---

Introduction

In the strictest sense, current concern about global climate change refers to the anticipated altering of temperature and precipitation resulting from the anthropogenically induced accumulation of greenhouse gases. Greenhouse gases, which include water vapor, carbon dioxide (CO₂), methane (CH₄), nitrous oxide (N₂O), tropospheric ozone (O₃) and chlorofluorocarbons (CFCs), are relatively transparent to incoming solar radiation but absorb outgoing infrared radiation emitted from the Earth's surface. All else being equal, increasing concentrations of these gases will cause temperatures to rise at the surface of the Earth (Manabe, 1983).

The gases themselves also influence the response of biotic systems to climatic change. Effects of carbon dioxide on plant growth varies among species and thus can change competitive interactions (Bazzaz and McConnaughay, 1992). CFCs destroy the ozone layer and lead to increased penetration of biologically active wavelengths of ultraviolet light (Liu et al., 1991). The effect of increasing greenhouse gas concentrations in the atmosphere on biotic systems is thus not simply one of changing climate over time, but includes complex and likely synergistic responses of biotic systems to climate, atmospheric gases and ultraviolet radiation.

The purposes of this paper are to: (1) briefly summarize knowledge of some direct and indirect effects of greenhouse gases on biotic systems; and (2) describe potential effects on bird populations and their habitats, with emphasis on the northern Great Plains.

Direct Effect: Carbon Dioxide Fertilization

Preindustrial atmospheric concentrations of CO₂, as determined from ice cores, were in the range of 270-290 parts per million (ppm) (Smith and Tirpak, 1990). The concentration in 1988, as measured at Mauna Loa Observatory, was approximately 350 ppm (MacDonald, 1990), with predicted increases of 2.4% per year continuing into the middle of the next century (Smith and Tirpak, 1990).

Atmospheric CO₂ has been shown to influence photosynthesis, transpiration, phenology and leaf nutrient composition of terrestrial plants. These effects have been thoroughly reviewed elsewhere (Bazzaz, 1990; Morison, 1990; Mooney et al., 1991); here I will summarize briefly the points salient to this discussion.

Plants with C₃ metabolism generally show higher rates of photosynthesis under elevated CO₂ than do C₄ plants. Plants having either metabolic pathway respond with declining stomatal conductance and transpiration, especially where water is limiting. In general, availability of light, water and nutrients modifies the magnitude of the response to CO₂ enrichment; if other resources are or become limiting, the fertilization effect is depressed.

Leaves grown under enhanced CO₂ in which a fertilization effect has occurred generally have lower nitrogen concentrations, a consequence of dilution by carbon-based products of photosynthesis (i.e., the carbon:nitrogen ratio of plant tissues increases). Invertebrate herbivores feeding on such leaves experience slower growth and maturation and in some cases increased mortality (Lincoln et al., 1993). Plants that defend themselves from herbivores with carbon-based allelochemicals may limit compensatory feeding to the extent that herbivore growth must slow due to inadequate nitrogen intake. On the other hand, nitrogen-based allelochemicals may be less abundant in foliage (Lincoln and Couvet, 1989).

Changes in growth and maturation rates of herbivorous invertebrates will affect organisms both higher and lower on the food chain (Fajer, 1989). Animals whose breeding cycles are endogenously timed to peaks in populations of herbivorous insects may find insufficient food supplies at critical periods. To secure the required nitrogen reserves, herbivores will have to consume more biomass, albeit over a longer period of time, causing further damage to their hosts.

To date, studies of effects of elevated CO₂ on herbivores have dealt primarily with phytophagous insects, but herbivorous vertebrates also may be expected to respond to changes in leaf nitrogen content. For example, sage grouse (Centrocercus urophasianus), a nonmigratory herbivorous bird of the western United States, feeds exclusively on sagebrush (Artemisia spp.) leaves and buds. Sage grouse have been shown to choose sagebrush subspecies and individuals having more nitrogen than plants sampled at random (Remington and Braun, 1985). If nitrogen is limiting to sage grouse, their vulnerability to changes in vegetation nutritional quality could be expressed as reduced over-winter survival or declining reproductive rates.

Animals that feed on pollen, nectar, fruit or seeds are often closely tied to one or a few plant species that flower or fruit at the appropriate time of year. Similarly, some plant species are dependent on a few species of animals for pollination or seed dispersal (Howe, 1984). Studies of flowering phenology under increased CO₂ have indicated the potential for changes in flowering date, but the direction and magnitude of the changes have been highly variable (St. Omer and Horvath, 1983; Garbutt et al., 1990; Reekie and Bazzaz, 1991; Rawson, 1992). Studies of fruiting phenology under increased CO₂ are lacking.

Interactive effects of temperature and CO₂ enrichment on plants for the most part elude generalization (Rawson, 1992). The magnitude and direction of the effects are highly dependent on availability of other resources and on each species' phenotypic plasticity.

Submerged and floating aquatic plants have received less attention with respect to direct effects of CO₂ enrichment than have terrestrial and emergent species. An early review (Lemon, 1983) suggested that, because submerged plants have access to both dissolved CO₂ and bicarbonate, relatively little response to increased atmospheric CO₂ might be expected. However, Barko et al. (1991) found evidence for potential shifts in competitive balance between the submerged Valisneria americana and Hydrilla verticillata under various levels of CO₂ concentration and nutrient availability. Work on water hyacinth (Eichhornia crassipes), a floating aquatic plant, has indicated substantial increases in biomass production under doubled CO₂ conditions (Idso and Kimball, 1985).

Indirect Effects: Ultraviolet B Radiation

The Earth is protected from short-wavelength, biologically destructive, ultraviolet radiation by the stratospheric ozone layer. Ozone absorbs all of the shortest wavelength, UVC (<290 nm), most of the intermediate wavelength, UVB (290 - 320 nm), but none of the longer wavelength ultraviolet radiation, UVA (320 - 400 nm). While destruction of the ozone layer by CFCs is not expected to influence transmittance of UVC, significant increases in UVB at the Earth's surface are expected. The transmittance of longer wavelength UVA radiation, important in photosynthesis, vitamin metabolism and animal orientation, will not be affected by changes in the ozone layer. Intensity of incident ultraviolet radiation is dependent on path length, such that higher-elevation and lower-latitude sites experience higher intensities (Caldwell et al., 1980).

Estimates of incident ultraviolet radiation can be made based on calculations combining measured depth of ozone and propagation of solar radiation through the atmosphere. These estimates suggest substantial increases in UVB radiation at middle and high latitudes, but little change in the tropics (Madronich, 1992). Satellite data for 1992 indicate declines in ozone over much of the Northern hemisphere (Gleason et al., 1993).

The amount of UVB reaching the Earth's surface also can be modified by anthropogenic aerosols, especially SO₂ (Liu et al., 1991) and smoke (Penner et al., 1992). Liu et al. (1991) estimated that rural areas of industrialized countries have experienced a 5 to 18% reduction in incident UVB since the industrial revolution. However, because SO₂ is not spread evenly over the globe, a simple canceling effect is not possible.

Wavelengths included within the UVB range are readily absorbed by proteins and nucleic acids, thus potentially disrupting fundamental cellular functions in both plants and animals. The majority of research on the effects of UVB has been conducted on crop plants and phytoplankton. No organism has yet been found with the ability to detect UVB radiation (Hader and Worrest, 1991). Some of the effects on plants of enhanced UVB radiation include smaller leaf area but higher chlorophyll concentration, decline in CO₂ fixation rate, chemical changes such as increases in various protective chromophores and in concentrations of pigments, increase in thickness of cuticular waxes and changes in overall growth and yield (Caldwell et al., 1989; Teramura and Sullivan, 1991; Tevini et al., 1991; Kramer et al., 1992; Sullivan et al., 1992; Krupa and Kickert, 1993).

As with the effect of CO₂, substantial interspecific differences in UVB sensitivity have been found (Hader and Worrest, 1991; Teramura and Sullivan, 1991; Tevini et al., 1991; Day et al., 1992; Sullivan et al., 1992; Krupa and Kickert, 1993). Even cultivars of the same crop plant may respond idiosyncratically to UVB (Teramura and Sullivan, 1991). Day et al. (1992) used a fiber-optic microprobe to show that the epidermis of conifer needles filters all incident UVB before it reaches the mesophyll, whereas the epidermis of herbaceous dicots transmits 18-41% of incident UVB to mesophyll tissue. Plants from high elevations at low latitudes have lower epidermal transmittance of UVB (Robberecht et al., 1980) and are generally less susceptible to UVB damage (Teramura and Sullivan, 1991; Sullivan et al., 1992), both of which are presumably adaptations to higher UVB in those environments.

Most research on effects of UVB on terrestrial animals has been concerned with public health. Exposure to UVB has been related to increased risk of skin cancer and cataracts in humans (Taylor, 1990; Longstreth, 1991). UVB also has been shown to suppress cellular immune function in shaved laboratory mice (Daynes, 1990). A recent review cites potential increases in a variety of diseases related to eyes and other exposed areas of domestic animals (Mayer, 1992).

Evidence exists that temperate aquatic organisms even now are negatively affected by current levels of UVB (Gala and Giesy, 1991). Aquatic organisms whose early life stages occur near the water's surface and have relatively little protective pigmentation may be exposed to lethal levels of UVB. In agricultural areas, byproducts of pesticides irradiated with UVB will result in new, and potentially more toxic, substances accumulating in wetlands (Oris and Giesy, 1985; 1986).

As with the CO₂ fertilization effect, response to increased UVB has the potential to change competitive interactions among plants, and perhaps among animals, depending on individual responses (Caldwell et al., 1989; Krupa and Kickert, 1993). Plants that respond to increased UVB by reduction in stature may experience greater shading by less sensitive neighbors (Sullivan et al., 1992). By changing the leaf area and relative size of some plants, interactions with herbivores may be altered, especially in those plants in which apparency (sensu Feeny, 1976; Rhoades and Cates, 1976) is related to attack. The disruption of nitrogen assimilation and increased cuticular resistance induced by UVB, in combination with the dilution effect of CO₂ fertilization, may further stress herbivores. Increased phenolics and cuticular waxes in vegetation also will slow decomposition (Bowes, 1993).

Data on interactions between increased UVB and CO₂ in plants under field conditions are lacking, although Krupa and Kickert (1993) discuss an experimental framework for exploring such interactions. Greenhouse experiments have demonstrated some compensation in biomass accumulation when a C₃ grass (Elymus athericus) was grown under both elevated UVB and CO₂ (van de Staaij, 1993). The effect of elevated UVB in combination with water stress has received more attention. In general, water stress additively modifies a plant's response to UVB by decreasing photosynthesis (Teramura and Sullivan, 1991). However, Balakumar et al. (1993) found smaller declines in growth response of cowpea (Vigna unguiculata) seedlings exposed simultaneously to drought and elevated UVB than to seedlings exposed to either stress singly.

Indirect Effects: Climate Change in the Great Plains

The potential effect of anthropogenic aerosols receiving the most attention is climatic forcing, actual shifts in temperature and precipitation caused by the "greenhouse effect." This is also the effect that is the least understood. Predictions of climate change 75 years from now are based on models that fail to simulate many aspects of the present climate (Grotch, 1988; Hulme, 1991; Karl et al., 1991b; Walsh and Crane, 1992). The most finely resolved of the global predictive models has a spatial resolution of 4° x 5° , roughly the size of Colorado, although more finely resolved models have been coupled to the general circulation models (GCMs; e.g., Dickinson, 1989). Regional predictions tend to vary greatly among the different models (Grotch, 1988; Cushman and Spring, 1989) and extrapolation to the finer scale more typical of traditional ecological research is fraught with uncertainty (Root and Schneider, 1993). With each added level of uncertainty in complex models, risk of erroneous conclusions is multiplied (Liebetrau and Scott, 1991). Nonetheless, all models agree that, on average, the Earth will warm in response to the buildup of anthropogenic aerosols in the atmosphere.

Climate in the Great Plains is a product of the strength of warm, dry, westerly winds (Borchert, 1950). Air descends from the peaks of the Rocky Mountains, having lost its moisture on the windward side (Manabe and Broccoli, 1990). When these winds are strong, they deflect moister air moving up from the Gulf coast and cold air masses moving down from the Arctic; diminished westerlies produce precipitation by allowing the moist tropical air and cold, dry, Arctic air to clash. The gradient of increasing precipitation from west to east across the Great Plains is due to greater intensity of storms in the east, rather than more days in which precipitation falls (Borchert, 1950).

Four of five GCMs agree about predicted climate in the northern Great Plains (Smith and Tirpak, 1990; Karl et al., 1991a). In general, the models predict higher average temperatures, a northward shift of the mid-latitude rain belt in the summer and a southward shift of the rain belt in the winter (Manabe and Wetherald, 1986). Precipitation is expected to decrease in the southern portion of the belt but increase in the northern portion, resulting in increasing winter and declining summer precipitation in the northern Great Plains. However, if temperatures warm earlier in the spring, snow will melt sooner, albedo will decrease and soil will absorb more solar energy. Soil moisture thus will be depleted sooner, and diminished summer rains are likely to be insufficient to replenish the earlier moisture deficit, which therefore will persist through the summer (Manabe and Wetherald, 1986, 1987; Wetherald, 1991). Increased extent of drought conditions at mid-latitudes is predicted, as is increased intensity of summer storms (Hansen et al., 1991).

While GCMs generally agree in their predictions of climate for the northern Great Plains, the effects of the predicted climate are far from certain. For example, when Mitchell and Warrilow (1987) added soil classification and soil temperature layers to a climate model, they found that the model's treatment of runoff over frozen ground could dramatically change the course of projected soil moisture depletion during the summer. Until such details of physical processes are understood, even the most accurate predictions of temperature and precipitation are of limited utility.

Despite projections that greenhouse warming will be detectable during the 1990s (Hansen et al., 1991), it now appears that it may be several decades before predicted changes in temperature and precipitation will be distinguishable from background variation, at least in the Great Plains (Karl et al., 1991a). In addition, a recent analysis of temperature trends has indicated that most of the 0.5 C warming currently observed in the Northern Hemisphere has occurred in the mean daily minimum temperatures, a result not anticipated in any of the GCMs (Karl et al., 1991b). This result also holds for temperatures from the Historical Climatology Network, analyzed for the Jamestown, North Dakota weather station (Todhunter, 1993). If temperatures are indeed increasing only at night, this clearly has important implications for biotic response, especially with respect to the advantages of different photosynthetic pathways.

Organisms may not respond in concert with increased temperatures. For example, Dewar and Watt (1992) explored the potential effect of higher temperatures on emergence of winter moth (Operophtera brumata) larvae and budburst of the larvae's host, Sitka spruce (Picea sitchensis). They found that budburst is far less sensitive to temperature than are the larvae, potentially leading to less coincidence between emergence of the herbivore and availability of its host plant as temperatures rise. Porter et al. (1991) modeled the effects of climate change on a crop pest, the European corn borer (Ostrinia nubilalis), and found that increased temperatures may lead not only to a northward shift in its distribution, but to an additional generation each growing season, which would have important consequences for agriculture.

Uncertainties in scale, extent and distribution of climatic change, coupled with our limited understanding of life histories of most organisms, make anticipating the response of biotic systems exceedingly difficult. An immense body of literature exists concerning modelled responses to climate change; these models have been widely reviewed (e.g. Shugart, 1990; Austin, 1992; Malanson, 1993). Malanson's review is unique in that it explicitly addresses information the models cannot provide.

Only a few quantitative models have been developed for wildlife response to climate change, and none deals explicitly with the fauna of the northern Great Plains. Nonetheless, it is worth considering at least one exemplary model. Rodenhouse (1992) constructed a detailed simulation model for black-throated blue warbler (Dendroica caerulescens) annual production in northeastern U.S. hardwood forests, by which he then predicted possible changes in production caused by varying temperature and precipitation. Importantly, this model explicitly deals with direct physiological effects of climate change on the breeding adults as well as effects of climate on predators and food abundance; by doing so it points out critical gaps in our understanding, such as the relationship between temperature and precipitation and abundance of the arthropod prey upon which most breeding birds rely. Such a model is needed for species common to other habitat types, if individualistic responses are to be assessed (Malanson, 1993).

Koford et al. (1992) have developed a detailed population model for female mallards (Anas platyrhynchos). Although this model has not been explored in terms of climatic change, it contains elements, such as spring temperatures and availability of wetland habitat, that rely on climate either directly or indirectly. The model would be a valuable tool for exploring various climate scenarios.

Some basic data exist concerning distribution of passerines with respect to climate over the broad geographic areas of interest in climate change research. Climatic limits to winter distributions of North American passerines have been described and related to the metabolic costs of existence (Root, 1988a, 1988b). Similarly, climatic limits of breeding grassland birds are being modeled and subjected to doubled CO₂ scenarios (J. Price, personal communication). Both of these studies examine the degree of year-to-year variation expected in species' ranges under the current climate. They also make obvious those species that do not fit a climate-mediated pattern, thus suggesting higher-order interactions.

Implications of the Buildup of Greenhouse Gases for Birds in the Northern Great Plains

Having considered the complexity of potential biological results of anthropogenically generated greenhouse gases, we can ask what effects the buildup of greenhouse gases may have on birds inhabiting the northern Great Plains. Because birds are relatively well studied, an examination of potential effects on them and their habitat should provide insights into possible effects of climate change on other, less well studied organisms.

Reproductive success is one of the most direct assessments of the effects of greenhouse gases on birds that nest in the northern Great Plains. With reproductive success as the endpoint, I have developed a schematic showing both direct and indirect effects of UVB, CO2, temperature and precipitation (Fig. 1). Although this is a highly simplified representation, the diagram makes evident several points at which interactions occur among variables. Rarely can the magnitude of the interaction be quantified; often the values of the variables contributing to the interactions are themselves unknown.

Avian reproductive success largely depends on four factors: habitat suitability, food availability, intensity of predation and/or brood parasitism and the ability of both adults and young to avoid disease. Each of these factors is itself the product of a cascading series of variables, many of which are vulnerable to the effect of changing concentrations of greenhouse gases. A discussion of each of the four factors in light of what we do and do not know about their response to greenhouse gases will help highlight areas in which research is needed.

Figure 1. Potential interactions among effects of ultraviolet B radiation, carbon dioxide, temperature and precipitation that may influence avian reproductive success in the northern Great Plains. Direct influences are indicated with solid lines, indirect with dashed lines.

Implications of the Buildup of Greenhouse Gases for Birds in the Northern Great Plains:
Habitat

The importance of the prairie pothole region as habitat for breeding and migrating waterfowl is well documented (Boyd, 1981; Krapu et al., 1983; Reynolds, 1987). More than half of the waterfowl production for North America occurs within the prairie pothole region (Batt et al., 1989) and persistence of wetlands through the breeding season is important for brood survival (Talent et al., 1982; Krapu et al., 1983; Leitch and Kaminski, 1985). Waterfowl occupy prairie wetlands for staging and stopover habitat during both spring and fall migration (Faanes, 1982; Kantrud, 1986; Alisauskas et al., 1988; Alisauskas and Ankney, 1992). Kantrud and Stewart (1984) estimated that 26% of the breeding birds other than waterfowl in North Dakota also depend on prairie wetlands.

Detailed investigations of individual wetlands or wetland complexes in the northern Great Plains have provided valuable information on hydrology, evapotranspiration and vegetation characteristics of prairie wetlands (Meyboom, 1966; Shjeflo, 1968; Eisenlohr, 1972; Poiani, 1987; Poiani and Johnson, 1988, 1991; Swanson et al., 1988; Merendino et al., 1990; Merendino and Smith, 1991). Because these studies concentrated on wetlands in relatively small geographic areas, their applicability to the entire pothole region is unknown.

The probability that a wetland will contain water is in part a product of evapotranspiration, precipitation and groundwater flow. The water balance of surrounding vegetation influences the direction of groundwater flow around wetlands (Meyboom, 1966). In general, water-use efficiency of plants will increase under elevated CO₂, with the magnitude modified by temperature and precipitation. Thus, plants can be expected to draw less water, leaving more groundwater for wetland recharge as the growing season progresses. However, higher temperatures will increase evaporation from soil and surface waters; larger and/or more leaves, as a result of CO₂ fertilization, could increase total evapotranspiration at a landscape level. The net amount of water available to fill wetlands will depend on the relative magnitudes of the changes in evaporation, transpiration, and leaf area as well as any changes in timing and magnitude of runoff events.

The value of a wetland as avian habitat depends not only on its water-holding capacity, but also on its chemical composition and the aquatic vegetation it supports. Each of these parameters will be influenced by changes in climate. Higher temperatures, by increasing evaporation, concentrate salts in wetlands, changing the invertebrate community along with the chemistry (Swanson et al., 1988). The direction of groundwater flow also influences wetland chemistry (LaBaugh et al., 1987); possible effects of climate change on groundwater in the closed basins typical of the northern Great Plains have not been well-characterized. Not all wetlands will respond identically to increased evaporation. Hydrology dictates that wetlands functioning as flow-through or groundwater recharge systems contain fewer salts than do those functioning as groundwater discharge systems (Swanson et al., 1988). If increased water use efficiency does not balance increased evaporation and the water table drops as temperatures rise, wetlands can be expected to have even lower water levels and greater ionic concentrations than would be expected from evaporation alone. Although the more saline wetlands are highly productive and constitute good habitat for migrating shorebirds and older waterfowl broods, very saline water is intolerable to young ducklings before their salt glands are fully functional (Swanson et al., 1984). Thus, although more saline wetlands are attractive to waterfowl, the presence of freshwater seeps is critical to successful reproduction. Until we have a clear understanding of the implications of climate change for local and regional water tables, the likelihood of persistence of freshwater seeps cannot be assessed.

Effects of greenhouse gases on aquatic plant communities are complex and very difficult to predict. The amount of water held in a basin has a direct effect on the extent of aquatic vegetation, but wetland chemistry, length of growing season, CO₂ fertilization and effects of UVB on plant growth form will also influence community composition and structure. For example, a comparison of semipermanently flooded wetlands of differing salinity showed that the more saline wetland contained only 40% of the total phytoplankton taxa found in the fresher wetlands (Kantrud et al., 1989); indeed, the general effect of increased salinity on prairie wetland vegetation is a decrease in species diversity. The most saline prairie wetlands support only a single aquatic macropyte: Ruppia maritima.

A simulation model developed for a semipermanent prairie wetland in North Dakota predicted substantial increases in overall vegetative cover under a climate change scenario that included only effects of temperature and precipitation (Poiani and Johnson, 1991; 1993). The model did not include possible changes in water chemistry. By the eleventh year of the simulation, the wetland was completely choked by emergent vegetation, substantially lowering its quality as waterfowl habitat. Blackbirds (Agelaius and Xanthocephalus spp.), rails (Laterallus, Rallus and Porzana spp.) and marsh-dwelling wrens (Cistothorus spp.), on the other hand, should benefit from such changes.

As discussed earlier, little is known about effects of elevated CO₂ on aquatic vegetation. This is an important limitation in our ability to anticipate the effects of climate change. For example, Typha spp. commonly occur along the margins of prairie wetlands, encroaching inward until stopped by either very high water or prolonged drawdown conditions. Because of their potential for eliminating waterfowl habitat, cattails are important components of prairie wetlands from a management perspective. Bowes (1993) pointed out that, because of their ability to transport large amounts of CO₂ derived from the sediment, cattails may respond less to CO₂ enrichment than other C₃ plants. However, the extent to which sediment-derived CO₂ contributes to leaf photosynthesis is unknown. Clearly, questions of basic biochemistry and physiology must be addressed before predictions can be made regarding species-specific responses to enhanced CO₂. Until such detailed mechanistic questions can be resolved, the extent to which wetland habitat will be altered remains uncertain.

Information on the effects of increased UVB exposure on aquatic vegetation in general, and in particular in the northern Great Plains, is virtually nonexistent. Thus, the first step must be to determine whether aquatic plants respond to UVB in ways similar to those expressed by terrestrial plants. Because water attenuates UVB, submerged vegetation, while being potentially less affected, may also be less tolerant, having had little evolutionary exposure to UVB. Caldwell et al. (1989) suggested that because UVB induces production of phenolics, terrestrial plants might become more resistant to detritivores. Such a result in prairie potholes has the potential to disrupt nutrient flow within the wetland from year to year, if decomposition is less complete before wetlands dry. Fallen and standing litter also inhibits germination of seeds of plants that would normally succeed as wetlands dry.

Indirect effects of increased UVB have, perhaps, greater potential for damage to prairie wetland habitat than do direct effects. Wetlands existing among cultivated fields often accumulate pesticides from runoff or drift. Photoenhanced toxicity of these pesticides (Oris and Giesy, 1985; 1986) could fundamentally alter community composition of both plants and animals. Toxicity of pesticides after UVB irradiation should be high on the list of research imperatives.

Most migratory birds in the Great Plains require upland habitat; many, such as ducks, require both upland and wetland habitat. Upland habitat in this region consists primarily of grassland, ranging from shortgrass prairie in the west, through mixed to tallgrass prairie along an increasing moisture gradient from west to east (Bazzaz and Parrish, 1982). These grasslands provide breeding habitat for ground-nesting passerines and raptors (Gilmer and Stewart, 1983; Kantrud and Kologiski, 1983), as well as nesting cover for otherwise wetland-dependent waterfowl (Stewart and Kantrud, 1974). Formerly expansive grasslands have already been fragmented by farming and livestock grazing, with resultant declines in some populations of upland-nesting birds.

The CO₂ fertilization effect will influence availability of upland habitat. If C₃ plants are favored over C₄, more spring cover could lead to higher nesting success for early-nesting species such as mallards, northern pintails (Anas acuta), greater prairie chickens (Tympanuchus cupido) and sharp-tailed grouse (T. phasianellus). These species are well-studied, so that efforts to model their response to CO₂-induced habitat changes should be fruitful.

Upland habitat is also sensitive to changing temperature and precipitation. Length of growing season will influence plant species composition and thus habitat structure. Coupland (1958), in a discussion of the effects of prolonged drought on the Great Plains, cited evidence that Opuntia cacti invade grasslands during periods of drought, and that grasses in general become sparser. Tilman and El Haddi (1992) found that plant species richness in grasslands fell an average of 37% following drought, with no significant recovery during two more normal years following. Such changes clearly influence habitat structure and seed availability for grassland animals.

Habitat structure also is vulnerable to the effects of UVB radiation, as plant species respond differently to increased levels. If, as in the Rocky Mountains, the leaf epidermis of grasses and woody dicots is better able to attenuate UVB than is the epidermis of herbaceous dicots (Day et al., 1992), considerable changes in upland habitat could ensue. For example, the woody subshrub buckbrush (Symphoricarpos occidentalis) tends to invade upland habitat to the detriment of the mixed-grass prairie community. Likewise, even unbroken prairie sod is subject to invasion by the exotic bluegrass (Poa pratensis). Should species such as these gain further competitive advantage, the entire prairie ecosystem, already compromised by cultivation and livestock grazing, may be at risk. Research is needed to assess the relative vulnerabilities of native and introduced grasses, and woody and herbaceous dicots, to increased UVB, as well as interactive effects of UVB with temperature, precipitation and CO₂.

Although fewer data exist for prairie wetlands than for upland habitat, models suggest that wetlands and uplands differ in their relative sensitivities to changes in temperature and precipitation. Two models illustrate these differences. Poiani and Johnson's (1993) model for a semipermanent wetland, discussed above, indicated that wetland size, depth and vegetation characteristics were more sensitive to increases in temperature than to either increases or decreases in precipitation. In contrast, Hunt et al.'s (1991) simulation model for temperate grasslands revealed that primary productivity was more responsive to precipitation than to temperature and that changes in productivity produced by changes in moisture continued up the food chain. Too few data exist for prairie wetlands to make similar comparisons of sensitivities to the effects of other greenhouse gases.

Implications of the Buildup of Greenhouse Gases for Birds in the Northern Great Plains:
Food Availability

All geese and most ducks feed to some extent on submerged and/or emergent aquatic vegetation. The fertilization effect of increased CO₂, especially in emergent vegetation, may be offset by shrinking wetland habitat and UVB-induced damage. If increased carbon storage dilutes nitrogen concentrations in foliage as well, the stage is set for a potential rapid collapse of the system, as birds increase their consumption of vegetation already in decline as a result of degraded and contracting habitat. Damage done by geese to vegetation is long-lasting, and can change plant communities (Kerbes et al., 1990), even without the effects of climatic change.

If vegetation responds to increasing CO₂ with diluted nitrogen concentrations in leaves, nutritional quality of the vegetation will decline, which together with the increased structural and chemical defenses caused by enhanced UVB, may alter the abundance or change the timing of population peaks of herbivorous invertebrates. Added to these effects of food supply for herbivorous invertebrates will be increased accumulation of degree-days, potentially advancing development of early stages so that emergence is no longer synchronized with food availability. Reproductive output of single-brooded shorebirds will be more susceptible to changes in peak invertebrate abundance than will that of multiple-brooded species such as most sparrows, thrushes and finches.

Ultraviolet radiation may lower densities of aquatic invertebrates by impairing development, especially in those with little pigmentation. Shallow, high pH waters typical of prairie wetlands promise little refuge from UVB. In addition, changes in climate will change both quality and quantity of habitat available for invertebrates.

Amphibians, important components of wetland food chains, are at risk to both increased UVB and higher temperatures. Larval tiger salamanders (Ambystoma tigrinum), for example, are a primary food of American white pelicans (Pelicanus erythrorhynchos) in North Dakota (Lingle, 1977). The eggs of the salamanders, deposited in masses near the water's surface in emergent vegetation, are likely to experience higher levels of UVB, the effects of which are as yet unknown. In addition, earlier drying of wetlands will exert selective pressure toward more rapid metamorphosis, limiting the size of the larvae and the amount of time they are available to pelicans.

Implications of the Buildup of Greenhouse Gases for Birds in the Northern Great Plains:
Susceptibility to Predators and Brood Parasites

Changes in habitat described above, including thinning of vegetation and increased interspersion of grassland with shrubs, may have an effect on the vulnerability of nesting birds to predation and brood parasitism. Several studies have demonstrated higher nest success for those nests located in large blocks of dense upland vegetation than for nests in more disturbed prairie habitat or cropland (Duebbert and Kantrud, 1974; Duebbert and Lokemoen, 1976; Higgins, 1977). Much of the decreased nest success on cultivated and disturbed land was attributed to mammalian predators in these studies.

Islands, which represent refuges from predators, are important nesting habitat for pelicans, cormorants, gulls, terns and some species of waterfowl (Lokemoen and Woodward, 1992). As water levels decline, formerly isolated nesting areas become accessible to red foxes (Vulpes vulpes), striped skunks (Mephitis mephitis), raccoons (Procyon lotor) and other terrestrial predators.

There is little evidence that predator populations in the northern Great Plains are limited by prey. Most changes in predator (Sargeant et al., 1993) and brood parasite (Mayfield, 1965) distribution and abundance since the arrival of European colonists can be related to human settlement and agricultural practices. At least one model has suggested that peak agricultural productivity will shift northward as the climate warms (Blasing and Solomon, 1983). To the extent that changes in climate influence human distribution across the Great Plains, predator and brood parasite distribution and abundance may also be affected, but the net outcome for avian prey and host populations is uncertain. Passerines in this region have a long evolutionary history with brown-headed cowbirds (Molothrus ater; Mayfield, 1965), making significant changes in the association less likely than in the newly colonized fragmented forests where brood parasites are proving so devastating to forest birds.

Ultimately, the effect predators will exert on birds as climate changes will depend on the availability of suitable habitat for the predators' own reproduction; on the abundance of alternative food such as rodents, insects, and the refuse of human habitation; and on the degree to which birds and their eggs are sought out in preference to other food sources. In addition, the disequilibrium of ecosystems associated with rapidly changing environmental conditions (Davis, 1984) may bring together potential predators and prey having no previous evolutionary contact, with unpredictable results.

Implications of the Buildup of Greenhouse Gases for Birds in the Northern Great Plains:
Susceptibility to Disease

If UVB exposure results in immunosuppression in birds as it does in fish (Fabacher et al., in press) and mammals (Daynes, 1990), birds dependent on wetlands may be more susceptible to such diseases as avian cholera and botulism, which already claim tens of thousands of birds each year. Earlier drying, combined with fewer numbers of wetlands, may concentrate birds into ever smaller areas, allowing rapid transmission of disease through populations.

Most birds depend exclusively on visual acuity to locate prey. If UVB affects avian vision over time as it does mammalian (Taylor, 1990; Mayer, 1992), not only will longevity decrease, but breeding success may be compromised because older, more experienced birds tend to have higher nest success than younger individuals. Longer-lived species such as raptors would be especially vulnerable.

Increased intensity of summer storms will have direct physiological effects on birds in the northern Great Plains. Summer storms, especially those accompanied by large hail, accounted for 7.4% of nonhunting waterfowl mortality across North America, according to one survey (Stout and Cornwell, 1976). Indeed, many species of birds are vulnerable to storms of the intensity currently seen in the Great Plains (Johnson, 1979).

Prospects

Before it is possible to make detailed predictions about the consequences of global climate change in the northern Great Plains, we must understand the basic mechanics of interactions among biotic systems, climate and greenhouse gases in the region. For example, we must understand how runoff from snowmelt affects the amount of water in wetlands before we can predict how changes in seasonal precipitation patterns will affect number of wetlands. How will changes in water-use efficiency of plants alter water retention in wetlands through the growing season? Can the combined consequences of increased water-use efficiency and greater winter precipitation balance a drier summer climate? An understanding of basic mechanics is critical to making quantitative predictions.

Of equal importance is an understanding of the ways invertebrate food resources will respond to habitat changes. How do herbivorous insects find their food plants? Is plant species or growth form important? In either case, how quickly can herbivores adapt to changes in these characteristics? Will some herbivores be extirpated if habitat changes? If so, will equally palatable and "catchable" species replace them? How will phytophagous insects respond to changes in nutritional quality of their host plants? What effect will this response have on the birds that depend on vast quantities of insects to raise their young? Answers to these questions hinge directly on an understanding of optimal foraging of both herbivores and their avian predators, and on possible feedback generated by changes in plant and herbivore species composition and abundance.

Vertebrate herbivores also may have to adapt to changes in forage quality. Will herbivorous birds such as grouse be able to consume and digest enough vegetation to sustain themselves during northern winters? Will geese be able to consume enough during their breeding season to maintain adequate productivity? If so, how will the added consumption change their habitat? To answer these questions, one must understand not only the energetics of the bird, but the physiological and anatomical plasticity of the digestive tract.

Timing is of major importance to migratory birds. Most migrations are timed by the constant progression of day length; timing of peak insect abundance on the breeding grounds is tied more closely to local temperatures. Will the two remain in synchrony as temperatures warm? Might earlier insect emergence (caused by higher temperatures) be balanced by slower growth (caused by declining nutrition)? Detailed studies of insect phenology and abundance with respect to avian reproductive success are needed to answer these questions.

Finally, we must determine the direct effects of a changing environment on the physical well-being of birds. Because many aspects of the environment are changing simultaneously, responses to individual stressors are unlikely to provide a complete picture of the net environmental stress a bird is experiencing. What is needed is a measure of the metabolic cost of adapting to environmental changes. If a quantitative estimate of this cost can be made, it should be possible to determine if adaptation is possible, or if a population will necessarily become extinct in a particular location, given anticipated changes in habitat and resource availability. Physiological studies should concentrate on measurements that integrate response to a range of environmental stresses. Stress hormones and shock proteins are certainly candidates. If such an integrator could be identified, long-term monitoring programs could be put in place to provide early warnings of populations imperiled by as-yet-unidentified stressors resulting from as-yet-unimagined environmental interactions.

---

Acknowledgments

Discussions with members of Northern Prairie Wildlife Research Center's Global Climate Change Interest Group greatly improved this manuscript. Members include: B. Bowen, N. Euliss, D. Johnson, R. Koford, P. Pietz, J. Price and L. Strong. R. Koford, E. Little and J. Price reviewed the manuscript in detail. I thank H. Hunt, S. Smith, and an anonymous reviewer for comments on earlier drafts of this manuscript.

---

---

NPWRC Home | Site Map | About Us | Staff | Search | Contact | Web Help | Copyright