Mapping of black-tailed prairie dog (Cynomys ludovicianus) colonies using National Agriculture Imagery Program (NAIP) 2015 imagery

Mapping of Black-tailed Prairie Dog (Cynomys
ludovicianus) Colonies using National Agriculture
Imagery Program (NAIP) 2015 Imagery

Prepared for:

Bureau of Land Management, Miles City Field Office

By:

Daniel A. Bachen, Bryce A. Maxell, Alexis L. McEwan, Boaz Crees

Montana Natural Heritage Program
Natural Resource Information System
Montana State Library

March 2017

MONTANA

) Natura sea erty

Mapping of Black-tailed Prairie Dog (Cynomys
ludovicianus) Colonies using National Agriculture
Imagery Program (NAIP) 2015 Imagery

Prepared for:
Bureau of Land Management, Miles City Field Office

Agreement Number:
L11AP20008

Prepared by:
Daniel A. Bachen, Bryce A. Maxell, Alexis L. McEwan, Boaz Crees

4 Montana

SS Fi2t Tary

P.O. Box 201800 « 1515 East Sixth Avenue ¢ Helena, MT 59620-1800 ¢ 406-444-3290

ike MONTANA
Z. Natural Heritage
C Program
t/

© 2017 Montana Natural Heritage Program

This document should be cited as follows:

Bachen, D.A., B.A. Maxell, A.L. McEwan, and B. Crees. 2016. Mapping of Black-tailed Prairie Dog
(Cynomys ludovicianus) colonies using National Agriculture Imagery Program (NAIP) 2015 Imagery.
Montana Natural Heritage Program, Helena, Montana. 18 p.

EXECUTIVE SUMMARY

Black-tailed Prairie Dogs (Cynomys ludovicianus)
alter vegetation and dig extensive burrows,
creating habitat for other species, and serve as
prey for both mammalian and avian predators.
Several animal species of conservation concern
at the state and federal level, including the
Mountain Plover (Charadrius montanus),
Burrowing Owl (Athene cunicularia), and Black-
footed Ferret (Mustela nigripes), are closely
associated with prairie dog colonies and make
use of burrows or the prairie dogs themselves.
Recently, prairie dog populations have been
impacted by sylvatic plague causing colony and
population dynamics to change. The Montana
Conservation Plan for Black-tailed and White-
tailed Prairie Dogs identifies the need to
monitor the distribution of these animals within
the state to aid in the conservation of prairie
dogs and dependent species.

We addressed the need for colony and complex
monitoring by digitizing potential colony
boundaries across Southcentral and Southeast
Montana using 2015 National Agriculture
Imagery Program (NAIP) imagery. In total, we
mapped 4,154 potential colonies covering
556,136 acres. Of these colonies, 4 were over
5,000 acres in size, 55 were between 1,000 and
5,000 acres, and the remaining 3,199 were less
than 1,000 acres. We aggregated colonies into
complexes using the 1.5 km rule, and found that
15 complexes contained at least 5,000 acres of
colonies and met the Category 1 criteria under
the Montana Conservation Plan for Black-tailed
and White-tailed Prairie Dogs.

To test the accuracy of our mapping, we ground
truthed a subset of our colonies to determine if

there was evidence of current or recent
occupancy by Black-tailed Prairie Dogs. Our
ground truthing efforts were based on 1 ha grid
cells overlaid on each colony. We walked the
perimeter of Category 1 complexes and
recorded whether grid cells of the colony were
active, inactive, or no evidence of occupancy. Of
the 56 colonies we ground truthed, 95% had
evidence of current or recent occupancy. At the
grid cell level, we found that cells along mapped
boundaries contained evidence of Black-tailed
Prairie Dogs only 37% of the time while 76% of
cells within the boundaries had evidence.

Our methods appear to have been effective at
identifying colonies, but less precise at
determining the exact boundaries of these
colonies. Some inaccuracies may have resulted
from the time delay between collection of the
NAIP imagery and collection of ground truthing
data. With future improvements in image
quality and resolution our precision and
accuracy will increase, thereby negating or
minimizing the need to ground truth digitized
colonies for future projects.

This project documented relatively large areas
occupied by Black-tailed Prairie Dogs, including
several complexes that may be suitable for
targeted conservation efforts for species such
as Black-footed Ferrets. With the development
of an oral plague vaccine, identification and
monitoring of colonies and complexes will be
increasingly important. Mapping colonies from
NAIP imagery may provide a cost effective way
to prioritize areas for vaccine deployment and
monitor the effects on colony size.

ACKNOWLEDGEMENTS

We extend considerable thanks and appreciation to all who conducted surveys for this project including:
Boaz Crees, Alexis McEwan, and Ellen Whittle. Thanks to Braden Burkholder for assistance with the
Prairie Dog database, Scott Blum for formatting the data and appending it to the Montana Natural
Heritage Program’s central animal observation database, and Bryce Maxell for providing valuable
feedback throughout the project. We thank the Miles City Field Office of the Bureau of Land
Management for providing funding and guidance for this project.

This project was supported by an agreement between the Bureau of Land Management and the

Montana Natural Heritage Program, a cooperative program of the Montana State Library and the
University of Montana (BLM L11AP20008)

TABLE OF CONTENTS

PEOCI UCT oah sceneries cate ea hte alicgs cosa cS depts nin oa Satoh Race a oh dav a Se guce seed vial caece nbd a eantd vuabede Reseahiees 1
PA) =1 64 aL G0 f=) (chen ee eo a 2
MCC OOS serecetised cents ecersk esa deepunc seers vivceaset dec cnaales taeda caus nauduenealsedveatadostuadescet eat deanansteas tosai ee tuates dee ab bventeatiathes 3
Digitization Of Colony BOUNCATIES ............:cccccccccesessessesecececesseseaesecececssseseaeaeeeeecesseseaaeaeceeecessessaaeaeeeesens 3
Ground Truthing of Select Colonies .............ccccesscccececessessnaececeesseesesaeaeeeeeceseeseaaeaeeeeecesseeaaeaeeeeseeeseseaaees 3
Data Preservation and DisS@MiMation .......ceeeeesceesseceececeeeeeesaeeeeaaeceeaeeceaeeeesaeeseaaeseeeeeceaeeeesaeeseaeeseeeees 4
RESUIES 6 a citccaesterie edt Ancitetnebtnas ner dt as ah etnies ace eet ie 5
AY Fe] ©) 9) f= te 5
GrOUNd FrUthin gn cas oecec celine iawsczeveosk hss as sesdvtn sae ct actan tees da oeaad duende cath sivas antetnsh th ibanecdevis lath Beietanteenaees 5
DISCUSSION iesxi setveesscegeeest sceecesssageeeet oaceeevebanevesh staecovesazevs et scaeeesebagevssh staecevssegees et sagveevebauevest staeeeveseteeset cevesesegeeeye. 6
Literature Cited sex sntituct A vecceesuleteti a aiiee atic anes habe es eave eae el a eae ae ee a ee es 9
List of Figures
Figure 1. Example of a mapped COlOny DOUNALY ........cccccccecessesssseeececeseesecnececeeecessessaeseeeeseeesesnsaees 10
Figure 2. Overview of mapped colonies using 2015 NAIP imagery .........:cccccccssssssceceeecessesssteseeeesens 11
Figure 3. Comparison across projects Of Mapped COIONY SIZEC...........:csescccccecsssesscneceeeeecessessnteaeeeesens 12
Figure 4. Map showing all 1.5 km COMPIEXES 00... cccecesesssssececeeeceseessaeeeeeeeeeseeseaaeaeeeeseuseeseaeaeeeesens 13
Figure 5. Complexes that meet Category 1 Criteria ...........cccccsccccecesssssssececececesseseeeseeeeecesseseaeaeeeesens 14
Figure 6. Ground truthed colonies showing proportion with evidence of Black-tailed Prairie Dogs
s ashagbcasnbcdedadeecgdn sabetedeebaneduvsa svkeiv ie lac ety aheGs Go shagedvedal cantenaa caleey ab seule svadeiaedoscdodsoacghesgabebedertanies 15
Figure 7. Proportion of interior and exterior grid cells with active, inactive and no evidence of
Black-tailed Prairie DOBS:..<c:.5. castes lees isc onda te dicta en nntses Ode davnovececs Wtosbivbenectind aes Gacemeea des 16
Figure 8. New colony detections ciccci scot ese ates cess te eee At ls ie 17
Figure 9. Proportion of cells surveyed compared to proportion with activity ............cccccccessessseeees 18

INTRODUCTION

Black-tailed Prairie Dogs (Cynomys ludovicianus)
alter vegetation and dig extensive burrows,
creating habitat for associated species, and
serve as prey for both mammalian and avian
predators. Although this species is still
widespread throughout much of its historic
range in Montana, sylvatic plague and
anthropogenic factors have reduced average
colony size and changed colony dynamics
(Augustine et al. 2008a), with concomitant
impacts on dependent species such as
Mountain Plover (Charadrius montanus),
Burrowing Owl (Athene cunicularia), and the
federally Endangered Black-footed Ferret
(Mustela nigripes) (Augustine et al. 2008b,
Desmond et al. 2000, Matchett et al. 2010). To
address the conservation needs of prairie dogs
within the state, the Montana Prairie Dog
Working Group released a conservation plan in
2002 to identify objectives for conservation
planning, including quantifying distribution and
abundance (MPDWG 2002).

Since the 1980’s several projects have sought to
quantify the area occupied by both Black-tailed
Prairie Dogs and White-tailed Prairie Dogs
(Cynomys leucurus) within the state (MPDWG
2002). Over the last decade, the Montana
Natural Heritage Program (MTNHP) has mapped
colonies using the 2005 and 2009 National
Agriculture Imagery Program (NAIP) imagery
(Maxell et al. 2010). Mapping from aerial
imagery can be a relatively cost efficient way to
identify colonies, particularly on inaccessible
public and private lands. However, two major
shortcomings exist when these methods are
used. First, colony boundaries are mapped on
the presence of burrows; therefore, it is not
possible to distinguish between active and
inactive colonies or active and inactive areas
within colonies. Second, features similar to
burrows may confound colony boundaries or
cause colonies to be mapped in areas without

Black-tailed Prairie Dogs. The 2005 and 2009
NAIP projects were 69% accurate at identifying
colonies and 80% accurate at identifying cells
within these colonies in areas of the state
where colonies are easily distinguished from the
background vegetation, geological, and
biological features (Maxell et al. 2010).
However, the 2005 NAIP imagery delineated
colonies in Custer, Phillips, and Rosebud
counties, had ground truthed accuracies of 51%,
52%, and 35% respectively for the proportion of
cells within a colony that had evidence of Black-
tailed Prairie Dogs. The low accuracies in these
counties may be caused by barren areas that
can superficially resemble the structure of
prairie dog colonies. Accuracy of mapping at the
colony level was also relatively inaccurate for
these same counties, with only 50%, 46%, and
32% of digitized colonies containing evidence of
occupancy. These inaccuracies are problematic
as these counties overlap core areas for this
species within the state. Therefore, better
information about colony coverage in these
areas is necessary as conservation and
management of Black-tailed Prairie Dog
colonies is essential for the conservation of a
number of Montana Species of Concern, and
directly effects reintroduction efforts for the
Endangered Black-footed Ferret.

Montana’s Conservation Plan for Black-tailed
and White-tailed Prairie Dogs (MPDWG 2002)
identifies two objectives that can be address
with mapping colonies from aerial imagery.
Objective #2 is to “develop statewide and
regional prairie dog distribution and abundance
standards”. To satisfy these objectives the plan
calls for inventory and monitoring of
distribution and abundance and identification
of complexes of colonies. While abundance
within colonies cannot be addressed with aerial
mapping, this technique is well suited to
quantifying distribution. Objective 3C: “identify

isolated prairie dog colonies in need of special
consideration, assess their needs, and
implement special management tasks, as
appropriate” can be addressed.

The plan also identifies specific goals for the
conservation of Black-tailed and White-tailed
Prairie Dogs and associated species. Specifically,
three categories are used to describe colony
complexes (MPDWG 2002; pages 15-16):

Category 1: A minimum of two Black-tailed
Prairie Dog complexes sufficient to maintain
viable populations of Black-footed Ferrets.
These should be at least 100 km apart, with
each encompassing at least 5,000 acres of
Black-tailed Prairie Dogs

Category 2: A total of 36,000 acres occupied
by Black-tailed Prairie Dogs, composed of at
least 20 complexes of at least 1,000 acres

Category 3: Complexes less than 1,000 acres
in size... plus scattered isolated colonies of
any acreage

To address these needs we mapped colonies of
Black-tailed Prairie Dogs using the 2015 NAIP
imagery across some or all of Golden Valley,
Stillwater, Musselshell, Bighorn, Yellowstone,
Treasure, Rosebud, Prairie, Powder River,
Carter, and Custer counties. To determine if the

improved resolution of the 2015 NAIP imagery
relative to earlier NAIP imagery impacted the
accuracy of identifying colonies in previously
problematic areas, we intensively ground
truthed colonies on accessible public lands in
central Custer County as well as confirming the
presence of Black-tailed Prairie Dogs at other
colonies across the rest of the mapped area.

PROJECT GOALS

Our primary goals for these efforts were to:

Use NAIP imagery to map areas with recent
evidence of Black-tailed Prairie Dog activity

Evaluate accuracy of this methodology
through ground truthing of 1-hectare grid
cells

Identify complexes for management and
potential Black-footed Ferret reintroduction
using a 1.5 km separation rule for defining a
complex

Confirm the presence of Black-tailed Prairie
Dogs at colonies outside of the boundaries
of historic colonies

Provide data products to partners for
management actions, planning, ground
truthing, and environmental reviews

METHODS

Digitization of Colony Boundaries

Colony boundaries were delineated on the 2015
NAIP 20 kilometer x 20 kilometer tiles following
methods used for the previous projects (see
Maxell et al. 2010). However, we chose to
delineate colony boundaries precisely instead of
classifying 1-hectare grid cells as occupied. As
the area of coverage was relatively small, and to
reduce observer bias, a single technician
conducted all of the mapping. All digitization
was conducted using Arc Map 10.4 © ESRI
Software. We examined the 2015 NAIP imagery
at a scale of approximately 1:3,000 for features
found within colonies such as mounding of dirt
outside of burrow entrances, and clipped
vegetation. When a potential colony was
identified, we digitized the boundary as a
polygon feature in a File Geodatabase (Figure
1). The colony boundary was then reviewed
with previously mapped colonies and point
observations overlaid, and any additional areas
of occupancy detected on the 2015 NAIP
imagery were mapped. Polygons were coded as
having been detected blind, with the aid of
previously mapped colonies, or reported
observations.

Ground Truthing of Select Colonies

To determine the accuracy of colony boundaries
delineated using the 2015 NAIP imagery, we
ground truthed a subset of mapped colonies. All
fieldwork was conducted over 30 person/days
in September 2016. Our goals for this effort
were to quantify the accuracy of mapped
colonies on two levels:

e Colony: Did the mapped colony contain
evidence of Black-tailed Prairie Dogs?

e Within colony: What was the
proportion of area with evidence of
prairie dog activity?

We prioritized ground truthing of colonies with
high value for conservation of associated
species, in particular Black-footed Ferrets. To
sustain a ferret population, large aggregations
of prairie dog colonies are required (Biggins et
al. 1993). To ensure that our ground truthing
results were most accurate within high priority
areas we allocated the most effort to colonies
within the largest identified complexes.

To create colony complexes (hereafter
complexes), we buffered all colonies by 750 m.
We then grouped colonies by overlap between
these buffers to identify complexes of colonies
within 1.5 km of each other. Finally, we
calculated the total size in acres of mapped
colonies within each complex. We focused our
ground truthing efforts on complexes meeting
the Category 1 criteria (i.e. > 5,000 acres of
active colonies). To facilitate surveys, we
overlaid a public lands layer on all colonies
within the Category 1 complexes and prioritized
publically accessible areas for survey.

Depending on the priority of each colony we
used one of two methods. For high priority
colonies in Category 1 complexes, we overlaid a
1-hectare x 1-hectare grid and identified cells in
proximity to the mapped boundary. Field
personal used GPS units with hectare centroids
pre-loaded and hard copy maps with the
hectare boundaries to identify each grid cell
within and adjacent to colonies. They then
walked the colony boundary and scored
whether each hectare grid cell contained active
burrows, inactive burrows, or no evidence of
occupancy. For colonies that were either within
these Category 1 complexes and visible, but not
accessible or outside of complexes, we sought
only to determine if there was evidence of
prairie dog activity at the colony level.
Personnel traveled as close as possible to these
colonies and looked for evidence of Black-tailed

Prairie Dogs or their burrows. As any inference
of absence using this survey method would be
untrustworthy due to incomplete survey
coverage, we only recorded if the colony was
active or inactive.

Data Preservation and Dissemination

To archive and make project data available to
land managers for use in conservation efforts
and land use planning, we added the mapped
colony boundaries and observation information
to the Montana Natural Heritage Program’s

Prairie Dog and Point Observation databases.
Spatial data containing both the mapped colony
boundaries and ground truthing was appended
to Montana’s statewide Prairie Dog Database
and information from this database is available
upon request. The observations of active Black-
tailed Prairie Dogs and other incidental species
observed during surveys were appended to the
MTNHP’s point observation database, and are
available upon request and through online tools
in the Natural Heritage Map Viewer web
application.

RESULTS

Mapping

Across the area of interest, we mapped the
boundaries of 4,154 potential colonies totaling
556,136 acres (Figure 2). Of these colonies, 4
were over 5,000 acres in size, 55 were between
1,000 and 5,000 acres, and the remaining 3,199
were less than 1,000 acres. Mapped colonies
averaged 134 acres with a minimum size of less
than 1 acre and a maximum size of 9,210 acres.
As in previous years, the majority of colonies
were between 11 and 50 acres (Figure 3). When
aggregated into complexes using the 1.5 km
rule, we identified 942 discrete complexes. Of
these complexes, 15 were greater than 5,000
acres in size and met the criteria for Category 1
designation (MPDWG 2002). An additional 53
complexes were between 1,000 and 5,000 acres
in size, and the other 874 were less than 1,000
acres (Figures 4 & 5). The smallest complex was
approximately 0.5 acres while the largest
exceeded 100,131 acres and encompassed 472
mapped colonies. As expected, many mapped
colonies were associated with previously
documented colonies, but 2,577 mapped
colonies were completely outside of those
previously mapped. The total coverage of these
new areas was 214,610 acres.

Ground Truthing

In total, we ground truthed 5,298 hectares in 56
prairie dog colonies within three Category 1
complexes (Figure 6). Of these colonies, 53 had

evidence of prairie dog activity (95%). Across
the ground truthed colonies, we surveyed 5 or
more hectares within 48 of these colonies. Of
these, 74% of the hectare grid cells that were
interior to the grid cells along the colony
boundary had evidence of prairie dog activity
(active or inactive). However, for grid cells that
overlapped the mapped colony boundary, only
37% had evidence of current or recent
occupancy (Figure 6). The proportion of interior
grid cells with evidence of activity varied across
colonies (Figure 7). Of the 3 colonies where no
prairie dog activity was recorded, all were
within 5 km along the Powder River, and 1 had
less than five cells surveyed. In addition to the
intensive ground truthing of these primary
colonies within Category 1 complexes, we also
documented evidence of prairie dog occupancy
at an additional 261 colonies (Figure 8). Of
these colonies, 193 did not overlap previously
mapped colonies.

Although we did not standardize the number of
cells surveyed for each colony, surveying a
greater proportion of cells within a colony did
not affect the proportion of interior cells with
evidence of activity (Figure 9). The apparent
lack of relationship between cells surveyed and
detections indicates that we did not bias our
detection of Black-tailed Prairie Dogs or
burrows through increased effort at some
colonies.

DISCUSSION

Mapping prairie dog colonies using the 2015
NAIP imagery was successful both in the
number of colonies delineated and the accuracy
of mapping colonies. Total acres of active and
inactive colonies exceeded those mapped in
previous years within the same area. Although
this could represent an increase in the area
occupied by Black-tailed Prairie Dogs, this could
also be due to more colony detections as a
result of better imagery. Although we detected
more colonies, using the 1.5 km separation rule
we delineated many of the same complexes
that were delineated using the 2005 and 2009
NAIP imagery (Maxell et al. 2010). As was the
case with previous efforts, the complex in
central and southern Custer County centered on
the Pumpkin Creek area was one of the largest
in the mapped area.

Compared to previous efforts in this part of the
state, identification of colonies from the 2015
NAIP imagery appears to have been more
accurate. However, methods used to ground
truth colonies for this project differed from
previous years, so the metrics are difficult to
compare directly. Both this and previous efforts
used 1-hectare grid cells as the sampling unit.
However, in past years, cells were sampled
randomly from within the colony and each cell
was surveyed with greater intensity. For this
effort, we sought to survey all cells along the
boundary of the mapped colony, and any others
within the colony should they be visible or the
opportunity arise. While surveys of each grid
cell were less intense, more cells were
surveyed.

Our analyses of the data also differed from
previous efforts. For this effort, we calculated
the proportion of cells with evidence of
occupancy on the boundary and in the interior
of the colony. These proportions differed, with
the interior cells being more accurate than

those on the boundary. Whether due to clearer
imagery or different survey methods, the
proportion of mapped colonies that contained
evidence of Black-tailed Prairie Dogs increased
over previous efforts. Ground truthing showed
that the mapping effort that used the 2005
imagery was correct at only 28 of the 56
colonies in Custer County. Our accuracy was
much higher, with 53 of 56 colonies having
observed evidence of occupancy by prairie
dogs. Additionally, all three colonies that were
mapped incorrectly are clustered within 5 km of
each other, which may indicate challenging
habitat or an isolated error in mapping.

Due to differences between efforts in
techniques used to ground truth colonies, we
cannot directly compare accuracy at the 1-
hectare level. However, the increase in number
of mapped colonies currently or previously
occupied by Black-tailed Prairie Dogs provides
evidence that this project was more successful
at correctly identifying recent evidence of Black-
tailed Prairie Dog activity. The 2005 NAIP
imagery mapping effort reported that 51% of
cells in Custer County were mapped incorrectly
(Maxell et al. 2010). These protocols also
included cells outside of colonies, which we did
not test with our methods. We found that 74%
of cells within the mapped interior of the
colonies were observed to be active or inactive,
which appears greater than the 2005 effort.
However, the cells mapped at the boundary
were only correctly identified 37% of the time.
One reason our boundary accuracy may have
been low was the temporal separation between
the 2015 imagery and the ground truthing in
2016. The spatiotemporal movement of plague-
affected colonies, like those in our study area, is
more dynamic and accelerated than colonies
not impacted by the disease (Augustine et al.
2008). Therefore, it is likely that the boundaries
of colonies changed to some degree in the time

between when the imagery was taken and the
ground truthing was performed. We would
expect mounds and other evidence of previous
occupancy under this scenario to be visible.
However, if these indicators are less detectable
than our expectations, this may contribute to
the inaccuracy of the mapped boundaries.

Because we did not ground truth a subset of
grid cells mapped outside of colony boundaries,
we cannot assess the degree to which our
mapping efforts missed colonies that may have
existed within the area. We did observe Black-
tailed Prairie Dogs outside of mapped
boundaries during ground truth efforts,
indicating that this error existed to some
degree. Given this and the high accuracy with
which we identified colonies, the number of
mapped colonies should be considered an
underestimate of the true number of colonies
present.

We prioritized assessing the boundary and
occupancy at the colony level over assessing
occupancy at the burrow level. Consequently,
using our methods makes it difficult to extract a
total number of acres that have evidence of
current or recent occupancy by Black-tailed
Prairie Dogs within the mapped colonies. The
area within colonies mapped using aerial
imagery is positively biased if the circumscribed
boundary is assumed to delineate an active
colony (McDonald et al. 2011). With this
assumption, the 556,136 acres of mapped
colonies should not be interpreted as an
estimate of active or inactive colonies. As we
did not attempt to estimate the proportion of
the colony that was occupied, we cannot infer
the actual number of acres of Black-tailed
Prairie Dogs present.

The increase in accuracy of detecting colonies
between this project and previous efforts may
be due to differences in imagery, personnel,

survey techniques, or the intrinsic differences
between years. Although the resolution of the

NAIP imagery was 1 m for all projects, the 2015
images appeared to be much higher quality
than the 2005 or 2009 NAIP imagery. We feel
that this is the most likely cause of increased
accuracy. As stated earlier, we also changed our
ground truthing protocols and although
unlikely, accuracy differences may have
resulted due to these differences in protocols.

In 2018, the USDA is expected to release the
2017 NAIP imagery at 0.5-meter resolution
(https://catalog.data.gov/dataset/national-
geospatial-data-asset-ngda-naip-imagery-2017-

2018-planned-acquisition). We anticipate that
increased resolution will allow increased

precision and accuracy of mapped colonies,
potentially making ground truthing
unnecessary. Future projects that use higher
resolution imagery should seek to test the
accuracy of detecting areas with and without
evidence of prairie dog occupancy to help
inform the need for future ground truthing.

The dynamic nature of colonies affected by
plague has presented a challenge for this and
other efforts that have used NAIP imagery to
identify prairie dog colonies due to the time it
takes for imagery to be processed and mapping
to be conducted. Although it is unlikely that
evidence of a colony would disappear between
the time the area was flown and the colony was
digitized, the likelihood of changes to the
colony increase with time. This is especially
problematic when trying to assess the accuracy
of mapping boundaries. Although the time it
takes for the USDA to release the NAIP imagery
is beyond our control, we did demonstrate that
it is possible to map colonies in a focal area
relatively quickly. In total, it took us a little over
a month to map our area of focus. For future
work that requires geospatial status and
distribution assessments of Black-tailed Prairie
Dog colonies, mapping from recent NAIP
imagery can provide this information relatively
quickly.

This project documented relatively large areas
occupied by Black-tailed Prairie Dogs, including
several complexes that may be suitable for
target conservation efforts for species such as
Black-footed Ferrets. With the development of
an oral plague vaccine (Rocke et al. 2010),
identification and monitoring of colonies and

complexes will be increasingly important. These
mapping methods may provide an efficient way
to identify complexes large enough to support
ferrets and efficiently monitor the landscape
level effects of a large-scale vaccination
program.

LITERATURE CITED

Augustine, D. J., M. R. Matchett, T. P. Toombs, J.
F. Cully Jr, T. L. Johnson, and J. G. Sidle. 2008.
Spatiotemporal dynamics of Black-tailed
Prairie Dog colonies affected by plague.
Landscape Ecology 23(3): 255-267.

Augustine, D.J., S.J. Dinsmore, M. B. Wunder,
V.J. Dreitz, and F. L. Knopf. 2008. Response of
Mountain Plovers to plague-driven dynamics
of Black-tailed Prairie Dog colonies.
Landscape Ecology 23(6): 689-697.

Biggins, D. E., B. J. Miller, L.R. Hanebury, B.
Oakleaf, A. H. Farmer, R. Crete, and A. Dood.
1993 A technique for evaluating Black-footed
Ferret Habitat. U.S. Fish and Wildlife Service
Biological Report 93(13) 73-88.

Desmond, M. J., J. A. Savidge, and K. M.
Eskridge. 2000. Correlations between
Burrowing Owl and Black-tailed Prairie Dog
declines: A 7-year analysis. The Journal of
Wildlife Management 64(4): 1067-1075.

Matchett, M.R., D. E. Biggins, V. Carlson, B.
Powell, and T. Rocke. 2010. Enzootic plague
reduces Black-footed Ferret (Mustela
nigripes) survival in Montana. Vector-Borne
and Zoonotic Diseases 10(1): 27-35.

Maxell, B.A., S. Blum, and K.V. Walker. 2010.
Preliminary Report: Mapping Black-tailed

Prairie Dog (Cynomys Iudovicianus) colonies
across Montana using the National
Agriculture Imagery Program (NAIP) 2005
imagery. Report to the Miles City Field Office
of the Bureau of Land Management and the
Nongame Program of the Montana
Department of Fish, Wildlife, and Parks.
Helena, MT: Montana Natural Heritage
Program. 27 pp. plus an appendix.

Montana Prairie Dog Working Group [MPDWG].

2002. Conservation plan for Black-tailed and
White-tailed Prairie Dogs in Montana. Helena,
MT: Montana Department of Fish, Wildlife,
and Parks. 51 p.

McDonald, L.L., T. R. Stanley, D. L. Otis, D. E.

Biggins, P. D. Stevens, J. L. Koprowski, and
W. Ballard. 2011. Recommended methods for
range-wide monitoring of prairie dogs in the
United States: U.S. Geological Survey
Scientific Investigations Report 2011-5063, 36
p.

Rocke, T.E., N. Pussini, S. R. Smith, J.

Williamson, B. Powell, and J. E. Osorio. 2010
Consumption of baits containing raccoon pox-
based plague vaccines protects black-tailed
prairie dogs (Cynomys ludovicianus). Vector-
Borne and Zoonotic Diseases 10: 53-58.

FIGURES

Figure 1. 2015 NAIP imagery with a potential colony boundary digitized and 1-hectare grid overlaid.

10

GARFIELD

FERGUS \

ry ee SUA 5
=
WHEATLAND - 2
GOLDEN ante , 5)

SWE

CARTER

ne
bre Pt POWDER RIVER

Montana State Library
Montana State Library

Figure 2. All colonies mapped from the 2015 NAIP imagery are shown in blue. Note that the highest density of colonies are in
central and southern Custer County.

11

0.6

™ NAIP_2005
— NAIP_2009
O NAIP_2015

Proportion of Colonies
0.2 0.3 0.4 0.5

0.1

0.0

<10 11-50 51-100 101-500 501-1000 >1000

Acre Class

Figure 3. The proportion of colonies within 6 size classes for previous projects using the 2005 and 2009 NAIP imagery, and the
current project which used the 2015 NAIP imagery. Proportions are used to compare previous efforts (range-wide)
with the current effort (limited to 11 counties).

12

GARFIELD

FERGUS

GOLDEN VALLE®
. v

STILLWATER

BIG HORN
POWDER RIVER

CARTER

Montana State Library
Monlana Slule Library

Figure 4. Complexes of mapped colonies using the 1.5 km separation rule that are greater than 5,000 acres (green), 1,000 to
5,000 acres (yellow) and less than 1,000 acres (red).

13

GARFIELD
FERGUS

GOLDEN VALLEY

fe}

STILLWATER

POWDER RIVER

CARTER

Montana State Library

Montana State Library

Figure 5. All Category 1 complexes (> 5,000 acres) identified using the 1.5 km separation rule. Note that the colonies
themselves are not shown, only the 750-meter buffer around each.

14

1.0

Cell Type
© Interior
foe) A Exterior
|)
©
(=)
Tt
oS
N
So
°o
oO

Active Inactive No Evidence

Figure 6. The proportion of 1-hectare grid cells with active prairie dogs observed, inactive burrows, or no evidence of current or
historic prairie dog occupancy; 95% confidence intervals shown for each point. All colonies used in this analysis had at
least 5 grid cells of each type surveyed. In total, 48 colonies were used to calculate statistics for interior cells and 53
for exterior.

15

yf RS

Miles City

Interior Cells
with Evidence of
Prairie Dogs

0-0.25
0.25 -0.5
: 0.5 - 0.75
ee ?# 0.75 - 1.0
~~ ROWDER. RIVER (%
» Not Sampled
oP Montana State Libra
na State Library

Figure 7. Ground truthed colonies within Category 1 complexes in central and southern Custer County. The proportion of
interior grid cells that had evidence of active or inactive prairie dog colonies are indicated by the color of the colony.

Note the three colonies shown in red and located south of Red Knob are the only colonies to have been delineated
that did not contain evidence of current or recent prairie dog occupancy.

16

J
pAVEON

GARFIELD

FERGUS
PRAIRIE

STILLWATER? 4
aa CARTER

POWDER RIVER

Montana State Library

Montana State Library

Figure 8. Colonies detected during ground truthing that did not overlap previously known colonies (red). These may represent
new colonies created since the 2009 NAPI effort, or colonies that were overlooked during previous efforts.

17

7.0

0
a = |
e
g
a
a ©
. 2
oO
c
oO
3
>
Ww
s 5
= oO
0
=
Oo : :
oN Colony Size (ac)} |
e* © 1-100
= 101-1000
4 >1000
=|
=

0.0 04 0.2 0.3 0.4 0.5 0.6
Prop. Surveyed

Figure 9. The proportion of cells surveyed plotted against the proportion of these cells with current or historic activity
documented during ground truthing. Note that there is no pattern in each of these metrics for any of the colony size
classes, indicating our results were likely not biased due to uneven survey effort.

18