Transcript If You Build It Will They Come?

If You Build It Will They Come?
Associations Between Birds and Vegetation Structure During
Ten Years After Thinning
Sveta Yegorova, Drs. Matt Betts , Joan Hagar* and Klaus Puettmann
Department of Forest Ecosystems and Society, Oregon State University
Golden-crowned Kinglet
*U.S. Geological Survey, Forest and Rangeland Ecosystem Science Center, Corvallis, OR
Figure 1. MacGillivray’s Warbler: Occupancy by Shrub Cover
1.5
0.0
0.5
GOF
1.0
1.5
2006
20
Year
40
60
80 100
Abundance
MacGillivray’s Warbler
500 1000
2000
500 1000
2000
1.5
0.5
GOF
1.5
0.5
GOF
0
Figure 4
2002
2006
50
Year
500 1000
500 1000
2000
3.5
3.0
2.5
1.5
1.0
Hammond’s Flycatcher
0
500 1000
2000
Figure 5
Year
30
40
50
Cumulative GOF Graphs
GOF for Abundant Species
60
GOF Less Abundant Species
Abundance
5
Shrub Cover
When pooled across species,
GOF is associated negatively
with abundance for relatively
Figure 6
abundant species (Fig 6).
GOF did not correlate with time since thinning (data not shown).
1.0
3
1998
0.0
1
2
Figure 2. Golden-crowned Kinglet: Occupancy by Conifer Density
GOF
4
Figure 1
Hammond’s Flycatcher
2.0
GOF
3.0
2.5
2.0
1.5
1.0
GOF
0.0 0.2 0.4 0.6 0.8 1.0
0
Swainson’s Thrush
Treatment GOF vs Abundance
3.5
2007
0.0 0.2 0.4 0.6 0.8 1.0
2006
2000
110
Hammond’s Flycatcher
1998 2000 2002 2004 2006
0
90
Abundance
Treatment GOF vs Year
2001
70
3.0
0
1998
2.0
2000
Shrub GOF vs Abundance
6
500 1000
1999
0.8
0.8
0.8
0.4
0.4
0.4
Occupancy
150
250
Abundance
350
50
100
150
Abundance
0
40
80
120
0.0
0.0
0.0
Golden-crowned Kinglet
0
40
80
120
0 20
60
100
Coniferous Stem Count
0 20
Figure 2
60
100
0.4
0.0
0.4
0.0
0.4
0.8
2007
0.8
2006
0.8
2001
0.0
Occurrence of eight species was modeled as a function of
vegetation structure or treatment (fixed effects) and block and
stand (random effects) using mixed-effects logistic regression.
2002
Shrub Cover
Occupancy
Analysis
Figure 3
1999
650
• 30-50 yr old Douglas-fir stands
• Mid-elevation (500-900ft) forests on west slopes of the central
Cascade Mountains in Oregon
• Detailed understory and overstory vegetation measurements
taken four times since thinning
• Breeding season point count surveys conducted six times
1998
Shrub GOF by Year
0.0 0.2 0.4 0.6 0.8 1.0
0.0 0.2 0.4 0.6 0.8 1.0
Occupancy
0
1997
Control
Shrub GOF vs Abundance
GOF
Heavy
Light
Light with Gaps
Density (trees per
hectare)
125
250-300
250-300 with 0.2 ha gaps
Occupancy vs Shrub Cover
0.0 0.2 0.4 0.6 0.8 1.0
1997
psi
Methods
Shrub GOF by Year
Swainson’s Thrush
Occupancy
I. If vegetation drives bird occurrence, then
a. There will be strong associations between bird occurrence
and vegetation structure immediately after thinning
b. These associations will be consistent over time
II. If other factors are contributing to habitat selection there may
be high year-to-year variability in bird-vegetation associations
MacGillvray’s Warbler
1.0
Vegetation effects varied among years and were
inconsistent in size
II. In some years vegetation models performed well (e.g.,
MacGillvray’s Warbler and Golden-crowned Kinglet, see
figures 1 and 2 below)
III. Changing population sizes may explain varying model
fits
0.5
I.
We used the ratio of variable estimate to its standard error
(GOF=Estimate/SE(Estimate)) to quantify the variability of the
size of the effect. GOF appears to be highly variable even for
species associated with vegetation structures (as in Fig. 1 and
2) in some years. See figures 3-6 for illustration.
0.0
Results
Hypotheses
Treatment
Quantifying Variability of
Vegetation/Treatment Effect on Occurrence
We selected best performing model, using Akaike
Information Criterion (AIC), from a priori set pool of
models that reflects the hypotheses stated above.
0.0 0.2 0.4 0.6 0.8 1.0
• Forest thinning encourages understory and overstory
development in young Douglas-fir forests
• Effects of thinning on birds have been shown to vary by
species and with time since thinning
• These effects are assumed to be mediated by the residual
vegetation structure but this assumption is rarely tested
• Alternatively, habitat selection may be driven by other
(non-vegetation) factors, including previous experience,
social information, and density of conspecifics
• Therefore disturbance (thinning) may cause a lag in birdvegetation association
• Large population size may result in poor fit between bird
occurrence and vegetation structure due to population
spillover into low quality habitat
Analysis continued
GOF
Background
Corvallis, Oregon 97330 USAUSGS, Corvallis, Oregon
0 20
60
100
Coniferous Stem Count
0 20
60
100
Discussion
Bird-vegetation associations appear to be highly variable in this longterm (10 years) study. Single-season bird surveys are unlikely to
properly reflect an underlying relationship between bird occurrence
and vegetation structure. Our results suggest that population size,
for abundant species, and possibly social factors interact with
vegetation structure to produce observed bird occurrence patterns.