Urban Naturalist 1 C.J. Gearin and J.D. Luscier 22001177 URBAN NATURALIST No. 14N:1o–. 2104 Short-term Responses of Fall Migratory Bird Communities to Suburban Riparian Management in Northeastern Missouri, USA Conor J. Gearin1,2,* and Jason D. Luscier1,3 Abstract - Riparian zones are important corridors for birds, and the structural diversity of intact vegetation along water courses that pass through fragmented landscapes may affect the diversity of birds that use such corridors. During fall migration in 2014, we surveyed 3 riparian sites in Adair County, MO, USA: a restored suburban site, a suburban site with mowed vegetation, and a rural site in a conservation area with oak-savannah vegetation. We hypothesized that, during fall migration, a suburban site restored to have higher levels of non-woody vegetation near stream banks would support bird communities with more species associated with hydric habitats when compared to the mowed suburban site. We also predicted that bird-community composition at the restored suburban site with taller vegetation would be more similar to that found at a rural site than would be expected at the unimproved suburban site. Due to the small scale of the restoration project, we were not able to replicate the treatments. Species richness estimates (± SE) were 10.6 ± 3.18 for the restored site, 12.2 ± 2.76 for the unrestored site, and 18.2 ± 3.06 for the conservation area. Although species richness was similar among suburban sites, species composition was not. The relative abundance of hydric-associated species (± 95% CI) were 24.68% ± 8.64% for the restored site, 8.33% ± 2.75% for the unrestored site, and 24.01% ± 6.90% for the conservation area. The unrestored site had a higher abundance of riparian species and the restored site had higher abundance of wetland-associated species. Differences in vegetative structure had significant impacts on community composition in the short te rm. Introduction North American bird populations are declining rapidly, and long-distance migrants have faced sharp declines (Murphy 2003). A suite of conservation measures are needed to address this decline, and riparian restoration could play an important role in in supporting avian migration through fragmented and developed landscapes (Hunter 2005). Riparian zones, the bands of vegetation adjacent to streams and rivers, can serve a number of functions to preserve avian diversity. In North America, the integrity and biodiversity of riparian zones are threatened by urbanization (Burton et al. 2005). Although birds are more-strongly associated with riparian habitats in the xeric southwestern and western US (Brand et al. 2008), the riparian zones of mesic regions of the eastern US similarly provide habitat for more species than nearby 1Biology Department, Truman State University, Kirksville, MO 65301, USA. 2 Department of Biology, University of Nebraska Omaha, Omaha, NE 68182, USA. 3Department of Biological and Environmental Sciences, Le Moyne College, Syracuse, NY 13214, USA. *Corresponding author - cgearin@unomaha.edu. Manuscript Editor: Mark Laska Urban Naturalist C.J. Gearin and J.D. Luscier 2017 No. 14 2 upland environments (Knopf et al. 1988). Riparian zones are often studied for their importance as breeding habitat (Bennet et al. 2014, Brand et al. 2008, Rosenberg et al. 1987, Sanders and Edge 1998, Small et. al 2007). However, in fragmented landscapes, riparian zones are also important habitat corridors for bird migration (Skagen et al. 2005, Tubbs 1980) and metapopulation maintenance through gene flow (Gillies and St. Clair 2008, Tubbs 19808). Furthermore, birds are reliable indicators of riparian quality (Bryce and Hughes 2002, Rich 2002, Vaughan et al. 2007). In an urban–rural gradient in Ohio, Pennington et al. (2008) found that resident and Neotropical migrant species made use of urban riparian corridors. In the desert southwest, Rosenberg et al. (1987) concluded that suburban riparian zones contained higher densities of native birds than riparian cottonwood–willow habitats outside the city because this novel setting provides increased availability of food and surface water. In the Great Plains, Rich (2002) reported that riparian habitat had greater resident and migrant bird densities than the surrounding andscape. Skagen et al. (2005) found that, in arid environments, rather than following the shortest route, riparian-associated species might alter their migration routes in order to move through the highest-quality riparian habitat available. This behavioral shift highlights the importance of conserving and restoring riparian corridors in fragmented landscapes. Riparian zones generally provide the most-productive habitat in the Southwest, West, Midwest, Northwest, and Great Plains of the US (Knopf and Samson 1994); thus, it is possible that urban and suburban riparian zones could have similar positive impacts for Midwestern bird communities as Rosenberg et al. (1987) found them to have in a more arid environment. Suburban areas—developed zones with intermediate levels of fragmentation and disturbance between rural and urban zones—may play a unique role in which native biodiversity is lost after land clearing, but can increase again after restoration (Rosenberg et al. 1987). The structure of riparian vegetation can influence bird diversity. Sanders and Edge (1998) emphasized the importance of mesic-shrub vegetation in the west, and Tewksbury et al. (2002) found that the width of the riparian corridor is correlated with a greater density of riparian-associated species. Brand et al. (2008) reported that higher species-richness is correlated with heterogeneity in riparian vegetation structure. Therefore, while restoring the tree community of a deforested urban or suburban stream takes many years, short-term changes in understory height and structure could have significant impacts on migratory bird diversity. We hypothesized that bird diversity during fall migration would be greater at a restored suburban riparian zone with higher structural diversity in the vegetation than in an unrestored suburban site with less structural diversity. Therefore, our objective was to compare bird species richness and composition at a suburban stream undergoing riparian restoration, a suburban stream in a mowed city park, and a rural stream in an oak-hickory savanna community within a conservation area. We predicted that species associated with hydric habitats would be most abundant at the restored site and that species composition at the restored site would be more similar to that found at the rural site than the unimproved suburban site. Urban Naturalist 3 C.J. Gearin and J.D. Luscier 2017 No. 14 Field-site Description We conducted the study within the city of Kirksville in Adair County, MO, USA (Fig 1). In 2013, Kirksville had an estimated population of 17,577 (US Census Bureau 2015). The nearest large urban area is Columbia, MO, about 150 km away. Kirksville is in a generally rural region, but the density of human developments near our field sites was essentially suburban (moderate-intensity urban land-use) rather than rural (Rosenberg et al. 1987). From September to November 2014, temperatures in Adair County ranged from -13.9 °C to 32.8 °C, and total rainfall was 33.1 cm (National Oceanic and Atmospheric Administration 2014). We studied 3 low-order streams within the Upper Salt River watershed (Fig. 1), in the Dissected Till Plains ecoregion (Dames and Todd 2015). The Salt River flows into Mark Twain Lake before its confluence with the Mississippi River (Dames and Todd 2015). Within the watershed, 70% of the land is used for agriculture (most of which is cultivated for crops), 14% of the land is forested, and urban land-use is low (Dames and Todd 2015). Severe erosion due to Figure 1. Study area (image from Google Earth©, accessed 17 March 2015). Truman State University (TSU) and Rotary Park (ROPA) are located within the suburban area of Kirksville, Adair County, MO. Big Creek Conservation Area (BCCA) is an oak-savannah restoration habitat managed by the Missouri Department of Conservation. Urban Naturalist C.J. Gearin and J.D. Luscier 2017 No. 14 4 cultivation has caused increased turbidity and led to declines in aquatic diversity and abundance (Dames and Todd 2015). Non-point source nutrient runoff from cropland has led to low water-quality and fish kills, but point-source pollution has not had a significant impact on watershed water-quality (Dames and Todd 2015). Birds that use riparian areas depend on the food resources of the stream; thus, they are affected by water quality (Vaughan et al. 2007). Degradation of water quality may have affected birds independent of the vegetation variables we studied; however, effects on water quality likely operate at the watershed scale, rather than the stream scale and would be equal across our sites (Palmer et al. 2010). In April 2014, students at Truman State University (TSU) in Adair County, MO, USA, began a restoration project of the riparian zone of Bear Creek, a low-order stream on the periphery of a college campus of about 6000 students (Fig. 2). The campus is surrounded by suburban development, and the stream runs close to several 2-lane roads. The students sought to improve the habitat and decrease erosion rates of the stream banks by planting native trees and shrubs (A. Kersten and M. Woolbright, TSU, Kirksville, MO, USA; pers. comm.). Beginning in April 2014, the university instituted a no-mow zone of 1.5 m on either side of the creek. By September 2014, the riparian understory at TSU consisted primarily of Ambrosia trifida L. (Giant Ragweed), Festuca sp. (ornamental fescue gass), and several species in the Asteraceae. No data on the bird community of the TSU stretch of Bear Creek existed previous to the restoration project. However, a stretch of stream riparianzone within the same landscape, which had not been restored and was subject to regular mowing, provided a comparison to an unrestored suburban site. Rotary Park (ROPA), a public park 2.3 km from TSU, exists within a suburban landscape and contains a low-order stream, Steer Creek (Fig 2). In fall 2014, the park was mowed approximately once per month, to the edge of Steer Creek’s bank. The riparian zone is near Baltimore Street, a 2-lane business route. The understory at ROPA consisted Figure 2. Photographs of the 3 study sites taken in September 2014. TSU is a suburban stream undergoing vegetation restoration on the edge of a college campus. ROPA is a suburban stream in a mowed public park. BCCA is a rural stream in an oak-savannah conservation area. Photographs © C.J. Gearin. Urban Naturalist 5 C.J. Gearin and J.D. Luscier 2017 No. 14 primarily of fescue, A. artemisiifolia L. (Common Ragweed), Giant Ragweed, and several Asteraceae species. Big Creek flows through Big Creek Conservation Area (BCCA), an oaksavannah habitat situated in a rural landscape ~4 km west of TSU (Fig. 2). The Missouri Department of Conservation (MDC) maintains open fields of Alopecurus pratensis L. (Meadow Foxtail), Andropogon gerardii Vitman (Big Bluestem), and other grasses through an annual burning regime. The riparian understory at BCCA consisted primarily of Giant Ragweed, Foxtail, Chasmanthium latifolium Michx. (River-oats), and Rosa multiflora Thunb. ex Murr. (Multiflora Rose). Methods Bird surveys At each site, we selected two 100-m stretches of riparian zone, based on which stretches were the most representative of the riparian vegetation-management practices at each site (i.e., we selected stretches with relatively continuous vegetation, avoiding bridges and culverts.) The stretches were at least 50 m apart. At ROPA, only 2 such stretches were available; thus, we selected 2 stretches at the other sites. The 2 transects at each site were not fully independent because the treatments were pseudo-replicated (Hurlbert 1984). Pseudo-replication does not necessarily invalidate a field study (Heffner et al. 1996), but it limited the scale of our inferences to the study area. Furthermore, we used confidence intervals (CI) to evaluate differences among sites, which is an approach appropriate for related samples (Foody 2009). We conducted bird surveys at 3 sites in Adair County for 10 weeks during fall migration, from 1 September to 5 November 2014. To meet the assumption of closed populations for species-richness analysis, we divided the 10 weeks into 5 two-week periods, and considered each survey period as a closed community. We surveyed TSU 85 times, ROPA 81 times, and BCCA 62 times (i.e., TSU and ROPA were surveyed about 16 times per period and BCCA 12 times per period). We increased our survey effort at TSU and ROPA after finding imprecision (large CIs) in period 1 species-richness estimates. We evaluated species-accumulation curves (with 95% confidence intervals) across the 5 two-week periods as a means for understanding how our uneven survey design might have affected our ability to detect species as fall migration progressed at the 3 sites (Colwell et al. 2004). We stopped surveying BCCA on 31 October before the beginning of deer-hunting season with rifles. All surveys took place within 1.5 h of sunrise (i.e., between 0700 and 0830 on 25 September 2014, when sunrise was at 0700). We rotated the order in which we surveyed sites to limit the effect of the time of the survey (i.e., if the order was ROPA-BCCA-TSU one day, the order the next survey date might be BCCA-TSUROPA. We did not conduct surveys on days with steady rain or winds >20 km/h (Martin et al. 1997) because avian behavior and migration are strongly dependent on weather (Richardson 1978). We conducted line-transect surveys, which allow for the detection of birds sheltering in dense vegetation either before or after they are flushed by the moving Urban Naturalist C.J. Gearin and J.D. Luscier 2017 No. 14 6 observer (Anderson et al. 2015) and which are better suited than point counts to the non-territorial behavior of migratory birds (Wilson et al. 2000). For each survey, we walked the stretch slowly and at a steady pace, stopping only when necessary to visually identify a species. We identified any bird within 50 m of the stream-bank by vocalizations or field marks. We used a laser rangefinder to measure the distance to each detection. After completing a survey in 1 direction, we waited 2 minutes before beginning to survey in the other direction. Surveys were always on the same side of the creek. At TSU and ROPA, surveys were on the southwest side of the creek, and at BCCA, surveys were on the east side of the creek because the stream bank was less obstructed by Multiflora Rose on that side. Vegetation surveys For each stretch, we used a Robel pole (Robel et al. 1970) to estimate understory vegetation height 1.5 m from the stream bank every 20 m, recording the highest ring of the pole obscured, as observed from 5 m away and at 1 m height. At this same point, we performed ocular estimation of canopy cover, and scored canopy cover according to the scale: 1 = 0–25% cover, 2 = 26–50%, 3 = 51–75%, and 4 = 76–100%. We estimated understory height and canopy cover in the middle of fall migration, during period 3 (28 September–11 October), and assumed these estimates to be representative of all 5 periods. Although riparian understory vegetation was limited to a band within a few meters of the stream, the riparian tree community extended further. To estimate tree density per ha, we surveyed 100-m long, 5-m wide transects extending to both sides of the stream and crossing the stream every 20 m, beginning at the endpoints of the bird-survey transects (adapted from Heitke et al. 2008). Thus, each stretch had 5 transects, and we averaged the 10 density measurements for each site. We counted every tree with >50% of its trunk within the transect belt. We classified trees 8–23 cm diameter at breast height (DBH) as small trees, and trees >23 cm DBH as large trees (adapted from Martin et al. 1997). If transects overlapped with each other due to curvature of the stream, we counted trees occurring in 2 transects both times. Analyses Species richness. For each survey period and each site, we calculated species richness estimates from detection/non-detection data using program SPECRICH2 (Hines 2011). SPECRICH2 uses species-accumulation curves to estimate the total number of species at a site (Hines 2011). Habitat-association assemblages. Estimates of species richness (the number of species at a survey site) are informative of bird diversity, but they do not reveal important differences in species’ life histories (Blair 2014, Wiens 1989). Dividing bird communities into their component habitat-association assemblages can allow assessment of habitat quality (Bryce and Hughes 2002). Calculating the percent of total detections accounted for by detections of a specific group of species at a site is an analysis that allows for examination of differences in habitat-association assemblages within bird communities, and is relatively free of mathematical bias (Weller 1999). Urban Naturalist 7 C.J. Gearin and J.D. Luscier 2017 No. 14 We used a nested categorization scheme to analyze the relative abundance of species assemblages within and among migratory communities. First, we grouped species into hydric-habitat affinity assemblages, using habitats reviewed in Poole (2005) for each species encountered (Skagen et al. 2005). We classified species as hydric-associated (H; species associated with wet areas, either wetland areas or riparian forests), non-hydric (N; species associated with drier habitats), or facultative (F; generalist species that use wet and dry habitats in proportion to their availability). We used quantitative data reviewed in Poole (2005) for fall and/or spring migration when it was available, and otherwise used breeding-season data. We found no quantitative studies on habitat preferences during breeding or migration for Geothlypis trichas (L.) (Common Yellowthroat), but observations strongly suggest that this species is associated with wet habitats (Poole 2005). We were conservative in categorizing species, using the facultative category whenever the data did not indicate a clear primary-habitat association. We excluded birds of prey from this analysis. Next, we sub-grouped species according to their primary-habitat associations as reported in Poole (2005). We divided the H species into wetland area (WET) species and riparian forest (RIP) species; and N species into urban (URB), open habitat (OPE), and upland woodland (WOO) species. We did not subdivide F species, but created an identical facultative (FAC) category for the habitat-generalist species. To compute relative abundances for the habitat-association assemblages, we calculated each species’ relative abundance at each site and each survey period as the number of detections for a species divided by the total detections at the given site for the given survey period. For example, Picoides pubescens L. (Downy Woodpecker) made up 12 out of 59 detections at BCCA during period 1, and, thus, had a relative abundance of 20.3%. We then summed the relative abundances of species in each of the hydric-affinity assemblages and each of the primary habitat-association assemblages to calculate the relative abundance for each assemblage. Averaging across survey periods for each site allowed us to estimate the variance in relative abundance for each assemblage. We used ordination via non-metric multidimensional scaling (NMDS) in program R (R Core Team 2015) to evaluate the relationship of bird detections of all assemblages across the 3 sites. This community-ordination technique allows for the measure of dissimilarity across species distributions within a given area (Oksanen 2016), and thus provides us with a graphical representation of species composition at each site. Species scaling is based on weights connected to the numbers of detections at each site. Shorter distances among species in the ordination plot indicated community-composition patterns. We based significance on non-overlapping CIs: if the CI of the difference included zero, we interpreted the difference as biologically insignificant (Foody 2009, Skagen et al. 2005). We calculated the variance of the difference between 2 estimates using the variance-sum law: var(Ŝ1 - Ŝ2) = var(Ŝ1) + var(Ŝ2) - 2cov(Ŝ1, Ŝ2), Urban Naturalist C.J. Gearin and J.D. Luscier 2017 No. 14 8 where Ŝ is the estimate, var is the variance, and cov is the covariance between 2 samples. When estimates were independent, we assumed that covariance was insignificant and omitted the covariance term, giving a more conservative estimate of the difference’s variance. After calculating the variance of the difference, we calculated 99%, 95%, and 90% confidence intervals to estimate levels of significance (Skagen et al. 2005). Results Migratory bird communities During fall 2014, we made 985 bird detections of 47 species (Table 1). Speciesarea curves across the 5 two-week intervals were relatively flat for TSU and ROPA, but showed a steady increase for BCCA as the fall migratory period progressed (Fig. 3). In all cases, the patterns plateaued, which indicated that our survey effort was adequate to detect species across the 5 periods. Increasing survey effort at TSU and ROPA did not decrease later CIs, which suggested that the migratory communities at those sites likely had high intrinsic variability. BCCA had smaller CIs, perhaps because of higher densities of resident birds within the less-disturbed rural landscape. Species richness. At TSU, the average species-richness estimate ± 95% CI was 10.6 ± 6.2 species. ROPA and BCCA had average species-richness estimates of 12.2 ± 5.41 and 18.2 ± 5.9, respectively. Within sites, species-richness estimates did not show any significant trends throughout the 5 survey periods (Fig . 4). Although BCCA had transiently higher species-richness than the other 2 sites, species richness was generally similar among sites (Table 2). BCCA had higher species- richness than TSU in the first 3 time periods, and higher species-richness than ROPA during period 5 (Table 2). Other differences were not significant, perhaps due to the large confidence intervals of species-richness estimates for TSU and ROPA. Species composition. At TSU, we most frequently observed Passer domesticus (House Sparrow), Common Yellowthroat, Turdus migratorius (American Robin), and Agelaius phoeniceus (Red-winged Blackbird) (Table 1). At ROPA, Figure 3. Species-area curves (95% confidence intervals) of bird detections for TSU, ROPA, and BCCA during fall migration 2014. Urban Naturalist 9 C.J. Gearin and J.D. Luscier 2017 No. 14 Table 1. Wetland-affinity categories, primary-habitat associations, and number of detections for each species during fall 2014 at the 3 survey sites for 47 species detected (n = 985 detections). H = species associated with wet areas, either wetland areas or riparian forests; N = species associated with drier habitats; F = generalist species that use wet and dry habitats; and BP = birds of prey (excluded from the analysis). WET = wetland, WOO = upland woodland, OPE = open habitat, FAC = facultative (habitat generalist species), RIP = riparian forest, and URB = urban. TSU = Truman State University, ROPA = Rotary Park, and BCCA = Big Creek Conservation Area. Hydric-habitat Primary-habitat Scientific name Common name assn association TSU ROPA BCCA Ardea herodias L. Great Blue Heron H WET 0 0 1 Scolopax minor Gmelin American Woodcock N WOO 0 0 2 Zenaida macroura L. Mourning Dove N OPE 2 0 0 Bubo virginianus Gmelin Great Horned Owl BP BP 0 0 1 Megaceryle alcyon L. Belted Kingfisher H WET 0 0 5 Melanerpes carolinus L. Red-bellied Woodpecker F FAC 0 1 5 Dryobates pubescens L. Downy Woodpecker H RIP 1 14 35 Colaptes auratus L. Northern Flicker N OPE 0 8 16 Hylatomus pileatus L. Pileated Woodpecker H RIP 0 0 4 Falco sparverius L. American Kestrel BP BP 1 0 0 Contopus virens L. Eastern Wood-Pewee F FAC 0 4 0 Sayornis phoebe Latham Eastern Phoebe F FAC 0 7 1 Vireo olivaceus L. Red-Eyed Vireo N WOO 0 0 3 Cyanocitta cristata L. Blue Jay F FAC 5 31 33 Poecile atricapillus L. Black-capped Chickadee F FAC 0 0 17 Baeolophus bicolor L. Tufted Titmouse N WOO 0 0 3 Sitta carolinensis Latham White-breasted Nuthatch H RIP 0 11 15 Troglodytes aedon Vieillot House Wren F FAC 0 0 1 Cistothorus palustris Wilson Marsh Wren H WET 4 0 0 Thryothorus ludovicianus Latham Carolina Wren F FAC 4 0 2 Polioptila caerulea L. Blue-gray Gnatcatcher F FAC 0 0 1 Regulus calendula L. Ruby-crowned Kinglet F FAC 0 0 4 Urban Naturalist C.J. Gearin and J.D. Luscier 2017 No. 14 10 Table 1, continued Hydric-habitat Primary-habitat Scientific name Common name assn association TSU ROPA BCCA Sialia sialis L. Eastern Bluebird N OPE 0 0 7 Turdus migratorius L. American Robin F FAC 23 199 33 Dumetella carolinensis L. Gray Catbird F FAC 0 0 17 Bombycilla cedrorum Vieillot Cedar Waxwing F FAC 6 18 0 Sturnus vulgaris L. European Starling N URB 0 67 0 Parkesia noveboracensis Gmelin Northern Waterthrush H WET 0 0 2 Mniotilta varia L. Black-and-white Warbler F FAC 0 1 0 Leiothlypis ruficapilla Wilson Nashville Warbler F FAC 0 0 1 Geothlypis trichas L. Common Yellowthroat H WET 26 0 3 Setophaga magnolia Wilson Magnolia Warbler N WOO 0 0 1 Setophaga petechia L. Yellow Warbler H WET 3 3 2 Setophaga coronata L. Yellow-rumped Warbler N WOO 3 37 0 Cardellina pusilla Wilson Wilson's Warbler H RIP 0 0 1 Pipilo erythrophthalmus L. Eastern Towhee F FAC 0 0 1 Spizella passerina Bechstein Chipping Sparrow F FAC 11 2 0 Melospiza melodia Wilson Song Sparrow F FAC 20 0 1 Melospiza georgiana Latham Swamp Sparrow H WET 6 0 0 Zonotrichia albicollis Gmelin White-throated Sparrow F FAC 3 0 23 Junco hyemalis L. Dark-eyed Junco F FAC 1 1 13 Cardinalis cardinalis L. Northern Cardinal F FAC 14 10 19 Agelaius phoeniceus L. Red-winged Blackbird H WET 22 2 0 Quiscalus quiscula L. Common Grackle F FAC 0 5 0 Haemorhous mexicanus Müller House Finch N URB 7 0 3 Spinus tristis L. American Goldfinch N OPE 9 1 40 Passer domesticus L. House Sparrow N URB 76 0 0 Urban Naturalist 11 C.J. Gearin and J.D. Luscier 2017 No. 14 Figure 4. Species-richness estimates with 95% CI for the 3 survey sites during each survey period, from 1 September to 7 November 2014. We calculated the estimates using program SPECRICH2 (Hines 2011). Total survey sample sizes: TSU = 85, ROPA = 81, BCCA = 62. Table 2. Comparisons of species-richness estimates among the 3 survey sites. We calculated estimates using program SPECRICH2 and presence/absence data for each survey period. Number of surveys for each site: TSU: n = 85; ROPA: n= 81; BCCA: n = 62. *denotes statistically significant differences (CI not including zero). CI at highest significance level Comparison (greater–lesser) Difference (if difference insignificant, then 95% CI reported) Period 1 ROPA vs. TSU 5.0 -3.8–13.8 (95% CI) BCCA vs. TSU 14.0 0.5–27.5 (99% CI) * BCCA vs. ROPA 9.0 -1.8–19.8 (95% CI) Period 2 ROPA vs. TSU 6.0 -0.7–12.7 (95% CI) BCCA vs. TSU 8.0 0.2–15.8 (95% CI) * BCCA vs. ROPA 9.0 -1.8–19.8 (95% CI) Period 3 TSU vs. ROPA 2.0 -9.1–13.1 (95% CI) BCCA vs. TSU 8.0 0.1–15.9 (95% CI) * BCCA vs. ROPA 10.0 -1.5–21.5 (95% CI) Period 4 ROPA vs. TSU 1.0 -5.7–7.7 (95% CI) BCCA vs. TSU 4.0 -3.0–11.0 (95% CI) BCCA vs. ROPA 3.0 -4.0–10.0 (95% CI) Period 5 TSU vs. ROPA 2.0 -7.1–11.1 (95% CI) BCCA vs. TSU 4.0 -5.4–13.4 (95% CI) BCCA vs. ROPA 6.0 0.1–11.9 (95% CI) * Urban Naturalist C.J. Gearin and J.D. Luscier 2017 No. 14 12 American Robins, Sturnus vulgaris (European Starling), and Setophaga coronata (Yellow-rumped Warbler) were most frequently observed. At BCCA, Spinus tristis (American Goldfinch), Downy Woodpeckers, Cyanocitta cristata (Blue Jay), and American Robins were most frequently observed. The NMDS analysis revealed that 58% (7 species) of all H species detected were closely associated with BCCA (Fig. 5). TSU had 2 H species and there were no H species associated with ROPA. Assemblage abundance. In comparing hydric-habitat affinity groups, survey sites differed significantly in the relative abundance of H species, and less significantly in abundances of F and N species (Fig. 6). Seven species did not have a close association with any of the 3 sites. TSU had higher relative abundance of H species than ROPA and similar abundance of H species compared to BCCA (Table 3). ROPA had lower H-species abundance than the other sites and higher-F species abundance than TSU. The differences in abundance between TSU and BCCA were not significant for both F and N species. Figure 5. Non-metric multidimensional scaling (NMDS) ordination of detections by species (NMDS1) among the 3 locations (NMDS2: Truman State University (TSU), Rotary Park (ROPA), and Big Creek Conservation Area (BCCA). Groups 1, 2, and 3 represent species that had similar scaling values. Urban Naturalist 13 C.J. Gearin and J.D. Luscier 2017 No. 14 At TSU, the 3 primary-habitat–association groups with the highest-average relative abundance (95% CI) were FAC, URB, and WET (Fig. 7). TSU had the greatest relative abundance of URB species. TSU and ROPA had similarly low averages for OPE relative abundance, at 4.4 ± 2.7% and 3.7 ± 3.6, respectively. ROPA’s dominant groups were FAC, URB, and WOO; however, the 95% CI for WOO included zero. BCCA’s community consisted mainly of FAC, RIP, and OPE. Figure 6. Average relative abundances for wetland (H), facultative (F), and non-wetland (N) hydric habitat affinity categories at the 3 survey sites for fall 2014. n = 983 detections. Averages are shown with their 95% CI. Birds of prey were excluded from this analysis. Table 3. Comparisons of relative abundances of hydric-habitat affinity groups among the 3 survey sites (n = 983 detections). *denotes statistically significant differences (CI not including zero. CI at highest significance level Comparison (greater–lesser) Difference (if difference insignificant, then 95% CI reported) Hydric TSU vs. ROPA 16.4% 4.4–28.3% (99% CI)* TSU vs. BCCA 0.7% -10.4–11.7% (95% CI) BCCA vs. ROPA 15.7% 5.9–25.5% (99% CI)* Facultative ROPA vs. TSU 21.4% 2.6–40.2% (95% CI)* BCCA vs. TSU 12.3% -4.3–28.8% (95% CI) ROPA vs BCCA 9.1% -5.4–23.7% (95% CI) Non-Hydric TSU vs. ROPA 5.0% -13.5–23.6% (95% CI) TSU vs. BCCA 11.6% -5.0–28.1% (95% CI) ROPA vs. BCCA 6.5% -7.5–20.5% (95% CI) Urban Naturalist C.J. Gearin and J.D. Luscier 2017 No. 14 14 The sites differed significantly in the relative abundance of wetland and riparian forest species (Table 4). TSU had higher relative-abundance of WET species than ROPA and BCCA, while BCCA had the highest abundances of RIP species. ROPA had higher abundances of RIP species than TSU. The 2 suburban sites, TSU and ROPA, had significantly greater relative abundance of URB species than BCCA. TSU had a greater abundance of URB species than ROPA. BCCA had a greater abundance of OPE species than TSU and ROPA. Riparian vegetation communities TSU had the greatest streamside-vegetation understory height, with an average height ± SE of 42.9 ± 6.3 cm (Table 5) and ROPA had the lowest, at 15.7 ± 1.4 cm. Although BCCA’s average understory height was less than TSU’s, it had the greatest variance, at 17.3 ± 10.0 cm; its vegetation structure was the most heterogeneous in height relative to the other 2 sites. Streamside-canopy cover was similar among sites, all scoring within the 0–25% range on average. BCCA had the highest treedensity, at 328 ± 104 per ha. ROPA and BCCA had similar densities of large trees, with 46 ± 24.1 and 70 ± 32.9 per ha, respectively. TSU’s vegetation understory was at least 18.7 cm higher than ROPA’s, and 10.1 cm higher than BCCA (Table 6). BCCA and ROPA did not differ significantly in understory height. The densities of the riparian-tree communities within 50 m of the stretch differed significantly among sites. BCCA had at least 22.0 more large Figure 7. Average relative abundances for wetland (WET), riparian (RIP), open (OPE), woodland (WOO), urban (URB), and facultative (FAC) at the 3 survey sites for fall 2014. n = 983 detections. Averages are shown with their 95% CI. Birds of prey were excluded from this analysis. Urban Naturalist 15 C.J. Gearin and J.D. Luscier 2017 No. 14 trees per ha than TSU, and ROPA had at least 6.5 more large trees per ha than TSU. ROPA and BCCA did not differ significantly in large-tree density. BCCA had at least 153.1 more small-trees per ha than TSU and at least 136.3 more than ROPA. ROPA and TSU did not differ significantly in small-tree densities. ROPA had at least 5.4 more total trees per ha than TSU. BCCA had at least 172.3 more total trees per ha than TSU and 117.3 more total trees per ha than ROPA. Discussion One year after the start of the riparian restoration project, the fall migratorybird community at a suburban restored site was more similar to a site with a mature riparian forest than it was to an unrestored site. This result suggests that in the short term, the higher riparian understory resulting from the no-mow zone at Table 4. Comparisons of relative abundance of 6 primary habitat-association groups among the 3 survey sites (n = 983 detections). *denotes statistically significant differences (CI not including zero). CI at highest significance level Comparison (greater–lesser) Difference (if difference insignificant, then 95% CI reported) Wetland TSU vs. ROPA 23.2% 11.6–34.7% (99% CI)* TSU vs. BCCA 19.9% 8.2–31.6% (99% CI)* BCCA vs. ROPA 3.3% 0.5 –6.0% (99% CI)* Riparian ROPA vs. TSU 6.8% 3.5–10.1% (99% CI)* BCCA vs. TSU 19.2% 10.6–27.8% (99% CI)* BCCA vs. ROPA 12.4% 3.2–21.6% (99% CI)* Open TSU vs. ROPA 0.7% -3.8–5.2 (95% CI) BCCA vs. TSU 14.7% 1.5–27.9 (99% CI)* BCCA vs. ROPA 15.4% 1.9–29.0% (99% CI)* Woodland ROPA vs. TSU 10.3% 0.2–20.4% (90% CI)* TSU vs. BCCA 2.1% -2.1–6.3% (95% CI) ROPA vs. BCCA 8.2% -4.4–20.8% (95% CI) Urban TSU vs. ROPA 14.7% 1.0–28.4% (95% CI)* TSU vs. BCCA 28.4% 11.5–45.3% (99% CI)* ROPA vs. BCCA 13.7% 6.6–20.9% (99% CI)* Table 5. Summary statistics for riparian-vegetation communities at the 3 survey sites during fall 2014. Values reported with 95% CI. Understory Average canopy Large-tree density Small-tree density Total tree density Site height (cm) -cover score (# per ha) (# per ha) (# per ha) TSU 42.9 ± 6.3 1.2 ± 0.3 14 ± 8.4 4 ± 5.2 18 ± 9.1 ROPA 15.7 ± 1.4 1.2 ± 0.3 46 ± 24.1 18 ± 18.9 64 ± 39.5 BCCA 17.3 ± 10.0 1.0 ± 0.0 70 ± 32.9 258 ± 76.5 328 ± 104 Urban Naturalist C.J. Gearin and J.D. Luscier 2017 No. 14 16 the suburban restored site had a biologically significant impact on hydric-associated birds migrating through the corridor, despite the fact that the unrestored site had greater total tree density than the restored suburban site. Although species-richness estimates were generally similar among sites throughout the fall migration in 2014, habitat-affinity assemblages differed significantly in relative abundance. The suburban restored site and the conservation area had similar abundances of hydric species, which were significantly greater than the unrestored site. However, at the restored suburban site, hydric species were mostly wetland-associated species, as opposed to species with an affinity for mature riparian forest. For example, Swamp Sparrows and Common Yellowthroats, 2 species we observed relatively frequently at the restored suburban site, are associated with swampy wetlands rather than riparian forests (Poole 2005). In contrast, hydric species at the conservation area and the unrestored site were mainly riparian species such as Downy Woodpeckers and White-breasted Nuthatches (Poole 2005), probably because both these sites had greater tree-densities than the restored suburban site. Thus, while it takes years to restore trees in a riparian buffer, a short-term impact on wetland species abundance is possible through the establishment of nomow zones in suburban riparian areas. Our results corroborate other evidence in the literature that riparian habitats allow birds to migrate through fragmented landscapes. Pennington et al. (2008) found that Neotropical migrants used urban riparian corridors in Ohio during spring migration. Tree cover was associated with migrant density (Pennington et al. 2008), Table 6. Comparisons of riparian-vegetation variables among the 3 survey sites during fall 2014. *denotes statistically significant differences (confidence intervals not including zero). CI at highest significance level Comparison (greater–lesser) Difference (if difference insignificant, then 95% CI reported) Vegetation understory (cm) TSU vs. ROPA 27.2 18.7–35.7 (99% CI)* TSU vs. BCCA 25.7 10.1–41.2 (99% CI)* BCCA vs. ROPA 1.5 -8.6–11.7 (95% CI) Canopy cover TSU vs. ROPA 0.0 -0.4–0.4 (95% CI) TSU vs. BCCA 0.2 -0.1–0.5 (95% CI) ROPA vs. BCCA 0.2 -0.1–0.5 (95% CI) Large trees (# per ha) ROPA vs. TSU 32.0 6.5–57.5 (95% CI)* BCCA vs. TSU 56.0 22.0–90.0 (95% CI)* BCCA vs. ROPA 24.0 -16.8–64.8 (95% CI) Small trees (# per ha) ROPA vs. TSU 14.0 -5.6–33.6 (95% CI) BCCA vs. TSU 254.0 153.1–354.9 (99% CI)* BCCA vs. ROPA 240.0 136.3–343.7 (99% CI)* All Trees (# per ha) ROPA vs. TSU 46.0 5.4–86.6 (95% CI)* BCCA vs. TSU 310.0 172.3–447.7 (99% CI)* BCCA vs. ROPA 264.0 117.3–410.7 (99% CI)* Urban Naturalist 17 C.J. Gearin and J.D. Luscier 2017 No. 14 which agreed with our finding that riparian-species abundance appeared dependent on tree density. However, landscape-level effects could explain some of our results (Bennet et al. 2014, Saab 1999). Species richness could have been higher at the conservation area because of its location in a rural matrix of relatively suitable habitats. In addition, the use of riparian zones as migratory corridors could depend on their relationship to regional migratory pathways and ancestral migration routes (Delmore et al. 2012). A potential concern for riparian restoration is the risk of creating ecological traps. Nest predation increases with edge density (Weldon and Haddad 2005). Due to the perimeter–area relationship, riparian corridors have a high edge-density. We observed Felis catus L. (Domestic cat) at the restored suburban site ([WHOSE?], pers. observ.). Depredation of birds by Domestic Cats can devastate bird populations (Woods et al. 2003). Increasing the width of the riparian belt could mitigate predation pressures (Weldon and Haddad 2005). We emphasize that only considering species richness could lead to a naïve perspective of bird diversity. Assessment of other factors is especially important given the phenomenon of increased species-richness at urban parks with moderate disturbance regimes, where there is an increase in generalist-species diversity but a decrease in native-species diversity (Blair 1996). Overemphasis on species-richness could confound investigations of the effects of urban restoration. An analysis of species composition allows ecologists to consider informative differences in species’ life histories that may provide more insights into the bird diversity that different habitats support. Management implications In the short-term, riparian no-mow zones in suburban landscapes can increase wetland quality in the Kirksville area. Managers should consider the potential importance of suburban and urban riparian zones as migratory corridors. No-mow zones are generally beneficial to bird diversity within fragmented Midwestern US landscapes (Frawley and Best 1991), and vegetation structural heterogeneity is correlated with greater bird-species richness (Brand et al. 2008). Restoring both the streamside understory and tree community could have the most impact on increasing both wetland and riparian-species abundance and overall bird diversity (Bennet et al. 2014). Managers can evaluate the habitat quality of restored riparian zones through bird surveys. Acknowledgments We thank the Missouri Department of Conservation; the city of Kirksville, MO, USA; the Truman State University Biology Department; Truman Facilities staff; and C. Cunningham, J. Gering, P. Goldman, A. Kersten, B. Thornton, and M. Woolbright. Literature Cited Anderson, A.S., T.A. Marques, L.P. Shoo, and S.E. Williams. 2015. Detectability in audio– visual surveys of tropical rainforest birds: The influence of species, weather, and habitat characteristics. PloS ONE 10:e0128464. Urban Naturalist C.J. Gearin and J.D. Luscier 2017 No. 14 18 Bennett, A.F., D.G. Nimmo, and J.Q. Radford. 2014. Riparian vegetation has disproportionate benefits for landscape-scale conservation of woodland birds in highly modified environments. Journal of Applied Ecology 51:514–523. Blair, R.B. 2014. Land use and avian-species diversity along an urban gradient. Ecological Applications 6:506–519. Brand, L.A., G.C. White, and B.R. Noon. 2008. Factors influencing species richness and community composition of breeding birds in a desert riparian corridor. The Condor 110:199–210. Bryce, S.A., and R.M. Hughes. 2002. Development of a bird integrity index: Using bird assemblages as indicators of riparian condition. Environmental Management 30:294–310. Burton, M.L., L.J. Samuelson, and S. Pan. 2005. Riparian woody-plant diversity and forest structure along an urban–rural gradient. Urban Ecosystems 8:93–106. Colwell, R.K., C.X. Mao, and J. Chang. 2004. Interpolating, extrapolating, and comparing incidence-based species-accumulation curves. Ecology 85:2717–2727. Dames, H.R., and B. Todd. 2015. Salt River watershed inventory and assessment. Missouri Department of Conservation. Kirksville, MO, USA. Delmore, K.E., J.W. Fox, and D.E. Irwin. 2012. Dramatic intraspecific differences in migratory routes, stopover sites, and wintering areas revealed using light-level geolocators. Proceedings of the Royal Society B 279:4582–4589. Foody, G.M. 2009. Classification-accuracy comparison: Hypothesis tests and the use of confidence intervals in evaluations of difference, equivalence, and non-inferiority. Remote Sensing of Environment 113:1658–1663. Frawley, B.J., and L.B. Best. 1991. Effects of mowing on breeding-bird abundance and species composition in alfalfa fields. Wildlife Society Bulletin 19:135–142. Gillies, C.S., and C.C. St. Clair. 2008. Riparian corridors enhance movement of a forestspecialist bird in fragmented tropical forest. Proceedings of the National Academy of Sciences of the United States of America 105:19,774–19,779. Heffner, R.A., M.J. Butler, and C.K. Reilly. 1996. Pseudoreplication revisited. Ecology 77:2558–2562. Heitke, J.D., E.J. Archer, D.D. Dugaw, B.A. Bouwes, E.A. Archer, R.C. Henderson, and J.L. Kershner. 2008. Effectiveness monitoring for streams and riparian areas: Sampling protocol for stream-channel attributes. PACFISH/INFISH Biological Opinion (PIBO) Effectiveness Monitoring Program, Multi-federal Agency Monitoring Program, Logan, UT. Available online at http://www.fs.fed.us/biology/fishecology/emp. Accessed 20 August 2014. Hines, J.E. 2011. SPECRICH2 Software to estimate the total number of species from species presence–absence data on multiple sample-sites or occasions. USGS-PWRC. Available online at http://www.mbr-pwrc.usgs.gov/software/specrich2.shtml. Accessed 28 August 2014. Hunter, M.L. 2005. A mesofilter conservation strategy to complement fine and coarse filters. Conservation Biology 19:1025–1029. Hurlbert, S.H. 1984. Pseudoreplication and the design of ecological field experiments. Ecological Monographs 54:187–211. Knopf, F.L., and F.B. Samson. 1994. Scale perspectives on avian diversity in western riparian ecosystems. Conservation Biology 8:669–676. Knopf, F.L., R.R. Johnson, T. Rich, F.B. Samson, and R.C. Szaro. 1988. Conservation of riparian ecosystems in the United States. The Wilson Bulletin 100:272–284. Urban Naturalist 19 C.J. Gearin and J.D. Luscier 2017 No. 14 Martin, T.E., C. Paine, C.J. Conway, W.M. Hochachka, P. Allen, and W. Jenkins. 1997. BBIRD field protocol. Montana Cooperative Wildlife Research Unit, University of Montana, Missoula, MT, USA. Available online at http://www.umt.edu/bbird/protocol/. Accessed 15 August 2014. Murphy, M.T. 2003. Avian population trends within the evolving agricultural landscape of eastern and central United States. The Auk 120:20–34. National Oceanic and Atmospheric Administration. 2014. Annual climatological summary for Kirskville, MO. Available online at http://www.webcitation.org/6Y10h3vOC. Accessed 23 April 2015. Oksanen, J. 2016. Vegan: An introduction to ordination. Available online at https://cran.rproject. org/web/packages/vegan/vignettes/intro-vegan.pdf. Accessed 15 September 2016. Palmer, M.A., H.L. Menninger, and E. Bernhardt. 2010. River restoration, habitat heterogeneity, and biodiversity: A failure of theory or practice? Freshwater Biology 55:205–222. Pennington, D.N., J. Hansel, R.B. Blair. 2008. The conservation value of urban riparian areas for land-birds during spring migration: Land cover, scale, and vegetation effects. Biological Conservation 141:1235–1248. Poole, A. (Ed.). 2005. The Birds of North America Online. Cornell Laboratory of Ornithology, Ithaca, NY, USA. Available online at http://bna.birds.cornell.edu/BNA/. Accessed 20 August 2014. R Core Team 2015. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available online at https://www.R-project. org/. Accessed 1 August 2016. Rich, T.D., 2002. Using breeding land birds in the assessment of western riparian systems. Wildlife Society Bulletin 30:1128–1139. Richardson, W.J. 1978. Timing and amount of bird migration in relation to weather: A review. Oikos 30:224–272. Robel, R.J., J.N. Briggs, A.D. Dayton, and L.C. Hulbert. 1970. Relationships between visual-obstruction measurements and weight of grassland vegetation. Journal of Range Management 23:295–297. Rosenberg, K.V., S.B. Terrill. and G.H. Rosenberg. 1987. Value of suburban habitats to desert riparian birds. The Wilson Bulletin 99:642–654. Saab, V.A. 1999. Importance of spatial scale to habitat use by breeding birds in riparian forests: A hierarchical analysis. Ecological Applications 9:135–151. Sanders, T.A., and W.D. Edge. 1998. Breeding-bird community composition in relation to riparian vegetation structure in the western United States. Journal of Wildlife Management 62:461–473. Skagen, S.K., J.F. Kelly, C. Van Riper, R.L. Hutto, D.M. Finch, D.J. Krueper, and C.P. Melcher. 2005. Geography of spring land-bird migration through riparian habitats in southwestern North America. The Condor 107:212–227. Small, S.L., F.R. Thompson III, G.R. Geupel, and J. Faaborg. 2007. Spotted Towhee population dynamics in a riparian restoration context. The Condor 109:721–733. Tewksbury, J.J., A.E. Black, N. Nur, V.A. Saab, B.D. Logan, and D.S. Dobkin. 2002. Effects of anthropogenic fragmentation and livestock grazing on western riparian-bird communities. Studies in Avian Biology 25:158–202. Tubbs, A.A. 1980. Riparian bird communities of the Great Plains. Pp. 419–433, In R.M. DeGraff and N.G. Tilghman (Compilers). Proceedings of the workshop: Management of western forests and grasslands for nongame birds. US Forest Service General Technical Report INT-86. Intermountain Forestry and Range Experiment Station, Ogden, UT, USA. 535 pp. Urban Naturalist C.J. Gearin and J.D. Luscier 2017 No. 14 20 US Census Bureau. 2015. State and county quick facts. US Department of Commerce. Available online at http://quickfacts.census.gov/qfd/states/29/2939026.html. Accessed 20 April 2015. Vaughan, I.P., D.G. Noble, and S.J. Ormerod. 2007. Combining surveys of river habitats and river birds to appraise riverine hydromorphology. Freshwater Biology 52:2270–2284. Weldon, A.J., and N.M. Haddad. 2005. The effects of patch shape on Indigo Bntings: Evidence for an ecological trap. Ecology 86:1422–1431. Weller, M.W. 1999. Wetland birds: Habitat Resources and Conservation Implications. Cambridge University Press, Cambridge, UK. 288 pp. Wiens, J.A., 1989. Spatial scaling in ecology. Functional Ecology 3:385–397. Wilson, R.R., D.J. Twedt, and A.B. Elliott. 2000. Comparison of line transects and point counts for monitoring spring migration in forested wetlands. Journal of Field Ornithology 71:345–355. Woods, M., R.A. McDonald, and S. Harris. 2003. Predation of wildlife by Domestic Cats, Felis catus, in Great Britain. Mammal Review 33:174–188.