Transcript Smith, Keeton, Twery, and Tobi_ESA 2007

Understory plant responses to
uneven-aged forestry alternatives in
northern hardwood-conifer forests
Kimberly J. Smith1, William S. Keeton1,
Mark J. Twery1,2, and Donald R. Tobi1
1Rubenstein
School of Environment and Natural Resources,
University of Vermont
2USDA Forest Service, Northeastern Research Station
 Structure-based forestry (Franklin et al. 2002; Keeton 2006):
Managing for multiple structural conditions
representative of natural successional dynamics
www.dred.state.nh.us
 Disturbance-based forestry
(Mitchell et al. 2002; Seymour et al. 2002):
Developing silvicultural
systems based on the scale and
pattern of natural disturbances
Stovall 2006
Stovall 2006
The Vermont Forest Ecosystem
Management Demonstration Project
(FEMDP)
 Structure and disturbance-based forestry practices can:
 Sustain a broad array of biodiversity and
ecosystem functions
 Provide for profitable timber management
Silvicultural Treatments (FEMDP)
 Single-tree selection (STS) and group selection (GS)
 Increased structural retention
 Modeled after the scale and pattern of natural disturbance
(based on the findings of Seymour et al. 2002)
 Structural Complexity Enhancement (SCE)
 Promote accelerated development of late-successional forest
characteristics:
 Increasing vertical and horizontal heterogeneity
 Elevating large snag and coarse woody debris (CWD) densities
 Reallocating basal area to larger diameter classes
 Enhance ecosystem services, including:
 Late-successional wildlife habitat (McGee et al. 1999)
 Carbon storage (Harmon et al. 1990)
 Riparian function (Keeton et al. 2007)
FEMDP Research
 Monitoring indicators
of biodiversity
response, including:





Birds
Small mammals
Amphibians
Soil invertebrates
Vegetation
© Al Sheldon
Previous Research
 Even-aged treatments
 Plantation forestry
(Ramovs and Roberts 2005)
 Clearcutting
(Gilliam et al. 1995; Halpern and Spies 1995;
Liu and Ashton 1999)
 Uneven-aged treatments
www.gov.ns.ca
 Single-tree and group selection
(Jenkins and Parker 1999; Scheller and
Mladenoff 2002; Kern 2006)
 Variable retention
(Halpern et al. 2005)
 Experimental canopy gaps
(Collins and Pickett 1988; Gray and Spies 1997)
www.gov.ns.ca
Hypotheses
 Uneven-aged, low-intensity silvicultural
systems can maintain understory plant
diversity and support late-successional species
 Retaining and enhancing stand structural
complexity can increase understory plant
diversity
 Plant responses are influenced by interactions
between canopy structure, soils, and climate
processes
Methods
 Study Areas:
 Mount Mansfield State Forest
 Jericho Research Forest
 Paul Smith’s College
 Mature, northern
hardwood stands with a
documented history of
previous timber management
Experimental Design
Mt. Mansfield State Forest
 2 ha treatment units
 0.1 ha permanent plots
(overstory structure)
1 m2 (vegetation)
 13 vegetation quadrats (1 m2)
 4 soil subplots (2 m2)
2 m2 (soils)
Data Collection
 Palmer Drought Severity Index (PDSI): multiple
climatic parameters condensed into a single index
• Temperature
• Evapotranspiration
• Precipitation
• Soil moisture loss
• Soil moisture recharge
• Runoff
PDSI
Analysis
Habitat guilds as defined by Ramovs and Roberts (2005):
 Early-successional
 Intermediate
 Late-successional
www.nrs.fs.fed.us
Response variables
 Diversity:
 Hill’s series of diversity indices (Hill 1973)
 Species richness
 Exponential Shannon Index
 Reciprocal Simpson Index
 Abundance:
 % cover by species
Analysis of Treatment Effects
Diversity and abundance:
 Linear mixed effects model
 fixed effects– treatment, site, year
 random effects– plots, units
 ANOVA models
 Pre- to post-harvest change of unit level means
 Test for differences among treatments
 Analyses performed for all species and by habitat guilds
Analysis of Treatment Effects
Compositional changes:
 Non-metric multidimensional scaling (NMS)
 interpret compositional patterns among
treatment units
 Multi-response permutation procedure (MRPP)
 pre- to post-harvest differences within treatments
 differences among treatments before and after harvest
 Locally impacted species
Sub-analysis of soil properties and overstory structure
 Five soil variables:
 % OM, % N, Ca, P, pH
 Linear mixed effects model
 fixed effects– treatment, site, year
 random effects– plots, units
 covariates- % OM, % N, Ca, P
 Pre- to post-harvest % change
 Soil properties
 Curtis’ relative density (RD)
 Diversity and abundance responses
 Explanatory variables related to
ordination axes in NMS
www.forestryimages.org
Analysis of Moisture Stress
 Palmer Drought Severity Index (PDSI)
 Period 1 (PDSI_1) = July-September, previous year
 Period 2 (PDSI_2) = April-June, current year
 Standardized understory responses to +/- unit mean
 Simple linear regressions
Results: Effects of Treatment
Treatment*time
p
All Species
Richness
Diversity_1
Diversity_2
% cover
<0.001
<0.001
0.004
<0.001
Late-Successional
Species
Richness
Diversity_1
Diversity_2
% cover
0.002
0.024
0.047
<0.001
Intermediate Species
Richness
Diversity_1
Diversity_2
% cover
0.348
0.117
0.217
0.467
Early-Successional
Species
Richness
Diversity_1
Diversity_2
% cover
<0.001
<0.001
0.001
<0.001
 Mixed effects model
 Understory responses
significantly affected
by treatment*time
interaction
Effects of Treatment: All Species
12
Control
4
8
SCE
3
Single-tree selection
Diversity
Richness
Group selection
4
0
2
1
0
-1
-4
-1
1
2
Year
 Richness:
p = 0.032
 SCE > CON
3
4
-1
1
2
3
Year
 Shannon Index:
p = 0.004
 SCE > CON
4
Effects of Treatment: Late-successional Species
2
1.2
0.9
Diversity
Richness
4
Control
Group selection
SCE
Single-tree selection
0
-2
0.6
0.3
0
-0.3
-0.6
-4
-1
1
2
Year
 Richness:
p = 0.012
 SCE > GS
3
4
-1
1
2
3
Year
 Shannon Index:
p = 0.009
 SCE > CON
4
Effects of Treatment: Species Composition
Pre-harvest
CON
GS
SCE
MRPP:
A = 0.009
p = 0.320
Axis 2 (33.2%)
STS
Axis 1 (44.8%)
Effects of Treatment: Species Composition
Post-harvest
CON
GS
SCE
A = 0.026
p = 0.142
Axis 2
MRPP:
(25.1%)
STS
Axis 1 (38.2%)
Locally Extirpated Species
Percent Species Lost by Treatment
% Species Lost
16
12
ANOVA:
8
p = 0.07
4
0
GS
STS
SCE
Treatment
CON
Locally Extirpated Species
Scientific Name
Actaea alba
Adiantum pedatum
Aralia nudicaulis
Arisaema triphyllum
Asarum canadense
Coptis trifolia
Eupatorium rugosum
Lonicera canadensis
Medeola virginiana
Osmunda claytoniana
Oxalis acetosella
Panax trifolia
Polygonatum pubescens
Polygonum cilinode
Pyrola elliptica
Sambucus racemosa
Smilacina racemosa
Trientalis borealis
Trillium erectum
Viburnum alnifolium
GS
Treatment*
SCE STS
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Fruit type
Dispersal
Habitat preference
perennial herb
fern
perennial herb
perennial herb
perennial herb
perennial herb
perennial herb
shrub
perennial herb
fern
perennial herb
perennial herb
perennial herb
perennial herb
perennial herb
shrub
perennial herb
perennial herb
perennial herb
shrub
berry
spore
drupe
berry
capsule
follicle
achene
berry
berry
spore
capsule
drupe
berry
achene
capsule
berry
berry
capsule
berry
drupe
biotic
abiotic
biotic
biotic
biotic
biotic
abiotic
biotic
biotic
abiotic
biotic
biotic
biotic
biotic
abiotic
biotic
biotic
biotic
biotic
biotic
late-successional
late-successional
late-successional
late-successional
late-successional
late-successional
intermediate
late-successional
late-successional
late-successional
late-successional
late-successional
late-successional
intermediate
late-successional
early-successional
late-successional
late-successional
late-successional
late-successional
CON
X
X
X
X
Growth form
*X indicates the species was present before harvest and absent from two or more units of this treatment type during all four
post-harvest years.
Effects of Soil Properties and Overstory Structure
 Mixed effects model:
 Significant effect of treatment
 Soil covariates not significant
 Exceptions: Intermediate species- % OM and % N
 Correlations:
 Unit level (coarse scale) and plot level (fine scale)
 Δ relative density related to Δ responses
 Δ soil properties generally not related to Δ responses
Effects of Soil Properties and Overstory Structure
Pre-harvest
CON
GS
SCE
RD:
τ = 0.439
%OM:
τ = -0.336
Axis 2 (33.2%)
Axis 2
STS
RD
Ca
P
N OM
Axis 1 (44.8%)
Effects of Soil Properties and Overstory Structure
Axis 21
Axis
%N:
Ca:
-0.336
ττ==0.520
SCE
STS
(25.1%)
%OM:
%N:
-0.362
ττ==0.388
GS
OM
Ca
NP
Axis 2
%OM:
RD:
0.368
ττ==0.441
Post-harvest
CON
RD
P:
τ = 0.494
Axis 1 (38.2%)
Effect of Moisture Stress
 PDSI_2
PDSI_1
(April-June
of current
(July-September
of
year)
previous year)
8
 Not
Not related
related to
to understory
understory
responses
responsesin controls or
treatments
GS:
p = 0.012, r2 = 0.747
STS: p = 0.018, r2 = 0.703
SCE: p = 0.024, r2 = 0.602
6
4
PDSI
 Exception:
Late-successional richness
FEMDP
VT
FERDA
NY
2
0
-2
-4
-3
-2
-1
1
2
year
3
4
Discussion
 Uneven-aged, low-intensity silvicultural systems
with increased structural retention maintain
understory plant species diversity
12
Control
4
8
SCE
3
Single-tree selection
Diversity
Richness
Group selection
4
0
2
1
0
-1
-4
-1
1
2
3
4
Year
-1
1
2
Year
All Species
3
4
Discussion
 Post-harvest increases in diversity
(Gilliam et al. 1995; Halpern and Spies 1995; Jenkins and Parker 1999)
 Increases in early-successional or ruderal species may
mask the loss of late-successional species
(Halpern and Spies 1995)
 Results of this study:
 Compositional differences not significant
 Pre- to post-harvest, within treatments
 Post-harvest, among treatments
Discussion
 Level and spatial pattern of retention may
influence loss of species following harvest
Percent Species Lost by Treatment
% Species Lost
16
12
ANOVA:
8
p = 0.07
4
0
GS
STS
SCE
Treatment
CON
Discussion
 Uneven-aged, low-intensity silvicultural systems
with increased structural retention support latesuccessional species
2
1.2
0.9
Diversity
Richness
4
Control
Group selection
SCE
Single-tree selection
0
-2
0.6
0.3
0
-0.3
-0.6
-4
-1
1
2
3
4
-1
Year
1
2
Year
Late-successional Species
3
4
Discussion
 Increasing post-harvest
stand structural complexity may
lead to increased plant diversity
 Previous studies:
 Plant diversity increases
with light availability
(e.g. Brosofske 2001)
 Results of this study:
 SCE:
 Lower light availability
 Greater light heterogeneity
 Greater increases in diversity
Stovall 2006
Discussion
 Retaining or enhancing
stand
Late-successional
structural complexity
northern
hardwood
forests
may increase
diversity:
characterized by:
 light heterogeneity
 heterogeneous light
environment
microhabitat diversity
 retaining
microsite canopy
variability
(Scheller and Mladenoff 2002)
Discussion
 Understory plant responses are influenced by interactions
between canopy structure, soils, and climate processes
 Moisture availability influences patterns of understory
vegetation
(Huebner et al. 1995; Hutchinson et al. 1999; Kolb and Diekmann 2004)
 Results of this study:
 Drought index correlated to late-successional richness in
treatment units, but not in controls
 Canopy removal may increase susceptibility of understory to
drought stress
Discussion
Percent Increase of Understory Responses in SCE Units
recovery
actual
60
8
FEMDP
VT
FERDA
NY
6
40
4
PDSI
% increase
 Post-harvest
diversity increases
in SCE units may be
partially due to
drought recovery
20
2
0
-2
-4
0
-3
-2
Richness
-1
1
year
Shannon
Diversity
2
3
4
Simpson
Diversity
Discussion
 Understory plant responses are influenced by interactions
between canopy structure, soils, and climate processes
 Soil nutrient availability influences patterns of
understory vegetation
(Kolb and Diekmann 2004; Fraterrigo et al. 2006)
 Harvesting overstory can affect soil nutrient availability
(Johnson et al. 1997; Elliott and Knoepp 2005)
 Results of this study:
 Overstory-vegetation relationships consistently
significant
 Soil-vegetation relationships highly variable
Conclusions
 Uneven-aged, low-intensity silvicultural treatments
with increased structural retention can maintain
understory plant diversity and support late-successional
species in northern hardwood-conifer forests during the
initial post-harvest recovery period
 Treatments that enhance stand structural complexity
may increase understory plant diversity by increasing
the heterogeneity of light and microsite variability
 Level and spatial pattern of retention may be
important to preserving understory plant species
 Plant responses are primarily influenced by changes
in overstory structure
Management Implications
 Sustainable forest management:
 Maintains biodiversity and ecosystem functioning
 Provides for timber harvest
 Best approaches for conserving understory plant diversity:
 Retain post-harvest structure
 Retain biological legacies (see Franklin et al. 2002)
 Enhance stand structural complexity
Acknowledgements
Vermont Monitoring Cooperative
 USDA McIntire-Stennis Forest Research Program
 USDA National Research Initiative
 Northeastern States Research Cooperative
Other helpful folks:
 Field crews of 2001-2006
 Alan Howard, UVM statistical counseling clinic
Questions?