Attach__No__1B_-_CCAFS_SEBM_Occup_Survey.pdf

Report on southeastern beach mouse

(Peromyscus polionotus niveiventris) habitat occupancy survey on

Cape Canaveral Air Force Station, March 2016.

FINAL

June 8, 2016

45th Space Wing

45 CES/CEIE

Environmental Conservation Patrick Air Force Base, FL

Delivery Order: FA2521-18-F-0059

IMSS Task Order: 201512-4913

Prepared by:

Donna M. Oddy and Eric D. Stolen Kennedy Environmental and Medical Contract (KEMCON)

IDIQ Environmental Projects IMSS Environmental Services Branch

IMSS-022

Kennedy Space Center, Florida 32899

Acknowledgements

This project was funded by the United States Air Force, 45th Space Wing, Environmental Conservation, Cape Canaveral Air Force Station and conducted through

IMSS’s KEMCON IDIQ Program. We thank the personnel at Cape Canaveral Air Force

Station for their support, especially Angy Chambers and Martha Carroll for their guidance. We thank Stephanie Legare for assistance with all phases of the logistics involved. We thank Stephanie Weiss, Brenton Back, and Tim Kozusko who assisted with tracking tube deployment, pickup, and site setup. Trade names and commercial enterprises or products are mentioned only for information.

Table of Contents

Acknowledgements

Background/Introduction

Methods

Results

Discussion

Literature Cited

Appendix A

Appendix B

Background/Introduction

The southeastern beach mouse (Peromyscus polionotus niveiventris), a federally listed threatened species, is a high priority for management on the federal lands encompassing and adjacent to the Cape Canaveral Air Force Station (CCAFS), including the Kennedy Space Center/ Merritt Island National Wildlife Refuge

(KSC/MINWR), and the Canaveral National Seashore (CNS). While the southeastern beach mouse is known to occupy sites ranging from Volusia County through Brevard

County, the distribution and habitat use within these federal lands is not well known.

The broad range of habitat conditions encompassing the species’ historical range, and the diversity of sites known to contain southeastern beach mice, require a landscape scale approach to monitoring the population and understanding species-habitat relationships.

Since 2008, biologists have monitored habitat occupancy of the southeastern beach mouse on the federal lands encompassing CCAFS/KSC/MINWR/CNS, with the goal of sampling habitat occupancy annually each winter/spring throughout the entire area of suitable coastal habitat. This monitoring program is utilizing habitat occupancy surveys based on rapid assessment techniques (i.e., tracking tubes or live-traps) to detect beach mouse occurrence at sampling stations, with repeated samples providing the information needed to estimate detectability. Measuring information on detectability allows the estimate of occupancy to account for the fact that surveys may fail to detect animals at some sites at which they are actually present (Mackenzie et al., 2002).

Through the use of covariates in occupancy models, this monitoring framework is very flexible, allowing numerous hypotheses about how various factors influence beach mouse habitat occupancy to be tested (Mackenzie et al., 2006).

A pilot study of the feasibility of such an occupancy study was conducted in

December 2009 along the 10 km beach on KSC/MINWR. This study utilized a spatial replication sampling design to provide the replicate surveys required to estimate the detection rate. In February 2010, a full-scale implementation of the habitat occupancy monitoring project was initiated on the nearly 70 km of beaches extending from the north jetty of the Port Canaveral inlet on CCAFS to the northern boundary of CNS using spatial replication. In March of 2011 a second year of data was collected from these same locations also using spatial replication. Between the second and third year of data collection, live trapping was conducted in a region encompassing the northern portion of the sampling locations. Results of this effort, combined with patterns of detections between the first and second years of the study raised questions about the scale of occupancy information that was being collected. In February 2012 a third year of sampling was performed at a subset (71) of the original sample points (43 sites on the

CNS were not sampled), using a modification of the study designed to tease apart questions regarding the scale of beach mouse habitat occupancy being measured.

Specifically, the original notion of using replicates in space (i.e., the eight tracking tube locations at each site) as the basic units of replication in occupancy models was expanded to also include replication in time by deploying tubes during 4 time periods.

This expanded sampling allowed investigation of the effects of temporary movements spatially at the scale of sample points (15 m radius), and temporally beyond the original three nights of sampling. In February-March 2013 a fourth year of data was collected from the same locations as in 2012, also using a combination of temporal and spatial replication.

In summer 2013 we conducted several tracking tube experiments designed to yield increased detection probability, while obtaining occupancy results comparable with

2012-2013, but with a decrease in effort (i.e., less visits). As a result of these experiments, the February-March 2014 survey once again used a combination of temporal and spatial sampling but consisted of two, six night periods rather than four, three night periods. In addition, all of the 114 original sites were sampled instead of the subset of 71 sites. In summer of 2014 we conducted additional tracking tube experiments also designed to yield increased detection probability, while obtaining occupancy results comparable with 2012-2014. These experiments attempted to exclude cotton rats (Sigmodon hispidus) from the tubes and decrease disturbance to the tubes, cards and ink pads from raccoons and skunks. The February-March 2015 survey used a combination of temporal and spatial sampling consisting of two, six night periods as in 2014 at all of the 114 original sites. Based on findings of the excluder experiments a 1” PVC reducer was placed in the 90º elbow attached to the tube. The

February-March 2016 survey also used a combination of temporal and spatial sampling consisting of two, six night periods with the 1” PVC reducer but with several changes 1) the number of tubes/site were reduced from eight to five, 2) all five tubes were placed 3 meters in the cardinal directions (N, S, E, W) and at the center pole, and 3) no tube covariates were taken. This report documents the results of that effort for the 35 sites

(approximately 21.4 linear km of coastal habitat) on CCAFS.

Methods

Sample sites and habitat covariates. All of the sites sampled in 2010-2015 on

CCAFS were resampled in 2016. Several of the sites experienced some form of alteration (i.e., prescribed fire {Site 4, 5, and 6}, chopping {Site 20}, removal of exotics

{Site 23}, etc.) since last year’s survey and some (Site 17, 19, 22, 25, 26 and 27) lost station poles due to erosion of the dunes at these locations. Sites were originally chosen using a combination of systematic and random selection of sampling locations.

Initial site locations on the beach were systematically spaced roughly 600 m apart along the coastline, from an arbitrarily selected start location at the northern boundary of CNS, southward to the Port Canaveral inlet on the southern end of the CCAFS beach. At each initial site location, we randomly selected a distance westward to locate the sample site, based on the width of the beach mouse habitat at the site (Appendix A).

Although we initially considered randomly selecting sites within all available beach mouse habitat, we decided that this method would be logistically too difficult. Also, because part of the focus of the study was to document the geographic range of the population within the combined federal lands, we preferred a systematic location of sites along the north to south extent rather than randomly selecting the initial location of sites

(in general, systematic designs with random start points can be substituted for completely random designs without limiting the resulting inferences that can be made back to the target population).

Once a site position had been determined, a GPS position was recorded for the center of the site using a Trimble GeoXT GPS unit (Trimble Navigation, LTD 1994), and a PVC pole was positioned to mark the site location. Four tracking tube locations were arranged in cardinal compass bearings (e.g., 0°, 45°, 90°, 180°) around the center pole and one was placed at the center pole, each 3 m from the center or at the center

(Appendix A). Tracking tube locations were marked with vinyl flags to allow them to be quickly relocated during tube deployment. Tracking tube covariates were not measured this year.

Tracking tube construction and deployment. Each tracking tube was made from a 30 cm long, 5.07 cm diameter section of PVC pipe, with an end cap at one end and a 90° elbow joint at the other. Two flexible 9-gauge wire legs each about 60 cm long were wrapped around the tube with the ends formed into legs. When deployed, a tracking tube was placed horizontally approximately 5 cm above the substrate using the wire legs, with the elbow joint pointed towards the ground. A 1 inch PVC reducer was inserted into the 90° elbow joint to reduce the opening from 2 inches to 1 inch. A 23 cm wooden dowel was stuck into the ground, extending upwards into the downward opening of the tube (the 1 inch reducer inserted into 90° elbow joint) and positioned so that it touched the edge of the tube opening. Tubes were baited with 10-15 (because they were small) sunflower seeds and placed at the closed end of the tube. Tracking papers were cut from cardstock paper to fit into tubes. The ink pad was made from a small felt pad glued and stapled to one side of a Tyvek sleeve (of the type used to protect credit card magnetic strips). The felt pad was pre-inked with a mixture of carbon lampblack ink suspended in food grade mineral oil, and the tracking paper was inserted into the end of the Tyvek sleeve and then slid into the tube with the inked end at the tube entrance (90° elbow joint). Papers were bent into a slightly convex shape prior to insertion so that they would fit into the bottom surface of the tube. This helped to prevent mice from crawling under the paper and thereby missing the ink pad and paper surface. Tracking tube papers were pre-labeled with the site number, tube location, and date. Further details of tracking tube and paper design can be found in Loggins et al., (2010) and Stolen et al., (2014).

The tubes were first deployed on the morning of 18 February 2016. The cards within the tubes were retrieved, replaced with new cards, re-inked and the seeds were replaced on the morning of 24 February, and retrieved without replacement on the morning of 1 March 2016. The tubes were also removed from the sites on 1 March

2016. When retrieved, tracking tube papers were secured to prevent smearing the footprints during transport. Potential problems noted during retrieval included:

disturbance (e.g., tube knocked over, card damaged, tube separated from dowel, etc.);

misdeployment (e.g., card in upside down, no ink pad, wrong card placed in tube, no dowel, etc.); and other concerns (fingerprint on card, ink smears on card, unable to be set for some reason). In the lab, each tracking tube paper was evaluated for evidence of a beach mouse visit independently (i.e., initially evaluated by two observers with any discrepancies being decided by a third observer), using a detailed protocol designed to separate beach mouse footprints from other small mammals residing on the federal properties (Appendix B). A database was used to store the results of the tracking tube paper evaluation, to perform data quality control measures, and to generate data reports for use in occupancy modeling.

Occupancy modeling. For this report, occupancy modeling was conducted using only the 35 sites on the CCAFS, and using only a limited set of potential occupancy models relating covariates to occupancy and detection. Occupancy models were formulated as hypotheses of how design features might affect detection or occupancy.

These models were then evaluated based on information-theoretic model selection using the unmarked package (Fiske and Chandler 2011) and custom programs in R (R

Development Core Team, 2015). Software also was used to generate parameter estimates (i.e., occupancy, detection and covariate effects) for each model.

Occupancy modeling utilizing the multiple years of data collected along the entire

70 km of the multi-agency study is underway and will be published elsewhere. Here we present results of simple occupancy models utilizing only stations on CCAFS sampled in 2016, to provide summary statistics to compare with previous CCAFS surveys.

Because we replicated samples in both space and time, the data collected in 2016 allows several different methods of analysis. One approach is to treat all detections within a station during each 6 night sampling period as a replicate survey; this is the single season occupancy design of MacKenzie et al. (2002). This formulation requires that occupancy be estimated for the entire period, but detection probabilities can be estimated for each replicate survey, or constrained to be equal for all surveys.

An alternative method that uses more of the information in the data treats the two sampling periods as “seasons” (in this case temporal replicates) and the five tube locations within a station as “surveys” at a finer level of replication (in this case spatial replicates). This formulation can be analyzed using the multi-season occupancy design of MacKenzie et al., (2003). This type of occupancy model estimates the rate of site occupancy for the first sampling period (survey), and also estimates the site extinction and colonization rates between the sampling periods. Occupancy rates for all sampling periods can also be calculated from the results.

Results

For beach mouse occupancy modeling, the criteria used to register a detection of a beach mouse at a tracking tube is that at least two of the three readers record a footprint less than 7.2 mm (Stolen et al., 2014). By this criterion, beach mice were detected at 59 of the 350 tracking tube sets (one tube active during one survey), 47 of the 175 tracking tube locations (five locations at each of the 35 stations), and 15 of the

35 stations (Table 1). In contrast with 2015, twenty sites did not have beach mice detections in 2016 (Table 2). In 2015, with the addition of excluders the number of disturbed tubes was reduced to only eight tubes or 1.4% (five of which were knocked over), the number of tubes with concerns decreased to 4 or 0.7% with human fingerprints, but the number of misdeployed tubes increased to 26 or 4.6% (24 tubes had incorrect cards placed in them and 2 tubes had cards that weren’t correctly assembled). This year the number of tubes disturbed was further decreased (1 during each survey period or 0.5%; both were knocked over tubes). In addition, the total number of concerned tubes also decreased to 12 (12 tubes had cards with fingerprints and/or ink smears on them but only one of the 12 also had footprints on it) but for analysis purposes the one with footprints on it is the only one that counted making it

0.3%. Furthermore, the number of misdeployed tubes was reduced to 5 or 1.4% (4 tubes had incorrect cards placed in them but the mistakes were corrected in the field and 1 tube was not correctly assembled).

Under the simplest design which pooled all detections within a station during each 6 night sampling period, the estimate of habitat occupancy was 64% (Table 3). The single season occupancy model treated all detections within a station during each 6 night sampling period as a replicate survey. The data used in this analysis are provided in

Table 1 and estimates from the various occupancy models in Table 3.

Figure 1. Locations at which southeastern beach mice were detected at least once during the occupancy survey on CCAFS, February-March 2016. Green circles indicate that beach mice were detected at a site and red circles indicate no detection at a site.

Numbers indicate site locations.

Table 1. Detections of southeastern beach mice in 2016 based on the 7.2 mm size criterion. A “+” indicates a southeastern beach mouse detection at each of the 5 tubes (N, S, E, W, C) for each station (at least 2 of the 3 tracking tube card readers recorded a footprint <7.2 mm in width). The total number of detections during each of the 2 sampling periods is given in columns separating the tracking tube detection columns.

2/18 – 2/24 2/24 – 3/1

S ta tio n

N S E W C

T o ta l

N S E W C

T o ta l

1 0 0

2 0 0

3 0 0

4 0 + 1

5 0 0

6 0 0

7 0 0

8 0 0

9 0 + + + 3

10 0 0

11 0 + + + + + 5

12 0 0

13 0 0

14 0 + + 2

15 0 + 1

16 + 1 + 1

17 0 0

18 0 0

19 0 0 20 0 + + + + + 5

21 0 + + + 3

22 0 0

23 0 0

24 0 0

25 + 1 + + + 3

26 0 + + 2

27 0 0

28 0 0

29 0 0

30 + + + + + 5 + + + + + 5 31 0 + 1

32 + + + + + 5 + + + + + 5

33 0 0

34 0 + + + + + 5 35 0 + + + + + 5

Table 2. Detections of southeastern beach mice in all years sampled based on the 7.2 mm size criterion. An “X” indicates a southeastern beach mouse was detected (at least 2 of the 3 tracking tube card readers recorded a footprint <7.2 mm in width) at each site in at least one of the tubes over all sampling periods. A summary column was used to show which sites had detections during any of the 7 years sampled. Note that different methods were used across years: 2010-2011 used 1 sample period of 3 nights with 8 tubes, 2012-13 used 4 sample periods of 3 nights each with 8 tubes, 2014-2015 used 2 sample periods of 6 nights each with 8 tubes and 2016 used 2 sample periods of 6 nights each with 5 tubes.

Site 2010 2011 2012 2013

Any Year

1 X X X X

2 X X X X X X

3 X X X X

4 X X X X X

5 X X X X

6 X X X X X

7 X X X X X

8 X X X X X

9 X X X X X X

10 X X X X

11 X X X X X X X X

12 X X X X

13 X X X X

14 X X X X X X X

15 X X X X X X

16 X X X X X

17 X X X X X

18 X X X X

19 X X X

20 X X X X X

21 X X X X X

22 X X X X X

23 X X X

24 X X X X

25 X X X X X X X X

26 X X X X X X

27 X X X X X

28 X X X X

29 X X X

30 X X X X X X X X

31 X X X X X X X

32 X X X X X X X

33 X X X X X X

34 X X X X X X X X

35 X X X X X X

Table 3. Parameter estimates from habitat occupancy models for southeastern beach mice sampled on CCAFS in February-March 2016, estimated by single season and multi-season occupancy models.

Design Parameter Estimate SE

Lower

CL

Upper

CL

first survey occupancy 0.41 0.09 0.26 0.58 period only detection 0.37 0.05 0.28 0.46 single-season, occupancy 0.64 0.20 0.24 0.91 constant detection detection 0.42 0.14 0.19 0.70 single-season, occupancy 0.43 0.08 0.28 0.59 survey-specific detection P1 0.27 0.11 0.10 0.53 detection detection P2 1.00 0.00 0.00 1.00

Multi-season, occupancy 0.12 0.05 0.04 0.27 constant detection detection 0.62 0.05 0.51 0.71 extinction 0.00 0.01 0.00 1.00 colonization 0.36 0.09 0.21 0.54

Multi-season, occupancy 0.12 0.05 0.04 0.27 survey-specific detection P1 0.59 0.11 0.37 0.79 detection detection P2 0.62 0.06 0.51 0.73 extinction 0.00 0.01 0.00 1.00 colonization 0.36 0.09 0.21 0.54

Discussion

The large amount of effort used to detect beach mice in the CCAFS occupancy study over the entire survey period resulted in a robust estimate of occupancy rate. The

CCAFS occupancy study sampling was integrated into a multi-agency study which includes 70 km of beaches extending from the north jetty of the Port Canaveral inlet on

CCAFS to the northern boundary of the CNS. This partnership is thriving with support from all federal land owners, and will continue to provide high quality information about the distribution and habitat use of the southeastern beach mouse. The occupancy modeling results presented here are not meant as a definitive analysis of beach mouse occupancy dynamics for CCAFS, but are intended to illustrate what is possible with occupancy models. Occupancy modeling utilizing the seven years of data and covariates collected along the entire 70 km of the multi-agency study is underway and will be published elsewhere.

When applied to 2016 data collected in the portion of CCAFS, the simplest occupancy model indicated a habitat occupancy rate of 64% for coastal habitat on

CCAFS. Because of the random placement of the sites, this can be interpreted as indicating that about 64% of coastal habitat on CCAFS was occupied by the southeastern beach mouse during the time period sampled. This is much higher than either of the first two year’s estimates (0.35 in 2010 and 0.30 in 2011) but less than the last four years (0.78 in 2012, 0.82 in 2013, 1.0 in 2014 and 0.79 in 2015). A complicating factor is the decrease in the number of tubes deployed at each site, although the theory behind occupancy modeling provides confidence that the occupancy measures are comparable despite differences in sampling effort. The increased rate of habitat occupancy in 2012-2016 compared with 2010-2011 was possibly due to changes in study design during later years which changed the scale at which occupancy was being measured. Occupancy rate calculated for just the first sample period in 2016 was 0.41 (Table 3); this is a better comparison with the first two years in which only one sample over a single period of three nights was used. Cyclical changes in small mammal populations are to be expected and this is one of the reasons why long-term monitoring is essential in understanding their population dynamics. The advantage of the occupancy approach is its ability to characterize spatial and temporal patterns in beach mouse habitat use over large geographic areas. This flexible and cost effective approach to monitor beach mouse populations will continue to undergo refinements to maximize its ability to inform management decisions.

In the summer of 2013-2014 we conducted several tracking tube experiments. The data collected in 2013 indicated that optimal sampling could be conducted using two, six night sampling periods in 2014 rather than the four, three night periods used in 2012 and 2013. This new design allowed us to keep the longer sampling period while obtaining occupancy results comparable with 2012-2013, but with a decrease in effort.

Results from 2015 which also used fewer sampling periods was similar to results from previous years and shows that the reduced sampling effort allows collection of data with the same quality of information with less effort. In addition to a better understanding of the effect of sampling duration, the experiments were also designed to determine how best to restrict certain small mammal species from reaching into or entering the tubes.

Specifically, raccoons and skunks often reach into tubes, while cotton rats often enter the tubes, causing disturbance (e.g., cards and/or ink pads removed from the tube, cards chewed or torn, etc.), increasing non-target footprints, and potentially reducing beach mouse use of tubes through competition or predator avoidance. Results from

2016 have shown that by excluding these non-target species we can decrease the rate of disturbance by non-target species and thereby increase the detection rate for beach mice and improve the precision of our estimate of habitat occupancy. The decrease in disturbance of the tubes in 2015-2016 as a result of the addition of the excluders may have brought to light another potential issue (i.e., knocking over the tube rather than the pulling out the card and ink sleeve) by raccoons and skunks. We believe this may be due to the lack of stability of the legs on the tubes (i.e., they are not made of the recommended 9-gauge wire). If not addressed, this may be further exacerbated when sampling is moved inland.

Literature Cited

Burnham, K. P., and D. R. Anderson. 2002. Model selection and multi-model inference:

a practical information-theoretic approach. Second edition. Springer, New York.

Fiske, I., and R. Chandler. 2011. Unmarked: An R Package for Fitting Hierarchical

Models of Wildlife Occurrence and Abundance Journal of Statistical Software 43:1-

23.

Keserauskis, M. M. 2007. Food habits of Peromyscus polionotus niveiventris

(southeastern beach mouse) on Cape Canaveral Air Force Station. M.S. Thesis, University of Central Florida, Orlando.

Loggins, R. E., J. A. Gore, L. L. Brown, L. A. Slaby, and E. H. Leone. 2010. A modified track tube for detecting beach mice. Journal of Wildlife Management 74:1154-1159.

Mackenzie, D. I., J. D. Nichols, G. B. Lachman, S. Droege, J. A. Royle, and C. A.

Langtimm. 2002. Estimating site occupancy rates when detection probabilities are less than one. Ecology 83:2248-2255.

Mackenzie, D. I., J. D. Nichols, J. E. Hines, M. G. Knutson, and A. B. Franklin. 2003.

Estimating Site Occupancy, Colonization, and Local Extinction When a Species Is

Detected Imperfectly. Ecology 84:2200-2207.

MacKenzie, D. I., J. D. Nichols, A. Royle, K. H. Pollock, L. L. Bailey, and J. E. Hines.

2006. Occupancy estimation and modeling: inferring patterns and dynamics of species occurrence. Elsevier, Amsterdam.

R Core Team (2015). R: A language and environment for statistical computing. R

Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/.

St. Johns River Water Management District. 2006. SJRWMD Land Use and Land Cover

(2004). URL (accessed July 2 2010).

http://floridaswater.com/gisdevelopment/docs/metadata/luse2004.htm#6

Stolen, E. D., D. M. Oddy, M. L. Legare, D. R. Breininger, S. L. Gann, S. A. Legare, S.

K. Weiss, K. G. Holloway-Adkins, and R. Schaub. 2014. Preventing tracking-tube false detections in occupancy modeling of southeastern beach mouse. Journal of

Fish and Wildlife Management 5(2):270-282.

Trimble Navigation, LTD. 1994. GPS pathfinder professional operating manual.

Sunnyvale, California, USA.

http://www.amazon.com/s/ref=ntt_athr_dp_sr_1?_encoding=UTF8&sort=relevancerank&search-alias=books&field-author=Darryl%20I.%20MacKenzie http://www.amazon.com/s/ref=ntt_athr_dp_sr_2?_encoding=UTF8&sort=relevancerank&search-alias=books&field-author=James%20D.%20Nichols http://www.amazon.com/s/ref=ntt_athr_dp_sr_3?_encoding=UTF8&sort=relevancerank&search-alias=books&field-author=J.%20Andrew%20Royle http://www.amazon.com/s/ref=ntt_athr_dp_sr_4?_encoding=UTF8&sort=relevancerank&search-alias=books&field-author=Kenneth%20H.%20Pollock http://www.amazon.com/s/ref=ntt_athr_dp_sr_5?_encoding=UTF8&sort=relevancerank&search-alias=books&field-author=Larissa%20L.%20Bailey http://www.amazon.com/s/ref=ntt_athr_dp_sr_6?_encoding=UTF8&sort=relevancerank&search-alias=books&field-author=James%20E.%20Hines http://floridaswater.com/gisdevelopment/docs/metadata/luse2004.htm#6

Appendix A.

Southeastern beach mouse occupancy study tube placement protocol

Placement of center pole at center point of site

To locate the site, travel along the beach north or south until the designated Y coordinate for the point is reached. Locate the point at which the dune vegetation begins along this line of latitude; this will be the starting point.

1) Standing at the starting point, face eastward and identify a line marking the transition between the open sand and the vegetation along the primary dune – this will be called the dune vegetation boundary (see figure below).

2) Locate the western edge of available habitat which is perpendicular to the dune vegetation boundary. Available habitat is habitat suitable for beach mice, and also that can be traversed quickly to ensure deployment of tubes in the allotted 15 minutes.

Generally, this will be where dense, woody vegetation or flooded salt marsh begins.

This will be called the unsuitable and/or inaccessible habitat boundary (see figure below).

3) Use the range finder to measure the distance between the dune vegetation boundary and the westernmost accessible habitat edge. This is the available habitat width.

4) If the available habitat width is greater than or equal to 90 m consult the random number datasheet and read the random east-west distance for the point. If the available habitat width is greater than or equal to 300 m, then use 300 m as the available habitat width (i.e., 300 m is the maximum distance for this study).

5) If the available habitat width is less than 90 m then the east-west distance for the point will be set at the center of the available habitat width (e.g., if the available habitat width is 50 m, then the east-west distance for the center point will be 25 m).

6) Using the rangefinder, one person stands at the starting point at the dune vegetation boundary and directs a second person to the position of the center point. The center point will be directly east from the starting point. The second person plants the center pole at the center point.

7) Flag path to center pole so that it can be located quickly from dune. Also flag position along dune.

8) Record GPS position of the center pole.

9) Record site covariates.

Placement of 5 tracking tubes at site

At each site, 5 tubes will be placed at the cardinal directions and at the center pole (N, S, E, W and C) 3 paces from the center point. The following directions begin with investigator standing at the center pole of a site.

1) At each cardinal direction, perform the following steps (use a compass to determine the direction by locating a distant landmark).

a. Walk the assigned number of paces (1 pace = 2 steps) into the dune vegetation in the appropriate cardinal direction.

b. Place the tracking tube at the position directly in front of your leading foot.

c. Flag site to allow relocation.

Appendix B.

Southeastern beach mouse habitat occupancy study tracking paper evaluation protocol

There should be an objective, repeatable method of deciding if a tracking paper has registered a beach mouse visit. The main problem is that registering a false positive

(i.e., recording a tracking paper as having beach mouse prints when it in fact does not) violates a basic assumption of the occupancy model (that occupancy is never recorded falsely). Conversely, a false negative does no harm statistically, beyond lowering the precision. Additionally, in the future, in sites where the Golden mouse and Florida mouse are also possible occupants, there could be footprint identification issues. Thus, a rule for objectively reading the tracking papers is necessary.

All tracking papers will be read once by each observer as follows:

1) Each tracking paper will be visually scanned for measurable prints (a measurable print is one which has a clear digit at either side of the front footprint).

2) If the tracking paper lacks measurable prints, the tracking paper is recorded as “read” on the data sheet but no measurement value is recorded. Any pertinent notes may also be recorded.

3) If the tracking paper has measurable prints, the tracking paper is recorded as “read” on the data sheet. All measurable prints will be measured with a caliper and the smallest value will be recorded on the data sheet. Other values will be recorded in notes if deemed necessary (e.g., if there appears to be several sets of prints of different sizes on the paper).

4) If there appears to be partial footprints that cannot be measured, the tracking paper is recorded as “other evidence” on the data sheet and notes about the nature of the prints is recorded.

5) Each tracking paper will be read 2 times in the blind (each reader reading once, unaware of any previous results).

6) Once each paper is examined by the 2 separate readers, the results will be tallied based on the size criterion (e.g., <7.2 mm).

7) All cards with only 1 reader measuring a print within the size range for beach mice will be re-examined by a 3rd reader.

8) The final detection results will be determined based on agreement of at least 2 out of 3 (e.g., either 2 readers measuring a print of suitable size, or 2 readers recording no measurable print).