Study Area

Clear Lake (39°01′N, 122°46′W) is a very large (180 km²) and relatively shallow (18 m maximum depth) lake in Lake County, California. The lake comprises 3 major arms, of which the northwestern is its largest and shallowest (Fig. 1). Major wetlands associated with the lake include its largest tributary, Rodman Slough, at the northwestern end of the lake, and its outlet, Cache Creek, which meanders through Anderson Marsh at the southeastern end of the lake (Fig. 1). The lake and its associated wetlands are bordered by tule beds (Schoenoplectus acutus and Schoenoplectus tabernaemontani) and riparian forest, rocks, gravel beaches, mud, and extensive urban development. The riparian forest is dominated by willows (Salix spp.), valley oak (Quercus lobata), Fremont cottonwood (Populus fremontii), and California sycamore (Platanus racemosa). The lake is severely impaired by a long history of human activities and is now extremely eutrophic, especially in areas with limited circulation (Suchanek et al. 2003).

View Figure

• Download as Powerpoint
• Download Figure

Methods

During the months of April through September from April 2010 to May 2018, we opportunistically recorded observations of turtles incidental to a research project on the breeding biology of the western grebe (Aechmophorus occidentalis) and Clark's grebe (Aechmophorus clarkii) in areas where they were most likely to nest. Once or twice during most weeks, averaging 21 trips/yr (range, 15–32) during 2010–2017 plus 3 trips in April and May 2018, we searched for Aechmophorus spp. nests from a canoe (most days, 5–25 km/d), motorboat (1–3 times/yr), vehicle, or by foot, and occasionally encountered turtles basking on emergent woody vegetation or floating Aechmophorus spp. nests. We attempted to identify each turtle encountered by scrutinizing it with 8× binoculars or by taking photographs through handheld cameras with a telephoto lens of up to 400 mm. Additional photos of turtles basking on floating Aechmophorus spp. nests constructed of macrophytic vegetation were obtained by motion-activated cameras (Bushnell Trophy Cam Bone Collector RTAP Night Vision and Bushnell Trophy Cam HD Aggressor No Glow) attached to metal stakes and focused on grebe nests for 11,309 hrs during 2014–2017.

We identified T. s. elegans by its distinctive yellow and red stripes on the sides of the head, yellow stripes on the carapace, or yellow stripes on the limbs and tail, which are absent in E. marmorata, or by the more jagged rear margin of the carapace, which is smoother in E. marmorata (Ernst and Lovich 2009; Thomson et al. 2016). We used the proportion of individuals belonging to each species as a measure of relative abundance. We included all turtles observed on different dates at the same locality, even though it is likely that some individuals were recounted during subsequent trips. We did this to increase the accuracy of our estimates of relative abundance for each locality, which requires as large a sample size as possible.

To measure the degree of urbanization of each locality, we used Google Earth (www.google.com/earth) to estimate the number of human homes (commercial buildings excluded) within 1 km of the center of observations at each locality. Our data for the number of homes should be regarded as estimates rather than exact counts because of the difficulty of distinguishing whether some buildings represented a commercial building or a residence, whether some buildings represented 1 or multiple residences, or whether multiple buildings and trailers adjacent to each other represented 1 or multiple residences.

Because of repeated sampling resulting in a lack of independent observations, inferential statistical tests were not used to test for differences in the proportions of species in different areas of the lake or in the use of Aechmophorus spp. nests vs. emergent vegetation, or to test for correlations between relative abundance of each species with the number of human homes. However, inferential statistical tests were used on a smaller subset of the data, based on the highest daily count of identified turtles at each site in order to exclude repeated (nonindependent) samples. A 2-sample χ² test (Zar 2010) was used to test for differences in the proportions of each species in the 3 major arms of the lake and a Pearson correlation coefficient (r statistic; Zar 2010) was used to test for a correlation between the proportion of turtles that were T. s. elegans and the number of homes, for sites with a minimum of 4 identified turtles during 1 or more daily counts.

Results

We identified 351 turtles and approximately 15% of all turtles observed were too distant or observed too briefly to identify. The turtles were relatively scarce except in a few areas with many basking sites in the channels of tributaries (Rodman Slough and Lakeside County Park), man-made canals (Clearlake Oaks), and the outlet of the lake (Anderson Marsh). Our highest counts were in the spring months of April and May, when basking may have been more prolonged than during the hotter summer months. Of the turtles identified, 12 (3%) were identified by photos from motion-activated cameras, 2 (< 1%) were dead (1 in the water, 1 on the road), and the remainder (96%) were directly observed. We searched for “melanistic” individuals of T. s. elegans with faded head and appendage stripes, which comprise a small proportion of T. s. elegans populations (Ernst and Lovich 2009), but confirmed only 1 individual.

Of the turtles identified, E. marmorata comprised 55% and T. s. elegans comprised 44% of the lake's basking turtle assemblage (Table 1). Both species of turtles were recorded in all 3 arms of Clear Lake and in associated wetland areas at each end of the lake (Fig. 1; Table 1). Based on data from all counts combined (Table 1), the proportion of T. s. elegans was higher than that of E. marmorata in the eastern arm of the lake (59% vs. 41%, n = 135), but not in the northwestern arm (40% vs. 60%, n = 181) or, especially, the southeastern arm (10% vs. 90%, n = 31). Based on the highest daily count for each site (Table 1), the ratios of the 2 species in the 3 arms of the lake differed significantly (χ²₂ = 27.4, p < 0.001). There was a positive but nonsignificant correlation between the proportion of turtles that were T. s. elegans and the number of homes for sites with a minimum of 4 identified turtles during one or more daily counts (r = 0.57, n = 5 sites, p = 0.32).

Table 1. Localities (numbers correspond to Fig. 1), coordinates, approximate number of human homes, and numbers of Trachemys scripta elegans and Emys marmorata observed at Clear Lake, Lake County, California, during 2010–2018. Localities 1–7 are in the northwestern arm of the lake, 8–9 in the southeastern arm, and 10 in the eastern arm.

View Fullscreen

One painted turtle (Chrysemys picta) with a carapace approximately 15 cm long was photographed at Lakeside County Park on 30 April 2017 (Herpetological Education and Research Project record 300004 at www.naherp.com). This species represented only 0.3% of the lake's basking turtle population.

In areas where turtles were observed basking on Aechmophorus spp. nests (Rodman Slough, north end of Clear Lake, Manning Creek, and Adobe Creek), the proportion of species basking on Aechmophorus spp. nests (59% for T. s. elegans vs. 41% for E. marmorata; n = 17) was similar to the proportion of species basking on emergent woody vegetation (55% for T. s. elegans vs. 45% for E. marmorata; n = 92).

We did not observe any nesting activities or record body size classes of the turtles, but we observed individuals estimated to be as small as 12 cm of both species. No hatchlings were observed.

Discussion

Our identifications of basking turtles may have been affected by 2 factors. First, T. s. elegans is more tolerant of approaching humans (Costa 2014) and is therefore more likely to be identified. Second, some melanistic T. s. elegans lacking stripes on the head and appendages may have been misidentified as E. marmorata. The extent to which our data was biased by these potentially counterbalancing variables is unknown.

Our observations reveal that T. s. elegans occurs in all 3 arms of Clear Lake. Differences in the number of turtles at each site resulted from some sites being surveyed more frequently than others and the turtles being easier to observe in areas with more basking sites. Geographic differences in the proportions of the species appear to be unrelated to the degree of urbanization. The proportion of T. s. elegans was highest in the eastern arm of the lake, where an extremely high level of urbanization occurs, and lowest at Anderson Marsh in the southeastern arm of the lake, where the density of human homes is very low, but the density of human homes was also very low at Rodman Slough, where T. s. elegans occurred at a relatively high density. Geographic differences in the proportions of the species may be due to differences in the timing or frequency of T. s. elegans introductions in each arm of the lake. The much lower proportion of T. s. elegans in the southeastern arm of the lake may be due to later or fewer introductions than in the other arms of the lake.

Although our data are limited, neither species appears to be more predisposed for basking on Aechmophorus spp. nests. Lambert et al. (2013) demonstrated that the 2 species differ in their use of basking substrates, with E. marmorata preferring sites with steeper midslopes and slopes at the water's edge, shallower water depths at the water's edge and near the edge, less disturbance, and concrete substrates.

Our data demonstrate that both species of turtles have coexisted at Clear Lake for at least 50 yrs since T. s. elegans was introduced, suggesting that E. marmorata may be relatively resilient to the presence of invasive T. s. elegans. It is unknown whether the population density of E. marmorata has declined owing to competition with T. s. elegans or if it has remained stable while the population density of T. s. elegans has increased. It is even possible that the population of E. marmorata has increased despite the presence of T. s. elegans. Because E. marmorata can thrive in highly eutrophic ecosystems, including wastewater treatment plants (Germano and Bury 2001; Germano 2010), it may be benefitting from an increased food supply due to the relatively rapid cultural eutrophication of Clear Lake in recent decades (Suchanek et al. 2003).

The large number of T. s. elegans at Clear Lake provides evidence that a breeding population has been established, although reproduction remains unconfirmed. The lack of observed hatchlings for both species suggests that their reproductive success may be low. The native American crow (Corvus brachyrhynchos) and northern raccoon (Procyon lotor), both of which are human commensals with increasing populations (Marzluff et al. 2001; Prange et al. 2003), are the primary predators of Aechmophorus spp. eggs in nests along the margins of Clear Lake (Hayes et al. 2018) and they may be major predators of turtle eggs and hatchlings as well. Several potential predators of turtle hatchlings have been introduced at Clear Lake and are now invasive, including several large species of fishes of the family Centrarchidae (Suchanek et al. 2003; Thompson et al. 2013) and the American bullfrog (Lithobates catesbeianus; Bury and Luckenbach 1976). The impact of these species on the reproductive success of turtles in California is poorly known and merits further study (Thomson et al. 2016).

Chrysemys picta has been reported as an introduced species at scattered localities in California (Banta and Morafka 1966; Bury and Luckenbach 1976; Spinks et al. 2003; Lemm 2006). Our observation provides the first record for Lake County, but because only 1 was observed, it is highly unlikely that a successfully breeding population has been established at Clear Lake.

Ideally, more intensive population and demographic studies should be conducted on Clear Lake to document population densities and trends of all turtle species. At the very least, simple monitoring with the use of our methods should be conducted in the future to document long-term trends in the proportions of the species.