Tier 1 Benthos: Laboratory Procedures
1. Sub-sampling
Benthic macrofaunal samples should whenever possible be processed in their entirety to ensure the best data quality. However, if samples have large volumes (ie. >2 x 1L jars, marine samples may be split to a maximum of ¼ as per EEM Pulp & Paper guidelines (Environment Canada, 2010), with a target count of >300 organisms in all subsamples whenever possible. Various methods of laboratory sub-sampling are available (Klemm et al. (1990) as well as Canton (1991), Mason (1991b), Plafkin et al. (1989), Wrona et al. (1982), Marchant (1989) and Sebastien et al. (1988), however many years of experimentation with local samples has yielded an optimum method for marine subtidal soft substrate benthos, as follows (and see Table 1).
Larger megafauna are picked out in their entirety and processed. Remaining sample debris is then washed through 2 sequential screens, 1cm and 1.0mm. The >1.0 cm fraction is retained and sorted in its entirety. Thus, split samples are fractioned into a ‘whole’ fraction (large debris and organisms, sorted and identified as whole) and a fine fraction (<1cm). Generally, only the fine fraction is split. Splitting is done using one of two methods. Most importantly, any macro organisms (>1cm in all dimensions) should be picked and retained from the sample. This subsampling approach meets EEM (Environmental Effects Monitoring; Environment Canada 2002) guidelines, but in addition includes a macro fraction (>1 cm) to account for the distribution of any large organisms.
Method 1. Standard sediment splitter. The 1.0mm fraction is then split into quarters using a sediment splitter (Rickly Hydrological Company, model #US LSS-72A). In samples with multiple jars, each jar is washed through a 1cm screen. The macro fraction is retained, and the fine fraction is split. Samples may be split to a maximum of ¼.
Method 2. Caton Tray (Canton, 1991). The sample undergoes a primary sort in a large tray to ensure any large organisms (>1cm in all dimensions) are collected from the sample. At this time, any large debris is removed and checked over. The remainder of the sample is spread on a large caton tray and a random ¼ is chosen to sort. Ideally a ¼ split will contain ≥ 300 organisms. If the split does not meet this guideline, more of the sample should be sorted, as the budget allows.
Subsamples are assessed visually for evenness of splitting, and re-split if there are obvious discrepancies. Where debris volumes exceeded 1L (i.e., more than 1 jar per sample), each jar was split independently to ensure the sample was split evenly.
Ten percent (10%) of split samples are then sorted in their entirety (with each split processed separately) to ensure that sub-sampling error, as indicated by variability in total abundance of organisms, is lower than 20%.
Results of a subsampling experiment to test the efficacy of this method are reported in Burd (2014). Precision estimates (SE/mean) between splits, as well as differences in estimated totals from each split and the actual totals from the summed splits was consistently less than 15%, with % differences in total abundance estimates from splits versus total counts less than 20% (positive and negative). This is a reasonable result for these types of samples (Elliott 1977). The highest error in this experiment was for a split was due to the very high sediment volume and stickiness as well as the extreme abundance of the dominant Capitella capitata complex in a sample taken near an organic deposition source (outfall). This extreme condition is not expected to occur in background or ambient stations in the SSAMEx program. The method is described below.
To ensure adequate representation of all faunal groups in the split sample, a subsampling accuracy report is required. For 10% of split samples, all subsamples taken from the same sample are processed in their entirety and compared to the actual total counts to obtain estimates of subsampling accuracy (see below). Environment Canada EEM (2002) guidelines state that the mean error across 10% of the split samples should be <20%, where;
% Error in the estimate = [1- (extrapolated # in sample)] x 100
Actual # in sample
Table 1 Example of recommended reporting for subsampling error. Example is for a volume based method using the Imhoff cone (Wrona et al., 1982) where up to 10 subsamples are sorted and the remainder of the sample was sorted. Accuracy of each subsample, the minimum, maximum and mean subsampling accuracy as well as the range in precision between individual subsamples are all reported.
Subsample no Number Inverts Predicted no. Pred – Expected % Diff from
| 1 | 218 | 3815 | 101 | 2.7 |
| 2 | 220 | 3850 | 136 | 3.7 |
| 3 | 230 | 4025 | 311 | 8.4 |
| 4 | 221 | 3868 | 154 | 4.1 |
| 5 | 221 | 3868 | 154 | 4.1 |
| 6 | 201 | 3518 | -197 | -5.3 |
| 7 | 219 | 3833 | 119 | 3.2 |
| 8 | 205 | 3588 | -127 | -3.4 |
| 9 | 221 | 3868 | 154 | 4.1 |
| 10 | 210 | 3675 | -39 | -1.1 |
| Total in remaining | 1548 | |||
| Total in sample | 3714 | Mean Absolute subsampling error (%) 4.0 | ||
| Range in Precision 0.5 – 9% | Min % error 1.1 | Max % error 8.4 |
2. Sorting
2.1. What to keep and what to discard
The “Memo” category in taxonomic processing of benthos samples typically includes auxilliary material collected in a given sample which is not normally part of the quantitative assessments of benthos. Using the basic principal that no data should be discarded, SSAMEx recommends that this information be recorded without undue processing effort in a separate category of data called “Memo” or “Comments”. Overall comments should be included from the taxonomist about unusual concentrations of any of these “Memo” categories in given samples. This information can be useful later to ecologists examining and explaining unusual features in samples. The following types of organisms are likely within this category and should be described at the time samples are picked, but specimens can be left in debris;
1. meiofauna – organisms that are considered meiofaunal and would usually pass through a 1 mm mesh screen (harpacticoid copepods, nematodes, foraminifera, etc.) can sometimes be captured on the 1 mm screen. These should be left unidentified past phylum or class, with general comments about abundance. They should not be included in macrofaunal enumerations for Tiers 1 or 2. If a Tier 3 survey analysis is required which includes meiofauna then separate, unscreened samples should be preserved and processed separately from macrofauna as described in the Tier 3 hyperlinked document.
2. Epifauna – In courser sediment samples, or those with considerable gravel, cobble or shell debris, attached epifaunal organisms may be present. In shallow samples in particular, it may not be cost-effective to fully enumerate and weigh these (particularly small barnacles and other colonial, encrusting organisms). These should be left unidentified past phylum or class, with general comments and subjective coding about abundance provided (categories low, medium, high). However, epifaunal forms which are typically found on sediments (anthozoa, some hydrozoa, mobile predators, etc.) should be identified and enumerated in the same manner as all other macrofauna (Tier 2) and not in the “Memo” category.
3. Organisms transported from elsewhere – (slumped or transported freshwater, intertidal or estuarine macroalgal mats and/or associated fauna). These should not be enumerated or extracted from sample debris. This judgement requires sample processing by experienced marine technicians.
4. Incidentals (egg cases, strings, off-casts of shells or tubes, zooplanktonic forms captured during sampler descent, fragments of organisms without heads). These should not be enumerated or extracted from sample debris.
5. Fragments (back ends of organisms and non-head pieces) should all be kept with their respective taxonomic groups for bulk weighing IF Tier 1 is the final stage of processing. If Tier 2 is to be done, fragments can be discarded.
2.2 Picking and sorting
Small fractions of a sample are placed in a gridded dish. Sorting is done with a dissecting scope at 10-40x magnification. Small amounts of debris (enough to cover a square petri dish in a single layer; <5mL) are sorted at one time. The dish must be scanned systematically and all animals and fragments removed using forceps. Each dish was sorted until no further organisms or fragments of organisms were recovered (each dish was sorted at least twice, regardless of the number of organisms recovered). The sample is washed gently and kept covered and wet throughout the sorting of the sample.
All vials of picked and sorted fauna are stored in 70% ethanol, with fragments placed in separate labelled jars for each group. All vials must have internal and external sample labels, which are double-checked when the preliminary numbers are entered into a database.
For samples which have large pieces of organic matter, the samples can be divided in the laboratory into appropriate size fractions to expedite the sorting process. The most commonly used fractions are coarse (> 1.00 mm) and fine (500 µm – 1.00 mm), which correspond to the divisions used to define coarse and fine particulate organic matter (CPOM and FPOM, respectively). If there are very large pieces it is sometimes beneficial to separate these from the rest of the sample with a 4.00 mm sieve. All fractions should then be sorted accordingly and the large numbers of organisms warrants it, the size fractions can be sub-sampled independently. Careful note taking is required for these more complex sorting procedures so that densities are calculated accurately. After the initial washing and fractionation of samples, the invertebrates should be sorted by trained technicians from the debris on a gridded tray or petri dish under a dissecting microscope at 10X to 20X magnification.
To minimize sorter bias, samples are distributed among trained personnel such that no person sorts all the replicates of a given sample, and/or no one person sorts >25% of a particular project. Fifty percent (50%) of all samples of a given sorter are spot-checked to ensure sorting efficiency of each particular sample is >90-95%. Spot-checks are generally performed on 25% of the sample, and the number of organisms recovered is scaled to the original sample volume (i.e., # found x 4 for a 25% resort). Samples with high volumes of wood debris (i.e., high sorting times and low abundances are accepted at 90% recovery; all other samples are accepted at 95%). Overall, the average estimated sorting efficiency must be >95%. Samples that do not pass the spot-check procedure are resorted in their entirety. If a particular sorter has a failed sample, the remaining samples completed by that sorter are spot-checked and resorts are done as necessary. If a sorter misses a particular type or group of organisms, all of their samples are resorted. Organisms recovered in spot checks are included in the data set.
It is recommended that at least 10 % of all samples be sorted in their entirety, and that the criteria for an acceptable sort be that < 10% of the total number of organisms were missed. If > 10% of the total number is found during the resort, then all the samples within that group of samples requires resorting. A further criterion which would require a resort is if an entire group of benthic invertebrates was missed by the sorter, (i.e. it wasn’t recognized as an organism), even if the missed organism constituted < 10 % of the total. The factors which should be considered when determining similar groups of samples include: 1) sampling area, 2) habitat class, and 3) individual sorters. The QA/QC guidelines apply independently to each group of samples sorted. Unsorted and sorted fractions are to be retained until taxonomy and sorting efficiency are confirmed and the data are reviewed.
2.3 Training and Quality Control
Many years of experience dealing with taxonomic analyses on the west coast have shown that the quality of initial picking and sorting of samples is the most critical (and problematic) stage of analysis for benthos samples. For this reason, it is recommended that benthos sample processing budgets include provision for external, arms length re-counts of the picked debris for up to 10% of samples by a recognized (with contact information) external laboratory unconnected to the primary taxonomic laboratory. In addition, it is critical that sample debris be kept for a period of at least 2 years following completion of the taxonomic work. This allows correction or verification of unexpected or extreme results.
This external verification should be IN ADDITION TO the following quality control procedures within the primary taxonomic laboratory. All laboratory personnel must have basic instruction and evaluation in the sample processing procedure by experienced laboratory staff. A Quality Control (QC) Officer in the taxonomic laboratory oversees the activities of inexperienced technicians. While 50% of all experienced technician’s samples are spot-checked by the QC officer, 100% of inexperienced technicians’ samples will be resorted, until a standard of 95% recovery is obtained.
The qualifications of a QC officer include consistent achievement >95% sorting efficiency and taxonomic knowledge of marine benthic macro-invertebrates. Verification of sorting by the QC officer should ensure that (1) >90-95% of all organisms are removed from marine samples (depending on type and protocol) and (2) no systemic bias is introduced. Organisms recovered in QA resorts are not included in the data set.
3. Data Endpoints
A preliminary counts of the major taxonomic groups from the sorted samples is the only requirement for identification of organisms in the Tier 1 sediment program. During sorting, organisms are separated into major taxonomic groups. If the samples are being further processed for Tier 2, this sorting should be done at a minimum to Arthropoda, Annelida, Mollusca, Echinodermata, and “other”. If the samples are only being processed at a Tier 1 level, it is imperative that the sorting be as detailed as possible and to family if possible. At a minimum the sample should be sorted to amphipods, ostracods, decapods, isopods, tanaids, other crustaceans; separate errantiate and sedentariate polychaetes, different types of echinoderms (ophiuroids, holothuroids, echinoids, etc.) and “other” to general phyla at least. This allows reasonable conversions to organic biomass and thus production estimates.
This sorting process is important for two purposes;
- to flag potential errors in the comparison of replicate samples returned by different sorters, and
- the comparison of preliminary counts and final numbers ensures all specimens in the sample are accounted for and the data recorded properly
4. Biomass estimates of major taxonomic groups
Large, megafaunal orgamisms (approximately 2 g wet weight or larger) are separated from the remainder of the sorted organisms. These will be weighed separately (see below). The remaining specimens sorted to major taxonomic groups for each sample (see “sorting” above) are then weighed as a bulk group for each sample. Wet weights (WW) are performed with a balance that is accurate to 0.1-0.01mg. The accuracy of 5 decimal places (0.01 mg) is optimal for low biomass groups. If the wet weight of rare or low biomass groups are beyond detection limits, these are reported as 0.00005 mg. Prior to weighing, organisms are removed from ethanol, blotted on a Kimwipe or filter paper, and the first weight upon placement in the balance recorded. One cannot wait for stabilization of the balance reading as ethanol evaporates quickly, and small organisms will dry out.
Larger megafaunal organisms (see above) are blotted for 30 seconds, placed on the balance pan, and then air-dried for 1.5 minutes (to allow some of the alcohol to evaporate) before the weight is recorded. The 1.5-minute waiting period was selected in 2001 after comparing the results of repeatedly weighing organisms using various methods. If the large organisms are being returned alive, this process should be a priority at the time of sample retrieval and washing on board. It is recommended that length and width measurements are taken as well for each specimen, to facilitate statistical weight/size calibrations for future streamlining of this process.
Summary of QA/QC steps (See appropriate sections and/or QA/QC benthos document for more detail)
1) A sample inventory is constructed, and the destination of these samples is confirmed upon survey completion, and appropriate paperwork drawn up for storage or return.
2) Careful and thorough laboratory washing, picking and sorting of samples. This factor has been recently proven in comparative studies by Metro Vancouver and Victoria CRD to be critical, and can result in order of magnitude errors in benthos enumerations. Re-examination of a percentage of picked samples is critical as described in the hyperlinked document, and should be done by experienced personnel other than the original sorter.
3) Ten percent (10%) of split samples are processed in their entirety (with each split processed separately) to ensure that sub-sampling error, as indicated by variability in total abundance of organisms, is lower than 20%.
4) Fifty percent (50%) of all samples of a given sorter are spot-checked to ensure sorting efficiency of each particular sample is >90-95%. Provision should be made for external, arms length re-counts of the picked debris for up to 10% of samples by a recognized (with contact information) external laboratory unconnected to the primary taxonomic laboratory.
5) Picked sample debris should be kept for at least 2 years in order to allow external re-examination of debris in the event of unusual results. If this is not feasible, debris should be disposed of following satisfactory EXTERNAL QA/QC (see QA/QC benthos document).
References
Burd, B.J. 2014. Re-analysis of CRD historical benthos data, and recommended Quality Control revisions for future monitoring surveys. Final report to Victoria Capital Regional District, environmental monitoring division, November, 2014.
Burd, B.J., Macdonald, T.A. and van Roodselaar, A., 2012a. Towards predicting basin-wide invertebrate organic biomass and production in marine sediments from a coastal sea. Plos One, 7(7): e40295. : Supplemental Material: http://www.plosone.org/search/simple;jsessionid=84E93FDB92ED534A87B60929F34392F6?from=globalSimpleSearch&filterJournals=PLoSONE&query=Burd&x=0&y=0
Canton, L.W. 1991. Improved subsampling methods for the EPA “Rapid Bioassessment” benthic protocols. Bull. N. Am. Ben. Soc. 8:317-319.
Elliott, J.M. 1977. Some methods for the statistical analysis of samples of benthic invertebrates. Freshwater Biological Assoc. Sci Publ. No. 25 160 pp.
Environment Canada. 2002. Revised Guidance for Sample Sorting and Subsampling
Protocols for EEM Benthic Invertebrate Community Surveys. https://www.ec.gc.ca/esee-eem/default.asp?lang=En&n=F919D331-1 accessed December 2012.
Environment Canada. 2010. Pulp and Paper Environmental Effects Monitoring (EEM)
Technical Guidance Document.
Klemm, D.J., P.A. Lewis, F. Fulk, and J.M. Lazorchak. 1990. Macroinvertebrate field and laboratory methods for evaluating the biological integrity of surface waters. EPA 600/4-90/030. U.S. Environmental Protection Agency, Environmental Monitoring Laboratory, Cincinatti, Ohio.
Marchant, R.A. 1989. A subsampler for samples of benthic invertebrates. Bull. Aust. Soc. Limnol. 12:49-52.
Mason, W.T. Jr. 1991b. Sieve sample splitter for benthic invertebrates. J. Freshwat. Ecol. 6:445-449.
Plafkin, J.L., M.T. Barbour, M.T. and K.D. Porter. 1989. Rapid bioassessment protocols for use in streams and rivers: benthic macroinvertebrates and fish. EPA/444/4-89-001, 162 pp.
Sebastien, R.J., Rosenberg, D.M. and A.P. Wiens. 1988. A method for subsampling unsorted benthic macroinvertebrates by weight. Hydrobiologia 157:69-75.
Wrona, F.J., Culp, J.M., and R.W. Davies. 1982. Macroinvertebrate subsampling: a simplified apparatus and approach. Can. J. Fish. Aquat. Sci., 39: 1051-1054.