2009 SOUTHEASTERN NATURALIST 8(1):107–120 Avian Response to Microclimate in Canopy Gaps in a Bottomland Hardwood Forest Tracey B. Champlin1,2, John C. Kilgo3,*, Marcia L. Gumpertz4, and Christopher E. Moorman1 Abstract - Microclimate may infl uence use of early successional habitat by birds. We assessed the relationships between avian habitat use and microclimate (temperature, light intensity, and relative humidity) in experimentally created canopy gaps in a bottomland hardwood forest on the Savannah River Site, SC. Gaps were 2- to 3-year-old group-selection timber harvest openings of three sizes (0.13, 0.26, 0.50 ha). Our study was conducted from spring through fall, encompassing four bird-use periods (spring migration, breeding, post-breeding, and fall migration), in 2002 and 2003. We used mist netting and simultaneously recorded microclimate variables to determine the infl uence of microclimate on bird habitat use. Microclimate was strongly affected by net location within canopy gaps in both years. Temperature generally was higher on the west side of gaps, light intensity was greater in gap centers, and relative humidity was higher on the east side of gaps. However, we found few relationships between bird captures and the microclimate variables. Bird captures were inversely correlated with temperature during the breeding and postbreeding periods in 2002 and positively correlated with temperature during spring 2003. Captures were high where humidity was high during post-breeding 2002, and captures were low where humidity was high during spring 2003. We conclude that variations in the local microclimate had minor infl uence on avian habitat use within gaps. Instead, habitat selection in relatively mild regions like the southeastern US is based primarily on vegetation structure, while other factors, including microclimate, are less important. Introduction Birds are often more abundant in and adjacent to gaps in forest canopies, especially larger gaps, than in adjacent mature forest (Kilgo et al. 1999, Moorman and Guynn 2001). However, the relative use of gaps by birds varies temporally from spring migration through fall migration (Bowen et al. 2007). For example, post-fl edging dispersal of birds from natal areas in mature forest with high canopy cover tends to be towards large openings or patches of early successional habitat with little to no canopy and a dense understory (Anders et al. 1998, Vega Rivera et al. 1998), thus yielding greater abundance there during the post-breeding period than the breeding period. Higher food abundance (Blake and Hoppes 1986, Levey 1988), increased 1Department of Forestry and Environmental Resources, North Carolina State University, Campus Box 8003, Raleigh, NC 27695. 2Current address - 10 Lepire Avenue, Westport, MA 02790. 3USDA Forest Service Southern Research Station, PO Box 700, New Ellenton, SC 29809. 4Department of Statistics, North Carolina State University, Campus Box 8203, Raleigh, NC 27695. *Corresponding author - jkilgo@fs.fed.us. 108 Southeastern Naturalist Vol. 8, No. 1 vegetative structure (Martin and Karr 1986), or higher temperature (Karr and Freemark 1983) may explain the greater use of gaps by birds. However, few studies have addressed the roles these habitat components play in temporal or spatial changes in bird use of early successional habitat. Karr and Freemark (1983) suggested the importance of microclimate as a factor in determining avian “physiological comfort.” Extremes in microclimate may have adverse effects on birds and their reproductive fitness (Martin and Ghalambor 1999, Wachob 1996a, Walsberg 1985). Conversely, selection for particular microclimatic conditions may minimize foraging costs in relation to benefits. Feeding in open areas such as forest canopy gaps where the incoming solar radiation and air temperature is greater than the adjacent forest (Chen et al. 1993) energetically reduces foraging costs (DeWoskin 1980). Most microclimate studies have focused on fixed points in time or space, such as single seasons, nest sites, roost sites, or breeding or wintering grounds (Calder 1973; Martin and Ghalambor 1999; Wachob 1996a, 1996b). Less is known about the effects of microclimatic factors on diurnal activities of birds across a spatial and temporal gradient, although microclimate studies have been conducted in clearcuts, where air temperature dramatically increases during the day relative to forest interiors (Chen et al. 1993). Birds may respond directly to microclimate changes or indirectly to changes in food or cover resulting from microclimate changes. For example, many arthropods, which constitute an important food resource for birds, respond positively to light and negatively to shade (White 1984). Abundance and diversity of certain arthropod orders may be high in gaps because of high light intensity (Ulyshen et al. 2004). Light, wind, temperature, radiation, humidity, and soil conditions define specific microclimates (Chen et al. 1999, Phillips and Shure 1990,Walsberg 1985). Whether natural or man-made, changes in forest structure affect the local microclimate (Minckler et al. 1973, Phillips and Shure 1990, Saunders et al. 1991). The creation of a canopy gap results in spatial and temporal differences in the microclimate (Chen et al. 1999, Phillips and Shure 1990, Saunders et al.1991, Xu et al. 1997), in plant species composition and structure (Phillips and Shure 1990), and hence in resource availability (Levey 1988). These effects of microclimate on components of habitat quality may be important in infl uencing aspects of avian life history such as survival, reproduction, timing of breeding seasons, species dispersal, and habitat selection (McCollin 1998, Perrins 1965, Perry 1994). Even within seemingly homogenous habitat patches, small-scale variability in microclimate may be important in habitat selection by birds (Sisk et al. 1997). Because microclimate metrics vary differentially with the type and size or length of edge (Chen et al. 1999, Saunders et al. 1991), gap size may similarly infl uence microclimate. For example, light intensity and air temperature increase from mature to successional forests (Chen et al. 1995, Sisk et al. 1997). Similarly, particular locations within a canopy gap likely 2009 T.B. Champlin, J.C. Kilgo, M.L. Gumpertz, and C.E. Moorman 109 have unique microclimates that may infl uence bird activity because of shading or distance to forest. The degree to which birds respond to microclimatic variation within canopy gaps and among gaps of different size is unknown. We assessed the relationships between avian use of gaps and microclimate measures (temperature, light intensity, and relative humidity) in experimentally created forest canopy gaps of three sizes in a bottomland hardwood forest on the Savannah River Site in South Carolina. We hypothesized that mist-net captures of birds would vary with temperature, light intensity, or relative humidity. Methods Study site and design We conducted the study during 2002 and 2003 at the US Department of Energy’s Savannah River Site, a 78,000-ha National Environmental Research Park in the Upper Coastal Plain of South Carolina. The climate of the area is humid subtropical, with a mean annual temperature of 18 ºC and mean annual precipitation of 122.5 cm (Blake et al. 2005). During April–October (our study period), long-term (1964–2001) mean monthly temperature ranged from 18.4 oC (April) to 28.1 oC (July), and mean relative humidity ranged from 56% (April) to 78% (August, September) (Blake et al. 2005). Our 120-ha study site was located in a 70- to 100-year-old, seasonally fl ooded bottomland hardwood forest. The forest canopy of the study site included typical bottomland hardwood species: Quercus falcata var. pagodaefolia Elliot (Cherrybark Oak), Q. laurifolia Michaux (Laurel Oak), Q. phellos Linnaeus (Willow Oak), Q. lyrata Walter (Overcup Oak) Q. michauxii Nuttall (Swamp Chestnut Oak), Liquidambar styracifl ua Linnaeus (Sweetgum), and Pinus taeda Linnaeus (Loblolly Pine). The understory was dominated by Sabal minor Persoon (Dwarf Palmetto) and Arundinaria gigantea Muhlenberg (Giant Switchcane), and a generally sparse midstory consisted primarily of Ilex opaca Aiton (American Holly), Morus rubra Linnaeus (Red Mulberry), and Carpinus carolinianus Walter (Ironwood). As part of a larger study examining bird-arthropod relations in canopy gaps (Champlin 2007), we created 12 experimental canopy openings (hereafter gaps) via group-selection timber harvest in August 2000. Four replicates of three sizes (0.13, 0.26, and 0.50 ha) were harvested, with the boundary of each gap at least 150 m from the nearest adjacent gap. Because arthropod abundance was manipulated in six of the gaps, herein we consider only data from the six unmanipulated control gaps (two of each size). Gaps were circular in shape and were defined to include all of the cleared area within the circumference delineated by the boles of trees left standing at the gap perimeter. During our study, which occurred during the second and third growing seasons post-harvest, gaps mainly contained early pioneering species such as grasses (Poaceae), sedges (Cyperaceae), Eupatorium capillifolium (Lam.) 110 Southeastern Naturalist Vol. 8, No. 1 Small (Dogfennel), Dwarf Palmetto, Giant Switchcane, and some woody stump sprout regenerative growth. Mist-netting We determined the relationship between bird use of various sized gaps and their microenvironments during four periods: spring migration (1 Apr–12 May), breeding (13 May–7 Jul), post-breeding (8 Jul–31 Aug), and fall migration (1 Sep–18 Oct). Dates for these periods were based on previous research in the area (Kilgo et. al 1999, Moorman and Guynn 2001). Mist-netting and microclimate measurements (see below) were conducted simultaneously, 5 days per week during all periods. Because vegetation structure was generally similar among our gaps and was <3 m tall (the height of our mist nets), we felt that biases typically associated with mist-net sampling were minimized (Remsen and Good 1996). Five nets (4-panel, 30-mm mesh) were deployed in each gap: one at center, one each on and perpendicular to the north and south edges, and two halfway to the east and west gap edges. Nets were operated in three gaps (one of each size) each day from daylight until approximately 3–5 hrs post sunrise, depending on weather conditions. Captured birds were identified to species, banded with a metal federal band, aged, sexed, measured, and released (Pyle 1997). Numbers of captures were standardized to captures per 100 net hours. Our sample size was too small to conduct analyses at the species level. Vegetation sampling Vertical vegetation structure was recorded at each mist net during July and early August in 2002 and 2003, using a modification of the techniques of Karr (1971) and Schemske and Brokaw (1981). Because most gap vegetation occurred within 3 m of the ground, we sampled only the lowest 3 m of vegetation. Two 12-m transects were established parallel to and 2 m distant from each side of each mist net. At 10 sampling points (1.2-m intervals) along each transect, the presence or absence of vegetation in each of 7 height intervals was recorded for a total of 20 points per height interval per net. Height intervals were 0–0.25, 0.25–0.50, 0.50–0.75, 0.75–1.00, 1.0–1.5, 1.5–2, and 2–3 m. Vegetation touches were recorded along a 2-cm x 2-m vertical pole at each sampling point and were tallied as grass/sedge, herb/forb, woody, or vine. For the 2–3-m height interval, we sighted along the pole and recorded the presence or absence of vegetation. We grouped all height intervals to calculate % cover for each of the four vegetation types (GRASS, FORB, WOODY, and VINE), and a gap’s total % cover (COVER) was calculated as the sum of % cover for the four vegetation types. Microclimate data collection We measured temperature, light intensity, and relative humidity. We selected these variables because they were easily measured and have been used commonly in other studies of bird response to microclimate. We could 2009 T.B. Champlin, J.C. Kilgo, M.L. Gumpertz, and C.E. Moorman 111 not measure the effects of wind because we could not sample bird response in windy conditions with mist nets. We used Four-channel HOBO® H-8 data loggers (Onset Computer Corporation, Bourne, MA) that had an accuracy and collection range for temperature, light intensity, and relative humidity of ±0.70 °C over -20–70 °C, ±21.5 lumens/m2 (±20% of reading) over 22–6458 lumens/m2, and ±5% over 5–50 °C, respectively. Data loggers were placed each day at fixed positions at the midpoint of each mist net (see Avian Response below), 1 m from the cleared net lane and 1 m above the ground (Sisk et al. 1997). North and south nets measured conditions at the gap edges, though data loggers were placed within the gap. East, west, and center nets measured conditions in the gap. Loggers were affixed to the top of a 1-m cane pole with Velcro® to minimize contact surface area that might infl uence the environmental readings. HOBO® logger data collection intervals were set to once every 12 sec, for a maximum collection time of 6 hr, 37 min (1985 data points per logger). We assumed that birds in the net at the time of extraction were caught within the climate recording hour interval (synonymous with mist-net checks every hour). Loggers were not deployed when conditions were foggy (i.e., visible water vapor in the air) to prevent damage to the humidity sensor and were removed if rain was threatening. We calculated the mean for each microclimate variable by averaging the values from points taken within the net-check hour each day for each logger. Data analysis Only those mist-net captures that had corresponding microclimate data were used in analyses. For example, if a data logger was removed early because of the threat of rain, any subsequent captures were not included in the analyses. Because of substantial differences in rainfall between sampling years, with drought in 2002 and the presence of surface water in spring 2003, we analyzed years separately. All data were tested for normality. We used a linear mixed model (PROC MIXED, SAS Institute 2001) to analyze the effects of gap size and net location on vegetation. We considered gaps as the replicate whole plot units, gap size as the whole plot factor, and net location as the subplot factor. Because vegetation data were collected once per year, this model did not include period. We used a linear mixed model (PROC MIXED, SAS Institute 2001) to compare mean temperature, light intensity, and relative humidity among periods, gap sizes, and net locations in each year. All microclimate measures were highly correlated (r > 0.9), so each microclimate parameter was analyzed in a separate model. We considered gaps as the replicate whole plot units, gap size and period as the whole plot factors, and net as the subplot factor. We analyzed bird response to microclimate separately for each bird-use period. We used a linear mixed model (PROC MIXED; SAS institute 2001) 112 Southeastern Naturalist Vol. 8, No. 1 to test avian response to the effects of temperature (°C), light intensity (lumens/ m2), and relative humidity (%). We considered gaps as the replicate whole plot units, gap size as the whole plot factor, and net location as the subplot factor. Birds considered winter residents (Hamel 1992), present only in early spring or late fall, and hummingbirds were not included in analyses. Only initial captures were used in analyses to minimize effects of autocorrelation. We could not test for daily response because the number of bird captures each day was small. Results Mist-net captures In 2002, we operated mist nests for 3842 net hours, during which time we obtained microclimate data and captured 263 individual birds (excluding recaptures) representing 41 species. In 2003, we operated mist nets for 4107 net hours and captured 282 individuals representing 33 species. However, six species comprised 70% of the captures: Baeolophus bicolor Linnaeus (Tufted Titmouse, 6%); Thryothorus ludovicianus Latham (Carolina Wren, 17%); Parula americana Linnaeus (Northern Parula, 7%), Geothlypis trichas Linnaeus (Common Yellowthroat, 22%), Cardinalis cardinalis Linnaeus (Northern Cardinal, 6%), and Passerina cyanea Linnaeus (Indigo Bunting, 12%). Vegetation Total % cover of vegetation (COVER) and the four categories of vegetation cover (GRASS, FORB, VINE, WOODY) did not differ among gap sizes (Table 1). In both years, total % cover (COVER) and % cover of FORB and GRASS were highest at center nets and lowest at the north and south edge nets (Table 1). Microclimate We deployed Hobo® data loggers 108 days in 2002 and 119 days in 2003, resulting in 7949 logger-hours. This produced 2,384,700 12-sec readings (7949 logger-hours x 60 min/hr x 5 readings/min) for each microclimate variable over the 2-yr study. During daylight hours (06:00–13:00), mean air temperature in canopy gaps averaged 22.56 ± 1.98 °C in 2002 and 22.08 ± 2.76 °C in 2003. Mean light intensity averaged 5884 ± 1173 lumens/m2 in 2002 and 5679 ± 1157 lumens/m2 in 2003. Mean relative humidity averaged 81.99 ± 6.94% in 2002 and 76.90 ± 8.71% in 2003. We detected few effects of gap size on microclimate. Relative humidity was highest in 0.13-ha gaps during the breeding period in 2003, and light intensity was highest in 0.50-ha gaps during the breeding, post-breeding, and fall migration periods in 2003 (Table 2). Temperature and relative humidity differed among net locations in all seasons of both years except during the 2002 post-breeding period (Table 3). 2009 T.B. Champlin, J.C. Kilgo, M.L. Gumpertz, and C.E. Moorman 113 Light intensity differed among net locations in every period of 2002 except for fall migration, but differed among net locations only during the breeding period and fall migration in 2003 (Table 3). Avian response to microclimate Bird captures decreased with temperature increases during the breeding and post-breeding periods in 2002 and increased with temperature increases during spring 2003 (Table 4). Bird captures were lowest where light intensity was highest during the breeding period in 2002 (Table 4). Bird captures and relative humidity were positively correlated during the post-breeding period in 2002 and negatively correlated during spring migration in 2003 (Table 4). During the fall migration period, no relationships between captures and microclimate were detected (Table 4). Table 1. Percent vegetation cover in three gap sizes (ha) and at five net locations in canopy gaps at the Savannah River Site, SC (2002–2003). a. Gap size Variable 0.13 ha 0.26 ha 0.50 ha SE F2,3 P COVER 2002 45.93 46.64 43.07 9.73 0.08 0.929 2003 53.79 48.14 58.79 11.23 0.45 0.675 FORB 2002 13.64 18.43 13.21 6.65 0.38 0.714 2003 12.50 12.21 12.57 3.42 0.01 0.994 GRASS 2002 22.57 24.79 19.57 5.89 0.40 0.704 2003 23.21 27.71 20.14 4.49 1.44 0.365 VINE 2002 4.93 1.07 3.57 3.21 0.74 0.547 2003 7.79 3.00 8.57 6.04 0.50 0.650 WOODY 2002 4.79 6.07 6.71 3.06 1.02 0.459 2003 10.29 5.21 17.50 7.53 1.35 0.383 b. Net location Variable North South East West Center SE F4,12 P COVER 2002 22.50 23.21 61.07 54.88 64.40 7.52 15.16 0.001 2003 24.17 34.29 65.12 64.40 79.88 5.99 30.33 <0.001 FORB 2002 2.50 1.55 27.02 17.74 26.67 5.64 9.82 0.001 2003 3.57 2.14 19.88 13.57 22.98 3.37 15.55 0.001 GRASS 2002 15.00 11.79 23.57 30.48 30.71 5.62 4.79 0.015 2003 11.90 16.19 24.40 32.50 33.45 5.49 6.11 0.006 VINE 2002 0.60 4.05 6.19 2.38 2.74 2.32 1.61 0.236 2003 1.79 7.98 9.05 6.67 6.79 3.40 1.34 0.311 WOODY 2002 4.40 5.83 4.29 4.29 4.29 2.11 0.21 0.929 2003 6.90 7.98 11.79 11.67 16.67 5.11 1.13 0.388 114 Southeastern Naturalist Vol. 8, No. 1 Discussion Microclimate varied much more within gaps than among gaps of different sizes. This within-gap variation was likely a result of differential shading or distance to mature trees. The progression of light entering a gap as the sun rose infl uenced the microclimate gradients in the gap. For example, the west nets were first to receive light as the sun penetrated the interior of gaps, thus beginning to warm the gaps from the west and progressing east. As the sun’s angle increased, the temperature within a gap increased from west to east. Temperature generally corresponded to light intensity with a slight time lag, because temperature usually increases as light intensity increases (Ritter 2006). As gaps warmed, relative humidity changed inversely with air temperature. Relative humidity, the ratio of the amount of water vapor in the Table 2. Microclimate variation by gap size (ha) in six canopy gaps at the Savannah River Site, SC (2002–2003). Year Period/variable 0.13 ha 0.26 ha 0.50 ha SE F2,3 P 2002 Spring Temp oC 20.74 22.57 21.92 1.11 0.69 0.566 LightA 5656 6398 6802 475 1.50 0.354 HumidityB 78.68 76.34 71.52 2.72 1.80 0.306 Breeding Temp oC 21.17 21.03 21.67 1.06 0.10 0.907 Light 5169 5696 6458 464 1.95 0.286 Humidity 85.00 82.50 80.58 3.00 0.55 0.627 Post-breeding Temp oC 23.78 23.92 24.39 0.62 0.27 0.780 Light 4663 5437 6316 636 1.69 0.323 Humidity 89.53 89.85 87.75 1.62 0.49 0.655 Fall Temp oC 22.57 23.04 23.92 0.82 0.70 0.564 Light 5229 5838 6555 829 0.64 0.586 Humidity 82.64 81.39 78.04 1.93 1.52 0.350 2003 Spring Temp oC 20.92 21.79 20.56 1.29 0.24 0.801 Light 5500 6488 6498 257 4.96 0.112 Humidity 82.84 76.21 74.78 4.46 0.93 0.485 Breeding Temp oC 20.95 22.03 22.23 0.54 1.65 0.329 Light 4778 5768 6406 262 9.81 0.048 Humidity 85.25 82.58 76.55 1.10 16.29 0.025 Post-breeding Temp oC 24.28 25.29 26.42 0.41 6.96 0.075 Light 4567 6009 6653 266 16.07 0.025 Humidity 67.76 66.98 66.52 2.27 7.78 0.065 Fall Temp oC 18.19 20.32 22.04 0.78 6.09 0.088 Light 4158 5479 5848 233 14.50 0.029 Humidity 67.76 66.98 66.52 2.27 0.08 0.929 ALight = lumens/m2. BHumidity = relative humidity (%). 2009 T.B. Champlin, J.C. Kilgo, M.L. Gumpertz, and C.E. Moorman 115 air to the amount of water vapor the air can hold for a given temperature, is dependent on air temperature (Critchfield 1983). The east side of gaps generally had cooler temperatures and thus experienced higher relative humidity. When the sun reached its zenith, the temperature differential across a gap abated. Variations in microclimate within gaps were fairly consistent. In nearly every season in both years, temperature was higher at the west nets, light intensity was greater at the center nets, and relative humidity was higher at the east nets. We found few relationships between bird captures and microclimate variables. Birds apparently were attracted to lower temperatures and higher humidity during the warmer seasons (breeding and post-breeding periods) of 2002, and they tracked higher temperatures and lower humidity during Table 3. Microclimate variation by net location in six canopy gaps at the Savannah River Site, SC (2002–2003). N = north, S = south, E = east, W = west, and C = center. Year Period/variable N S E W C SE F4,12 P 2002 Spring Temp oC 21.93 20.04 19.90 24.20 22.65 0.68 52.59 <0.001 LightA 5404 5789 6385 6949 6900 305 20.58 <0.001 HumidityB 74.64 79.48 80.99 68.73 73.73 1.69 50.03 <0.001 Breeding Temp oC 20.52 20.84 20.50 22.91 21.69 0.66 15.20 <0.001 Light 4534 5286 6083 6289 6681 314 22.13 <0.001 Humidity 84.10 83.12 86.29 77.77 82.20 1.91 11.97 0.001 Post-breeding Temp oC 23.93 23.81 23.37 24.94 24.10 0.47 2.97 0.064 Light 4857 5734 5102 5510 6157 439 3.68 0.035 Humidity 88.28 89.26 91.53 86.23 89.92 1.40 2.82 0.074 Fall Temp oC 24.18 21.73 22.43 23.82 23.72 0.56 9.66 0.001 Light 5798 5752 5780 5998 6041 536 0.25 0.904 Humidity 77.15 83.95 83.94 78.49 79.91 1.51 7.57 0.003 2003 Spring Temp oC 21.40 19.39 19.42 22.49 22.74 0.80 25.23 <0.001 Light 5420 5844 6205 6407 6934 332 2.97 0.064 Humidity 77.60 81.76 83.16 73.13 74.07 2.69 26.96 <0.001 Breeding Temp oC 21.13 21.07 20.62 23.13 22.71 0.38 21.34 <0.001 Light 4564 6004 5742 5739 6203 338 3.56 0.039 Humidity 82.34 82.33 84.59 77.98 80.07 0.92 11.67 <0.001 Post-breeding Temp oC 24.92 24.56 24.36 26.77 26.04 0.36 11.74 0.001 Light 5028 5510 6021 6224 5932 301 2.72 0.081 Humidity 65.04 68.50 75.46 62.10 64.34 2.18 5.34 0.011 Fall Temp oC 20.91 18.41 18.81 21.75 21.05 0.61 10.53 0.001 Light 4725 4639 5113 5967 5364 301 3.18 0.053 Humidity 65.04 68.50 75.46 62.10 64.34 2.18 7.22 0.003 ALight = lumens/m2. BHumidity = relative humidity (%). 116 Southeastern Naturalist Vol. 8, No. 1 the cooler weeks of spring migration 2003. However, in most seasons, relationships were not significant. Birds avoided areas of the greatest light intensity during the breeding period in 2002, but nets with the highest light intensity probably were most visible to birds. Also, mist-net captures were lowest later in the morning when light intensity was highest. In general, relationships between avian habitat use and microclimate were weak. Two factors that may have limited our ability to detect avian-microclimate relationships include the necessity of pooling all species and the scale of investigation. Pooling species may have obscured any species-specific responses. However, because our sample was dominated by shrub-scrub Table 4. Slopes of the lines describing the relationships between microclimate variables and mist-net captures in six canopy gaps at the Savannah River Site, SC (2002–2003). Also presented are the effects of gap size and net location, included in the ANOVA model to control variance contributed by these effects. Mist-net Captures Gap size Net Year Period/variable Slope SE F1,11 P F2,3 P F4,11 P 2002 Spring Temp ºC -0.0619 0.208 0.09 0.771 2.03 0.277 0.78 0.560 LightA -0.0059 0.004 1.87 0.199 3.86 0.148 1.24 0.351 HumidityB 0.1327 0.068 3.78 0.078 5.23 0.105 1.66 0.228 Breeding Temp ºC -0.3207 0.131 5.98 0.033 3.73 0.153 2.90 0.073 Light -0.0070 0.003 7.18 0.021 8.13 0.062 2.74 0.083 Humidity 0.1017 0.054 3.57 0.085 3.12 0.185 2.94 0.070 Post-breeding Temp ºC -0.5613 0.246 5.20 0.044 0.96 0.477 2.84 0.077 Light -0.0041 0.003 1.57 0.237 1.20 0.414 1.71 0.218 Humidity 0.1729 0.075 5.28 0.042 0.67 0.575 2.62 0.093 Fall Temp ºC -0.1566 0.153 1.05 0.328 22.37 0.016 3.69 0.038 Light -0.0013 0.002 0.57 0.467 20.67 0.018 3.72 0.038 Humidity 0.0384 0.058 0.44 0.520 18.24 0.021 3.03 0.066 2003 Spring Temp ºC 0.4391 0.142 9.62 0.010 3.04 0.190 5.38 0.012 Light -0.0044 0.003 2.68 0.130 0.31 0.754 5.93 0.009 Humidity -0.1393 0.047 8.85 0.013 2.30 0.248 5.15 0.014 Breeding Temp ºC 0.2938 0.263 1.25 0.288 2.76 0.209 0.88 0.509 Light -0.0019 0.003 0.32 0.582 3.27 0.176 1.88 0.184 Humidity -0.0215 0.118 0.03 0.859 0.52 0.638 1.05 0.424 Post-breeding Temp ºC -0.4362 0.331 1.74 0.214 2.85 0.202 1.02 0.441 Light -0.0061 0.004 2.45 0.146 4.15 0.137 1.27 0.341 Humidity 0.1176 0.055 4.61 0.055 6.30 0.084 1.63 0.236 Fall Temp ºC 0.0451 0.178 0.06 0.805 0.42 0.693 0.38 0.821 Light -0.0004 0.004 0.01 0.971 0.24 0.802 0.33 0.854 Humidity 0.0031 0.050 0.00 0.951 0.77 0.539 0.37 0.823 ALight = lumens/m2. BHumidity = relative humidity (%). 2009 T.B. Champlin, J.C. Kilgo, M.L. Gumpertz, and C.E. Moorman 117 species, we suspect that these species likely had similar tolerances for the microclimatic conditions of the early successional habitat in our gaps, and thus might have responded in similar ways. Likewise, the fact that we sampled only in gaps, and not in adjacent mature forest, prevented us from detecting birds that may have left gaps completely for potentially more desirable microclimates in nearby forested areas. Although we did not sample forested areas because our focus was on within-gap habitat use, we may have detected more consistent patterns in avian habitat use had we encompassed a broader range of microclimatic conditions. Although microclimate does not appear to be a primary determinant of habitat use in our study, it may be important in some seasons or conditions. During some warmer seasons, captures were greater in areas of lower temperatures and higher relative humidity, whereas during some cooler seasons, captures were greater in areas of higher temperature and lower relative humidity. These within-patch shifts in avian habitat use during various seasons suggest that birds occasionally tracked microclimate for physiologic comfort (Karr and Freemark 1983), even if these patterns were not consistent throughout the study. Similarly, Phainopepla nitens Swainson (Phainopepla) in the southwestern US shifted to cooler microclimates to balance thermoregulatory demands during summer (Walsberg 1985), and Poecile gambeli Ridgway (Mountain Chickadee) in Wyoming preferentially selected warmer foraging sites during winter (Wachob 1996b). Both of these studies were conducted in more extreme environments than ours. The infl uence of microclimate on avian habitat selection may be less important at our study site because of the relatively mild climate there. Despite the occurrence of both high and low rainfall conditions between the two years of our study, the microclimate variables we measured were within long-term norms for the area (Blake et al. 2005). The Southeast generally lacks environmental extremes, and this milder climate may favor selection of habitats by birds via factors other than microclimate, except perhaps on the coldest and warmest days. The lack of consistent relationships between bird captures and microclimate may be evidence that vegetation structure has more infl uence on avian use of canopy gaps than microclimate. Bowen (2004) and Champlin (2007) concluded that avian use of forest canopy gaps was in response to the complex vegetation structure there rather than arthropod prey availability. Birds may select the dense vegetation in forest canopy gaps during migration and post-breeding dispersal, periods of increased mobility and vulnerability to predators. During the post-breeding period, molting adults and newly fl edged young may seek out gaps for the protective cover they offer (Anders et al. 1998, Vega Rivera et al. 1998). We conclude that microclimate variability is not the primary determinant of gap habitat selection by birds in bottomland forests of the relatively mild southeastern US. Rather, habitat selection is most likely based primarily on vegetation structure, with microclimate being less important. 118 Southeastern Naturalist Vol. 8, No. 1 Acknowledgments The US Department of Energy provided funding for this study through the USDA Forest Service-Savannah River and the USDA Forest Service Southern Research Station under Interagency Agreement (DE-AI09-00SR22188). Thanks also to the USDA Forest Service-Savannah River for logistical support, particularly J. Blake, K. Wright, and E. Olson. North Carolina State University provided logistical and financial assistance. We thank D. Jones, S. Junker, F. Spilker, and M. Olsen for help with fieldwork, S. Donaghy and K. Gross for statistical assistance, and C. Sorenson for reviewing the manuscript. Literature Cited Anders, A.D., J. Faaborg, and F.R. Thompson III. 1998. Postfl edging dispersal, habitat use, and home-range size of juvenile Wood Thrushes. Auk 115:349–358. Blake, J.G., and W.G. Hoppes. 1986. Infl uence of resource abundance on use of treefall gaps by birds in an isolated woodlot. Auk 103:328–340. Blake, J.I., C.H. Hunter, Jr., and B.A. Bayle. 2005. Climate and air quality. Pp. 20– 30, In J.C. Kilgo and J.I. Blake (Eds.). Ecology and Management of a Forested Landscape: Fifty Years on the Savannah River Site. Island Press, Washington, DC. 479 pp. Bowen, L.T. 2004. Seasonal relationships between birds and arthropods in bottomland forest canopy gaps. Ph.D Dissertation. North Carolina State University, Raleigh, NC. 86 pp. Bowen, L.T., C.E. Moorman, and J.C. Kilgo. 2007. Seasonal bird use of canopy gaps in a bottomland forest. Wilson Journal of Ornithology 119:77–88. Calder, W.A. 1973. Microhabitat selection during nesting of hummingbirds in the Rocky Mountains. Ecology 54:127–134. Champlin, T.B. 2007. Do food availability and microclimate determine bird use of forest canopy gaps? M.Sc. Thesis. North Carolina State University, Raleigh, NC. 68 pp. Chen J., J.F. Franklin, and T.A. Spies. 1993. Contrasting microclimates among clearcut, edge, and interior of old-growth Douglas-fir forest. Agricultural and Forest Meteorology 63:219–237. Chen, J., J.F. Franklin, and T.A. Spies. 1995. Growing-season microclimate gradients from clearcut edges into old-growth Douglas-fir forests. Ecological Applications 5:74–86. Chen, J., S.C. Saunders, T.R. Crow, R.J. Naiman, K.D. Brosofske, G.D. Mroz, B.L. Brookshire, and J.F. Franklin. 1999. Microclimate in forest ecosystem and landscape ecology. BioScience 49:288–297. Critchfield, H.J. 1983. General Climatology, 4th Edition. Prentice-Hall, Inc., Englewood Cliffs, NJ. 446 pp. DeWoskin, R. 1980. Heat exchange infl uence on foraging behavior in Zonotrichia fl ocks. Ecology 61:30–36. Hamel, P.B. 1992. Land Manager’s Guide to the Birds of the South. The Nature Conservancy, Chapel Hill, NC. 367 pp. Karr, J.R. 1971. Structure of avian communities in selected Panama and Illinois habitats. Ecological Monographs 41:130–148. 2009 T.B. Champlin, J.C. Kilgo, M.L. Gumpertz, and C.E. Moorman 119 Karr, J.R., and K.E. Freemark. 1983. Habitat selection and environmental gradients: Dynamics in the “stable” tropics. Ecology 64:1481–1494. Kilgo, J.C., K.V. Miller, and W.P. Smith. 1999. Effects of group-selection timber harvest in bottomland hardwoods on fall migrant birds. Journal of Field Ornithology 70:404–413. Levey. D.J. 1988. Tropical wet forest treefall gaps and distributions of understory birds and plants. Ecology 69:1076–1089. Martin, T.E., and C.K. Ghalambor. 1999. Males feeding females during incubation. I. Required by microclimate or constrained by nest predation? American Naturalist 153:131–139. Martin, T.E., and J.R. Karr. 1986. Patch utilization by migrating birds: Resource oriented? Ornis Scandinavica 17:165–174. McCollin, D. 1998. Forest edge and habitat selection in birds: A functional approach. Ecography 21:247–260. Minckler, L.S., J.D. Woerheide, and R.C. Schlesinger. 1973. Light, soil moisture, and tree reproduction in hardwood forest openings. Research Paper NC-89, USDA Forest Service, North Central Experimental Station, St. Paul, MN. 9 pp. Moorman, C.E., and D.C. Guynn, Jr. 2001. Effects of group-selection opening size on breeding-bird habitat use in a bottomland forest. Ecological Applications 11:1680–1691. Perrins, C.M. 1965. Population fl uctuations and clutch size in the Great Tit Parus major L. Journal of Animal Ecology 34:601–647. Perry, D.A. 1994. Forest Ecosystems. Johns Hopkins University Press, Baltimore, MD. 649 pp. Phillips, D.L., and D.J. Shure. 1990. Patch-size effects on early succession in southern Appalachian forests. Ecology 71:204–212. Pyle, P. 1997. Identification Guide to North American Birds. Slate Creek Press, Bolinas, CA. 732 pp. Remsen, J.V., and D.A. Good. 1996. Misuse of data from mist-net captures to assess relative abundance in bird populations. Auk 113:381–398. Ritter, M.E. 2006. The physical environment: An introduction to physical geography. Available online at http://www.uwsp.edu/geo/faculty/ritter/geog101/textbook/ title_page.html. Accessed December 10, 2007. SAS Institute. 2001. SAS users guide version 8.2. SAS Institute, Inc., Cary, NC. 1848 pp. Saunders, D.A., R.J. Hobbs, C.R. Margules. 1991. Biological consequences of ecosystem fragmentation: A review. Conservation Biology 5:18–32. Schemske, D.W., and N. Brokaw. 1981. Treefalls and the distribution of understory birds in a tropical forest. Ecology 62:938–945. Sisk, T.D., N.M. Haddad, and P.R. Ehrlich. 1997. Bird assemblages in patchy woodlands: Modeling the effects of edge and matrix habitats. Ecological Applications 7:1170–1180. Ulyshen, M.D., J.L Hanula, S. Horn, J.C. Kilgo, and C.E. Moorman. 2004. Spatial and temporal patterns of beetles associated with coarse woody debris in managed bottomland hardwood forests. Forest Ecology and Management 199:259–272. Vega Rivera, J.H., J.H. Rappole, W.J. McShea, and C.S. Haas. 1998. Wood Thrush post-fl edging movements and habitat use in northern Virginia. Condor 100:69–78. 120 Southeastern Naturalist Vol. 8, No. 1 Wachob, D.G. 1996a. A microclimate analysis of nest-site selection by Mountain Chickadees. Journal of Field Ornithology 67:525–533. Wachob, D.G. 1996b. The effect of thermal microclimate on foraging site selection by wintering Mountain Chickadees. Condor 98:114–122. Walsberg, G.E. 1985. Physiological consequences of microhabitat selection. Pp. 389–413, In M.L. Cody (Ed.). Habitat Selection in Birds. Academic Press, New York, NY. 558 pp. White, T.C.R. 1984. The abundance of invertebrate herbivores in relation to the availability of nitrogen in stress food plants. Oecologia 63:90–105. Xu, M., J. Chen, and B.L. Brookshire. 1997. Temperature and its variability in the oak forests of southeast Missouri’s Ozarks. Climate Research 8:209–223.