No single descriptor of aquatic macroinvertebrate assemblages is generally accepted as better than all others (i.e., most accurate and sensitive) (Resh and Jackson 1993, Barbour et al. 1999). Thus, the macroinvertebrate data were summarized as estimates of density for individual taxa or groups of taxa, and with community structure metrics that are commonly used in water quality monitoring programs. Each of the variables described below is calculated with information from the same data set, which results in a certain degree of redundancy among the descriptors. Thus, when meaningful changes in aquatic macroinvertebrate assemblages occur, we would expect that those changes would be apparent in changes for more than one descriptor. Not all macroinvertebrates were identified to the family level because of specimen size, damage, or taxonomic limitations. Thus, our estimates of richness may slightly underestimate actual richness.
The macroinvertebrate samples are quantitative because the animals are collected from a known area of stream bottom enclosed by the modified Hess sampler. This allows the macroinvertebrate counts to be expressed as a density (e.g., number per unit area or individuals per square meter), which can be then used to compare across sites and years. A large range of macroinvertebrate density was observed in this study, from < 1000/m2 to >100,000 per m2. We used density estimates to identify the 10 most common macroinvertebrates at each site, and to note when abundance seemed unusually low or high. Density is often extremely low at sites receiving significant Acid Mine Drainage. Density can be very high when selected pollution-tolerant species increase in abundance in response to pollution with significant organic and nutrient enrichment.
Rather than relying only on one or a few individual metrics, multimetric indices have been developed that integrate various types of information into a single number that can be used to compare streams. The Macroinvertebrate Aggregated Index for Streams (MAIS) was developed by Smith and Voshell (1997) based on benthic macroinvertebrate data from streams in the Mid-Atlantic Highlands of Maryland (51 sites), Pennsylvania (53 sites), Virginia (126 sites) and West Virginia (200 sites). The MAIS summarizes the values of 10 metrics:
- Ephemeroptera Richness,
- EPT Richness
- Intolerant Taxa Richness
- % Ephemeroptera
- % EPT
- % 5 Dominant Taxa
- Simpson Diversity
- HBI (Hilsenhof Biotic Index)
- % Scrapers
- % Haptobenthos
While the MAIS is relatively new, 9 of the 10 individual metrics (the exception is % Haptobenthos) used to calculate a MAIS Score have a long history in aquatic insect ecology and water quality monitoring programs. Values for the individual metrics are transformed into a score of 0, 1 and 2, and then combined into a MAIS Score. MAIS Scores are predicted to decrease in response to a decrease in water/habitat quality. Streams are classified based on MAIS Scores as follows:
- 13.1-20 classify a site as "Good"
- 6.1-13 classify a site as "Fair"
- 0-6 classify a site as "Poor"
The difference between Good and Poor sites is dramatic. For example, EPT Richness (the number of mayfly, stonefly, and caddisfly families) might be 11-12 at the highest scoring Good sites, but only 1-3 at the Poor sites.
The abundance of macroinvertebrate species within a community often differs between years. This difference is referred to as annual variation, and the cumulative affect can be evident in the MAIS Scores. Some of this variation may be related to natural phenomena such as droughts or floods; some of it may be related to human activities such as an unexpected pollution discharge; most of it is often unexplained. In the case of water quality monitoring, it can have a significant effect on stream classifications. For example, the MAIS Scores and site classifications presented for Sites 1-19 are the average across eleven years. However, the figure below illustrates how MAIS Scores for each site varied among years, and this annual variation often resulted in different stream classifications (e.g., Good versus Fair or Fair versus Poor). One site has been classified as Good, Fair or Poor, depending on the year. Across years, a site that was Good on average was never Poor in any year, and a site that was Poor on average was never Good in any year. Thus, one can be confident in Good or Poor classifications based on several years of data. However, if only one or a few years of data are available, it is important to be conservative in the use of stream classifications because annual variation may have influenced that classification.
Annual Variation of MAIS scores at individual sites