Ohio Caddisfly Safari

1994 Summary

Five years of assessing wetland structure and function using adult insects: what have we learned ?

R. J. Garono, Ph.D.
Oregon State University Tillamook Bay National Estuary Project Garibaldi, OR 97118

James G. Kooser
Rettew Associates, Inc. Mechanicsburg, PA 17055

This study was initiated in 1990 in order to determine if it was possible to use adult insect assemblages to characterize wetland structure and function. The success of compensatory wetland replacement is frequently judged on the basis of percent vegetation cover. Measuring wetland vegetation cover may not necessarily reflect the suite of wetland "functions" that were intended to be replaced. Measuring cover of carefully selected and planted species, especially after only 1 or 2 years, seems to be tautological. Water quality parameters are frequently not measured at all. Aquatic insects may provide an integrative mechanism by which wetland structure and function can be measured. The purpose of our current work is to determine if (1) wetland insect assemblages are distinct from those of adjacent upland areas, (2) different types of wetlands can be distinguished from one another, and (3) wetland plant community structure and composition is related to insect assemblages.

We chose light trapping as a method to collect insects from wetland areas. This method traps positively phototaxic insects from a sampling area that is known to be affected by meteorological conditions. However, light trapping offers advantages in that it is non-invasive and therefore, can be used in fragile wetland areas; it integrates over a relatively wide area and adult insects (especially, members of Trichoptera) can be identified to the species level. Aquatic larvae, especially members of the order Trichoptera, cannot be readily identified to species. In 1991 and 1992, we compiled light trap data from the literature and supplemented those data with collections from several additional wetland sites (including adjacent upland control areas and several replacement wetlands). We found that caddisfly species could be used to distinguish between freshwater coastal wetlands, inland lakes and ponds, Sphagnum-dominated wetlands and saturated-soil wetlands. Later ordinations showed that replacement wetlands were more similar to upland control sites than to other wetlands. We modified our sampling protocol in 1994 to include light trapping on three consecutive nights at each site, to help account for differences between sample nights. We also began to analyze higher taxonomic groups (order and family level) in an attempt to reduce the level of effort necessary for sample processing. In addition, we also began to collect detailed information on the structure and composition of wetland plant communities.

Thousands of acres of wetlands are lost each year due to filling, diking and development. To offset the loss of wetland function and values, wetlands are often created or existing degraded wetlands are augmented or enhanced. These created or enhanced wetlands are intended to replace those ecological 'functions' that were lost when the original wetland was destroyed. However, the success of wetland creation /enhancement actions is frequently evaluated by measuring acreage of wetland created, numbers of plant or of plant cover. Most often, the ecological functions that were intended to be replaced by wetland construction or enhancement are not directly considered in the success criteria; simply measuring acreage, plant cover or plant survival does not measure ecosystem function. Furthermore, plant species and densities are often directly manipulated by wetland creators.

Aquatic insect larvae are commonly used as bioindicators in stream ecosystems (Plafkin et al, 1989; Ohio EPA, 1988), we propose to extend the use of insects as bioindicators to wetlands. Aquatic insect larvae and adults have specific ecological requirements that must be met in order for a particular species to be represented in a wetland insect assemblage. For example, aquatic insect larvae have specific requirements for dissolved oxygen, pH, water temperature, water velocity, and food items (predator, herbivore, particle size, etc.). We have asked whether adult insects collected by light trap exhibit preferences for specific types of wetlands and whether insect assemblages can be related to aspects of wetland vegetation (i.e., composition, structure, etc.). By relating insect inventory data to specific wetland attributes we hope to develop a tool by which to evaluate nature preserve management actions and wetland manipulations.

Caddisflies are well suited to be aquatic bioindicator organisms because the larvae of this order, except for a few species that have secondarily adapted to terrestrial environments (Flint, 1958; Anderson, 1967), are wholly aquatic. Trichoptera is a species-rich order; for example, there are approximately 250 species of caddisflies known from Ohio (Huryn and Foote, 1983). Since caddisfly larvae occur in a wide variety of aquatic habitats: ponds, lakes, streams and in saturated soils (Borror et al., 1964), and belong to several different trophic categories: predators, scraper, shredders, collectors, piercers, etc. (Merritt and Cummins, 1984), caddisflies may be useful as bioindicators. The large number of species, each with its own ecological requirements have made it possible for caddisflies to be used to characterize wetlands in Ohio wetlands (Garono and MacLean, 1988; Garono and Kooser, 1994). Furthermore, since most caddisfly larvae are confined to aquatic ecosystems, they are exposed to the entire suite of conditions present in the ecosystem being sampled, and therefore may serve as integrative measures of habitat quality and water quality functions. Thus, representation of various caddisfly trophic guilds in light trap collections may reflect wetland structure and heterogeneity.

Early work by Garono (1986) indicated that caddisflies collected from 14 Ohio peatlands could be used characterize those wetlands (Table 1). In his study, 37,061 caddisflies were collected at bimonthly intervals from May-October of 1984-85 and identified to species.

In order to evaluate the use of wetland insect populations as a tool for evaluating constructed wetlands, Garono and Kooser (1994) sampled adult insects at two replacement wetlands during 1991-93 and combined those data with light trap collection records from other sites. Wetlands sampled included Lake Erie marshes, peatland and replacement wetlands. At each site light traps were operated at adjacent upland sites in order to compare upland and wetland collection results. To minimize seasonal differences, only August light trap collection results were used in the analysis. Instead of caddisfly species, genera were used in this analysis to see whether wetlands could be characterized on the basis of Trichoptera genera.

Results of the 1994 study showed that wetlands could be characterized on the basis of their caddisfly assemblages: sites with open water appeared separate from those with sedge or Sphagnum mats. Caddisflies collected in newly created wetlands were similar to those collected at upland sites.

Beginning in 1994, the sampling methodology was modified to sample each wetland (both wetland and upland sites) for two to three evenings on each visit. In order to determine how well a single light trapping event represented the insect fauna at each site. We expanded the scope to of the study to include all insects collected by the light trap, aquatic and terrestrial. We also were interested in whether an ordination of insect orders could be used to characterize wetland sites; this would greatly facilitate the processing of insect material by persons with little taxonomic training. Finally, to determine if this technique could be used in wetlands in other parts of the us, studies were initiated by our group in Texas (1993) and in Oregon (1994: Figure 1).

Figure 1. Study locations where our group is collecting adult aquatic insects with light traps in order to characterize aquatic ecosystems (rivers, bogs, fens, marshes, lakes, flatwood ponds, baygalls, playa lakes, desert springs, forested wetlands, replacement wetlands, reservoirs).

The results presented here are the results from our 1994 ohio wetland study. the objectives of this study are:

1. to determine if adult insects collected by light trap can be used to characterize wetland and upland sites in ohio;
2. to determine how well represented the wholly aquatic orders are in light trap collections;
3. to determine if insect assemblages are related to the structure and composition of wetland vegetation.

Sampling insect populations: Insect populations were sampled using portable, battery powered fluorescent light traps (Figure 2). At least two traps per site, one in the wetland and one in an adjacent upland, were operated from dusk to dawn. To preserve specimens, light traps were filled to a depth of approximately 2 cm with 70% ETOH. At each site, light trap collections were made on three evenings during a 10-day period (11- 20 August 1994). All insects collected were sorted to order. Caddisflies will be identified to species, when possible. The average number of individuals in each order was calculated from replicate trapping events made at each site. Since we are interested in developing an assessment tool, this study reports results of analyses made at the order-level.

Figure 2. A light trap set up at Morgan Swamp, a forested wetland in NE Ohio.

Ordination of insect data: The average number of individuals per order collected at each site was ordinated using detrended correspondence analysis (DCA) (Hill, 1979). DCA is a descriptive statistical technique that reduces patterns inherent in large, complex, data sets to low-dimensional ordination space. DCA is an eigen analysis procedure, performing both Q-type (in this case, order-level) ordination and R-type (sites) ordination, in which axis scores are derived solely from the data matrix (Gauch, 1982). DCA ordinated wetland sites based solely on the taxonomic composition of insect assemblages. DCA produces axes along which sites with similar compositions cluster.

Evaluation of vegetation structure and composition: Vegetation structure and composition in wetland trap areas was measured using a modified line-intercept method (Kooser and Garono, in prep.). Four 30 m long line transects were established along cardinal directions with the light trap as the center. At two meter intervals along each transect, we recorded the species and height (cm) of each plant part which touches a PVC pole calibrated in cm. This techniques generates a vertical and horizontal representation of the plant community surrounding the light trap. While the effective distance of a fluorescent light trap is not known, evidence from North Carolina and from previous studies completed by our team suggest this distance may be on the order of 20 to 50 m. We selected 30 m as a reasonable transect length for this study.

The data were plotted as cross section profiles showing the distribution of plant "hits" at each sampling point. We calculated the mean number of "hits" per sample point, as an indicator of the vertical density of the plant community. We calculated a "roughness coefficient" as the ratio of the 60 m (total of the 2 adjacent 30 m) transect length to the upper surface of the plant community. In plant communities with a rather flat upper surface, as might be found in a wetland with a large open water component, this number approaches 1. As the upper surface becomes more complex, the coefficient decreases. We also determined whether the distribution of holes, or vertical and horizontal spaces in the community, is clumped, random or uniform using an index of dispersion.

Fifty two light trap events were made at ten Ohio sites (Figure 3). The study resulted in the collection of 96,125 individuals from 52 light trap events. Thirteen orders were represented in the light trap collections, including several members of the order Odonata which are not generally considered to be night time fliers. Diptera was the most well represented order at all sites: the average number of individuals belonging to Diptera was greatest all sites, except for one upland site, Rosemont Country Club. Not all members of Diptera are aquatic. Additional work is underway to determine which families or genera are likely to be important in wetlands.

Figure 3. Map showing light trap locations at the 10 Ohio sites sampled during this study.

The wholly aquatic orders, Trichoptera, Ephemeroptera, Plecoptera and Odonata, accounted for 28.0% of the total number of individuals collected during this study. More than 25%of the individuals collected belonged to the order Ephemeroptera. The great number of mayflies was collected at Winous Point (Lake Erie) at a lake and wetland site. Although large numbers of mayflies were observed in all three collections at Winous Point, large numbers of Ephemeroptera were not observed at other similar Lake Erie sites (Arcola Creek, Old Woman Creek). Insect material from 1995 and 1996 will help us to determine if the large number of mayflies observed in 1994 are characteristic of Winous Point or the result of an episodic hatch.

Of the insects observed in all light trap collections, only 2.09 percent of the individuals belonged to the order Trichoptera . However, Trichoptera was the fifth ranked order at all sites and members of this order were observed in 45 (out of 52) collections. Caddisflies were the third ranked order from all wetland sites and the fifth ranked order from all upland sites. Wetland and upland insect assemblages differ even at the order level.

A detrended correspondence analysis was performed on the data matrix of sites vs average number of individuals per order to better describe patterns in the insect assemblages among sites. DCA Axis I accounted for 95% and Axis II for 4% of the variability in the original data matrix, respectively. While there were no distinct clusters along either Axis I or II, several patterns can be seen (Figure 4). Lake Erie wetland sites range along Axis I from 0 to 150. Upland sites range along Axis I from 175 to 400. Replacement wetlands and inland bogs and fens appear near the center of Axis I. Additional ordinations are planned to investigate subsets of this data matrix.

Figure 4. Ordination resulting from DCA of Avg. No. Insect Orders by site.

Finally, to determine if vegetation structure was related to light trap results, correlations were calculated between the vegetation roughness coefficient and DCA Axis scores, the average number of individuals and the average number of orders at each site. Vegetation was sampled at eight wetlands. The roughness coefficient was strongly and negatively correlated with DCA Axis I (- 0.811) and correlated with the average number of individuals at each site (0.610). Wetlands with low Axis I scores had higher roughness coefficients, meaning the upper surface of the vegetation formed a relatively flat surface. Examples include Nelumbo- or Pontedaria-dominated wetlands with large open water components. Wetlands with high Axis I scores had low roughness coefficients, rougher upper surfaces, and therefore greater structural heterogeneity. Examples include fens and shrub dominated bogs. Axis I may represent an open water-shrub-tree gradient. Additional work is necessary to relate the structure and composition of plant communities to insect assemblages, especially at our upland sites.

1. The order Diptera was the most well-represented order at all sites.
2. There are differences in the insect assemblages between upland and wetland sites, even at the order level. Wholly aquatic insect orders accounted for 28 percent of the total number of insects collected. The order Trichoptera was ranked fifth in the overall collections, third for the wetland collections and fifth for upland collections.
3. Detrended Correspondence Analysis (DCA) was used to visualize patterns in a data matrix constructed from the average number of individuals for each order from each site. DCA Axis I accounted for 95% of the variability in the original data matrix. Lake Erie wetland sites had low Axis I scores and upland sites had high Axis I scores.
4. The ratio of the length of the transect to the "roughness" of the vegetation surface was strongly correlated to DCA Axis I and to the average number of individuals at eight of the wetland sites. Axis I may represent an open water---shrub---tree gradient.

Literature Cited

Anderson, N. H. 1967. Life cycle of a terrestrial caddisfly, Philocasca demita (Trichoptera, Limnephilidae), in North America. Ann. Ent. Soc. Am. 60: 320-323.

Borror, D. J., D. M. Delong, and C. A. Triplehorn. 1976 (1964??). Study of Insects. Holt, Rinehard and Winston Publishers, N.Y., New York. 852 p.

Flint, O. S. 1958. The larva and terrestrial pupa of Ironoquia parvula (Trichoptera, Limnephilidae). J. N. Y. Ent. Soc. 66: 59-62.

Garono, R. J. 1986. A survey of Trichoptera of Ohio remnant bog and fens. M.S. Thesis, Youngstown State University, Youngstown, Ohio. 204 pp.

Garono, R. J. and D. B. MacLean. 1988. Caddisflies (Trichoptera) of Ohio wetlands as indicated by light-trapping. Ohio J. Sci. 88: 143-151.

Garono, R. J. and J. G. Kooser. 1994. Ordination of wetland insect populations: evaluation of a potential mitigation monitoring tool. In: W.J. Mitsch, ed. Global Wetlands: Old World and New. Elsevier Science Publishers. Amsterdam, The Netherlands. pp. 681-691.

Gauch, H. G. 1982. Multivariate analysis in community ecology. Cambridge University Press, N. Y., New York.

Hill, M. O. 1979. DECORANA - A FORTRAN program for detrended correspondence analysis and reciprocal averaging. Cornell University, Ithaca, N. Y.

Huryn, A. D. and B. A. Foote. 1983. An annotated list of the caddisflies (Trichoptera) of Ohio. Proceedings of the Entomological Society of Washington. 85: 783-796.

Kooser, J.G. and R.J. Garono. In prep. Vertical and horizontal structure of natural and constructed wetlands in Ohio.

Kooser, J.G. and R. J. Garono. 1993. Ordination of wetland insect populations: continued evaluation of a potential mitigation monitoring tool. The Ohio Journal of Science 93(2):44.

Merritt, R. W. and K. W. Cummins. 1984. An introduction to the aquatic insects of North America. Kendall Hunt Publishing Co. Dubuque, Iowa. 722 p.

Ohio Environmental Protection Agency. 1988. Biological Criteria for the Protection of Aquatic Life. Volume II, Users Manual for Biological Field Assessment of Ohio Surface Waters. Ohio EPA, Columbus, OH.

Pielou, E. C. 1966. The measurement of diversity in different types of biological collections. J. Theo. Bio. 13: 131-144.

Plafkin, J.L., M.T. Barbour, K.D. Porter, S.K. Gross and R.M. Hughes. 1989. Rapid Bioassessment Protocols for use in Streams and Rivers. United States Environmental Protection Agency, Washington, D.C.

We wish to thank the following for their financial contributions: The Nature Conservancy of Ohio, Ohio Biological Survey, The Nature Conservancy of Texas, US EPA, Region V, Texas Natural Resource Conservation Commission, Oregon Dept. of Fish & Wildlife, and Annie's Kayaks.

We also with to thank the following for their assistance: Board of Trustees of Rosemont Country Club, Ohio Department of Natural Resources, Division of Natural Areas & Preserves, Dr. J. Strittholt, P. Frost, Dr. R. L. Kiesling, Dr. E. B. Long, Dr. Roy Kroll, and E. Gienger.

Return to Dr. R.J. Garono's Homepage