HomeMy WebLinkAbout01 Jan 2025 packet
Agenda items are subject to change – Times are approximate and are also subject to change
Agenda | January 16, 2025
River Park Center
Second Floor Conference Room
123 Emma Road, Basalt CO 81621
TIME ITEM
Administrative Items
4:00 PM Board Comment Public Comment Additions/Deletions to Agenda 2025 Calendar
4:10 PM Approval of Minutes December 19, 2024 Appointment of Chair and Vice Chair Outreach/Education/Grants
4:15 PM White River USFS-Healthy Rivers Program: Beaver Inventory 2023-2024
– R. Clay Ramey, Aspen-Sopris District Fisheries Biologist, White River
National Forest
4:45 PM Priorities for 2025- Healthy Rivers Board and Staff Staff Reports
5:15 PM Monthly Staff Report- Lisa MacDonald and Lisa Tasker
5:20 PM Open Discussion
5:30 PM Adjourn
Upcoming 2025 Regular Meetings Feb 20, Mar 20, Apr 17, May 15, Jun 19, Jul 17, Aug 21, Sep
18, Oct 23, Nov 20, Dec 18
1
1
HEALTHY RIVERS AND STREAMS CITIZENS ADVISORY BOARD
Regular Meeting December 19, 2024 4:00 PM 123 Emma Road, 2nd Floor Basalt, CO 81621
River Board members present: Andre Wille, Bill Jochems, Chris Lemons, Kirstin Neff, Wendy Huber, Maciej Mrotek Mr. Andrews appeared via zoom River Board members absent:
Attendees in addition to presenters: Lisa MacDonald, Lisa Tasker, Matt Annabel
Media Present: _____________________________________________________________________________________________________________________________________ Board Comment: Chair Huber thanked the county for the 2025 Aspen Historical Calendars. Public Comment: None Additions/Deletions to the Agenda: Shift agenda items due to zoom technical difficulties.
2025 Calendar
Staff prepared a draft of a 2025 calendar
Approval of Minutes November 21, 2024
Mr. Mrotek moved to approve the Nov 21, 2024 minutes. Ms. Neff seconded the motion. The motion passed
6/0. Mr. Andrews joined at approximately 4:20 due to zoom difficulties on our end. Lincoln Creek Phase 1 Presentation - LRE Water for the Lincoln Creek Workgroup – Brent Johnson and Chip Fisher Mr. Fisher and Mr. Johnson presented the results of the phase 1 work on Lincoln Creek funded by the Healthy Rivers Board and the Independence Pass Foundation.
WECo Water Fluency Program – Kirstin Neff Ms. Neff provided information on the grant program.
Aspen 82, Aspen Mountain Interview – Matt Annabel Board received spreadsheet and sign up for topics and dates for the interviews. Scholarship Committee Mr. Wille and Ms. Neff will make up the 2025 scholarship committee.
Staff Report Staff discussed the addition of the communication stats for 2024.
Open Discussion Change grant cycles to once a year
2
2
Mr. Jochems toured a Navajo Tribal Area in the Southwest area. Ms. Neff moved to adjourn the meeting. Mr. Wille seconded the motion. The motion passed 7/0. The meeting adjourned at approx. 6:20 pm.
Approved: Attest:
______________________________________________ _________________________________________________ Wendy Huber – Chair Lisa Tasker
3
Pitkin County Healthy Rivers Program and White River Na�onal Forest:
Beaver Inventory, 2023-2024
R. Clay Ramey
U.S. Forest Service
White River Na�onal Forest
Aspen-Sopris District Fisheries Biologist
robert.ramey@usda.gov
4
2
Summary
The Pitkin County Healthy Rivers Board funded a beaver- and habitat inventory on Na�onal Forest lands
in the upper Roaring Fork Watershed for the 2023 and 2024 field seasons to increase our understanding
of the current status of beaver in the watershed. Ul�mately, this work seeks to inform beter riparian
management and watershed health in the Roaring Fork Valley, and project work in furtherance of those
goals.
The Healthy Rivers Board provided $100,000 in funding via agreement to field seasonals to conduct this
inventory and describe its findings.
A random sampling protocol was developed to inventory roughly 353,016 acres of Roaring Fork
headwaters managed by White River Na�onal Forest. Over both field seasons, 296 sites were surveyed
in 66 streams in 15 HUC6s. Random surveys found 138 dams and 16 lodges. We found 17% of random
survey sites to be currently occupied by beaver. About 34% of surveys observed some sign of beaver.
Currently occupied sites were found in 11 of the 15 HUCs and “past” evidence of beaver were found in
all 15 HUCs.
Sites that were or had been occupied by beaver had a different mix of vegeta�on than unoccupied sites.
We found that as the percentages of willow and aspen available at a site increased, and as stream slope
and eleva�on decreased, the probability of beaver occupa�on increased. We found that while beaver
occupied areas both within and outside of catle grazing allotments, willow was more available outside
of grazing allotments than within them.
Habitat suitability criteria were developed based on the condi�ons we found to be associated with
beaver. Those criteria were applied to es�mated and measured vegeta�on condi�ons throughout the
study area to determine which sub-watersheds were likely to have the most available habitat for beaver,
and which would be good candidates for addi�onal project work to improve their suitability.
This work provided empirical data related to beaver, their habitat, and their frequency and spa�al
distribu�on on the Forest. The methodology used by this project is now a tool that the Forest can use to
assess beaver habitat in its ongoing land management efforts.
5
3
Introduc�on
The American beaver (Castor canadensis) is a na�ve keystone species within the 2.3 million acres
managed by the White River Na�onal Forest (the Forest, hereina�er). Prior to American setlement of
the Roaring Fork Valley, beavers were believed to be common, if not nearly ubiquitous. The size and
spa�al distribu�on of the current beaver popula�on is believed to be considerably smaller as a result of
past and current land- and wildlife management prac�ces. The reduc�on in beaver and the subsequent
loss of their dams may represent a loss on the order of 40% of beaver-related surface water and
sediment storage capacity in Colorado (Scamardo et al., 2022). In the local snowmelt driven hydrology,
beaver dam construc�on affects how water moves through the landscape and consequently influences
instream and riparian condi�ons and the subsequent ecology.
In the winter of 2022, the Pitkin County Healthy Rivers Program approached the West Zone Fisheries
Program to cooperate in efforts to promote beaver occupa�on on the Forest in the interest of
maintaining and improving watershed health in the Roaring Fork Valley. It was determined that an
important first step was to collect empirical data related to beaver presence and spa�al distribu�on,
their habitat use, and the availability of those habitats on the Forest. The intent of data collec�on was to
provide informa�on to inform future efforts to promote beaver occupa�on on the Forest.
The general habitat requirements of beaver are well known based on literature going back to at least the
1950s (as summarized in Allen, 1983). Modern remotely sensed data and analy�cal tools can illustrate
the current extent of beaver impoundments across the landscape (Marshall et al., 2024). We could thus
iden�fy the current spa�al distribu�on of beaver on the Forest. What we lacked were empirical data
that would help provide some explana�on of that distribu�on related to riparian vegeta�on and physical
habitat. Secondly, we were looking to iden�fy factors limi�ng beaver u�liza�on and distribu�on so that
where feasible, treatment(s) and Forest management could beter facilitate beaver occupa�on to
improve watershed health.
Agreement
Pitkin County Health Rivers allocated $50,000 in 2023 and $60,000 in 2024 to the Forest via an
Agreement for these purposes (table 1). Funds were used to hire and equip two Forest Service seasonal
personnel to conduct a beaver inventory in the Roaring Fork headwaters to collect and summarize
empirical data related to beavers and their habitat on the Forest.
Table 1. Actual costs billed to Agreement.
2023 2024 Totals
Agreement Funds $50,000 $50,000 $100,000 Salary $35,530 $41,292 $76,822 77%
Rent $5,000 $0 $5,000 5%
Truck $8,369 $5,474 $13,843 14%
Supplies $1,101 $2,712 $3,813 4%
Sub-Totals $50,000 $49,478 $99,478
Funds Remaining $0 $522 $522 1%
6
4
Methods and Analyses
Because the area of interest was large (353,016 acres in 15 6th field hydrologic unit codes, or HUCs), a
random sampling protocol was developed. The Forest u�lized spa�al so�ware (ESRI ArcPro) to iden�fy
stream reaches on the Forest with less than approximately 10% slope based on the assump�on that
steeper streams were less likely to be frequently u�lized by beaver. Random sampling points were
created on the iden�fied stream reaches throughout the Roaring Fork headwaters and a 100 m buffer
was applied to each point. Using images from GoogleEarth, the major woody vegeta�on types that were
assumed to be relevant to beaver use (willow, aspen, cotonwood, alder, and conifer) were iden�fied and
their area calculated as a percentage of the area within the 100 m buffer around each point. Surveyors
conducted each survey with these es�mates of the rela�ve frequency of cover types in hand which they
either confirmed or modified based on condi�ons found at the �me of the survey. The Forest devised a
survey protocol that characterized the beaver, stream, and riparian condi�ons within the 100 m buffer
around each survey point. Survey metrics included the presence of beaver sign (cut stems, tracks, etc.;
may or may not have been evidence of actual, “permanent” occupa�on of the site); the occupancy
status of the site (new, current, past, or unoccupied); the number of dams and lodges observed; the
average bankfull- and floodprone widths of the stream and the resul�ng confinement ra�o; the channel
slope; and the Rosgen stream type (Rosgen, 1994). In 2024, channel slope was the only stream channel
metric measured.
Among the goals of the inventory was to atempt to define the range of habitat condi�ons that could be
associated with beaver use. Owing to the random sampling method used and the patchy distribu�on of
beaver, the number of currently occupied sites surveyed was small. Late in the 2023 sampling season, it
was decided to bolster the sample size of occupied sites. Fi�een sites were surveyed in 2023 that were
known to be currently occupied by beaver to beter describe the habitat condi�ons at occupied sites.
Only randomly selected sites were surveyed in 2024.
While comple�ng the surveys in the field, beaver occupa�on of sites was categorized as either new,
current, past, or unoccupied (ever, so far as could be observed). These labels were assigned based on
the judgment of the surveyors. During data analysis, condi�ons at each type of site (new, current, past,
unoccupied) were compared to each other for each of the habitat metrics (vegeta�on frequencies,
stream slope, etc.) using analysis of variance (aov in RStudio (R Core Team, 2024)), and Tukey’s HSD was
used to compare specific differences between groups. In terms of the vegeta�on, channel slopes, and
bankfull widths at each type of site, new, current, or past sites were not sta�s�cally different from one
another. Condi�ons at the new, current, or past sites were almost always sta�s�cally different than
those at unoccupied sites. Therefore, further ANOVA models lumped new/current/past into a new
category of “occupied” and compared those values to unoccupied sites. It was known that beaver can
significantly and durably affect the habitats that they occupy and therefore it was not surprising to find
that sites formerly occupied by beaver were more like sites that were recently or currently occupied than
they were to sites that have never been.
Habitat suitability criteria were defined using data from “new” and “current” sites because they
represented condi�ons observed to be used by beaver in their current state. The specific values were
based in large part on the 10th and 90th percen�les for each metric. Percen�les were used to trim off
extreme values on either end that may represent marginal or rare circumstances. The assump�on was
7
5
that the central por�on of the distribu�ons were more likely to represent typical habitat use and
therefore be more useful in terms of assessment or predic�on.
Addi�onal metrics for each site and HUC were determined in the office. Each site was given a site code,
its 6th field HUC determined, eleva�on, coordinates (UTM), the name of the stream it was on, the Forest
Management Area it was within (USDA Forest Service, 2002), and whether permited livestock grazing
(by catle, or not) occurred at that point.
For each 6th field HUC, its area, the percentage of that area that was managed by the Forest and that
under Wilderness management, the kilometers of stream within and the percentage of that length that
was less than 10% slope, and the average eleva�on, was determined. Road- and trail densi�es were also
summarized by HUC. Addi�onally, a spa�al layer of features believed to be beaver impoundments as of
2021 was provided by Sarah Marshall (Ecohydrologist, Colorado Natural Heritage Program, Colorado
State University; Marshall et al., 2024). These data were referred to for a variety of purposes.
Data were stored in an Access database (Microso�) and queries were writen to summarize as needed.
Sta�s�cal analyses were conducted using R (using linear and generalized linear regression, and ANOVA,
lm, glm, and aov, respec�vely) and were considered significant at alpha = 0.05. Some analyses at the
HUC6 level referred to alpha = 0.1 because of the small sample size and a greater degree of uncertainly
was acceptable. Only data from randomly sampled sites were used to characterize the spa�al
distribu�on and density of beaver across the sampling area. Data from non-random sites that targeted
occupied stream reaches were only used to beter describe condi�ons at occupied sites.
Results
Over both field seasons, 296 sites were surveyed over 77 field days (3.8 surveys per day), in 66 streams
in 15 HUC6s. Random surveys found 138 dams and 16 lodges.
Beaver Frequency and Spa�al Distribu�on
We found 17% of random survey sites to be currently occupied by beaver (beaver were obviously
residing at those loca�ons). About 34% of surveys observed some sign of beaver. Currently occupied
sites were found in 11 of the 15 HUCs sugges�ng that the current popula�on may be widely, if sparsely,
distributed. Table 2 summarizes the frequency of beaver sign within each HUC. Sites in in the
Headwaters Fryingpan River sub-watershed rarely had beaver sign, while sites in the headwaters of
Snowmass Creek frequently did.
Evidence of “past” beaver occupa�on was found at 37% of sites in all 15 subwatersheds sugges�ng that
there used to be more beaver in more places than there are currently.
8
6
Table 2. Frequency of beaver sign within each HUC. Only HUCs with 10 or more surveys were considered
here.
HUC6
# Sites
Surveyed
# Sites with
Beaver Sign
% Sites with
Beaver Sign # Dams # Lodges
Headwaters Fryingpan River 23 2 9% 1 1
Lime Creek 13 2 15% 0 0
North Fork Fryingpan River 27 6 22% 0 0
Fourmile Creek 38 9 24% 17 3
Deeds Creek-Fryingpan River 25 7 28% 21 1
Hunter Creek 10 3 30% 3 0
Thompson Creek 79 27 34% 45 6
Headwaters Maroon Creek 10 4 40% 11 1
Mill Creek-Cattle Creek 19 12 63% 13 1
Headwaters Snowmass Creek 12 9 75% 17 3
Habitat Used
Among the goals of the inventory was to try to associate specific habitat condi�ons with beaver
occupa�on. As a result of the random sampling methodology used for the inventory and what appeared
to be a patchy distribu�on of beaver, few currently occupied sites were randomly surveyed. Fi�een sites
that beaver were known to occupy were surveyed in 2023 in an effort to bolster the suitable habitat
por�on of this project. These sites were iden�fied using data from Marshall et al. (2024) of current (as
of 2021) beaver dams/ponds and were in the Castle, Express, and North Thompson creek drainages.
Vegeta�on Characteris�cs
Sites that were or had been occupied by beaver had a different mix of vegeta�on than unoccupied sites.
Table 3 summarizes the percentages of each major type of vegeta�on found within a 100 m radius of
each survey site. With the excep�on of those for cotonwood, the percentages of each vegeta�on type
were significantly different between occupied and unoccupied sites.
Sites occupied by beaver averaged 16% willow, 6% more willow than at unoccupied sites. Only 3 “past”
sites had zero willow, and there wasn’t even water at one of those sites currently. Occupied sites had an
average of 11% aspen, 6% more than unoccupied sites.
Twenty four percent of sampled sites had 1-17% alder. The difference in alder between occupied and
unoccupied sites was sta�s�cally significant, but not par�cularly useful at only 0.74%. Where alder were
found, beaver were found almost 70% of the �me, sugges�ng that it’s a veg type that they u�lize. Given
this very small difference however, from a prac�cal perspec�ve, alder wasn’t very useful to discriminate
site suitability.
Cotonwood were only noted at 17 out of the 296 sites. Occupied sites with cotonwood had 1—25%
cotonwood, more willow and aspen, and less conifer, while unoccupied sites with cotonwood had only
1-5%, with less willow and aspen and more conifer.
9
7
Table 3. Vegetation differences between occupied and unoccupied sites. All differences were significant
at alpha = 0.05, except for cottonwood.
Habitat Metric Group n Min Mean Mean + Sd.Dev
Willow (%) Occupied 158 0 16 30
Unoccupied 138 0 10 24
Aspen (%) Occupied 158 0 11 26
Unoccupied 138 0 5 15
Alder (%) Occupied 158 0 1 4
Unoccupied 138 0 0 2
Cottonwood (%) Occupied 158 0 1 4
*Not sig. diff. Unoccupied 138 0 0 1
Conifer (%) Occupied 158 0 27 48
Unoccupied 138 0 41 72
Unoccupied sites had 14% more conifer on average. Surveyors noted that they never saw any conifer
with chew marks. Higher percentages of conifer necessarily excluded the other vegeta�on types so the
absence of deciduous vegeta�on more so than the presence (or absence) of conifer per se is probably a
more likely explana�on.
A model was run using logis�c regression (glm in R) to gauge the rela�ve importance of each of the
vegeta�on types on the probability of beaver occupa�on. The percentages of willow, aspen, and alder
were the only significant terms in the model which suggested that the prevalence of those species were
important factors in beaver occupa�on. As the percentages of willow, aspen, and alder increased, the
probability of beaver occupa�on increased. An alterna�ve model was run that combined willow, aspen,
alder and cotonwood into “woody veg”. The model that included each species individually was a beter
fit (the AIC was significantly lower), more explanatory, than the model that used the aggregated woody
veg metric. That would imply that these woody species are important individually and are not
interchangeable.
Physical Habitat Characteris�cs
There were differences in some of the physical habitat characteris�cs of occupied and unoccupied sites.
At occupied sites, channel slopes were flater, bankfull widths were wider, and eleva�ons were lower
(table 4). Surveyed eleva�ons ranged from 7,444 to 11,581 feet. The highest current beaver sign we
found was at 10,561 feet. We found past sign at 10,909 feet.
10
8
Table 4. Differences in physical habitat between occupied and unoccupied sites. All differences were
significant at alpha = 0.05.
Habitat Metric Group n Minimum Mean Mean + Sd.Dev
Elevation (ft) Occupied 132 7,444 9,222 9,954
Unoccupied 129 7,531 9,791 10,565
Channel Slope (%) Occupied 132 0 4 7
Unoccupied 129 0 6 9
Bankfull Widths
(ft) Occupied 56 1 7 12
Unoccupied 61 1 5 7
Confinement ra�os (floodprone width/bankfull width; Rosgen, 1994) were not different between
occupied and unoccupied sites. This measurement appeared to have been confounded by the
geomorphic effects of beaver use of sites. At sites with extensive past beaver habitat modifica�on,
o�en�mes we saw the stream channel ou�lank the beaver structures and the stream channel thus
tended to be steeper and more confined in its current state than it might have been when beaver first
colonized the site. In other cases, formerly occupied sites that had been extensively modified by beaver
had large confinement ra�os (i.e. were unconfined), but that was a consequence of beaver use and not
its cause. Sa�sfied with this informa�on collected in 2023, we did not measure these variables in 2024.
Forest management areas, categories for specific areas of the Forest that spell out the types of uses
allowed, did not appear to have any influence on beaver occupa�on.
Rosgen stream types appeared to be differen�ally occupied (table 5; data from 2023). Channel slopes
are steepest at type “A” and get flater as you proceed down. The Forest contains many miles of type B
and therefore it’s not surprising that we surveyed more B’s than the other types. Type C have slopes <
2%, wider floodplains, and sinuous channels and were occupied more frequently than the other channel
types. The F type is o�en a degraded state where a sinuous, low gradient E became unstable and
entrenched, loosing access to its floodplain. Both F sites surveyed had past beaver sign and an F-type
morphology seemed to be a common form following beaver ac�vity a�er the dams failed. Following
dam failure, the stream eroded down through the material accumulated above the dam resul�ng in an F-
type morphology. So as was seen related to confinement, the associa�on with type F was probably more
a consequence than a cause.
Table 5. Rosgen stream types and the percentage of sites that were occupied.
Rosgen Type n % Occupied A 23 26% B 57 44% C 29 69% E 6 50% F* 2 100% * ”Past” sites
11
9
Another logis�c regression was run that included both the vegeta�on and physical habitat metrics in an
effort to integrate them. As willow and aspen increased, and stream slope and eleva�on decreased, the
probability of beaver occupa�on increased.
Beaver Distribu�on and Catle Grazing
Just over half of the sites surveyed were within catle grazing allotments (56%) and just over half of the
sites we found that did or used to have beaver were also within grazing allotments (Current/New/Past
averaged 58%; table 6). Slightly more “past” sites were within grazing allotments (59%). A chi-square
test of these numbers was insignificant (at alpha = 0.1; chisq.test in R) indica�ng that there was no
patern.
Table 6. Beaver and cattle grazing.
Occupancy n Catle No Grazing % Catle Grazing
Current + New 37 20 17 54%
Past 106 63 43 59%
Unoccupied 134 71 63 53%
Total 277 154 123 56%
Vegeta�on Rela�onships with Catle Grazing
Grazing was related to the vegeta�on present. Sites within grazing allotments had an average of 7% less
willow than un-grazed areas (table 7). Sites within grazing allotments had an average of 4% more aspen.
There was no difference in alder by grazing condi�on. Willow at unoccupied (“No Beaver”) sites within
grazing allotments averaged only 6.8% (figure 1).
Table 7. Differences in willow and aspen by grazing condition.
Habitat Metric Group n Min Mean Mean + Sd.Dev
Willow (%) Cattle 173 0 10 20
No grazing 123 0 17 35
Aspen (%) Cattle 173 0 10 23
No grazing 123 0 6 19
12
10
Figure 1. Difference in willow by beaver occupation and grazing condition.
Observed Beaver Sign
A generalized linear model with beaver sign as the outcome included willow, aspen, alder, cotonwood,
conifer, grazing condi�on, eleva�on, and slope. Stepwise regression (stepAIC in R) parsed the model
down to the most important terms. The resul�ng model found that the probability of finding beaver sign
at a site was posi�vely related to willow, aspen, and no grazing, and nega�vely related to eleva�on and
slope.
Habitat Suitability Criteria
Habitat suitability criteria (HSC) for beaver on the Forest were defined based on the vegeta�on and
habitat informa�on thus described. Only data from “new” and “current” sites were used (n = 48). Only
the values for willow, aspen, and stream slope were u�lized because those metrics appeared to be the
most relevant and those with the most variability. The eleva�ons used encompass nearly all of the
Forest for example, and therefore vegeta�on condi�ons were expected to constrain beaver to a greater
degree than eleva�on.
The HSC referred to in the rest of this analysis are summarized in table 8. Suitable sites were defined as
having at least 5% willow, at least 1% aspen, and slopes under 5.5%. The 10th percen�le of aspen was
s�ll zero as there were some sites where no aspen was noted. Notes related to those surveys included
references to the presence of other minor woody species like mountain mahogany and red osier
dogwood for example, that were beyond the precision of our habitat surveys. Because aspen was
repeatedly noted throughout data analyses as being an important correlate to beaver habitat use, the
lower bound of aspen was subjec�vely assigned as 1%. Upper bounds of posi�vely related metrics, and
13
11
the lower bound of the nega�vely associated metric, were ignored based on the idea that a site couldn’t
have too much willow or a slope too flat.
Table 8. Habitat suitability criteria.
Habitat Metric Lower bound Upper bound
Willow (%) 5.0 NA
Aspen (%) 1.0 NA
Channel Slope (%) NA 5.5
Habitat Availability
We had es�mated vegeta�on condi�ons for about 50 sites that were not surveyed. We incorporated
those sites into the dataset to es�mate the rela�ve frequency of habitats in each HUC that fit the HSC
(table 9. Only HUCs with more than 10 total sites were included.). The HUCs with the highest
percentage of sites with HSC therefore are likely to have the greatest area of suitable habitat available
for beaver (values in table 9 are in rank order of “% HSC”). The HUCs in the botom half of the table may
present the best opportuni�es to reintroduce beaver as it’s more likely that the habitat could support
them.
Table 9. The percentage of sites within each HUC6 with HSC, ranked in ascending order. “% Past” is the
percentage of surveyed sites with “past” beaver occupancy. “% New-Current” are the percentages of
surveyed sites with “new” or “current” occupation. “Avg % willow” and “aspen” are the average
percentages of willow and aspen in each HUC.
HUC6
Surveyed
Sites
% HSC all
sites
%
Past
% New-
Current
Avg %
Willow
Avg %
Aspen
Available
Habitat
Lime Creek 13 7% 15% 0% 9% 4% -
North Fork
Fryingpan River 27 7% 26% 0% 6% 2% Headwaters
Fryingpan River 23 8% 52% 4% 25% 0% Thompson Creek 79 18% 46% 23% 11% 8% Headwaters
Maroon Creek 10 18% 20% 20% 25% 18% Deeds Creek-
Fryingpan River 25 19% 8% 20% 19% 1% Fourmile Creek 38 24% 61% 3% 8% 17%
Hunter Creek 10 27% 10% 20% 18% 8% Headwaters
Snowmass Creek 12 38% 58% 17% 16% 19% Mill Creek-Cattle
Creek 19 47% 21% 58% 8% 3% +
Averages 21% 32% 16% 14% 8%
14
12
Metrics at the HUC6 Scale
Analyses examined HUC6-level correla�ons between summary sta�s�cs. Because very small sample
sizes distorted these analyses, only HUCs sampled more than 10 �mes were evaluated. Analyses at the
HUC level were conducted at alpha = 0.1 because sample sizes were small and the results were therefore
merely exploratory.
Wilderness and Forest Extent
The percentages of each HUC6 that were managed by the Forest or within Wilderness areas and their
rela�onship to observed beaver sign was considered. A linear model was run with % Wilderness, %
Forest, and eleva�on, with beaver sign as the outcome. The percentage of a HUC6 that was Wilderness
was posi�vely related to the frequency of observed beaver sign. Beaver sign decreased with eleva�on
which confounded the Forest term as the eleva�on and % Forest terms were correlated and eleva�on
had the most signal. The percentage of willow was posi�vely related to the percentage of Wilderness in
each HUC6.
Road and Trail Density
Road densi�es ranged from 0.02 to 0.76 km road/km2. Road density was nega�vely related to the
frequency of beaver sign, and was not related to either willow or aspen, so therefore a vegeta�on effect
of roads was not found. Trail densi�es were posi�vely related to beaver sign.
Discussion
Beaver Frequency and Spa�al Distribu�on
We found beaver sign in every subwatershed we surveyed which indicated that beaver are widely
distributed across the Forest. That about 1/3rd of the random sites had some evidence of beaver use was
encouraging.
At some sites, beaver had extensively modified the habitat in the past, but it was not clear if the current
habitat was conduc�ve to beaver occupa�on. It was not obvious therefore that because a given
subwatershed had a lot of past sign that it was necessarily well suited to beaver now.
Habitat Used
Because of the zero-sum methodology of our survey, where a lot of any vegeta�on type necessarily
meant less of others, the effect of conifer was in a way offset by the effects of what appeared to be the
more favorable deciduous vegeta�on types. Occupied sites had fewer conifer, but that was because they
had higher percentages of willow, aspen, and alder. Paying more aten�on to the desirable deciduous
species rather than conifer thus seems appropriate. The general paterns of vegeta�on and physical
habitat we found to be u�lized, as well as the 100 m scale of our surveys, were consistent with the
literature (as summarized in Allen, 1983).
Vegeta�on and Physical Habitat Characteris�cs
The presence of aspen appeared to be important. Anecdotally, beaver seem to occupy many large, wet
meadow-type habitats where aspen doesn’t tend to grow. In those types of habitats willow and alder
15
13
tend to be the dominant woody vegeta�on and they seem to be readily exploited. Current sites with
zero aspen were noted to have mountain mahogany and other minor “woody tree and shrub” species
that were not singled out in this survey that may have served the same purpose as aspen. It may be that
with “enough” willow and alder and other woody riparian vegeta�on, aspen becomes less important.
Aspen didn’t seem to need to be dominant, but at least present. Where it was present, there were
chewed stumps and aspen visible in the dams. We did not account for the size of available aspen and
there’s literature to suggest that smaller aspen are more readily u�lized than very large aspen (Jenkins,
1980). At sites in Babish Gulch for example, it was noted that there was lots of old sign but no new, and
that aspen were very common, but very large. Many “new” beaver sites and their residents were in
proximity to, and consuming, patches of young, small aspen. It may be that long periods without stand-
replacing disturbance are resul�ng in aspen of a size class that is not readily exploitable by beaver, and
that those old stands are preven�ng the development of younger ones.
We tended to find beaver where we found cotonwood which would suggest that condi�ons suitable for
cotonwood overlap those suitable for beaver, but that was not a common habitat type on the Forest.
Our surveys did not take herbaceous or aqua�c vegeta�on into account, which beaver are known to
u�lize (Jenkins, 1981, as cited in Allen, 1983). We assumed that the extent of riparian woody vegeta�on
would be more limi�ng on the Forest than herbaceous vegeta�on.
It’s important to recall that this survey was limited to rela�vely high eleva�on Forest lands, therefore the
“flater, wider” streams on the Forest might have been “steeper and narrower” if compared to those
available on downstream, private lands in the valley floor, for example.
With the excep�on of stream slope, physical habitat seemed less a constraint on beaver than vegeta�on
condi�ons on the Forest. While beaver were found throughout the eleva�on gradient on the Forest,
evidence of beaver was more common in the lower por�ons. While we did assess bankfull- and
floodprone widths and the resul�ng channel confinement, we did not measure valley floor widths.
Wider, flater valleys with more sinuous channels are known to be preferen�ally preferred habitats
(Scrafford et al., 2017). It’s plausible that in our analyses, eleva�on was serving as a proxy for these
wider valley types which are more typical further down the stream network and therefore at lower
eleva�ons. This geomorphic patern to physical habitat is likely to be more relevant than eleva�on per
se.
Grazing Effects
Beaver occupied (lived in) grazed areas more or less propor�onally with their availability. While “past”
sites were slightly more common within grazing allotments, the difference was small enough that it’s not
clear if it represents a patern or not.
It would appear that while grazed areas had less willow, many had enough to be exploitable by beaver.
Willow at unoccupied (“No Beaver”) sites within grazing allotments averaged only 6.8%, which is over
the 5% defined by the HSC, but not by much (figure 2). Also note that occupied sites averaged 16%
willow (table 3), while sites within grazing allotments only averaged 10% (table 7). It’s plausible that the
lessor extent of willow within grazing allotments is a contribu�ng factor to the current restric�on in the
distribu�on of beaver. On the other hand, the 10% average willow available at unoccupied sites might
16
14
represent underu�lized habitat on the Forest, that could support beaver were beaver to colonize those
places.
There were 4% more aspen within grazing allotments. This probably reflects the types of areas that are
made available to catle for grazing (mixed- and open country favored by aspen vs. dense conifer stands),
rather than a causal effect of grazing.
Habitat Availability and Rela�onships at the HUC6 Scale
There were differences in habitat availability between the HUC6s (table 9). In the Lime Creek HUC for
example, only 7% of sites appeared to have beaver habitat as defined by the HSC. None of the surveyed
sites were occupied and only 15% appeared to have ever been. In the Mill Creek-Catle Creek HUC6,
47% of sites were noted to contain beaver habitat as defined by the HSC, 58% of surveyed sites were
currently occupied by beaver, while an addi�onal 21% of sites had signs of past beaver occupa�on. Both
of these observa�ons are consistent with our anecdotal experience in these sub-watersheds, where Lime
Creek seems to have few beavers and Catle Creek many.
It was not surprising that Wilderness management and road density were related (posi�vely and
nega�vely, respec�vely) to observa�ons of beaver sign. The samples sizes of these HUC analyses were
small and conducted with a higher level of uncertainty so these findings may or may not hold up with a
bigger dataset, but both findings are intui�vely appealing. Our methodology of random sampling and
characterizing vegeta�on by aerial imagery within 100 m of specific sites is likely a useful tool to beter
assess riparian condi�ons in Wilderness areas where by defini�on, we aspire to manage for ecologically
appropriate condi�ons of “reference-like” quality. Given some�mes nega�ve effects of extensive beaver
near the road network, concentra�ng our efforts away from roads seems a reasonable approach when
feasible.
Conclusions and Future Work
These data, while helpful, do not explain why we have the beaver popula�on size that we have or why
there are many unoccupied areas that appear habitable. We think vegeta�on has a lot to do with it.
Some of those vegeta�on condi�ons are likely related to past- and current land use prac�ces, while
others may be related to natural succession of vegeta�on types from aspen to conifer, and from small
aspen to large, and how that relates to the fire- and natural disturbance regime. While we have some
es�mate now of the availability of habitat that we can relate to beaver, it remains unknown as to why
beaver aren’t seemingly colonizing all available habitat or where the popula�on is with regard to the
current carrying capacity of the landscape.
We found some sugges�on of a grazing effect on willow availability and we know that willow is
important to beavers. It’s also within the realm of possibility that willow in the types of environments
where grazing occurs are also affected by elk, moose or other wildlife, or some other influence outside
the scope of our surveys. But because permited catle grazing is within the management purview of the
Forest, we believe that the range management regime should con�nue efforts to quan�fy and monitor
the extent of willow within grazing allotments with the aim of increasing it. We know that willow is
important to beaver, that there’s less of it in grazing allotments, and that increasing willow is likely to
contribute towards beaver habitat suitability and stream- and watershed health. This survey also
17
15
provides an example of an analy�cal means to assess the spa�al extent of willow at the site- and
subwatershed scales, and some context to evaluate riparian vegeta�on in terms of beaver habitat
suitability.
Importantly, this work is already furthering efforts to increase riparian habitat suitability on the Forest.
For ease of reference, table 9 is par�ally reproduced below:
HUC6 Surveyed Sites
% HSC
all sites % Past % New Current Avg % Willow Avg % Aspen
Thompson Creek 79 18% 46% 23% 11% 8%
Fourmile Creek 38 24% 61% 3% 8% 17%
Overall Averages 21% 32% 16% 14% 8%
Surveys and subsequent analyses noted that the Thompson Creek and Fourmile HUCs had near-average
rates of habitat suitable to beaver as defined by our HSC, high percentages of “past” sites that indicated
formerly more extensive beaver use, and average percentages of willow and aspen that were
approximately what we’re looking for vegeta�on-wise. Fourmile in par�cular looks under-u�lized (only
3% of sites had “new” or “current” beaver occupa�on) and that may be a reflec�on of its low average
percentage of willow. In part because of this view, the Forest and partners are pursuing a program of
work to increase the spa�al extent of willow and improve instream condi�ons via low-tech process-
based restora�on, and to expand the area of riparian-adjacent aspen via prescribed fire. This work has
also highlighted some areas where that type of work may not be very frui�ul. In the headwaters of the
Fryingpan for example, it would seem that while beaver were once common there, the vegeta�on has
shi�ed from aspen to conifer and therefore that area doesn’t appear to have a lot of poten�al for beaver
in its current state. In the headwaters of Snowmass Creek and Mill Creek-Catle Creek, those areas
appear to be suppor�ng beaver and their habitat and therefore may not need much in the way of
interven�on.
Thank You
We’d like to thank Lisa Tasker, Pitkin County, and the Healthy Rivers Board for funding this work. While
important and needed, it would not have goten done with this level of aten�on without funding from
the Board. Thank you for your support.
18
16
References
Allen, A. W. 1983. Habitat suitability index models: Beaver. US Fish and Wildlife Service. FWS/OBS-
82/10.30. Revised. 20 p. Available online: htps://docs.streamnetlibrary.org/USFWS/hsimbeaver.pdf.
Accessed 16 December 2024.
Marshall, S., D. Bull, M. Lythe, B. Spinner, and M. Chapman. 2024. Beaver Pond Data from the Colorado
Beaver Ac�vity Mapper (Version 1). Colorado Natural Heritage Program, Fort Collins, Colorado, U.S.A.
[producer and distributor] and Lynker Technologies, Boulder, Colorado [producer].
R Core Team (2024). R: A Language and Environment for Sta�s�cal Computer. R Founda�on for
Sta�s�cal Compu�ng, Vienna, Austria. htps://www.R-project.org.
Rosgen, David. 1994. A Classifica�on of Natural Rivers. Catena Vol. 22: 169-199. Available online:
htps://www.sciencedirect.com/science/ar�cle/pii/0341816294900019. Accessed 13 October 2023.
Scamardo, J., S. Marshall, E. Wohl. 2022. Es�ma�ng widespread beaver dam loss: Habitat decline and
surface storage loss at a regional scale. Ecosphere 13(3): e3962. Available online:
htps://doi.org/10.1002/ecs2.3962. Accessed 17 December 2024.
Scrafford, M., D. Tyers, D. Paten, and B. Sowell. 2017. Beaver Habitat Selec�on for 24 Yr Since
Reintroduc�on North of Yellowstone Na�onal Park. Rangeland Ecology and Management 71:266-273.
Available online: https://doi.org/10.1016/j.rama.2017.12.001. Accessed 17 December 2024.
USDA Forest Service. 2002. Land and Resource Management Plan, 2002 Revision, for the White River
Na�onal Forest. Rocky Mountain Region. Available online:
htps://www.fs.usda.gov/Internet/FSE_DOCUMENTS/fsbdev3_000999.pdf.
19
MEETING DATE: January 16, 2025
TO: Healthy Rivers Board FROM: Lisa Tasker and Lisa Macdonald CC: Ry Neiley, County Attorney AGENDA ITEM: Monthly Staff Report Administrative: Attend Town of Basalt POST meeting 1/8/24 WWP presentation
Projects and Infrastructure:
Beaver Project with the USFS Continue to work with USFS on planning future projects. USFS modified collection agreement to River Board in May 2024 and to BOCC for approval May 28 with formal acceptance via resolutions in June. Meetings ongoing. USFS interns had a successful 2024 season on the forest with Clay Ramey’s guidance. Mr. Ramey will attend the January 2025 meeting.
Wild and Scenic Efforts Wild and Scenic Subcommittee calendar invite out for the 4th Thursdays of each month beginning January 23rd. The Instream Flow or “Peaking Instream Flow” Subcommittee met January 9th for it’s 8th meeting with the next meeting planned for February.
Crystal River Restoration Project – Town of Carbondale Partnership on Project Sign design work continues with DHM and all partners to the project. Sam Caudill State Wildlife Area Property still closed through CPW construction estimated work completion in March 2025. CPW awarded road work to Johnson Construction. Johnson has started work. Preparing a request for proposals for the landscape work with bids due March 2025
Lincoln Creek/Ruby Mine Board heard LRE presentation of contract work and updates at Dec. meeting. Whitewater/ Fish Park Streamside Amenities 2024 Staff presented plan to the Town of Basalt’s POST Board 1/8/25 RFP out for streamside work. Bids due 1/10/2025 Bid review week of Jan 13, 2025 Outreach/Education/Grants
• Ads are running on KDNK and KMTS, APR and the Mountain stations. • Aspen 82 (LOCAL) – sign up and topics submitted to Board Dec, 2024.
• 2025 Scholarship information provided to the schools. Kirstin and Andre on sub-committee. Education and Training 2025:
• Colorado Water Congress Winter Conference Denver January 29-31 - Neff in person Ongoing - Staff continues to participate and/or collaborate with the following groups:
20
• Colorado Water Congress State Affairs Committee meetings Mondays 8:00 am when CO State Congress is in session
• Roaring Fork Wildfire Collaborative t as applicable
• Watershed Collaborative Meetings, Board is on listserv and receives updates
• Colorado Basin Roundtable half day meetings every 2 months
• Crystal River Wild and Scenic Core Group, Collaborative, Coalition/Advocacy,
Stakeholder meetings
• Local Beaver group including CPW, USFS, BLM, RFC, Sierra Club, EcoFlight, The Roaring Fork Biodiversity Initiative, National Forest Foundation and others - no scheduled meetings
• Lincoln Creek/Grizzly/Ruby Mine Local working group ongoing Three teams have been formed; Public Comms, Short Term Water Quality Monitoring, Phase II Scope of Work is being created with completion of the Phase I project with LRE Water in December.
National Wild and Scenic Coalition Zoom every two months
• National W&S Rivers Virtual Hill Week is in March
• Meetings on the first Friday every two months Riverscape Restoration Network Meetings with American Rivers and Audubon Rockies–
• Beaver Working Group for the state of Colorado
• Pitkin County Vision 2050 Comprehensive Plan Counties are required by the state of Colorado to adopt comprehensive plans as advisory documents. They provide guidance for land use and other county programs and policies. Staff
• Aspen Parks & Open Space Strategic Plan focus group
21