Response of stream invertebrates to a global-warming thermal regime: an ecosystem-level manipulation

Ecology, March, 1996 by Ian D. Hogg, D. Dudley Williams

INTRODUCTION

The human-enhanced greenhouse effect (global warming) may result in one of the most rapid changes in temperature ever experienced by the earth's biota (Ojima et al. 1991, Esser 1992). Global-warming scenarios forecast increases in mean global air temperatures of 1.5-4.5 [degrees] C following a doubling of atmospheric C[O.sub.2] (IPCC 1990). For latitudes [greater than]30 [degrees] N the temperature increase may be even larger, particularly during winter (Hengeveld 1990). These temperatures will translate directly into changes in the temperature of running waters (Meisner et al. 1987). The need for research on the ecological consequences of such a climate change has received formal recognition through the Sustainable Biosphere Initiative of the Ecological Society of America (Lubchenco et al. 1991). Unfortunately, our ability to predict the consequences of elevated temperatures in lotic systems is limited by our lack of knowledge of natural phenomena (Sweeney et al. 1992).

Previous studies examining the effects of temperature on stream invertebrates have been restricted to laboratory experiments (Sweeney et al. 1986), studies using artificial outdoor channels (Rempel and Carter 1987), or field surveys (Harper 1973). For example, it has been shown that increased temperatures may increase the respiratory rate of animals at the expense of growth (Sweeney and Schnack 1977) and fecundity (Sweeney 1978), and that significant changes in growth rate, adult size, and onset of adult insect emergence can occur as a result of relatively small shifts in temperature (e.g., 1-3 [degrees] C; Langford 1975, Sweeney et al. 1986, Rempel and Carter 1987, Baker and Feltmate 1989). While such studies provide useful insights into the potential response of ecosystems, the direct extrapolation of their results to natural ecosystems is often confounded by the complexity of natural systems in responding to perturbation. Perhaps the most frequently advocated approach to evaluating the effects of such natural phenomena is the use of large temporal- and spatial-scale field experiments (Hall et al. 1980, Odum 1990, Schindler 1990, Mooney 1991). However, to date there have been few attempts to undertake large-scale field experiments in lotic systems, perhaps because of the considerable logistic challenges that are inherent to these habitats (Hairston 1989, Hogg et al. 1992, 1995).

Here, we present the results of a large temporal- and spatial-scale field experiment that was designed to test the effects of increased temperature on stream invertebrates. We heated a permanent first-order stream (Valley Spring) near Toronto, Ontario, Canada, in accord with global-warming predictions for the southern Great Lakes Region (Environment Canada, unpublished data), and examined the effects of the manipulation on total animal densities, biomass, and species composition. Further, we performed a detailed analysis of life history parameters for three resident species [Nemoura trispinosa Claassen (Plecoptera), Lepidostoma vernale (Banks) (Trichoptera), and Hyalella azteca (Saussure)(Amphipoda)] including: (1) growth patterns (immature to adult); (2) size at maturity; (3) fecundity (total number of eggs per female); and (4) emergence patterns for adult insects. The three species are widely distributed in eastern North America (Flint and Wiggins 1961, Strong 1972, Harper 1973), and represent taxa that are widely separated phylogenetically - a hemimetabolous insect, a holometabolous insect, and a crustacean, respectively. Both N. trispinosa and L. vernale tend to be restricted to cool-stream habitats (0-20 [degrees] C; Flint and Wiggins 1961, Harper 1973), whereas H. azteca is found in a variety of habitats including cool streams and subtropical ponds (Edwards and Cowell 1992). Accordingly, we expected that N. trispinosa and L. vernale would respond negatively to increased temperature whereas H. azteca would be more adaptable in response to our experimental thermal regime.

On the basis of species' distributions and previous literature, we predicted that our temperature manipulation would result in: (1) decreased total animal densities, total biomass, and taxon richness (Odum 1985); (2) increased animal growth rates (shorter life cycle) (Odum 1985); (3) earlier emergence of adult insects (Langford 1975); (4) smaller size at maturity (Vannote and Sweeney 1980); and (5) decreased fecundity (Sweeney 1978).

METHODS

Study area

Valley Spring is a small (1 m wide x 60 m long x 3.5 cm deep) first-order stream located in southern Ontario (43 [degrees] 45[minutes] N, 79 [degrees] 15[minutes] W; 150 m elevation). The stream has a single point of issue (25 cm across), and has an annual discharge of 1800-2300 L/h. The substratum consists of a mixture of coarse sand/silt with [approximately equal to]20-30% aquatic macrophyte cover (primarily Nasturtium officinale). The life histories of resident invertebrate species and their temporal and spatial zonation are known based on 2 yr of pre-manipulation baseline data (Norton et al. 1988, Williams and Hogg 1988, I. D. Hogg and D. D. Williams, unpublished data). Annual water temperature ranges from [approximately equal to] 7 to 18 [degrees] C at the stream source. For further description see Williams and Hogg (1988).