Southeastern Naturalist 299 D.B. McNair 22001166 SOUTHEASTERN NATURALIST 1V5o(2l.) :1259,9 N–3o1. 42 Population Status of the Eastern Phoebe in South-Central North Carolina: Breeding Increase at Water-Based Anthropogenic Sites Congruent with Breeding Bird Survey (BBS) and Christmas Bird Count (CBC) Data Douglas B. McNair* Abstract - Few studies in southeastern North America have compared local data sets to locally based results from 2 national surveys (North American breeding bird survey [BBS], Christmas bird count [CBC]). In 2012, I reexamined nest-site type use and nest type of Sayornis phoebe (Eastern Phoebe) at 109 water-based anthropogenic structures originally studied in south-central North Carolina in 1981. In 2012, Eastern Phoebes still strongly preferred breeding at small bridges with ledges, especially at the same structures where I studied them in 1981, even though use of other nest-site types slightly increased except at circular culverts, where no Eastern Phoebes nested. During the 31-year interval between studies, ~1/3rd of the bridges (20 of 62; 32%) were replaced with structures less favorable as nest-sites; thus, a lower number and proportion of small bridges with ledges were available in 2012. Although breeding Eastern Phoebes are still slowly increasing in abundance at water-based anthropogenic sites in south-central North Carolina, I project that this population will reach zero growth in 2027 as replacement of small bridges with ledges by other structures continues. The findings from this local water-based anthropogenic nest-site survey in south-central North Carolina was congruent with results from 2 national surveys (BBS, CBC) in a portion of the Pee Dee region in documenting a modest increase in the number of Eastern Phoebes during a time span ranging over 30 years. Introduction Nest-sites are the major limiting resource for Sayornis phoebe Latham (Eastern Phoebe, hereafter, Phoebe) during the breeding season when breeding-site fidelity is very high (Beheler et al. 2003, Weeks 2011). Phoebes prefer nest-site types located over or near water, and breeding directly over water at water-based anthropogenic sites (bridges, culverts) has been documented in many areas of central and eastern North America (Coffey 1963, 1976; Faanes 1980; Jackson and Weeks 1976; McWhorter 2001; Ware and Duncan 1989; Weeks 1979, 2011), including North Carolina in 1981 (McNair 1984). In south-central North Carolina, centered on Richmond County that borders South Carolina, Phoebes were uncommon in 1981, and in the Sandhills they only nested along woodland or forest streams at waterbased sites (McNair 1984). Since the turn of the 21st century, Phoebes have nested at some land-based anthropogenic sites in the Sandhills (McNair and Campbell 2013), but they still primarily breed at water-based sites. Phoebe use of water-based nest-site types has rarely been re-assessed within the same study area, and never in *35 Rowell Road, Wellfleet, MA 02667; dbmcnair@gmail.com. Manuscript Editor: Michael Steinberg Southeastern Naturalist D.B. McNair 2016 Vol. 15, No. 2 300 southeastern North America. In this study I compare nest-site type use, its availability, and nest type of Phoebes at water-based sites in Richmond County, NC, including adjacent areas of neighboring counties in North and South Carolina in 2012, thirty-one years after the original study. Extensive rehabilitation and replacement of ~30% of bridges and culverts occurred within my study area from 1981 to 2012 (NCDOT 2015, SCDOT 2015), including changes in design and construction materials. In the mid-Atlantic region, by the early 1970s, corrugated metal and concrete circular culverts were being employed to replace small bridges with steel I-beams that contain ledges (Whitaker 1974). Weeks (1984) stated that changing bridge and culvert design, especially replacement of small wooden or concrete bridges and square concrete culverts with unusable circular, corrugated metal or plastic culverts can significantly reduce available nest-sites for Phoebes in Indiana. The availability of small bridges with ledges, especially with steel I-beams that provide complete shelf support for statant nests, was by far the most important factor that determined if Eastern Phoebes nested at water-based sites in 1981 in south-central North Carolina (McNair 1984). Phoebes were much less numerous at the other nest-site types I documented: (1) box culverts (Phoebes uncommon), (2) small bridges without ledges (Phoebes scarce), and (3) large bridges/railroad trestles/circular culverts (Phoebes absent). In 1981, Phoebes reached the southeastern periphery of their breeding range along and near the Fall Line in south-central North Carolina; their breeding limit including adjacent South Carolina was still truncated along or near the Fall Line in the late 1980s and early 1990s (Cely 2003, McNair 1990, McNair and Post 1993, Post and Gauthreaux 1989). As documented by the North American breeding bird survey (BBS) of the US Fish and Wildlife Service, the Phoebe’s breeding-range front has slowly expanded from the Piedmont into the Coastal Plain (see summer distribution maps in Sauer et al. 2014). This range expansion has been accompanied by an increase in the relative abundance of Phoebes in most areas of the Piedmont and Coastal Plain including the Carolinas (see trend maps in Sauer et al. 2014); these trends were especially pronounced in North Carolina from 2003 through 2013. The relative abundance of Phoebes in the Coastal Plain of southeastern North Carolina and South Carolina remains low, at 0.05 to 1 bird per BBS route, rising to ~1–3 birds along the Fall Line in south-central North Carolina, and higher still at 3–10 birds in the Piedmont of North Carolina (Sauer et al. 2014). Phoebes are behaviorally plastic and adapted very early on to use of water-based anthropogenic nest-sites (Weeks 2011), which may be an ecological driver that has permitted them to expand their breeding-range front and increase in abundance in southeastern North America. Alternative explanations for an increase in their abundance, such as enhanced survivorship in North Carolina during winter over the past several decades because of global warming (LeGrand 2015), is also possible. Both the BBS and Christmas bird count (CBC) in south-central North Carolina and a portion of the Pee Dee region are expected to document a recent population increase of the Phoebe in comparison to historical data circa 1981. However, the BBS increase Southeastern Naturalist 301 D.B. McNair 2016 Vol. 15, No. 2 should be a closer fit to my population sample at water-based anthropogenic nestsites compared to the CBC because the BBS samples a similar population during the breeding season. If Phoebes have indeed increased in abundance in south-central North Carolina in recent decades (Sauer et al. 2014), and if nest-site type availability has remained unchanged at the same structure, especially at small bridges with ledges, I predicted that occupancy of water-based anthropogenic nest-site types Phoebes use would also have increased over the same period. If preferred nest-site type availability had decreased by 2012 because the structures surveyed in 1981 had been replaced by any of the other nest-site types in 2012, especially circular culverts, then breeding Phoebes would decrease at these different structures. The overall population change of Eastern Phoebes breeding in south-central North Carolina would then be determined by the relative contributions of the predicted increase at the same structures versus the predicted decrease at different structures in 2012. Methods Field-site description The study area comprised Richmond County, NC, and adjacent areas of 6 neighboring counties (North Carolina: eastern Anson, southern Montgomery, southwestern Moore, western Scotland; South Carolina: northeastern Chesterfield, northern Marlboro). The study area included the Piedmont, the Sandhills subregion of the Coastal Plain, and the Inner Coastal Plain (McNair 1984). Eastern Phoebe surveys Nest-site types. I examined 163 water-based anthropogenic sites in 1981 (Mc- Nair 1984). In late April through May of 2012, I re-examined 109 (67%) of these sites for which I had complete documentation, 104 sites in North Carolina (59 in Richmond County) and 5 sites in South Carolina (NCDOT 2015, SCDOT 2015). Of these sites, 68, 39, and 2, respectively, were located in the Piedmont, Sandhills, and Inner Coastal Plain (adjacent to the Fall Line). Of the 109 sites, 75 included the same water-based anthropogenic structure in 1981 and 2012 (= same structure), and 34 sites included different structures in 1981 and 2012. After 1981, the different structure was usually a replacement of the older structure, but occasionally represented extensive rehabilitation of the existing structure. I defined small bridges as those less than 6 m in height at their center above ground or water and with a span usually of less than 30 m; most small bridges were less than 5 m in height at their center (McNair 1984). I defined nest types based on the appearance of the nest (for an alternative definition, see Schukman et al. 2011). Breeding bird surveys. The roadside bias of the BBS, a large-scale, long-term survey (Pardieck et al. 2015, Sauer et al. 2014) should not underestimate numbers of Phoebes in south-central North Carolina during the breeding season because nesting Phoebes are closely associated with anthropogenic structures along or near roads. Twelve routes for the BBS recorded at least 1 Phoebe within or near a portion of the Pee Dee region of the Carolinas (NC: 8 routes; SC: 4 routes). All routes were Southeastern Naturalist D.B. McNair 2016 Vol. 15, No. 2 302 in predominantly rural areas and included 3 physiographic provinces, the Piedmont, Sandhills, and Inner Coastal Plain, with some overlap between physiographic provinces on one route (see results). Christmas bird counts. The CBC is another large-scale, long-term survey, but is primarily recreational (Carolina Bird Club 2015, National Audubon Society 2015), in contrast to the BBS that was explicitly created to detect population trends (Dunn et al. 2005). Five CBCs have been conducted within or near the Pee Dee region of North and South Carolina. They were carried out at Pee Dee National Wildlife Refuge, Southern Pines, and Cumberland County (Fayetteville) NC; and at Carolina Sandhills National Wildlife Refuge and the Pee Dee area (Florence) in SC. All 3 physiographic provinces are represented. Count circles within Cumberland County and the Pee Dee area include metropolitan statistical areas, defined by the US Office of Management and Budget as a central urban area or cluster; the Pee Dee area count circle is ~20% urban, 20% semi-rural, and 60% rural (S. Smolen-Morton, Francis Marion University, Florence, SC, SSmolenMorton@fmarion.edu, pers. comm.). The Southern Pines count circle includes a portion of the Pinehurst-Southern Pines micropolitan statistical area, defined as a smaller or less-dense central urban area or cluster; this count circle is ~10% urban and 90% rural (S. Campbell, susan@ncaves.com, 910-949-3207, pers. comm.). The remaining 2 count circles, Pee Dee NWR and Carolina Sandhills NWR, are almost entirely (~95%) rural areas, the remainder (~5%) being urban or semi-rural (N. Jordan, US Fish and Wildlife Service, nancy_jordan@ fws.gov, 843-335-6026, pers. comm.; D.B. McNair, pers. observ.). Data analysis Nest-site types. I divided the 109 water-based anthropogenic sites into 6 nestsite types (small bridge with ledges, small bridge without ledges, large bridge, box culvert, circular culvert, railroad trestle), grouped by whether Phoebes nested at the same or a different structure after 1981. For all nest-site types combined, I performed 2 x 2 chi-square tests with Yate’s correction to examine whether these structures with or without active nests were proportionally more or less numerous in either year (1981, 2012). I also examined whether each of the 6 nest-site types with or without nests at the same structure in both years or at a different structure in 2012 were proportionally more or less numerous. I then performed 1 x 1 chi-square tests with Yate’s correction to examine 4 outcomes: with and without nests at the same structure in 1981 and 2012 (75 sites) and at different structures (34 sites); the 4 outcomes were reduced for both chi-square tests to 2 categories—unchanged and changed outcomes. For nest types, I performed a 2 x 2 chi-square test with Yate’s correction to examine if statant nests compared to all other nest types combined were proportionally more or less numerous at the same structure in both years or at a different structure in 2012. Otherwise, I present descriptive statistics. Breeding bird surveys. I retrieved raw data collected by the BBS (Pardieck et al. 2015) to calculate the relative abundance of Phoebes on individual routes that were surveyed by 1–4 observers per route. The small bias of observer effects Southeastern Naturalist 303 D.B. McNair 2016 Vol. 15, No. 2 (5% underestimation of counts in the first year; Link et al. 2008) is expected to be negligible for Phoebes. This species is generally solitary, easily detected by sight and sound (not ventriloquial), and no more than 2 phoebes would usually be expected per stop; during their first survey, any observer should easily distinguish the 2 individuals. I calculated results using arithmetic and geometric means, which were similar. I chose the arithmetic mean because results were slightly more conservative (even though the use of the geometric mean for proportional and ratio data which are used in this study can have advantages). Only 2 routes had data from before the 1990s and preceding my first survey of nesting Eastern Phoebes in south-central North Carolina in 1981. Therefore, for all routes I compared the mean count of birds for 2 non-exclusive year-groups, the first group for all years of data available on each route versus the second group of only 2009– 2014 (3 years before-and-after I collected data in 2012). For all years of data, I also calculated Pearson’s product correlation coefficients (r) to examine which routes had a significant increase in the number of Phoebes. I next calculated the difference in mean counts between the 2 year-groups and rounded to the nearest 0.5 bird to examine how much Eastern Phoebes had increased or decreased on each route. I used this conservative approach between the 2 year-groups (which are not independent) to reduce the contrast in mean counts so any consistency in results would be more robust, given potentially confounding factors such as differences between physiographic provinces, the span of years, the number of years run, and observer ability within and between the routes (Link et al. 2008). Finally, I calculated the ratio (proportional change) of the mean count for 2009–2014 divided by the mean count of all years of data for each BBS route to compare to ratios from my nest-site data and CBC results. Christmas bird counts. The Phoebe is generally a solitary species that is easily located and identified, and rarely occurs at feeders (see Dunn et al. 2005). In my study area, Phoebes are more widely distributed in winter than during the breeding season; thus, CBC data should be useful for the study of this species in the Carolinas (McNair 1987; cf., Arenaria interpres L. [Ruddy Turnstone] in California, Pandolfino and Helmericks 2005; Calidris alpina L. [Dunlin] in North America, Xu et al. 2015). The number of birds counted on CBCs can be confounded by variation in effort (Butcher and McCulloch 1990, Link et al. 2008). Large CBC data sets have generally followed an over-dispersed Poisson distribution with complex adjustments made for variation in effort (Dunn et al. 2005, Link and Sauer 2007, Link et al. 2008, Sauer and Link 2002; though see Xu et al. 2015). For my small CBC data set, I selected 3 winters before and after each of the 2 years I sampled breeding Phoebes for a total of 6 years for each year-group (historical: 1978–1979 to 1983– 1984; recent: 2009–2010 to 2014–2015) to ensure an adequate and representative sample for each CBC. The total number of birds and party-hours for each year from all CBCs (n = 44), as well as CBCs partitioned into historical (n = 17) and recent (n = 27) year-groups, each approximated a normal distribution at an alpha value of 0.05 (six Shapiro-Wilk tests: W = 0.93–0.95 [range], P = 0.06–0.41 [range]), which I confirmed by examination of Q-Q plots. Pearson’s product correlation coefficient Southeastern Naturalist D.B. McNair 2016 Vol. 15, No. 2 304 between the total number of birds and party-hours for each year from all CBCs was significant (r = 0.37, P = 0.01), indicating that additional effort influenced counts and that the relationship was approximately linear. The relationship between the number of birds and party-hours remained approximately linear for the recent yeargroup (r = 0.40, P = 0.04), but was non-linear for the historical year-group (r = 0.01, P = 0.96) even though the mean number of party-hours from all counts within each year-group was similar (historical year-group: 42.01; recent year-group: 43.65), unlike most CBCs where the mean count effort has increased over a long time interval (Butcher and McCulloch 1990). This disparity between the 2 year-groups indicates that observers had improved in their ability to efficiently detect Phoebes on recent CBCs. Dunn et al. (2005) cautioned that the relationship between the number of birds and effort can vary among CBC circles. Examination of CBC circles in the historical year-group revealed that 1 (Carolina Sandhills NWR) approximated a linear relationship whereas the other 2 did not. Nonetheless, I did not make any additional effort adjustment to the number of party-hours for the 2 historical CBCs that were non-linear, given the small sample size and linear distribution of the other historical and all recent CBCs, but this shortcoming should be recognized. Consequently, I calculated the number of Phoebes per party-hour, a simple index of relative abundance, for each year of the 2 year-groups for each of the 5 CBCs. I also compared these results after converting counts to a common scale corresponding to the mean level of effort for each of the 2 year-groups, a simple additional effort adjustment (Blem 1995, Butcher and McCulloch 1990, Link et al. 2008). The mean effort between the 2 year-groups was very similar (see above), effort on counts was generally low (maximum of 86 party-hours on 1 count), and the range of effort was low on all 5 count circles, which should reduce concerns about the scaling effects of effort when effort is high (Link et al. 2008, Sauer and Link 2002). Regardless, results from this comparison using unadjusted versus scaled party-hours were virtually identical, so I chose the simpler method of using unadjusted party-hours. After also examining the distribution of the number of birds and party-hours for each year on each CBC, I rejected 3 outlier values on 2 count circles from the recent year-group that were associated with an abnormally low number of party-hours which distorted results. I then calculated the mean number of birds per party-hour over all years within each year-group for each CBC, which reduced annual variability in counts and count effort, including any potential weather effects (Link and Sauer 2007, Link et al. 2008). I also converted these counts to represent an 8-h party-hour/day and then calculated the difference between the mean counts of the 2 year-groups, rounded off to the nearest 0.5 bird, for each CBC. Finally, I calculated the ratio (proportional change) of the mean number of birds per party-hour between the recent year-group divided by the mean number of the historical year-group for each CBC. Using only 5 count circles and the availability of only incomplete data for several of them weaken this examination of CBC data, but consistent results here may strengthen interpretation of results from the other 2 data sources (mine at nest-sites, BBS). Southeastern Naturalist 305 D.B. McNair 2016 Vol. 15, No. 2 Results Nest-site types Of the 218 structures representing 6 nest-site types used over both years (1981, 2012), 124 (57%) were small bridges with ledges which contained 80 of the 97 active Phoebe nests (82%) at all structures (Table 1), much greater occupancy than at any of the other 5 nest-site types. In 1981, 45 structures contained active nests and 64 structures did not, whereas in 2012, 52 structures contained active nests and 53 did not; the difference in proportion of nests between the 2 years for all nestsite types combined was not significant (χ2 = 1.15, df = 1, P > 0.2). In contrast, the difference in the proportion of nests at the same structure in both years compared to different structures in 2012 was highly significant (χ2 = 15.4, df = 1, P < 0.001). All 6 nest-site types at a different structure in 2012 contained proportionally fewer nests, including small bridges with ledges (Table 1). There were 62 individual small bridges with ledges, but only 8 of these were built after 1981 (13%). Only 2 (3%) small bridges with ledges were extensively rehabilitated or replaced with another small bridge with ledges after 1981, whereas 20 (32%) small bridges with ledges were replaced by other nest-site types by 2012. Structural materials that included steel alone or in combination with other materials (wood, concrete, stone) comprised 60 of the 62 small bridges with ledges (97%); steel and wood was the most common combination. In contrast, there were 36 individual small bridges without ledges, and 12 of these were built after 1981 (33%), a much higher number and proportion than small bridges with ledges. Structural materials of small bridges without ledges that included steel alone or in combination with other materials (concrete, wood), comprised only 5 of 36 (14%) of these bridges and none contained steel beams. Table 1. Eastern Phoebe water-based anthropogenic nest-site types with and without nests at the same structure in both years (1981, 2012; n = 75) or at a different structure in 2012 (n = 34) in south-central North Carolina. Total = totals for each type over both years (i.e., each structure that was available in both years counts twice). All structures Same structure Different structure With Without With Without With Without Nest-site type Total nest nest nest nest nest nest Small bridge with ledgesA 124 80 41 67 28 13 13 Small bridge without ledges 51 9 42 7 23 2 19 Large bridgeB 19 2 16 2 11 0 5 Box culvert 13 5 8 3 3 2 5 Circular culvert 8 0 8 0 0 0 8 Railroad trestle 2 1 1 1 1 0 0 No structureC 1 0 1 0 0 0 1 Total 218 97 117 80 66 17 51 AThree outcomes in 2012 were undetermined. BOne outcome in 2012 was undetermined. CBox culvert along seasonal feeder rivulet into Hitchcock Creek was removed after 1981, road bed was built up, and area was largely filled in, supporting only a very short, small square concrete cap that allows water to pass through. Southeastern Naturalist D.B. McNair 2016 Vol. 15, No. 2 306 Small bridges with ledges comprised 49 of the 75 (65%) sites with the same structure in both years (Table 2). I excluded 4 sites that had undetermined outcomes in 2012; 49 of 71 (69%) outcomes at the same structure remained unchanged (Table 2), whereas 22 outcomes (31%) changed (χ2 = 9.52, df = 1, P < 0.005). Of the 22 outcomes that changed, 18 structures (82%) without a nest in 1981 contained a nest in 2012, representing 5 nest-site types, especially small bridges with ledges (Table 2). In contrast, only 4 of 22 (18%) changes were of structures that contained a nest in 1981 but were without one in 2012. The net-gain differential was greatest for small bridges with ledges (Table 2). Thus, 32 of 71 sites (45%) contained nests in 1981, whereas 46 of 71 sites (65%) contained nests in 2012. This net gain of 14 nests at the same structures between the 2 years over a time interval of 31 years represents a net percentage gain of 20% (or 0.65%/year). Small bridges with ledges were present in only 4 of the 34 (12%) sites with a different structure in both years (Table 3). Twenty-one (21) outcomes at 34 sites (62%) with a different structure remained unchanged (Table 3), whereas 13 outcomes (38%) changed (χ2 = 1.44, df = 1, P > 0.2). Of the 13 outcomes that changed, 4 structures (31%) without a nest in 1981 contained a nest in 2012. In contrast, 9 of 13 (69%) changed outcomes represented structures without a nest in 2012 that contained a nest in 1981 (Table 3). Thus, 11 of 34 sites (32%) contained nests in 1981, whereas 6 of 34 sites (18%) contained nests in 2012. This net loss of 5 nests at different structures between the 2 years over a time interval of 31 years represents a net percentage loss of 14% (or 0.45%/year). Therefore, adding together the net difference for sites that changed structures and those that remained the same, Phoebes in south-central North Carolina centered on Richmond County had a net gain of 9 nests at all sites between the 2 years over a time interval of 31 years, which represents a net percentage gain at all water-based anthropogenic nest-sites of 6% (or 0.2%/year). The ratio (proportional change) of the number of all nest-sites with nests in 2012 divided by the number of nests in 1981 is 1.21, and for only sites Table 2. Four outcomes, with and without nests in 1981 and 2012, at 75 sites with the same structure for 5 water-based anthropogenic nest-site types in south-central North Carolina. Outcome No nest With nest in in 1981, 1981, and Total Nests in No nests and with nest no nest Nest-site type sites both years in both years in 2012 in 2012 Small bridge with ledgesA 49 26 7 10 3 Small bridge without ledges 15 1 9 4 1 Large bridgeB 7 4 2 Box culvert 3 1 1 1 0 Railroad trestle 1 0 0 1 0 Total 75 28 21 18 4 ATwo small bridges with ledges contained nests in 1981, and one did not, but all three outcomes in 2012 were undetermined. BOne large bridge without a nest in 1981 had an undetermined outcome in 2012. Southeastern Naturalist 307 D.B. McNair 2016 Vol. 15, No. 2 with the same structures in both years, a better measure of unhindered population increase of Phoebes, is 1.44. Nest types Only statant nests were present at 79 of 93 structures (85%). The remaining nest types were 2 different types in the same year, statant and adherent nests (plastered to a rough, irregular vertical surface) on the same structure (n = 2), semi-statant (incomplete shelf support; n = 2), adherent (n = 7), and undetermined (n = 3). The proportion of statant nests compared to all other nest types combined at the same structure in both years or at a different structure in 2012 was similar (χ2 = 0.88, df = 1, P > 0.2). All statant nests were on ledges, whereas support for 11 non-statant nests was Crabronidae (organ-pipe mud-dauber wasp) nests (n = 6), single screw-bolts (n = 3), 1 weathered protuberance from an old outer pillar of a railroad trestle, and 1 on top of a foundation of a Hirundo rustica L. (Barn Swallow) nest that was adherent. Breeding bird surveys Phoebes significantly increased over time spans ranging from 17 to 46 years on 4 of 12 BBS routes, and there was a difference in the mean count between the 2 year-groups (all years, 2009–2014) of at least an increase of 1 bird for these 4 routes (Table 4). Phoebes on Biscoe and the other 7 routes with smaller numbers did not increase. Two counts showed an increase of 0.5 bird and there were no differences in relative abundance on the other 6 routes (Table 4). The highest mean counts for both year-groups were on 2 of the 3 Piedmont routes, where the greatest increase Table 3. Four outcomes, with and without nests in 1981 and 2012 at 34 sites with a different structure (of a same or different type as was used at that site in 1981) in 2012 for four 1981 water-based anthropogenic nest-site types in south-central North Carolina. Outcome No nests No nest in With nest in 1981 nest-site type Total Nests in in 1981, and with 1981, and 2012 nest-site type sites both years both years nest in 2012 no nest in 2012 Small bridge with ledges small bridge with ledges 2 1 1 small bridge without ledges 9 1 4 4 large bridge 1 1 box culvert 3 1 2 circular culvert 7 2 5 Small bridge without ledges small bridge with ledges 2 2 small bridge without ledges 3 2 1 box culvert 3 3 circular culvert 1 1 Large bridge large bridge 2 2 Box culvert no structure 1 1 Total 34 2 19 4 9 Southeastern Naturalist D.B. McNair 2016 Vol. 15, No. 2 308 also occurred. The general pattern within and between year-groups was for mean counts and differences in mean counts highest in the Piedmont, followed by the Sandhills, and lowest in the Inner Coastal Plain, where the total number of birds on each of the 4 routes was less than 10, regardless of differences in the span of years, the number of years run, and observer ability. Nine of 12 ratios (proportional change) of the mean count for 2009–2014 divided by the mean count for all years on BBS routes ranged from 0.96 to 1.73 (Table 4). The other 3 routes represented outlier values (0 on 2 routes and 1 ratio of 2.63 on the Sandhills route), all 3 of which had a total number of birds less than 10 each. The largest valid ratio of 1.73 had the longest span of years. Christmas bird counts The mean number of Phoebes per party-hour in the recent year-group (2009– 2010 to 2014–2015) was 2–3x higher on the single CBC circle in the Piedmont (0.98) than any of the very similar results on the 4 count circles in the Sandhills or Inner Coastal Plain (range = 0.36–0.43) within or near a portion of the Pee Dee region of North and South Carolina (Table 5). The lowest mean number of Phoebes per party-hour in the historic year-group (1978–1979 to 1983–1984) was in the Sandhills (range = 0.09–0.15) compared to the Inner Coastal Plain (0.34), although historic data were not collected on 2 of the 5 CBCs. When converted to Table 4. The total number of birds, mean count for 2 groups of years (all years, 2009–2014), difference in mean count between the 2 groups of years, and the ratio of the mean count between both groups of years for the Eastern Phoebe on 12 Breeding Bird Survey (BBS) routes within or near the Pee Dee region of North and South Carolina. Physiographic province (Prov.): P = Piedmont, S = Sandhills, and I = Inner Coastal Plain. Probability for Pearson’s correlation coefficient: * = 0.01–0.05; ** = less than 0.01. Differences in mean count (Diff.) are rounded to the nearest 0.5 bird. Ratio = Ratio 2009–2014/all years. Mean count BSS route Years Total All 2009– Prov. Name number State Years run birds years 2014 Diff. Ratio P Wilgrove 63216 NC 1995–2012 16 94 5.88* 8.25A 2.5 1.40 P Oakboro 63217 NC 1969–2014 19 106 5.58* 9.67 4.0 1.73 P Biscoe 63215 NC 1990–2011 21 55 2.62 2.67B 0.0 1.02 S Lake Surf 63314 NC 1998–2014 13 42 3.23* 5.33 2.0 1.65 S Hamlet 63207 NC 1995–2014 19 33 1.74** 2.83 1.0 1.63 S Raeford 63900 NC 1990–2014 21 28 1.33 1.50 0.0 1.13 S Sandhills 80900 SC 1994–2014 21 8 0.38 1.00 0.5 2.63 S/I Mount Pisgah 80053 SC 1999–2010 7 11 1.57 1.50C 0.0 0.96 I Bethel 63315 NC 2002–2013 9 4 0.44 0.75D 0.5 1.70 I Rowland 63106 NC 1990–2013 22 3 0.14 0.00E 0.0 0.00 I Bennettsville 80054 SC 1999–2010 10 9 0.90 1.00C 0.0 1.11 I Dillon 80010 SC 1970–2010 28 1 0.04 0.00C 0.0 0.00 AOnly 2009–2012. BOnly 2009–2011. COnly 2009–2010. DOnly 2010–2013. EOnly 2009–2013. Southeastern Naturalist 309 D.B. McNair 2016 Vol. 15, No. 2 an 8-h party-hour/day, the difference in mean counts between the 2 year-groups was greatest in the Sandhills (2–2.5 birds) compared to the Inner Coastal Plain (0.5 bird). Phoebes increased on all 3 CBCs with data for both year groups, where the ratio (proportional change) of the mean number of Phoebes per party-hour between recent and historic year-groups was greatest in the Sandhills (2.6, 4.33) compared to the Inner Coastal Plain (1.26). Discussion As expected, small bridges with ledges that contained statant nests remained the preferred nest-site type and nest type of Phoebes in forested habitats centered on Richmond County, NC, as is true for the mountains of Virginia (Clapp 1993). Adherent nests are still rather scarce in south-central North Carolina (McNair 1984, this study), unlike western Alabama and the northern Piedmont of Virginia where many nests under bridges were attached to mud-dauber wasp nests or old Barn Swallow nests (Clapp 1993, Jackson and Weeks 1976). Nonetheless, Phoebes have again demonstrated their adaptability (Weeks 2011) in south-central North Carolina by using at least some alternative water-based anthropogenic nest-site types, such as larger concrete-box culverts. However, replacement of small bridges with ledges by less-suitable or entirely unsuitable alternative nestsite types, especially circular culverts, is inappropriate for the conservation of Phoebes unless improvements such as inverted-T wood structures that support Table 5. The total number of birds, mean number per party-hour, and the ratio of the mean number between 2 groups of 6 years each (1978–1983 and 2009–2014) for the Eastern Phoebe on 5 Christmas Bird Counts (CBC) within or near the Pee Dee region of North and South Carolina. nd = no data (counts not conducted) and na = not applicable. Ratio = Ratio 2009–2014/1978–1983. Physiographic Total Mean number/ CBC province birds party-hour Ratio Pee Dee NWR, NC Piedmont na 1978–1983 nd nd 2009–2014A 100 0.98 Southern Pines, NC Sandhills 4.33 1978–1983 28 0.09 2009–2014 174 0.39 Carolina Sandhills NWR, SC Sandhills 2.60 1979–1983B 15 0.15 2009–2014C 52 0.39 Cumberland County [Fayetteville], NC Sandhills/Inner Coastal Plain na 1978–1983 nd nd 2009–2014 85 0.36 Pee Dee Area [Florence], SC Inner Coastal Plain 1.26 1978–1983 58 0.34 2009–2014 114 0.43 AOne outlier value removed (2010–2011 winter). BOne count not conducted (1978–1979 winter). CTwo outlier values removed (2011–2012 and 2014–2015 winters). Southeastern Naturalist D.B. McNair 2016 Vol. 15, No. 2 310 statant nests are installed within larger circular culverts with a diameter of at least 1.2 m (Weeks 1984, Whitaker 1974). The decks, superstructures, and usually substructures of most small bridges are now primarily built of reinforced concrete with no ledges, i.e., slab bridges (McNair 1990, Weeks 1984; this study). Similar structural improvements for nesting Phoebes could also be installed within slab bridges and concrete-box culverts. North Carolina has a backlog of ~40% of bridges and culverts that require replacement or extensive rehabilitation due to inadequate funding (NCDOT 2015), although the proportion of deficient bridges in Richmond County (17%) is much lower than the statewide average. Construction and installation costs for structures that support nesting Phoebes would add to the funding deficit for replacement bridges and culverts in North Carolina, a problem reported from many other states such as Iowa (TRIP 2015). If the projected decline of Phoebes becomes too sharp at some time in the future (see below), the NCDOT should revisit the issue of providing artificial platforms in structures that are otherwise unsuitable for nesting Phoebes (and several other species that nest at water-based anthropogenic sites; cf., Indiana; Weeks 1984). Even though Phoebes are highly susceptible to severe winter weather, such as the region experienced in 1976–1977 (Robbins et al. 1986), breeding Phoebes in south-central North Carolina have slowly increased in abundance (this study), a trend which has not yet been affected by a sharp reduction in small bridges with ledges over the 31-y time interval (1981–2012). At the 109 water-based anthropogenic sites within the study area, 34 changes (31%) to a different structure over 31 years represents ~1 replaced or extensively rehabilitated structure every year. As such, assuming the number of Phoebes breeding at these water-based anthropogenic sites is a reliable population sample, my projections suggest that in south-central North Carolina they may reach zero population growth in 2027. Based on my analyses, by 2027, the number of sites with the same structures will decline to 56 and those with different structures will increase to 48 if conditions otherwise remain unchanged. Breeding Phoebes have been scarce in the extreme eastern Piedmont at land-based anthropogenic sites within the study area (cf. Coffey 1963; D. McNair, unpubl. data) and have only had a limited recent spread to land-based anthropogenic sites in the Sandhills, including Richmond County (McNair and Campbell 2013). Despite this scarcity, the recent occupancy of Phoebes at land-based anthropogenic sites suggests that saturation of suitable water-based nest-sites may be occurring even though Phoebes are still relatively uncommon in south-central North Carolina compared to many other areas of the central and eastern US such as Appalachia (Hill and Gates 1988, Sauer et al. 2014, Weeks 2011). This local-scale study in south-central North Carolina where water-based anthropogenic nest-sites form a preponderance of all nest-sites can help inform solutions for Phoebe breeding populations stressed or potentially stressed by a sharp reduction of suitable nest-sites in other areas of North America. Counts made by the BBS and the CBC surveys in the Pee Dee region likely under-represent bird-population numbers because of low route densities and low number of count circles. The results of these surveys are also weakened by missing Southeastern Naturalist 311 D.B. McNair 2016 Vol. 15, No. 2 data for some years, inconsistent route coverage because of difficulty of observer recruitment, and inconsistencies in effort efficiency on several count circles. Sauer et al. (2013) also flagged very small samples of BBS routes (less than 14), as in this study, as a cause of possibly unreliable data. Nonetheless, despite differences between some routes and count circles in their degree of urbanization, increases in relative abundance of Phoebes on half of the BBS routes, with no decreases, and increases on all 3 CBC circles is likely not due to chance alone but represent a true increase in Phoebes. The local increase of breeding Phoebes at water-based anthropogenic nest-sites over a span of 31 years in south-central North Carolina is one factor that is responsible for increases documented on BBS and CBC data sets from a portion of the Pee Dee region used herein. Peterjohn and Sauer (1995) documented increasing complexity in the temporal patterns of population trends of Progne subis L. (Purple Martin) as their analyses focused on national, to regional, and then to state levels (and Bird Conservation Regions within states; Sauer et al. 2013). They did not examine whether many population increases or declines were locally distributed (e.g., of a small region or area within a state), although Peterjohn and Sauer (1995) stated that Purple Martins exhibited a wide variety of temporal patterns in their trends in the southern portion of their breeding range. The study reported here is the first in the Carolinas and one of the few in southeastern North America, that documents that a local independent data set that measured the temporal pattern of a population increase is congruent with BBS and CBC data even though my analyses of the latter 2 data sets are primitive (cf., Link and Sauer 2007, Link et al. 2008, Sauer et al. 2013, Wilson et al. 2013). This robust result strongly suggests that additional analyses should be performed on a variety of permanent resident and temperate migrant species in North America, using other local data sets in conjunction with BBS and CBC data (Wetzel and Krupa 2013), to determine local factors responsible for temporal population changes in the Carolinas or elsewhere in southeastern North America. The 2 national surveys, using data only from a portion of the Pee Dee region, were also consistent in their general documentation of larger numbers of Phoebes in the Piedmont compared to the Sandhills and Inner Coastal Plain, which is in agreement with BBS results on the relative abundance of Phoebes in these physiographic provinces along the Atlantic slope of southeastern North America (Sauer et al. 2014; see Introduction, this study). The mean counts in the Inner Coastal Plain compared to the Sandhills were lower in the BBS surveys but comparable in the CBCs, especially recently, which suggests proportionally more Phoebes are temperate migrants wintering in the Inner Coastal Plain compared to a higher proportion of permanent residents in the Sandhills. In addition, ignoring 1 BBS route with a very low number of birds, BBS ratios of mean counts that measure proportional change in all 3 physiographic provinces aligned themselves more closely with my nest-site ratios, compared to much higher ratios of mean counts in the Sandhills on 2 CBCs. These results suggest that BBS data fit my data more closely than CBC data, as expected, but differences in methodology between the 2 national surveys and my local survey and differences in the Southeastern Naturalist D.B. McNair 2016 Vol. 15, No. 2 312 time span of my nest-site data and most BBS routes prevent a more precise analysis. Firmer evidence is needed from marked birds to authoritatively determine the precise population status of Phoebes in each of the 3 physiographic provinces in the Pee Dee region of the Carolinas. Acknowledgments BBS and CBC data were provided by, respectively, the Patuxent Wildlife Research Center of the US Fish and Wildlife Service and the National Audubon Society and through the generous efforts of volunteers in the Carolinas. I thank B. Bockhahn, S. Campbell, N. Jordan, S. Smolen-Morton, and J. Walker, and R.J. Davis, respectively, who supplied me additional information on their CBC circles and on 4 BBS routes in North Carolina; D.B. Cook, R.L. Floyd, and J.M. Tucker at the SCDOT who supplied me with bridge information (Pee Dee Region bridges-responsedoc.xls) through a Freedom of Information Act request; J.C. Kilgo, M.T. Stanback, and H. Weeks Jr. for their reviews of a penultimate version of the manuscript; and 2 anonymous individuals for their reviews of the submitted version of the manuscript. Literature Cited Beheler, A.S., O.E. Rhodes, and H.P. Weeks. 2003. Breeding site and mate fidelity in Eastern Phoebes (Sayornis phoebe) in Indiana. Auk 120:990–999. Blem, C.R. 1995. Winter abundance of vultures in Virginia 1965–1993. Raven 66:83–86. Butcher, G.S., and C.E. McCulloch. 1990. Influence of observer effort on the number of individual birds recorded on Christmas bird counts. Pp. 120–129, In J.R. Sauer and S. Droege (Eds.). Survey designs and statistical methods for the estimation of avian population trends. Biological Report 90(1). US Fish and Wildlife Service, Laurel, MD. 166 pp. Carolina Bird Club. 2015. Christmas bird counts. Available online at http://www.carolinabirdclub. org/christmas. Accessed October 2015. Cely, J.E. 2003. The South Carolina breeding bird atlas 1988–1995. South Carolina Department of Natural Resources, Columbia, SC. 305 pp. Clapp, R.B. 1993. Eastern Phoebes lay eggs in more than one nest-cup. Raven 64:84–87. Coffey, B.B., Jr. 1976. The Eastern Phoebe as a summer resident in Mississippi. Mississippi Kite 6:38–41. Coffey, J.W. 1963. A nesting study of the Eastern Phoebe. Migrant 34:41–49. Dunn, E.H., C.M. Francis, P.J. Blancher, S.R. Drennan, M.A. Howe, D. Lepage, C.S. Robbins, K.V. Rosenberg, J.R. Sauer, and K.G. Smith. 2005. Enhancing the scientific value of the Christmas bird count. Auk 122:338–346. Faanes, C.A. 1980. Breeding biology of Eastern Phoebes in northern Wisconsin. Wilson Bulletin 92:107–110. Hill, S.R., and J.E. Gates. 1988. Nesting ecology and microhabitat of the Eastern Phoebe in the central Appalachians. American Midland Naturalist 120:313–324. Jackson, J.A., and R.E. Weeks. 1976. Nesting of the Eastern Phoebe and Barn Swallow in western Alabama. Alabama Birdlife 24:7–9. LeGrand, H.E., Jr., J. Haire, A. Iyoob, and T. Howard. 2015. Birds of North Carolina: Their distribution and abundance. Eastern Phoebe. Available online at http://www.carolinabirdclub. org/ncbirds/view.php. Accessed October 2015. Southeastern Naturalist 313 D.B. McNair 2016 Vol. 15, No. 2 Link, W.A., and J.R. Sauer. 2007. Seasonal components of avian population change: Joint analysis of 2 large-scale monitoring programs. Ecology 88:49–55. Link, W.A., J.R. Sauer, and D.K. Niven. 2008. Combining breeding bird survey and Christmas bird-count data to evaluate seasonal components of population change in Northern Bobwhite. Journal of Wildlife Management 72:44–51. McNair, D.B. 1984. Nest placement of the Eastern Phoebe under bridges in south-central North Carolina. Oriole 49:1–6. McNair, D.B. 1987. Status and distribution of the Fish Crow in the Carolinas and Georgia. Oriole 52:28–45. McNair, D.B. 1990. Eastern Phoebe breeds in the northeast upper coastal plain of South Carolina. Chat 54:59–61. McNair, D.B., and S. Campbell. 2013. Eastern Phoebes (Sayornis phoebe) breed at landbased anthropogenic sites in the North Carolina Sandhills. Chat 77:69–73. McNair, D.B., and W. Post. 1993. Supplement to status and distribution of South Carolina birds. Charleston Museum Ornithological Contribution No. 8. Charleston, SC. 49 pp. McWhorter, V. 2001. Eastern Phoebe: First documented nesting at the Okatibbee Wildlife Management Area, Lauderdale County, Mississippi. Mississippi Kite 31(2):39–40. National Audubon Society. 2015. Christmas bird counts. Available online at http://www. audubon.org/conservation/science/christmas-bird-count. Accessed October 2015. North Carolina Department of Transportation (NCDOT). 2015. North Carolina bridge information: Statewidebridges.xls. Available online at http://www.ncdot.gov/projects/ ncbridges/default.html. Accessed June 2012 and October 2015. Pandolfino, E.R., and J.W. Helmericks. 2005. Changes in winter abundance of the Ruddy Turnstone along the coast of California. Western Birds 36:121–130. Pardieck, K.L., D.J. Ziolkowski Jr., and M.-A.R. Hudson. 2015. North American breeding bird survey dataset 1966–2014. Version 2014.0. US Geological Survey, Patuxent Wildlife Research Center. Available online at http://www.pwrc.usgs.gov/BBS/RawData/. Accessed October 2015. Peterjohn, B.G., and J.R. Sauer. 1995. Purple Martin population trends from the North American breeding bird survey, 1966–1994. Purple Martin Update 6(2):2–8. Post, W., and S.A. Gauthreaux Jr. 1989. Birds of South Carolina. Charleston Museum, Charleston, SC. 83 pp. Robbins, C.S., D. Bystrak, and P.H. Geissler. 1986. The breeding bird survey: Its first fifteen years, 1965–1979. Resource publication 157. US Fish and Wildlife Service, Washington, DC. 196 pp. Sauer, J.R., and W.A. Link. 2002. Using Christmas bird count data in analysis of population changes. American Birds 56:10–14. Sauer, J.R., W.A. Link, J.E. Fallon, K.L. Pardieck, and D.J. Ziolkowski Jr. 2013. The North American breeding bird survey 1966–2011: Summary analysis and species accounts. North American Fauna 79:1–32. Sauer, J.R., J.E. Hines, J.E. Fallon, K.L. Pardieck, D.J. Ziolkowski Jr., and W.A. Link. 2014. The North American breeding bird survey: Results and analysis 1966–2013. Version 01.30.2015. US Geological Survey Patuxent Wildlife Research Center, Laurel, MD. Available online at http://www.mbr-pwrc.usgs.gov/bbs/bbs.html. Accessed October 2015. Schukman, J.M., A. Lira-Noriega, and A.T. Peterson. 2011. Multiscalar ecological characterization of Say’s and Eastern Phoebes and their zone of contact in the Great Plains. Condor 113:372–384. Southeastern Naturalist D.B. McNair 2016 Vol. 15, No. 2 314 South Carolina Department of Transportation (SCDOT). 2015. Available online at http:// www.dot.state.sc.us. Accessed June 2012 and October 2015. TRIP. 2015. Iowa’s top transportation challenges: Meeting the state’s need for safe and efficient mobility. Available online at tripnet.org. Accessed October 2015. Ware, D.M., and R.A. Duncan. 1989. First record of the Eastern Phoebe nesting in Florida. Florida Field Naturalist 17:22. Weeks, H.P., Jr. 1979. Nesting ecology of the Eastern Phoebe in southern Indiana. Wilson Bulletin 91:441–454. Weeks, H.P., Jr. 1984. Importance and management of riparian bridges and culverts for nesting passerines. Pp. 163–175, In W.C. McComb (Ed.). Proceedings of a workshop on management of nongame species and ecological communities. University of Kentucky, Department of Forestry, Lexington, KY. 404 pp. Weeks, H.P., Jr. 2011. Eastern Phoebe (Sayornis phoebe). No. 94, In A.A. Poole (Ed.). The Birds of North America online. Revised 26 April 2011. Cornell Lab of Ornithology, Ithaca, NY. Available online at http://bna.birds.cornell.edu/bna/species/094/biblio. Accessed June 2012 and October 2015. Wetzel, D.P., and J.J. Krupa. 2013. Where are the bluebirds of the Bluegrass? Eastern Bluebird decline in central Kentucky. American Midland Naturalist 169:398–408. Whitaker, G.A. 1974. Phoebe and Barn Swallow nesting structures. Transactions of the Northeastern Fisheries and Wildlife Conference 31:57–62. Wilson, S., E.M. Anderson, A.S.G. Wilson, D.F. Bertram, and P. Arcese. 2013. Citizen science reveals an extensive shift in the winter distribution of migratory Western Grebes. PLoS ONE 8(6): e65408. doi:10.1371/journal.pone.0065408. Xu, C., J. Barrett, D.B. Lank, and R.C. Ydenberg. 2015. Large and irregular population fluctuations in migratory Pacific (Calidris alpina pacifica) and Atlantic (C.a. hudsonica) Dunlins are driven by density-dependence and climatic factors. Population Ecology 57:551–567.