Bat use of abandoned mines throughout Nevada
Abstract
Abandoned mines offer important roosting habitat for several species of bats throughout the western United States. Currently, abandoned mine reclamation programs are tasked with closing abandoned mines to ameliorate safety, health, and environmental hazards found in and around these sites. Without appropriate pre-closure evaluations to determine use of mine workings prior to closure, bats that depend on abandoned mines may be negatively impacted. To mitigate impacts of abandoned mine reclamation on bats, surveyors typically conduct pre-closure biological evaluations and recommend wildlife compatible closures (e.g., bat gates) for ecologically important sites. Due to hazardous conditions found in many abandoned mines, internal surveys cannot always be conducted, and external surveys are not reliable for determining underground habitat or inferring past, future, or potential use of mines by bats when they are absent during external surveys. The purpose of our study was to use internal mine surveys to examine relationships between abandoned mine use by bats and characteristics of the mine and landscape, including portal area and shape, number of portals, mine depth, elevation, proximity to water and land use type. We found that surface features including land use type, distance to water, and elevation were associated with bat use, as were several mine features including depth of workings and portal shape. To best conserve sensitive species of bats, it is essential that pre-closure biological evaluations be as detailed as possible to enhance biological understanding of species' roosting associations and distribution throughout the landscape. Further information will best facilitate development of ecologically sound closure recommendations for abandoned mine openings.
Underground mining has greatly impacted landscapes of North America through modifications of landscapes, intensified use of resources (e.g., timber, water), and creation of new subterranean habitat across the continent (Francaviglia 1997, Green 1997). As underground mining expanded westward during the 19th and 20th centuries, bats rapidly began occupying the new, open niche spaces (McAney 1999, Sherwin et al. 2009). As mining operations became unprofitable, many of the complexes were abandoned and left exposed to the elements. Lack of maintenance following abandonment has led to destabilization of abandoned mine openings and subterranean workings, making them prone to collapse. Additionally, lack of internal ventilation increases likelihood of dangerous internal atmospheres (e.g., anoxic conditions; Sherwin et al. 2009). The Surface Mining Control and Reclamation Act supports efforts to close abandoned mines to protect people from hazardous conditions within and around mine workings (Green 1997). However, for many bat species, including Townsend's big-eared bats (Corynorhinus townsendii), and several species of Myotis, abandoned mines offer important roosting habitat for hibernation (overwintering sites), maternity use (where females give birth and raise young), and day (where males and sometimes nonreproductive females roost) and night roosts (resting stops during the night; Humphrey and Kunz 1976, Kunz 1982, Sherwin et al. 2000, Morrison and Fox 2009, Ingersoll et al. 2010).
One method to mitigate loss of habitat within abandoned mines is installation of bat-compatible closures (White and Seginak 1987). When abandoned mines are identified as ecologically important habitat for bats, portals are typically secured with steel gates. Bat-compatible closures allow bats continued access to roosts, while precluding human access (White and Seginak 1987, Sherwin et al. 2009). To determine which mines should be gated, wildlife managers require pre-closure surveys to diagnose how each site is used by bats (Altenbach and Pierson 1995, Sherwin et al. 2009, Neubaum et al. 2017). During internal surveys, researchers enter the underground environment and search for bats and associated signs of use (guano, culled insect parts, and other biological residues) that reveal past use even when bats aren't present at time of survey. Additionally, internal surveys reveal the underground habitat available in each mine, allowing for more informed understanding of the potential for future use by bats (Altenbach and Pierson 1995, Sherwin et al. 2009, Loeb et al. 2015). However, due to the dangerous conditions inside mines, such as internal instability and collapse, sudden drop offs, and bad air, full internal surveys are not always possible or desired (Altenbach and Pierson 1995, Sherwin et al. 2009, Loeb et al. 2015).
In contrast, external surveys take place outside of the mine where openings are monitored for bat activity. Bats must be present and exiting through monitored openings during external surveys to confirm actual use of the mine by bats. There are limitations to external activity surveys and results are based on the assumptions that all mine openings are known and concurrently monitored, any use of a site by bats can be captured at the time of survey, and that the method of survey (night vision, thermal imaging, netting, etc.) can accurately detect use (Loeb et al. 2015). External surveys can be a valuable method for use at confirmed roosts when colony estimates are needed to monitor trends or to determine which of multiple entrances are used by bats (Sherwin et al. 2009, Loeb et al. 2015). Presence data (bats present at time of survey), depending on the time of year, can help to assess species composition or how known roosts are used (i.e., maternity, swarming; Loeb et al. 2015, Neubaum et al. 2017); however, there is no way to interpret importance of absence data. Absence of bats exiting the opening at time of survey cannot be used to infer a lack of use by bats and so external surveys are not generally an effective method for assessing biological importance of abandoned mines and/or developing closure recommendations (Sherwin et al. 2009). Because of the dangerous conditions within abandoned mines, internal surveys are not always possible, and closure recommendations must be developed based on externally collected data. Therefore, if more permanent external characteristics, such as portal size and shape, can provide a reliable indication of potential use by bats, surveyors may more efficiently target sites in need of internal survey or develop more robust closure recommendations when an internal survey is not possible.
Of 23 bat species found in Nevada, 13 were included on the Nevada Wildlife Action Plan's 2012 Species of Conservation Priorities list. Of these 13, 8, including Allen's big-eared bat (Idionycteris phyllotis), California leaf-nosed bat (Macrotus californicus), cave myotis (Myotis velifer), fringed myotis (Myotis thysanodes), western small-footed myotis (Myotis ciliolabrum), long-eared myotis (Myotis evotis), little brown bat (Myotis lucifugus), and C. townsendii use subterranean features as roosts during all or some parts of the year (Wildlife Action Plan Team 2012). Myotis lucifugus is categorized as endangered by the IUCN red list following large population declines in eastern portions of its range due to white nose syndrome (Cheng et al. 2021, Solari 2021). Additionally, M. ciliolabrum is listed as at risk or vulnerable in Alberta and British Columbia, and M. evotis is included in the Mexican regulation for species conservation NOM-59-SEMARNAT-2001 (Holloway and Barclay 2001; Arroyo-Cabrales and Álvarez-Castañeda 2017a, b). Due to the strong association of these species of bats with abandoned mines throughout Nevada, the reclamation of biologically critical roosting habitat poses a potential threat to the conservation of subterranean-roosting bats throughout the state (McAney 1999, Morrison and Fox 2009, Sherwin et al. 2009, Wildlife Action Plan Team 2012).
The purpose of our study was to test whether surface variables, including mine features (i.e., portal size, portal shape, number of portals, and elevation) and surface conditions (i.e., land use type and proximity to water), collected without requiring entry into abandoned mines, were correlated with actual use of abandoned mines by bats in Nevada. We also included one internal variable, mine depth, in our analysis. Actual use was determined through internal surveys based on observation of roosting bats and biological residues (i.e., guano, culled insect parts, staining) resultant from roost use. If surface variables are correlated with use of abandoned mines by bats, these variables could be used as predictors of bat use, possibly reducing the need for internal survey of potentially dangerous mine workings. Additionally, surface variables correlated with bat use of mines could better inform large scale reclamation projects by providing landscape scale predictors of bat occupancy across a broader geographical context.
STUDY AREA
Our study was conducted throughout the state of Nevada, USA, with greatest intensity in Clark, Nye, Pershing, and White Pine counties. Most of Nevada is classified as desert shrubland and most of the locations in this study were in open woodland (grazed) and desert shrubland (grazed and ungrazed) areas (U.S. Department of Agriculture [USDA] 2018).
METHODS
Data were collected at 178 abandoned mines between September 2016 and October 2018. Across all years, surveys were conducted between July and December, and most (84%) were conducted between July and October. Each site was surveyed once, following internal survey and safety protocols developed by Sherwin et al. (2009), including wearing proper personal protective equipment, having backup light sources, and using gas meters. Surveyors walked slowly through each mine scanning for hazards, such as winzes or collapsing areas, and then looked for bats, guano, or insect parts. Global Positioning System coordinates were measured using a GPSMAP® 64sx (Garmin®, Olathe, KS, USA) handheld GPS. Portal size, measured using a laser distance meter, was calculated as height × width while portal shape was calculated as the ratio of height to width. For mines with multiple portals, portal area was considered the combined area of all portals. Total area of all openings quantified the spatial breadth of subterranean accessibility from surface habitats. Elevation was determined using the USGS Elevation Point Query Service, which was accessed through the Elevatr package in R Studio (Hollister et al. 2022, R Core Team 2022) and depth was calculated via step count and walking stride length measurement. Abandoned mines were considered to be used by bats if they contained bats, guano, substrate staining, and/or culled insect parts.
We used a vector shapefile (scale = 1:7,500,000) from the USDA in ArcMap to determine which land use type in which each mine was located (USDA 2018; Figure 1). To analyze the relationship among mines used by bats and proximity to water, we used a shapefile layer, USA Detailed Water Bodies, in ArcMap (Esri, Redlands, CA, USA; https://www.arcgis.com/home/item.html?id=48c77cbde9a0470fb371f8c8a8a7421a) from the USGS in partnership with the U. S. Environmental Protection Agency. The USA Detailed Water Bodies layer provides a database of perennial water bodies and reaches that comprise the surface water drainage systems of the United States. We used Euclidean distances of mines to water sources in our analysis.
Before developing our models, we tested explanatory variables, including portal area and shape, number of portals, mine depth, elevation, land use, and distance to water, for correlations to avoid model overfitting. We used a Pearson's product moment correlation with a threshold of 0.6 to test all pairs of continuous variables for correlation. We then used Kruskal-Wallis Tests and one-way ANOVAs to test for correlations between categorical variables and other variables. If any pair of variables were correlated, we did not include them together in the same model. According to a Kruskal-Wallis Test, proximity to water is correlated to land use (X2 = 18.69, df = 2, P < 0.001). Although not linearly correlated, a one-way ANOVA revealed that there was a statistically significant difference in elevation between land use types (Fdf,df = 85.98, P < 0.001). Using a Tukey's HSD Test, we found that the mean value of elevation differed significantly between each land use type (desert shrubland grazed—open woodland grazed, P < 0.001; desert shrubland mostly ungrazed—open woodland grazed, P < 0.001; desert shrubland mostly ungrazed—desert shrubland grazed, P < 0.001; Figure 2). Therefore, we built one model that excluded land use and a second model that included land use and excluded distance to water and elevation.
We utilized logistic regression analyses to determine the effects of land use, proximity to water, portal size (height × width), portal shape (ratio of height/width), number of connected portals, mine length/depth, and elevation on bat use of mines. We ran individual models on all variables and models containing all variables excluding land use (M1) and all variables excluding elevation and distance to water (M2). We used a Wald Chi-square Test to determine if land use was significant in models in which it was included. We evaluated models based on Akaike's Information Criteria (AIC) and considered models within ∆AIC < 2 informative (Burnham and Anderson 2002). All statistical analyses were conducted using R (v. 4.1.2; R Core Team 2022), with a significance threshold of 0.05.
RESULTS
Approximately 56% (100/178) of mines surveyed exhibited signs of use by bats. Eighteen of the mines were occupied by bats at the time of the survey. Corynorhinus townsendii were identified in 13 of these mines, Myotis sp. were identified in 2 sites, and specific use of the remaining 3 sites could not be diagnosed. All surveyed abandoned mines occurred within 1 of 3 land use types, as categorized by the USDA: open woodland grazed (piñon, juniper, aspen groves, chaparral and brush), desert shrubland grazed, or desert shrubland mostly ungrazed (Figure 1). Because ∆AIC was <2 for both multivariate models, they were considered equally informative (Table 1). According to the Wald Chi-square tests, land use type was a significant predictor of bat use of abandoned mines in multivariate (M1: X2 = 10.1, P < 0.01) and univariate (X2 = 6.8, P < 0.05) logistic regressions. In the univariate model, the open woodland grazed land use type was positively associated with use (M9: Z = 2.34, P < 0.05) and the desert shrubland mostly ungrazed land use type was negatively associated with use (M9: Z = −2.55, P < 0.01; Figure 3). Distance to water was also significant in both the multivariate (M2: Z = −2.48, P < 0.05) and univariate (M8: Z = −2.74, P < 0.05) logistic regression models, with mines closer to water exhibiting more signs of use. Additionally, higher elevation mines were more associated with use as elevation was significant in both multivariate (M2: Z = 2.44, P < 0.05) and univariate models (M7: Z = 1.96, P ≤ 0.05; Table 1).
| Model | Variables | P-value | z-score | AIC |
|---|---|---|---|---|
| M1 | Area | 0.03 | −2.22 | 219.10 |
| Shape | 0.01 | 2.77 | ||
| Portal number | 0.42 | 0.79 | ||
| Depth | <0.001 | 3.40 | ||
| Land use | Wald test: P < 0.01 | X2 = 10.1 | ||
| M2 | Area | 0.08 | −1.75 | 219.10 |
| Shape | 0.01 | 2.61 | ||
| Portal number | 0.44 | 0.77 | ||
| Depth | <0.01 | 2.99 | ||
| Elevation | 0.01 | −2.44 | ||
| Distance to water | 0.01 | 2.48 | ||
| M3 | Area | 0.19 | −1.32 | 246.15 |
| M4 | Shape | 0.05 | 1.93 | 243/67 |
| M5 | Portal number | 0.39 | 0.87 | 247.11 |
| M6 | Depth | <0.01 | 3.06 | 231.16 |
| M7 | Elevation | 0.05 | 1.96 | 244.11 |
| M8 | Distance to water | <0.01 | −2.74 | 239.65 |
| M9 | Land use | Wald test: P < 0.03 | X2 = 6.8 | 243.00 |
| Open woodland grazed | 0.02 | 2.34 | ||
| Desert shrubland grazed | 0.22 | −1.22 | ||
| Desert shrubland mostly ungrazed | 0.01 | −2.55 |
Portal area ranged from just over 0.3 m2 to 183.79 m2 (Table 2). Alone, portal area was not a significant predictor of mine use, nor was it significant in a model with elevation and distance to water, but it was significant in the model with land use (M1: Z = −2.22, P < 0.05; Table 1). Portal shape, a ratio of height/width, ranged from 0.2143 to 4.375 (Table 2). Shape was also not a significant predictor of use in the univariate model but was in both multivariate models (M1: Z = 2.77, P < 0.05; M2: Z = 2.61, P < 0.05; Table 1). Number of portals, which ranged from 1 to 9, was not a significant predictor of use in any models (Tables 1 and 2). Depth, which ranged from 1.83 m to 320.04 m, was significant in all models, with deeper mines exhibiting a greater likelihood of use (M1: Z = 3.40, P < 0.001; M2: Z = 2.99, P < 0.05; M6: Z = 3.06, P < 0.001; Tables 1 and 2).
| Continuous variables | Use | No Use | ||||||
|---|---|---|---|---|---|---|---|---|
| Min | Max | Mean | Standard deviation | Min | Max | Mean | Standard deviation | |
| Portal area (m2) | 0.91 | 183.79 | 13.03 | 25.34 | 0.38 | 121.92 | 18.04 | 22.78 |
| Portal shape | 0.21 | 4.38 | 0.98 | 0.55 | 0.25 | 1.67 | 0.84 | 0.34 |
| Number of portals | 1 | 9 | 1.25 | 1.00 | 1 | 3 | 1.14 | 0.39 |
| Depth (m) | 4.57 | 320.04 | 33.25 | 44.21 | 1.83 | 233.78 | 14.98 | 26.71 |
| Elevation (m) | 1,004.01 | 2,975.15 | 1,917.14 | 503.01 | 978.41 | 2,715.55 | 1,832.46 | 405.80 |
| Distance to water (m) | 1,512.72 | 41,724.07 | 12,029.85 | 6,686.68 |
1,264.01 | 41,703.96 | 15,275.97 | 8,165.04 |
DISCUSSION
Of the 100 abandoned mines that exhibited signs of use, only 18 were occupied by bats at the time of the survey. If surveyors were to solely rely on external activity surveys, then only 18% or fewer of the 100 mines in our study in which we documented use would have been considered occupied. External activity surveys are dependent on bats being present and recorded at time of survey. As a result, presence data are meaningful and provide insight into actual use of surveyed workings; however, lack of bat activity at the time of survey (negative data) cannot be used to diagnose a site as unused by bats. Assumptions due to lack of observed bat activity are particularly problematic as some species, such as C. townsendii, exhibit a high degree of movement among roosts as well as intraseasonal and interseasonal variation in timing of roost occupancy (Sherwin et al. 2000, 2003, 2009). Similarly, M. lucifugus have exhibited a high degree of movement with 35.5% of tagged females using multiple roosts, separated up to 6.1 km, in a maternity season (Slough and Jung 2020). Roost lability as well as intraseasonal variation in timing of roost occupancy means that even biologically critical roosts can be vacant much of the time.
We did find that some external variables were positively correlated with use of abandoned mines by bats. Abandoned mines in open woodland grazed habitat (pinyon, juniper, aspen groves, chaparral, and brush) were more likely to be used than those in the desert shrubland mostly ungrazed habitat. This association with open woodland grazed habitat is consistent with historical netting and tracking data of some Nevada bat species (Kuenzi et al. 1999, Fellers and Pierson 2002). Corynorhinus townsendii are associated with riparian habitats within open woodland areas (Fellers and Pierson 2002). Similarly, in mist netting surveys conducted over water sources in west central Nevada, ranging from approximately 1,524 to 2,460 m, more bat species were associated with higher-elevation water sources surrounded by pinyon-pine juniper (Kuenzi et al. 1999). Mines in our study ranged from 978 to 2,975 m. Mines used by bats were associated with relatively higher elevation, which was also correlated with the open woodland grazed land use type.
In addition to habitat type, the data support a positive association between proximity of abandoned mines to perennial water sources and roost use by bats. The closer a site was to water, the more likely it was to be used by bats. However, this could vary depending on roost type, which was not considered in this study. Night roosts are often chosen opportunistically within close range of resources such as foraging areas and water (Hirshfeld et al. 1977, Anthony et al. 1981, Kunz 1982, Lacki et al. 2007, Ormsbee et al. 2007). Some biologically important roosts, such as maternity or hibernacula, may not be as close to these resources (Arbuthnott and Brigham 2007, de Boer et al. 2013). Big brown bats (Eptesicus fuscus), for instance, have been documented traveling up to 11 km from maternity roosts to foraging sites, with the insect communities around these foraging sites differing from those around the roosting area (Arbuthnott and Brigham 2007). Species distribution models suggest that landscape structure influences bat species distribution and that scale and landscape pattern affects species differently (Ducci et al. 2015). Ultimately, because of the importance of land use and distance to water, proximity to resources are likely good predictors of bat use of abandoned mines in this region.
Bat use of mines was associated with smaller portals that had a larger height to width ratio (i.e., tall and narrow). The size and shape of the entrance, as well as internal passages, influence microclimatic conditions such as airflow, temperature, and humidity, which often influences roost selection, especially of hibernation and maternity roosts (Kunz 1982, Raesly and Gates 1987, Hurst and Lacki 1999, Johnson et al. 2006, Ingersoll et al. 2010). Entrance shape and size preference may vary by species and roost type. For instance, C. townsendii in Utah were found to be associated with lower entrance heights (Sherwin et al. 2000), while research in West Virginia found that bat occupancy was more strongly associated with taller and narrower mine openings (Johnson et al. 2006), and tricolored bats (Perimyotis subflavus) in Iowa prefer hibernacula with vertical openings (i.e., shafts; Dixon 2011). Smaller, taller portals may be related to optimal internal microclimate conditions for bats in this region (Kunz 1982, Raesly and Gates 1987, Hurst and Lacki 1999, Johnson et al. 2006, Ingersoll et al. 2010).
Larger mines were also more likely to be used by bats than mines with smaller internal dimensions. Associations with larger mines may be due to a greater array of microclimatic conditions typically found in larger mines or may simply result from additional space available to house larger colonies (Kunz 1982, Tuttle and Stevenson 1982, Willis et al. 2006). Roosting-group size of E. fuscus, for example, was positively correlated with cavity size (Willis et al. 2006). If substrate surface area is positively correlated with bat use, the increased use of large mines may be a simple reflection of their increased contribution of roosting habitat relative to smaller mines. In some landscapes, the largest mines may be only modestly larger than the smallest mines in the area. As such, the largest mines in these areas may be used sporadically, and less predictably than truly large, complex mines (Sherwin et al. 2003), in which case discrete external surveys are more likely to miss seasonal activity and roost occupancy. It should also be noted that smaller or shorter mines do not necessarily mean that they are unimportant as many of the smaller mines included in this study had evidence of bat use. In reality, many smaller workings are ecologically important to bats, often as night roosts for consuming prey, though a larger proportion of large mines are used by bats than smaller mines (Altenbach and Pierson 1995, Sherwin et al. 2009, Loeb et al. 2015). Because smaller mines are more abundant in historically mined landscapes, their net contribution to the roosting landscape is similar to that of the fewer, large mines available (Kunz 1982, Sherwin et al. 2003, Dixon 2011).
Unfortunately, internal depth and complexity of workings cannot be directly measured unless an internal survey is conducted. Similarly, direct evidence of bat occupancy (guano) remains unavailable without internal survey. If internal surveys are not possible, lack of external activity at any given time is not evidence of nonuse. An internal survey, on the other hand, would detect indicators of bats even when no bats are present or when bats are hibernating (which cannot be diagnosed from the outside of the mine). However, due to the dangerous conditions within abandoned mines, internal surveys are not always possible, and closure recommendations must be developed based on externally collected data. However, data are not as comprehensive as internal survey data, so it is critical that external surveyors carefully evaluate the surrounding landscape for any clues from which internal structure and habitat may be inferred. Airflow at openings is typically indicative of interconnected openings and should be carefully considered during surface evaluations (Altenbach and Pierson 1995, Sherwin et al. 2009). Additionally, a large mine dump or rock waste pile implies extensive internal workings, as can historical objects such as rails, wire ropes, and remains of support structures (Francaviglia 1997, Sherwin et al. 2009). Oftentimes, however, surface artifacts have been removed and even large mines can lack external evidence proportional to internal workings.
To best conserve sensitive populations of bats, it is essential that biological evaluations of potential roosts and closure recommendations be as detailed as possible and that biologists understand species' roosting associations and distribution throughout the landscape. Mine structure, including depth and portal size and shape, has an effect on roost associations, possibly because of the effect these features have on internal microclimate conditions (Kunz 1982, Raesly and Gates 1987, Hurst and Lacki 1999, Johnson et al. 2006, Ducci et al. 2015). When internal surveys cannot be conducted, land use type, proximity to water, and evidence of complex inner workings should be carefully considered when making closure recommendations as these factors are indicative of bat use. However, if these features are lacking, more external monitoring (video, game cameras, acoustic detectors inside portals, etc.) should be conducted to determine use, or lack thereof, before making a closure recommendation.
ACKNOWLEDGMENTS
We would like to thank L. Hancock, K. Ekholm, J. Danielson, C. Devine-Rosser, B. Thompson, and B. O'Conner for their help on this project. We thank the National Park Service, the Cooperative Ecosystems Studies Unit, the Nevada Department of Wildlife, and Christopher Newport University for helping fund this project.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflicts of interest.
ETHICS STATEMENT
All animals in this study were treated humanely following recommendations of the American Society of Mammalogists and in accordance with the approval of the Christopher Newport University Animal Care and Use Committee (IACUC-2018-7) and Nevada Department of Wildlife scientific collection permit #39645.
Open Research
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
REFERENCES
-
Associate Editor: A. Piaggio.
