Blank et al. • WINTERING BIRD RESPONSE TO FALL MOWING 
61 
three passes. We communicated regularly and 
watched for birds moving within study sites so 
that individual birds were not counted twice. By 
using these methods, at least one observer walked 
within 10 m of all points in the study sites. 
Diefenbach et al. (2003) reported nearly 100% 
detection of breeding grassland birds within 25 m 
of observers, and Roberts and Schnell (2006) 
recommended that observers walk within 10 m of 
all points in fixed areas when calculating density 
of wintering grassland birds. Thus, we assumed 
100% detection during our surveys. One obser¬ 
vation of an American Kestrel ( Falco sparverius) 
observed foraging above a study site during a 
survey was included in the counts. 
Statistical Analyses. —We used three bird 
community metrics to compare bird use of mowed 
and unmowed buffers: total abundance, species 
richness, and total avian conservation value 
(TACV). The latter is an index used to assess 
the relative conservation value of different sites 
that incorporates the biological vulnerability and 
the regional importance of each species (Nuttle et 
al. 2003). We calculated TACV by multiplying 
each species’ abundance by its Partners in Flight 
conservation priority rank (Carter et al. 2000, 
Nuttle et al. 2003) for the Mid-Atlantic Bird 
Conservation Region (Partners in Flight 2008), 
and then summing the species-specific TACV 
scores within a site (Conover et al. 2007, 2009). 
We categorized each bird species as either a 
grassland or scrub-shrub species based on litera¬ 
ture of species assemblages (Askins 1993, Vick¬ 
ery et al. 1999, Hunter et al. 2001, Sauer et al. 
2008, Schlossberg and King 2008, Poole 2010). 
We calculated the mean of each bird commu¬ 
nity metric and species’ abundance across the 
three rounds of bird surveys, and used the means 
as response variables in statistical analyses. Bird 
and vegetation metrics were not normally distrib¬ 
uted within treatments, and we used generalized 
linear mixed models (GLMM) in Proc GLIMMIX 
(SAS Institute, Cary, NC, USA) to compare 
responses in mowed and unmowed treatments. 
We specified a Poisson distribution for models of 
bird metrics and either a log-normal or a Poisson 
distribution for models of vegetation metrics. We 
treated management type (mowed or unmowed) 
as a fixed factor, buffer as a random block (to 
account for the paired study sites), and grass type 
(cool- or warm-season) as a random factor. We 
included study site area as an offset in all bird 
models because study sites differed in area, and 
TABLE 1. Vegetation characteristics (mean ± SE) in 
mowed and unmowed buffers on the Eastern Shore of 
Maryland, winter 2007. 
Management type 
Vegetation-- 
characteristic Mowed Unmowed F P 
Vertical density 
5.5 
+ 
0.9 
21.9 
2.7 
115.4 
<0.001 
Maximum height, 
cm 
3.2 
+ 
0.1 
4.6 
0.1 
158.3 
<0.001 
Litter depth, cm 
4.7 
+ 
0.7 
4.4 
0.7 
0.1 
0.72 
Percent cover 
Grass 
3.2 
H- 
0.2 
3.6 
-i- 
0.2 
5.1 
0.045 
Forbs 
4.1 
+ 
2.1 
5.7 
3.0 
8.2 
0.016 
Trees 
0.1 
0.1 
0.6 
0.3 
4.0 
0.070 
Litter 
3.9 
0.4 
3.5 
-4- 
0.4 
3.7 
0.078 
Bare ground 
5.1 
± 
1.4 
2.9 
2.6 
3.5 
0.086 
included width as a covariate because buffer 
width influences bird communities (Best 2000, 
Clark and Reeder 2005, Blank et al. 2011). We 
only analyzed the species-specific responses of 
Savannah Sparrow ( Passerculus sandwichensis). 
Song Sparrow ( Melospiza melodia ), and White- 
throated Sparrow ( Zonotrichia albicollis) because 
we could not fit appropriate models to the 
distribution of other species due to a lack of 
detections in most study sites. We considered a 
test result statistically significant at P < 0.05. 
RESULTS 
Vertical vegetation density, maximum height, 
percent cover of grass, and percent cover of forbs 
were significantly greater in unmowed than in 
mowed buffers (Table 1). We detected 412 birds 
in buffers, of which 98% were in unmowed 
buffers. We observed five species in mowed 
buffers and 14 species in unmowed buffers. Eight 
species were grassland or scrub-shrub birds 
(Table 2) and constituted 92% of all detections. 
The Song Sparrow was the most abundant species 
(45% of detections), followed by Field Sparrow 
(Spizella pusilla ; 19%), and Savannah Sparrow 
(10%). Savannah Sparrow (F ltI2 = 6.36, P = 
0.027), Song Sparrow {F\^ 2 = 16.54, P = 0.001), 
and White-throated Sparrow (F U2 = 5.68, P = 
0.035) were all more abundant in unmowed than 
in mowed buffers. Total abundance, species 
richness, and TACV were all greater in unmowed 
than in mowed buffers (Table 3). 
DISCUSSION 
Wintering bird use of mowed buffers was less 
than in unmowed buffers. All bird community 
