DISPERSAL OF SAND DUNE LIzA1ws (ScELoPoRusAiNIcoLus) IN THE MEscALERO SANDS ECOSYSTEM

DISPERSAL OF SAND DUNE LIzA1ws (ScELoPoRusAiNIcoLus) IN THE MEscALERO SANDS ECOSYSTEM Interim report Field studies July 2002 through September 2003 Prepared by Lee A. Fitzgerald Assistant Professor and Curator of Amphibians and Reptiles Section of Ecology and Evolutionary Biology Texas A&M University February 2004 Charles W. Painter Endangered Species Program New Mexico Department of Game and Fish 1 Previous research on Sceloporus arenicolus, the Sand Dune Lizard, demonstrated these lizards are restricted to shinnery oak (Quercus havardii) sand dunes with active dune complexes (Mescalero Sands ecosystem). Anthropogenic habitat disturbance, specifically stabilizing dunes through shinnery oak removal with herbicides, and oil and gas development, were shown to have an effect on the numbers of Sand Dune Lizards at study sites (Painter et al. 1999). We determined the geographic range and distribution of the species in great detail in New Mexico, and found the distribution of S. arenicolus is small and composed of occupied and unoccupied areas of potential and suitable habitat (Fitzgerald et a!. 1997). Sand dune lizards are habitat specialists at several spatial scales. Within the limits of their geographical range, sand dune lizards have an extremely strong affinity for bowl-shaped depressions in active dune complexes refelted to as sand dune blowouts. During 3 field seasons of the distribution study, and two seasons of meticulous walking transect surveys designed to study effects of oil and gas development on Sand Dune Lizards, not a single individual was observed outside blowouts in the shinnery matrix (Painter etal. 1999). Although the importance of sand dune blowouts is clear, we do not have information about the importance of the shinnery oak matrix (the stands of shinnery oak surrounding sand dune blowouts), to sand dune lizards. Shinnery oak is clearly important for its role in maintaining active dune complexes that contain the mosaic of blowouts, but it is unknown if hatchlings (neonates) and juvenile lizards occupy the matrix as well as blowouts. Sand dune lizards lay their eggs during late June and early July, but we do not have information on habitat use and activity of neonates, nor how they may disperse away from hatching sites in and around the edges of blowouts. Moreover, we only have scant information on activity of adults during August and September. The first step to understanding dispersal in Sand Dune Lizards is to learn to what extent the hatchling or adult lizards may use the shinnery oak matrix. To effectively conserve populations of Sand Dune Lizards we need to understand what causes the patchy distribution of Sand Dune Lizards within its geographic range. For example, why are areas of suitable habitat unoccupied? Should we expect areas of restored habitat to become colonized by Sand Dune Lizards? The answer to these questions hinge on information about how the lizards may disperse from occupied sites to unoccupied sites, and what may constrain their dispersal. The research reported herein is aimed at understanding patterns of dispersal of these lizards. We intensively surveyed areas of shinnery oak adjacent to dune complexes occupied by sand dune lizards to gather and analyze information on potential use of the shinnery oak matrix by dispersing neonate and adult sand dune lizards. We used an extensive design of pitfall trapping to detect neonates. Additionally, we conducted a pilot study designed to test the feasibility of radio-tracking sand dune lizards. This pilot study was justified because female sand dune lizards may disperse across shinnery oak

flats to find nesting sites. Hence trapping neonates may oniy be part of the dispersal pattern in sand dune lizards. Gravid female S. undulatus are known to disperse outside their normal home range to find suitable nesting habitat (M. Angilletta, personal communication) and such is probably the case for many Sceloporus species. Pitfall trapping and visual detection methods would have an exceedingly small chance of detecting dispersing gravid females in the shinnery oak, and the advent of miniature radio transmitters enables individual females to be tracked. This research is important for the conservation of S. arenicolus because it will give critical information about the habitat requirements ofjuveniles as well as adults. The combination of this research with previous studies can help piece together the reasons underlying the patchy distribution of S. arenicolus, and help understand the conditions that allow populations to persist. Study site Methods The study site is within The Caprock Wildlife Area, located 48 km (30 miles) east of Roswell, NM on US Hwy 380. The area is managed by the Bureau of Land Management, with adjoining lands owned and administered by the State of New Mexico. Caprock Wildlife Area is a relatively undisturbed area of shinnery oak sand dune habitat containing many dune complexes occupied by sand dune lizards as well as flatter areas between the dune complexes. The area contains both treated and untreated areas of shinnery oak sand dune habitat. The land is used for cattle grazing and open to the public for hunting and camping. To our knowledge, this area is not scheduled for future oil and gas development or herbicide spraying, although several tracts were treated with tebuthiuron herbicide to remove shinnery oak in the early 1 990s (D. Baggao, personal communication). Pitfall-trap transects and trapping protocol We established 17 pitfall trap transects in the study area. Each transect consisted of a straight line of 15 pitfall traps made from 20-liter buckets (5 gallon) placed level with the surface of the sand (Figure 1). Each trap was covered by a 41 cmx 41 cm (16 in x 16 in) plywood square, elevated slightly off the surface to provide an entrance and shade for trapped lizards. Traps were spaced 15 m apart resulting in transects 210 m total length. The 17 transects used 255 pitfall traps. Transects were placed haphazardly throughout the study area, avoiding the central areas of dune complexes with extensive sand dune blowouts. Because habitat use by adult and juvenile sand dune lizards in sand dune blowouts was studied previously (Fitzgerald et al. 1997), it was not necessary to sample within sand dune blowout habitat known to be occupied by sand dune lizards. The transects provided good trapping coverage of all the habitats in the study area including edges of dune complexes, areas of shinnery oak flats, and sand dune blowouts. One transect (transect 1) covered a transition from untreated to treated (with tebuthiuron herbicide) shinnery habitat, and transect 2 was entirely within treated habitat. All remaining transects were entirely within untreated habitat. Each trap was precisely located with a Trimble GPS unit and the data entered into a GIS. Figure 2 shows a map of transect locations. Pitfall traps were opened on day 1 of each sampling period and checked daily on days 2 through 6, and closed on day 6, resulting in 5 full days of continuous trapping, except for the June 2003 sampling period that consisted of 8 trapping days. The sampling periods thus produced 75 trap-days per transect and 1,275 total trap-days for each sampling period. 2

3 Figure 1. A pitfall trap made from a 20 1 bucket. Each trap was covered with a plywood top to provide shade and create a refuge that would attract active lizards. N 1 a i 2 Kilometers A Figure 2. Map of the study area showing the locations of arranged in 17 transects consisting of 15 pitfalls each for a total of 255 traps. White areas are sand dune blowouts in a matrix of shinnery oak. Transects 9, 14, 15, 16, and 17 are closest to large dune complexes, and not coincidentally resulted in 40 of 57 (70%) of sand dune lizard captures.

The transect number and pitfall location where each lizard is captured was noted. The snout-vent length of all captured lizards was measured, and gender was determined when possible. (It is difficult to determine the gender of neonates and juveniles.) Each lizard received a cohort mark (toe clip) to identify it as having been captured during a given study period, and released 3 meters away from the trap where it was captured. All procedures were carried out in accordance with an approved animal use protocol from Texas A&M University. Radio tracking Miniature radio transmitters glued to the backs of 8 sand dune lizards. The radios weigh approximately 0.23 g and were temporarily attached to the lizards dorsum using Super Glue (Figure 3). The radio-tagged subjects were monitored several times daily, and their locations noted and entered into a GIS using ArcView. 4 Figure 3 A gravid sand dune lizard that was radio-tagged and monitored for several days. The radios weigh 0.23 g and were attached to the dorsum with a drop of glue. The image on right shows the radio next to a US $0.10 coin for scale. General patterns Results and Discussion We established the trapping transects during July 2002 and completed 7 trapping periods, totaling 9,690 trap-days. Trapping periods coltesponded to the activity period of sand dune lizards, including late July and September to correspond to emergence and activity of neonates (Table 1). Table 1. Dates of trapping periods, number of trapping days, and total trap-days during each trapping period. Trapping period Trapping days Total trap-days and date 1. 24-29 July2002 5 1,275 2.11-l6September2002 5 1,275 3, 12-17 April 2003 5 1,275 4. 14-19 May 2003 5 1,275 5. 10-17 June 2003 8 2,040 6. 22-27 July 2003 5 1,275 7. 17-22 September 2003 5 1,275

5 A total of 646 lizards of 7 species were captured during the 7 sampling periods, corresponding to 0.067 lizard captures/trap-day over the entire study period. Of this total 57 were sand dune lizards, corresponding to a trapping rate of 0.006/trap-day (Table 2). Three lizards found dead were not identifiable and were not included in the analyses. In addition to the lizards, 7 snakes (4 Arizona elegans, 1 Pituophis catenfer, 1 Heterodon nasica, and 1 Masticophis flagellum), 8 pocket-mice (probably Perognathus sp.), 2 Scaphiopus couchi, 11 Spea bombfrons, and 1 Ambystoma tigrinum were found in traps. Crotaphytus collaris (collared lizard), Phrynosoma modestum (round-tailed homed lizard), and a variety of snake species such as Crotalus viridis (prairie rattlesnake) and Sistruru$ catenatus (Massasauga) were also present at the study site, but were not captured in the pitfall traps. Table 2. Number of each lizard species captured during each of the 7 trapping periods. Dates for each trapping period are listed in Table 1. The bottom row shows the data expressed by trapping effort (captures/trap-day) Trapping period 1 2 3 4 5 6 7 Captures July Sept April May June July Sept per Species 2002 2002 2003 2003 2003 2003 2003 Totals trap-day Hoibrookia maculata Lesser Earless Lizard 32 4 24 39 32 99 31 261 0.027 Sceloporus undulatus Prairie Lizard 20 20 27 31 49 15 25 187 0.019 Cnemidophorus sexlineatus Six-lined Racerunner 4 2 2 9 44 11 5 77 0.008 Sceloporus arenicolus SandDune Lizard 4 3 6 14 8 5 17 57 0.006 Uta stansburiana Side-Blotched Lizard 9 4 1 5 7 16 11 53 0.005 Eumeces obsoleta GreatPlainsSkink 3 3 1 7 0.001 Phrynosoma cornutum Texas Homed Lizard 1 1 1 1 4 0.000 Grand Total 73 33 60 99 141 150 90 646 Total capture rates pertrap-day 0.057 0.026 0.047 0.078 0.111 0.074 0.071 0.067 Hoibrookia maculata (lesser earless lizard) and S. undulatus (prairie lizard) were the most commonly trapped lizards. These species preferred habitat is fiat, relatively open areas in the case of IL maculata, whereas S. undulatus prefers shinnery oak cover outside of blowouts. Hence, it is almost certain the trap numbers reflect the relative abundance of these species in the interdune areas where the bulk of traps were located. Cnemidophorus sexlineatus is a wide-ranging lizard in the whiptail family that forages actively throughout the patchy landscape. It was the third most commonly trapped lizard. S. arenicolus (sand dune lizard) and U stansburiana (side-blotched lizard) were almost equally abundant in the samples. More lizards were caught during trapping period 6 (June 2003) because the traps were checked over 8 days instead of 5 days. However, the capture rate was comparable to several other transects when the trapdays were taken into account. Trapping period 5 (May 2003) had the highest capture rate, and period 2 (September 2002) had the lowest.

6 The transects varied in the number of lizards captured (Figure 3, Table 3) It appears there were more lizard captures on transects that were closer to dune complexes, although detailed GIS analyses will enable us to quantify the distances between blowouts and pitfalls. Transect 9, for example, runs southwesterly through an area skirting sand dune blowouts (Figure 2). Six of the 7 species were captured on transect 9, including 11 sand dune lizards. As seen in Figure 2, we have base maps and trap locations in a GIS. Analyses to be completed after the 3 year of data collection will quantify the number and density of sand dune blowouts near each transect. 80 Number of lizards captured by transect 60-0 40 z 20-0 1 2 3 4 5 6 7 8 9 1011 121314151617 Transect number (corresponds to Figure 2) Figure 3 The number of lizards captured on each transect was variable. It appears the capture numbers were highest on transects that are closest to large dune complexes Table 3. The number o eacn iizara species capturea on eacn 01 me i i transects am ng the study period. The transect number corresnonds to the numbered transects on Figure 2. Species_abbreviation* Transect C. E. H. P. S. S. U. Grand No. sex obs mac corn aren und stans Total 1 1 1 13 1 9 2 28 2 16 2 13 31 3 2 1 18 4 11 2 38 4 8 1 21 1 8 1 40 5 22 1 1 14 6 44 6 6 8 2 13 29 7 2 19 1 5 27 8 5 16 11 32 9 13 1 25 1 11 18 69 10 7 23 2 12 6 50 11 3 15 3 14 5 40 12 1 1 10 1 12 3 28 13 3 14 1 15 12 45 14 3 1 17 6 19 2 48 15 13 1 9 1 8 8 1 41 16 4 6 11 10 31 17 6 9 1 4 6 26 Grand Total 77 7 261 4 57 187 53 646 *Species abbreviations: Cnemidophorus sexlineatus, Eurneces obsoleta, Hoibrookia maculata, Phrynosoma cornutum, Sceloporus arenicolus, Sceloporus undulatus, Uta stansburiana. Common names are listed in Table 2.

7 Sand dune lizard captures Sand dune lizards were most numerous in the traps on transects 9, 15, and 16. These 3 transects all traverse sand dune blowouts, and 40 of 57 (70%) of all sand dune lizards were captured on those transects (Table 3). Sand dune lizards were present on all transects except transects 2 and 8. Transect 2 is entirely within an areas treated with tebuthiuron in the mid 1 990s. Shinnery oak is sparse in that area and former sand dune blowouts are now vegetated with grasses. Transect 8 is a flat area with grasses and shinnery, relatively distant from clusters of sand dune blowouts (Figure 2). It is important to note we have not captured a sand dune lizard in a pitfall trap more than 20 meters from a sand dune blowout. Almost all the sand dune lizard captured were in pitfalls in blowouts, or on the vegetated facing slopes of dunes suffounding blowouts. The size of sand dune lizards captured reflected the emergence of neonates in July. Body size of the 29 sand dune lizards captured in July and September reflects that almost all these individuals were neonates or young juveniles. Table 4. Mean snout-vent lengths of 57 sand dune lizards captured in pitfall traps by sample month. There were 2 sampling periods in July and September during 2002 and 2003, respectively. Month Mean snout-vent length (n) April 47.80 (6) May 52.79 (14) June 54.38 (8) July 31.38 (9) September 32.10 (20) Overall 41.74 (57) The overall number of recaptures for all species during the study was 36 (5.9%) (Table 4). Five of the 57 (9.6%) captures of sand dune lizards were recaptures, resulting in the highest recapture rate among the 5 most common species. However, the recapture rate is relatively low overall, and coupled with small sample sizes, it is doubtful that strong conclusions will be drawn from recapture data. Table 5. The number of recaptures for each lizard species. The sand dune lizard had Species abbreviations are the same as Table 3 Species no yes total % recaptures H.mac 250 11 261 4.4% S.und 173 14 187 8.1% C. sex 75 2 77 2.7% S. aren 52 5 57 9.6% U. stan 49 4 53 8.2% E.obs 7 7 0.0% P. corn 4 4 0.0% 610 36 646 5.9%

Radiotrackin pilot study 8 We conducted a pilot study to test the feasibility of radio-tracking sand dune lizards, and were encouraged by the results. We used miniature radio transmitters weighing approximately 0.23 g. According to the manufacturer these may be the smallest wildlife radio transmitters available. The miniature radios worked well, allowing us to detect radio signals from several hundred meters and locate lizards in their resting sites, even when buried in sand. Battery life on these radios is approximately 10 days. This equipment will enable us to locate nesting sites and track movements of gravid female sand dune lizards. Eight lizards were tracked and monitored from 3 to 7 days. The location of each lizard was plotted onto large scale maps of the study area, and the points were later entered into a GIS. Figure xx shows locations of radio-tagged lizards, mostly within sand dune blowouts. The inset shows the movements of a gravid female over a 2-day period. This subject left the sand dune blowout where she was living and apparently traversed about 200 m of shinnery oak to arrive at a different set of blowouts. She remained there for a day, then returned to her territory. Figure 4. Map of locations of sand dune lizards revealed by radio tracking. The inset shows the movements of a gravid female over a 2-thy period. She left a sand dune blowout, traversed the shinnery habitat, and presumably nested, She then returned to the blowout where she had been living. 1

Detailed studies of movements of adult sand dune lizards, particularly gravid females, is important to understanding the patterns of dispersal of sand dune lizards. It appears based on our results to date, that gravid females leave their normal home range and disperse several hundred meters to nest. The reasons for this dispersal are probably complex. We do not understand nesting requirements of sand dune lizards (to our knowledge, nests have never been found), and females may wander in order to find a suitable microenvironment for nesting. Additionally, females may select nesting sites in habitat that favors neonate survival after hatching. Finally, nesting outside the parent s home ranges may be a dispersal-related adaptation to reduce competition among kin. At this time we are designing a detailed study of female movements to be undertaken in subsequent field seasons. In addition to quantifying the extent of movements of sand dune lizards, we expect to describe nesting behavior, the nesting environment, and how sand dune lizards negotiate the landscape while nesting. All these aspects of the lizards natural history clearly play an important role in understanding dispersal within and among populations. Visits to treated area study sites We visited several of areas used as part of the study by Snell et al. (1995) to evaluate effects of shinnery oak removal on sand dune lizard populations. The shinnery oak sand dune habitat in areas treated with tebuthiuron herbicide in the 1 980s and 1 990s have not recovered, and we did not detect sand dune lizards on the herbicide study plots. Conclusions and Recommendations The pitfall study is revealing limited use of interdune areas by juvenile sand dune lizards. Dispersal events are naturally rare, and it was not surprising to find small numbers of sand dune lizards outside of their preferred sand dune blowout habitat. We plan to continue the trapping protocol for a season in order to improve sample sizes and account for annual variation in lizard numbers. To date, no sand dune lizards have been captured more than 20 meters from sand dune blowouts, indicating that long-distance movements by neonates may be extremely rare. Neonates, like adults, seem to prefer sand dune blowouts, and neonates are easily observed in sand dune blowouts from late July through September. It is important to note, however, that we have only reported on 57 captures of sand dune lizards, and we need more data to have greater confidence in our understanding of interdune habitat use by neonate sand dune lizards. Additionally, it is abundantly clear that sand dune blowouts are patches in a shinnery oak matrix, and obviously the shinnery oak matrix must exist in order for blowouts to form and persist. Despite the species affinity for sand dune blowouts, dispersal obviously occurs at some frequency. Dispersal can occur by movements of neonates themselves, but also can occur if females leave their home ranges and disperse to nesting sites. The initial findings from the radiotracking pilot study showed a gravid female traversed shinnery oak for more than 150 meters to reach another area of sand dune blowouts. Hence the patterns of dispersal in sand dune lizards presumably involve both the movements of neonates and adults, particularly nesting females. Future research should be directed at understanding detailed movements by individual adult lizards. Use of shinnery oak by adults is clearly infrequent, which explains why sand dune lizards are never observed away from sand dune blowouts during walking surveys. However, if infrequent use of shinnery oak is a requirement during nesting, the shinnery oak habitat adjacent to dune complexes is clearly an important part of the lizards life cycle. It is important to understand the freqency and limits to dispersal in order to predict how construction of roads, oil well pads, and other forms of land use may alter the movement patterns of sand dune lizards. Dispersal studies of sand dune lizards at several scales are needed to design conservation protocols for this endemic habitat specialist. The pitfall study described here is aimed at small-scale dispersal, i.e., the use of interdune areas by adults and neonates in an area of optimal shinnery oak sand dune habitat. Sand 9

0 dune lizards have a patchy geographic distribution within their range, and common sense tells us that lizards have either dispersed in the past to populate areas, or the lizards were once homogenously distributed and have gone locally extinct in many sites without re-colonization. The answer to this question requires a large scale, metapopulation approach using genetic infonnation, and we have already embarked on studies of the population genetics of sand dune lizards throughout their range. By understanding the limits to local mechanisms of dispersion by neonates and females, we will be able to link small-scale, local patterns of dispersal, to long-term patterns of gene flow. This information will enable us to make powerful predictions about how land use in the Mescalero Sands Ecosystem may affect the viability of sand dune lizard populations in the future. 10 Acknowledgements We sincerely thank Melissa Bennett, Rebecca Grey, Cristina Solis, David LaFever, David Laurencio, David Henderson, James Hoim, Devra Jones, Larry Kamees, Bruce Christman, Wade Ryberg, Mike Sears, Lauren Chan, and Laura Bakken for help with all sorts of field tasks including installing and checking pitfall traps, radiotracking, and camp life. We also thank Steve Belinda, Kate Belinda, Rand French, and Dan Baggao of the Bureau of Land Management. Kate Belinda and Mike Sears helped assemble the GIS files, and make maps of the study area. Mike Sears created the maps in this report. References C.W. Painter, Fitzgerald, L.A., D.A. Sias, L. Pierce, H.L. Snell. 1999. Management Plan for Sceloporus arenicolus in New Mexico. Management Plan for New Mexico Department of Game and Fish, Bureau of Land Management, US Fish and Wildlife Service. 45 pp + 9 appendices. Fitzgerald, L.A., C.W. Painter, D.A. Sias, H.L. Snell. 1997. The range, distribution and habitat of Sceloporus arenicolus in New Mexico. Final report to New Mexico Department of Game and Fish, Santa Fe, NM. 30 pp + appendices.