:

                                                               i.      Seven of the metrics are based on numbers of taxa rather than numbers of individuals; for these metrics, specimen counts are immaterial.

                                                             ii.      The remaining three metrics which are based on numbers of individuals have to do with “percent dominance” of tolerant individuals, predators, or three-most-populous taxa. These are likely to reflect the sample population fairly accurately from within the subsample, because the taxonomist is expected to have homogenized the sample well before laying it out on the grid, and because dominant taxa are likely to have already been well-represented in the subsample and therefore not subject to inclusion in the “Large and Rare” count.

Case II.            One or more replicates have a total count of fewer than the target minimum of 250.  In this case, Streamkeepers combines and/or discards the original replicates to create “revised replicates” according to the following protocol:

a.       If just one replicate is <250, either:

                                                               i.      Combine it with the smaller of the other replicates if the other replicate has <450 specimens and the combined total would be ≤550; or

                                                             ii.      If the above does not apply, discard the low sample and run the B-IBI with 2 replicates (see below).

b.      If 2 replicates are <250:

                                                               i.      If those 2 replicates total to ≥250 specimens, combine them into a single replicate and calculate the B-IBI with the two “revised replicates,” as described below.

                                                             ii.      If not, and the largest replicate is <450, try to add one or both of the small replicates (first the larger, then the smaller) to the largest replicate (to try to bring the largest replicate into the target range).  Stop once the total ≥500, and do not exceed 550.  Then calculate the B-IBI with that single “revised replicate,” as described below.

                                                            iii.      Otherwise, discard the two low replicates and calculate the B-IBI with the single remaining replicate, as described below.

c.       If all 3 replicates are <250:

                                                               i.      Program managers need to decide whether the sample is valid and reflects a sampling site with low biological activity, and/or whether to label the sample “questionable” or “unacceptable”—in the latter case, explanatory comments should be included with the sample in the database.

                                                             ii.      If managers determine the sample is “acceptable” or “questionable,” calculate the B-IBI metrics using all 3 replicates as is.  If “questionable,” include the qualifier in any database report.

Procedure for calculating B-IBI metrics with <3 “revised replicates”:   METRIC CALCULATION METHOD Total # of taxa, total # of mayfly/stonefly/caddisfly/clinger taxa & % predatory, % tolerant, % dominance Calculate for each revised replicate* and then take the average; use B-IBI metrics as usual.   *”Revised Replicate” = Replicates as combined or discarded according to Steps 1 & 2 above. Total # of long-lived (NOTE:  for a standard 3-replicate sample, a B-IBI score of “3” is assigned when the total # of unique long-lived taxa across all 3 replicates is between 4 and 8, inclusive.) Average the # of taxa across all revised replicates*, then analyze as follows:   ·         Assign a B-IBI score of 3 if 2.5 ≤ average ≤ 5.5 ·         Assign a 1 if average < 2.5 ·         Assign a 5 if average > 5.5 Total # of intolerant (NOTE:  for a standard 3-replicate sample, a B-IBI score of “3” is assigned when the total # of unique intolerant taxa across all 3 replicates is between 2 and 4, inclusive.) Average the # of taxa across all revised replicates*, then analyze as follows:   ·         Assign a B-IBI score of 3 if 1 ≤ average ≤ 3 ·         Assign a 1 if average < 1 ·         Assign a 5 if average > 3