Elk monitoring in Lewis and Clark National Historical Park: 2008-2012 synthesis report

Maintaining elk (Cervus elaphus roosevelti) herds that frequent Lewis and Clark National Historical Park (NHP) is central to the park’s purpose of preserving the historic, cultural, scenic, and natural resources associated with the winter encampment of the Lewis and Clark expedition. Elk were critically important to the Lewis and Clark expedition in providing food and hides that sustained the expedition during the winter of 1805-06 and supplied them for their return east during 1806. Today, elk remain a key component of interpreting the Lewis and Clark story to over 200,000 park visitors each year at the Fort Clatsop visitor center.

In 2008, the US Geological Survey (USGS) began collaborating with Lewis and Clark NHP and
the NPS North Coast and Cascades Network to develop a protocol for monitoring long-term
changes in the magnitude and spatial patterns of elk use within and adjacent to Lewis and Clark
NHP (Griffin et al. 2011). Specific objectives of the monitoring program were to measure trends
in (1) relative use of the Fort Clatsop unit by elk during winter; (2) the proportion of areas where
elk sign is present in the Fort Clatsop unit in winter; and (3) the frequency of elk sightings from
roads in and around the Fort Clatsop unit. This report synthesizes the results of the first four
years of monitoring elk distribution and use in Lewis and Clark NHP from 2008-2012. We also
present data from FY2012 (Appendix 1), in lieu of an annual report for that year.

We used fecal pellet group surveys as the cornerstone for monitoring trends in both relative use
of the Fort Clatsop Unit by elk and the proportion of areas where elk sign was present at the end
of winter. We estimated pellet group density based on data collected from a network of fecal
pellet plots distributed systematically throughout the unit. We developed a double observer
sampling scheme that enabled us to estimate detection biases and improve the accuracy of pellet
group density estimates. We computed the estimated detection probability for any pellet group
observed; this probability was a function of the pellet group size and stage of decay, as well as
lighting and vegetation conditions, and the number of observers (one or two) searching for
pellets in that subplot. We then used these estimated detection probabilities to adjust the raw
counts of the detected pellet groups to account for groups that likely went undetected under
similar pellet and environmental conditions (each observed pellet group was weighted by the
inverse of its estimated detection probability). We also used results from the late winter fecal
pellet surveys to quantify the proportion of areas where elk pellets occurred (PAO), which was
based on the presence of fecal pellet groups and estimation of detection biases (i.e., accounting
for pellet groups that likely went undetected by both observers). In this synthesis, we report
temporal trends in both pellet group density and PAO from 2008-2012, based on weighted linear
regression analyses as well as spatial variation of pellet group densities over time.

We completed late winter fecal pellet surveys at 61-66 plots annually, depending on yearly
variation in access. We cleared fecal pellets at survey points in late October / early November
each year and returned in late February / early March to count pellet groups left by elk over the
winter. The estimated probability that a pellet group was detected by any one observer during
late winter was affected most by the pellet group size and was less affected by decay class and
lighting conditions. Per-observer detection probabilities ranged from as low as ~10-15% for
single pellets to ~85-90% for pellet groups with 50 pellets. Average pellet group density in the
Fort Clatsop unit ranged annually from 0.58 (+/- 1.43 standard error [SE]) to 0.93 (+/- 2.25 SE)
pellet groups per 3-m radius subplot. Pellet group density declined over time, at approximately 8.8% per year (+/- 2.5% SE), but that slope was not statistically distinguishable from zero (2-
tailed P=0.16). Following correction for detection biases, the proportion of surveyed points used
by elk (i.e., PAO) ranged from 0.44 (+/- 0.07 SE) to 0.53 (+/- 0.07 SE) during the 4 winters. The
estimated proportion of areas where elk pellets occurred (PAO) declined at a rate of 2.6% per
year (+/- 1.2% per year SE), but that trend also was not statistically distinguishable from zero (2-
tailed P=0.17). Statistical significance of a measure’s trend depends on both the magnitude (i.e.,
slope) of the observed trend and the number of years the trend continues in the same increasing
or decreasing direction. Through simulation modeling we determined how many additional years
of surveys would be required to reveal a statistically significant trend, based on the same trends
in pellet group density and PAO, and associated variation, observed from 2009-2012. Assuming
the same trends persist in the future, simulations indicated that there is a 70% probability that a
statistically significant trend would be detected after two more years of conducting pellet group
surveys.

Relative use by elk during winter, as indexed by elk pellet group density, was generally greatest
in the southeast region of the Fort Clatsop unit in or near the large freshwater marsh at the mouth
of Colewort Creek and adjacent upland areas. Pellet group density was also higher than average
in the north-central forested area, not far from a privately-owned pasture north of the park
boundary. This spatial pattern in pellet group densities across the Fort Clatsop unit was
consistent across all four years, although specific pellet group densities varied from year to year.
Pellet group density declined significantly over time at two points in the southeast of the Fort
Clatsop unit, even though pellet group density at those points remained higher than the unit
average. Pellet group density increased significantly over time at one point in the north-central
region, and at one point in the south-central region of the unit, indicating a slight shift in the
distribution of elk use within the Fort Clatsop Unit over the four years.

As an index of visitors’ opportunities to see elk in and around the Fort Clatsop Unit, we
conducted replicated roadside elk surveys 3-5 times monthly during February, April, June,
August, October and December 2008-2012. During each morning of survey, we searched for elk
along four routes that totaled 32 km. We examined bimonthly trends in the numbers of elk
groups seen, the total number of elk seen, and the observed composition ratios for those six
months of the year. The average number of elk groups seen per survey ranged from 0.75 (+/-
0.32 SE) during February to a peak of 1.95 (+/- 0.36 SE) during June. Despite this seasonal
variation in numbers of elk groups seen, the average total number of elk seen per morning was
less variable. The average ratios of antlered elk to antlerless adult elk (i.e., bulls:cows) and
calves to antlerless adult elk (i.e. calves:cows) varied seasonally, with the highest of both
average ratios observed in August. We detected no significant trends in the average number of
elk groups and total numbers of elk seen per survey from 2008-2012. Similarly, ratios of calves
and antlered elk per antlerless elk did not differ over time.

Elk groups were frequently seen from January to August in the southeast region of the Fort
Clatsop unit, in the vicinity of Colewort Creek. Outside of NPS lands, we observed elk most
frequently in open areas near the Astoria regional airport, in the pastures and forests immediately
north of the Fort Clatsop unit and, prior to the construction of a residential development, in a
pasture northwest of the Fort Clatsop unit.

Elk monitoring at Lewis and Clark NHP is still in its initial years and additional monitoring will
be required to verify trends that appear to be emerging. For example, the initial monitoring
suggested incipient declining trends in both pellet group density and proportion of plots with
pellets present, as well as, potentially, a small shift in elk distribution away from a new trail that
was recently constructed in the southeast portion of the Fort Clatsop unit. Continued monitoring
will aid in determining whether this local change in distribution persists (or, alternatively,
resulted from short-term random variation), and whether there will be any positive or negative
effect in the northern portion of the unit where a new trail has been constructed. High variability
in road counts prevented our ability to find any clear trend in numbers or composition of elk
observed in and near Fort Clatsop, but changes in the patterns of observations of elk from
roadways suggest that residential development outside the park has reduced the available habitat
for elk in some of the areas surveyed, and may have affected spatial use patterns of elk adjacent
to some areas of the park. In addition to monitoring future effects of land use changes outside the
park, continued monitoring may also prove useful for assessing elk responses to natural
succession in forests disturbed by windthrow in December 2007 and to NPS vegetation
management activities such as variable density thinning in the forest, trail development, and
restoration at Otter Point tidal area and Colewort Creek Slough.