Northeastern Naturalist Vol. 25, No. 2 K.G. Slankard, D.L. Baxley, and G.L. Sprandel 2018 277 2018 NORTHEASTERN NATURALIST 25(2):277–290 The Impacts of Native-Grassland Restoration on Raptors and their Prey on a Reclaimed Surface Mine in Kentucky Kate G. Slankard1,*, Danna L. Baxley2, and Gary L. Sprandel1 Abstract - We evaluated the effects of native-grass restoration on the spatial distribution and density of raptors, vegetative characteristics, and small-mammal communities at Peabody Wildlife Management Area, a large reclaimed surface-coal mine in western Kentucky. We surveyed raptors from 2008 to 2012 via distance sampling at roadside points, and conducted vegetation and small-mammal surveys. We found no associations between total small-mammal relative abundance and native-grass restoration or vegetative characteristics. However, management for native grass positively affected the density of Circus cyaneus (Northern Harrier) and influenced the local distribution of Northern Harriers and Buteo jamaicensis (Red-Tailed Hawk). These results suggest that restoration and management of native grass on reclaimed mine lands can enhance habitat for grassland raptors, including the Northern Harrier, a species of conservation concern throughout its range. Introduction Grasslands are declining at a rapid pace in North America (Jones and Bock 2002), resulting in a negative trend in grassland bird populations (Brennan and Kuvlesky 2005). Many grasslands contain exotic vegetation, and the restoration and management of native grasslands has been cited as a top objective for grassland bird recovery (Dechant et al. 2003). The Northern Bobwhite Conservation Initiative, Conservation Reserve Program (CRP) and other multi-state programs have committed much effort and expense to restore native grasslands in the eastern US in recent years (Morgan and Robinson 2008, NABCI 2017). Reclaimed surfacecoal mines can provide expansive areas of grassland in Kentucky and elsewhere in the eastern US (Brennan and Kuvlesky 2005); however, they often contain an abundance of exotic vegetation. Due to the aforementioned initiatives, restoring native grasses on reclaimed mine sites to benefit wildlife has become more popular in recent years (Yeiser et al. 2015). Circus cyaneus (L.) (Northern Harrier) is a species of conservation concern throughout much of its range in North America (Hamerstrom 1979, NatureServe 2015), including Kentucky (KDFWR 2013). Northern Harriers are known to use reclaimed mine lands during both the nonbreeding and breeding seasons in Kentucky (Palmer-Ball 1996, Vukovich 2004), along with more common species such as Buteo jamaicensis (Gmelin) (Red-tailed Hawk) and Falco sparverius L. (American Kestrel). In fact, the consistent nesting range of Northern Harriers in Kentucky is nearly limited to reclaimed mine areas in roughly 4 counties in the west-central part 1Kentucky Department of Fish and Wildlife Resources, 1 Sportsman’s Lane, Frankfort, KY 40601. 2The Nature Conservancy of Kentucky, 114 Woodland Avenue, Lexington, KY 40502. *Corresponding author - kate.slankard@ky.gov. Manuscript Editor: Noah Perlut Northeastern Naturalist 278 K.G. Slankard, D.L. Baxley, and G.L. Sprandel 2018 Vol. 25, No. 2 of the state, including Peabody Wildlife Management Area (PWMA) (Palmer-Ball 2003). Much of PWMA has been intensively managed for native grass, but prior to our study little was known about how these habitat-management actions affect Northern Harriers and other grassland raptors. Raptor surveys at PWMA were first conducted during 2002–2003 by Eastern Kentucky University (Vukovich and Monroe 2005) and were continued year-round through 2007 by Kentucky Department of Fish and Wildlife Resources (KDFWR) personnel and Kentucky State Nature Preserves Commission staff. The data collected from 2002 to 2007 suggested that raptor densities might differ between areas that have undergone management for native warm-season grasses (NWSG) and unmanaged areas (KDFWR 2007), but more study was needed to draw confident conclusions. The diet of most grassland raptor species is typically high in small mammals (Baker and Brooks 1981) and reclaimed mine areas can support small-mammal densities similar to those on unmined grassland (Hingtgen and Clark 1984). Northern Harriers, in particular, are known to eat mostly Microtus spp. (voles) and other small mammals throughout their range (MacWhirter and Bildstein 2000) and on PWMA (Stewart 2004, Vukovich and Ritchison 2006). Several researchers have investigated the effects of native-grass restoration on small-mammal populations (Mengak 2004, Stone 2007). There have also been studies investigating the relationships between prey abundance and the winter abundance (Baker and Brooks 1981) and nest productivity of raptors, including Northern Harriers (Hamerstrom 1979). Variables influencing nest-site selection for Northern Harriers have also been well documented, especially the preference to nest in idle grasslands (not recently managed; Herkert et al. 1999, Toland 1986). However, there has been little research on the effects of native-grassland restoration at reclaimed surface mines on the ecology of wintering raptors in North America. Furthermore, none of the aforementioned studies attempted to simultaneously examine the relationships between native-grass restoration, wintering and nesting raptors, their prey base, and vegetative structure. Recent research has shown that increasing the amount of grassland available results in rapid increases of wintering grassland raptors. For instance, Wilson et al. (2010) found that the installation of Conservation Reserve Enhancement Program (CREP) grasslands modestly increased the numbers of wintering Red-tailed Hawks and American Kestrels and considerably increased the numbers of Northern Harriers in the CREP region of Pennsylvania. Another study in Texas found that migrating and wintering Northern Harriers used CREP grasslands more than other types of grassland (Littlefield and Johnson 2005). These success stories led us to pose the question of where we should focus monetary resources for grassland raptors: the conversion of land under other uses to grassland or the management of existing grasslands, many of which contain non-native vegetation. The goals of this study were to determine if the restoration of NWSG affects the density and spatial distribution of raptors and to identify the mechanisms that drive patterns of raptor-habitat use. We built on past experiences with raptor surveys at PWMA and expanded the study design in 2008 to assess the relationships among Northeastern Naturalist Vol. 25, No. 2 K.G. Slankard, D.L. Baxley, and G.L. Sprandel 2018 279 raptor habitat use, grassland management, small-scale vegetation structure, and small-mammal abundance. We were particularly interested in the Northern Harrier because it is a species of greatest conservation need listed in Kentucky’s State Wildlife Action Plan (KDFWR 2013) and the Red-tailed Hawk because it is the most common raptor species at our field-site. Field-site Description Our study site included the KDFWR-owned portions of PWMA in Muhlenberg and Ohio counties in western Kentucky. PWMA comprised 15,891 ha—mostly reclaimed mine land—22% of which was open vegetation (grassland, agricultural land, etc.). KDFWR began restoring NWSG at PWMA in 1997, and a total of 1451 ha (42% of the open area) had been planted in NWSG at the time of this study. PWMA is made up of several large land tracts interspersed among privately owned agricultural land and active surface mines. We conducted our study on 2 separate regions of PWMA. The eastern units (9476 ha) contained much area that was undergoing intensive management for NWSG before and during the study. The western units (6415 ha) served as the control because they contained mostly unmanaged, reclaimed mine land, planted in non-native vegetation (Fig.1). Areas managed for NWSG were dominated by planted native grasses including Sorghastrum nutans (L.) Nash (Indiangrass), Andropogon gerardii Vitman (Big Bluestem), and Schizachyrium scoparium (Michx.) Nash (Little Bluestem), as well as a few encroaching non-native species, especially Lespedeza cuneata (Dum. Cours.) G. Don (Sericea Lespedeza). Open, unmanaged areas contained several exotic species including Sericea Lespedeza and Festuca spp. (fescue grasses), as well as some native forbs and grasses such as Solidago spp. (goldenrod), Ambrosia artemisiifolia L. (Common Ragweed), and Andropogon virginicus L. (Broomsedge). Figure 1. Raptor-survey points and areas managed for native warm-season grasses on Peabody Wildife Managment Area. Inset map shows the location of the study area in Kentucky referenced in the text. Northeastern Naturalist 280 K.G. Slankard, D.L. Baxley, and G.L. Sprandel 2018 Vol. 25, No. 2 KDFWR identified PWMA as a Kentucky Quail Focus Area in 2008 (Morgan and Robinson 2008), and a grant to restore Colinus virginianus (L.) (Northern Bobwhite) habitat was the driving force behind most of the management activities on the eastern units. KDFWR intensively managed the open areas on the eastern units using NWSG planting, herbicide, disking, and prescribed fire. Methods Raptor surveys We conducted raptor surveys once each month, during the 3rd week of the month, from December 2008 to July 2012. We did not conduct surveys on days with precipitation or fog, when wind speeds were greater than 24 kph, or when snow depth was greater than 3 cm. We employed a distance-sampling, roadside, point-count survey methodology for replicability and to calculate density estimates (Bibby et al. 2000, Buckland et al. 2001). A single observer surveyed 29 points for raptors in a single day beginning at least ½ h after sunrise and ending at least ½ h before sunset. We did not stratify points, but instead assumed that a constrained random-sampling design would survey areas with varying amounts of managed and unmanaged vegetation. We identified suitable roadside sampling points by intersecting all roads with the open-land cells from the 2005 National Landcover Dataset (KDGI 2007a) in ArcGIS 9.3 software (ESRI, Redlands, CA). From the resulting 2644 suitable points, we randomly selected points that were at least 1000 m apart, leading to 73 potential points. We then manually examined each of the 73 potential points via 0.6-m resolution aerial imagery from 2006 (KDGI 2007b) and on the ground. We eliminated any point did not contain at least 75% open grassland vegetation within the survey area (500-m radius of the point) or was not reliably accessible. We randomly selected 29 of the remaining 42 points for the final surve y route. During the first survey, we used a GPS unit to locate each point and then marked points permanently with steel posts. We visited the same points on each survey date and reversed the order in which we visited the points on alternate surveys. To minimize observer bias, the same observer collected all of the raptor data. At each point, the observer got out of the vehicle and watched for raptors for 3 min. For each individual detected, the observer recorded the species, age, sex, time, and location. The observer based all information recorded on the first detection of the individual and marked the approximate location of each raptor on a 1:7000 scale map with 2006 Farm Service Agency 0.6-m resolution aerial imagery (KDGI 2007b). Later, we digitized the detection locations for analysis using the same imagery in ArcGIS. Small-mammal trapping To obtain estimates of small-mammal relative abundance, we conducted smallmammal trapping, using snap-traps. We conducted trapping at 6 grid locations for 2 consecutive trap-nights in both March and July each year from 2009 to 2012. We sampled new grids each year, for a total of 24 grids sampled; 50% of grids sampled each season were within areas managed for NWSG and 50% were in unmanaged areas. Grids located in managed areas had not been managed more recently than Northeastern Naturalist Vol. 25, No. 2 K.G. Slankard, D.L. Baxley, and G.L. Sprandel 2018 281 6 months prior to the trap session. To minimize the effects of weather, we did not conduct small-mammal trapping when there was more than 3 cm of snow on the ground or on days/nights where there was substantial rainfall or temperatures below -12 °C. We generated random grid-locations for each vegetation type (Beyer 2004). We employed aerial imagery and ground-level observation to verify the vegetation composition of each grid. If any grid did not contain at least 75% open (not forest or water) vegetation, we eliminated it and used the next random point. We focused trapping efforts on field interiors to avoid edge effects by buffering each grid border by at least 100 m from the field edge and any roads. The border of each grid was also at least 100 m from any other grid to ensure they were spatially independent. We set small-mammal traps (laid within 1 m of each other in sets of 2) in 7 x 7 grids, that consisted of 49 double traps placed 10 m apart. We baited traps with peanut butter and oatmeal and set each trap on the mineral soil by removing the litter layer if necessary. We set 20 museum-special traps and 78 mouse traps (Victor M-040; Woodstream Corp., Lititz, PA) at each grid location. We collected captured mammals and placed them in a bag labeled with the grid number, trap type (mouse trap or museum special) and date. We assigned a unique ID number to, identified (to species), measured, aged, and sexed all specimens. Quantitative soil and vegetation measurements We collected vegetation measurements at each small-mammal trapping grid during mammal sampling periods in both winter and summer. We sampled vegetative variables at 5 designated trap locations within each trap grid (the 4 corners and the middle). To avoid sampling trampled vegetation, we offset the vegetation plot 2 m northeast from the trap location. We measured soil compaction at each vegetation plot center, using a pocket-sized penetrometer (E-280; Geotest Instrument Corp., Burr Ridge, IL). We also measured 4 vegetation variables at each plot: vegetation height, vegetation heterogeneity, vegetation density, and percent cover of vegetation type. We measured vegetation height and heterogeneity at the same time using 2 meter-sticks tied together in a cross. We dropped the cross at the plot center; at each tip of the cross, we measured vegetation height (4 measurements) using a separate meter stick. We used an average of these 4 measurements as the vegetation-height value in analyses and calculated vegetation heterogeneity as: (max height - min height) / (mean height) (Bibby et al. 2000). We estimated vegetation density at the plot center by using a portable light meter (photometer) (Extech 401025; FLIR Commercial Systems, Nashua, NH) to measure light penetration at the heights of 1 cm, 5 cm, 20 cm, 50 cm, and at an open location. We recorded light-meter measurements in foot candles (Fc x 100) and calculated indices of vegetation density for each height (Fc of light penetration/Fc at an open location; Bibby et al. 2000). To obtain precise vegetation-density measurements, we did not measure this variable in rain, fog, extreme cloudiness or after sunset. We estimated vegetation composition to the nearest 5%, within a 10-m radius of the plot center, from eye level looking down. The observer estimated the percentage Northeastern Naturalist 282 K.G. Slankard, D.L. Baxley, and G.L. Sprandel 2018 Vol. 25, No. 2 of cover in each of the following categories: trees, shrubs, bare ground, forbs, NWSG, cool-season grass, and Sericea Lespedeza. This measurement was subjective; thus, the same observer conducted this estimate for all the plots. Statistical analysis Although there were several observations for Buteo lineatus (Gmelin) (Red-shouldered Hawk), Haliaeetus leucocephalus (L.) (Bald Eagle), Pandion haliaetus (L.) (Osprey), Ictinia mississippiensis (Wilson) (Mississippi Kite), Falco columbarius L. (Merlin), Accipiter cooperii (Bonaparte) (Cooper’s Hawk), and Buteo lagopus (Pontoppidan) (Rough-legged Hawk), we did not include these species in the analysis due to small sample size or their small likelihood to respond to grassland-habitat management. We opted to exclude American Kestrel (n = 61) from the analysis due to a borderline sample size and a potential bias resulting from an uneven distribution of American Kestrel nest boxes on PWMA. We also excluded from the analysis individuals we suspected to have been double-counted. We calculated Red-tailed Hawk and Northern Harrier densities in Program Distance (Buckland et al. 2001, Laake et al. 1993), and truncated observation data at 500 m from each survey point for this analysis. Within Program Distance, we modeled potential detection functions (half normal cosine, half normal simple polynomial, half normal hermite polynomial, hazard rate simple polynomial, and uniform polynomial) to estimate densities. We used Akaike’s information criterion (AIC) model selection to evaluate and select the best density and detection-function fit for each season and species. We assumed factors impacting detection would be consistent within species and season; consequently, detection functions were constrained to be of the same type for the same species for seasonal analyses. However, density comparisons across NWSG categories were not constrained within a species group because results indicated that landscape factors may have impacted detectability. For example, Red-tailed Hawks in areas with a high prevalence of NWSG had a different detection function than those in areas with a low prevalence of NWSG. For seasonal analysis, we considered December–February the wintering season and March–July the nesting season. Raptors utilize a landscape at a scale much larger than most birds, and presumably, foraging areas are quite large (Slater and Rock 2005). A small field of NWSG is likely not large enough to contain a home range or viable population. For this reason, during the following analyses, we attempted to take into account the variation in size and layout of available NWSG fields at the study site, and minimized arbitrary categorizing of raptor observations when 2 vegetative-cover types occurred in close proximity. To evaluate the effects of NWSG management on raptor density and detection, we used GIS to calculate the percentage of each survey area (within 500 m of the survey point) that had been planted in NWSG. We then used Jenks natural breaks (de Smith et al. 2009) within ArcGIS to divide raptor-survey areas into 3 categories: high prevalence of NWSG (39–96%), medium prevalence of NWSG (14.5–39.1%), and low prevalence of NWSG (0–14%). We compared the density Northeastern Naturalist Vol. 25, No. 2 K.G. Slankard, D.L. Baxley, and G.L. Sprandel 2018 283 and detectability of Red-tailed Hawks and Northern Harriers within these 3 categories in Program Distance. We also conducted analyses using distance-based metrics to assess the relationship between spatial raptor-distribution and management for NWSG. To conduct this analysis, we calculated the mean distance to the closest NWSG field for raptor observations compared to random points in open habitat. To create the random points, we identified open landscape using the National Landcover Data Set (Homer et al. 2015) and the Create Random Points tool in ArcGIS to generate 1600 random points within open, 30 m x 30 m cells. We selected 50 random points for each of the 32 survey days of the study to pair with real raptor detections. We then used the Near tool in ArcGIS to calculate the distance of both the actual raptor observations and the paired random points to the closest NWSG field. We were careful to calculate these distances using only the NWSG fields that were present on the survey date (some fields were planted during the study); seasonal analysis was completed for Northern Harrier only. We did not truncate the distance between raptor observations and NWSG, assuming that the resulting means would represent true patterns of habitat use. We further assumed that large distances were still relevant to the local distribution of raptors based on past research on the large nesting home range and daily movements of Northern Harriers (Slater and Rock 2005). We conducted non-parametric Wilcoxon sign-rank tests in JMP statistical software version 12 (SAS Institute, Cary, NC) to determine differences in mean distance to NWSG fields between raptor locations and random locations. To reduce the probability of Type II error, we employed sequential Bonferroni correction (initial alpha = 0.05; Rice [1989]). To understand parameters that might be driving patterns of raptor density and distribution, we conducted separate, stand-alone analyses to evaluate differences in small-mammal relative abundance between managed and unmanaged grids, and differences in vegetation variables between managed and unmanaged grids. We conducted these analyses in JMP Statistical Software. We tested assumptions of normality and equal variances and conducted non-parametric Wilcoxon rank-sum tests (if assumptions were not met) or parametric t-tests (if assumptions were met) to evaluate differences in datasets. We pooled capture data for the summer and winter sampling sessions for each small-mammal trapping grid and compared species richness and capture rate between managed and unmanaged plots. We defined capture rate as the number of mammals per unique trap-grid sampled and present this as an index of abundance. We analyzed species-specific seasonal and management associations for Peromyscus maniculatus bairdii (Prairie Deer Mouse), Mus musculus (House Mouse), and Microtus ochrogaster (Prairie Vole). Sample sizes were not large enough to conduct species-specific analyses for other small-mammal species. Results Raptor surveys We calculated and compared seasonal densities for Red-tailed Hawk (n = 209) and Northern Harrier (n = 131). For these analyses, the detection functions with Northeastern Naturalist 284 K.G. Slankard, D.L. Baxley, and G.L. Sprandel 2018 Vol. 25, No. 2 the best fit were limited to HN (half normal cosine) or HC (hazard cosine). Northern Harrier (winter: 0.28–0.58/km2; summer: 0.11–0.25/km2) and Red-tailed Hawk (winter: 0.28–0.58/km2; summer: 0.20–0.38/km2) density ranges across years were higher in winter than the nesting season. Northern Harriers and Red-Tailed Hawks had highest densities within survey areas categorized as having a high prevalence of NWSG (Table 1). Northern Harrier and Red-tailed Hawk observations were also significantly closer to areas that had been planted in NWSG when compared to observations at random points. Northern Harriers were significantly closer to areas planted in NWSG than random points during both the winter and nesting seasons. This association was particularly strong during the nesting season (Table 2). Small-mammal trapping During 2352 trap-nights, we captured 745 individuals representing 10 species. We were unable to identify some individuals (n = 12) due to partial predation or decay, and we only identified some Peromyscus spp. to genus because of the difficulty of differentiating Peromyscus leucopus (Rafinesque) (White-footed Mouse) and Prairie Deer Mouse. Prairie Vole, House Mouse, and Prairie Deer Mouse were the most frequently captured species, respectively, representing 68% of total captures (Table 3). Data suggested that mammal-capture rates (captures per individual Table 1. Density estimations, standard errors, sample sizes, and detectability of Northern Harriers and Red-Tailed Hawks within areas managed for high (39–96%), medium (14.5–39.1%), and low (0–14%) prevalence of native warm-season grasses (NWSG) at Peabody Wildlife Management Area 2008– 2012. Detection functions (det. func.) are reported as HN (half normal cosine) or HC (hazard cosine). Species and management type n/km2 SE n Detectability Det. func. Northern Harrier High-prevalence NWSG 0.66 0.12 40 0.3–0.6 HN Medium-prevalence NWSG 0.18 0.05 34 0.4–1.0 HN Low-prevalence NWSG 0.16 0.04 47 0.4–1.0 HN Red-Tailed Hawk High-prevalence NWSG 0.38 0.05 47 0.4–1.0 HC Medium-prevalence NWSG 0.16 0.05 31 0.4–1.0 HN Low-prevalence NWSG 0.32 0.10 102 0.6–1.0 HN Table 2. Mean distance to the closest area managed for native warm-season grass for raptor locations and random locations on Peabody Wildlife Management Area, Kentucky, 2008–2012 and Wilcoxon rank-sum test results (P-values). Seasonal analysis was completed for Northern Harrier only. Species Season Random (m) Real (m) P-value Northern Harrier All 1140 520 0.0001 Nesting 1130 303 0.0001 Winter 1156 682 0.0001 Red-tailed Hawk All 1140 903 0.0002 Northeastern Naturalist Vol. 25, No. 2 K.G. Slankard, D.L. Baxley, and G.L. Sprandel 2018 285 sampling grid over a 2-d period (n = 24) were higher in summer (summer: mean = 20.75, SD = 19.42; winter: mean = 10.3, SD = 8.2; χ2 = 3.38, P = 0.07). Capture rate of mammals was consistent in 2009–2010 and 2012; however, mammal abundance was lower in 2011 than in other years of the project (2009: mean = 40.0, SD = 27.4; 2010: mean = 41.7 41.7, SD = 13.3; 2011: mean = 3.5, SD = 3.3; and 2012: mean = 39, SD = 24.2; χ2 = 13.1, P = 0.004). Capture rates for all mammals did not differ between grids managed for NWSG and unmanaged grids (managed: mean = 28.8, SD = 18.7; unmanaged: mean = 33.3, SD = 29.8; t (11) = 4.4, P = 0.44). Small-mammal species richness in managed plots also did not differ from unmanaged plots (managed: mean = 5.0, SD = 2.0; unmanaged: mean = 4.6, SD = 2.2, t (11) = 0.09, P = 0.77). A couple of relationships were statistically significant for 3 species-specific (House Mouse, Prairie Deer Mouse, and Prairie Vole) capture-rate analyses assessing season and vegetation type (Table 4). Prairie Vole exhibited significantly higher capture rates in summer compared to winter (summer: mean = 10.95, SD = 14.2; winter: mean = 2.29, SD = 3.4; t (11) = 2.34, P = 0.02). None of the 3 small-mammal species were captured more in managed stands; we captured Prairie Deer Mouse more often in unmanaged stands compared to managed stands (managed: mean = 0.58, SD = 1.38; unmanaged: mean = 4.17, SD = 4.47; t (11) = 2.87, P = 0.004). Table 4. Mean, standard deviation, and P-values for small mammals captured during the summer and winter and in managed and unmanaged stands on Peabody Wildlife Management Area 2008–2012. Means are reported per survey grid to compare summer (n = 12) to winter (n = 12) capture rates and capture rates in managed (n = 12) and unmanaged (n = 12) grids. * = statistically significant. Summer Winter Test Managed Unmanaged Test Species Mean SD Mean SD statistic P Mean SD Mean SD Statistic P Mus musculus 2.8 5.4 1.1 2.6 1.32 0.19 3.4 6.2 2.2 4.6 -0.47 0.06 Peromyscus maniculatus 2.8 3.7 1.1 2.4 -1.81 0.07 0.6 1.4 4.2 4.5 2.87 0.004* Microtus ochrogaster 11.0 14.2 2.3 3.4 2.34 0.02* 7.5 8.3 14.4 18.1 0.58 0.56 Table 3. Mean annual raw detections, standard deviations, and sample sizes, across all years and seasons for all small-mammal species recorded on Peabody Wildlife Management Area, Kentucky, 2008–2012. Species Mean SD n Reithrodontomys humilis Audubon & Bachman (Eastern Harvest Mouse) 12.3 10.9 49 Mus musculus L. (House Mouse) 23.5 26.0 94 Cryptotis parva (Say) (North American Least Shrew) 13.3 15.3 55 Microtus pennsylvanicus (Ord) (Meadow Vole) 0.3 0.5 1 Blarina brevicauda (Say) (Northern Short-tailed Shrew) 2.0 3.4 8 Microtus pinetorum (Le Conte) (Pine Vole) 9.5 9.7 38 Peromyscus maniculatus bairdii Hoy and Kennicott (Prairie Deer Mouse) 23.0 13.0 92 Michrotus ochrogaster (Wagner) (Prairie Vole) 79.5 68.6 318 Synaptomys cooperi (Baird) (Southern Bog Lemming) 8.0 7.5 32 Peromyscus leucopus (White-footed Mouse) 10.5 6.8 42 Northeastern Naturalist 286 K.G. Slankard, D.L. Baxley, and G.L. Sprandel 2018 Vol. 25, No. 2 Vegetation characteristics We found significant differences between managed and unmanaged grids in percent cover of NWSG (managed: mean = 54.41, SD = 26.17; unmanaged: mean = 4.18, SD = 6.51; t (11) = 16.51, P = 0.0001) and percent cover of cool-season grass (managed: mean = 8.49, SD = 10.06; unmanaged: mean = 36.73, SD = 23.77; t (11) = 10.86, P = 0.001). No statistical differences existed for other fine-scale vegetation variables including percent forbs, percent shrubs, percent trees, percent bare ground, soil compaction, vegetation density, vegetation heterogeneity, or vegetation height. Percent bare ground was notably low for both managed and unmanaged grids (managed: mean = 2.46, SD = 1.97; unmanaged: mean = 4.03, SD = 4.54) and ground-level vegetation-density indices (Fc of light penetration at 5 cm/Fc at an open location) indicated dense vegetation for both managed and unmanaged grids (managed: mean = 0.20, SD = 0.13; unmanaged: mean = 0.42, SD = 0.30). Discussion We found no associations between small-mammal relative abundance and native-grass restoration or vegetative characteristics. However, management for native grass positively affected the density of Northern Harriers and influenced the local distribution of Northern Harriers and Red-Tailed Hawks. Small-mammal abundance was higher in summer than winter. Nonetheless, counts and densities for Red-tailed Hawk and Northern Harrier were highest in the winter. This finding is not surprising because Kentucky’s resident birds are likely joined by migratory northern birds during this time (Palmer-Ball 2003). Although we focused on smallmammal abundance as a measure of prey availability, the raptor species we focused on for this study might also be responding to differences in availability of other prey types. More study is needed to understand how NWSG restoration affects songbird abundance. The temporal variation we measured in small-mammal abundance (i.e., low abundance in 2011) indicates a likely relationship between raptors and small mammals at PWMA that might be influenced by landscape-level factors like weather and cyclical mammal-population variation, even more so than small-scale vegetative or management factors. Red-tailed Hawk densities were highest in survey areas with a high-prevalence of NWSG. However, it is puzzling that mid-prevalence NWSG areas exhibited lower densities for Red-tailed Hawk than low-prevalence NWSG areas. Analysis for this species may be confounded by their dependency on perch sites (Preston 1990), a variable for which we did not account. Perch availability might also have caused the difference in the detection function we found for this species in areas with high-prevalence NWSG versus those with low-prevalence NWSG. Nevertheless, we also found that Red-tailed Hawks occurred significantly closer to NWSG than random points, so we do think they are exhibiting some level of habitat preference for NWSG areas. Despite the fact that we documented increased raptor densities at survey points with a high prevalence of NWSG, we found no significant differences between the small-scale vegetation structure or prey abundance in managed vs. unmanaged Northeastern Naturalist Vol. 25, No. 2 K.G. Slankard, D.L. Baxley, and G.L. Sprandel 2018 287 vegetation plots that might explain raptor preference for these areas. Preston (1990) had similar results in Arkansas, noting no direct relationship between raptor foraging distribution and prey biomass. Still, our spatial-distribution analysis indicated that raptors prefer areas managed for NWSG; we assume that there is a reason for this trend. We surmise that larger-scale vegetation structure, associated with the management for NWSG, may be driving the spatial distribution of raptors. In a similar study, Toland (1987) found that American Kestrels used disturbed grassland more often than expected, hunted more successfully in disturbed areas, and cued in on human-related disturbances in managed-grassland habitat in Missouri. We documented high vegetation-density and low coverage of bare ground at both the managed and unmanaged field interiors in our study area. Dense vegetation makes it difficult for raptors to see prey beneath the top vegetative layer. Prescribed fire (early spring), disking, and herbicide applications, which occurred in PWMA NWSG fields, probably provided small openings more often than unmanaged habitat (e.g., firebreaks). These openings likely provide areas of superior prey visibility for raptors but can be difficult to account for with GIS analysis or vegetation-sampling protocols. We focused our vegetative sampling on field interiors and habitat that had not been managed within 6 months, which limited our ability to account for the vegetative openings and field edges created by management. Nevertheless, we now theorize that these disturbance-dependent field edges and small openings may be quite important and likely increase the availability of prey to raptors. Conservation strategies for birds of prey often focus on increasing prey abundance. Except for documenting the Prairie Deer Mouse’s preference for unmanaged areas, we were not able to determine any effects on small mammal communities by the restoration of NWSG. It is possible that the dense nature of NWSG plantings provided similar vegetative structure to non-managed vegetation; consequently, there were no real differences in relative abundance between vegetation types. Our results are contrary to one study that found native-grass restoration increased smallmammal abundance in Virginia hayfields (Mengak 2004). Stone (2007) reported a decline in small mammals immediately after native-grass restoration efforts in Colorado, with a rebound ~2–5 years later. The same study suggested that weather may have as great or a greater influence on small-mammal abundance than habitat restoration efforts (Stone 2007). Bird abundance and distribution is often used to assess habitat quality (Johnson 2007). Our results indicated that habitat management for NWSG can improve densities of Northern Harriers and influence their local distribution. We thus infer that management for NWSG provides more-valuable foraging habitat than unmanaged, reclaimed mine land. At first glance, our results may seem to conflict with past studies on the habitat needs of Northern Harrier, which tend to emphasize the need to leave some idle grassland for nesting (Hamerstrom and Kopney 1981, Herkert et al. 1999) or place little value in NWSG restoration for Northern Harrier habitat. For example, in Illinois and Kentucky, nest placement was not dependent on whether fields were comprised of native or non-native grass (Herkert et al. 1999, Vukovich and Ritchison 2006). Meanwhile, Toland (1986) found that management Northeastern Naturalist 288 K.G. Slankard, D.L. Baxley, and G.L. Sprandel 2018 Vol. 25, No. 2 for expansive tracts of homogenous native grasses was not compatible with conservation of Northern Harrier nesting habitat, and nesting success of the species in idled grasslands was higher than in managed grasslands. Most of these studies recognized that management would be necessary to keep vegetation in an early successional state, but left its importance at that, focusing more on the Northern Harrier’s needs for nesting structure. Nonetheless, the habitat that a raptor selects serves several important functions, such as areas for nesting/roosting and foraging. Our study did not take into account the locations of nests, but our results suggest that Northern Harriers prefer more recently managed vegetation for foraging during both the nesting and winter season. Furthermore, our study demonstrates the compatibility of native grass restoration with habitat management for grassland raptors. Acknowledgments We thank the US Fish and Wildlife Service for funding this project through the State Wildlife Grants Program. We also thank the many field staff and volunteers that assisted with field work on this project including: Jim Barnard, Megan Connor, Jared Handley, Ben Leffew, John MacGregor, Tonya Mammone, Brainard Palmer-Ball Jr., Daniel Stoelb, Shawchyi Vorisek, Eric Williams, and PWMA staff. We acknowledge John MacGregor and Keith Wethington for their advice on the study design and Jim Barnard for his review of the manuscript. We also thank this journal’s anonymous reviewers. Literature Cited Baker, J.A., and R.J. Brooks. 1981. Raptor and vole populations at an airport. Journal of Wildlife Management. 45:390–396. Beyer, H.L. 2004. Hawth’s analysis tools for ArcGIS. Available online at http://www.spatialecology. com/htools. Accessed 3 February 2016. Bibby, C.J., N.D. Burgess, D.A. Hill, and S.H. Mustoe. 2000. Bird Census Techniques. Academic Press, London, UK. 302 pp. Brennan, L.A., and W.P. Kuvlesky. 2005. North American grassland birds: An unfolding conservation crisis? Journal of Wildlife Management 69:1–13. Buckland, S.T., D.R. Anderson, K.P. Burnham, J.L Laake, D.L. Borchers, and L. Thomas. 2001. Introduction to Distance Sampling. Oxford University Press, Oxford, UK. 448 pp. Dechant, J.A., D.H. Sondreal, D.H. Johnson, L.D. Igl, C.M. Goldade, M.P. Nenneman, and B.R. Euliss. 2003. Effects of management practices on grassland birds: Northern Harrier. Northern Prairie Wildlife Research Center, Jamestown, ND. 19 pp. de Smith, M., M. Goodchild, and P. Longley. 2009. Geospatial Analysis: A Comprehensive Guide to Principles, Techniques, and Software Tools, 3rd Edition. Splint, Leicester, UK. 516 pp. Hamerstrom, F. 1979. Effect of prey of predator: Voles and harriers. The Auk 96:370–374. Hamerstrom, F., and M. Kopney. 1981. Harrier nest-site vegetation. Journal of Raptor Research 15:86–88. Herkert, J.R., S.A. Simpson, R.L. Westmeier, T.L. Esker, and J.W. Walk. 1999. Response of Northern Harriers and Short-eared Owls to grassland management in Illinois. Journal of Wildlife Management 63:517–523. Hingtgen, T.M., and W.R. Clark. 1984. Small-mammal recolonization of reclaimed coal surface-mined land in Wyoming. Journal of Wildlife Management 48:1255–1261. Northeastern Naturalist Vol. 25, No. 2 K.G. Slankard, D.L. Baxley, and G.L. Sprandel 2018 289 Homer, C.G., J.A. Dewitz, L. Yang, S. Jin, P. Danielson, G. Xian, J. Coulston, N.D. Herold, J.D. Wickham, and K. Megown. 2015. Completion of the 2011 National Land Cover Database for the conterminous United States: Representing a decade of land-cover–change information. Photogrammetric Engineering and Remote Sensing 81:345–354. Johnson, M.D. 2007. Measuring habitat quality: A review. Condor 109:489–504. Jones, Z.F., and C.E. Bock. 2002. Conservation of grassland birds in an urbanizing landscape: A historical perspective. Condor 104:643–651. Kentucky Department of Fish and Wildlife Resources (KDFWR). 2007. Avian monitoring grant report (unpublished report). Frankfort, KY 40601. 60 pp. KDFWR. 2013. Kentucky's comprehensive wildlife conservation strategy. Available online at http://fw.ky.gov/WAP/Pages/Default.aspx. Accessed 1 August 2016. Kentucky Division of Geographic Information (KDGI). 2007a. Kentucky 2005 land-cover: Anderson level II. Available online at http://kygisserver.ky.gov/geoportal/catalog/search/ resource/details.page?uuid=%7B47906D61-FFB5-4A13-BF2A-EB6A14F17040%7D. Accessed: 1 December 2015. KDGI. 2007b. Kentucky 2006 US Department of Agriculture, Farm Service Agency National Agriculture Imagery Program imagery. Available online at http://kygisserver. ky.gov/geoportal/catalog/search/resource/details.page?uuid=%7B0CC25313-CA4C- 46D4-9FA7-F4448E03749D%7D. Accessed 20 January 2015. Laake, J.L., S.T. Buckland, D.R. Anderson, and K.P. Burnham. 1993. Distance: User’s guide, version 2.0. Colorado Cooperative Fish and Wildlife Research Unit, Colorado State University, Fort Collins, CO. 72 pp. Littlefield, C.D., and D.H. Johnson. 2005. Habitat preferences of migrant and wintering Northern Harriers in northwestern Texas. Southwestern Naturalist 50:448–452. MacWhirter, R.B., and K. Bildstein. 2000. Northern Harrier (Circus cyaneus). Number 210, In A. Poole and F. Gill (Eds.). The Birds of North America, Inc., Philadelphia, PA. 32 pp. Mengak, M.T. 2004. Response of small-mammal populations to fescue-hayfield conversion to native warm-season grasses in Bath County, Virginia. Virginia Journal of Science 55:169–176. Morgan, J., and B. Robinson. 2008. Road to recovery: The blueprint for restoring the Northern Bobwhite in Kentucky. Available online at http://fw.ky.gov/Hunt/Documents/ quailplan08.pdf. Accessed 1 August 2016. NatureServe. 2015. NatureServe Explorer: An online encyclopedia of life, Version 7.1. Available online at http://explorer.natureserve.org. Accessed 3 December 2015. North American Bird Conservation Initiative (NABCI), US Committee. 2017. The state of the birds 2017: A farm-bill special report. Cornell Lab of Ornithology, Ithaca, NY. 4 pp. Palmer-Ball, B.L., Jr. 1996. The Kentucky Breeding Bird Atlas. University Press of Kentucky, Lexington, KY. 372 pp. Palmer-Ball, B.L., Jr. 2003. The Annotated Checklist of the Birds of Kentucky. Gateway Press, Louisville, KY. 182 pp. Preston, C.R. 1990. Distribution of raptor foraging in relation to prey biomass and habitat structure. Condor 92:107–112. Rice, W.R. 1989. Analyzing tables of statistical tests. Evolution 43:223–225. Slater, G.L., and C. Rock. 2005. Northern Harrier (Circus cyaneus): A technical conservation assessment. Available online at https://www.fs.usda.gov/Internet/FSE_DOCUMENTS/ stelprdb5182007.pdf. Accessed 1 February 2015. Stewart, J.N. 2004. Pellet analysis of wintering Northern Harriers on reclaimed strip mines in western Kentucky. Undergraduate Thesis. Eastern Kentucky University. Richmond, KY. 4 pp. Northeastern Naturalist 290 K.G. Slankard, D.L. Baxley, and G.L. Sprandel 2018 Vol. 25, No. 2 Stone, E.R. 2007. Measuring impacts of restoration on small mammals in a mixed-grass Colorado prairie. Ecological Restoration 25:183–190. Toland, B.R. 1986. Nesting ecology of Northern Harriers in Southwest Missouri. Transactions of the Missouri Academy of Science 20:49–57. Toland, B.R. 1987. The effect of vegetative cover on foraging strategies, hunting success, and nesting distribution of American Kestrels in Central Missouri. Journal of Raptor Research 21:14–20. Vukovich, M. 2004. Use of reclaimed surface-mine grasslands by Northern Harriers, Shorteared Owls and other raptors: Abundance, behavior, and nesting success. M.Sc. Thesis. Eastern Kentucky University. Richmond, KY. 81 pp. Vukovich, M., and M. Monroe. 2005. Winter abundance of Northern Harriers, Short-eared Owls, and other raptors on reclaimed grasslands in West-Central Kentucky. Kentucky Warbler 81:46–54. Vukovich, M., and G. Ritchison. 2006. Nesting success and behavior of Northern Harriers on a reclaimed surface-mine grassland in Kentucky. Journal of Raptor Research 40:210–216. Wilson, A., M. Brittinghan, and G. Grove. 2010. Association of wintering raptors with Conservation Reserve Enhancement Program grasslands in Pennsylvania. Journal of Field Ornithology 81:361–372. Yeiser, J.M., D.L. Baxley, B.A. Robinson, J.J. Morgan, J.N. Stewart, and J.O. Barnard. 2015. A comparison of coal-mine reclamation seed mixes in Kentucky: Implications for grassland establishment in Appalachia. International Journal of Mining, Reclamation, and Environment 30:257–267.