Monitoring Program for Pygmy Rabbits

Development and Evaluation of a Monitoring Program for Pygmy Rabbits


The recent petition for Endangered Species Act (ESA) listing for pygmy rabbits was, in part, based on a perceived decline in the species, however, data to evaluate this supposition are not available. Efforts during the past 2-3 years have documented numerous new occurrences of the species in Idaho, which have helped to fill out the statewide distribution of pygmy rabbits. However, it is not known if populations of pygmy rabbits fluctuate or cycle, as documented in other lagomorphs, and some observations suggest that populations may shift across a landscape over time. Therefore, an understanding of population trends over time requires information on changes in both abundance and distribution. This work addressed the first of these population criteria.
 

Monitoring burrow systems over the past 7 years in the Lemhi Valley has documented marked fluctuations in density of active burrows, which likely reflect fluctuations in population density of rabbits. Although burrow entrance counts are commonly used to estimate population abundance for semi-fossorial mammals, this relationship has not been evaluated for pygmy rabbits. Our work investigated the relationship between density of burrow systems and density of rabbits, and used this information to evaluate an index of rabbit abundance that could be employed by wildlife biologists to monitor changes in abundance of pygmy rabbit populations over time.


Objectives

The purpose of this research on the INL Site was to develop and evaluate a standardized method to monitor abundance of pygmy rabbits

 

Specific objectives were to:

  • Calibrate an index of abundance based on burrow systems by correlating the index with estimates of population density

  • Design standardized protocols for monitoring abundance.

 

Methods and Results
 

Project Design
In the Lemhi Valley, 2 sites were delineated during fall 2007 and 4 additional sites were delineated during fall of 2008. Of the 6 total sites in the Lemhi Valley 5 were approximately 100 ha in size and 1 was approximately 50 ha. A census of all burrow systems and mark-resight surveys were completed on all 6 sites. Census of burrow systems provide an evaluation of the density and activity status of rabbit burrows, and mark-resight surveys provide estimates of the abundance of rabbits. A second method for estimating abundance based on observations of tracks at burrow systems immediately following snow also was used. Mark-resight and snow-tracking techniques were used to evaluate and calibrate an index based on burrow systems.

 


On the INL Site, with the cooperation of WCS, 24 16-ha sites were censused for burrows during 2008 in addition to a larger 100 ha site delineated during fall 2007. Mark-resight and snow surveys were used to estimate abundance on the large site, however, snow-track surveys were used as the sole method of estimating abundance on 16-ha sites.

 


Burrow Censuses
A complete census of burrow systems was conducted on all sites during autumn of 2007 and 2008. For each burrow system, GPS locations and the number of burrow entrances were recorded, pellets were collected at a selection of active burrow systems for species confirmation, and each system classified based on sign/activity as described by Sanchez et al. (2009) modified from Roberts (2001). This classification system ranks burrows into four classes of activity status: active, recently active, old, very old.  A map of burrow systems was then compiled using ArcGIS 9.


As expected, systematic censuses indicated a difference in the number of burrow systems in each activity class and the density of burrow systems on each site. For the large sites, Cedar Gulch had the lowest total number of burrow systems (131) and the lowest density of burrow systems overall, while Warm Springs had the largest number of burrow systems (748) and the highest density burrow systems. When we compared the ratios of the densities of active and recently active burrow systems to old and very old, it was apparent that there was more current activity on some sites and possibly more past activity on others. These differences might indicate that current habitat conditions are not as favorable, that the population has declined for some other reason, or that the population might have shifted over time.

 

Table 1. Criteria used to assess activity status of pygmy rabbit burrows (Sanchez et al. 2009, adapted from Roberts 2001).  This system has been used at the Lemhi Valley study sites since 2002.

 

Criteria

 

#1 (Active)

#2 (Recently    Active)

 

#3 (Old)

 

#4 (Very Old)

Burrow Entrances:

intact/open

X

X

X

 

intact/debris

 

 

X

 

collapsed

 

 

 

X

Fecal Pellets:

fresh

X

 

 

 

old/weathered

 

X

 

 

absent

 

 

X

X

Digging/Tracks:

fresh/abundant

X

 

 

 

absent/old/few

 

X

X

X

 

On the smaller plots on the INL Site, a similar trend was documented, with sites containing differing ratios of density of burrows in each activity class.  Again, when separated out by density of active and recent versus old and very old burrow systems, some sites contained more active burrows and fewer older ones, while other sites exhibited the opposite trend. 

 

Table 2. Summary of active and recently active burrow systems and burrow densities from burrow censuses conducted on 24 16-ha plots Sept-November 2008.

 

 

 

Site ID

 

Active Burrow Systems

Recently Active Burrow Systems

Total Active and Recent Burrows

 

Burrow Density (Active/ha)

 

Max Number Rabbits

 

Rabbit Density (Rabbits/ha)

Index Predicted Density (Rabbits/ha)

O

0

0

0

0.00

-

-

0

A

2

7

9

0.13

-

-

         -

D

0

0

0

0.00

-

-

0

2

4

8

12

0.25

-

-

0.03

60

3

6

9

0.19

-

-

0.01

66

0

2

2

0.00

-

-

0

111

7

13

20

0.44

-

-

0.12

112

0

10

10

0.00

-

-

0

141

5

19

24

0.31

-

-

0.07

163

9

17

26

0.56

-

-

0.16

174

3

6

9

0.19

-

-

-

K

0

2

2

0.00

0

0.00

0

18

0

1

1

0.00

0

0.00

0

63

0

9

9

0.00

0

0.00

0

84

0

7

7

0.00

0

0.00

0

L

6

12

18

0.38

3

0.19

0.10

37

17

6

23

1.06

1

0.06

0.26

50

6

18

24

0.38

1

0.06

0.10

51

10

14

24

0.63

0

0.00

0.18

99

7

20

27

0.44

0

0.00

0.12

104

12

26

38

0.75

3

0.19

0.21

114

39

51

90

2.44

2

0.13

0.39

115

10

18

28

0.63

2

0.13

0.18

226

1

0

1

0.06

0

0.00

 -

 

Trapping and Radio-collaring

Trapping was conducted from 4-14 days at larger sites in the Lemhi Valley and on the INL Site. At sites in the Lemhi Valley, a visual search and chase technique was the primary method used to capture animals in both seasons, however due to low success of this technique on the INL Site other methods were employed.  Additional techniques included use of drift fences, spotlighting, and placement of traps in active locations during daylight hours.  Captured animals were fitted with 4.2 g radio transmitters (Holohil Inc., Toronto), and we implanted PIT tags for animal identification and collected standard mammalian measurements (weight, hind foot, ear length). 

 

Trapping in the Lemhi Valley resulted in 1-24 rabbits collared per site with a total of 79 rabbits collared over 2 seasons.  The day after capture we located rabbits to visually check collar fit.  It was observed that several animals made relatively large movements (100-300 meters) after capture and that they returned to an area close to their capture within a few days.  For this reason, resight events started at least 2 days after the last trapping event.

 

Trapping effort at Atomic City during 2007 yielded only one animal captured.  The visual technique used in the Lemhi Valley proved to be inefficient; therefore, other techniques were attempted, but without success.  The one rabbit that was radio-collared remained in the same complex of burrows where it was captured.

 

Mark-resight Surveys

Upon completion of trapping events, mark-resight surveys commenced.  Animals were resighted by using maps of burrow locations and GPS to navigate to all active and recently active burrow systems as determined from the previous burrow censuses.  Using this technique allowed us to maximize resight probabilities of all animals.  When an animal was sighted, we recorded (1) the presence of a mark (i.e., radio collar); (2) relative location of the rabbit (at a burrow system, under sagebrush, etc.); and (3) GPS location of sighted rabbit.  A portable receiver was used to identify whether a rabbit was collared and to record which rabbit (if collared) had been located.  Other measurements were taken at each resighting occasion from approximately the center of the study site; these included weather (categories: rain, snow, sunny, overcast), temperature, snow cover (4 cover classes), wind (categorized into 4 classes), and date/time.  Upon completion of each resight occasion, all collared animals that were not detected were located to determine if they were onsite for survey.  Resight occasions were completed for 5-19 occasions per site.

 

Mark-resight surveys in the Lemhi Valley and INL Site yielded from 0-136 resights of collared rabbits per site and 0-80 uncollared rabbits per site over the course of 5-18 resight occasions.  On several occasions 1-3 rabbits were documented offsite and thus not available for resight.  Also, there were a few mortalities between the resight occasions on Cedar Gulch, Warm Springs, and MID.  Animals that were either offsite or had died were removed from the pool of available marked rabbits for calculations of population estimates.

 

Snow-track Surveys

Snow-track surveys were conducted by using maps of burrow systems and GPS locations of active and recently active burrows (as determined during earlier burrow censuses).  Observers navigated back to these sites within 1 day after fresh snowfall.  Rabbit occupancy was determined by presence of rabbits, tracks, and/or evidence of digging at burrow entrances.  On the larger sites where rabbits were radio-collared, estimates of the number of rabbits were adjusted for any collared rabbits that were not detected because they were in a burrow and had not yet emerged, or they occupied a burrow that was not categorized as active or recently active.   On sites without radio-collared rabbits, the numbers of rabbits documented via snow-track surveys were used as the total number of rabbits onsite.   

 

A limited number of snow surveys were completed for large sites during 2007 and 2008, and less than ½ of the smaller 16-ha plots were surveyed in winter during 2008.  Two preliminary snow-track surveys were completed on Cedar Gulch and Rocky Canyon, 2 surveys on Atomic City, and 3 surveys on Bull Creek and Rocky Canyon.  Additionally, 9 of the 22 smaller 16-ha sites on the INL Site were surveyed 3 times each.

Snow-track surveys on the larger sites, Cedar Gulch, Rocky Canyon, Atomic City, Bull Creek, and Ten Mile, resulted in estimates ranging from 2-29.2 rabbits per site.  On all sites we had 100 percent detection rate for rabbits that had emerged from burrows; however, on the Cedar Gulch and Rocky Canyon sites, we had 1-2 radio-collared rabbits per survey that were located in burrows that were not identified as active or recently active, and thus were used to adjust our estimates for total rabbits on site.


 

Table 3. Numbers of rabbits estimated via snow-track survey results on the large study sites in the Lemhi Valley and on the INL, and 16-ha plots on the INL completed during winters 2007-2009.

Study Area

Site/Plot

Survey 1

Survey 2

Survey 3

Lemhi

BC

2

2

3

Lemhi

TM

5

4

4

INL

AT

13

10

-

Lemhi

CG

21.9

-

-

Lemhi

RC

29.2

-

-

INL

K

0

0

0

INL

63

0

0

1

INL

84

0

0

2

INL

18

0

0

0

INL

L

2

2

3

INL

37

1

1

1

INL

50

1

1

0

INL

51

0

0

0

INL

99

0

0

3

INL

104

2

2

3

INL

114

2

1

1

INL

115

1

2

0

INL

226

0

0

0

 

On the smaller 16-ha sites on the INL Site, results of surveys yielded from 0-3 rabbits per plot per survey.  Across several of the surveys, the number of rabbits varied from 0-3, so we used the maximum number of rabbits recorded onsite during the 3 replicate surveys as the number of rabbits per site for that plot.  Plots with the highest number of active burrows were not consistently rated as those with the largest number of rabbits from the snow-track surveys (Table 2).  Maximum numbers of rabbits onsite were then converted into a density estimate for each 16-ha plot (Table 2).

 

Index Development

We used the density of active burrows and the density of rabbits from the larger sites on Lemhi Valley and the one large site on the INL Site to develop an index of abundance for pygmy rabbits.  For additional details see Price (2009).  We found that a curvilinear relationship fit the data best and the resulting regression is shown in Figure 8.  We fit a curved relationship rather than a linear one because we documented that individual rabbits used more burrow systems as the density of burrow systems available to them increased, and therefore, rabbit density would be expected to increase at a slower rate as burrow density increases. 

Relationship between density of pygmy rabbits and density of active burrow systems at 7 sites in east central Idaho, 2007 and 2008.  The curvilinear relationship reflects the use of a greater number of burrows by individuals as density of available burrows increased.

 

We chose not to include the 16-ha plots from the INL Site in the index development with the larger sites for several reasons.  First, we did not have any rabbits collared on those plots, and therefore, results of the snow-track surveys could not be adjusted for rabbits that might not have been counted.   Second, the snow-track survey results indicated that the numbers of rabbits on the 16-ha plots was extremely low, and consequently, variation among repeated surveys of even 1 or 2 rabbits markedly changed the density estimates.  The variation among surveys also suggested that there might be substantial movement on and off of these plots because of their limited size.  Sanchez and Rachlow (2008) documented rabbit home ranges to vary from 0.4 to 32.1 ha, therefore, the 16-ha plots might only be able to provide habitat for a limited number of rabbits even under the most favorable habitat conditions.

When treated as a separate analysis or data for index development we found that there was not a strong relationship between the density of active burrow systems and the density of rabbits on the 16-ha plots (r2 = 0.16).  We believe that this was a function of the plot size and low density of rabbits. 

 

We then used the index developed on the larger sites to predict the number of rabbits based on active burrow systems on each 16-ha plot.  Results are summarized in Table 2.  As expected, the number of rabbits predicted by the index and the number counted during snow-track surveys were not tightly correlated.  The site with the highest number of active burrow systems (39) was predicted to have the highest density of rabbits (0.39 rabbits/ha), but density based on the snow-track surveys was 3 times lower (Table 2).  It is important to recognize, however, that the differences in the absolute numbers of rabbits between predicted estimates and snow-track counts were relatively small (0-4 rabbits) because the densities were low and the plot sizes were relatively small.  Because the curve intersects the x-axis at 0.2508, densities of burrows below this value will result in negative estimates for rabbit densities.  Therefore, at extremely low densities of burrow systems (i.e., densities below 0.025 burrows/ha) this index will be unreliable at estimating rabbit densities.

 

Conclusions and Recommendations

 

We documented a significant positive relationship between the density of active burrow systems and the density of rabbits.  The relationship is not linear because rabbits used more burrow systems as the density of burrow systems available to them increased.  Because of the curved shape of the relationship, caution should be used when estimating abundances in areas with high densities of burrows – relatively large changes in burrow densities will reflect relatively small differences in densities of rabbits. 

 

Like any index, this one should be validated in other areas.  We suggest that for the Lemhi Valley, this index can be used to estimate abundance of rabbits.  Until validated for other areas, however, the index should be used only to provide relative estimates of abundance at individual sites over time.  Many environmental factors likely differ among regions and study sites (e.g., soil and vegetation characteristics, densities of other small mammals, diversity and abundance of predators).  Because these factors could influence the relationship between rabbits and their burrow systems, making comparisons of rabbit abundances based on burrow densities across sites is not advisable.  The index is suitable, however, for monitoring changes in relative abundance over time within sites, assuming that the environmental factors that significantly influence the rabbit-burrow relationship are relatively constant over time within a site.  In this respect, use of the burrow system index on the INL Site can provide a monitoring tool for estimating changes in relative abundance over time.  More rigorous validation at multiple sites within the INL Site is recommended before using the index to estimate absolute abundances. 

 

We suggest that larger plot sizes than those used on the INL Site during this study (i.e., 16 ha) be monitored to track changes in relative densities of rabbits on the INL Site over time.  Our smallest “large” site in the Lemhi Valley was about 50 ha, however, that was also a site with the highest density of burrow systems (3.46 active burrows/ha).  The average density of active burrow systems on the 24 16-ha plots on the INL Site that were included in this study was 0.37/ha (0.55/ha for the 16 plots that contained >1 active burrow system), and 10 plots contained <3 active burrow systems.  Larger plot sizes would reduce the number of plots that contained too few active burrow systems to adequately estimate relative abundance of rabbits, and would decrease variability among estimates across time.

We recommend that annual burrow censuses be conducted during the autumn to estimate relative abundance of rabbits.  Burrow censuses could be conducted at other times of year; however, our index was calibrated with data collected during autumn, and it is unknown how behaviors at other times of year might influence the index (e.g., dispersal of juveniles, Estes-Zumpf and Rachlow 2009; creation of natal burrows; Rachlow et al. 2005).  The protocol used to conduct burrow censuses is described in Price (2009) and Sanchez et al. (2009).

 

Because the index of rabbit abundance is based on the presence of fresh pellets to categorize active burrow systems, persistence of pellets that appear fresh will influence estimates of densities of active burrows and hence, rabbit abundance.  It is important to recognize that this could result in a detection lag for rapid population declines.  Based on estimated rates of pellet degradation (Sanchez et al. 2009), this could result in a lag of one or more years between large population declines and detection of those declines based on annual burrow