Long-term Responses of Planted and Naturally Regenerating Conifers to Variable-Retention Harvest
in the Pacific Northwest, U.S.A.
Lauren S. Urgenson1, Charles B. Halpern1, Paul D. Anderson2
1
University of Washington, Seattle, WA; 2USFS Pacific Northwest Research Station
Experimental Treatments
Cumulative mortality
40%
15%
Cleared areas between five 1-ha residual patches
Cleared areas between two 1-ha residual patches
Dispersed trees with basal area equivalent to 40%A
Dispersed trees with basal area equivalent to 15%A
Systematic sampling grid: 40 × 40 m spacing
Numbers of plots: n = 12 (40%A-c) and 22 (15%A-c)
n = 32 (40%D and 15%D)
40%A-c, 15%A-c
Measurements: mortality and height growth of four conifer species
(species and density vary by block)
Sampling periods: early (1 or 2 yr), late (5 to 11-12 yr), and the full
study period
H2 & H3. Natural regeneration
●
15%D
Experimental treatments at Butte, Washington. The 15%A
treatment is out of the image to the west. The 75%A treatment
(three circular 1-ha gaps) was not included in this study.
Treatment dimensions are 320 m x 400 m.
DP
CF
BU
LW
PH
36-52
Tshe ground-based; submerchantable stems
cut if damaged
1460-1710 34-62 165
258-475 72-106 Abco helicopter
210-275 28-52
65
221-562 54-73 Tshe suspension cables
975-1280 40-53 70-80 759-1781 48-65 Tshe helicopter
825-975 40-66 140-170 182-335 61-77 Abgr helicopter
850-1035 9-33 110-140 512-1005 59-87 Abam ground-based; all submerchantable stems cut
*Abam = Abies amabilis, Abco = A. concolor, Abgr = A. grandis, Tshe = Tsuga
heterophylla
0
2
4
8
10
12
0.8
0.7
not tested
0.5
0.4
0.3
ns
Early
6
Pinus ponderosa
(n = 2)
0.6
Total
0.2
1
4
BU
3
2
1
2
4
6
8
10
12
0
2
4
6
8
10
LW
Measurements: density of seedlings (<10 cm) and saplings
(>10 cm, <5 cm dbh) by species or seral group
Sampling periods: pre-treatment, post-treatment (1 or 2 yr), and final
(11 or 12 yr)
● Local (plot-scale) predictors of density: overstory density and basal area,
shrub cover, slash cover and depth
●
50
Pinus monitcola
(n = 5)
40%D
40%A-c
15%D
15%A-c
40
●
●
●
H1. GLMM and a priori contrasts testing effects of retention level and pattern
on planted species survival and growth for early, late, and full study periods
H2. GLMM and a priori contrasts testing effects of retention level and pattern
on natural regeneration density at final sampling
H3. AIC and model averaging to identify the best supported models explaining
final density of natural regeneration in the context of retention treatments
and post-treatment densities
50
10
The results of this large-scale experiment illustrate how level and pattern of overstory
retention influence early regeneration processes in Pseudotsuga-dominated forests of
the Pacific Northwest. By testing a common set of treatments on forests of differing
structure, they highlight critical sources of variation in regeneration response that can
have implications for the longer-term development of these forests.
5
40%D
40%A-c
15%D
15%A-c
H2. Density and species composition of natural regeneration. Density and
species composition of natural regeneration was highly variable within and
among blocks before treatment. Much of this variation persisted through harvest.
Pseudotsuga menziesii
(n = 6)
a b b b
40
a b b b
ab b b
30
30
Density and seral composition of natural regeneration
a b b b
Total seedlings
20
10
20
10
ns
0
50
ns
0
Early
Late
Mean
Early
Abies spp.
(n = 3).
50
Late
Mean
Pinus ponderosa
(n = 2)
30
b
30
b
ab
ab
20
a a
20
aa
10
10
ns
0
Early
Late
Mean
Early
Late
Mean
H1. Growth of planted seedling growth. Early growth (1-2 yr) growth
was unaffected by treatment. Consistent with expectation, late growth
(5-12 yr) was greatly reduced at higher levels of dispersed retention
(40%D).
Late-seral seedlings
ns
20
ns
ns
10
Pre
40
30
Pseudotsuga seedlings
40%D
40%A-c
15%D
15%A-c
40
0
not tested
40
0
Analyses
Conclusions and Management Implications
Abies amabilis
Pinus monticola
Pseudotsuga menziesii
Thuja plicata
Tsuga heterophylla
Other species
15
0
50
14
Post
Final
Total saplings
Pre
Post
Final
Pseudotsuga saplings
Pre
Post
Final
Late-seral saplings
12
10
a
8
ab
bc
6
a
c
bb
4
b
a ab
b ab
2
0
Pre
0.10
0.13
0.11
0.08
0
PH
20
Growth of planted seedlings
Pseudo-R2
* BA = basal area, SHC = shrub cover, SLC = slash cover, SLD = slash
depth, TPH = trees per ha
Abies grandis
Pinus monticola
Pseudotsuga menziesii
Taxus brevifolia
Thuja plicata
Tsuga heterophylla
Other species
Number of growing seasons (yr)
H1. Mortality of planted seedlings. Consistent with expectation, early
mortality (1-2 yr) was reduced in dispersed treatments (Abies) and at
higher levels of retention (Pinus and Pseudotsuga) compared to the cleared
areas of aggregated treatments. 50% of this mortality occurred in the first
growing season. Counter to expectation, late mortality (5-12 yr) did not
differ among treatments.
Shrub cover (SHC)
Predictors*
H3. Local predictors of natural regeneration density. Results were consistent with
expectation. Sapling densities were negatively correlated with slash cover (SLC) or slash
depth (SLD). Negative correlations for saplings—but not seedlings—suggest a legacy of
initial slash-loading effects, reducing establishment soon after harvest. Density of
Pseudotsuga was negatively correlated with shrub cover (SHC) but density of late-seral
species was not.
0
0.5
12
Slash cover (SLC)
Slash depth (SLD)
Seral group
Saplings (>10 cm tall, Pseudotsuga +TPH, –SLD, –SHC
<5 cm dbh)
Late seral
+TPH, –SLC
Seedlings (<10 cm tall) Pseudotsuga –SHC, –TPH, +BA
Late seral
None
Abies amabilis
Abies procera
Pinus monticola
Pseudotsuga menziesii
Thuja plicata
Tsuga heterophylla
Other species
0.0
0
Size class
0 0
0
1.0
0.1
0.0
Annual height growth (cm / yr)
●
●
110-130 310-500
12
b
b
40%D, 15%D
H1. Planted seedlings
4-7
10
a
a
Late
BU
945-1310
8
0.1
PH LW
WF
2
Pre
Po
Finst
al
●
Tree
Basal
Yarding method and
Elevation Slope Stand density
area Forest treatment of subBlock
(m)
(%) age (yr) (no./ha) (m 2/ha) zone* mechantable stems
6
Trees per ha (TPH)
Basal area (BA)
Abies concolor
Abies magnifica var. shastensis
Calocedrus decurrens
Pinus monticola
Pseudotsuga menziesii
Taxus brevifolia
Tsuga heterophylla
DP
3
Pre
Po
Finst
al
●
40%A
4
Pre
Po
Finst
al
Sampling design
WF
DP
0
0.0
CF
40%D
4
b
b
0.2
5
ns
0.0
a
a
0.3
ab
ab
b
0.1
0.5
0.4
a
0.3
0.2
0.7
0.6
0.4
Number of growing seasons (yr)
Field Methods and Analyses
The experiment consists of operational-scale (13-ha) harvest units replicated at
six locations in western Oregon and Washington. Treatments contrast both the
level of retention (15% vs. 40% of original basal area) and its spatial
distribution (trees dispersed vs. aggregated in 1-ha patches). Replication
across a broad range of latitudes and forest zones offers a rare opportunity to
study responses to a common set of treatments in a diversity of physical and
biotic settings.
ns
Abies spp.
(n = 3)
0.8
Early
0.5
Total
10
Pre
Po
Finst
al
Locations and characteristics of the six experimental blocks
15%D
2
Total
ns
Thousand seedlings / ha
Federal standards and guidelines for regeneration harvests in the Pacific
Northwest require at least 15% overstory retention in each harvest unit, with
70% of this in the form of large (0.2-1.0 ha) aggregates. The Demonstration of
Ecosystem Management Options (DEMO) Study was initiated in 1994 to test
the ecological and silvicultural implications of level and pattern of retention,
including the sufficiency of the 15% minimum standard.
0.2
Late
0.6
Early
a
ab
ab
b
Abies concolor
Pinus ponderosa
Pseudotsuga menziesii
Tsuga heterophylla
Other species
20
15
0.7
0.3
WF
Pseudotsuga menziesii
(n = 6)
0.8
Late
ns
0.4
0
40%D
Results H3.
Local predictors of natural regeneration density
Thousand saplings / ha
Study Areas
0.5
0.0
40%A-c
15%A-c
40%D
15%D
The DEMO Study
0.6
0.1
H3. Local predictors of natural regeneration density. Local (plot-scale)
density of natural regeneration will be reduced by greater cover and depth of
logging slash and by greater cover of competing shrubs. Effects will be stronger
for early-seral species that prefer mineral substrates for germination and higher
levels of understory light. Late-seral species will benefit from greater overstory
influence (density or basal area), consistent with H2.
Results H2.
Density and species composition of natural regeneration
40%D
40%A-c
15%D
15%A-c
0.7
Annual height growth (cm / yr)
We present results from long-term (12-yr) studies of conifer regeneration as part
of a regional-scale experiment in variable-retention harvest in mature coniferous
forests of the Pacific Northwest, U.S.A.—the DEMO Study. We examine
mortality and height growth of planted conifers and density and seral composition
of natural regeneration in treatments with differing levels and patterns of
retention across a range of latitudes and forest zones. We also assess withintreatment (local) influences on natural regeneration, including variation in
overstory structure, competing vegetation, and slash accumulation.
Results H1.
Mortality of planted seedlings
Pinus monitcola
(n = 5)
0.8
Cumulative mortality
H2. Density and composition of natural regeneration. Responses will differ
for early-seral (shade-intolerant) and late-seral (shade-tolerant) species, consistent
with their physiological requirements. Greater retention and presence of a
dispersed overstory will inhibit density of early-seral species, including
Pseudotsuga menziesii, the dominant canopy species in this system. However,
both factors will increase the density of late-seral species—enhancing survival of
advanced regeneration and facilitating post-harvest establishment.
Aggregated
Partial retention of the overstory is hypothesized to have both direct and indirect
effects on the survival, growth, density, and composition of regenerating trees.
Two elements of residual forest structure—the level (amount) and spatial pattern
of retained trees—are likely to play important roles in the post-harvest
regeneration of forests, but long-term studies of regeneration responses are rare.
15%A-c
40%A-c
Density (thousand saplings / ha)
Hypotheses
H1. Mortality and growth of planted seedlings. Early survival of planted trees
will be greater in dispersed treatments and at higher levels of retention, consistent
with the amelioration of microclimatic stress by retained trees in the post-harvest
environment. Longer-term growth and survival will be reduced in these
environments, particularly for shade-intolerant species, as resource limitations
(shading and root competition) become more important.
Dispersed
Introduction
Variable-retention harvest has become a key component of ecologically
sustainable forest management in many regions of the world. It emphasizes the
retention of forest structure through harvest to mimic processes and outcomes of
natural disturbance and succession. In contrast to clearcut logging, an implicit
goal of variable retention is to balance the ecological and economic values of
managed forests.
Post
Final
Pre
Post
Final
Pre
Post
Final
H2. Density and seral composition of natural regeneration. Seedling densities
did not differ among treatments at final sampling, although mean densities tended
to be greater in dispersed treatments. Sapling densities were 1.5-2.5× greater in
dispersed than in cleared areas of aggregated treatments. Consistent with
expectation, low-level dispersed retention (15%D) promoted Pseudotsuga, the
early-seral dominant, but high levels of dispersed retention (40%D) favored
survival and subsequent establishment of late-seral conifers.
● Survival and growth of planted seedlings. Early (1-2 yr) survival was improved
by greater retention of dispersed trees (40%). However, longer-term survival was not,
and growth was distinctly reduced. Without further intervention or natural disturbance
to enhance light availability in 40%D, planted seedlings are unlikely to contribute to
future overstory development. In contrast, growth was not reduced at lower retention
(15%D). Future observations are needed to assess the potential for planted trees to
contribute to overstory development in these forests.
● Composition of natural regeneration. Level and pattern of retention can interact
to influence the seral composition of regenerating trees. At low density, dispersed
retention promoted establishment of Pseudotsuga, presumably by enhancing local seed
rain within a relatively high-light environment. At higher density it favored late-seral
conifers, increasing survival of advanced regeneration and facilitating establishment
and growth of shade-tolerant species.
● Density of natural regeneration. Treatments with dispersed retention supported
significantly greater (1.5-2.5×) regeneration than did the cleared areas of aggregated
treatments. Final densities in dispersed treatments greatly exceeded minimum
stocking levels (312/ha) in all but one experimental unit (15%D at DP). Depending on
management objectives, dispersed treatments that regenerate to high local densities
(e.g., WF and PH) may require thinning to reduce competitive interactions and to
prevent extinction of the understory.
Although densities in the cleared areas of aggregated treatments generally exceeded
minimum stocking, they often fell below this threshold (e.g., DP and LW). Dispersal
limitation, lack of advanced regeneration, deep accumulations of slash, and rapid
recovery of a dense shrub layer may constrain regeneration in these forests. The slow
pace of conifer establishment in these clearings highlights the potential to create and
maintain early-seral habitats using retention harvests—habitats that are underrepresented in landscapes traditionally managed to achieve rapid reforestation.
Acknowledgement: This a product of the Demonstration of Ecosystem Management Options
(DEMO) Study, a joint effort of the USDA Forest Service Region 6 and Pacific Northwest
Research Station. Research partners include the University of Washington, Oregon State
University, University of Oregon, Gifford Pinchot and Umpqua National Forests, and the
Washington State Department of Natural Resources (http://www.fs.fed.us/pnw/rmp/demo/).
Funding was provided by the USDA Forest Service, PNW Research Station.