Livestock Grazing Influences Habitat and Local Extirpations of the Arizona Montane Vole in New Mexico

The montane vole Microtus montanus (M. m.) is a locally abundant small mammal with a geographic range that extends across much of the higher elevations of the western United States, from south-central British Columbia, Canada, south to Arizona and New Mexico (Hall 1981). The Arizona montane vole M. m. arizonensis is the southernmost subspecies, which is restricted to the upper headwaters of the Little Colorado and Gila rivers in the White Mountains region of east-central Arizona and adjacent areas of west-central New Mexico (Figure 1; Anderson 1959; Sera and Early 2003). The Arizona montane vole is highly isolated from the next nearest population of the species located in northern New Mexico and, consequently, it is morphologically and genetically divergent (Anderson 1959; Frey 2009). In New Mexico, the Arizona montane vole is known based on two published locations in the upper San Francisco River watershed—a major tributary to the Gila River in west-central Catron County. It was first discovered in New Mexico at Centerfire Bog in 1978 (Hubbard et al. 1983). A second population was discovered in 1994 at Jenkins Creek (Frey et al. 1995). Subsequently, NMDGF (2022) captured montane voles at several locations on Jenkins Creek in 1998 and 2000 near the original 1994 location, though no details were provided. The Arizona montane vole is listed as endangered by the state of New Mexico based on its small population and restricted habitat (NMDGF 2022).

The montane vole is a typical “grass-tunneling” vole that makes runways under the cover of graminoids and associated forbs that are also used as food (Sera and Early 2003). Throughout the species’ range, grass cover is the principal feature that characterizes the habitat of montane voles, and they are most abundant at locations with abundant herbaceous cover (Anderson 1959; Getz 1985; Sera and Early 2003). This habitat relationship is likely due to the vulnerability of voles to predation by a variety of vertebrate predators when cover is not available. In the American Southwest, montane voles are restricted to high elevation meadows within the coniferous forest zone, which receives more precipitation and has higher primary productivity (Anderson 1959; Findley and Jones 1962). Habitat requirements of the Arizona montane vole have not been defined, but based on its extreme southern distribution, its habitat may be even more restricted to mesic locations that can support high herbaceous cover. Indeed, Findley and Jones (1962) described Arizona montane voles as occupying mesic grass and sedge dominated meadows, while Hoffmeister (1986) described them as associated with damp to wet grassy or marshy areas that provided dense herbaceous cover for their runways. In New Mexico, Arizona montane voles were caught in wet sedge and grass meadows surrounding open water and emergent marshes at Centerfire Bog (Hubbard et al. 1983), while at Jenkins Creek they were captured in a mesic meadow with moist soil and tall, dense grass (Frey et al. 1995).

The New Mexico jumping mouse Zapus luteus luteus is sympatric with the Arizona montane vole and it is associated with tall, dense herbaceous vegetation on mesic soil (Frey 2017). The New Mexico jumping mouse has suffered substantial population declines due to loss of habitat from livestock grazing and other threats, warranting its listing (as Z. hudsonius luteus) as endangered under the U.S. Endangered Species Act (ESA 1973, as amended; USFWS 2014 Reference S1; Frey 2017). Thus, it is possible the Arizona montane vole also has experienced habitat loss and population declines. Information on the vole’s habitat and status are necessary to manage the species in New Mexico. The overarching goal of this study was to determine the distribution and limiting factors for Arizona montane voles in New Mexico to inform its conservation and management. Our specific objectives were to 1) evaluate museum records and survey for Arizona montane voles in New Mexico in 2004 and 2020 to evaluate their distribution and any changes in occurrence; 2) use habitat data collected from throughout the range of the Arizona montane vole in Arizona and New Mexico to describe its landscape scale habitat selection and microhabitat; 3) evaluate influence of livestock grazing on their occurrence and habitat; and 4) make management recommendations on the basis of these results to guide its conservation.

The lowest elevation record for Arizona montane voles reported by Hoffmeister (1986) was 2,125 m. Therefore, we confined our field surveys to the upper San Francisco River watershed on lands managed by the Gila National Forest in west-central Catron County (Figure 2). Elevations in this watershed ranged from 2,125 m in the Luna Valley at the southern edge of the study area to over 2,830 m on the Canovas Rim that defines a portion of the northern boundary of the watershed. The vegetation was predominantly ponderosa pine Pinus ponderosa forest. Lower elevations contained grassland or pinyon Pinus edulis-juniper Juniperus spp. woodland, while higher elevations were primarily mixed coniferous forest (Pseudotsuga menziesii, Abies concolor, Populus tremuloides). Meadows and grasslands were found primarily in valley bottoms. The main uses of the area were livestock grazing, timber extraction, and recreation.

We searched the ARCTOS database (https://arctosdb.org/ [August 2024]) for museum specimen records of M. montanus from Catron County, New Mexico. Because the montane vole is easily misidentified (Hubbard et al. 1983), we examined specimens of other vole species from the study area and held in the University of New Mexico, Museum of Southwestern Biology. We used morphological characteristics described in Frey (2010) to identify the specimens. In 2020, we attempted to resurvey locations from which these specimens came.

Montane voles require areas with dense graminoid plants such as grasses and sedges (Sera and Early 2003). In arid environments, these areas occur where there is moist soil, such as at springs, ciénegas, and fens, and along streams and drainage bottoms. We searched for areas with lush graminoids, using maps, satellite imagery, prior field experience, and opportunistically as we travelled throughout the study area. Grass-tunneling voles construct distinctive runways on the surface of the ground that are approximately the width of the vole’s body and slightly depressed into the ground due to repeated foot-traffic; runways are kept cleared of vegetation growth by the vole’s activities. Voles use their runway system as travel lanes connecting their nest to foraging areas with a majority of movements confined to a short distance (<5 m) from the nest (Pearson 1960). The presence of clipped green vegetation and moist green feces, which is a product of the vole’s graminivory, in the runway indicate that a vole is actively using the runway. Thus, we determined presence of grass-tunneling voles, which in our study area included montane voles and Mogollon voles Microtus mogollonensis, by finding active runways (Yarborough and Chambers 2007). We walked areas with relatively lush graminoid vegetation and visually searched for runways. In areas with dense vegetation, we used our hands to part the vegetation to see the ground at frequent (e.g., 1- to 2-m) intervals. Once we found vole runways, we placed live traps directly on the runways and at burrow entrances to determine which vole species was using it. Because voles use small well-defined areas that are easily observable, it requires relatively little trapping effort to verify presence, and thus trapping efforts are much lower than used to survey for most other species of small mammals where large numbers of traps must be set in arrays in hopes of intersecting moving animals.

We conducted field surveys in June and early July 2004 and in June 2020. When vole runways were found at a site, we set Sherman live traps (model LFATDG; H.B. Sherman, Tallahassee, FL) baited with commercial horse sweet feed approximately 1–5 m apart on runways whenever possible and clustered traps in areas with high concentrations of active runways. We attempted to saturate potential habitat with traps even if vole runways were not found, and we curtailed trapping at a site once a montane vole was captured. Thus, sampling effort was variable among sites. In 2020, we resampled locations that we sampled in 2004. If no vole runways were found at a historical site, we set traps in the best-developed herbaceous vegetation cover available. We did not trap historical sites if there was no vole habitat and no vole runways present and regarded voles as absent from such sites. To prevent mortalities, we protected traps from the sun and checked them frequently. We identified captured animals and released them at their capture locations. For all voles captured, we measured hindfoot length, tail length, mass, and sex. We collected at least one Arizona montane vole from each new location and prepared it as a voucher specimen. For these specimens, we also collected total length and ear length. Capture rate was calculated as the number of captures at a site per 100 trap-nights (a trap-night is a measure of survey effort wherein one trap-night is equivalent to one trap set for one night), which provides an index of relative abundance (Wilson et al. 1996).

Due to a confidentiality agreement with the New Mexico Department of Game and Fish, specific locality data are not provided in this paper. Researchers can obtain details of survey locations from the lead author, New Mexico Department of Game and Fish, or New Mexico Heritage Program. We use quotations to designate unofficial place names.

We collected quantitative habitat data at two scales, landscape (i.e., along an up to ca 1-km reach of stream or valley bottom), and microhabitat (i.e., on a 4-m radius plot at trap locations). Landscape scale data were collected in 2020, while microhabitat data were collected during both 2004 and 2020. Both methods were previously described and used in studies conducted in Arizona in 2008 and 2009 on the New Mexico jumping mouse (Frey 2017) and western water shrew Sorex navigator (Frey and Calkins 2020), during which Arizona montane voles were the most common mammal captured (Frey 2011). We used habitat data collected in Arizona to increase sample sizes for statistical analyses and to better represent the range of habitat conditions used by Arizona montane voles across their distribution.

For the landscape “stream reach” scale, at each survey site we established paired transects paralleling the stream or drainage bottom and located 0.5 m (i.e., “stream-edge transect”) and 4.5 m (i.e., “inland transect”) from the edge of the green-line (i.e., vegetation closest to the water or area where water would flow if present). We located the paired transects on the same side of the drainage that trapping occurred, or we randomly determined the side if trapping occurred on both sides of the stream or drainage bottom. We established sample stations each 20 m along transects. We aimed for transects to be 1 km in length, but lack of continuous habitat or lack of access reduced the length at most sites. We measured vertical cover with a Robel pole (Robel et al. 1970). At each station, we read the Robel pole from the opposing transect (i.e., 4 m away) with our eyes at 1 m height and reading the lowest 1-inch (25.4-mm) segment of the pole that was not obstructed by cover. In addition, we recorded the dominant plant or other structure that covered the pole: conifer, rush Juncaceae, sedge Carex spp., grass Poaceae, forb, willow Salix spp., alder Alnus spp., shrub cinquefoil Dasiphora fruticosa, Wood’s rose Rosa woodsii, other shrub, dead standing plant, coarse woody debris, rock, bank, and other (e.g., cattle feces). We recorded if survey sites were permitted for livestock grazing (https://www.fs.usda.gov/rangeland-management/grazing/ [August 2024]) and if there was sign of recent livestock grazing (presence of cattle, feces, or footprints). Physical counts of livestock on an allotment by the U.S. Forest Service are uncommon and therefore actual livestock use at any one time is generally unknown. Allotments with permitted grazing reflect long-term grazing patterns.

We collected microhabitat data at traps where we captured Arizona montane voles. If none were captured, we collected microhabitat data at a representative location where Mogollon voles were captured since the two species occur in the best available vole habitat at a site. Mogollon voles tend to use drier grasslands than montane voles but also use mesic areas where there are no competitive vole species, such as the Arizona montane vole (Frey and LaRue 1993; JK Frey, personal observation). If no voles were captured, a microhabitat plot was established at a representative location in the best-developed herbaceous cover. Because the sample unit was the trap location, at some sites more than one microhabitat plot was measured. If animals were caught in close proximity (<8 m apart) we only collected microhabitat data at one trap to prevent plots from overlapping. At the trap, we visually estimated slope and aspect. We measured canopy cover with a spherical convex densiometer (Forestry Suppliers, Jackson, MS) in the four cardinal directions as determined by a compass. We obtained an index of soil moisture ranging from 1 (dry) to 10 (saturated) using a soil moisture probe (Lincoln Irrigation, Lincoln, NE) inserted into the ground approximately 4 cm. We assessed vertical cover with a Robel pole (read in inches) from a 4-m distance at a 1-m eye level: we read the Robel pole at the trap from three random azimuths as well as 4-m away from the trap along three random azimuths. We established four 4-m perpendicular transects at a random azimuth from the trap. At each 1-m interval along a transect, we used a Daubenmire (1959) frame to assess the percent cover of sedges, rushes Juncaceae, field horsetail Equisetum arvense, forbs, grass, willow, alder, redosier dogwood Cornus sericea, shrub cinquefoil, Wood’s rose, moss, coarse woody debris, litter, rocks, gravel, bare ground, open water, and other (e.g., cattle feces). Cover classes were 1 (0–5% cover), 2 (5–25% cover), 3 (25–50% cover), 4 (50–75% cover), 5 (75–95% cover), and 6 (95–100% cover). In addition, we recorded soil moisture, litter depth, and stubble height in each frame. We measured stubble height with a ruler (mm) and recorded it in two ways. We measured laid-over stubble height as the representative height of the vegetation as it naturally lay. We measured vertical stubble height by measuring the height of a representative blade of the dominant herbaceous vegetation that was fully extended vertically from the ground. Finally, we identified and counted each tree and shrub within 1 m of transects.

We calculated statistics using GNU PSPP version 1.4.1 (GNU Project 2015). We tested variables for normality using one-sample Kolmogorov-Smirnov tests and used Pearson and Spearman correlations to assess relationships among variables for normal and non-normal variables, respectively. For the landscape scale, we tested for differences in survey sites where Arizona montane voles were captured or not captured, and we tested for differences in survey sites based on livestock grazing policy or sign of livestock, by using two-tailed t-tests (equality of variances not assumed) and two-sample Kolmogorov-Smirnov tests for parametric and non-parametric data, respectively. We interpreted strength of evidence based on P-values following recommendations of Ramsey and Schafer (2002), where P < 0.01 was convincing evidence, P = 0.01–0.05 was moderate evidence, P = 0.05–0.10 was suggestive but inconclusive evidence, and P > 0.10 was no evidence. Because this was a mensurative study design that did not control for confounding variables, use replication, or manipulate the system, the scope of inference is limited to the study sites (Morrison et al. 2008). For microhabitat variables, we report descriptive statistics for capture locations.

In the museum collections, there were nine specimens of Arizona montane vole from three locations in August and September 2008 that had not previously been published. The three locations were on Stone Creek (MSB 195586), the San Francisco River above Stone Creek (MSB 196253, 196635), and an unnamed creek that is a tributary to the San Francisco River (“Unnamed Creek” herein; MSB 198312, 198419, 198421, 198735, 198747, 198425; Figure 2). Three of these had been incorrectly identified as Mogollon voles, while conversely one specimen (MSB 198414) cataloged as an Arizona montane vole was actually a Mogollon vole. Two vole specimens collected in July 1974 from near Aragon on the Tularosa River (MSB 164728, 164729) and originally identified as montane voles are actually Mogollon voles (see Frey 2010 for details). In 2020, we resampled the San Francisco River and “Unnamed Creek” sites; we did not resample the Stone Creek site due to access limitations.

In 2004, we trapped 13 sites using 915 trap-nights and captured Arizona montane voles at five sites: Jenkins Creek lower, SA Creek, Romero Creek, “Flanagan Creek” upper, and San Francisco River (Figure 2; Table 1). We resurveyed the original 1994 capture location for the vole on Jenkins Creek but did not capture it there. In 2004, capture rates of Arizona montane voles were high at “Flanagan Creek” upper (15.4 captures/100 trap-nights), low at San Francisco River and SA Creek (ca 6.5 captures/100 trap-nights), and very low at Romero Creek and Jenkins Creek lower (<3.0 captures/100 trap-nights; Table 1). In 2020, we trapped 13 sites using 1,466 trap-nights and captured Arizona montane voles at four sites (Figure 2, Table 1). The 2020 surveys included six sites where montane voles were previously known to occur: Jenkins Creek original, all sites where Arizona montane voles were captured in 2004 except “Flanagan Creek” upper (i.e., sites 5, 7, 12, 15), and “Unnamed Creek”, which was documented by a specimen collected in 2008. We captured Arizona montane voles at only one, Romero Creek. “Flanagan Creek” upper was not trapped in 2020 because there was no vegetation cover or vole runways; therefore, voles were not present. Three of the sites where we captured Arizona montane voles in 2020 were new locations: Jenkins Creek middle, “Flanagan Creek” lower, and “Trap Creek” (Figure 2, Table 1). In 2020, capture rates of Arizona montane voles were moderate at the historical Romero Creek site (11.3/100 trap-nights) and very low at the three new sites (<3.0/100 trap-nights; Table 1).

At the landscape scale, reaches where Arizona montane voles were captured in Arizona and New Mexico were dominated by grasses (28.0%), sedges (20.6%), and forbs (19.4%; Table 2). Compared to reaches where montane voles were not captured, there was strong evidence that reaches where voles were captured had less cover of “other”, moderate evidence that capture reaches had higher mean vertical cover and more cover of conifers, and suggestive evidence that capture reaches were at higher elevation and had more alder and other shrub cover. On the stream-edge transects, which best represent conditions of a riparian zone, the mean vertical cover on reaches where montane voles were captured was 50 cm (19.6 in; 95% confidence interval (CI): 41–59 cm; Figure 3), but was 30 cm (11.8 in; 95% CI: 15–45) on reaches where montane voles were not captured. All sites surveyed in New Mexico in 2020 had mean vertical cover of the stream-edge transect less than the lower 95% confidence interval for this transect collected throughout the range of the Arizona montane vole (Figure 3).

Microhabitat at traps that captured Arizona montane voles in Arizona and New Mexico was typified by saturated soil (mean = 9.1), low canopy cover (mean = 8.4%), high vertical cover (mean = 40 cm [15.9 in] at the trap), and ground cover dominated by sedges (Table 3). Microhabitat also consisted of relatively high cover provided by grasses, forbs, litter and rushes (Table 3). On average, vertical cover and soil moisture were higher at the trap than on the surrounding 4-m radius plot, suggesting fine scale habitat selection for the wettest spots with highest cover of herbaceous vegetation. The lower 95% confidence interval of vertical cover at traps that captured Arizona montane voles in Arizona and New Mexico was 28 cm (10.9 in; Figure 4).

Captures of Arizona montane voles were more often in areas where livestock grazing was not permitted (z = 5.72, P = 0.017; Figure 5), although there was no pattern of presence of livestock or their sign with captures of montane voles. All but two sites, both livestock grazing exclosures, had sign of recent livestock grazing. The capture rate of Arizona montane voles was higher in areas where grazing was not permitted (z = 3.42, P = 0.064).

Livestock grazing policy influenced vegetation characteristics at the landscape scale. There was convincing evidence that mean vertical cover was higher at sites where livestock grazing was not permitted (streamside transect: 53 cm [20.8 in]; upland transect: 45 cm [17.6 in]) compared to sites where grazing was permitted (streamside transect: 21 cm [8.2 in]; upland transect 14 cm [5.4 in]). There was moderate to suggestive evidence that cover provided by forbs, alder, willow, and other shrubs was greater where livestock grazing was not permitted, while cover provided by rush and “other” was greater where livestock grazing was permitted. There were eight surveys where Arizona montane voles were captured and livestock grazing was permitted. Two were in Arizona (Fish Creek and West Fish Creek). In New Mexico, voles were captured at three of these sites (Jenkins Creek lower, “Flanagan Creek” upper, San Francisco River) during 2004, but not 2020. Montane voles were captured at two sites with permitted grazing in 2020 (Jenkins Creek middle, “Flanagan Creek” lower). Romero Creek was the only site where livestock grazing was permitted and Arizona montane voles were captured in both 2004 and 2020.

The successful captures of Arizona montane voles fill in important distributional gaps for the subspecies in New Mexico and suggest that the vole’s historical distribution encompassed most of the drainage network of the upper San Francisco River watershed. Despite extensive searches for vole habitat, we found few places with vole signs and captured Arizona montane voles at a minority of those sites. Our results indicate that the Arizona montane vole in New Mexico has declined since 2004 and that it is currently in jeopardy of becoming extirpated in the state. Compared to distribution in 2004 and 2008, there was an 83% decline in occurrence of Arizona montane voles in New Mexico in 2020. Of particular concern, in 2020, we captured montane voles at only one location within the Centerfire Creek subdrainage, where they were rare. The Centerfire Creek subdrainage is the largest watershed in the study area and is likely isolated from populations in other subdrainages by intervening low elevations.

Habitat for Arizona montane voles is characterized by moist soil with tall, dense herbaceous plant cover, especially provided by sedges. Based on 95% CIs, there appears to be a minimum mean vertical cover requirement for these voles of approximately 43 cm (=17 in) across the stream reach and 28 cm (=10.9 in) at local used patches (represented in this analysis by the 4-m radius microhabitat plots). In New Mexico in 2020, none of the stream reaches met these conditions, and many sites lacked any patches of vegetation that met the lower threshold for microhabitat, both indicating highly degraded riparian conditions. The evidence suggests that the loss of adequate vertical cover is a proximate driver of population extirpation of Arizona montane voles. Further, the evidence suggests that the ultimate reason for this loss of vertical cover was livestock grazing. Across the range of the Arizona montane vole, most occupied sites were in areas where livestock grazing was not permitted, and capture rates of the voles was greater in such areas.

Results of this study demonstrated that livestock grazing management on Forest Service allotments generally did not provide the vegetation requirements for Arizona montane voles in New Mexico. Similar conclusions have been made for other species of mammals in Arizona and New Mexico (e.g., Frey and Malaney 2009; Small et al. 2016; Horncastle et al. 2019). Vegetation conditions and the associated mammal community at most sites sampled in this study during 2004 and 2020 had declined. With the exception of grazing exclosures, the only instance of improved vegetation conditions between 2004 and 2020 was Romero Creek (Figure 6). In 2004, the stream was mostly dry, there were few riparian plants that did not provide adequate cover, and only a single Arizona montane vole was captured that was using a log as cover for its runways. In contrast, in 2020 the stream was mostly continuous and had abundant fishes, while the riparian zone included extensive moist soil and patches of tall, dense herbaceous riparian vegetation and Arizona montane voles were relatively abundant. The improved conditions were likely due to a change in the way cattle were permitted to graze in the Romero Pasture that occurred in 2003 and 2004. In 2003, permitted livestock numbers were decreased to 185 cow/calf pairs, an almost 17% decrease from prior numbers. Then in 2004, the season of use changed from the growing season (portions of May-August) to the post-monsoon growing season period (portions of September-October). Late growing season livestock grazing is generally considered detrimental to riparian vegetation and is similar in impact to growing season grazing, which is universally considered detrimental to riparian vegetation in the American Southwest (Baker 2001; Cram et al. 2018). Thus, the overall benefit of the changed management is likely attributed to the overall total reduction of use (i.e., combination of reduction in time and numbers) with possible additional benefits due to avoiding grazing in the hot season (Swanson et al. 2015). Although riparian conditions at Romero Creek had vastly improved, the vertical cover in 2020 was still low compared to occupied sites throughout the vole’s range. The relatively low vertical cover on Romero Creek is likely because it continues to be depauperate in riparian shrubs. The near absence of riparian shrubs may be due to the late season grazing, which is considered especially detrimental to woody plant species (Baker 2001; Cram et al. 2018; Swanson et al. 2015) and possibly lack of upstream sources for reestablishment of alders and willows. Regardless, the upward trends of habitat conditions on Romero Creek indicate that range management can improve conditions for the vole. The upward trends also suggest that the deteriorating conditions at other sites is not due to general climate patterns and further bolsters the argument that habitat declines and extirpations of local populations of Arizona montane voles is due to livestock grazing.

Exclusion of cattle from riparian zones by using corridor fencing can result in rapid restoration of riparian vegetation in the American Southwest (Baker 2001; Cram et al. 2018). In our study area, vegetation in the grazing exclosures we sampled sharply contrasted with the denuded nature of the majority of stream reaches. It is likely that recovery of Arizona montane voles will require additional restoration efforts in the upper San Francisco River watershed. However, recovery of Arizona montane voles requires more than restoration of the habitat. For instance, we captured both Arizona montane voles and Mogollon voles in a grazing exclosure on SA Creek in 2004, but in 2020 did not capture any voles in that exclosure or a more recently erected exclosure in an adjacent area. Vertical cover appeared similar across years. One possible reason for the apparent extirpation of this population was unauthorized livestock grazing as we observed ample signs of cattle grazing inside the exclosure. In addition, we observed signs of flooding and debris accumulation that appeared to derive from erosion of the degraded uplands in the watershed. Vole populations are prone to large fluctuations in population size in unstable environments (Pinter 1988; Sera and Early 2003). Thus, extirpation of populations is possible in grazing exclosures if they are small and prone to disturbance. Voles have limited dispersal capabilities (Wolff 1985) and, consequently, if no nearby source population is present, restored habitat cannot be naturally recolonized. In these instances, translocations would be necessary to restore the populations.

Horncastle et al. (2019) evaluated occupancy of Arizona montane voles and other small mammals in the White Mountains of Arizona. Arizona montane voles were the second most commonly captured species in their study, and they captured twice as many in ungrazed meadows compared to grazed meadows. Their top model contained only vegetation height as a predictor of occupancy of Arizona montane voles and this variable also appeared in most models with ΔAICc < 4.0. However, this set of models contained the null (i.e., ΔAICc = 1.37), as well as more complex versions of the top model, which may signal that the additional variables are uninformative (Arnold 2010; Leroux 2019). There are several possible reasons that their study did not find stronger predictors of occupancy for Arizona montane voles. First, as demonstrated herein, vole species are difficult to identify, even with prepared museum specimens. Misidentification of Mogollon voles as montane voles during their field study could have obscured habitat relations of the montane vole in their occupancy models, especially given the Mogollon vole’s tolerance for more arid conditions. Second, compared to New Mexico, Arizona has higher elevations and relatively more areas where livestock grazing is not permitted; thus, it is possible their study did not include as much marginal habitat as was included in this study and, consequently, Arizona montane voles were more uniformly present at their survey sites. Occupancy models do not perform well when occupancy probability is high (MacKenzie et al. 2018). Occupancy estimates were not presented in Horncastle et al. (2019) in order to further evaluate this hypothesis, although the abundance of voles reported in their study suggests that high occupancy could have influenced results. Lastly, in our study, vertical cover was an important variable distinguishing sites where Arizona montane voles were captured or not captured, while in Horncastle et al. (2019) their top variable was vegetation height, which was measured the same as vertical stubble height in this study. Vertical cover and vertical height do not measure the same structural aspect of vegetation. Where both measurements have been used to describe small mammal habitat, vertical cover was the more important variable for species that require vegetation cover (Frey and Malaney 2009). Vegetation height (= vertical stubble height) is a measurement taken with a ruler from the base of a plant to the tip of the leaves when fully extended vertically, while vertical cover is measured by viewing a Robel pole from 4 m away and recording the lowest 1-inch band not obscured by vegetation or other habitat structure. In this study, the two variables taken on microhabitat plots had only moderate correlation (r = 0.637), and they became more dissimilar with taller vegetation (Figure S1). It is possible for tall plants to fail to supply ample vertical cover if the plants are sparse. In addition, in some species of plants, such as some sedges, the leaves extend outward and droop over, forming a natural dense ceiling above the ground level. Thus, pulling the leaves vertical obscures an important aspect of the vegetation that may be functional to voles. The critical habitat element for Arizona montane voles is dense vegetation cover above the ground (i.e., vertical cover), which is necessary to conceal their runways.

The Arizona montane vole is a bellwether of herbaceous vegetation conditions. In healthy riparian zones throughout its range, Arizona montane voles are typically the most abundant small mammal species. Because of their usual high population densities, they are likely critical to ecosystem function, for instance serving as prey to a host of small predators such as hawks and owls, improving soil conditions through burrowing activity, and influencing plant community composition and structure, such as through the dispersal of hypogeous fungi (Frey 2018 Reference S2). Further, riparian zones that are incapable of supporting Arizona montane voles are likely incapable of supporting an entire range of other species, such as shrews Sorex spp., jumping mice, weasels Neogale frenata, and leopard frogs Lithobates spp. The current status of Arizona montane voles in New Mexico is a testament to these degraded stream systems. Restoration of habitat for Arizona montane voles is likely to benefit myriad species.

We conclude that the Arizona montane vole is at risk of becoming extinct in New Mexico and that urgent management actions are needed to prevent its extirpation in addition to action-oriented conservation planning to recover the species. Based on our results and the conclusions of other studies, restoring habitat for these voles requires curtailment of livestock grazing though practices such as riparian exclosures and allotment retirement, or use of increased management to limit livestock numbers and duration of grazing in riparian zones. We recommend excluding livestock from springs, bogs, and fens while piping drinking water to adjacent uplands. Riparian exclosures could be implemented with fencing gaps or piped water troughs to allow livestock access to water. We recommend regular monitoring of existing exclosures to remove livestock and repair fencing. We recommend implementing stream restoration techniques such as construction of beaver dam analogs that can restore incised streams and create more riparian habitat (Pollock et al. 2014). Restoration of upland conditions may be necessary in some watersheds (upper Centerfire Creek) to improve riparian health by reducing erosion. The population of Arizona montane voles on Jenkins Creek is especially important to the overall conservation of the vole in New Mexico as it is the only known persisting population within the Centerfire Creek subdrainage, which is likely isolated from other populations. We recommend urgent management actions to prevent extirpation of this population. We recommended additional surveys, especially within the Centerfire Creek subdrainage to locate any other persisting populations. Further, because montane vole populations are influenced by spring precipitation (Pinter 1988) and may fluctuate, we recommend long-term monitoring of populations. Because vast areas of likely historical habitat appear to no longer have populations of montane voles, translocations may be necessary to restore populations and recover the species. Any such translocations would benefit from genetic information. Finally, we recommend that riparian restoration should achieve a minimum of 17 in mean vertical cover along stream reaches taken with a Robel pole every 20 m, 0.5 m from the stream edge. This target will ensure that riparian vegetation is functioning to support most small mammal populations, including Arizona montane voles.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Table S1. Archive of landscape stream-reach cover data collected at sites surveyed for the Arizona montane vole Microtus montanus arizonensis in Apache and Greenlee counties, Arizona, 2008–2009, and Catron County New Mexico, 2020.

Available: https://doi.org/10.3996/JFWM-24-020.S1 (18.4 KB)

Table S2. Archive of microhabitat data collected at sites surveyed for the Arizona montane vole Microtus montanus arizonensis in Apache and Greenlee counties, Arizona, 2008–2009, and Catron County New Mexico, 2004 and 2020.

Available: https://doi.org/10.3996/JFWM-24-020.S2 (13.1 KB)

Reference S1. Frey, J.K. 2018. Beavers, livestock, and riparian synergies: bringing small mammals into the picture. Pages in Riparian Research and Management: Past, Present, and Future. Volume 1. US Forest Service, Rocky Mountain Research Station, General Technical Report, RMRS-GTR-377.

Available: https://doi.org/10.2737/RMRS-GTR-377-CHAP6 (30 August 2024; 363 KB PDF)

Reference S2. [USFWS] U.S. Fish and Wildlife Service. 2014a. Final rule: determination of endangered status for the New Mexico meadow jumping mouse throughout its range. Federal Register 79(111):33119–33137; 10 June 2014.

Available: https://www.gpo.gov/fdsys/pkg/FR-2014-06-10/pdf/2014-13094.pdf (30 August 2024; 285 KB PDF)

Figure S1. Relationship between mean vertical cover (in; 1 in = 2.5 cm) and mean vertical stubble height (mm) on microhabitat plots. Above 10 in of vertical cover, the relationship between the two variables becomes diffuse. Vertical cover is a more meaningful measure of habitat for Arizona montane voles Microtus montanus arizonensis than vertical stubble height.

Available: https://doi.org/10.3996/JFWM-24-020.S1 (32.5 KB)

We thank the Share with Wildlife Program and State Wildlife Grant T-67-R-1 for funding this research and V. Seamster for assistance. We thank the Gila National Forest, Quemado Ranger District for supporting this study and T. Hendricks for helpful information about livestock management. We thank three reviewers and the associate editor for helpful comments that improved the paper.

Any use of trade, product, website, or firm names in this publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.