Ecology of a Remnant Population of Oregon Spotted Frogs (Rana Pretiosa) in Thurston County, Washington

The Oregon spotted frog (Rana pretiosa) is listed as a State Endangered species under Washington State law, and is a candidate for listing under the federal Endangered Species Act. The first Oregon spotted frog verified in Washington since 1968 was captured in 1990 along Dempsey Creek, a tributary of the Black River in Thurston County, Washington (McAllister et al. 1993). Only three additional populations of this species are known in Washington. Suggested reasons for population declines include altered hydrology, predation by exotic fish and amphibians, and physiologic effects from changes in water chemistry and ultraviolet radiation (Hayes et al. 1997).

Until recently, virtually all that was known about Oregon spotted frogs was from a study conducted in British Columbia along the Little Campbell River (Licht 1974). In 1996, we initiated a study of Oregon spotted frogs at Dempsey Creek to better understand this species ecology. The Dempsey Creek population was ideal for study because of the mutual interest of the landowner, Port Blakely Tree Farms, Ltd., in developing a better understanding of site-specific characteristics of the frog population. Furthermore, because the study area was grazed throughout the year by about 20 cows, it provided an opportunity to study the effects of grazing on frog habitat. We identified four topics of study: refinement of marking and capture techniques; population characteristics; home range and habitat use; and surveys for new frog populations on tributaries near Dempsey Creek.

Our initial study effort focused on refinement of capture and marking techniques for this species since Oregon spotted frogs are a highly aquatic ranid and difficult to capture (Licht 1986a), and marking techniques for amphibian research are not as well developed compared to other terrestrial wildlife (Ferner 1979). Intensive surveys for frogs revealed that capture success by hand or nets was enhanced by taking slow, infrequent steps in search pools, and scanning the surface of the water for floating or resting frogs before proceeding. Searches also were more productive when temperatures were >4°C (40°F) and under clear skies when frogs surfaced to bask and feed. We captured 568 different frogs at Dempsey Creek from October, 1996 through June, 1999. Thirtysix frogs were marked with numbered fingerling (knee) tags, 462 with subcutaneous passive integrated transponders (PIT-tags), and 60 with VHF transmitters attached to nylon waistbelts. We detected no difference in survival or body weight changes among frogs marked with the different methods (P $0.315). However, knee-tags caused serious lacerations on four of 12 (33%) frogs that were recaptured, and this marking method was discontinued after March, 1997. PIT-tags were very effective for long-term marking of frogs, with no observed side-effects. Frogs were monitored via telemetry an average of 57 days (SE = 10), but we were only able to redeploy transmitters on 17 frogs (28%), largely because of transmitter belt slippage (35%) or our inability to recapture frogs before transmitters expired prematurely (21%). Nonetheless, telemetry provided unbiased location information that could not have been gleaned from marking alone.

The second study, Oregon spotted frog population characteristics, was designed to establish population size from egg mass counts, mark-recapture trials, and age class data. During six seasons between 1997-99, we conducted 48 mark-recapture surveys totaling 372 hours. Markrecapture analyses for closed population models that assumed unequal capture probabilities between survey periods, yielded seasonal population estimates of between 101 to 674 males in the winters from 1997 and 1999, and between 50 and 200 females in fall and spring. Because most mark-recapture estimates were less than the known marked population (e.g., 341 animals), this method underestimated the population. This probably resulted from our failure to sample all frog habitats in the study area, even though our searches were conducted as completely as possible, so we tended to resample a subpopulation with increased numbers of marked frogs. In another mark-recapture analysis, we combined data by year to estimate annual population size using an open model. This analysis provided a more realistic population estimate (514), with reasonable precision (SE = 129), and a realistic annual survival rate (37%). We suggest application of the mark-recapture method for this species could be improved with use of capture methods that eliminate search bias in dense habitats (e.g., crayfish or funnel traps). The second method we used to estimate population size was by counting egg masses, and projecting the number of adults by assuming one egg mass was laid per female and that there was a 1:1 sex ratio of adults. We also estimated the subadult (<2 yr-old) population, projected from the ratio of subadults to adults that were captured. We differentiated sizes of suspected adults from subadults by measuring the snout-vent length of 27 frogs captured in amplexus. Total population estimates averaged 496 animals and ranged between 410 and 555 individuals for the period 1996-99. We believe estimates from egg mass counts could be improved by determining the annual percent of adult females that do not oviposit. The third method used to estimate the frog population in 1999 combined capture summary data and knowledge of a total lack of recruitment in 1998. This resulted in the highest population estimate of 853 adult frogs. This was a reasonable upper estimate of the Dempsey Creek population, and we conclude, based on the three methods we used, that there were between 500 and 850 adult frogs in the population during the study.

We assessed other population characteristics of Oregon spotted frogs including minimal actual survival rates, annual changes in age cohorts, the association of spring rainfall to juvenile recruitment, individual growth rates, and sources of mortality. Minimal adult survival based on numbers of frogs recaptured from the marked population was 28% in 1997, and 16% in 1998. Proportions of adults in the population based on annual egg mass counts varied from 53% to 67%, and were greatly affected by larval survival from the previous spring. Larval survival, in turn, was related to spring rainfall. Rates at which we encountered juvenile frogs were higher (>1 frog more/hr), though not statistically different (P = 0.146), in years of high rainfall in April and May (12.7 cm/mo) compared to years of low spring rainfall (4.4 cm/mo). However, we believe these differences would be magnified in true drought years, and that drought conditions in April and May over 2 to 3 years could seriously affect population numbers. Based on recaptures of 88 frogs, we estimated that adult males grew an average of 2.2 mm/yr, and females grew an average of 6.2 mm/yr. Predictive models suggested males ceased to grow at 57.2 mm, and females at 75.9 mm. Growth rates and predicted maximum adult sizes were between those identified for Oregon spotted frogs in British Columbia, and for Columbia spotted frogs at higher elevations and colder climates in Wyoming. Observed mortality of Oregon spotted frogs was rare at Dempsey Creek. Of 11 mortalities identified, four (36%) were attributed to predation by endemic birds and mammals, four (36%) to unknown causes, two (18%) to disease, and one (9%) to humans. No bullfrogs (Rana catesbeiana), potential predators of Oregon spotted frogs, were observed at Dempsey Creek.

Our third study used radio-telemetry to investigate the year-long and seasonal distributions of the frog population, to determine the movement patterns and ranges of individuals, and to determine frog selection of habitats. The telemetered population of 60 frogs occupied a range that was a mosaic of 38.5 acres of wetlands and 32.5 acres of upland pasture. Spatial use of the range was closely related to seasonal behaviors and changing surface water conditions. During the breeding season (February through May), frogs occupied >50% of the population range, and selected shallow, backwater pools for oviposition sites. In the dry season (June through August), shallow pools disappeared and frogs were forced to move up to several hundred meters to deeper remnant pools. Frogs significantly reduced movements, and occupied an average of <10% of their home ranges. During the wet season (September through January) frogs again moved exceptional distances back up drainages and reoccupied the breeding area and peripheral shallow areas. Three frogs, that wintered in shallow water that inundated dense vegetation, buried themselves at the base of plants during the coldest period. We found that different life stages of Oregon spotted frogs had different critical seasons associated with aquatic needs: adequate water levels for egg and tadpole survival were most important in the breeding season; deep pools were most critical for survival of juveniles and adults in the dry season; and adequate water levels over emergent vegetation were important for survival of all age classes during the wet season and coldest time of the year. Further, a topographic gradient with overall gradual relief that maintained adequate water for inter-pool movements was vital to provide for all needs during the annual cycle. We documented only one instance of upland movement by an Oregon spotted frog, and concluded that disjunct, land-locked, shallow breeding pools and deep-water pools do not provide yearround needs of Oregon spotted frogs populations.

Individual frogs exhibited two types of annual movement patterns: infrequent, long-distance movements between widely separated pools, and frequent movement between pools in closer proximity. Home ranges of four frogs averaged 5.4 ac (100% fixed kernels), and on average frogs moved a minimum of 5 m/da (SE =1) throughout the year.

Plant structure within habitat communities, particularly wetlands influenced by grazing, were key influences on frog habitat use at Dempsey Creek. For coarse-scale analysis of habitat selection we used digital imagery to estimate proportions of nine wetland habitats, and determined habitat use from 654 telemetry locations. Frogs avoided dry uplands. Frogs selected sedge (Carex obnupta, and C. utriculata)/rush (Juncus effusus) habitat during the breeding season, which was closely associated with shallow and ephemeral waters at oviposition pools, and was grazed (i.e., sedge) more intensively by cows relative to other types. Hardhack (Spiraea douglasii) was an important emergent vegetation in the hottest part of the dry season as it shaded and maintained the remnant pools. Reed canary-grass (Phalaris arundinacea) was widely used (30%), but statistically found to be avoided because of its extensive distribution relative to other types. Microhabitat analysis, <0.25 m² from frog locations, showed reed canary-grass communities were used conditionally, depending on the degree of cattle grazing. Grazing created penetrable, open habitat in reed canary-grass communities that were otherwise too dense for frog use. Frogs preferred a moderate to high degree of water surface exposure (i.e., 50% to 75% water), or conversely, a low to moderate degree of emergent vegetation (i.e., 25% to 50%). Consequently, low emergent cover (0% to 25%), potentially resulting from overgrazing, was also unsuitable for frog use. We recommend further study of cattle grazing patterns, possibly through use of radiotelemetry, to investigate the timing and intensity of grazing that are the most beneficial to Oregon spotted frogs in reed canary-grass communities. Additional habitat analysis revealed that subsurface aquatic habitat associated with frog locations was largely (89%) open water, with the remainder submergent plants above a bottom layer of herbs, detritus, and sediment. Throughout the year frogs were located in water that averaged 19.0 cm in depth (mean monthly range = 8.5 cm to 26.2 cm), with average surface temperatures of 14.7 °C (mean monthly range = 5.6 to 19.1 °C) and average subsurface temperatures of 13.4 °C (mean monthly range = 6.1 to 18.5 °C).

Our final study was a survey for Oregon spotted frogs along portions of the Black River and the associated Chehalis River drainage using volunteers and agency personnel. During 58 searches totaling 123 hours Oregon spotted frogs were located at two new locations. The first, at Stony Creek was closely associated with the Dempsey Creek population. The second, at Beaver Creek was significant since it was located 16 km from the known population. Oregon spotted frog surveys, to date, have sampled suitable habitat over much of the Chehalis River basin. A systematic attempt to identify suitable habitat within the basin should seek to identify habitat with the following features: 1) extensive (at least 20 acres) contiguous and shallow emergent palustrine wetland habitat; 2) low gradient stream course or ditch draining the wetlands; and 3) high seasonal hydrologic fluctuations such that surface water is extensive in winter and early spring, and extremely limited in late summer.