Amphibians may have relatively great vulnerability to environmental perturbation due to several factors. Each species has a complex of specific habitat and dietary needs related to complete metamorphosis, including the transitional period between larva and adult. Problems with pollutants may be exacerbated because amphibians’ permeable skin makes transdermal movement of toxins easy. Eggs of many species require pure, well-oxygenated water and are susceptible to siltation, pollution, and predation. Predators may be especially damaging because amphibian larvae often are awkward and lack effective defensive or escape mechanisms; thus amphibians may require breeding habitat that affords temporal or spatial isolation from predators. A key problem arises in that the aquarium trade includes several species of introduced, exotic fish, amphibians, and reptiles that may be released into the wild, inadvertently dispersing parasites and pathogens that affect native amphibians. In light of their apparent great sensitivity, amphibians may serve as an early warning system for environmental degradation.

Modern industrial farming systems lead to emission of greenhouse gases (such as NOx resulting from fertilizer application), use of ozone-depleting technology (e.g. methyl bromide fumigation), destruction of native vegetation and general simplification of the landscape (e.g. removal of dead wood, rocks, rodents), changed hydrology (irrigation and drainage needs must be met), and use of pesticides and nitrate fertilizers. All of these actions may adversely affect amphibians, and some may be allayed or mitigated.

In this two-part article, we summarize the hypothesized threats to amphibians; highlight important findings from the international scientific literature; summarize the situation for a few key species with different primary threats and ranges in California: red-legged frog (Rana aurora draytonii), California tiger salamander (Ambystoma californiense), and mountain yellow-legged frog (Rana muscosa); and recommend management options for farmers, to benefit amphibians and other wildlife.

Global Climatic Change and Ultraviolet Radiation
Because many populations of amphibians disappeared or declined essentially synchronously from regions of North America, Central America, and Australia, globally distributed potential causes have been sought, including climatic and other atmospheric issues (Blaustein et al. 2001, Beebee 2002).

In Costa Rica, climate change has been convincingly implicated in the well-documented declines of many species, apparently due to changes in precipitation and temperature. In temperate regions amphibian breeding has been shifting to earlier dates in response to climate warming (Gibbs and Breisch 2001, Corn and Muths 2002). The distributions of mobile species such as birds are predicted to change in response to shifting climates, however, many less-mobile species such as amphibians may disappear completely if the climate changes dramatically.

Numerous experiments have shown that exposure to ultraviolet radiation results in elevated mortality and malformations in developing embyros and larvae of some species (Tietge et al. 2001; Ankley et al. 2002). Because the amount of ultraviolet radiation reaching the Earth’s surface has increased in recent years, there is suspicion that ultraviolet radiation may have caused or contributed to amphibian declines.

Metapopulation Dynamics and Landscape Ecology
Although many salamander species live their entire lives on land, our most familiar amphibians begin life as aquatic larvae, and then metamorphose into adults. Adults return to aquatic habitats to breed, but many of these species live the majority of their lives distant from water. As a result, modifications to either aquatic or terrestrial habitats can negatively impact amphibian populations. Ideal landscapes for these amphibians contain abundant aquatic habitats, situated close enough to one another that animals can move among them, and nested within upland habitats suitable for adult survival. Amphibians may require distinct habitats for breeding, adult feeding, hibernation, and aestivation (summer dormancy).

A metapopulation has been defined as a set of local populations connected by migrating individuals and with varying degrees of isolation from one another. Local populations, besides being depleted by emigrants and augmented by immigrants, may be subject to extinction through stochastic (probabilistic) events or deterministic events, resulting from habitat degradation. Locales where extinctions occur can also be re-colonized, if habitat remains or becomes suitable. Marsh and Trenham (2001) wrote that in systems where stochastically driven extinctions prevail, landscape-ecological considerations are essential to conservation planning (e.g. patches and corridors must be considered to allow re-colonization). In cases where deterministic extinctions prevail, the emphasis should be on local habitat conservation. An interpretive difficulty lies in the distinction between “landscape” and “local” scales, because dispersal, homing, and colonizing abilities and tendencies vary among species and are still poorly understood.

Marsh and Trenham (2001) emphasized the importance of terrestrial habitat to the typical amphibian life cycle, and cautioned that plans that focus strictly on maintaining breeding habitat (e.g. ponds and wetlands) will probably fail to conserve populations. In related work, Pope et al. (2000), studying northern leopard frog (Rana pipiens) in Ontario, Canada, found that inclusion of non-breeding summer habitat (meadows) in a statistical model was essential in discerning metapopulation structure; data for frog occurrence in breeding habitat (ponds) were insufficient by themselves.

Semlitsch (2000) presented a conceptual overview of principles for managing populations and communities of aquatic-breeding amphibians, emphasizing (1) the number or density of individuals dispersing from individual wetlands; (2) the diversity of wetlands with regard to hydroperiod or timing and duration of inundation; and (3) the probability of dispersal among adjacent wetlands or the rescue and re-colonization of local populations. A special concern was the loss of small, ephemeral wetlands (<4.0 ha) because they harbor high abundances and species diversities; loss of these wetlands can interfere with metapopulation dynamics of re-colonization and is expected to lead to more local extinctions. Wetlands can be impaired as amphibian habitat through changes in hydrologic cycles. For example, increased predation may occur if hydroperiod (time of flooding) is lengthened or fish are allowed to enter through anthropogenic connections with other bodies of water. By contrast, early drying may prevent amphibians from completing metamorphosis. Natural habitats can be fragmented through logging, farming, road-building, canal construction, and urbanization.

In Minnesota, Lehtinen et al. (1999) evaluated amphibian assemblages in wetlands occurring in tallgrass prairie and northern hardwood forest, using a geographic information system with land-use variables quantified at scales of 500, 1000, and 2500 m. Ten species of amphibians were found; the three commonest were northern leopard frog (Rana pipiens), tiger salamander (Ambystoma tigrinum), and American toad (Bufo americanus). Increasing wetland isolation and road density were correlated with lower amphibian species richness at all scales in both ecoregions. Proportion of urban land use showed a negative relationship with species richness at all scales in the hardwood forest ecoregion.

Knutson et al. (1999) found that the presence of urban land was negatively correlated with abundance and species richness of frogs and toads. Positive associations were found for upland and wetland forests and emergent wetlands. Edge and diverse habitats that included wetlands showed positive correlations. For Iowa, length of wetland/forest edge showed the greatest positive correlation, whereas the presence of urban land showed the greatest negative. For Wisconsin, the two most significant associations with relative abundance were positive correlations with forest area and agricultural area. Frogs and toads were positively associated with agriculture in Wisconsin but not in Iowa.

In southern California, Griffin and Case (2001) studied arroyo southwestern toad (Bufo microscaphus californicus) attached by external belts to radiotransmitters to assess seasonal preferences by males vs. females for areas of differing substrate, land form, land use, and vegetation types. In general, the toads foraged in and dispersed through a wide range of habitat types found in southern California. The observed patterns showed that female toads preferred terrace and channel habitats over campground, agricultural, or upland habitats, and that male toads occupied channel habitats during breeding season. Toads burrowed preferentially in sand as opposed to other soil types; above-ground activity was not limited by substrate type. Males occurred on agricultural lands increasingly after the breeding season. This toad appears to avoid dense, tall vegetation for burrowing.

In Connecticut, Gibbs (1998) used drift fences fitted with pitfall traps to explore the concept of filters and conduits governing movement by amphibians. This work suggests that dispersal by some but not all frogs and salamanders is influenced by forest borders and streambeds and that edges between forests and roads are less permeable than forest/open land edges.

As reviewed by Trombulak and Frissell (2000), roads have a range of influences on terrestrial and aquatic life. Fahrig et al. (1995) found that road traffic was an important mortality factor for amphibians in Ontario, Canada, and that intensity of traffic on adjoining roads was correlated with reduced density and reduced chorus intensity of frogs and toads. Ashley and Robinson (1996) in Ontario, Canada, near Lake Erie, found that road kill of amphibians was significantly associated with adjacent roadside vegetation, that dispersing young of the year were the principal victims, and that northern leopard frog (Rana pipiens) was the amphibian most commonly killed. Carr and Fahrig (2001) in Ontario, Canada, found that populations of a more vagile (=mobile) species, northern leopard frog, was more adversely affected by road traffic than were populations of the less vagile green frog (Rana clamitans). Both species move among habitats during the year, but leopard frog requires three types of habitat, whereas green frog makes use of two.

Gibbs (1998), working in southern Connecticut, correlated traits of various amphibians with their apparent tolerance of habitat fragmentation. He suggested that species with low density, coupled with habitat restrictedness and high mobility could predispose for local extinction due to habitat fragmentation.

In a coastal watershed of Santa Cruz County, Calif., Bulger et al. (2003) documented the seasonal dispersal behavior of radio-tagged adult California red-legged frog, showing that adults traveled to and from breeding sites by night and in approximately straight lines, irrespective of topographic or vegetational features. This study occurred mainly within a matrix of diverse native vegetation and did not formally assess the role of roads or other anthropogenic disturbances.

Changed Hydrology
As indicated by Blaustein and Kiesecker (2002), in the arid West extensive seasonal wetlands have been drained or replaced by small permanent ponds. In California, many extensive wetlands have been converted to farmlands, sometimes with associated small catchments. This pattern may not only remove sensitive amphibian species outright, but it may also exacerbate the interference by non-native invaders that fare better and have greater negative impacts on native species in the new environments.

California owes its agricultural productivity and competitiveness in large part to irrigation technology that enables farmers to put water where they want it, when they want it. In turn, irrigated agriculture in California is enabled by a vast and intricate system of dams, reservoirs, canals, small catchments, ditches, and irrigation systems. This infrastructure and its management have greatly affected temporal and spatial patterns of water availability to natural communities in much of California (Bugg et al. 1998). Water districts alter flow patterns in all the major rivers flowing into the Great Central Valley except the Cosumnes River. The presence and management of dams affect not only flow but also sediment size and texture, and impact stream dynamics both upstream and downstream. Recharge of groundwater in flood plains can also be impaired. All this can affect the viability of native plants and animals. Reduced groundwater recharge as a result of stream channel incision and reduced flooding has been invoked as a partial explanation for the decline of valley oak (Quercus lobata). Also, dam releases cause unnaturally high flows in late sping and early summer, washing away eggs and larvae of foothill yellow-legged frog (Rana boylii), a species that reproduces in channels with low flow (Ashton et al., year unspecified).

Dam removal is becoming an accepted practice in river management; although this can reverse some of the trends mentioned above, it carries its own set of potential problems (Stanley and Doyle 2003).

In some cases, artificial wetlands and catchments that enable or result from agricultural water use may seem to mimic natural features of California, yet there are important differences. For example, due to patterns of impoundment, re-channeling, and release, the seasonal abundance of water in many areas may differ profoundly from natural patterns: Many artificial wetlands and ponds contain water during seasons when it is lacking in nearby, more natural, systems. The converse may also be true. Morever, wide and rapid fluctuation in water levels, caused by drawdown and refill of catchments, may inhibit the development of native emergent and riparian vegetation, both of which are important in the ecology of amphibians.

Adams (2000), in the Puget Lowlands of Washington state, used enclosure cages to test survival of native northern red-legged frog (Rana aurora ssp. aurora) or Pacific tree frog in permanent vs. temporary ponds, in combination with bullfrog (Rana catesbeiana) larvae or bluegill sunfish (Lepomis macrochirus). Native frog larvae survived better in temporary ponds. There was no survival when larvae were caged with sunfish. Caging with bullfrog larvae did not affect Pacific tree frog (Hyla regilla) larval survival, and showed mixed results with northern red-legged frog larvae. Pond permanence and introduced species led to reduced survival of native amphibians, but neither explained all the variability in frog larval survival.

Hazell et al. (2003) compared historical and current conditions of stream morphology and dynamics in New South Wales, Australia, to determine changes that have occurred since European colonization, and how these are likely to influence survival of native frogs. The earliest written accounts by European settlers in New South Wales show that streams appeared as chains of permanent ponds with short, shallow, grassy connecting channels that flowed seasonally. Winter flooding created adjoining ephemeral wetlands. The permanent ponds and ephemeral wetlands would have supported differing complexes of frogs, with differing times required for larval development. Flooding also provided wet soil and associated vegetational cover, which are important to newly metamorphosed frogs. Under management by the European colonists, stream management changed: channels are now typically incised, flooding restricted, and vegetation altered. Thus, conditions now differ profoundly from those under which many native frogs are presumed to have evolved.

Kolozsvary and Swihart (1999) in the midwestern U.S. found that the greatest amphibian species diversity was correlated with the presence of wetlands of intermediate permanence.

Tailed frog (Ascaphus truei) requires shaded, rocky headwaters for breeding. Increased water temperatures, siltation, and in-channel woody debris associated with clear-cut logging greatly diminish reproduction, according to a study by Dupuis and Steventon (1999), who recommend that creeks with favorable substrates and temperatures be protected by old-growth buffers.

Mineral Enrichment of Water
Nitrogen enrichment of freshwater aquatic systems can occur as the result of land-use patterns and farming systems (Honisch et al. 2002). Data from Marco et al. (1999) suggest that larval amphibians may be especially sensitive to such perturbations, and that adverse effects occur below the thresholds recommended by the U.S. Environmental Protection Agency. This should be considered in recommending practices for agriculture in the region.

In Pennsylvania, Laposata and Dunson (2000) found that, in contrast to natural ponds, ponds with 14-year histories of addition of treated wastewater had higher median conductance, pH, and concentrations of Na, K, Ca, Mg, and N-NO3. Such ponds developed mats of duckweed (Lemna spp.) and, for three amphibian species, had lower densities of egg masses, hatching success, and larval survival.

De Solla et al. (2002), working in the Sumas Prairie of the Lower Fraser Valley of British Columbia, Canada, implicated agricultural runoff and correlated increased ammonia, total phosphate, and biological oxygen demand with decreased hatching of eggs of northwestern salamander (Ambystoma gracile) and northern red-legged frog (Rana aurora ssp. aurora).

[Note: In the next issue of Sustainable Agriculture, this two-part article will continue with discussion of pesticides, associated organisms (e.g. parasites, pathogens, plants, predators, and competitors), and on-farm modifications that may favor native amphibians. The list of references will also be provided.]