Use of Terrestrial Habitat by Western Pond Turtles, Clemmys marmorata: Implications for Management

Abstract

Despite its extensive range, the western pond turtle, Clemmys marmorata, is currently a can- didate for federal listing. Understanding its use of the landscape has become increasingly important to the development of appropriate management plans. Using radiotelemetry, we examined movements of turtles in a two-mile stretch of the Trinity River (Trinity County, California). We observed frequent and prolonged use of terrestrial habitat for both nesting and overwintering activities; the turtles travelled into upland areas as far as 500 m from the river. Males utilized terrestrial habitat in at least ten months of the year, and females were on land every month as a result of their additional terrestrial behavior while gravid. Hatchlings over- wintered in the nest. These observations suggest that the terrestrial habitat is as important as the aquatic habitat to the viability of western pond turtle populations. The implications for management are significant, considering the variety of development pressures on lands adjacent to waterways.

Full text

Proceedings: Conservation, Restoration, and Management of Tortoises and Turtles. An International Conference, pp. 352-357
held 1997 by the New York Turtle and Tortoise Society
Use of Terrestrial Habitat by Western Pond Turtles, Clemmys marmorata:
Implications for Management
DEVIN A. REESE
1
,
2
AND HARTWELL H. WELSH
1
1USDA Forest Service. PSW Redwood Science Laboratory, Arcata, CA 95521, USA
2Current Address: USAID/G/ENV, The Ronald Reagan Building, Room 308. Washington. D.C. 20523-1812
ABSTRACT: Despite its extensive range, the western pond turtle, Clemmys marmorata, is currently a can-
didate for federal listing. Understanding its use of the landscape has become increasingly important to the
development of appropriate management plans. Using radiotelemetry, we examined movements of turtles in
a two-mile stretch of the Trinity River (Trinity County, California). We observed frequent and prolonged use
of terrestrial habitat for both nesting and overwintering activities; the turtles travelled into upland areas as
far as 500 m from the river. Males utilized terrestrial habitat in at least ten months of the year, and females
were on land every month as a result of their additional terrestrial behavior while gravid. Hatchlings over-
wintered in the nest. These observations suggest that the terrestrial habitat is as important as the aquatic
habitat to the viability of western pond turtle populations. The implications for management are significant,
considering the variety of development pressures on lands adjacent to waterways.
The western pond turtle, Clemmys marmorata, is the
only extant aquatic turtle native to California. Western pond
turtles have an extensive range (western Washington to
northwest Baja California: Stebbins. 1985) and appear to fill
a variety of aquatic niches. They are found in ponds, rivers,
vernal pools, streams. ephemeral creeks. reservoirs, agricul-
tural ditches. sewage treatment ponds. and estuaries. De-
spite the fact that they are widespread habitat generalists,
western pond turtles are declining (Holland, 1992). C. mar
Figure 1. Ten-gram radio transmitter with whip antennae attached with epoxy cement. (Photo
by Douglas Welsh).
morata became a candidate for federal listing in 1991 (Fed-
eral Register, 1991). It has been given legal status as
"Threatened" in Washington State, "Sensitive" in Oregon,
and "Of Special Concern" in California.
All other members of the genus Clemmys are semi-
terrestrial. C. insculpta feeds in alder thickets and corn
fields (Kaufmann, 1992) and C. guttata aestivates in ter-
restrial habitats (Ward, 1976). C. muhlenbergii can be found
in riparian vegetation (Chase et al., 1989). Preliminary ob-
servations of western pond turtles indi-
cated that they may also have terrestrial
affinities:
• They frequently cross roads in agri-
cultural areas of California (Reese.
pers. obs.).
• Their underwater numbers decrease
appreciably in the main stem Trinity
River as winter approaches (Reese
and Welsh, 1992, unpubl. data), sug-
gesting that they probably leave the
river.
• They nest on land, a feature com-
mon to all but one known species of
aquatic turtle. Nesting sites for C.
marmorata can be as far as 400 m
from water (Storer. 1930, Holland.
1991a, 1991c; Rathbun et al.. l992 ).
Understanding the function of ter-
restrial movements could be helpful in

Figure 2. Location of study reach along the main fork of the Trinity
River, Trinity County, California.
To minimize potential differences in observed behavior
between males and females that may result from the location
of capture rather than sex, we designated males and females
found within 50 m of each other as pairs. We also required
that pairs be separated by at least 300 m to minimize the de-
gree of home range overlap, while keeping the study area to
a manageable size for hiking to turtle locations. Juvenile
turtles (<110 mm carapace length) were excluded.
Turtles were fitted with 160 MHz radios that have
approximately a 2 km base range (AVM Electronics, Liver-
more, California). The 10 g radios (4.2 cm long, 3.3 cm
wide, 1.0 cm thick) were affixed to the carapace using PC-7
overnight-drying epoxy cement (Protective Coating Com-
pany, Allentown, Pennsylvania) with the whip antennae at-
tached to the marginals (Figure 1). The batteries were esti-
mated to last approximately 12 months with a pulse rate of
85/min. The radio-equipped turtles were located weekly on
a random day to avoid errors from systematic telemetry
monitoring. (A systematic schedule could, for example, bias
the results towards particular cyclical behaviors.) After a
turtle was located its position and behavior were noted.
Turtles were disturbed only to the extent necessary to es-
tablish their locations on the ground. For terrestrial loca-
tions, this occasionally required manual searching through
leaf and needle litter. For aquatic locations, triangulation
was used to situate the turtle as precisely as possible. The
USE OF TERRESTRIAL HABITAT BY CLEMMYS MARMORATA: IMPLICATIONS FOR MANAGEMENT 353
the interpretation of range-wide trends. It may also allow us
to predict the extent and nature of terrestrial habitat use by
turtles at poorly known sites. Movements of emydid turtles
onto land for nesting are documented for numerous species
(Burger, 1975; Congdon et al., 1983; Schwarzkopf and
Brooks, 1987: Quinn and Tate. 1991). Terrestrial overwin-
tering is less well understood (Bennett et al., 1970). Other
potential triggers for land travel include movements to es-
cape unsuitable conditions (Gibbons, 1986) or movements
within a home range that includes multiple bodies of water.
The objective of this study was to investigate the role of
the terrestrial environment in the life history of the western
pond turtle. Although information exists on location of nest
sites, there are few data regarding the frequency and dura-
tion of terrestrial movements. With the advent and refine-
ment of radiotelemetry for small animals, collection of this
data became possible. Our intent was to investigate over-
land movements associated with nesting as well as identify
new terrestrial destinations, such as overwintering sites.
Study Area
The study area was a 2.8 km stretch of the main stem
Trinity River (Trinity County, California). The stretch (Fig-
ure 2) runs between Douglas City and Junction City in a
relatively unpopulated section of river. Surveys conducted
previously (Lind et al., 1992) found high densities of west-
ern pond turtles along this stretch. The surrounding land is
divided in ownership between the U.S. Bureau of Land
Management, the U.S. Forest Service. and private owners.
The dominant riparian tree canopy species are white alder
(Alnus rhombifolia) and yellow willow (Salix lasiandra).
The adjacent upland habitat is characterized as montane
hardwood-conifer and montane hardwood (Mayer and Lau-
denslayer, 1988). The former applies primarily to the north-
facing slopes, which harbor a diverse mix of hardwood and
conifer species. Conifers, including Douglas fir (Pseudo-
tsuga menziesii) and Ponderosa pine (Pinus ponderosa) form
a high canopy. The subcanopy consists of Pacific madrone
(Arbutus menziesii), tan oak (Lithocarpus densiforus), and
California black oak (Quercus kelloggii). In contrast, south-
facing slopes are dominated by hardwoods, including man-
zanita (Arctostaphylos sp.), Pacific madrone, several oak
species, and gray pine (Pinus sabiniana).
METHODS
This three-year, ongoing study of western pond turtle
movements and habitat use was initiated in May 1992. Re-
sults presented here were generated from radiotelemetry data
collected during the first year of study. Results from the re-
mainder of the study are reported in Reese (1996). Our in-
tent was to radio track 12 western pond turtles for the dura-
tion.

354 D. A. REESE AND H. H. WELSH
following data were recorded for all terrestrial locations:
shortest distance to the water, slope, aspect, canopy cover,
and habitat type.
In addition to the 12 radio-tagged individuals, seven fe-
males were equipped with short-term radios to monitor nest-
ing behavior. These females were checked every three hours
from dawn to dusk while gravid. If they remained active
after dusk, they were checked every three hours throughout
the night. Monitoring continued for at least one week after
nesting. For comparison, males were checked on a similar
schedule during two weeks of the nesting season.
Active nests were covered with mesh cages to prevent
predation of eggs or hatchlings (Holland, 1992), and this
also ensured that we could ascertain the time of emergence.
son. The winter peak represents movements of both sexes
to overwintering sites. The graph shows a high level of ter-
restrial behavior: females spent time on land during every
month of the year, whereas males spent tune on land during
all months except July and August. As described above,
these are likely to be minimum estimates of the amount of
actual terrestrial behavior in the monitored population.
Nesting
Although we observed only one female actually deposit-
ing its eggs, we were able to estimate the times the others
nested. based on their behavior and reproductive status.
Nesting occurred from mid-June to mid-July. Females were
highly terrestrial while gravid, making multiple trips (rang-
Figure 3. Clemmys marmorata partially buried beneath leaf and needle litter.
(Photo by Douglas Welsh.)
ing from 2 to 11) onto land. which were initiated as early as
one month prior to nesting. While on land each female bur-
rowed and was partially or completely concealed beneath
leaf or needle litter (Figure 3). Females remained buried
in single locations for as long as three days. occasionally
changing orientation. The amount of time spent on land was
greatest just before nesting and declined thereafter (Fig-
ure 5 ).
The female that was observed nesting had travelled at
least 31 m from the river’s edge assuming that it had taken
the shortest route. Located in a clearing surrounded by
mixed hardwood, the nest was excavated in hard-packed
silty soil on a slight. east-facing; slope only a few meters
from the riparian zone. Oviposition occurred in the evening
and the female remained by the nest overnight before re-
turning to the river. Hatching emerged from the nest the
following March (eight months later). Hatchlings did not
immediately leave the nest area after emerging, spending as
many as nine days under leaves.
Data Analysis
We calculated the proportion of turtles that
spent time on land during each month from June
1992 to June 1993. A single location on land at
least one meter from the water was considered
sufficient to be designated a terrestrial sighting.
This generated a conservative estimate of terres-
trial behavior, considering that turtles may have
made additional land trips during the week-
long intervals between checks. We included the
gravid females as well as females from the year-
round set that were not gravid or of unknown
reproductive status.
We calculated how many days gravid fe-
males spent on land during the weeks before and
after nesting. The more intensive monitoring
schedule (every 3 hours) during nesting season
allowed us to generate these estimates. Analysis
of radio-tracking data requires certain assump-
tions regarding the positions of individual turtles in the in-
terim periods between successive locations. The following
assumptions were chosen because they required the least
conjecture and thereby seemed parsimonious:
1. If a turtle was in the same medium (river, land, or pond)
at consecutive sightings. it was assumed to have re-
mained there during the interim; and
2. If a turtle changed medium between consecutive sight-
ings, the first location at the new medium was assumed
to be the first time it was there.
RESULTS
Terrestrial movements were most common during the
summer and winter (Figure 4). The result is best described as
a bimodal distribution representing seasonal changes in level
of terrestrial activity. The peak in June of each year
represents movements of gravid females during nesting sea-

USE OF TERRESTRIAL HABITAT BY CLEMMYS MARMORATA: IMPLICATIONS FOR MANAGEMENT 355
Overwintering
Following the high level of terrestrial activity
associated with nesting in June, there was a lull
during which few turtles were found on land (Fig-
ure 4). A second period of terrestrial behavior be-
gan in September. when all 12 radio-equipped tur-
tles left the river. It was not possible to establish
the exact departure dates. as the turtles were moni-
tored weekly.
Between September and early December, tur-
tles made as many as four changes in position on
land. At each location, they were found com-
pletely buried under leaf or needle litter. The lo-
cations eventually occupied for the duration of the
winter were all in upland habitat beyond the ri-
parian zone at a mean distance of 203 m from the
water. However, they varied with respect to mi-
crohabitat features (Table 1). Vegetation type in-
cluded both hardwood-dominant and conifer-
dominant woodlands. Two turtles overwintered in
lentic bodies of water. whereas the other ten re-
mained on land.
Return movements to the river from overwin-
tering sites began in February 1993 and were not
completed until as late as June. Turtles visited a
variety of locations along the way; the average re-
turn time to the river was seven weeks. Although
travel speed (meters travelled per day on average)
was not related to distance of overwintering sites
from the river, there was a significant correlation
between travel speed and order of initiation of
return movements (P < .02, Spearman rho, two-
tailed). Specifically, turtles that initiated their re-
turn trips later travelled faster. The two individ-
uals that overwintered in other bodies of water
(slough and lake) were the last to initiate their re-
turn trips, and their movements overland were the
fastest (Figure 6). The routes followed to over-
wintering sites were in some cases different from
the return routes. For example, two individuals
stopped for a few weeks at a vernal pool that had
been dry in the fall during their outbound trip.
DISCUSSION
The western pond turtles we monitored exhib-
ited considerable terrestrial activity. Females nat-
urally travel onto land for nesting, but the large
number of overland trips they made before actu-
ally ovipositing is noteworthy. Rathbun et al.
(1992) also reported multiple trips by gravid fe-
males. It is possible that females gain a thermo-
regulatory advantage by spending time buried on
Figure 4. Proportion of radio-tagged turtles in the Trinity River study reach that
travelled onto land during each month.
Figure 5. Mean number of days spent on land by gravid females during the
weeks before and after ovi
p
osition.
Figure 6. Travel speeds of turtles from overwintering sites in spring of 1993.

land during preovipositional development of the embryo
(e.g., Podocnemis expansa). The air on the Trinity River is
consistently warmer than the water at this time of year
(USFWS temperature records, Lewiston station). It is also
possible that females are responding to our presence with
preovipositional arrest, a mechanism that allows them to
retain eggs until conditions are favorable (Ewert, 1985). In
either case, their tendency to burrow under litter is consistent
with known behavior of other species (e.g., Kinosternon
subrubrum and Clemmys insculpta) during periods spent on
land (Bennett, 1972; Kaufmann, 1992).
It is of interest that hatchlings did not emerge from the
nest until spring. Hatchlings of other emydid species are
known to remain in terrestrial nest cavities through their first
winter (Hartweg, 1946; Gibbons and Nelson, 1978; Gibbons,
1990). Feldman (1982), who based his observations on cap-
tives, suggested that C. marmorata hatchlings may do the
same. It is also possible that egg development was sus-
pended and hatching delayed until just prior to the March
emergence time. This phenomenon of embryonic diapause
occurs in a number of turtle species (Ewert, 1985). Either
strategy may be an adaptive response to unfavorable condi-
tions on the Trinity River in the fall, such as high water lev-
els or low temperatures. Gibbons and Nelson (1978) sug-
gested that delayed emergence may provide the benefit of
sanctuary during a period when the growth benefits gained
from early emergence are likely to be outweighed by preda-
tion or by mortality from harsh environmental conditions.
All of the radio-equipped turtles spent seven months of the
year away from the river at overwintering sites. These
included both terrestrial refuges and lentic bodies of water
as far as 500 m from the river. This overwintering strategy
may be an adaptive response to winter flooding. However,
reasons for the long distances the turtles travelled, well be-
yond the flood zone, are unknown. It is also of interest that
most turtles overwintered on relatively cool north- and east-
facing slopes rather than south- and west-facing slopes.
Microhabitat characteristics of the overwintering sites
were variable. There did not appear to be an association
with any single habitat type. The timing and duration of
movements during the 1992-1993 winter varied with the in-
dividual turtles. The departure from the river was asyn-
chronous and the spring return even more so. Turtles that
left the river later travelled faster, which may be attributable
to the warmer temperatures and consequently higher activity
potential.
CONCLUSIONS
This study provides a preliminary framework for devel-
oping management programs for the long-term survival of
this species. Western pond turtles travel onto land for a va-
riety of reasons and consequently occur in the terrestrial en-
vironment during all times of the year. These findings pre-
scribe a management strategy that provides protection for
not only waterways but also adjacent lands. The upland area
used by turtles at this study site far exceeds the size of tradi-
tionally protected buffer zones alone rivers. Unfortunately.
with our current knowledge we cannot predict specifically
which portions of the terrestrial environment are critical for
356 D. A. REESE AND H. H. WELSH
TABLE 1
Overwintering habitat of turtles on main stem Trinity River during 1992-1993. Douglas fir dominant = ≥ 75% Douglas firs;
hardwood dominant = ≥ 75% hardwoods; conifer dominant = mix of conifers with none comprising more than 75%; mixed
hardwood/conifer = neither comprising more than 75%; mixed alder/willow = neither comprising more than 75%.
Turtle/Sex Slope aspect Habitat type Canopy cover (%) Distance from shore (m)
698 ♀ 25 NE Douglas fir dominant 75 480
773 ♀ 15 NE Hardwood dominant 50 255
949 ♀ 5 E Hardwood dominant 50 75
749 ♀ None Mixed hardwood/conifer 75 85
215 ♀ 10E Mixed hardwood/conifer 90 126
868 ♀ 5 N Conifer dominant 75 215
678 ♂ None Conifer dominant 80 245
725 ♂ 20 E Conifer dominant 15 95
528 ♂ None Hardwood dominant 70 145
560 ♂ Lake Hardwood dominant 0 500
377b ♂ 14 NE Mixed hardwood/conifer 50 65
335b ♂ Slough Mixed alder/willow 59 65

USE OF TERRESTRIAL HABITAT BY CLEMMYS MARMORATA: IMPLICATIONS FOR MANAGEMENT 357
Feldman, M. 1982. Notes on reproduction in Clemmys marmorata.
Herpetol. Rev. 13:10-11.
Gibbons, J. W. 1986. Movement patterns among turtle populations:
Applicability to management of the desert tortoise. Herpetologica
42(1):104-113.
Gibbons, J. W. 1990. Life History and Ecology of the Slider Turtle.
Smithsonian Institution Press. Washington. D.C. i-xiv + 368 pp.
Gibbons, J. W. and D. H. Nelson. 1978. The evolutionary signifi-
cance of delayed emergence from the nest by hatchling turtles.
Evolution 32(3) 297-303.
Hartweg, N. 1946. Confirmation of overwintering in painted turtle
hatchlings. Copeia 1946:255.
Holland, D. C. 1991a. Status and reproductive dynamics of a popu-
lation of western pond turtles (Clemmys marmorata) in Klickitat
County, Washington in 1991. Report to the Washington Depart-
ment of Wildlife.
Holland, D. C. 1991c. Distribution and current status of the western
pond turtle (Clemmys marmorata) in Oregon. Report to the Ore-
gon Department of Fish and Wildlife.
Holland, D. C. 1992. A synopsis of the ecology and current status of
the western pond turtle (Clemmys marmorata). Report prepared
for USDI Fish and Wildlife Service, San Simeon, California.
Kaufmann, J. H. 1992. Habitat use by wood turtles in central Penn-
sylvania. J. Herpetol. 26(3):315-321.
Lind, A. J., R. W. Wilson. and H. H. Welsh. Jr. 1992. Distribution
and associations of the willow flycatcher, western pond turtle, and
foothill yellow-legged frog on the main fork Trinity River. Interim
report submitted to the Wildlife Task Group. Trinity River Resto-
ration Project, USDI Fish and Wildlife Service, and Bureau of
Reclamation, Weaverville, California.
Mayer, K. E. and W. F. Laudenslayer, Jr. (eds.). 1988. A Guide to
Wildlife Habitats of California. California Department of Forestry
and Fire Protection, Sacramento, California.
Quinn, N. W. A. and D. P. Tate. 1991. Seasonal movements and hab-
itat of wood turtles (Clemmys insculpta) in Algonquin Park, Can-
ada. J. Herpetol. 25:217-220.
Rathbun, G. B., N. Seipel, and D. Holland. 1992. Nesting behavior
and movements of western pond turtles. Clemmys marmorata.
Southwest. Nat. 37(3):319-324.
Reese, D. A. 1996. Comparative demography and habitat use of west-
ern pond turtles in northern California: The effects of damming
and related alterations. Ph.D. dissertation, University of California,
Berkeley. 253 pp.
Schwarzkopf, L. and R. J. Brooks. 1987. Nest-site selection and off-
spring sex ratio in painted turtles, Chrysemys picta. Copeia 1987:
53-61.
Stebbins, R. C. 1985. Field Guide to Western Reptiles and Amphi-
bians. Houghton Mifflin Co., Boston. 336 pp.
Storer, T. I. 1930. Notes on the range and life-history of the Pacific
freshwater turtle, Clemmys marmorata. Univ. Calif. Publ. in Zool.
32:429-441.
Ward, F. P., C. J. Hohmann, 1. F. Ulrich, and S. E. Hill. 1976. Sea-
sonal microhabitat selections of spotted turtles (Clemmys guttata)
in Maryland elucidated by radioisotope tracking. Herpetologica
32:60-64.
LITERATURE
CITED
Bennett, D. H. 1972. Notes on the terrestrial wintering of mud turtles
(Kinosternon subrubrum). Herpetologica 28(3):245-247.
Bennett, D. H., J. W. Gibbons, and J. C. Franson. 1970. Terrestrial
activity in aquatic turtles. Ecology 51:738-740.
Burger, J. and W. A. Montevecchi. 1975. Nest site selection in the
terrapin Malaclemys terrapin. Copeia 1975:113-1 19.
Chase, J. D., K. R. Dixon, J. E. Gates, D. Jacobs, and G. J. Taylor.
1989. Habitat characteristics, population size, and home range of
the bog turtle, Clemmys muhlenbergii, in Maryland. J. Herpetol.
23(4):356-362.
Congdon, J. D., D. W. Tinkle, G. L. Breitenbach, and R. C. van Loben
Sels. 1983. Nesting ecology and hatching success in the turtle
Emydoidea blandingii. Herpetologica 39:417-429.
Ewers, M. A. 1985. Embryology of turtles. In C. Gans, F. Billets, and
P. F. A. Maderson (eds.), Biology of the Reptilia, Vol. 14, pp.
75-268. John Wiley and Sons, New York.
western pond turtles. They utilize a variety of upland habi-
tats as well as the network of creek. ponds. and ephemeral
bodies of water associated with riverine systems.
The riparian habitat serves an integral role in the life
history of the western pond turtle. and many other species
may cross the uncertain boundary between river and land
because of their specialized. seasonally varying require-
ments. We must therefore reevaluate our view of riparian
habitat as "buffer zone;" which connotes a supportive rather
than primary role in the ecosystem. Terrestrial riverine
habitat warrants consideration aside from its function of
buffering the aquatic habitat from external impacts. Man-
agement strategies that address the functioning of entire wa-
tersheds are more likely to afford adequate protection for
these vagile, semi-terrestrial species.
ACKNOWLEDGEMENTS
The U.S. Bureau of Reclamation Trinity River Restora-
tion Project provided the majority of funding for this re-
search. Additional funding was provided by the U.S. Forest
Service Redwood Science Lab. Shasta-Trinity National
Forest, and U.S. Forest Service Fish-Habitat Relationships
program. We wish to acknowledge Amy J. Lind for her
continuing role in providing logistical assistance as well as
scientific expertise. Thanks to Harry W. Greene, Reginald
H. Barrett, and Mary Power for their review of the manu-
script. The following persons helped in the field: Jenny
Glueck. Polly Taylor, Randy Wilson, Sarah Mook. Douglas
Welsh, Logan Olds, and Amy Lind. We are grateful to all
of them.