Agriculture Canada Logo

1989 - 1994

A Protocol for Monitoring and Assessment of Water Quality In Agricultural Streams Using Benthic Invertebrates

D.R. Barton and M.E. Farmer,
Department of Biology, University of Waterloo,
Waterloo, Ontario, Canada, N2L 3G1.
M.E. Dillon and D.R. Oliver,
Centre for Land and Biological Research Resources, Agriculture Canada,
Ottawa, Ontario, Canada K1A 0C6.


Download report  (32 KB pdf)



The primary objective was to develop a set of techniques, a protocol, for monitoring and assessment of surface water quality in agricultural regions of southern Ontario based on benthic invertebrates.

This involved determining the best way to collect appropriate samples, how closely the animals should be identified, and how the results can be interpreted. The protocol will be used to assess the status of small streams relative to other streams draining land under the same crop, or to monitor changes in water quality due to changes in farming practices. The assessment function is intended to identify sites which are better or worse than the average for southern Ontario; monitoring will be valuable in evaluating the results of mitigative measures such as leaving riparian buffer strips or adopting conservation tillage. It was tested in a study of four pairs of streams. One of each pair drained land primarily under conservation tillage and the other under conventional tillage.


Agriculture has profound local and regional effects on both the quality and quantity of surface waters. Removal of the forest which originally covered most of southern Ontario led to higher summer and lower winter water temperatures, as well as destabilization of hydrologic regimes, increased erosion and siltation. Subsequent use of this cleared land for agriculture has additional effects on the quality of surface waters, with the nature and magnitude of the effects reflecting different farming practices.

In addition to each of the basic suites of stresses which are characteristic of each major kind of crop, the relative impact of fields under the same crop may differ substantially because of individual farming practices, variations in soil conditions or chance events such as heavy rain just after application of fertilizers. Such variation can make it very difficult, and expensive, to monitor the effects of agricultural practices on surface water quality through chemical assays. Pulsed inputs of chemicals resulting from spills or heavy precipitation may easily be missed in routine surveys, but may have severe biological impacts.

The problem of rapid fluctuations in the chemistry of running waters has long been recognized with respect to discharges of wastes from industrial and municipal sources, so the alternative approach of measuring biological effects directly has been widely accepted (Rosenberg and Resh 1992). Benthic invertebrate communities are nearly ideal for this purpose because they are easily sampled and include large numbers of individuals belonging to many species, each with its own response to any particular type of stress. These animals live on the bottom of the stream or lake and have life-cycles lasting months to years, so are exposed to the conditions in the stream over fairly long periods of time. Schindler (1987) and Gray (1989) have suggested that biological surveillance of communities, characterizing taxonomic richness and composition, is a very sensitive tool for the detection of alterations in aquatic ecosystems.

For point sources, the usual approach is to sample upstream and downstream of the point of discharge and to express the differences in terms of changes in total abundance, numbers of species, some expression of "diversity", or a biotic index. In such situations, the benthic fauna reflects average conditions as well as intermittent discharges, and biological monitoring tends to be much cheaper and more sensitive than routine chemical analyses.

The potential utility of benthic invertebrates for monitoring and assessment of the effects of agriculture on water quality should be equally great, but the situation differs in one very fundamental way. With a few exceptions such as runoff from feed lots or manure storage, agricultural effects do not originate as point sources, thus upstream-downstream comparisons are rarely practical. Furthermore, agricultural activities tend to inflict a number of stresses simultaneously so it is not easy to ascribe the composition of the fauna at any particular point to any single aspect of farming practice.


This study has demonstrated that the biodiversity of benthic invertebrates in small streams in agricultural southern Ontario is very high. These diverse communities respond in distinctive ways to inputs of eroded soil, pesticides, fertilizers and other organic matter. There is a strong seasonality in the composition of the benthic fauna, and those species which grow more actively during the warmer months appear to be more sensitive to agricultural activities than those which are most active in winter.

Percent Model Affinity proved to be the most versatile and sensitive way to summarize invertebrate data for the purposes of water quality monitoring and assessment. Samples from additional streams can be compared to the expected community from forested reference streams to identify any which are significantly impacted. Inspection of the detailed composition of the fauna usually gives an indication of the nature of the problem and the kind of subsequent detailed monitoring or analysis which might be warranted. Where remedial actions or changes in farming practices have occurred, monitoring of the biological effects would involve comparing benthic samples with the average communities from both forested streams and streams draining the same crops. Changes would be seen as decreasing PMA over time relative to the same crop; increasing PMA relative to forested streams would indicate improving water quality.


The most important result of this work is the Monitoring Protocol which has been developed for invertebrate-based monitoring and assessment of surface water quality in southern Ontario. The procedure for this protocol is as follows:

1. Field Collecting.

  1. Record location of stream (e.g. UTM), adjacent landuse and general land use in area.

  2. Take sample by disturbing portions of all microhabitats (mid-channel riffles and pools, marginal vegetation, undercut banks, etc.) in the stream with a booted foot.

  3. Capture the dislodged substrate in a dip net with 300 m mesh bag.

  4. Rinse the contents of the net to remove fine sediments, discarding large pieces of vegetation, wood and stones.

  5. Save 400 ml of contents in net in a jar and preserve with 10% formalin.

2. Laboratory analysis

  1. Empty sample into a fine-meshed (100 m aperture) net; rinse to remove the preservative and fine sediments not washed through in the field.

  2. Examine sample under a dissecting microscope and remove, at least, the first 200 animals encountered.

  3. Briefly examine unsorted sample for large rare animals which may not have been included in the first 200 individuals.

  4. Preserve animals in 70% ethanol.

  5. Identify and enumerate the animals to the lowest practical taxonomic level.

  6. Convert the counts to percentages of the total number of animals sorted from the sample.

  7. Calculate the Percent Model Affinity (PMA) between the sample and the appropriate expected reference community.

  8. Values of PMA which are more than 2 standard deviations away from the mean for the reference community are significantly different.

The cost of invertebrate biomonitoring can be estimated from the time and skill of the personnel needed for each step. Collection of the sample requires about 10-15 minutes at the site. Sorting the sample in the laboratory takes 2-3 hours. Personnel performing each of these steps require, at most, a few hours of training. The time and expertise required for the identification stage vary with the taxonomic level. An experienced benthic taxonomist should be able to identify the specimens to the family level in about half an hour, to the lowest practical level in about 1-3 hours, depending on the number of taxa present. Less experienced personnel will need more time and should have their identifications confirmed by experts in the field.


Agriculture is the predominant landuse in the lower Great Lakes basin, so any inputs to surface waters enter the Lakes themselves, either directly or indirectly. Our results suggest that about 80% of the small streams in southern Ontario are negatively impacted by agriculture.

Substantial investments over the past 25 years have reduced the inputs of nutrients and organic wastes from municipal sources, and these efforts appear to be resulting in significant improvements in water quality. Some progress has also been made with respect to industrial effluents. Such point sources are easily identified so remedial efforts can be concentrated.

The Monitoring Protocol presented here can be used to identify streams subjected to significant non-point source pollution, and to monitor improvements which result from changes in landuse practices. Agricultural streams receive inputs of eroded sediment, inorganic and organic nutrients, and a variety of agro-chemicals. Broad chemically based surveys to locate problem sites are prohibitively expensive; the invertebrate fauna can provide much more information, at a much lower cost.


The Monitoring Protocol is useful to any organization interested in the effects of agriculture on the water quality of streams. User groups include conservation authorities, provincial Ministries (Environment, Natural Resources, Agriculture), federal Departments (Agriculture, Environment, Fisheries), as well as private organizations (e.g. Ducks Unlimited, OFAH), and private consultants.

The taxonomic descriptions and keys will increase the capability of environmental impact workers to distinguish and name Chironomidae living in small streams.


In order to increase the sensitivity of PMA for monitoring of agricultural impacts, expected communities could be refined to reflect narrower spatial and temporal scales. This can be accomplished through extension of the existing database by sampling at different times at some of the sites already surveyed, as well as new ones. Additional data from forested reference sites are especially needed. Similarly, sensitivity will be enhanced with improvements in our ability to identify invertebrates to the species level. This will require continuing taxonomic studies.

The database should also be expanded to include larger streams. Concentration on headwater streams during this phase, in order to isolate the effects of individual crops, has left several questions to be answered:

  1. Does the fauna of larger rivers integrate all landuses upstream?

  2. Are certain agricultural inputs processed (i.e. removed from circulation) within the stream?

Answers to these questions would allow more meaningful interpretations of the effects of agriculture on the Great Lakes themselves, and could increase the efficiency of biomonitoring programs by reducing the number of sites needed for initial surveys.




Table of Contents | GLWQ Home

Thursday, May 05, 2011 04:02:41 PM