Southeastern Naturalist C.G. Crouch, et al. 2019 Vol. 18, No. 2 240 2019 SOUTHEASTERN NATURALIST 18(2):240–255 Comparative Habitat Use of Wintering American Kestrels and Loggerhead Shrikes Along South Texas Roadways Carter G. Crouch1,6,*, Aidan J. Flores2, Anastasia Krainyk3, Leonard A. Brennan1, David B. Wester1, Eric D. Grahmann1, Robert H. Benson4, Fidel Hernández1, and Jeffrey F. Kelly5 Abstract - Falco sparverius (American Kestrel) and Lanius ludovicianus (Loggerhead Shrike) have undergone major population declines, which may be partially due to habitat changes on their wintering grounds. The objectives of this study were to compare quantitative habitat metrics and landcover data at sites used by American Kestrels and Loggerhead Shrikes during the winter. We conducted this study by observing individuals of both species along private and public roads across 8 counties in South Texas and recording GPS coordinates and various habitat characteristics at each location. We calculated woody canopy cover by digitizing all woody cover within 100 m of the sighting locations and at random points, and obtained landcover data from the National Land Cover Database (NLCD) for 2001 and 2011. Female American Kestrels were associated with shorter herbaceous height (median = 7.4 cm) than male American Kestrels (16.4 cm) or Loggerhead Shrikes (13.6 cm). Woody canopy cover was lower around sites used by female American Kestrels (q0.9 = 1.9%), compared to sites used by male American Kestrels (q0.9 = 17.8%), Loggerhead Shrikes (q0.9 = 21.3%), or at random points (q0.9 = 51.6%). Although cultivated crops were the most common cover type within 100-m radius buffers of each species group, this cover type was more abundant (73.2% ± 1.5) in sites used by female American Kestrels than sites used by male American Kestrels (47.7% ± 4.1) or Loggerhead Shrikes (41.1% ± 2.7). In contrast, sites used by female American Kestrel had lower amounts of pasture/hay, grassland/ herbaceous, and shrub/scrub types compared to those used by male American Kestrels or Loggerhead Shrikes. Two of the most important cover classes for American Kestrels and Loggerhead Shrikes, cultivated crops and pasture/hay, still comprised 26.0% (481,882 ha) and 22.3% (413,396 ha) of these 8 South Texas counties in 2011, but had decreased by 0.77% and 0.84% since 2001. Medium- and high-intensity development has increased, for a total of 3360 ha (15.49% change), over the same timeframe. Slowing the loss of these favorable open-cover habitat types should be a priority for the conservation of American Kestrels and Loggerhead Shrikes in South Texas. Introduction Falco sparverius L. (American Kestrel; hereafter, Kestrel) have experienced severe long-term declines in portions of their range in North America (Bird 2009, 1Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville, Kingsville, TX 78363. 2Dick and Mary Lewis Kleberg College of Agriculture, Natural Resources and Human Sciences, Kingsville, TX 78363. 3USGS Patuxent Wildlife Research Center, Laurel, MD 20708. 4Laboratory of Bioacoustics, Department of Physical and Environmental Sciences (Emeritus), Texas A&M University-Corpus Christi, Corpus Christi, TX 78412. 5Oklahoma Biological Survey and Department of Zoology, University of Oklahoma, Norman, OK. 73019. 6Current address - Natural Resource Department, 100 Pasigo Street, Burns, OR 97720. *Corresponding author - carterGcrouch@gmail.com. Manuscript Editor: Douglas McNair Southeastern Naturalist 241 C.G. Crouch, et al. 2019 Vol. 18, No. 2 Sauer et al. 2017, Smallwood et al. 2009). Lanius ludovicianus L. (Loggerhead Shrike) (hereafter, Shrike) are also experiencing continent-wide declines (Cade and Woods 1997; Sauer et al. 2017; Yosef 1994, 1996). While many hypotheses have been proposed for these declines, land-use changes on wintering grounds may be an important factor. Smallwood et al. (2009) speculated that declines of Kestrels in Florida, Georgia, Virginia, Maryland, New Jersey, Massachusetts, Pennsylvania, Saskatchewan, and the Yukon Territory were primarily caused by factors outside of breeding areas, either along migration routes or on the wintering grounds. Habitat loss along the Gulf Coast is linked to declines in Shrike populations that breed in the Midwest (Lymn and Temple 1991), and loss of pastureland was linked to a decline of Shrikes in South Carolina and nationwide (Froehly et al. 2019, Gawlik and Bildstein 1993). Yosef (1996) recommended further investigation into degradation of the wintering-ground habitat and its effects on migrating populations as a priority for future research on Shrikes. According to Yosef (1996), key management priorities for future research on Shrikes should be to determine the degree of niche overlap between Shrikes and potential competitors and to investigate whether the presence of these competitors affects Shrike productivity. Although habitat studies on Kestrels and Shrikes are common (Ardia and Bildstein 2001, Bohall-Wood 1987, Bohall-Wood and Collopy 1986, Gawlik and Bildstein 1993, Koplin 1973, Mills 1976, O’Brien and Ritchison 2011, Prescott and Collister 1993, Stinson et. al 1981), comparative habitat-use studies between the 2 species are not (but see Mills 1979). Mills (1979) believed that competition between Kestrels and Shrikes should occur due to their similarity in foraging behavior. If prey items are limiting populations in open landscapes, Kestrels and Shrikes may be competitors, given that both species eat arthropods and small vertebrates, including small mammals, birds, and reptiles (Smallwood and Bird 2002, Yosef 1996). Bildstein and Grubb (1980) determined that, although Kestrels and Shrikes inhabited some of the same areas, these 2 species showed spatial segregation and were not randomly distributed in relation to each other. However, those authors did not measure habitat variables or document differences in habitat metrics between the 2 species. Herbaceous vegetation height is an important factor for habitat use in Kestrels (Ardia and Bildstein 2001, Sheffield et al. 2001, Smallwood 1987) and Shrikes (Bohall- Wood 1987, Fornes 2004, Gawlik and Bildstein 1993, O’Brien and Ritchison 2011, Prescott and Collister 1993), although measurements are frequently categorical and qualitative in nature. Mills (1979) reported that Kestrels and Shrikes hunt exclusively over short grass, even though grasshoppers appeared more abundant in taller grass. Toland (1987) found that hunting success decreased with increasing vegetation height for Kestrels. The literature is inconsistent when it comes to the relationship between vegetation height and habitat use by Shrikes. While some studies indicate that short vegetation is important for Shrike habitat (Bohall-Wood 1987, Gawlik and Bildstein 1993, O’Brien and Ritchison 2011), Prescott and Collister (1993) indicated that occupied Shrike territories had a greater percentage of tall grass (≥20 cm) than nearby unoccupied areas and a taller average grass height Southeastern Naturalist C.G. Crouch, et al. 2019 Vol. 18, No. 2 242 (20.0 cm vs. 15.8 cm, based on back-transformed means) than unoccupied areas. Results of studies on mowing effects are also inconclusive, as both mowed and unmowed patches of grassland received similar Shrike use in Texas (Chavez-Ramirez et al. 1994). In Florida, Shrikes altered foraging behavior in response to manipulations of vegetation height through mowing, but did not adjust their territory sizes (Yosef and Grubb 1993). Canopy cover of woody vegetation is also a major component in habitat selection and use for Kestrels (Ardia and Bildstein 2001, Bohall-Wood and Collopy 1986, Koplin 1973, Mills 1976, Smallwood 1987, Stinson et. al 1981) and Shrikes (Bohall-Wood 1987). Sexual segregation based on woody canopy cover is well established in Kestrels, with females inhabiting areas that are more open and have less woody cover. Open areas with less woody cover are assumed to provide better habitat than areas that are less open (Ardia and Bildstein 2001, Bohall-Wood and Collopy 1986, Koplin 1973, Mills 1976, Stinson et. al 1981). Smallwood (1987) showed that male Kestrels occupied sites with more woody cover and spent more time foraging than females, were more likely to attack novel prey items, and experienced greater loss of body mass than females. Furthermore, male Kestrels wintering in Pennsylvania had lower and declining body condition compared to females, which maintained and did not lose their condition (Ardia 2002). Ardia (2002) attributed this difference in part to site differences. Quantitative studies of the use of woody canopy cover by these 2 species have been limited to Smallwood’s (1987) findings of the proportion of a Kestrel’s territory covered by woody cover in south-central Florida. Other studies exploring the use of woody canopy cover by Kestrels and Shrikes tend to be categorical and qualitative in nature and thus lack some detail (Ardia and Bildstein 2001, Bohall-Wood 1987, Bohall-Wood and Collopy 1986, Koplin 1973, Mills 1976). The objectives of this study were to: (1) measure and compare several important winter habitat variables among sites used by female and male Kestrels and Shrikes in South Texas, including the extent of woody canopy cover among sites used by these birds and at randomly selected sites; and (2) compare landcover use categories at Kestrel and Shrike locations and changes in landcover data over a 10-y period to examine if suitable cover categories were declining in our study area. We limited inferences of habitat use to roadsides, as we did not sample off of roadways due to the inaccessibility of private property bordering roadways. Following Bildstein and Grubb (1980), we hypothesized that sites used by Kestrels and Shrikes would differ in vegetation structure (i.e., herbaceous height and woody canopy cover). Due to the importance of pasture to Shrikes (Froehly et al. 2019, Gawlik and Bildstein 1993) and the abundance of Kestrels in plowed agricultural areas, we hypothesized that, compared to Kestrels, Shrikes would use taller herbaceous vegetation at sites that included a higher percentage of pasture. We also hypothesized that sites used by male Kestrels would have greater average woody canopy cover than sites used by females. Finally, we predicted that randomly selected sites would have greater woody canopy cover than sites used by both species because they prefer more-open country. Southeastern Naturalist 243 C.G. Crouch, et al. 2019 Vol. 18, No. 2 Methods Field-site description Our study area was along private property roadways on the San Christoval Ranch, the Rob and Bessie Welder Wildlife Refuge, the Mad Island Marsh Preserve, and public roadways (county roads in 7 counties: Bee, Jim Wells, Karnes, Kleberg, Live Oak, Nueces, and San Patricio; farm-to-market roads) in South Texas. The San Christoval Ranch is located in Karnes and Live Oak counties; the Welder Wildlife Refuge is located in San Patricio County; and the Mad Island Marsh Preserve is a property of The Nature Conservancy located in Matagorda County. Historical management on the properties includes grazing, prescribed fires, and mechanical treatments. Agricultural fields (>60%) and cattle pastures comprised the majority of the study area, with farmhouses and farming operations interspersed throughout. Common crops planted in these agricultural fields were Gossypium spp. (cotton), Sorghum spp. (sorghum), and Triticum spp. (winter wheat). Non–native grasses common in pastures and fence rows included Dicanthium and Bothriochloa spp. (Old World bluestems) and Cynodon dactylon (L.) Pers. (Bermudagrass). The study area included patches of woody cover interspersed between agricultural fields, within cattle pastures, along fence rows, and planted around farm houses. Common woody plants included Prosopis glandulosa Torr. (Honey Mesquite), Vachellia farnesiana (L.) White et Arn. (Huisache), Quercus spp. (oak), and Celtis spp. (hackberry). Total precipitation in Corpus Christi, TX, from November to February was 18.6 cm in 2014–2015 and 14.2 cm in 2015–2016. Average high and low temperatures from November to February were 19.6 °C and 9.5 °C in 2014–2015 and 22.0 °C and 11.0 °C in 2015–2016, respectively (US Climate Data 2017). We preferentially sampled roads with powerlines because of the high visibility and density of individual birds. We sampled private and public roadways each once over the course of 16 d during November 2014–February 2015 and 18 d during November 2015–February 2016. We sighted Kestrels and Shrikes by driving along private and public roads at speeds of ≤60 kph and recorded GPS locations with a GPSMAP64s and GPSMAP62stc (GARMIN®, Olathe, KS) for the area in which the bird was first observed. We identified Kestrels to sex, but did not differentiate sexes for Shrikes due to the similarity of sexes (Yosef 1996). We treated all individuals independently. If the bird was perched, we noted perch structure and measured perch height with a tape measure if it was <2.5 m or used a Nikon Pro Staff 3 Rangefinder (Nikon Inc., Tokyo, Japan) if it was >10 m. For heights from 2.5 m to 10 m, we used ocular estimation to determine height. If the bird was flying, we recorded a GPS location where we first observed the bird. We excluded flying birds and birds perched on the ground for estimates of perch height. Habitat use We measured herbaceous height, woody cover, and distance to woody plants (for Kestrels only) centered on the points where birds were observed. We excluded points >30 m from roadways, by using the roadways layer (The Transportation Southeastern Naturalist C.G. Crouch, et al. 2019 Vol. 18, No. 2 244 Planning and Programming Division of the Texas Department of Transportation 2016), digitizing private roadways travelled, and using the “Near” tool in ArcMap 10.4.1 (ArcGIS, ESRI, Redlands, CA). We took all measurements immediately after finding the bird, with the exception of woody cover and landcover data, which we collected later via ArcMap. We measured herbaceous (broad-leaved plant or graminoid) height in the location where the bird was first spotted and locations 10 paces in the 4 cardinal directions (N, E, S, W) for a total of 5 sampled points per bird. We used a tape measure to determine effective herbaceous height within a 25-cm2 area around the point, excluding obvious outlier stalks or blades that stood out above the remainder of the vegetation. If a point fell on a roadway or on bare soil, we recorded herbaceous height as 0 cm. If we could not access a point (private property), we determined herbaceous height using ocular estimation. We did not collect herbaceous height data if the bird location was >20 m from an accessible point (i.e., across fencerows on private property). We used the average (5 points/bird) herbaceous height around the sighting location as the estimate of herbaceous height. To compare woody canopy cover use, we took a subset of female Kestrel and Shrike locations to compare to our sample of male Kestrels (n = 80). We separated Kestrel and Shrike locations into 8 geographical sections based on location spacing. There were no female Kestrel locations in 1 of the geographical sections, so we used 7 sections for female Kestrels. We then randomly selected 80 female Kestrels and 80 Shrikes using geographic stratification by the 7 and 8 sections to ensure the random subset of samples was geographically representative of all of the data collected. To obtain random available points, we used ArcMap to generate 80 random points (no minimum distance between points) within 30 m of roadways that were within 500 m of points used. To calculate woody canopy cover at sites used (n = 80 for male Kestrels, female Kestrels, and Shrikes) and randomly selected sites (n = 80), we digitized woody cover within a 100-m radius centered at either the used point or a randomly selected point. We digitized woody cover based on the 2015 Texas Orthoimagery Program Data (50 cm resolution NC/CIR data; Texas Natural Resources Information System, Austin, TX). We digitized all visible woody cover at a scale of 1:705. We used the 2014 National Agriculture Imagery Program (NAIP) Orthoimagery (1-m resolution; US Department of Agriculture Farm Service Agency) and GoogleEarth (Google, Inc., Mountain View, CA) to confirm woody cover classifications. After digitizing woody cover, we calculated the woody canopy cover percentage at 3 spatial extents (10-m–, 50-m–, and 100-m–radius buffers) by dividing the area of the digitized woody cover by the total area of the buffered circle. We used multiple spatial extents (10 m, 50 m, and 100 m) because habitat selection can occur at multiple extents (Johnson 1980) and Ardia and Bildstein (2001) previously compared male and female Kestrel habitat at 10 m and 100 m extents. We measured the distance to woody cover that was ≥1.5 m tall in each quadrant (NE, SE, SW, and NW) from each Kestrel’s location (Cottam and Curtis 1956) using a Nikon Pro Staff 3 Rangefinder (Nikon, Inc., Melville, NY). Unfortunately, we only collected this data for Kestrels, due to a higher percentage of early Southeastern Naturalist 245 C.G. Crouch, et al. 2019 Vol. 18, No. 2 observations of Shrikes perching in or above trees. The rangefinder only accurately recorded distances of ≤500 m; thus, we truncated all values ≥500 m as 500 m. We report the average (4 quadrants) and minimum distances to woody cover (≥1.5 m). Landcover use and changes For reporting the NLCD cover types, we condensed the categories provided by Homer et al. (2015) into 8 new categories. We kept developed (open space), developed (low intensity), shrub/scrub, grassland/herbaceous, pasture/hay, and cultivated crops, and we combined developed (medium intensity) and developed (high intensity) into a new category called developed (≥medium intensity). We also combined all other categories into an “other” category. To calculate the percentage of land-use categories found at sites used by female Kestrels, male Kestrels, and Shrikes, we intersected the 2011 condensed NLCD data with 100-m buffers around all bird locations found within 30 m of a roadway. To calculate landcover change in our study area, we clipped both the 2001 and 2011 condensed NLCD data sets to the boundaries of the 8 counties included in this study. Statistical analysis We considered the linear distance of the roadways layer clipped by a 500-m–radius buffer around all use points as a metric of the roadway distance sampled. This was an underestimate of the actual sampling distance traveled as it only included stretches of roadway where Kestrels and Shrikes were observed, but we did not track total sampling distance. Habitat measurements are assumed to be independent because we only sampled each road once and collected vegetation data once for each bird. We pooled habitat data between years (2014–2015 and 2015–2016). We compared perch use among female Kestrels, male Kestrels, and Shrikes by categorizing these observations into: (1) powerlines, and (2) other (fence, flying ground-pipes and posts, power structure [excluding line], road sign, and trees and shrubs), and analyzing with a chi-square test of independence. A complete breakdown of perches used by each group can be found in Crouch (2017). For other habitat variables, a Shapiro-Wilk test of normality (Shapiro and Wilk 1965) indicated a normality assumption violation for the raw data; thus, we natural-log–transformed the data. Therefore, we analyzed perch height, herbaceous height, and woody canopy cover within buffers with 10-m, 50-m, and 100-m radii using a Kruskal-Wallis test (Kruskal and Wallis 1952) on the complete set (all groups). We performed pairwise comparisons (Mann-Whitney U test) between groups (male Kestrels, female Kestrels, Shrikes, randomly selected sites) following a significant Kruskal-Wallis test (Carmer and Swanson 1973, Smith and Han 1981). We used the Mann-Whitney U test to compare the average (4 quadrants) and minimum distance to woody cover (≥1.5 m tall) between male and female Kestrels.. Due to the similarity of results for woody canopy cover for the 3 buffers, we report only the results for the 10-m–radius buffer, but results for the 50-m and 100-m buffers can be found in Crouch (2017). To test the hypothesis that habitat composition did not differ among the 3 species groups, we compared habitat composition (condensed cover classes) at sites used by female Kestrels, male Kestrels, and Shrikes with a permutational multivariate Southeastern Naturalist C.G. Crouch, et al. 2019 Vol. 18, No. 2 246 analysis of variance using a Euclidean distance similarity measure (Anderson 2001, 2017). This omnibus hypothesis was rejected; thus, we followed up comparisons among species groups for each habitat type with permutational univariate analysis of variance, and when this test was significant, we followed up with pairwise comparisons among species groups. We report the total acreage and the percentage of the 8 counties in important cover types using NLCD 2011 data. We also report the change in acreage and the percentage change for these important classes over a 10-y period (NLCD 2001– NLCD 2011). Results For this study we sampled 299.43 km of roadways for Kestrels and Shrikes. Powerlines were the most frequent type of perch, and their use did not differ (χ2 2df = 2.82, P = 0.24) among female Kestrels (83.8%), male Kestrels (76.3%), and Shrikes (84%). Perch height also did not differ (H2df= 2.83, P = 0.24) among groups with a median of 7 m for all groups. Habitat use Herbaceous height at used sites differed (H2df = 29.3, P < 0.001) among the 3 groups (female Kestrels, male Kestrels, Shrikes; Fig. 1). The median herbaceous height used by female Kestrels was significantly shorter (7.4 cm) than heights used by male Kestrels (16.4 cm; U1df = 17.2, P < 0.001) or Shrikes (13.6 cm; U1df = 18.7, P < 0.001). We detected no difference (U1df = 0.73, P = 0.39) in the median herbaceous heights at sites used by male Kestrels and Shrikes. Female Kestrels were located farther from woody canopy cover ≥1.5 m tall than male Kestrels with respect to average (4 quadrants) and minimum distances; U1df = 38.14, P < 0.0001; U1df = 30.67, P < 0.0001; Fig. 2). Woody canopy cover at 10 m differed (H3df = 16.58, P = 0.0003) among the 4 groups (female Kestrels, male Kestrels, Shrikes, random points; Fig. 3), although the distribution is strongly skewed for all 4 groups. Woody canopy cover did not differ (P > 0.30) among male Kestrels (q0.9 = 17.8%), Shrikes (q0.9 = 21.3%), or random points (q0.9 = 51.6), but was significantly lower (P < 0.0008) at sites used by female Kestrels (q0.9 = 1.9%) compared to any of the other 3 groups. Landcover use and changes Habitat composition differed (F2,627 = 43.4, P < 0.0001) among species groups. Although cultivated crops were the most common cover type within 100-m radius buffers of each species group, this cover type was more abundant (mean ± SE; 73.2% ± 1.5) in sites used by female Kestrels than sites used by male Kestrels (47.7% ± 4.1) or Shrikes (41.1% ± 2.7). In contrast, sites used by female Kestrel had lower amounts of pasture/hay, grassland/herbaceous and shrub/scrub types than sites used by male Kestrels or Shrikes (Table 1). Two of the most important cover classes for Kestrels and Shrikes, cultivated crops and pasture/hay, still comprised 26.0% (481,882 ha) and 22.3% (413,396 ha) of these 8 South Texas counties in 2011 (Table 1). Developed (open space) Southeastern Naturalist 247 C.G. Crouch, et al. 2019 Vol. 18, No. 2 comprised 3.7% (67,940 ha) and shrub/scrub comprised 26.7% (495,024 ha) in 2011. A class not used much by either species, developed (≥medium intensity), only comprised 1.3% (25,047 ha) of the 8 counties in 2011. Cultivated crops and pasture/ hay have decreased by 3733 ha (-0.77% change) and 3501 ha (-0.84% change) since 2001, while developed (≥medium intensity) has increased for a combined total of 3360 ha (15.49% change). Developed (open space) and shrub/scrub have increased by 0.11% (73 ha) and 0.01% (49 ha), respectively. Discussion Powerlines were the most commonly used perch structures for Kestrels and Shrikes in this study, although our results are very likely biased due to our protocol. Nonetheless, powerlines have also been the most commonly used perch for Kestrels Figure 1. Herbaceous height frequency and quantiles at sites used by female American Kestrels, male American Kestrels, and Loggerhead Shrikes. Habitat data collected across private property and public roadways in 8 South Texas counties in November–February, 2014–2015 and 2015–2016. Southeastern Naturalist C.G. Crouch, et al. 2019 Vol. 18, No. 2 248 Table 1. The percentage of the National Land Cover Database (NLCD 2011) condensed classes found within 100-m radius buffers of female American Kestrel, male American Kestrel, and Loggerhead Shrike locations. We obtained the area (ha) of classes in 8 counties (Area) from NLCD 2011 data and the cover class % change from 2001 to 2011 (Cover class change) by comparisons of NLCD 2001 and 2011 data. Means among species groups within a habitat type followed by the same letter are not significantly different (protected LSD test: P > 0.05). Female Kestrel Male Kestrel Shrike Cover class Habitat type (n = 388) (n = 80) (n = 162) Area (ha) change Cultivated crops 73.2A ± 1.5 47.7B ± 4.1 41.1B ± 2.7 481,882 -0.77% Developed (low intensity) 4.2A ± 0.3 5.4AB ±1.1 6.5B ± 0.7 33,684 +2.63% Developed (≥ medium intensity) 0.4A ± 0.1 0.6 A ± 0.3 0.5A ± 0.5 25,047 +15.49% Developed (open space) 11.0A ± 0.5 13.3 AB ± 1.1 15.8B ± 0.8 67,940 +0.11% Grassland/herbaceous 1.8A ± 0.4 4.3B ± 1.3 4.5B ± 0.9 87,011 -0.51% Other 1.2A ± 0.4 3.5B ± 1.2 2.5AB ± 0.8 252,474 +1.27% Pasture/hay 6.9A ± 0.9 15.2B ± 2.7 22.3B ± 2.2 413,396 -0.84% Shrub/scrub 1.2A ± 0.3 10.0B ± 2.2 6.9B ± 1.2 495,024 +0.01% Figure 2. Frequency and quantiles of the average (4 quadrants) and minimum distances to woody cover (brush) at sites used by female and male American Kestrels. Habitat data collected as in Figure 1. Southeastern Naturalist 249 C.G. Crouch, et al. 2019 Vol. 18, No. 2 (Bauer 1982, Bildstein 1978, Bildstein and Grubb 1980, Langley 1999, Worm et. al 2013) and Shrikes (Bildstein and Grubb 1980, Bohall-Wood 1987, Gawlik and Bildstein 1993) in many other studies across the US. However, trees and fences were the most common perch for Shrikes in Kentucky (O’Brien and Ritchison 2011) and mesquite was the primary perch for Shrikes in the Rolling Plains of Texas (Becker et al. 2009). The Becker et al. (2009) and O’Brien and Ritchison (2011) studies were not road-based surveys, and the differences in perches used in these 2 studies are likely related to this. The median perch height for Kestrels (7 m) in this study is similar to previously reported values of perch height in roadside surveys (Bildstein 1978, Worm et al. 2013). The median perch height for Shrikes (7 m) in this study is similar to the perch heights observed in South Carolina (Gawlik and Bildstein 1993), but higher than the average perch heights observed in non-roadside surveys (Becker et al. 2009, O’Brien and Ritchison 2011). Habitat use As we hypothesized, sites used by Kestrels differed in herbaceous height compared to sites used by Shrikes, but only when we compared female Kestrels and Figure 3. Frequency and quantiles of the canopy cover (10-m buffer) at sites used by female American Kestrels, male American Kestrels, Loggerhead Shrikes, and random points. Woody canopy cover given in percent. Habitat data collected as in Figure 1. Southeastern Naturalist C.G. Crouch, et al. 2019 Vol. 18, No. 2 250 Shrikes. In this study, we observed female Kestrels at sites that had lower herbaceous height than both male Kestrels and Shrikes; however, a majority of all 3 groups were in sites with herbaceous height of ≤25 cm. Other researchers have considered less than 25-cm–height herbaceous cover as suitable foraging substrate for these 2 species (Ardia and Bildstein 2001; Smallwood 1987, 1988). In an experimental study in Oregon, Kestrels preferred short vegetation (8 cm) over tall vegetation (~20 cm; Sheffield et al. 2001). Although an average herbaceous height of less than 25 cm has been considered suitable, Kestrels may prefer shorter vegetation. Herbaceous heights around sites used by Shrikes in this study (median = 13.6 cm) were taller than herbaceous height under Shrikes perched in South Carolina during the winter (mean = 0.4 cm) and other seasons (mean = 0.7 cm) (Gawlik and Bildstein 1993) and around sites used by Shrikes wintering in Kentucky that were associated with continuous perches (mean = 7.1 cm; O’Brien and Ritchison 2011). Herbaceous heights around sites used by Shrikes in this study were similar to those reported for Shrikes wintering in Kentucky that were associated with isolated perches (mean = 12.7 cm; O’Brien and Ritchison 2011), and shorter than the herbaceous heights used by breeding Shrikes in Alberta (20.0 cm, based on back-transformed means; Prescott and Collister 1993). We limited herbaceous height measurements to a small buffer (10-m radius) around points used by Kestrels and Shrikes, although Kestrels and Shrikes likely select sites based on herbaceous height at a larger spatial extent than 10 m around perch points. As we hypothesized, sites used by Kestrels differed in woody canopy cover compared to sites used by Shrikes. However, this difference was only the case when comparing female Kestrels and Shrikes. The use of greater woody canopy cover differentiated male Kestrels and Shrikes from female Kestrels. Similar to findings of other researchers (Ardia and Bildstein 2001, Bohall-Wood and Collopy 1986, Koplin 1973, Mills 1976, Smallwood 1987, Stinson et. al 1981), male Kestrels in this study used areas with more woody cover than female Kestrels and perched closer to woody cover. Although woody canopy cover differed between male and female Kestrels at the 10-m, 50-m, and 100-m spatial extents, the average woody canopy cover was more similar at the 10-m spatial extent than the 100-m spatial extent (Crouch 2017). This result is similar to Ardia and Bildstein’s (2001) findings of greater differentiation in habitat used at the 100-m spatial extent compared to the 10-m spatial extent. At the 100-m spatial extent, those authors found that as the percentage of woodlot increased, the probability the bird was male increased. In our study, increasing woody canopy cover with an increase in spatial extent supports the hypothesis of Ardia and Bildstein (2001) that males select smaller areas devoid of woody vegetation relative to what is available within a 100-m radius of perch sites. Similarly, Mills (1976) found male Kestrels in heavily wooded areas tended to be found in openings within this landscape. The randomly selected sites also differed from sites used by female Kestrels, with greater woody canopy cover at the randomly selected sites. Although the average woody canopy cover at randomly selected sites was greatest, it was not statistically different from woody canopy cover at sites used by male Kestrels or Shrikes. The lack of a difference in woody Southeastern Naturalist 251 C.G. Crouch, et al. 2019 Vol. 18, No. 2 canopy cover between randomly selected sites and sites used by male Kestrels and Shrikes is likely due in part to uniform habitat characteristics at even larger spatial extents than we sampled. That is, the extent of open habitats was much larger than the 500-m buffer we sampled around sites used by Kestrels and Shrikes. Landcover use and changes The finding that cultivated crops made up the highest percentage of sites used by female Kestrels, male Kestrels, and Shrikes is unsurprising, as cultivated crops made up the majority of our survey area. Gawlik and Bildstein (1993) also found Shrikes increasing their use of cropland in the winter compared to the breeding season. However, the differences in landcover in sites used by these 3 groups are telling. As we predicted, pasture/hay was more important for Shrikes than Kestrels but was significantly different only when compared to female Kestrels. Froehly et al. (2019) found occupancy of breeding Shrikes was best predicted by the percentage of pasture within 1 km. Smallwood (1988) determined that later arrival of male Kestrels to the wintering grounds in Florida led to them using lower-quality habitat with a woodier canopy cover. The higher woody canopy cover in sites used by male Kestrels compared to female Kestrels may not be a result of selection as much as establishing territories in the unoccupied areas that have more woody cover. By arriving earlier than males, females can select more-open areas with shorter vegetation, which are abundant in the agricultural areas in the winter. We surveyed more open areas; thus, it is logical that we would detect more females. Female Kestrels also migrate farther distances than males (Heath et al. 2012), so there may also be fewer males arriving in South Texas. Figure 4. Numbers per party-hour of American Kestrel and Loggerhead Shrike from Christmas Bird Count data in Texas, 1980–2015 (National Audubon Society 2010). Southeastern Naturalist C.G. Crouch, et al. 2019 Vol. 18, No. 2 252 Wintering Kestrels and Shrikes are declining in Texas. Statewide Christmas Bird Count data documents a steady decline for both species from 1980 to 2015 (National Audubon Society 2010; Fig. 4). Recent eBird data (2007–2016) from Texas documents a decline in non-breeding Kestrels, including within much of South Texas (Fink et al. 2018). The 8-county area of South Texas that we studied still provides a large amount of land in cultivated crops and pasture/hay, but we did detect a small decline from 2001 to 2011 in these cover categories. We also found an increase in developed land (≥medium intensity). While these losses were small, any decrease in available habitat for these 2 declining species should be a concern. McClure et al. (2017) called for demographic studies of Kestrels outside of the breeding season, and such research should be a priority for both species. Crouch (2017) examined the wintering ecology of Kestrels in agricultural areas of South Texas, including 3 of the 8 counties included in this study. That study found high site fidelity, small territories, and an annual apparent survival rate of 48.2%. Given these results, he suggested agricultural areas in South Texas were an important region for wintering Kestrels. Replication of that study in other vegetation communities and other regions would be helpful in better understanding the demographics of wintering Kestrels. While Chabot et al. (2016) studied demographics of breeding Shrikes on tallgrass prairie in Illinois, demographic studies on wintering Shrikes are sorely needed. Relating winter demographics to habitat use on the wintering grounds would also be helpful to understanding the decline of these 2 species. Management Recommendations Kestrels and Shrikes are able to use different open, herbaceous cover types in South Texas, including cultivated crops and pastures or hay fields. The current South Texas landscape still provides a large amount of habitat for Kestrels and Shrikes. Maintaining these open and herbaceous cover types along roadsides while limiting medium- and high-intensity development in occupied areas should be top conservation priorities for management of Kestrels and Shrikes. Clean-farming practices in this region keep large areas of crop fields fallow for months at a time. Although these open areas that are devoid of most vegetation are used by both species, the narrow corridors of grassy and herbaceous vegetation along roadsides may be important for supporting Kestrels, Shrikes, and their food sources. Experimental habitat manipulations of roadside vegetation using native species of grasses and forbs along roadside corridors would be a useful next step for evaluating how Kestrels, Shrikes, and other grassland birds use such areas in landscapes dominated by clean-farming agriculture. Acknowledgments We thank the South Texas Chapter of the Quail Coalition, the Williams family, the Richard M. Kleberg Jr. Center for Quail Research, and the C.C. Winn Endowed Chair for Quail Research for project funding. We also thank the Sportsmen’s Club of Fort Worth for the Harry Tennison Graduate Scholarship and the René Barrientos Tuition Assistance for help with C. Crouch’s tuition costs. We thank the Nature Conservancy and the Mad Island Southeastern Naturalist 253 C.G. Crouch, et al. 2019 Vol. 18, No. 2 Marsh Preserve, the Rob and Bessie Welder Wildlife Refuge, and the San Christoval Ranch for access to the properties for data collection. Comments by K.L. Bildstein, Masahiro Ohnishi, Janel Ortiz, and one anonymous reviewer helped improve this manuscript. This is publication number 19-105 of the Caesar Kleberg Wildlife Research Institute. The methods used for this study were exempt from Texas A&M University-Kingsville IACUC protocol. Literature Cited Anderson, M.J. 2001. A new method of non-parametric multivariate analysis of variance. Austral Ecology 26:32–46. Anderson, M.J. 2017. Permutational multivariate analysis of variance (PERMANOVA). WileyStatsRef: Statistics Reference Online. DOI:10.1002/9781118445112.stat07841. Accessed 15 January 2019. Ardia, D.R. 2002. Energetic consequences of sex-related habitat segregation in wintering American Kestrels (Falco sparverius). Canadian Journal of Zoology 80:516–523. Ardia, D.R., and K.L. Bildstein. 2001. Sex-related differences in habitat use in wintering American Kestrels. Auk 118:746–750. Bauer, E.N. 1982. Winter roadside raptor surveys in El Paso County, Colorado, 1962–1979. Raptor Research 16:10–13. Becker, M.E., P.E. Bednekoff, M.W. Janis, and D.C. Ruthven III. 2009. Characteristics of foraging-perch sites used by Loggerhead Shrikes. Wilson Journal of Ornithology 121:104–111. Bird. D.M. 2009. The American Kestrel: From common to scarce? Journal of Raptor Research 43:261–262. Bildstein, K.L. 1978. Behavioral ecology of Red–tailed Hawks (Buteo jamaicensis), Rough–legged Hawks (B. lagopus), Northern Harriers (Circus cyaneus), American Kestrels (Falco sparverius), and other raptorial birds wintering in south central Ohio. Ph.D. Dissertation. The Ohio State University, Columbus, OH. 364 pp. Bildstein, K.L., and T.C. Grubb Jr. 1980. Spatial distributions of American Kestrels and Loggerhead Shrikes wintering sympatrically in eastern Texas. Raptor Research 14:90–91. Bohall-Wood, P. 1987. Abundance, habitat use, and perch use of Loggerhead Shrikes in north-central Florida. Wilson Bulletin 99:82–86. Bohall-Wood, P., and M.W. Collopy. 1986. Abundance and habitat selection of two American Kestrel subspecies in north-central Florida. Auk 103:557–563. Cade, T.J., and C.P. Woods. 1997. Changes in distribution and abundance of the Loggerhead Shrike. Conservation Biology 11:21–31. Carmer, S.G., and M.R. Swanson. 1973. An evaluation of ten pairwise multiple comparison procedures by Monte Carlo methods. Journal of the American Statistical Association 68:66–74. Chabot, A.A., F. Harty, J. Herkert, and W. Glass. 2016. Population demographics of the Loggerhead Shrike: Insights into the species decline from a long-term study in the Midewin National tallgrass prairie. North American Prairie Conference 13:69–78. Chavez–Ramirez, F., D.E. Gawlik, F.G. Prieto, and R.D. Slack. 1994. Effects of habitat structure on patch use by Loggerhead Shrikes wintering in a natural grassland. Condor 96:228–231. Cottam, G., and J.T. Curtis. 1956. The use of distance measures in phytosociological sampling. Ecology 37:451–460. Crouch, C.G. 2017. Habitat and ecology of grassland birds in South Texas. Ph.D. Dissertation. Texas A&M University-Kingsville, Kingsville, TX. 142 pp. Southeastern Naturalist C.G. Crouch, et al. 2019 Vol. 18, No. 2 254 Fink, D., T. Auer, A. Johnston, M. Strimas-Mackey, M. Iliff, and S. Kelling. 2018. eBird Status and Trends. Version: November 2018. Cornell Lab of Ornithology, Ithaca, NY. Available online at https://ebird.org/science/status-and-trends. Accessed 13 January 2019. Fornes, G.L. 2004. Habitat use by Loggerhead Shrikes (Lanius ludovicianus) at Midewin National Tallgrass Prairie, Illinois: An application of Brooks and Temple’s habitat suitability index. American Midland Naturalist 151:338–345. Froehly, J.A., A.K. Tageler, C.M. Bodinof Jackowski, and D.S. Jachowski. 2019. Effects of scale and land cover on Loggerhead Shrike occupancy. Journal of Wildlife Management 83:426–434. DOI:10.1002/jwmg.21587. Gawlik, D.E., and K.L. Bildstein. 1993. Seasonal habitat use and abundance of Loggerhead Shrikes in South Carolina. Journal of Wildlife Management 57:352–357. Heath, J.A., K. Steenhof, and M.A. Foster. 2012. Shorter migration distances associated with higher winter temperatures suggest a mechanism for advancing nesting phenology of American Kestrels, Falco sparverius. Journal of Avian Biology 43:376–384. 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 landcover change information. Photogrammetric Engineering and Remote Sensing 81:345–354. Johnson, D.H. 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology 61:65–71. Koplin, J.R. 1973. Differential habitat use by sexes of American Kestrels wintering in northern California. Raptor Research 7:39–42. Kruskal, W.H., and W.A. Wallis. 1952. Use of ranks in one-criterion variance analysis. Journal of the American Statistical Association 47:583–621. Langley, W. 1999. Perch and habitat use by Red–tailed Hawks and American Kestrels along a highway in eastern Kansas. Transactions of the Kansas Academy of Science 102:92–99. Lymn, N., and S.A. Temple. 1991. Land-use changes in the Gulf Coast region: Links to declines in Midwestern Loggerhead shrike populations. Passenger Pigeon 53:315–325. McClure, C.J.W., S. E. Schulwitz, R.V. Buskirk, B.P. Pauli, and J.A. Heath. 2017. Commentary: Research recommendations for understanding the decline of American Kestrels (Falco sparverius) across much of North America. Journal of Raptor Research 51:455–464. Mills, G.S. 1976. American Kestrel sex ratios and habitat separation. Auk 93:740–748. Mills, G.S. 1979. Foraging patterns of kestrels and shrikes and their relation to an optimal foraging. Ph.D. Dissertation. University of Arizona, Tucson, AZ. 73 pp. National Audubon Society. 2010. The Christmas Bird Count Historical Results. Available online at http://www.christmasbirdcount.org. Accessed 5 January 2019. O’Brien, E., and G. Ritchison. 2011. Non–breeding ecology of Loggerhead Shrikes in Kentucky. Wilson Journal of Ornithology 123:360–366. Prescott, D.R., and D.M. Collister. 1993. Characteristics of occupied and unoccupied Loggerhead Shrike territories in southern Alberta. Journal of Wildlife Management 57:346–352. Sauer, J.R., D.K. Niven, J.E. Hines, D.J. Ziolkowski, Jr, K.L. Pardieck, J.E. Fallon, and W.A. Link. 2017. The North American Breeding Bird Survey, Results and Analysis 1966–2015. Version 2.07.2017. USGS Patuxent Wildlife Research Center, Laurel, MD. Available online at https://www.mbr–pwrc.usgs.gov/bbs/. Accessed October 2017. Shapiro, S.S., and M.B. Wilk. 1965. An analysis of variance test for normality (complete samples). Biometrika 52:591–611. Southeastern Naturalist 255 C.G. Crouch, et al. 2019 Vol. 18, No. 2 Sheffield, L.M., J.R. Crait, W.D. Edge, and G. Wang. 2001. Response of American Kestrels and Gray-tailed Voles to vegetation height and supplemental perches. Canadian Journal of Zoology 79:380–385. Smallwood, J.A. 1987. Sexual segregation by habitat in American Kestrels wintering in southcentral Florida: Vegetative structure and responses to differential prey availability. Condor 89:842–849. Smallwood, J.A. 1988. A mechanism of sexual segregation by habitat in American Kestrels (Falco sparverius) wintering in south-central Florida. Auk 105:36–46. Smallwood, J.A., and D.M. Bird. 2002. American Kestrel (Falco sparverius). In A. Poole (Ed.). The Birds of North America Online. Cornell Lab of Ornithology, Ithaca, NY. Available online at https://birdsna.org/Species–Account/bna/species/amekes. Accessed 13 January 2019. Smallwood, J.A., M.F. Causey, D.H. Mossop, J.R. Klucsarits, B. Robertson, S. Robertson, J. Mason, M.J. Maurer, R.J. Melvin, R.D. Dawson, G.R. Bortolotti, J.W. Parrish Jr., T.F. Breen, and K. Boyd. 2009. Why are American Kestrel (Falco sparverius) populations declining in North America? Evidence from nest-box programs. Journal of Raptor Research 43:274–282. Smith, W.C., and D. Han. 1981. Error rate for testing a contrast after a significant F test. Communications in Statistics-Simulation and Computation 6:545–556. Stinson, C.H., D.L. Crawford, and J.L. Lauthner. 1981. Sex differences in winter habitat of American Kestrels in Georgia. Journal of Field Ornithology 52:29–35. The Transportation Panning and Programming Division of the Texas Department of Transportation. 2016. TxDOT Roadways. Texas Natural Resources Information Systems. Available online at https://tnris.org/data-catalog/entry/txdot-roadways/. Accessed 1 April 2017. 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. US Climate Data. 2017. Climate Corpus Christi, Texas. Your weather service. Available online at http://www.usclimatedata.com/climate/corpus–christi/texas/united–states/ ustx0294. Accessed 13 January 2019. Worm, A.J., M.M. Bobowski, and T.S. Risch. 2013. Perch-type characteristics of overwintering Red–tailed Hawks (Buteo jamaicensis) and American Kestrels (Falco sparverius). Journal of the Arkansas Academy of Science 67:159–162. Yosef, R. 1994. Conservation commentary: Evaluation of the global decline in the true shrikes (Family Laniidae). Auk 111:228–233. Yosef, R. 1996. Loggerhead Shrike (Lanius ludovicianus), In A. Poole (Ed.). The Birds of North America Online. Cornell Lab of Ornithology, Ithaca, NY. Available online at https://birdsna.org/Species–Account/bna/species/logshr/introduction. Accessed 13 January 2019. Yosef, R., and T.C. Grubb Jr. 1993. Effect of vegetation height on hunting behavior and diet of Loggerhead Shrikes. Condor 95:127–131.