Content uploaded by Erica Nol
Author content
All content in this area was uploaded by Erica Nol
Content may be subject to copyright.
Reproductive success of cavity-nesting birds in
partially harvested woodlots
Melissa A. Straus, Kata Bavrlic, Erica Nol, Dawn M. Burke, and Ken A. Elliott
Abstract: Cavity-nesting birds are dependent on large declining and dead trees that are frequently removed during partial
harvesting. We compared breeding densities, nest survival, nest site characteristics, food abundance, and reproductive param-
eters of six species of cavity-nesting birds in partially harvested and reference woodlots in southwestern Ontario, Canada.
Silvicultural practices significantly altered woodlot structure, with treatment-specific effects on bark arthropod biomass,
fledging dates for the Red-bellied Woodpecker (Melanerpes carolinus (Linnaeus, 1758)), and site suitability for the Yellow-
bellied Sapsucker (Sphyrapicus varius (Linnaeus, 1766)). Red-bellied Woodpecker, Downy Woodpecker (Picoides pubescens
(Linnaeus, 1766)), and Hairy Woodpecker (Picoides villosus (Linnaeus, 1766)) experienced lower breeding densities in re-
cently cut sites. Daily survival rates were generally greater for nests positioned higher up in large trees and for Northern
Flicker (Colaptes auratus (Linnaeus, 1758)) nests excavated in healthy and hard wood. Conversely, the Black-capped Chick-
adee (Poecile atricapillus (Linnaeus, 1766)) had higher daily survival rates in low, small trees (<10 cm diameter at breast
height) and sites with lower arthropod abundance. We conclude that although partial harvesting has the potential to decrease
cavity-nesting bird breeding densities, conscientious cavity tree retention during harvest may provide suitable nesting sites
that maintain high rates of nest success, regardless of the silvicultural treatments that we examined. However, further re-
search is required to monitor these trends beyond a single harvesting rotation.
Résumé : Les oiseaux nicheurs de cavités dépendent des gros arbres dépérissants ou morts qui sont souvent abattus lors de
coupes partielles. Nous avons comparé la densité de couples nicheurs, la survie des nids, les caractéristiques des sites de ni-
dification, l’abondance de nourriture et les paramètres reproducteurs de six espèces d’oiseaux nicheurs de cavités dans des
boisés ayant fait l’objet de coupe partielle et des boisés de référence du sud-ouest de l’Ontario, au Canada. Les pratiques
sylvicoles ont significativement altéré la structure des boisés, ce qui a eu des effets spécifiques aux traitements sur la bio-
masse d’arthropodes de l’écorce, les dates d’envol du pic à ventre roux (Melanerpes carolinus (Linnaeus, 1758)) et la qua-
lité des sites de nidification du pic maculé (Sphyrapicus varius (Linnaeus, 1766)). La densité de couples nicheurs de pics à
ventre roux, de pics mineurs (Picoides pubescens (Linnaeus, 1766)) et des pics chevelus (Picoides villosus (Linnaeus,
1766)) a diminué dans les sites coupés récemment. Le taux de survie quotidienne était en général meilleur pour les nids si-
tués haut dans les grands arbres et en particulier pour les pics flamboyants (Colaptes auratus (Linnaeus, 1758)) nichant
dans du bois sain et dur. À l’inverse, le taux de survie quotidienne de la mésange à tête noire (Poecile atricapillus (Lin-
naeus, 1766)) était meilleur pour les nids situés dans de petits arbres (diamètre à hauteur de poitrine <10 cm) et dans les si-
tes où l’abondance des arthropodes est faible. Bien que la coupe partielle puisse éventuellement réduire la densité des
oiseaux nicheurs de cavités, nous concluons qu’un effort consciencieux de rétention d’arbres à cavités lors de la coupe peut
fournir des sites de nidification de qualité qui maintiendront de hauts taux de succès de nidification, sans égard aux traite-
ments sylvicoles que nous avons étudiés. Toutefois, des recherches supplémentaires sont requises pour effectuer un suivi de
ces tendances au-delà d’une première période de coupe.
[Traduit par la Rédaction]
Introduction
Anthropogenic activities have resulted in fragmentation,
forest degradation, and severe loss of forest cover throughout
deciduous forests of Europe and North America (Mönkkönen
and Welsh 1994). To combat further forest fragmentation and
degradation, silvicultural guidelines for remaining and recov-
ering upland tolerant hardwoods recommend the use of par-
tial harvesting and uneven-aged management as well as the
preservation of specific habitat features for wildlife (Franklin
et al. 1997). Cavity trees are critical for roosting, nesting,
denning, food storage, and foraging habitat for a variety of
mammals, birds, reptiles, and amphibians (Aitken and Martin
2004). Cavity retention regulations are often integrated with
forest management plans to conserve habitat for cavity-de-
pendent species (Franklin et al. 1997; Hutto 2006). However,
Received 6 April 2010. Accepted 22 November 2010. Published at www.nrcresearchpress.com/cjfr on 20 April 2011.
M.A. Straus and K. Bavrlic. Environmental and Life Sciences Graduate Program, Environmental Science Centre, 1600 West Bank Drive,
Trent University, Peterborough, ON K9J 7B8, Canada.
E. Nol. Biology Department, Environmental Science Centre, 1600 West Bank Drive, Trent University, Peterborough, ON K9J 7B8,
Canada.
D.M. Burke and K.A. Elliott. Ontario Ministry of Natural Resources, 659 Exeter Road, London, ON N6E 1L3, Canada.
Corresponding author: M.A. Straus (e-mail: melissastraus@trentu.ca).
1004
Can. J. For. Res. 41: 1004–1017 (2011) doi:10.1139/X11-012 Published by NRC Research Press
Can. J. For. Res. Downloaded from www.nrcresearchpress.com by Dalian Nationalities University on 06/05/13
For personal use only.
guidelines and bylaws differ across geographic regions and
may not always provide appropriate habitat safeguards (Hutto
2006).
Cavity-nesting birds often select nesting and foraging sites
in large declining or dead trees (Raphael and White 1984).
These trees are also targeted during selection system harvests
to improve timber quality and growth (Costello et al. 2000)
and ensure worker safety. As cavities may be limited even in
the absence of timber extraction (Brawn and Balda 1988), sil-
viculture can contribute to the removal of a resource that is
already in short supply (Vanderwel et al. 2006). Although
the loss of cavity trees can result in severe declines in breed-
ing bird density, the use of density as the sole measure of
habitat suitability is potentially problematic (Van Horne
1983), particularly if high-density sites are ecological traps
(i.e., experience poor reproductive success and (or) survival).
Cavity nesters typically experience high rates of nest suc-
cess, but predation remains the primary cause of failure (Li
and Martin 1991). Changes to forest structure induced by sil-
viculture may influence the predator community (Simard and
Fryxell 2003) and therefore nest success (Duguay et al.
2000). Food abundance may be altered by the removal of
large dead and dying trees that often house high densities of
bark- and wood-boring insects (Raphael and White 1984).
This in turn may alter clutch size, timing of nest initiation
and (or) fledging dates, nestling growth and survival, and
daily nest survival (Duguay et al. 2000).
Nest site characteristics of primary cavity nesters (i.e.,
birds capable of excavating their own cavities) are well docu-
mented. They typically select large-diameter declining or
dead trees with broken tops (Raphael and White 1984) in
areas with high tree density (Swallow et al. 1986; but see
Rolstad et al. 2000). However, nest site features that confer
high success for cavity-nesting birds have not been well
documented (Li and Martin 1991). Specific nest characteris-
tics (e.g., substrate hardness) can decrease predation risk by
providing a barrier to predator penetration (Tozer et al.
2009).
This study was designed to examine (i) the effect of silvi-
cultural intensity on vegetation structure, food abundance,
and four key nest reproductive parameters of cavity-nesting
birds (clutch size, date the first egg was laid, nestling period,
and date of fledging) and (ii) the effects of silviculture, nest
age/date, nest microhabitat, and food abundance on daily sur-
vival rates and nest success of these birds nesting in eastern
deciduous forests in a heavily fragmented landscape in south-
ern Canada.
Materials and methods
Study sites
Nineteen study sites were located within 75 km of London
(42°59′N, 81°14′W), Ontario, Canada, in Elgin, Middlesex,
Haldimand-Norfolk, and Oxford counties (Fig. 1). The land-
scape is highly fragmented by agriculture and development,
with an average forest cover of 14% (Ontario Ministry of
Natural Resources 2006). Land ownership is predominantly
private (approximately 87%) with high turnover rates (i.e.,
11 year average ownership).
Forests were dominated by sugar maple (Acer saccharum
Marsh.) with American beech (Fagus grandifolia Ehrh.), red
oak (Quercus rubra L.), red maple (Acer rubrum L.), and
white ash (Fraxinus americana L.) as codominants.
Study sites ranging in size from 12 to 70 ha were located
within much larger woodlands of up to 270 ha (see Supple-
mentary Material1) and typically delineated by property or
silvicultural treatment boundaries. Sites were categorized
into one of five silvicultural treatments defined according to
time since harvest and silvicultural intensity. “Reference”
sites (n= 4) were those that had not been harvested within
the past 15–50 years and were characterized by high basal
area (29.7 ± 0.6 m2/ha) (Table 1). Four partial harvesting
techniques were studied: group (n= 3), standard (n= 4),
old standard (n= 3), and heavy cuts (n= 4) (but see foot-
note in Supplementary Material1), with the first three in-
tended to match current guidelines for selection harvesting
(Ontario Ministry of Natural Resources 2000). These sites
had up to one third of the pre-harvest basal area removed,
maintained a residual basal area greater than 20.5 ±
1.5 m2∕ha, and retained at least six living cavity trees per
hectare. “Group”selection harvesting confined the removal
of trees to 12 small, clearcut patches scattered across a treat-
ment area of approximately 20 ha. This represented a 5% re-
moval rate of both volume and basal area (28.4 ± 1.3 m2/ha)
(Table 1) with little or no removal between gaps. Group cuts
consisted of the creation of five small gaps (20 m diameter,
gap size area 0.03 ha), four medium gaps (30 m, 0.07 ha),
and three large gaps (42 m, 0.14 ha). “Standard”sites were
harvested using single-tree selection where unhealthy trees
were targeted during removal and basal area was reduced to
a current mean of 23.2 ± 1.1 m2/ha (Table 1). Both group
and standard cut sites were harvested between 2002 and
2004 (<5 years post-harvest at the time of study). “Old
standard”cuts were harvested using the same prescription as
for standard sites between 1994 and 1998 (6–10 years post-
harvest at the time of study) with a current mean basal area
of 25.5 ± 0.9 m2/ha. “Heavy”cuts were harvested between
1994 and 1998 and complied with municipal diameter-limit
tree-cutting bylaws but failed to meet residual basal area
guidelines (current mean basal area of 18.9 ± 1.5 m2/ha) (Ta-
ble 1) through targeting large trees and (or) removing greater
than one third of the pre-harvest basal area.
Cavity-nesting bird breeding densities
Breeding densities were determined in 2004 and 2005 us-
ing territory mapping (Bibby et al. 2000) for five primary ex-
cavators, Red-bellied Woodpecker (Melanerpes carolinus
(Linnaeus, 1758)), Downy Woodpecker (Picodes pubescens
(Linnaeus, 1766)), Hairy Woodpecker (Picodes villosus (Lin-
naeus, 1766)), Yellow-bellied Sapsucker (Sphyrapicus varius
(Linnaeus, 1766)), and Northern Flicker (Colaptes auratus
(Linnaeus, 1758)), and one weak excavator, Black-capped
Chickadee (Poecile atricapillus (Linnaeus, 1766)). Study
sites were visited a minimum of nine times per year between
1 April and 15 July and systematically searched from sunrise
to 11 a.m. Densities were calculated as the number of territo-
ries detected divided by search area (territories per hectare)
and included partial territories.
1Supplementary data are available with the article through the journal Web site (http://www.nrcresearchpress.com/cjfr).
Straus et al. 1005
Published by NRC Research Press
Can. J. For. Res. Downloaded from www.nrcresearchpress.com by Dalian Nationalities University on 06/05/13
For personal use only.
Table 1. Structural comparison of vegetation variables between silvicultural treatments in southwestern Ontario (mean ± SE with nin par-
entheses).
Silvicultural treatment
Parameter Reference
(≥15 years old) Group
(≤5 years old) Old standard
(≥6 years old) Standard
(≤5 years old) Heavy
(≥6 years old) Fp
% shrub cover* 16.0±1.5 (101) 16.2±3.4 (41) 14.2±1.6 (76) 12.5±1.6 (80) 9.7±2.9 (32) 1.86 0.120
% sapling cover* 11.5±1.0 (101)a 15.3±2.4 (41) 18.0±2.3 (76)b 9.4±1.5 (80)a 18.0±3.0 (32)b 4.37 <0.001
% understory cover* 23.6±1.7 (101)a 27.1±3.1 (41) 25.8±2.2 (76) 22.9±2.3 (80)a 31.9±2.6 (32)b 3.12 0.040
% canopy cover* 77.7±1.4 (101)a 65.4±3.4 (41)b 70.0±3.0 (76)c 57.9±3.3 (80)d 58.8±6.2 (32)d 11.68 <0.001
Average closest tree (m)* 6.1±0.4 (101)a 5.2±0.3 (41)a 5.6±0.2 (76)a 10.5±3.1 (80)b 6.7±0.6 (32) 3.11 0.020
Live tree basal area (m2/ha)* 29.7±0.6 (101)a 28.4±1.3 (41)b 25.5±0.9 (76)c 23.2±1.1 (80)d 18.9±1.5 (32)d 5.02 <0.001
Mast tree basal area (m2/ha)†11.4±1.5 (157)a 8.4±0.8 (47)a 6.1±0.4 (75) 4.9±1.1 (45)b 2.4±0.7 (95)b 6.23 <0.006
Decayed downed woody debris (m2/ha)‡31.2±9.0 (4) 24.4±9.8 (3) 18.2±4.0 (3) 8.6±1.4 (3) 19.5±3.7 (4) 1.07 0.410
Fresh downed woody debris (m2/ha)‡15.2±15.0 (4) 42.5±12.7 (3) 11.1±5.5 (3) 79.5±22.6 (3) 30.1±11.7 (4) 1.43 0.230
Density healthy trees >25 cm (no./ha)†119.4±4.5 (157)a 105.6±1.7 (47) 121.7±5.4 (75)a 105.7±7.2 (45) 98.2±5.7 (95)b 3.08 0.020
Density declining trees >25 cm (no./ha)†59.6±3.5 (157)a 42.0±5.7 (47)b 29.0±3.5 (75)c 46.7±5.3 (45)b 26.1±2.8 (95)c 17.92 <0.001
Density dead trees >25 cm (no./ha)†14.2±1.6 (157)a 4.3±1.6 (47)b 8.3±1.6 (75)b 5.6±1.6 (45)b 5.8±1.2 (95)b 6.17 <0.001
Density dead trees (no./ha)†51.8±3.4 (157)a 33.5±5.8 (47)b 38.3±4.6 (75)b 49.4±1.4 (45)c 28.4±3.4 (95)d 6.01 <0.001
Woodlot patch area (ha)§201.4±8.8 (4)a 217.2±26.8 (3)a 80.6±46.2 (3)b 186.7±11.8 (4) 81.3±23.9 (5)b 5.09 0.010
Note: Different letters indicate a significant difference (p< 0.10) between treatments using Tukey’s unequal N HSD post hoc test.
*Nested ANOVA, site nested within treatment, F[4,17].
†Nested ANOVA, site nested within treatment, F[4,14].
‡One-way ANOVA, F[4,15].
§One-way ANOVA, F[4,14].
Fig. 1. Distribution of study sites near London, Ontario, Canada.
1006 Can. J. For. Res. Vol. 41, 2011
Published by NRC Research Press
Can. J. For. Res. Downloaded from www.nrcresearchpress.com by Dalian Nationalities University on 06/05/13
For personal use only.
Nest searching and monitoring
Active cavity nests were located between mid-April and
early August each year from 2004 to 2008. Nests were lo-
cated by (i) territory mapping to determine the number and
location of breeding pairs, (ii) listening for excavation activ-
ities and begging nestlings, (iii) tracking adults carrying food
or nesting material, (iv) checking nests from previous years
for cavity, tree, or territory reuse, and (v) chance encounters.
Due to high site fidelity of many of our study species, nests
from previous years were inspected for signs of reuse using a
TreeTop Peeper (Sandpiper Technologies, Manteca, Califor-
nia). When old nests were too high to peep (>14.5 m), we
examined the tree for fresh excavations. Nests were peeped
once in early to mid-May and again in June. Due to an in-
creasing number of old nests, time constraints only permitted
nest checks once in 2007, conducted throughout May and
June when nest activity is at its peak. Although we attempted
to spend equal time searching in each study site, old nest
sites and territory maps were used to target efforts such that
more productive sites (i.e., those with more territorial birds)
tended to receive greater attention as efforts were made to
find every nest of birds with established territories. Each
nest with a minimum of two confirmed visits was included
in the nest survival analyses, regardless of reuse status.
Once located, active nests were monitored every 5–7days
(more frequently as nests neared completion) using the Tree-
Top Peeper (nests ≤14.5 m) or 20–50 min observation peri-
ods (nests >14.5 m). Details were recorded on the presence–
absence of brooding, size of food, cavity depth of parent(s)
penetration during feeding, and if nestlings were visible at
the cavity entrance. Nest age was then estimated based on
correlations between behavior and nest content established
from 750 h of observations made in 2004–2005. A nest was
considered successful if at least one young fledged. Fledging
was confirmed by (i) fledglings seen and (or) heard near the
nest, (ii) an intact nest and young were capable of flying at
the previous visit, or (iii) parents seen carrying food in the
vicinity of nests.
Potential nest predators present in our study sites included
raccoon (Procyon lotor (Linnaeus, 1758)), southern flying
squirrel (Glaucomys volans (Linnaeus, 1758)), eastern grey
squirrel (Sciurus carolinensis Gmelin, 1788), eastern chip-
munk (Tamias striatus (Linnaeus, 1758)), red squirrel (Ta-
miasciurus hudsonicus (Erxleben, 1777)), long-tailed weasel
(Mustela frenata Lichtenstein, 1831), white-footed mouse
(Peromyscus leucopus (Rafinesque, 1818)), deer mouse (Per-
omyscus maniculatus (Wagner, 1845)), Pileated Woodpecker
(Dryocopus pileatus (Linnaeus, 1758)), Red-bellied Wood-
pecker, gray ratsnake (Pantherophis obsoleta), and fox snake
(Elaphe vulpina (Baird and Girard, 1853)). Potential compet-
itors included House Wrens (Troglodytes aedon Vieillot,
1809) and European Starlings (Sturnus vulgaris Linnaeus,
1758).
Vegetation assessment
Site vegetation characteristics were assessed in 2004–2005
(n= 137) using random habitat plots stratified across silvi-
cultural treatment and study sites. These plots (0.04 ha,
11.28 m radius) were centered on a nonuse tree (>10 cm di-
ameter at breast height (DBH)) and measurements included
percent cover of shrub (0.5–1.3 m height), sapling (>1.3 m
height, <2.5 cm DBH), understory (2.5–9.9 cm DBH), and
canopy cover (>10 m height). Average closest tree was meas-
ured using the point-quarter method.
Basal area and density variables presented in Table 1 were
calculated from tree species and DBH data recorded for all
standing trees (≥10.0 cm DBH) alive or dead located within
the random habitat plot. Living trees were further classified
according to provincial tree marking guidelines as healthy
(equivalent to Acceptable Growing Stock) or declining
(equivalent to Unacceptable Growing Stock) by documenting
the presence of defects indicative of internal rot or poor
health (Ontario Ministry of Natural Resources 2004). In addi-
tion to documenting potential foraging and nesting sites (i.e.,
densities of declining and dead trees), the basal area of mast
producing tree species (e.g., oak, hickory, cherry, walnut, and
beech) was calculated as potential food sources for both the
Red-bellied Woodpecker (Shackelford et al. 2000) and small
mammal nest predators (e.g., southern flying squirrel).
Tree data (species, DBH, and health) were available from
previously established permanent growth plots (n= 276) and
were included in basal area and density calculations (Table 1)
to increase sample size for sites harvested before 2000 (refer-
ence, heavy, and old standard). These plots (0.04 ha, 11.28 m
radius) were established in a stratified sampling design dis-
tributed across each study site. Although these data were col-
lected in 2000, changes in tree diameter, or health status, are
quite slow and the resulting basal area or density calculations
vary little between years. For group selection sites, since only
5% of the total area in which we searched for nests was har-
vested, we calculated post-harvest site basal area as 95% of
the precut measure. This was done in an effort to not over-
or underestimate site-level basal area resulting from a silvi-
cultural method that creates a patchy network of uneven har-
vested gaps and an unharvested matrix.
A second set of plots were assessed that were centered on
random subset of nest trees stratified by species and site
(Red-bellied Woodpecker n= 42, Yellow-bellied Sapsucker
n= 61, Northern Flicker n= 64, and Black-capped Chicka-
dee n= 35). Centered on the nest tree, these nest plots col-
lected the same habitat data as at random plots as well as
nest tree and nest microhabitat characteristics. Nest height,
nest tree DBH, species, and number of additional excavated
cavities were recorded. Tree health (healthy, declining, and
dead) was assigned in the field. For nesting sites where reuse
occurred, vegetation characteristics were not considered inde-
pendent and thus were only entered into statistical analyses
once.
One hundred metre transects (n= 17 (see footnote in Sup-
plementary Material1); 2003–2005) were placed randomly in
study sites post-harvest to calculate downed woody debris by
tallying, categorizing, and measuring all intersecting pieces
of downed woody debris. Fresh downed woody debris was
hard with intact bark, while decayed downed woody debris
was soft with little or no bark. The volume of downed woody
debris was calculated using equations developed by Van Wag-
ner (1968).
Land cover data (Ontario Ministry of Natural Resources
2006) were used to calculate site metrics including study site
area (i.e., the area in which we searched for nests), woodlot
patch area (i.e., continuously forested area surrounded by
Straus et al. 1007
Published by NRC Research Press
Can. J. For. Res. Downloaded from www.nrcresearchpress.com by Dalian Nationalities University on 06/05/13
For personal use only.
hard edges), and nest distance to hard edge (e.g., agriculture,
road, railroad).
Arthropod biomass
We sampled arthropod biomass in 2007 to assess cavity
nester food abundance across silvicultural treatments and tree
health as well as model its influence on nest success. Four
randomly selected plots were established in each study site
and tangle-traps were placed on the closest healthy, declin-
ing, and dead tree of any species (>10 cm DBH). This re-
sulted in a total of 12 trees sampled per site. Tangle-traps
were established by applying a layer of Tangle-trap Insect
Trap Coating (The Tanglefoot Company, Grand Rapids,
Michigan) on 30 cm wide transparent plastic wrapped around
each tree at 1.3 m above the ground. Two sampling periods
of 2 weeks duration were separated by a 2-week gap. The
timing of arthropod sampling was designed to correspond to
the nestling period of our study species when adults are feed-
ing young and any food limitation would presumably be most
influential (initial setup range 11–17 May and repeated 8–15
June).
A standardized subsampling method was used to analyze
data from the tangle-traps. Each trap was laid flat and a grid
consisting of 45 9 cm × 9 cm quadrats was placed over the
trap. Eleven quadrats were then randomly selected for sam-
pling. This resulted in a total sample area of 891 cm2
(11 cm × 9 cm × 9 cm), an area comparable with that from
other published studies that sampled 600–1300 cm2(Mosley
et al. 2006). Insects greater than 2 mm in length were identi-
fied to Order and their length recorded to the nearest milli-
metre. For five of these quadrats (selected at random),
individuals less than 2 mm were also counted. Dry biomass
was calculated using published Order-specific regression
equations (Sage 1982; Sample et al. 1993).
Statistical analyses
All statistical analyses were performed in SAS 9.0 (SAS
Institute Inc., Cary, North Carolina) and Statistica 7 (StatSoft
Inc., Tulsa, Oklahoma) and tests were considered significant
at a= 0.10 to reduce the likelihood of Type II errors (Naka-
waga 2004) and because the number of sites per treatment
was small. Data were transformed where applicable and as
indicated.
We selected 14 vegetation variables (Table 1) a priori for
their potential influence on cavity-nesting bird site suitability
(Raphael and White 1984) to characterize the structural dif-
ference among silvicultural treatments. Amount of vegetative
cover, tree densities (e.g., density of dead trees), average
closest tree, and basal area measures (e.g., mast tree basal
area) were assessed using nested ANOVAs with site as a ran-
dom factor nested within treatment. Downed woody debris
and woodlot patch area were analyzed using a one-way AN-
OVA due to a lack of replicate samples within each study
site. We used Kruskal–Wallis ANOVA by ranks to analyze
breeding densities, as these data failed to meet normality as-
sumptions.
A logistic exposure model (Shaffer 2004) was used to as-
sess the explanatory power of suites of hierarchical models
(see Davis et al. 2006) on nest success and compared using
Akaike’s information criterion (AIC) adjusted for small sam-
ple size (AICc). Eighteen variables were included in the can-
didate model set (i.e., those within 2 AICcunits of the top
model) and were grouped by type to facilitate analyses: nest
age/date (suite 1), woodlot-specific effects (suite 2), and nest
microhabitat (suite 3). Nest age/date models with DAICcval-
ues less than 2 were applied to the second suite (i.e., wood-
lot-specific effects) and then the third (i.e., nest microhabitat).
Suite 1 analysis followed the work of Grant et al. (2005) to
examine the effects of nest age, date, and year on cavity-nest-
ing bird daily survival rates (DSRs). We modeled nest date
and age as linear, quadratic (x+x2), and, in the case of nest
age, cubic effects (x+x2+x3). Twenty-four candidate mod-
els were created based on all possible combinations of nest
age, year, and date (Grant et al. 2005).
Suite 2 examined woodlot-specific effects (arthropod bio-
mass (pooled by study site), woodlot patch area, and silvicul-
tural treatment) that were modeled individually and in all
possible additive combinations. Since Northern Flickers pri-
marily forage on the ground for ants (Moore 1998), arthropod
biomass was not included in the model set for this species.
Also, the Black-capped Chickadee never nested in healthy
trees; therefore, we removed this health class from the analy-
sis.
Suite 3 examined seven nest microhabitat vegetation char-
acteristics selected a priori based on demonstrated effects on
cavity-nesting bird habitat selection (Bavrlic 2008) and DSR
(Li and Martin 1991). Nest tree species (hard maple (e.g.,
sugar), soft maple (e.g, red), ash, beech, birch, fruit and nut
producers, or unidentifiable), DBH, health (dead, declining,
or healthy), and number of additional excavated cavities in
the nest tree as well as the height of the nest, the percent
cover of the canopy (>10 m) in the nest plot, and distance
of the nest to a hard edge were modeled.
Model-averaged parameter estimates (to account for model
uncertainty) were generated from all models and fitted using
PROC GENMOD (SAS Institute Inc., Cary, North Carolina).
DSRs were calculated using parameter estimates (b0and b1)
from the logistic exposure model:
SuccessModel ¼ðDSRModel Þnest period
where
DSRModel ¼eb0þb1x=1þeb0þb1x
For comparison purposes, and for the Downy and Hairy
Woodpeckers, the Mayfield (1975) method was used to cal-
culate overall nest success regardless of silvicultural treat-
ment or other nest-specific variables:
SuccessMayfield ¼ðDSRMayfield Þnest period
where
DSRMayfield ¼1number of failed nests
total number of exposure days
Apparent nest success, also for comparison purposes, was
calculated as the number of successful nests divided by the
total number of nests.
Goodness of fit was determined for the data set using a
likelihood ratio c2test to compare the global model with the
null (Tabachnick and Fidell 1996). Variables in the candidate
set of models were ranked according to their importance val-
1008 Can. J. For. Res. Vol. 41, 2011
Published by NRC Research Press
Can. J. For. Res. Downloaded from www.nrcresearchpress.com by Dalian Nationalities University on 06/05/13
For personal use only.
ues by summing the AICcweights for all models in which
those variables appeared. Importance values within a factor
of 2 of the top variables were considered to be influential on
nest survival.
Arthropod biomass was log-transformed and analyzed us-
ing a factorial ANOVA to test for effects of (i) sampling pe-
riod (i.e., early, late), (ii) silvicultural treatment, (iii) tree
health (i.e., healthy, declining, live), and (iv) all interactions.
Post hoc analyses were conducted using Tukey’s unequal N
HSD test. All mean values are presented as the log10 of bio-
mass (originally calculated in milligrams).
First egg date was calculated using mean published incu-
bation periods and clutch sizes (see Smith 1993; Moore
1998; Shackelford et al. 2000; Jackson and Ouellet 2002;
Jackson et al. 2002; Walters et al. 2002) for inaccessible
nests and backdated. Clutch size was considered as the max-
imum number of eggs or young recorded in accessible nests
(i.e., could not be calculated for nests ≥14.5 m or Downy
Woodpecker cavities). Nestling period was calculated as time
from hatching until fledging, backdating where necessary.
Date of fledging was calculated as 2 days after a nest check
in which young were capable of fledging unless fledging was
observed directly.
The effect of silvicultural treatment on nest reproductive
parameters was analyzed using a one-way ANOVA, remov-
ing from analyses any treatments for which n< 4. Post hoc
analyses were conducted using Tukey’s unequal N HSD test
by species for each parameter.
Results
Vegetation structure
Vegetation characteristics varied significantly among treat-
ments for most structural attributes and particularly for den-
sities of dead and declining trees (Table 1). Reference sites
had higher canopy cover and densities of large (>25 cm
DBH) dead trees than all partially harvested sites. Heavily
logged sites had denser growth in the understory but less
canopy closure and lower densities of large, healthy, and de-
clining trees. Standard sites were similar to reference sites in
vertical structure but had significantly lower basal area of
live trees and density of large dead and declining trees. Old
standard cuts were similar to reference sites in both vertical
structure and basal areas but had significantly lower densities
of dead and declining trees. Although similar in structure and
basal area measurements, reference sites had higher densities
of declining and dead trees than group selection cuts.
Breeding densities
Breeding bird densities varied significantly among silvicul-
tural treatments for the Red-bellied, Hairy, and Downy
Woodpeckers. The Red-bellied Woodpecker was more abun-
dant in old standard cut sites compared with both standard
and group cuts. The Downy Woodpecker was also more
abundant in old standard cut sites, but only when compared
with group or heavy cuts (Table 2). Hairy Woodpeckers
were more abundant in reference sites than in group or heavy
cuts (Table 2). Yellow-bellied Sapsuckers were absent from
heavy cut sites.
Overall nest success
Approximately 6080 h were spent locating 466 nests of six
species of cavity-nesting birds. Of these nests, 83 were re-
moved from analyses, the majority due to abandonment dur-
ing construction or unknown factors before egg-laying was
confirmed. High rates of nest success were experienced
across all study sites. Downy and Hairy Woodpeckers had
the highest rates of nest success over the entire nesting cycle
(0.956 and 0.960, respectively), having only one confirmed
failure each (Table 3). Therefore, modeling was impossible
for these two species, as the number of interval failures was
negligible and nest success was therefore not influenced by
any of the variables examined. Black-capped Chickadees ex-
perienced the lowest nest success (0.617) followed by the
Yellow-bellied Sapsucker, Northern Flicker, and Red-bellied
Woodpecker (Table 3). Nest success was similar across silvi-
cultural treatments for each species (Table 3).
Nest site microhabitat
As a group, woodpeckers tended to nest high (12–16 m) in
large-diameter (39–48 cm DBH) trees with numerous addi-
tional feeding and nest cavities (9–13) in areas of moderate
Table 2. Comparison of breeding pair densities (pairs per hectare) in reference and partially harvested forest patches (mean ± SE with
ranges in parentheses).
Silvicultural treatment
Species Reference Group Old standard Standard Heavy H[4,18] p
Red-bellied Woodpecker 0.5±0.2 0.3±0.06a 0.8±0.15b 0.1±0.03a 0.5±0.02 9.17 0.06
(0.3–0.8) (0.2–0.4) (0.7–1.0) (0–0.5) (0.3–0.8)
Yellow-bellied Sapsucker 0.8±0.04 0.2±0.12 0.5±0.01 0.4±0.02 0 —*—*
(0.2–1.2) (0.1–0.3) (0.3–0.8) (0.4–0.9)
Downy Woodpecker 0.8±0.05 0.6±0.09b 1.0±0.06a 0.8±0.09 0.7±0.08b 8.82 0.07
(0.7–1.0) (0.4–0.7) (0.9–1.2) (0.6–1.1) (0.5–0.9
Hairy Woodpecker 0.8±0.01a 0.4±0.03b 0.7±0.02 0.6±0.07 0.4±0.01b 10.25 0.05
(0.5–1.0) (0.3–0.5) (0.6–0.7) (0.5–0.8) (0–0.6)
Northern Flicker 0.7±0.03 0.5±0.03 0.8±0.12 0.7±0.11 0.6±0.08 7.23 0.12
(0.6–0.8) (0.4–0.6) (0.7–1.2) (0.5–1.1) (0.4–0.9)
Black-capped Chickadee 1.5±0.06 1.5±0.03 2.0±0.11 1.7±0.50 2.3±0.25 6.98 0.14
(1.3–1.6) (1.2–1.9) (1.7–2.3) (1.1–2.1) (1.8–2.8)
Note: Different letters indicate a significant difference (p< 0.10) between treatments using post hoc tests.
*Sample size too low to test for significance.
Straus et al. 1009
Published by NRC Research Press
Can. J. For. Res. Downloaded from www.nrcresearchpress.com by Dalian Nationalities University on 06/05/13
For personal use only.
canopy cover (48%–67%) more than 150 m from edges (Ta-
Table 4). Nests were typically built in dead (39%–69%) or
declining (23%–54%) soft maples (either Freeman (Acer ×
freemanii E. Murr.), silver (Acer saccharinum L.), or red,
37%–53%) (Table 4). Black-capped Chickadees nested much
closer to the ground than woodpeckers, in the smallest diam-
eter trees, frequently in short dead stubs of trees of unidenti-
fiable species (Table 4).
Model selection
Three hundred and eighty-three nests had sufficient obser-
vations to be included in the logistic exposure analysis mod-
eling. These nests had effective sample sizes of 304 trials for
the Red-bellied Woodpecker, 433 for the Yellow-bellied Sap-
sucker, 195 for the Downy Woodpecker, 248 for the Hairy
Woodpecker, 434 for the Northern Flicker, and 138 for the
Black-capped Chickadee.
There was no overall effect of nest age/date, woodlot patch
area, or silvicultural treatment on nest survival for any of the
species studied. For the Red-bellied Woodpecker, 50 of the
106 models examining the influence of nest microhabitat var-
iables on nest success failed to converge. This is attributable
to high success (up to 100%) in a number of the treatments
(Table 3), making modeling difficult, as the sample size of
failed nests was small. Of the models that did converge, the
set of candidate models indicated that DSR was affected by
nest height, canopy coverage, DBH of the nest tree, number
of additional cavities in the nest tree, and distance to edge
(Table 5). Red-bellied Woodpecker nests high in large diam-
eter trees, containing additional cavities, in microhabitats
with greater canopy cover and farther from edges experi-
enced greater success (Table 6). Overall model fit for the
Red-bellied Woodpecker could not be assessed because the
global model failed to converge and nest height and tree di-
ameter were positively correlated (r= 0.478, p= 0.007).
Arthropod biomass and nest tree DBH were the most im-
portant variables influencing nest survival for the Yellow-bel-
lied Sapsucker (Table 6). Candidate models included
combinations of all nest microhabitat variables examined,
with the exception of tree species. Overall, nests in trees
with large DBH, plots with lower canopy cover, and greater
arthropod biomass experienced greater DSR (Table 6). Nest
microhabitat model fit was significant (c2= 34.4, p<
0.001).
All nest variables appeared in the candidate set of models
for the Northern Flicker except distance to edge (Table 5).
Importance values for nest height and tree species were high
(>0.9 each) (Table 6), with further support for tree health
and canopy cover. Nests in healthy trees and hard maples
(e.g., sugar and black maple (Acer nigrum Michx.)) experi-
enced greater DSR than nests in other tree species or those
of poorer health. Higher nests and lower canopy coverage
also increased DSR for Northern Flickers (Table 6). Overall
model fit was significant (c2= 59.5, p<0.001).
The Black-capped Chickadee candidate set of models, con-
trary to the other primary cavity nesters, included a negative
effect of nest height, canopy coverage, tree DBH, and arthro-
pod biomass (Table 6). Tree health and distance to edge were
less supported as indicated by their low importance values
(Table 6). Fit of the global model was significant (c2=
30.0, p< 0.001).
Table 3. Number of nests, apparent nest success (number of nests successful/total number of nests, expressed as a percentage), and calculated nest success using parameter estimates (b0
and b1) from the model treatment using the logistic-exposure method for six species of cavity-nesting birds in southwestern Ontario.
Silvicultural treatment
Reference Group Old standard Standard Heavy Total
Species No. % Success* No. % Success* No. % Success* No. % Success* No. % Success* No. % Success†
Red-bellied Woodpecker 17 94.1 0.911 6 100.0 1.000 11 72.7 0.810 12 100.0 1.000 9 100.0 1.000 55 92.7 0.902
Yellow-bellied Sapsucker 48 83.3 0.745 8 75.0 0.738 9 77.8 0.732 20 90.0 0.761 0 —— 85 82.4 0.721
Downy Woodpecker 24 95.8 0.882†5 100.0 1.000 14 100.0 1.000 11 100.0 1.000 4 100.0 1.000 58 98.3 0.956
Hairy Woodpecker 23 100.0 1.000 13 100.0 1.000 10 100.0 1.000 22 95.5 0.879†10 100.0 1.000 78 98.7 0.960
Northern Flicker 14 71.4 0.612 10 70.0 0.630 16 81.3 0.709 23 87.0 0.842 6 100.0 1.000 69 81.2 0.759
Black-capped Chickadee 14 85.7 0.587 7 71.4 0.587 6 83.3 0.589 6 66.7 0.562 5 60.0 0.559 38 76.3 0.617
*SuccessModel = (DSRModel)nest period, DSRModel =e
b0þb1x=1þeb0þb1x.
†SuccessMayfield = (DSRMayfield)nest period, DSRMayfield =1–(number of failed nests/total number of exposure days).
1010 Can. J. For. Res. Vol. 41, 2011
Published by NRC Research Press
Can. J. For. Res. Downloaded from www.nrcresearchpress.com by Dalian Nationalities University on 06/05/13
For personal use only.
Arthropod biomass
From two sampling periods (2007), 48 230 arthropods
greater than 2 mm long from 20 Orders were counted. Dip-
tera accounted for 61.6% of the arthropods sampled followed
by Coleoptera (26.7%). Biomass was significantly (F[1,117] =
11.48, p= 0.001) lower in the first sampling period (2.61 ±
0.03 (log) mg) than in the second (2.73 ± 0.02 (log) mg).
Treatment effects (F[4,117] = 7.63, p< 0.001) indicated that
trees in standard selection sites (2.83 ± 0.04 (log) mg) con-
tained significantly greater arthropod biomass than those in
heavy cuts (2.64 ± 0.04 (log) mg) or reference sites (2.61 ±
0.03 (log) mg) (Fig. 2) for all health types.
There was a slight interaction effect between treatment and
tree health (F[8,131] = 2.03, p= 0.048) (Fig. 2) but these ef-
fects were minor relative to the strength of the main effects.
There was no overall difference in biomass between trees of
differing health status while controlling for sampling period
and treatment effects (F2,131 = 0.71, p= 0.794).
Nest reproductive parameters
There were no significant differences in clutch size, date of
first egg, or nestling period among silvicultural treatments for
any of the six species studied. However, Red-bellied Wood-
pecker fledging dates varied significantly (F[4,46] = 2.84, p=
0.035), as nests tended to fledge later in standard and old
standard sites than in other treatments (Table 7). Birds in
these standard treatments tended to delay egg-laying,
although clutch size and nestling period lengths were similar
across treatments.
Discussion
This is one of the first studies to document nest success
rates of cavity-nesting birds in southwestern Ontario. Our re-
sults indicate that silvicultural practices significantly altered
vegetation structure and some cavity-nesting bird breeding
densities. Nest survival remained uniformly high across the
range of silvicultural treatments for most of the species
studied. Exceptionally, the Yellow-bellied Sapsucker appears
to be highly sensitive to intensive logging and is absent on
the heavily cut sites. To determine whether silvicultural prac-
tices are creating ecological traps, further research is re-
quired, as there are currently no reliable adult or juvenile
survival estimates for the woodpecker species studied.
Despite the apparent lack of an effect of silvicultural inten-
sity on nest success, we found that partial harvesting did re-
duce the number of large dead and, to a lesser extent,
declining trees, which had some consequences for cavity-
nesting bird breeding densities. For the Downy and Hairy
Woodpeckers, densities were most severely reduced in group
and heavy cuts, while Red-bellied Woodpecker densities were
lowest in group and standard cut sites, with old standard sites
often equal to or superior to reference sites. The exact mech-
anism behind this reduction in woodpecker densities is un-
clear, as old standard sites supported lower densities of large
declining and dead trees than reference sites. Overall, the
maintenance of cavity trees outlined by current silvicultural
guidelines, with the possible exception of the case of group
selection, appears to provide reasonably adequate nesting
habitat to support high densities of cavity nesters, especially
as time since harvest elapses (e.g., 6–10 years post-harvest).
Table 4. Nest microhabitat (% or mean ± SE with nin parentheses) characteristics for six species of cavity-nesting birds in southwestern Ontario.
Species HT (m) DBH (cm) CAV (no.) CAN (%) EDGE (m) DD
(%) DC
(%) HE
(%) Mh
(%) Ms
(%) Ash
(%) BB
(%) FN
(%) Unk (%)
Red-bellied Woodpecker 16.2±0.6 (55) 47.0±2.5 (46) 11.5±3.3 (43) 61.9±3.4 (43) 151.1±13.3 (55) 48.2 45.8 6.0 20.4 37.0 13.0 13.0 9.3 7.3
Yellow-bellied Sapsucker 11.8±0.4 (85) 39.3±1.5 (76) 9.4±2.0 (64) 66.4±2.3 (65) 194.1±12.7 (85) 48.2 45.8 6.0 6.5 37.9 26.7 17.7 5.6 5.6
Downy Woodpecker 14.2±0.8 (58) 41.8±2.3 (55) 13.4±5.0 (50) 67.2±2.8 (50) 181.0±15.5 (58) 39.6 46.5 13.9 20.7 39.7 22.4 1.7 10.3 5.2
Hairy Woodpecker 14.9±0.7 (78) 44.0±2.3 (68) 12.7±6.4 (66) 60.3±2.6 (65) 201.5±11.3 (78) 33.8 53.5 12.7 9.3 53.4 9.3 8.0 12.0 8.0
Northern Flicker 12.4±0.5 (72) 47.9±2.7 (67) 9.8±1.4 (65) 48.0±3.3 (65) 153.4±12.8 (72) 69.0 22.5 8.5 16.9 40.8 26.8* —*—* 15.5
Black-capped Chickadee 1.7±0.2 (38) 12.2±0.4 (31) 7.1±2.0 (36) 65.1±4.3 (35) 166.4±17.1 (38) 94.7 5.3 0.0 5.4 18.9 2.7 29.7 8.1 35.2
Note: Given are HT (nest height), DBH (diameter of nest tree at breast height), CAV (number of additional cavities in the nest tree), CAN (percent canopy coverage in nest plot), EDGE (distance of nest to a
hard edge), tree health (DD, dead; DC, declining; HE, healthy), and tree species (Mh, hard maple; Ms, soft maple; Ash, all ash species; BB, beech and birch; FN, fruit and nut trees; Unk, unidentifiable tree
species),
*To facilitate model convergence during the logistic exposure analyses, tree species were further grouped by combining Ash, BB, and FN.
Straus et al. 1011
Published by NRC Research Press
Can. J. For. Res. Downloaded from www.nrcresearchpress.com by Dalian Nationalities University on 06/05/13
For personal use only.
Furthermore, pairs able to procure a suitable nest site in these
woodlots are able to achieve high nest success.
Of the seven nest microhabitat variables studied, all were
found to have a weak effect on DSR for one or more cavity-
nesting bird species. Nest height may influence nest success,
as small mammals are capable of depredating lower nests
more quickly and efficiently by not providing parent birds
time for detection and defense (Nilsson 1984). Higher nest
cavities also help to obscure auditory and olfactory cues pro-
vided by nestlings and (or) vocalizing adults to ground pred-
ators (Mahon and Martin 2006).
Black-capped Chickadees were unique from the other cav-
ity nesters in our study, as lower nests in small-diameter trees
experienced greater DSR than higher nests in larger trees.
The mechanism behind these results is not understood but
may be an artifact of habitat selection. As Black-capped
Chickadees nest in snags with broken tops (93% in Vermont;
Runde and Capen 1987) that are shorter than random nonuse
trees (Bavrlic 2008), nests in short trees will always be lower
regardless of their success. Other studies have failed to find
an influence of nest height on nest survival for this species
(Christman and Dhondt 1997).
Excavating in harder substrate (e.g., healthy trees; Schepps
et al. 1999) and large-diameter trees (i.e., with thicker nest
walls) may make nests less susceptible to predation (Christ-
man and Dhondt 1997). These characteristics may make the
cavity inaccessible, as smaller, weaker predators are unable
to scratch or chew through and (or) the energetic input for
the predator exceeds an acceptable threshold. Northern Flick-
ers, Yellow-bellied Sapsuckers, and Red-bellied Woodpeckers
experienced higher reproductive success in larger diameter
trees. Furthermore, the Northern Flicker had greater nest suc-
cess in healthy and hard maple (e.g., sugar and black) trees.
High canopy cover in the vicinity of the nest was nega-
Table 5. Final logistic-exposure candidate model set examining the effects of site-averaged arthropod biomass and nest microhabitat on daily
survival rates for four species of cavity-nesting birds.
Species Rank Candidate model loge(L)KAICcDAICcwi
Red-bellied Woodpecker 1 HT+CAN –10.45 3 26.91 0.000 0.130
2 DBH+HT –10.62 3 27.25 0.340 0.110
3 DBH+HT+CAV –10.13 4 28.28 1.370 0.066
4 DBH –12.30 2 28.60 1.690 0.057
5 DBH+HT+EDGE –10.40 4 28.81 1.900 0.050
41 CONSTANT –23.19 1 50.07 23.160 0.000
Yellow-bellied Sapsucker 1 ARTH+DBH –53.96 3 113.93 0.000 0.071
2 ARTH+CAV –54.14 3 114.29 0.360 0.060
3 ARTH+DBH+CAV –53.15 4 114.32 0.390 0.059
4 ARTH+CAN –54.21 3 114.43 0.500 0.056
5 ARTH+DBH+CAN –53.23 4 114.48 0.550 0.054
6 ARTH+HT+CAV –53.55 4 115.11 1.180 0.040
7 ARTH+DBH+HT –53.74 4 115.50 1.570 0.032
8 ARTH+EDGE+CAV –53.76 4 115.54 1.610 0.032
9 ARTH+HT+CAN –53.80 4 115.61 1.680 0.030
10 ARTH+EDGE+DBH –53.80 4 115.62 1.690 0.031
11 ARTH+EDGE+CAN –53.81 4 115.64 1.710 0.030
12 ARTH+DBH+HEALTH+CAV –51.87 5 115.78 1.850 0.028
102 CONSTANT –69.32 1 140.64 26.710 0.000
Northern Flicker 1 TSP+HEALTH+HT+CAN –39.93 8 95.92 0.000 0.157
2 TSP+HEALTH+HT+CAV+CAN –39.07 9 96.23 0.310 0.135
3 DBH+TSP+HEALTH+HT –40.17 8 96.41 0.490 0.123
4 DBH+TSP+HEALTH+HT+CAN –39.24 9 96.57 0.650 0.114
5 TSP+HEALTH+HT+CAV –40.39 8 96.86 0.940 0.099
6 TSP+HT+CAN –42.93 6 97.90 1.980 0.058
103 CONSTANT –68.37 1 138.73 42.810 0.000
Black-capped Chickadee 1 ARTH+DBH+HT+CAN –24.86 5 59.82 0.000 0.089
2 ARTH+DBH+HT+CAN+EDGE –24.13 6 60.41 0.590 0.066
3 ARTH+DBH+HEALTH+HT+CAN –24.15 6 60.43 0.610 0.065
4 ARTH+HEALTH+HT+CAN –25.33 5 60.77 0.950 0.055
5 ARTH+PATCH+DBH+HT+CAN+EDGE –23.41 7 61.01 1.190 0.049
6 ARTH+HT+CAN –26.48 4 61.02 1.200 0.049
7 ARTH+PATCH+HT+CAN –26.48 5 61.02 1.200 0.049
8 ARTH+PATCH+DBH+HT+CAN –24.69 6 61.52 1.700 0.038
9 ARTH+PATCH+DBH+HEALTH+HT+CAN –23.75 7 61.68 1.860 0.035
203 CONSTANT –39.64 1 81.28 21.460 0.000
Note: ARTH, arthropod biomass. Nest microhabitat variables: DBH, diameter at breast height; TSP, tree species; HEALTH, nest tree health; HT, nest
height; CAV, number of additional feeding or nest cavities; CAN, percent canopy coverage; EDGE, distance to hard edge. Kis the number of parameters
(categorical variables: K= number of categories –1) in the model plus an intercept.
1012 Can. J. For. Res. Vol. 41, 2011
Published by NRC Research Press
Can. J. For. Res. Downloaded from www.nrcresearchpress.com by Dalian Nationalities University on 06/05/13
For personal use only.
tively related to DSR for the Black-capped Chickadee and
Northern Flicker but improved DSR for the Red-bellied
Woodpecker. Areas with lower canopy coverage have more
open understories that will often develop dense sapling layers
(Jobes et al. 2004), providing nest screening from potential
predators for the Black-capped Chickadee. For species nest-
ing higher in the trees (e.g., Red-bellied Woodpecker), denser
high canopies would also provide screening from predators
(Mahon and Martin 2006).
For the Northern Flicker, nest plots supported significantly
greater densities of declining and dead trees than random
(Bavrlic 2008). Higher densities of dead and declining trees
result in low canopy cover at the nest plot. This abundance
of unoccupied nesting sites (i.e., dead and declining trees)
relative to occupied sites could reduce the efficiency of
search strategy driven predators (Martin 1988).
Distance to edge appeared in the candidate model set for
the Red-bellied Woodpecker, Yellow-bellied Sapsucker, and
Black-capped Chickadee. However, importance values for
this variable in the models of DSR for all three species were
low. Previous research on primary cavity nesters has also
failed to find edge effects on nest success for cavity nesters
in fragmented forests (Deng and Gao 2005).
Arthropod biomass varied over the breeding season and
between silvicultural treatments but not between trees of
varying health. Invertebrate seasonality is attributable to fe-
cundity and the recruitment of invertebrate young (Duguay
et al. 2000). The decline of prey availability later in the
breeding season is well recognized as a constraint for birds
that postpone nesting (reviewed in Martin 1987).
Woodpeckers select large snags as foraging substrate
(Swallow et al. 1988) due to an increased density of wood-
Table 6. Importance values and model-averaged parameter estimates (± unconditional SE) for arthro-
pod and nest microhabitat influences on daily nest survival for four species of cavity-nesting birds.
Species Rank Variable Importance value Parameter estimate ± SE
Red-bellied Woodpecker 1 DBH 0.653 0.008±0.04
2 HT 0.591 0.249±0.28
3 CAN 0.490 0.005±0.02
4 EDGE 0.254 –0.001±0.00
5 CAV 0.212 0.059±0.15
Yellow-bellied Sapsucker 1 ARTH 1.000 2.68±3.04
2 DBH 0.769 0.018±0.02
3 CAV 0.521 0.003±0.02
4 CAN 0.406 –0.001±0.00
5 EDGE 0.259 0.003±0.00
6 HT 0.247 0.016±0.04
7 HEALTH 0.196
DD 0.094±0.33
DC 0.271±0.55
HE 0.000±0.00
Northern Flicker 1 TSP 0.998
Mh 28.122±15.34
Ms 4.608±2.03
Other 0.832±0.78
Unk 0.000±0.00
2 HT 0.917 0.307±0.15
3 HEALTH 0.773
DD –17.829±8.22
DC –20.583±9.33
HE 0.000±0.00
4 CAN 0.673 –0.015±0.02
5 DBH 0.434 0.0148±0.03
6 CAV 0.332 –0.018±0.03
Black-capped Chickadee 1 ARTH 1.000 –5.195±4.97
2 HT 0.858 –0.705±0.47
3 CAN 0.757 –0.110±0.06
4 DBH 0.619 –0.106±0.17
5 HEALTH 0.386
DC 9.918±8.19
DD 0.000±0.00
6 EDGE 0.269 0.002±0.00
Note: ARTH, arthropod biomass. Nest microhabitat variables: DBH, diameter at breast height; TSP, tree species;
HEALTH, nest tree health; HT, nest height; CAV, number of additional feeding or nest cavities; CAN, percent ca-
nopy coverage; EDGE, distance to hard edge. Tree health categories: DD, dead; DC, declining; HE, healthy. Tree
species: Mh, hard maple; Ms, soft maple; Unk, unidentifiable; Other, all others.
Straus et al. 1013
Published by NRC Research Press
Can. J. For. Res. Downloaded from www.nrcresearchpress.com by Dalian Nationalities University on 06/05/13
For personal use only.
borers and catch per unit effort (Raphael and White 1984).
Contrary to our expectations, we failed to find an effect of
tree health on arthropod biomass. However, we did find an
effect of silviculture on arthropod biomass in which standard
cuts had greater abundance than reference or heavily cut
sites. Time since harvest likely explains this trend, as stand-
ard cut sites incurred an influx of downed woody debris
most recently (<5 years) and, although not significant, these
sites contained the greatest amount of fresh downed woody
debris. This resulted in a simultaneous flush of invertebrate
breeding habitat and food availability (Ulyshen et al. 2004).
Arthropod biomass entered top models describing DSR for
the Black-capped Chickadee and Yellow-bellied Sapsucker.
Counterintuitively, Black-capped Chickadee nests in areas
with greater food availability (e.g., standard sites) were less
likely to succeed than nests in areas with lower food avail-
ability (e.g., reference and heavy cuts). Previous research has
shown foraging location plasticity for chickadees in response
to variability in prey abundance (Holmes and Schultz 1988),
potentially making them resilient to the decreases in food
abundance observed in our study sites. Also, House Wrens
are aggressive birds that usurp chickadees from their nest
sites (Johnson 1998) and in our study sites nested almost ex-
clusively in Black-capped Chickadee excavated nests (88%;
K. Bavrlic, unpublished data). With significantly higher den-
sities of House Wrens in old standard cut sites than in refer-
ence, group, and heavy cuts and intermediate densities in
recent standard cuts (Bavrlic 2008), increased competition
for nest sites could decrease Black-capped Chickadee success
regardless of the increased prey abundance in standard cuts.
Further study would be required to determine the mechanism
behind these patterns.
Although the Yellow-bellied Sapsucker selected habitat
characteristics similar to the other woodpecker species
studied, discriminate function analyses indicated high re-
source overlap at the nest tree and nest microhabitat level
with the Hairy Woodpecker (Bavrlic 2008). Kilham (1971)
also acknowledged the potential for competition between the
Yellow-bellied Sapsucker and Hairy Woodpecker. As Hairy
Woodpeckers are year-round residents and Yellow-bellied
Sapsuckers are migratory, we hypothesize that Yellow-bellied
Sapsuckers returning to heavily cut woodlots are unable to
secure critical nesting habitat because of increased competi-
tion due to (i) the post-harvest loss of dead and declining
trees and (ii) high resource overlap with resident Hairy
Woodpeckers that establish territories prior to sapsuckers re-
turning to the breeding grounds.
Primary cavity nesters play a crucial role in forested eco-
systems by providing a vital resource (i.e., cavities) to a suite
of secondary cavity nesters (i.e., those incapable of excava-
tion; Martin and Eadie 1999). Diminished densities attributed
to insufficient retention of quality cavity trees would further
Fig. 2. Arthropod biomass collected from 161 trees 11–17 May (Round 1: solid circles, solid line) and 8–15 June (Round 2: open circles,
broken line) in 2007 using tangle-traps on dead, declining, and healthy trees from 19 study woodlots in southwestern Ontario.
1014 Can. J. For. Res. Vol. 41, 2011
Published by NRC Research Press
Can. J. For. Res. Downloaded from www.nrcresearchpress.com by Dalian Nationalities University on 06/05/13
For personal use only.
Table 7. Mean ± SE (nin parentheses) reproductive parameters for six species of cavity-nesting birds in southwestern Ontario.
Silvicultural treatment
Species Parameter Reference Group Old standard Standard Heavy Fp
Red-bellied Woodpecker* Julian date of first egg 123.0±1.8 (17) 121.3±3.0 (6) 130.5±3.1 (11) 128.5±3.6 (12) 122.6±2.2 (9) 1.94 0.118
Clutch size 2.8±0.4 (5) na 4.0±1.1 (4) 3.0±0 (1) 4.0±1.0 (2) 1.33 0.286
Length of nestling period 25.9±0.5 (16) 25.8±0.5 (6) 24.6±0.4 (10) 25.5±0.6 (11) 26.1±0.4 (8) 1.45 0.232
Julian fledge date 163.9±1.5 (16) 163.2±3.1 (6) 172.2±3.1 (10) 171.5±3.6 (11) 163.6±2.5 (8) 2.84 0.035
Yellow-bellied Sapsucker Julian date of first egg 136.5±1.0 (48) 136.0±1.1 (8) 139.2±2.4 (9) 139.0±2.2 (20) na 0.78 0.507
Clutch size 3.9±0.2 (27) 5.3±0.5 (4) 4.0±0.6 (5) 3.8±0.2 (17) na 1.66 0.188
Length of nestling period 26.7±0.4 (38) 26.7±0.7 (6) 26.1±0.8 (7) 25.5±0.3 (18) na 1.70 0.176
Julian fledge date 178.5±0.8 (38) 178.5±1.5 (6) 177.6±1.0 (7) 179.8±2.4 (18) na 0.24 0.865
Downy Woodpecker Julian date of first egg 130.8±1.3 (24) 130.4±2.7 (5) 132.8±1.2 (11) 132.5±2.0 (14) 132.3±2.7 (4) 0.35 0.843
Length of nestling period 20.9±0.1 (22) 20.8±0.5 (4) 20.5±0.3 (11) 20.4±0.2 (14) 20.3±0.5 (4) 1.32 0.276
Julian fledge date 166.4±1.4 (22) 168.3±2.3 (4) 168.5±1.3 (11) 167.8±2.0 (14) 167.0±2.5 (4) 0.25 0.907
Hairy Woodpecker Julian date of first egg 112.8±2.1 (23) 111.8±2.6 (13) 114.9±4.5 (10) 112.7±1.6 (22) 108.6±1.5 (10) 0.60 0.662
Clutch size 2.8±0.3 (8) 3.0±0.6 (3) 4.0±0 (1) 2.9±0.3 (9) 3.3±0.3 (3) 0.15 0.708
Length of nestling period 27.0±0.4 (22) 28.0±0.5 (12) 27.4±0.5 (10) 27.5±0.3 (20) 27.0±0.5 (10) 1.05 0.391
Julian fledge date 155.2±2.0 (22) 154.1±2.2 (12) 158.4±4.3 (10) 154.9±1.6 (20) 151.2±1.2 (10) 0.88 0.479
Northern Flicker Julian date of first egg 135.4±2.6 (14) 134.5±2.5 (10) 137.7±1.9 (16) 141.8±3.1 (26) 147.3±6.5 (6) 1.72 0.155
Clutch size 5.6±0.4 (12) 5.2±0.8 (6) 4.9±0.7 (9) 6.2±0.4 (14) 6.3±0.3 (4) 1.26 0.302
Length of nestling period 24.9±0.7 (11) 25.4±0.7 (7) 24.5±0.7 (13) 24.9±0.6 (20) 23.0±0.8 (5) 0.93 0.452
Julian fledge date 178.6±2.9 (11) 178.7±2.9 (7) 178.6±2.0 (13) 183.4±3.3 (20) 183.2±6.5 (5) 0.56 0.642
Black-capped Chickadee Julidan date of first egg 134.5±4.1 (14) 145.9±6.2 (7) 135.2±5.6 (6) 135.5±5.7 (6) 132.2±7.8 (5) 0.82 0.524
Clutch size 5.2±0.5 (12) 6.8±0.8 (4) 6.0±0 (3) 5.0±0.6 (5) 6.0±1.0 (5) 1.02 0.404
Length of nestling period 17.3±0.4 (12) 17.3±0.6 (4) 11.0±0 (1) 16.5±0.5 (4) 16.5±2.5 (2) —†—†
Julian fledge date 165.1±2.6 (12) 182.5±8.7(4) 187.0±0 (1) 169.3±8.6 (4) 178.0±16.0 (2) —†—†
Note: Different letters indicate a significant difference (p< 0.10) between treatments using Tukey’s unequal N HSD post hoc test. A Julian date of 110 corresponds to 20 April. na, not available.
*Treatments with n< 4 were removed from the analyses.
†Sample size too low to test for significance.
Straus et al. 1015
Published by NRC Research Press
Can. J. For. Res. Downloaded from www.nrcresearchpress.com by Dalian Nationalities University on 06/05/13
For personal use only.
limit cavity availability in the forest for other species. Similar
to other jurisdictions, the highly fragmented landscape of
southern Ontario imposes added challenges including unsus-
tainable tree-cutting bylaws that do not follow provincial
guidelines. Additionally, landowners who follow these guide-
lines (e.g., standard and group selection cuts) may still nega-
tively affect cavity-nesting birds, as densities of some key
primary cavity nesters decline even with retention of six cav-
ity trees per hectare. However, our work does show that for
all but one species, 10 years after single-tree selection har-
vesting, densities of cavity-nesting birds are either greater
than or equal to those in reference sites. Further research is
needed to monitor these trends beyond a single harvesting
cycle as well as determine the need for refinement of silvicul-
tural practices outside southern Ontario.
Acknowledgements
We thank J. Lau, D. Geleynse, K. Gilroy, C. Van Ness, B.
Taylor, and numerous others for help collecting field data.
We also thank E. Proctor for assistance in the bug lab. Dis-
cussions on statistical concepts from J. Balsdon and D. Tozer
were invaluable. The cooperation provided by private land-
owners and funding from Trent University, Ontario Ministry
of Natural Resources, Canadian Wildlife Services, Environ-
ment Canada, Long Point Regional Conservation Authority,
Science Horizons, Summer Experience Program, and the
Quebec Exchange program were crucial to this project.
References
Aitken, K.E.H., and Martin, K. 2004. Nest cavity availability and
selection in aspen–conifer groves in a grassland landscape. Can. J.
For. Res. 34(10): 2099–2109. doi:10.1139/x04-086.
Bavrlic, K. 2008. The effects of partial harvesting on cavity-nesting
birds in the Carolinian forests of southwestern Ontario: habitat
responses and species interactions. M.Sc. thesis, Trent University,
Peterborough, Ont.
Bibby, C.J., Burgess, N.D., and Hill, D.A. 2000. Bird census
techniques. Academic Press, London, U.K.
Brawn, J.D., and Balda, R.P. 1988. Population biology of cavity
nesters in northern Arizona: do nest sites limit breeding densities?
Condor, 90(1): 61–71. doi:10.2307/1368434.
Christman, B.J., and Dhondt, A.A. 1997. Nest predation in Black-
capped Chickadees: how safe are cavity nests? Auk, 114: 769–773.
Costello, C.A., Yamasaki, M., Pekins, P.J., Leak, W.B., and Neefus,
C.D. 2000. Songbird response to group selection harvests and
clearcuts in a New Hampshire northern hardwood forest. For. Ecol.
Manag. 127(1–3): 41–54. doi:10.1016/S0378-1127(99)00131-0.
Davis, S.K., Brigham, R.M., Shaffer, T.L., and James, P.C. 2006.
Mixed-grass prairie passerines exhibit weak and variable
responses to patch size. Auk, 123(3): 807–821. doi:10.1642/
0004-8038(2006)123[807:MPPEWA]2.0.CO;2.
Deng, W.H., and Gao, W. 2005. Edge effects on nesting success of
cavity-nesting birds in fragmented forests. Biol. Conserv. 126(3):
363–370. doi:10.1016/j.biocon.2005.06.013.
Duguay, J.P., Wood, P.B., and Miller, G.W. 2000. Effects of timber
harvests on invertebrate biomass and avian nest success. Wildl.
Soc. Bull. 28: 1123–1131.
Franklin, J.F., Berg, D.R., Tornburgh, D.A., and Tappeiner, J.C.
1997. Silvicultural approaches to timber harvesting: variable
retention harvesting systems. In Creating a forestry for the 21st
century: the science of ecosystem management. Island Press,
Washington, D.C.
Grant, T.A., Shaffer, T.L., Madden, E.M., and Pietz, P.J. 2005. Time-
specific variations in passerine nest survival: new insights into old
questions. Auk, 122(2): 661–672. doi:10.1642/0004-8038(2005)
122[0661:TVIPNS]2.0.CO;2.
Holmes, R.T., and Schultz, J.C. 1988. Food availability for forest
birds: effects of prey distribution and abundance on bird foraging.
Can. J. Zool. 66(3): 720–728. doi:10.1139/z88-107.
Hutto, H. 2006. Towards meaningful snag-management guidelines for
postfire salvage logging in North American conifer forests.
Conserv. Biol. 20(4): 984–993. doi:10.1111/j.1523-1739.2006.
00494.x. PMID:16922215.
Jackson, J.A., and Ouellet, H.R. 2002. Downy Woodpecker (Picoides
pubescens). In The birds of North America. The Birds of North
America, Inc., Philadelphia, Pa.
Jackson, J.A., Ouellet, H.R., and Jackson, B.J.S. 2002. Hairy
Woodpecker (Picoides villosus). In The birds of North America.
The Birds of North America, Inc., Philadelphia, Pa.
Jobes, A.P., Nol, E., and Voigt, D.R. 2004. Effects of selection cutting
on bird communities in contiguous Eastern hardwood forests. J.
Wildl. Manag. 68(1): 51–60. doi:10.2193/0022-541X(2004)068
[0051:EOSCOB]2.0.CO;2.
Johnson, L.S. 1998. House Wren (Troglodytes aedon). In The birds of
North America. No. 380. The Birds of North America, Inc.,
Philadelphia, Pa.
Kilham, L. 1971. Reproductive behavior of Yellow-bellied Sapsuck-
ers. I. Preference for nesting in Fomes-infected aspens and nest
hole interrelations with flying squirrels, raccoons, and other
animals. Wilson Bull. 83(2): 159–171.
Li, P., and Martin, T.E. 1991. Nest-site selection and nesting success
of cavity-nesting birds in high elevation forest drainages. Auk,
108: 405–418.
Mahon, C.L., and Martin, K. 2006. Nest survival of chickadees in
managed forests: habitat, predator, and year effects. J. Wildl.
Manag. 70(5): 1257–1265. doi:10.2193/0022-541X(2006)70
[1257:NSOCIM]2.0.CO;2.
Martin, T.E. 1987. Food as a limit on breeding birds: a life-history
perspective. Annu. Rev. Ecol. Syst. 18(1): 453–487. doi:10.1146/
annurev.es.18.110187.002321.
Martin, T.E. 1988. On the advantages of being different: nest
predation and coexistence of birds species. Proc. Natl. Acad. Sci.
U.S.A. 85(7): 2196–2199. doi:10.1073/pnas.85.7.2196.
Martin, K., and Eadie, J.M. 1999. Nest webs: a community-wide
approach to the mangement and conservation of cavity-nesting
forest birds. For. Ecol. Manag. 115(2–3): 243–257. doi:10.1016/
S0378-1127(98)00403-4.
Mayfield, H.F. 1975. Suggestions for calculating nest success. Wilson
Bull. 87: 456–466.
Mönkkönen, M., and Welsh, D.A. 1994. A biogeographical
hypothesis on the effects of human cause landscape changes on
the forest bird community of Europe and North America. Ann.
Zool. Fenn. 31:61–70.
Moore, W.S. 1998. Northern Flicker (Coleoptus auratus). In The
birds of North America. The Birds of North America, Inc.,
Philadelphia, Pa.
Mosley, E., Holmes, S.B., and Nol, E. 2006. Songbird diversity and
movement in upland and riparian habitats in the boreal mixedwood
forest of northeastern Ontario. Can. J. For. Res. 36(5): 1149–1164.
doi:10.1139/X06-010.
Nakagawa, S. 2004. A farewell to Bonferroni: the problems of low
statistical power and publication bias. Behav. Ecol. 15(6): 1044–
1045. doi:10.1093/beheco/arh107.
Nilsson, S.G. 1984. The evolution of nest-site selection among hole-
nesting birds: the importance of nest predation and competition.
Ornis Scand. 15(3): 167–175. doi:10.2307/3675958.
1016 Can. J. For. Res. Vol. 41, 2011
Published by NRC Research Press
Can. J. For. Res. Downloaded from www.nrcresearchpress.com by Dalian Nationalities University on 06/05/13
For personal use only.
Ontario Ministry of Natural Resources. 2000. A silvicultural guide to
managing southern Ontario forests. Queen’s Printer for Ontario,
Toronto, Ont.
Ontario Ministry of Natural Resources. 2004. Ontario tree marking
guide. Queen’s Printer for Ontario, Toronto, Ont.
Ontario Ministry of Natural Resources. 2006. SOLRIS phase 1
wooded areas —2006 release (WOODAREA). Ontario Land
Information Directory of Metadata for Geospatial Information
Holdings, Toronto, Ont.
Raphael, M.G., and White, M. 1984. Use of snags by cavity-nesting
birds in the Sierra Nevada. Department of Forest and Resources
Management, University of California, Berkley, Calif.
Rolstad, J., Rolstad, E., and Saeteren, O. 2000. Black woodpecker
nest sites: characteristics, selection, and reproductive success. J.
Wildl. Manag. 64(4): 1053–1066. doi:10.2307/3803216.
Runde, D.E., and Capen, D.E. 1987. Characteristics of northern
hardwood trees used by cavity-nesting birds. J. Wildl. Manag. 51
(1): 217–223. doi:10.2307/3801659.
Sage, R.D. 1982. Wet and dry-weight estimates of insects and spiders
based on length. Am. Midl. Nat. 108(2): 407–411. doi:10.2307/
2425505.
Sample, B.E., Cooper, R.J., Greer, R.D., and Whitmore, R.C. 1993.
Estimation of insect biomass by length and width. Am. Midl. Nat.
129(2): 234–240. doi:10.2307/2426503.
Schepps, J., Lohr, S., and Martin, T.E. 1999. Does tree hardness
influence nest-tree selection by primary cavity nesters? Auk, 116:
658–665.
Shackelford, C.E., Brown, R.E., and Conner, R.N. 2000. Red-bellied
Woodpecker (Melanerpes carolinus). In The birds of North
America. The Birds of North America, Inc., Philadelphia, Pa.
Shaffer, T.L. 2004. A unified approach to analyzing nest success.
Auk, 121(2): 526–540. doi:10.1642/0004-8038(2004)121[0526:
AUATAN]2.0.CO;2.
Simard, J.R., and Fryxell, J.M. 2003. Effects of selective logging on
terrestrial small mammals and arthropods. Can. J. Zool. 81(8):
1318–1326. doi:10.1139/z03-129.
Smith, S.M. 1993. Black-capped Chickadee (Parus atricapillus). In
The birds of North America. The Birds of North America, Inc.,
Philadelphia, Pa.
Swallow, S.K., Gutierrez, R.J., and Howard, R.A., Jr. 1986. Primary
cavity-site selection by birds. J. Wildl. Manag. 50(4): 576–583.
doi:10.2307/3800966.
Swallow, S.K., Howard, R.A., Jr, and Gutierrez, R.J. 1988. Snag
preferences of woodpeckers foraging in a northeastern hardwood
forest. Wilson Bull. 100: 236–246.
Tabachnick, B.G., and Fidell, L.S. 1996. Using multivariate statistics.
California State University, Northridge, Calif.
Tozer, D.C., Nol, E., Burke, D.M., Elliott, K.E., and Falk, K.J. 2009.
Predation by bears on woodpecker nests: are nestling begging and
habitat choice risky business? Auk, 126(2): 300–309. doi:10.1525/
auk.2009.07204.
Ulyshen, M.D., Hanula, J.L., Horn, S., Kilgo, J.C., and Moorman, C.
E. 2004. Spatial and temporal patterns of beetles associated with
coarse woody debris in managed bottomland hardwood forests.
For. Ecol. Manag. 190: 259–272.
Van Horne, B. 1983. Density as a misleading indicator of habitat
quality. J. Wildl. Manag. 47(4): 893–901. doi:10.2307/3808148.
Van Wagner, C.E. 1968. The intercept method of forest fuel
sampling. For. Sci. 14:20–26.
Vanderwel, M.C., Caspersen, J.P., and Woods, M.E. 2006. Snag
dynamics in partially harvested and unmanaged northern hard-
wood forests. Can. J. For. Res. 36(11): 2769–2779. doi:10.1139/
X06-173.
Walters, E.L., Miller, E.H., and Lowther, P.E. 2002. Yellow-bellied
Sapsucker (Sphyrapicus varius). In The birds of North America.
The Birds of North America, Inc., Philadelphia, Pa.
Straus et al. 1017
Published by NRC Research Press
Can. J. For. Res. Downloaded from www.nrcresearchpress.com by Dalian Nationalities University on 06/05/13
For personal use only.
