BioOne.org will be down briefly for maintenance on 17 December 2024 between 18:00-22:00 Pacific Time US. We apologize for any inconvenience.
Open Access
How to translate text using browser tools
1 September 2010 Bat community structure within riparian areas of northwestern Georgia, USA
Joshua B. Johnson, W. Mark Ford, John W. Edwards, Michael A. Menzel
Author Affiliations +
Abstract

Although it is well known that bats commonly forage in riparian areas, which provide water resources and insect concentrations, the role that the physical structure of riparian areas plays in influencing local bat communities is less certain. In 2000–2002, we used acoustic monitoring to determine bat species presence at 338 riparian sites in northwestern Georgia, USA. We used a 2-dimensional nonmetric multidimensional scaling (NMDS) ordination to assess how separations among species were partially associated with riparian conditions. Our NMDS analysis found some degree of habitat partitioning among bat species occurring in northwestern Georgia and was dictated in part by riparian condition. Myotis grisescens and M. septentrionalis were associated with low-elevation lotic waterways, whereas M. lucifugus, Lasiurus borealis, and Eptesicus fuscus were associated with high-elevation lentic waterways with sparse canopy cover. However, riparian conditions had weak relations with NMDS axes, possibly resulting in coincidental associations in some cases. Regression tree analysis indicated that higher bat species richness was associated with apparently uncommon small, high-elevation waterways with sparse canopy cover as well as larger streams and rivers that had wetlands adjacent to them. Including high-elevation waterways with existing management recommendations for endangered M. grisescens foraging areas (large, low-elevation streams and rivers) will be the most effective conservation strategy to benefit the most bat species in northwestern Georgia and probably elsewhere in the southern Appalachians.

Introduction

It is widely accepted that bat foraging activity is commonly greatest in areas associated with water sources, both lentic and lotic systems, as well as wetlands (Racey 1998, Grindal et al. 1999). These areas provide drinking water and abundant insects for foraging bats (Racey et al. 1998). Moreover, roost trees often occur in adjacent riparian zones, reducing commuting distances to foraging areas (Daniel et al. 2008). Linear riparian features such as wooded buffers along streams and rivers that serve as foraging habitats can also serve as efficient commuting corridors (LaVal et al. 1977, Menzel et al. 2005). Despite the importance of riparian areas as bat habitat, very few studies have examined the association between bat activity patterns and the physical structure and condition of riparian areas (Ciechanowski 2002, Lloyd et al. 2006, Williams et al. 2006, Ober & Hayes 2008). It has been suggested that retaining a diversity of seral stages, i.e., shrub-stage and mature forest, in riparian areas could provide foraging habitat for most bat species in a given area (Lloyd et al. 2006, Williams et al. 2006, Ober & Hayes 2008). A diversity of riparian habitat condition is necessary because of the differences in foraging habitat selection among bat species, which spatially partition themselves among various cover types to reduce competition for food resources (Kunz 1973, LaVal et al. 1977).

Foraging habitat selection by bats is dictated by physical, i.e., wing loading and jaw morphology, as well as echolocation characteristics (Kusch & Schotte 2007, Lacki et al. 2007). Bats with high wing-loading and low-frequency echolocation calls typically select open areas such as fields or large waterways over which to forage (Fenton & Bell 1981, Ford et al. 2005). Bats with low wing-loading and broadband, high-frequency echolocation calls are capable of efficiently foraging in cluttered conditions such as forest canopies and headwater streams (Ratcliffe & Dawson 2003, Lacki et al. 2007). Bats with intermediate wing-loading commonly are more plastic in the types of foraging areas that they use; they are capable of efficiently foraging in relatively cluttered as well as open areas (Lacki et al. 2007). A bat community, therefore, is an assemblage of species able to efficiently forage in various cover types or environments of variable amounts of physical complexity or structure. Some bat species, for example those with high wing-loading, are more likely to occur together when foraging than with species with low wing-loading. Species with intermediate wing-loading may occur in many different cover types, thereby increasing the likelihood that they will occur with species with both high or low wing-loading. Considered together, the various foraging strategies exhibited by different bat species suggest that a diversity of riparian conditions is beneficial for more species (Lloyd et al. 2006, Williams et al. 2006, Ober & Hayes 2008).

Riparian areas serve numerous ecological functions, not only as habitat for bats and other wildlife, but also for erosion and flood control, and biogeochemical cycling (Naiman & Décamps 1997). In the southeastern United States, there is growing concern regarding urbanization and other development near riparian zones and potential impacts on riparian function and wildlife (Zhang et al. 2008). Similar to other southeastern states, Georgia, particularly the northern, mountainous portion, is experiencing increasing development and urbanization (Turner & Ruscher 1988). The Cumberland Plateau and Ridge and Valley physiographic provinces of northwestern Georgia together exist as a region that is high in bat species richness compared to other areas of the state (Menzel et al. 2000), as well as other plant and animal groups (Wharton 1978). Regionally, 12 bat species potentially occur, including Corynorhinus rafinesquii, Eptesicus fuscus, Lasionycteris noctivagans, Lasiurus borealis, L. cinereus, Myotis grisescens, M. leibii, M. lucifugus, M. septentrionalis, M. sodalis, Nycticeius humeralis, and Perimyotis subflavus (Menzel et al. 2000). Myotis grisescens and M. sodalis are federally-listed endangered species, and Corynorhinus rafinesquii and M. leibii are rare species in the southeastern United States. Because the importance of conserving riparian areas in northwestern Georgia cannot be overemphasized from a biodiversity standpoint, it is imperative that riparian areas are identified and characterized, and that their relevance to bat communities is understood in terms of how physical structure of riparian areas and adjacent land uses affect bats. Accordingly, our objectives were to determine the degree to which different bat species occurred together and if their occurrence was related to riparian condition in northwestern Georgia. Also, we determined if bat species richness was influenced by riparian condition. We predicted that within riparian zones, bat species would be partitioned according to the physical structure of the vegetation adjacent to water sources and the size of the water source itself; E. fuscus, N. humeralis, L. cinereus, and M. grisescens would be associated with larger water sources with less canopy cover (LaVal et al. 1977, Ford et al. 2005). Myotis septentrionalis, Corynorhinus rafinesquii, M. sodalis, and M. leibii would occur at small, headwater streams and small pools with dense canopy cover (Ford et al. 2005, Menzel et al. 2005). Lasiurus borealis, P. subflavus, L. noctivagans, and M. lucifugus would occur in a variety of riparian conditions, with no preference for either large or small water sources (Ratcliffe & Dawson 2003, Ford et al. 2005). All species would occur throughout the area, regardless of elevation, with exception of M. grisescens, which commonly is associated with larger waterways in valleys (LaVal et al. 1977). Species such as E. fuscus and M. lucifugus would occur at riparian areas with surrounding development, as these species are known to roost in anthropogenic structures (Barbour & Davis 1969).

Methods

Study area

We conducted our study in the Cumberland Plateau and Ridge and Valley physiographic provinces (34º10′ – 35º00′ N; 84º49′ – 85º37′ W) in northwestern Georgia (Fig. 1). Long, linear ridges characterized the Ridge and Valley province, with sharp escarpments rising 100–300 m above valley floors. River valleys were wide and flat in the Ridge and Valley province, with waterways forming more deeply incised, narrow gorges to the northwest in the Cumberland Plateau province, which was predominated by the Lookout Mountain massif. Our study area encompassed approximately 8820 km of streams and rivers, and approximately 5700 ponds and lakes (Simley & Carswell 2009). Quercus spp. — Pinus spp. forests were predominate in uplands, and riparian areas included forest communities of Liquidambar styraciflua, Platanus occidentalis, Liriodendron tulipifera or shrub wetlands consisting of Cephalanthus occidentalis and Salix nigra. Most alluvial valleys in the Ridge and Valley province have been cleared for pasture and row crops. Limestone geology underlies most of the area, and caves occur throughout (Wharton 1978).

Fig. 1.

Acoustic sampling sites in northwestern Georgia, 2000–2002.

f01_192.jpg

Acoustic monitoring

In summers 2000–2002 (June–August), we conducted active acoustic monitoring at water sources with an Anabat II (Titley Electronics, Ballina, Australia) bat detector linked to a Compaq Presario 1200 laptop computer (Compaq Computer Corporation, Houston, Texas, USA) via an Anabat V Zero-Crossing Analysis Interface Module (ZCAIM) and Anabat 6.2d software (Corben 1999a, Broders et al. 2003). In 2000, we established an 8.3 × 9.4 km sampling grid over a 5100 km2 study area in the northwestern corner of Georgia. We randomly sampled 3–4 water sources within 4 km of each grid intersection (n = 69) (Johnson et al. 2002). We sampled a variety of water sources, from small streams to large rivers, and ponds and lakes of various sizes. Moreover, we sampled at locations with different riparian vegetation, e.g., clearings and forests. In 2001, we focused our sampling along streams and rivers with varying vegetative complexity. In 2002, we sampled at streams in the Chickamauga and Chattanooga National Military Park. We monitored bat activity at each site once for 20 minutes between 21.15 and 02.00 hours (3–10 sites per night) (Schirmacher et al. 2007, Johnson et al. 2008). We used active monitoring, sweeping the detector to sample over the water source until a bat was detected. We attempted to capture as much of the echolocation pass sequence as possible by orienting the detector to follow the detected bat's flight. Sampling was not conducted during periods of heavy rain, high wind (≥ 20 kph), or cold temperatures (< 10°C), as these conditions may negatively affect bat activity (Erickson & West 2002). Bat echolocation passes, or a series of echolocation pulses or calls emitted by bats as they navigate and search for food, were analyzed with Analook 4.7j software (Corben 1999b). The senior author identified all echolocation passes to avoid variation in identification accuracy among personnel (O'Farrell et al. 1999) and only identified echolocation passes containing ≥ 3 pulses. We qualitatively identified echolocation passes of non-Myotis bats by comparing structures and frequencies of echolocation passes recorded at sampling sites to a library of identified echolocation passes of hand-released bats captured throughout the southeastern United States (Fenton & Bell 1981). We identified Myotis species based on frequency and slope characteristics of echolocation passes (O'Farrell 1999, Britzke & Murray 2000, Murray et al. 2001, Britzke et al. 2002).

Riparian conditions

We measured 18 variables at each acoustic monitoring site that we believed influenced presence of bats either directly or indirectly based on previous studies (Table 1). Each waterway was categorized as lentic or lotic (variable = Lotic). The surface of the water was qualitatively classified as pool, run, riffle, or rapid because some bat species have been found to avoid rapids and riffles of lotic waterways (Flow) (Mackey & Barclay 1989). We measured the width of each water source with a laser range finder (Leica Camera Inc., Solms, Germany) or tape measure (Width) (Seidman & Zabel 2001). Water depth was categorized (Depth; 0–15 cm, 16–90 cm, > 90 cm). Substrate was classified according to approximate aggregate size, including clay/mud, sand, gravel, cobble, or bedrock (Substrate). We measured the temperature of the water's surface with a thermometer (Temp). Turbidity was categorized (Turbidity; clear, semi-turbid, turbid). Because some bats avoid structural clutter (e.g., vegetation, tree limbs) when foraging over water (Mackey & Barclay 1989), we estimated the amount of clutter over water sources that could potentially impede bat flight by separating streams into 20 1 m segments; 10 upstream and 10 downstream from the sampling site. Each segment received a score of one if clutter was present < 3 m from the water's surface or zero if otherwise, resulting in a potential score of 0–20 for each sample site (Shrub). We estimated percent (0, 25, 50, 75, or 100) overstory forest canopy cover at 5 locations (5 m spacing): 1 at the sampling site, 2 upstream and 2 downstream from the sampling site, and averaged the values (% Canopy). We tallied the number of dead trees (Snags) and trees with exfoliating bark (Exfobark) within 50 m of each sampling site.

Table 1.

Relation between 2 dimensional nonmetric multidimensional scaling (NMDS) axes and riparian variables measured at acoustic sampling sites (n = 300) in northwestern Georgia, 2000–2002.

t01_192.gif

We determined the elevation (Elevation), and latitude and longitude of each sampling site with a GeoExplorer II global positioning unit (Trimble Navigation, Ltd., Sunnyvale, California) and entered them into a GIS. We used ArcMap 9.2 (ESRI 2006) to acquire and analyze land use/land cover types from the 2001 National Land Cover Database (NLCD; Homer et al. 2007). We determined the percent of 15 possible land use/land cover types in a 1 ha area (56.4 m radius) surrounding each sampling site (USGS 2003). An area of 1 ha was chosen to examine the land use/land cover types immediately surrounding each sampling site while retaining independence among sampling sites. We summed percentages of NLCD land cover classes into 6 new land cover types as follows: (1) developed open space + low intensity developed + medium intensity developed + high intensity developed (Developed); (2) open water (Water); (3) coniferous forest + deciduous forest + mixed forest (Forest); (4) barren land + unconsolidated shore + scrub or shrub + grassland (Open); (5) pasture or hay + crops (Ag); and (6) woody wetland + forested wetland + emergent herbaceous wetland (Wetland).

Statistical analysis

We examined ordination of acoustic sampling sites according to bat species assemblages with nonmetric multidimensional scaling (NMDS; Appendix A; McCune & Grace 2002). To conduct the analysis, we used the metaMDS function in the R vegan package (Oksanen et al. 2008, RDCT 2008). We performed NMDS only on acoustic sampling sites where bats were positively identified (300 of 338 sites). We compiled a data matrix with bat species (presence or absence) and riparian variables as columns and sites as rows. Bray-Curtis dissimilarity distances among sites were determined for the bat species presence/absence portion of the data matrix. Because some acoustic sampling sites had identical bat species assemblages, a small positive value (0.071) was automatically added by R to these sites to allow Bray-Curtis distances to differentiate among them. Nonmetric multidimensional scaling hierarchically orders sites by their Bray-Curtis distances, and seeks to determine and represent the optimum position of n entities in k-dimensional space. Moreover, NMDS optimizes the position of the entities by minimizing the measure of stress, which is the magnitude entities must be moved in k-dimensional space to preserve monotonicity, or the original hierarchical ordering of the sites. The starting location of the entities is constructed at random in k-dimensional space. The entities are then moved to preserve monotonicity, and stress is measured. This process is repeated several times within each k-dimensional space to ensure stress stability. The n entities can be perfectly represented (zero stress) in n-1 dimensions. To represent n entities in > n-1 dimensional space, it is likely that the entities would have to be moved to preserve monotonicity, consequently increasing stress. Although lower stress values equate to a better representation of entities and may require a larger number of dimensions, interpretability is sacrificed if too many dimensions are retained in order to minimize stress values (McCune & Grace 2002). For our study, we iteratively examined stress levels of 1–6 dimensional ordinations of the data. A maximum of 20 runs was conducted for each dimension. Within each run, stress was averaged over 5 iteration-intervals for a maximum of 200 iterations and examined for stability. We chose to proceed with the dimension that provided the best compromise between stress (< 20) and interpretability (Clarke 1993).

We used the envfit function in the R vegan package to determine if riparian conditions were correlated with the NMDS ordination axes based on 1000 permutations of the data (Oksanen et al. 2008, RDCT 2008). Prior to analysis, we used Spearman's rank correlation analysis to examine all riparian condition variables for collinearity, and considered any variable pair with a correlation (ρ) exceeding 0.7 to be significantly correlated. No variable pairs were significantly correlated. We considered riparian condition variables to be significantly correlated with NMDS ordination axes at the P ≤ 0.10 level. We calculated partial correlation coefficients for each variable.

To determine if riparian condition affected bat species richness, we used a regression tree analysis performed in R mvpart package (McCune & Grace 2002, Oksanen et al. 2008, RDCT 2008). We used a regression tree instead of a classification tree approach in order to detect trends in species richness based on riparian variables rather than attempting classification of absolute species richness values. We used all acoustic sampling sites (n = 338) in the analysis, including sites where only unidentifiable bats or unidentifiable Myotis spp. were recorded. We considered bat species richness to be one (1) at sites where only unidentifiable bats, unidentifiable Myotis spp., or undetermined L. borealis/P. subflavus were recorded because we were uncertain of exactly how many bat species were present at these sites. It would be incorrect to consider these sites as having no bat species present, making interpretation of the results problematic. If only unidentifiable Myotis spp. and undetermined L. borealis/P. subflavus were recorded, species richness was two (2) for the site. We performed 50 multiple cross validations on 10 random subsets of the data. The procedure was performed 5 times to ensure stability of the results in terms of tree size and cross-validation error rate. We determined classification tree size, i.e., number of leaves, by selecting the largest tree with a cross-validation error within one standard error of the minimum (Faraway 2006).

Fig. 2.

Association between nonmetric multidimensional scaling (NMDS) ordination axes of bat species and riparian conditions in northwestern Georgia, 2000–2002. Directions of arrows indicate increasing variable value.

EPFU = Eptesicus fuscus; LABO = Lasiurus borealis; LACI = Lasiurus cinereus; MYGR = Myotis grisescens; MYLU = Myotis lucifugus; MYSE = Myotis septentrionalis; MYSO = Myotis sodalis; PESU = Perimyotis subflavus.

f02_192.jpg

Results

Acoustic monitoring

We conducted acoustic monitoring at 338 sampling sites, including 31 sites at ponds and lakes, 170 sites at streams (≤ 4th order) and 137 sites at rivers (≥ 5th order). We recorded 12018 echolocation passes in summers 2000–2002, and identified P. subflavus at 246 sites, M. grisescens at 126 sites, L. borealis at 67 sites, E. fuscus at 45 sites, M. lucifugus at 38 sites, M. septentrionalis at 23 sites, L. cinereus at 6 sites, and M. sodalis at 2 sites. Due to call sequences that were comprised of too few pulses or were of insufficient quality, we were unable to identify to species Myotis at 166 sites, and L. borealis/P.subflavus at 6 sites. At 16 (7.5%) sampling sites, no echolocation passes were recorded.

Statistical analysis

Solutions for NMDS ordinations at 1–6 dimensions were achieved within 18 runs of the data. Final stress values for all dimensions ≥ 2 were < 20. The stress value of the 2-dimensional NMDS ordination stabilized at 17.17 after 3 runs of the data. Among the 8 recorded species, only M. septentrionalis and M. grisescens were positively associated with NMDS axis 1; all other species showed a negative association. Lasiurus borealis, E. fuscus, M. lucifugus, and M. grisescens were negatively associated with NMDS axis 2. Myotis septentrionalis, L. cinereus, M. sodalis, and P. subflavus were positively associated with NMDS axis 2 (Fig. 2).

Of the 18 riparian variables that we measured, 8 were correlated (P ≤ 0.10) with the 2-dimension species space axes (Table 1). However, individually, all variables poorly explained (r2 ≤ 5.0%) variation in species space. All riparian variables except Elevation were positively associated with NMDS axis 1. All riparian variables except % Canopy and Lotic were negatively associated with NMDS axis 2 (Fig. 2).

Fig. 3.

Bat species richness predicted by riparian condition at acoustic monitoring sites (n = 338) in northwestern Georgia, 2000–2002.

f03_192.jpg

Lasiurus borealis, E. fuscus, and M. lucifugus were associated with high-elevation lentic waterways with sparse canopy cover. Myotis septentrionalis were associated with low-elevation lotic waterways with dense canopy cover. Myotis grisescens were associated with low-elevation, turbid, deep lotic waterways with adjacent wetlands. Lasiurus cinereus and M. sodalis showed a negative association with high amounts of water and wetland areas, but sample sizes for these species were 6 and 2 respectively, indicating that these results should not be considered robust. Perimyotis subflavus were the most ubiquitous species recorded, resulting in a nearly central location on the NMDS plot. However, there was some indication of negative association between P. subflavus presence and amount of water and wetland areas (Fig. 2).

Regression tree analysis indicated that up to 6 leaves were within 1 cross-validation error of the minimum and were meaningful in determining species richness according to riparian variables (Fig. 3). The model error was 0.83, indicating that 17% variation in species richness was explained by the regression tree. Relatively high (> 2) species richness was associated with 2 riparian conditions; small, high-elevation water sources with sparse canopy cover, and large streams and rivers with adjacent wetland areas (Fig. 3).

Discussion

The bat community that we observed was somewhat consistent with expectations, but was absent of N. humeralis, C. rafinesquii, L. noctivagans, and M. leibii. Nycticeius humeralis probably are more common on the Coastal Plain to the south, and were rarely documented during mist net surveys at the Chickamauga and Chattanooga National Military Park within our study area (Menzel et al. 2000, Ford et al. 2004). Lasionycteris noctivagans likely occur in northwestern Georgia only as seasonal transients or as winter residents (Menzel et al. 2000). Corynorhinus rafinesquii and M. leibii are considered rare throughout their respective geographic ranges and, in northwestern Georgia, are most likely occur near rock outcrops and talus slopes associated with the Lookout Mountain massif, where our surveys did not intensively focus (Menzel et al. 2000). Moreover, acoustic sampling is not appropriate for detecting C. rafinesquii because their low-intensity echolocation calls are difficult to detect (Menzel 2003). Notwithstanding the few species that we did not record, our surveys documented a species assemblage that was partitioned to some degree according to riparian condition.

Despite the low variation in species-riparian relations explained by the variables we measured, several species met our expectations in the types of riparian areas wherein they were associated. Myotis grisescens were associated with deep, turbid streams and rivers in the valleys. Chickamauga Creek and the Oostanaula River previously have been identified as primary foraging areas for M. grisescens in northwestern Georgia (Johnson 2002). M. grisescens typically forage over larger waterways throughout their range, and may competitively exclude other Myotis species from these areas (LaVal et al. 1977). Indeed, the closest Myotis species to M. grisescens on the NMDS ordination was M. septentrionalis, which was associated more with waterways that had dense canopy cover, likely headwater streams. M. septentrionalis typically forage in intact forests and upland streams (Ford et al. 2005). However, contrary to our expectations, M. septentrionalis in northwestern Georgia were more commonly detected at lower elevations. In the Central Appalachian Mountains region, for example, M. septentrionalis are more prevalent in upland areas (Ford et al. 2005). In northwestern Georgia, the underlying sandstone geology and droughty soils of high elevation areas, i.e., the Lookout Mountain massif, do not provide the necessary conditions for an abundance of forested headwater streams and seeps (Wharton 1978). This is supported to some extent by the results of the regression tree analysis that indicate that higher species richness is associated with narrow, high-elevation waterways that have sparse canopy cover. M. lucifugus were at the opposite side of NMDS axis 1 from M. grisescens and M. septentrionalis, and were associated with high-elevation lentic waterways with sparse canopy cover. Foraging areas of M. lucifugus commonly are riparian areas with less canopy cover than those used by sympatric M. septentrionalis or M. sodalis (Ford et al. 2005, Schirmacher et al. 2007). Similar to M. lucifugus, E. fuscus and L. borealis foraged at high-elevation lentic waterways with sparse canopy cover. Eptesicus fuscus and Lasiurus borealis are commonly associated with open areas, including large pools and streams, when foraging (Ford et al. 2005, Schirmacher et al. 2007, Brooks 2009). However, it is unclear why these species were more commonly detected at high-elevation waterways, as we expected that they would be associated with low-elevation riparian areas (Ford et al. 2002, 2005). Perhaps the association is merely coincidently, considering the relatively weak explanatory power of elevation. Perimyotis subflavus were the most ubiquitous species detected in northwestern Georgia, resulting in a central location on the NMDS ordination. P. subflavus are considered a generalist species capable of efficiently foraging over both open and cluttered waterways of all types (Schirmacher et al. 2007). Low sample sizes of M. sodalis and L. cinereus limit strong inferences that can be drawn regarding the types of riparian areas they use in northwestern Georgia. L. cinereus, a rare summer resident of Georgia, typically forage in open areas, including large ponds and reservoirs (Menzel et al. 2000, Brooks 2009). Myotis sodalis typically forage in forests with intact canopies and near headwater streams (Menzel et al. 2005, Schirmacher et al. 2007). Given the wide range of riparian conditions used by the various bat species in northwestern Georgia, we believe our results not only support the importance of riparian areas as bat habitat in general, but also corroborate recommendations to maintain a variety of riparian areas to foster the full bat community (Lloyd et al. 2006, Williams et al. 2006, Ober & Hayes 2008). Results of the regression tree analysis indicate that two distinct types of riparian areas are associated with high bat species richness. Riparian areas at small, high-elevation waterways had high bat species richness. Also, these areas had sparse canopy cover, allowing use by open-area foragers, including E. fuscus, M. lucifugus, and L. cinereus (Brooks 2009). Water sources at high elevations may be an uncommon resource, increasing the importance of these areas to a wide range of bat species that may be forced to concentrate their activity in a relatively small area. Indeed, only seven of the sampling sites met these criteria, indicating that they are relatively uncommon riparian types in northwestern Georgia. High bat activity and species diversity has been documented at high-elevation wetlands and intermittent streams (Seidman & Zabel 2001, Francl et al. 2004, Lloyd et al. 2006). Also important to high species richness in northwestern Georgia were larger streams and rivers that had wetlands, i.e., inundated floodplains, adjacent to them. There may be an abundance of snags associated with these wetlands where bats can day- and night-roost (Carter 2006). Similar to some high-elevation waterways, many of the larger streams and rivers in the valleys also have sparse canopy cover, providing foraging conditions amenable to a higher overall number of bat species (Lloyd et al. 2006, Williams et al. 2006, Ober & Hayes 2008). These larger waterways, including the Oostanaula River and Chickamauga Creek, are known foraging areas of endangered M. grisescens, adding to their importance to bats (Johnson 2002). Although the regression tree explained 17% of variation in bat species richness, the results provide broader merit from a biological and conservation standpoint. By implementing conservation strategies for the M. grisescens, i.e., protecting larger streams and rivers, many bat species will benefit as a consequence. Therefore, protection of both small, open, high-elevation waterways and large streams and rivers and adjacent wetlands would be the most effective strategy to benefit the most bat species in northwestern Georgia.

A fuller understanding of the association between bats and riparian conditions would be gained by incorporating assessments of water quality and aquatic and aerial macroinvertebrates at sampling sites. Moreover, the NMDS and regression tree analyses that we used undoubtedly could be improved with the incorporation of data from a broader suite of ecological conditions. For example, urbanization may negatively impact water quality and arthropod resources for bats at riparian areas (Kusch et al. 2004, Kalcounis-Rueppell et al. 2007). More research is needed to determine the relation between arthropod resources and anthropogenic influences to form a more complete picture of the ecology of these riparian areas. Indeed, our models could be improved by a more complete analysis of the relation between riparian areas and bats with the addition of water chemistry and arthropod availability data to the existing dataset.

Acknowledgements

Georgia Department of Natural Resources (GA DNR) and National Park Service (NPS) provided funding, housing, and support for this project. J. Ozier, N. Holcomb, A. Padgett, T. Gilbreath, and T. Touchstone of the GA DNR and J. Szyjkowski of the NPS provided invaluable assistance and logistic support. We thank M. Strager of West Virginia University (WVU) and S. Koonce of the United States Department of Agriculture (USDA) Forest Service for assistance with Geographic Information Systems. G. Merovich (WVU) provided much appreciated assistance with statistical coding in R and a review of this manuscript. We thank B. Sparks and S. Sutton for providing field assistance. In-kind support was provided by USDA Forest Service Northern Research Station and WVU Division of Forestry and Natural Resources. We thank 2 anonymous reviewers of this manuscript for their helpful comments.

Literature

1.

Barbour R.W. & Davis W.H. 1969: Bats of America. University Press of Kentucky , Lexington, KentuckyGoogle Scholar

2.

Britzke E.R. & Murray K.L. 2000: A quantitative method for selection of identifiable search-phase calls using the anabat system. Bat Res. News 41: 33–36. Google Scholar

3.

Britzke E.R., Murray K.L., Heywood J.E. & Robbins L.W. 2002: Acoustic identification. In: Kurta A. & Kennedy J. (eds.), The Indiana Bat: Biology and management of an endangered species. Bat Conservation International , Austin, Texas : 221–225. Google Scholar

4.

Broders H.G., Quinn G.M. & Forbes G.J. 2003: Species status, and the spatial and temporal patterns of activity of bats in southwest Nova Scotia, Canada. Northeast. Nat. 10: 383−398. Google Scholar

5.

Brooks R.T. 2009: Habitat-associated and temporal patterns of bat activity in a diverse forest landscape of southern New England, USA. Biodivers. Conserv. 18: 529–545. Google Scholar

6.

Carter T.C. 2006: Indiana bats in the Midwest: Importance of hydric habitats. J. Wildl. Manage. 70: 1185–1190. Google Scholar

7.

Ciechanowski M. 2002: Community structure and activity of bats (Chiroptera) over different water bodies. Mamm. Biol. 67: 276–285. Google Scholar

8.

Clarke K.R. 1993: Non-parametric multivariate analyses of changes in community structure. Aust. J. Ecol. 18: 117–143. Google Scholar

9.

Corben C. 1999a: Anabat, Version 6.2d [computer software]. Ballina , AustraliaGoogle Scholar

10.

Corben C. 1999b: Analook, Version 4.7j [computer software]. Ballina , AustraliaGoogle Scholar

11.

Daniel S., Korine C. & Pinshow B. 2008: Central-place foraging in nursing, arthropod-gleaning bats. Can. J. Zool. 86: 623–626. Google Scholar

12.

Environmental Systems Research Institute (ESRI), Inc. 2006: ArcMap™ 9.2 [computer software]. Redlands , CaliforniaGoogle Scholar

13.

Erickson J.L. & West S.D. 2002: The influence of regional climate and nightly weather conditions on activity patterns of insectivorous bats. Acta Chiropterol. 4: 17–24. Google Scholar

14.

Faraway J.J. 2006: Extending the linear model with R. Chapman & Hall , New York, New YorkGoogle Scholar

15.

Fenton M.B. & Bell G.P. 1981: Recognition of species of insectivorous bats by their echolocation calls. J. Mammal. 62: 233–243. Google Scholar

16.

Ford W.M., Edwards J.W. & Johnson J.B. 2004: Battlefield bat survey. Unpublished report prepared for the National Park Service , Fort Oglethorpe, GeorgiaGoogle Scholar

17.

Ford W.M., Menzel M.A., Menzel J.M. & Welch D.J. 2002: Influence of summer temperature on sex ratios in eastern red bats (Lasiurus borealis). Am. Midl. Nat. 147: 179–184. Google Scholar

18.

Ford W.M., Menzel M.A., Rodrigue J.L., Menzel J.M. & Johnson J.B. 2005: Relating bat species presence to simple habitat measures in a central Appalachian forest. Biol. Conserv. 126: 528–539. Google Scholar

19.

Francl K.E., Ford W.M. & Castleberry S.B. 2004: Bat activity in central Appalachian wetlands. Georgia J. Sci. 62: 87–94. Google Scholar

20.

Grindal S.D., Morissette J.L. & Brigham R.M. 1999: Concentration of bat activity in riparian habitats over an elevational gradient. Can. J. Zool. 77: 972–977. Google Scholar

21.

Homer C., Dewitz J., Fry J., Coan M., Hossain N., Larson C., Herold N., McKerrow A., VanDriel J.N. & Wickham J. 2007: Completion of the 2001 National Land Cover Database for the conterminous United States. Photogrammetric Engineering and Remote Sensing 73: 337–341. Google Scholar

22.

Johnson J.B. 2002: Spatial and predictive foraging models for gray bats in northwest Georgia and a comparison of two acoustical bat survey techniques. M.S. thesis. West Virginia University , Morgantown, West VirginiaGoogle Scholar

23.

Johnson J.B., Gates J.E. & Ford W.M. 2008: Distribution and activity of bats at local and landscape scales within a rural-urban gradient. Urban Ecosyst. 11: 227–242. Google Scholar

24.

Kalcounis-Rueppell M.C., Payne V.H., Huff S.R. & Boyko A.L. 2007: Effects of wastewater treatment plant effluent on bat foraging ecology in an urban stream system. Biol. Conserv. 138: 120–130. Google Scholar

25.

Kunz T.H. 1973: Resource utilization: Temporal and spatial components of bat activity in central Iowa. J. Mammal. 54: 14–32. Google Scholar

26.

Kusch J. & Schotte F. 2007: Effects of fine-scale foraging habitat selection on bat community structure and diversity in a temperate low mountain range forest. Folia Zool. 56: 263–276. Google Scholar

27.

Kusch J., Weber C., Idelberger S. & Koob T. 2004: Foraging habitat preferences of bats in relation to food supply and spatial vegetation structures in a western European low mountain range forest. Folia Zool. 53: 113–128. Google Scholar

28.

Lacki M.J., Amelon S.K. & Baker M.D. 2007: Foraging ecology of bats in forests. In: Lacki M.J., Hayes J.P. & Kurta A. (eds.), Bats in forests. Johns Hopkins , Baltimore, Maryland : 83–127. Google Scholar

29.

LaVal R.K., Clawson R.L., LaVal M.L. & Caire W. 1977: Foraging behavior and nocturnal activity patterns of Missouri bats, with emphasis on the endangered species Myotis grisescens and Myotis sodalis. J. Mammal. 58: 592–599. Google Scholar

30.

Lloyd A., Law B. & Goldingay R. 2006: Bat activity on riparian zones and upper slopes in Australian timber production forests and the effectiveness of riparian buffers. Biol. Conserv. 129: 207–220. Google Scholar

31.

Mackey R.L. & Barclay R.M.R. 1989: The influence of physical clutter and noise on the activity of bats over water. Can. J. Zool. 67: 1167–1170. Google Scholar

32.

McCune B. & Grace J.B. 2002: Analysis of ecological communities. MjM Software Design , Gleneden Beach, OregonGoogle Scholar

33.

Menzel J.M., Ford W.M., Menzel M.A., Carter T.C., Gardner J.E., Garner J.D. & Hoffman J.E. 2005: Summer habitat use and home-range analysis of the endangered Indiana bat. J. Wildl. Manage. 69: 430–436. Google Scholar

34.

Menzel M.A. 2003: An examination of factors influencing the spatial distribution of foraging bats in pine stands in the southeastern United States. Ph.D. Dissertation, West Virginia University , Morgantown, West VirginiaGoogle Scholar

35.

Menzel M.A., Chapman B.R., Ford W.M., Menzel J.M. & Laerm J. 2000: A review of the distribution and roosting ecology of bats in Georgia. Georgia J. Sci. 58: 143–178. Google Scholar

36.

Murray K.L., Britzke E.R. & Robbins L.W. 2001: Variation in search phase calls of bats. J. Mammal. 82: 728–737. Google Scholar

37.

Naiman R.J. & Décamps H. 1997: The ecology of interfaces: Riparian zones. Annu. Rev. Ecol. Syst. 28: 621–658. Google Scholar

38.

Ober H.K. & Hayes J.P. 2008: Influence of vegetation on bat use of riparian areas at multiple spatial scales. J. Wildl. Manage. 72: 396–404. Google Scholar

39.

O'Farrell M.J. 1999: Blind test for ability to discriminate vocal signatures of the little brown bat Myotis lucifugus and the Indiana bat Myotis sodalis. Bat Res. News 40: 44–48. Google Scholar

40.

O'Farrell M.J., Miller B.W. & Gannon W.L. 1999: Qualitative identification of free-flying bats using the Anabat detector. J. Mammal. 80: 11–23. Google Scholar

41.

Oksanen J., Kindt R., Legendre P., O'Hara B., Simpson G.L., Solymos P., Stevens M.H.H. & Wagner H. 2008: Vegan: Community ecology package. R package version 1.15-1. Available at:  http://cran.r-project.org/,  http://vegan.r-forge.r-project.org/  Google Scholar

42.

R Development Core Team (RDCT) 2008: R: A language and environment for statistical computing. R Foundation for Statistical Computing. Vienna, AustriaGoogle Scholar

43.

Racey P.A. 1998: The importance of the riparian environment as a habitat for European bats. Symp. Zool. Soc. London 71: 69–91. Google Scholar

44.

Racey P.A., Swift S.M., Rydell J. & Brodie L. 1998: Bats and insects over two Scottish rivers with contrasting nitrate status. Anim. Conserv. 1: 195–202. Google Scholar

45.

Ratcliffe J.M. & Dawson J.W. 2003: Behavioral flexibility: The little brown bat, Myotis lucifugus, and the northern long-eared bat, M. septentrionalis, both glean and hawk prey. Anim. Behav. 66: 847–856. Google Scholar

46.

Schirmacher M.R., Castleberry S.B., Ford W.M. & Miller K.V. 2007: Habitat associations of bats in south-central West Virginia. Proc. Southeast. Assoc. Fisheries Wildl. Agencies 61: 46–52. Google Scholar

47.

Seidman V.M. & Zabel C.J. 2001: Bat activity along intermittent streams in northwestern California. J. Mammal. 82: 738–747. Google Scholar

48.

Simley J.D. & Carswell W.J., Jr. 2009: The national map — hydrography. U.S. Geological Survey Fact Sheet : 2009–3054. Google Scholar

49.

Turner M.G. & Ruscher C.L. 1988: Changes in landscape patterns in Georgia, USA. Landscape Ecol. 1: 241–251. Google Scholar

50.

United States Geological Survey (USGS) 2003: National Land Cover Database Zone 60 Land Cover Layer. Available at:  http://www.mrlc.gov  Google Scholar

51.

Wharton C.H. 1978: The natural environments of Georgia. Bulletin 114, 174. Geologic and Water Resources Division and Georgia Department of Natural Resources , Atlanta, Georgia, USAGoogle Scholar

52.

Williams J.A., O'Farrell M.J. & Riddle B.R. 2006: Habitat use by bats in a riparian corridor of the Mojave Desert in southern Nevada. J. Mammal. 87: 1145–1153. Google Scholar

53.

Zhang Y., He H.S. & Yang J. 2008: The wildland-urban interface dynamics in the southeastern U.S. from 1990 to 2000. Landscape Urban Plann. 85: 155–162. Google Scholar
Joshua B. Johnson, W. Mark Ford, John W. Edwards, and Michael A. Menzel "Bat community structure within riparian areas of northwestern Georgia, USA," Folia Zoologica 59(3), 192-202, (1 September 2010). https://doi.org/10.25225/fozo.v59.i3.a4.2010
Received: 2 November 2009; Accepted: 1 January 2010; Published: 1 September 2010
KEYWORDS
Anabat
Eptesicus
Lasiurus
Myotis
nonmetric multidimensional scaling
Perimyotis
regression tree
Back to Top