Content uploaded by Stephen Pecot
Author content
All content in this area was uploaded by Stephen Pecot on Jul 02, 2014
Content may be subject to copyright.
Gaps in a gappy forest: plant resources, longleaf
pine regeneration, and understory response to
tree removal in longleaf pine savannas
John P. McGuire, Robert J. Mitchell, E. Barry Moser, Stephen D. Pecot,
Dean H. Gjerstad, and Craig W. Hedman
Abstract: Resource availability and planted longleaf pine (Pinus palustris Mill.) seedling and understory vegetation re
-
sponse within and among three sizes of experimentally created canopy gaps (0.11, 0.41, 1.63 ha) in a mature longleaf
pine savanna were investigated for 2 years. Longleaf pine seedlings and understory vegetation showed increased growth
in gaps created by tree removal. Longleaf pine seedling growth within gaps was maximized approximately 18 m from
the uncut savanna. Increased longleaf pine seedling survival under the uncut savanna canopy observed after the first
year suggests that the overstory may facilitate establishment of longleaf pine seedlings rather than reduce survival
through competition. Despite the relative openness of the uncut longleaf pine forest, light quantity was increased by
tree removal. Light was also the resource most strongly correlated with seedling and understory vegetation growth. Al
-
though net N mineralization was correlated to seedling response, the amount of variation explained was low relative to
light. Belowground (root) gaps were not strong, in part because of non-pine understory roots increasing in biomass fol
-
lowing tree removal. These results suggest that regeneration of longleaf pine may be maximized within gap sizes as
small as approximately 0.10 ha, due largely to increases in light availability.
Résumé : Pendant 2 ans, les auteurs ont étudié la disponibilité des ressources et la réaction de la végétation de sous-
étage et de semis transplantés de pin des marais (Pinus palustris Mill.) à l’intérieur et entre des trouées de trois dimen-
sions (0,11, 0,41 et 1,63 ha) créées artificiellement dans une savane de pin des marais. Les semis de pin des marais et
la végétation de sous-étage ont eu une croissance accrue dans les trouées créées par l’enlèvement des arbres. La crois-
sance des semis de pin des marais dans les trouées était maximale à 18 m approximativement de la limite de la savane
non coupée. L’augmentation de la survie des semis de pin des marais, qui a été observée après un an sous couvert
dans la savane, laisse croire que l’étage dominant pourrait faciliter leur établissement plutôt que de réduire leur survie
à cause de la compétition. Malgré que la forêt non coupée de pin des marais soit relativement ouverte, l’enlèvement
des arbres augmente la quantité de lumière. La lumière est la ressource la plus étroitement corrélée avec la croissance
des semis et de la végétation de sous-étage. Quoique la minéralisation nette de N soit corrélée avec la réaction des se
-
mis, la part de variation expliquée est faible comparativement à la lumière. Les trouées souterraines (racines) ne sont
pas importantes, en partie à cause de l’augmentation de la biomasse racinaire des plantes de sous-étage autres que le
pin des marais, suite à l’enlèvement des arbres. Ces résultats indiquent que la régénération de pin des marais pourrait
être maximisée par des trouées aussi petites qu’environ 0,10 ha, surtout à cause de l’augmentation de la disponibilité
de la lumière.
[Traduit par la Rédaction] McGuire et al. 778
Introduction
While closed-canopy forests and tropical savannas have
been well studied with respect to how the structure of woody
plants influences resource availability and vegetation re
-
sponse (Belsky and Canham 1994), temperate savannas of
the southeastern Coastal Plain are not as well understood
(McPherson 1997). In the absence of disturbance, closed-
canopy forests can be viewed as a matrix of trees in which
most of the crowns overlap to form a continuous canopy
with only small holes or gaps (Endler 1993; Franklin 1985;
Franklin et al. 1987). When disturbances create gaps in these
forests, a pronounced disparity is also created between the
environments associated with the intact canopy and that of
the gap (Van Pelt and Franklin 1999). Savannas are an open,
grassland landscape interspersed with trees or shrubs in vari
-
able density but with crowns that do not overlap (Belsky and
Canham 1994; Scholes and Archer 1997). The presence of
Can. J. For. Res. 31: 765–778 (2001) © 2001 NRC Canada
765
DOI: 10.1139/cjfr-31-5-765
Received May 11, 2000. Accepted December 20, 2000. Published on the NRC Research Press Web site on April 21, 2001.
J.P. McGuire and D.H. Gjerstad. School of Forestry, Auburn University, Auburn, AL 36849-5418, U.S.A.
R.J. Mitchell
1
and S.D. Pecot. Joseph W. Jones Ecological Research Center, Newton, GA 31770, U.S.A.
E.B. Moser. Department of Experimental Statistics, Louisiana Agricultural Experiment Station, Louisiana State A&M University,
Baton Rouge, LA 70803-5606, U.S.A.
C.W. Hedman. International Paper, Southlands Experiment Forest, Bainbridge, GA 31717, U.S.A.
1
Corresponding author (e-mail: rmitchel@jonesctr.org).
I:\cjfr\cjfr31\cjfr-05\X01-003.vp
Tuesday, April 17, 2001 3:15:45 PM
Color profile: Disabled
Composite Default screen
individual trees in savannas breaks the continuity of the
grassland-matrix environment (Belsky and Canham 1994),
and, similar to gaps, the resulting environmental heterogene
-
ity influences the structure and function of savannas.
Gaps in closed-canopy forests are caused by the death (in
-
dividually or in groups) of branches or trees. In gaps where
large canopy trees are killed, reduced competition for light
(Canham 1988; Canham et al. 1990; Chazdon and Fetcher
1984) and sometimes soil resources (Parsons et al. 1994;
Mladenoff 1987; Vitousek and Denslow 1986) increases
overall resource availability to the forest floor. While gaps
increase resources, these resources are not necessarily con
-
gruent within gaps, either spatially or temporally (Belsky
and Canham 1994; Van Pelt and Franklin 1999). The re
-
duced competition for resources by the overstory in a gap
increases seedling and sapling growth. However, the hetero
-
geneous resource response within gaps alters the ratio of re
-
sources, influencing species-specific responses (Denslow 1987;
Pacala et al. 1993).
Individual trees also alter resource levels and modify mi
-
croclimate in tropical savannas. Light is attenuated under
-
neath individual savanna trees but much less than that of a
closed-canopy forest, ranging from 55% of full sun under
-
neath Acacia tortilis (Forssk.) Hayne (Belsky et al. 1989,
1993) to 7% for Quercus douglasii H. & A. (Jackson et al.
1990). In addition, the duration and intensity of solar radia-
tion reaching the surrounding grassland is influenced over a
considerably larger area than that represented by the vertical
projection of the crowns of trees. On the equator in Kenya,
isolated trees of A. tortillis cast shade as far as 40 m from
the tree in the early morning and late afternoon, and the
daily integrated light values increase as distance from the
tree increases (Belsky et al. 1989).
In contrast to light, soil nutrients are often greater under-
neath individual trees in savannas. Increased soil fertility has
been attributed to a variety of influences. Nitrogen fertility of
soils may be increased by the presence of legumes (Bernhard-
Reversat 1988; Radwanski and Wickens 1967; Vitousek and
Walker 1989). Increased N mineralization has also been re
-
ported underneath trees where the litter dynamics differ from
that of the grassland matrix (Reich et al. 2001). Trees can
act as nutrient pumps, drawing nutrients from deep horizons
or laterally from areas well beyond the canopy and deposit
-
ing them underneath the canopy via litter fall (Belsky et al.
1989; Vetaas 1992; Weltzin and Coughenour 1990) or by
canopy throughfall (Kellman 1979). Dry deposition of nutri
-
ents (due to the aerodynamic roughness of tree canopies in
-
tercepting dust and nutrients) (Bernhard-Reversat 1982) and
less leaching of nutrients underneath a canopy (Belsky and
Canham 1994) have also been cited as additional factors re
-
sulting in fertility islands underneath isolated trees. More
-
over, nutrient additions by birds perching on limbs or animal
defecation from those that take cover in the shade of a can
-
opy have also been suggested as another factor yielding
greater soil fertility underneath savanna trees (Belsky 1994;
Georgiadis 1989). In addition to soil fertility, soil moisture is
often increased underneath individual savanna trees as com
-
pared with the grassland matrix (Joffre and Rambal 1988;
Kennard and Walker 1973; Parker and Muller 1982). The
impact of individual trees varies seasonally with rainfall and
temperature patterns. Also, the interactions between soil mois
-
ture and isolated trees can be influenced by interception,
evapotranspirational losses from trees, modification of the
microclimate, and alteration of evapotranspirational losses
from vegetation and soils influenced by the tree canopy
(Belsky et al. 1989; Vetaas 1992). Nevertheless, the net re
-
sult of isolated trees in tropical savannas is often to increase
soil resources, facilitate the productivity of tree patches, and
promote the recruitment of new woody plants relative to that
experienced in the open grassland. When woody plants in
-
crease in density to some threshold, however, productivity
within the tree patches often declines due to resource com
-
petition (Belsky and Amundson 1992; Obot 1988).
Longleaf pine (Pinus palustris Mill.) savannas are intermedi
-
ate in density to closed-canopy forests and tropical savannas.
The extent that adult trees compete for resources with regen
-
erating seedlings or facilitate regeneration is not well under
-
stood. Furthermore, the response of the grass-dominated
understory to reductions in overstory abundance has not
been reported. Regeneration of longleaf pine seedlings has
been shown to be negatively related to adult tree size and
positively related to distance from adults (Platt et al. 1988).
Whether the impact of adult trees on longleaf pine seedlings
is due to resource competition or influences on fire behavior
has not been well documented (Grace and Platt 1995). Palik
et al. (1997) reported that seedling growth was strongly in-
fluenced by the greater light in gaps and the increased N
availability found in the center of gaps. More recently,
Brockway and Outcalt (1998) found no differences in light
within gaps of longleaf pine woodlands. They speculated
that root gaps resulting from tree mortality would increase
soil moisture and positively affect growth and survival of
longleaf pine seedlings.
This study addresses the mechanisms regulating longleaf
pine seedling and understory response within and among
various-sized canopy gaps created by removal of overstory
trees. Gap sizes ranged from those previously reported to in
-
crease seedling response (Palik et al. 1997) to larger open
-
ings expected to exceed those needed to maximize seedling
response. The specific objectives of this study were to
(i) quantify plant resources (soil nitrogen, soil moisture, and
light availability) and (ii) determine the growth of planted
seedlings and understory vegetation as influenced by the cre
-
ation of gaps. We hypothesize that the understory vegetation
and planted longleaf pine seedlings will respond positively
to increased light availability within gaps. Furthermore, we
propose that, when the overstory is disturbed, increased under
-
story growth (grasses and herbs), particularly belowground,
will mediate any increased availability in soil resources as
-
sociated with overstory disturbance.
Methods
Study site
This study was conducted at the Silver Lake Tract of Interna
-
tional Paper’s Southland’s Experiment Forest (SEF), southwestern
Georgia, U.S.A. (30°48
′
N, 84°39
′
W). Elevations of SEF range from
about 23 to 95 m above mean sea level. The climate is character
-
ized as humid-subtropical. Average precipitation is 140 cm/year,
with 67% of the precipitation falling between March and October.
Summer droughts, however, are not unusual. Average daily temper
-
atures range from 21–34°C in summer to 5–17°C in winter (an av
-
erage of 260 frost-free days; C. Hedman, unpublished data).
© 2001 NRC Canada
766 Can. J. For. Res. Vol. 31, 2001
I:\cjfr\cjfr31\cjfr-05\X01-003.vp
Tuesday, April 17, 2001 3:15:46 PM
Color profile: Disabled
Composite Default screen
The Silver Lake Tract is dominated by 60- to 90-year-old, second-
growth, longleaf pine stands. The longleaf pine stands are found
predominantly on well-drained, sandy-loam soils with a gently roll
-
ing terrain (Buckner et al. 1979). These forests are characterized as
an open canopy of longleaf pine with a rich understory dominated
by C
4
grasses. Soils are classified in the Orangeburg and Norfolk
series. Orangeburg soils are characterized as well drained with a
sandy surface layer and 20–33 cm of sandy loam over a sandy
clay-loam subsoil. Norfolk soils are also well drained and moder
-
ately permeable with up to 35 cm of a sandy-loam surface layer
over a sandy clay-loam subsoil. Both soils are strongly acidic and
low in natural fertility and organic matter (USDA 1986).
Experimental design
In September 1996 canopy openings were created with a rubber-
tired feller-buncher. Felled trees’ boles were removed from the
gaps using a rubber-tired log skidder. Each of six 5-ha blocks con
-
tained three circular canopy openings and an uncut control plot.
The canopy openings ranged in size from 0.11 ha (18 m radius or
0.75 times tree height), 0.41 ha (36 m radius or 1.5 times tree
height), and 1.63 ha (72 m radius or 3 times tree height). The uncut
control plot was 1.63 ha. A 50-m buffer of trees was maintained
between all treatments.
Three of the six experimental blocks were used for measuring
resource availability, understory biomass, and belowground struc
-
ture. Seedlings, however, were planted in all six experimental blocks.
Transects were arranged along N, E, S, and W directions within
each overstory removal treatment and uncut control. Subplots 1 m
2
in area were established on all transects starting at the gap edge
and proceeding inward at 9, 18, 36, and 72 m (for the 1.63-ha gap
and control). The 0.41-ha gap had transects that extended 36 m
from the edge, while the 0.11-ha gap transects were 18 m. Each
gap center contained four 1-m
2
subplots where the four transects
(cardinal directions) intersected. In addition to subplots established
in the four cardinal directions, light measurements were also taken
on northwest, northeast, southwest and southeast bearings at simi-
lar distances from gap edge to better capture the directional varia-
tion of light within gaps. A low-intensity fire was prescribed for all
blocks in February 1996 (prior to gap creation) and again in the
winter of 1998. Understory vegetation recovery began a few weeks
after each fire.
Light resource measurements
The distribution of light within gap treatments and the uncut sa
-
vanna was quantified using hemispherical photographs (Mitchell
and Whitmore 1993). Photographs were recorded at all subplot lo
-
cations in eight directions (N, NE, E, SE, S, SW, W, and NW) on
400 ASA black and white film on uniformly cloudy days (Easter
and Spies 1994). Film negatives were converted to a digital format
and then analyzed using Hemiview
®
(Delta-T Devices Ltd., Burwell,
Cambridge, U.K.) to calculate a gap light index (Canham 1988).
Soil resource measurements
Nitrogen availability and gravimetric soil moisture was assessed
monthly from March 1997 to September 1998 using in situ buried
bag incubations of soil (Eno 1960). Soil cores were taken (2 cm di
-
ameter × 10 cm depth) from each subplot and composited across
each cardinal direction. A subsample of each composited sample
was taken to estimate initial pools of inorganic nitrogen and soil
moisture. Composited samples were put in gas-permeable plastic
bags, buried in their corresponding subplot locations, i.e., distance
from the edge, and incubated in the soil for 25–35 days. Initial and
incubated soil samples were extracted with 2 M KCl and subse
-
quently analyzed for nitrate and ammonium using a Lachat Auto
-
analyzer (Lachat Instruments, Inc., Milwaukee, Wis.). Net nitrogen
availability was determined by subtracting the initial pool from the
final pool of extractable inorganic nitrogen (Hart et al. 1994). Soil
subsamples were also dried to a constant mass at 100°C to deter
-
mine gravimetric soil moisture content (Gardner 1986). Values
from the initial and final pools of gravimetric soil moisture were
averaged by sample period to obtain net soil moisture.
Fine root standing crop
Root cores were used to measure fine root biomass and specific
root length. In August 1998, one root core (7.6 cm diameter ×
30 cm length) was taken adjacent to each subplot location in the
0.11-ha, 1.63-ha, and uncut control treatments. Since the 0.36-ha
treatment was assumed to fall somewhere in between the 0.11-ha
and 1.63-ha gap treatments, it was not sampled. Samples were
composited across cardinal directions according to distance from
the edge. Samples were transported and temporarily stored in cold
storage until processing. Roots were removed from the soil by
sieving through a 2.0-mm screen. All live roots were sorted into
pine and non-pine root groups (<3.0 mm diameter). Root biomass
was recorded after drying roots at 70°C for 48 h to a constant
mass. Root subsamples were placed in a muffle furnace at 500°C
for4htoconvertbiomassvalues to an ash-free basis.
A subsample of living pine (when present) and living non-pine
roots were collected from each treatment and control to determine
specific root length. To prevent desiccation, roots used in scanning
were kept cool (near 10°C) and moist. Root groups were analyzed
for root length using a Comair root length analyzer (Hawker de
Haviland Ltd., Melbourne, Australia). Prior to scanning, root sam
-
ples were finely cut and spread thoroughly across the analyzer’s
scanning surface to prevent overlap of individual roots (Richards et
al. 1979). Roots were then dried and root mass recorded. Specific
root lengths were expressed as ratios of length per root biomass.
These ratios were multiplied by the standing root crop biomass
value from each soil core to yield root length density by volume of
soil (cm/cm
3
).
Understory vegetation response
In October 1997 and September 1998, all aboveground biomass
of understory plants was destructively sampled along NW and SE
transects at intervals consistent with the seedling subplots. In Octo
-
ber 1997, only 0.11-ha, 1.63-ha treatments, and controls were sam
-
pled for understory biomass. Sampling in September 1998 included
all treatments and controls. Vegetation was dried to a constant
mass, weighed, and recorded.
Seedling planting and measurements
Nine 1-year-old container-grown longleaf pine seedlings were
planted by hand within every 1-m
2
subplot in February 1997. Only
seedlings with dark green foliage and root-collar diameters of at
least 8.0 mm were planted. The seed source for the containerized
seedlings was within 50 km of the experimental site (Ichauway
Plantation, Baker County, Ga.). Approximately 3 weeks after plant
-
© 2001 NRC Canada
McGuire et al. 767
Source* Effects df
G Fixed 3
R(G) Random 8
T
†
Fixed 11
G×T
†
Fixed 33
T × R(G)
†
Random 111
Note: The example presented is gravimetric soil
moisture.
*G, gap; R, replicate; T, time.
†
df varied with sampling frequency.
Table 1. Basic ANOVA table for the split-
plot, repeated measures design.
I:\cjfr\cjfr31\cjfr-05\X01-003.vp
Tuesday, April 17, 2001 3:15:46 PM
Color profile: Disabled
Composite Default screen
ing, seedling mortality was assessed. Dead seedlings were replaced
with living seedlings. Seeding survival was again assessed in Octo
-
ber 1997 (one growing season) and September 1998 (two growing
seasons).
In September 1998 above- and below-ground portions of seed
-
lings were harvested (Palik et al. 1997). Seedling root-collar diam
-
eters were recorded immediately following harvest. Root systems
were washed over a 2.0-mm sieve to remove adhering soil. Both
shoots and roots were dried at 70°C to a constant mass and
weighed. Ash-free dry masses of roots were recorded.
Data analysis
The data from the randomized complete block design were ana
-
lyzed using a split-plot mixed model (Littell et al. 1996) with
SAS
®
, version 8.00 (SAS Institute Inc. 2000) to account for re
-
peated measures within gap treatments (Table 1). The observed
intragap correlation as measured by the gap random effect was
small (r
≤
0.1) suggesting only a small degree of dependence
among sampling points within gaps. Further, alternative covariance
structures (exponential and spherical variogram models) for the
spatial dependence of sampling points were not needed based upon
likelihood ratio goodness-of-fit tests.
Data were tested for normality and examined for homogeneity
of variances using univariate analysis (Snedecor and Cochran 1989).
In cases where these assumptions were violated, the data were ad
-
justed using log transformations. Since overstory (pine) roots in all
gaps were intrinsically non-normal (pine roots were never observed
in plots 18 m from gap edge), our analysis of within-gap differ
-
ences focused on pine roots 0 and 9 m from the edge. These sub
-
plots, however, were compared against the pine root average of all
subplots in the uncut savanna.
Contrast comparisons were constructed to account for the fact
that not all gaps contain the same distances from the gap edge. For
some variables, location was nested within bearing (direction).
However, if no significant differences were found in bearing or the
sample size was too small, e.g., number of longleaf pine seedlings,
© 2001 NRC Canada
768 Can. J. For. Res. Vol. 31, 2001
Treatment
Net ammonification
(kg·ha
–1
·year
–1
)
Net nitrification
(kg·ha
–1
·year
–1
)
Net N mineralization
(kg·ha
–1
·year
–1
)
Light index
(%)
Uncut savanna 5.18 (0.71)b 2.60 (0.5)c 7.78 (1.18)b 48.2 (0.75)a
0.11-ha gap 2.26 (0.47)a 11.99 (1.89)a 14.25 (2.14)a 67.25 (1.13)b
0.41-ha gap 2.46 (0.58)a 8.15 (1.40)ab 10.62 (1.80)ab 79.13 (1.07)c
1.63-ha gap 2.43 (0.63)a 5.60 (1.01)bc 8.04 (1.55)b 84.18 (1.14)d
Note: Values are mean with SE given in parentheses. Values with different letters are significantly different (p < 0.01).
Table 2. Responses of soil and light resources to overstory gap formation.
Fig. 1. Light availability (gap light index (GLI)) in the uncut savanna (a) and 0.11-, 0.41-, and 1.63-ha gap treatments (b, c, and d,re
-
spectively). Gradations of grayscale and contours indicate changes in light availability. Note that the scales for treatments vary.
I:\cjfr\cjfr31\cjfr-05\X01-003.vp
Tuesday, April 17, 2001 3:16:12 PM
Color profile: Disabled
Composite Default screen
then variables were pooled among locations. Because our model
contained nested error terms, denominator degrees of freedom
were estimated using the Kenward and Roger (1997) approxima
-
tion. Adjusted confidence intervals were calculated based upon
these corrected degrees of freedom.
An asymptotic monomolecular growth curve (Seber and Wild
1989) was fit using non-linear regression to model seedling growth
response within gaps as a function of the distance from the edge.
This model was reparameterized so that we could determine at
what distance seedling growth was 75% of maximum growth
within gap openings. Linear regression models were also run to
test for relationships between (i) resource effects and seedling
growth response and (ii) resource effects and understory plant re
-
sponse. Residuals were examined, and nonlinear regression was
used where appropriate.
Results
Light availability
All gaps received more light than the uncut savanna, and
light increased with the size of gap opening (Table 2). In ad
-
© 2001 NRC Canada
McGuire et al. 769
Fig. 2. Gravimetric soil moisture among gap treatments and the uncut savanna for two growing seasons (1997–1998). Significant dif
-
ferences from the uncut savanna are shown with asterisks: *, p < 0.10; **, p < 0.005; ***, p < 0.001.
Fig. 3. Extractable (2 M KCl) N (NO
3
–
-N+NH
4
+
-N) (±1 SE) in soils of gap treatments and uncut savanna. Significant differences
from the uncut savanna are shown with asterisks: *, p < 0.10; **, p < 0.005.
I:\cjfr\cjfr31\cjfr-05\X01-003.vp
Tuesday, April 17, 2001 3:16:13 PM
Color profile: Disabled
Composite Default screen
dition, relationships between location and compass bearing
(direction) and light availability were observed within gaps
but not within the uncut savanna matrix (Fig. 1). Light avail
-
ability in 0.11-ha gaps quickly increased from the gap edge
to 9 m out from the edge (F
1,311
= 41.47, p<0.0001). How
-
ever, light availability did not vary from9mtothegapcen
-
ter. Similar increases in light availability from the gap edge
were noted within 0.41-ha gaps (F
1,311
= 119.16, p < 0.0001);
however, light availability increased from the edge to 18 m
then became asymptotic. Light increased from the gap edge
within 1.63-ha gaps (F
1,311
= 262.94, p < 0.0001) but was
not maximized until after 36 m from the gap edge. Southern
(t
43
= 3.74, p = 0.0143), southwestern (t
43
= 3.16, p =
0.0640), and southeastern (t
43
= 3.58, p = 0.0222) sections of
0.11-ha gaps received less light than the gap center. Addi
-
tionally, more light was available in the northern portion
than the southern (t
43
= –2.83, p = 0.0431) and southeastern
(t
43
= –2.67, p = 0.0633) portions of 0.11-ha gaps. Light was
similarly affected in other gap sizes, but the proportion of
gap area that received full sun increased with gap size.
Soil resources
In some sampling periods, soil moisture differed signifi
-
cantly among treatments; however, no pattern in timing (wet,
dry, or intermediate moisture periods) or treatment response
was evident through time (Fig. 2). Also, soil moisture did
not vary from the gap edge to center in any gap (data not
shown). It is worth noting (i) severe drought periods during
the summers of 1997 and 1998 and (ii) broad fluctuations in
soil moisture that were observed during the 2 years of the
study (Fig. 2). Thus, no patterns in competition for soil
moisture were consistently observed in wet, dry, or interme
-
diate soil moisture conditions.
Mean concentrations of extractable inorganic N (NO
3
–
-N +
NH
4
+
-N) were higher under canopy gaps (1.52 ± 0.25 kg/ha
(mean ± SE) in 0.11-ha gaps, 1.73 ± 0.31 kg/ha in 0.41-ha
gaps, and 1.53 ± 0.28 kg/ha in 1.63-ha gaps) than under the
uncut savanna (1.16 ± 0.15 kg/ha). These differences, how-
ever, were due predominantly to higher N mineralization
rates observed in gaps immediately following a drought dur-
ing the second growing season (Fig. 3). Throughout normal
rainfall events, N concentrations did not differ in gaps in any
predictable manner from that of the uncut savanna. Inorganic
N also did not vary in distance from the edge of gaps (data
not shown).
Mineralization of N, however, did vary with treatment
(Table 2). Compared with that under the uncut savanna, soil
net NH
4
+
production (ammonification) was lower in gaps
(0.11-ha gaps: F
1,138
= 12.65, p = 0.0005; 0.41-ha gaps:
F
1,138
= 10.95, p = 0.0012; and 1.63-ha gaps: F
1,138
= 11.20,
p = 0.0011) but did not vary with gap size. Soil net NO
3
–
production (nitrification) was highest in the 0.11-ha treat
-
ment and lowest in the 1.63-ha treatment. Both the 0.11-ha
gaps (F
1,138
= 25.94, p < 0.0001) and 0.41-ha gaps (F
1,138
=
9.06, p = 0.0031) had higher nitrification levels than that ob
-
served in the uncut savanna. Similar to nitrification, net ni
-
trogen mineralization also increased with gaps but was
inversely related to gap size (Table 2).
Nitrogen mineralization rates also varied within gap treat
-
ments. Net ammonification decreased from the gap edge to
center within 0.11-ha gaps (F
1,586
= 23.66, p < 0.0001) and
0.41-ha gaps (F
1,586
= 7.91, p = 0.0051) but did not differ
spatially within the 1.63-ha gaps (F
1,586
= 1.01, p = 0.3165)
(Fig. 4a). Net nitrification increased from the gap edge to
center within the 0.11-ha gaps (F
1,586
= 19.07, p < 0.0001),
0.41-ha gaps (F
1,586
= 5.01, p = 0.0256), and 1.63-ha gaps
(F
1,586
= 4.09, p = 0.0435) (Fig. 4b). Spatial differences in
net N mineralization were not detected from the edge to cen
-
ter within any gaps (Fig. 4c).
Fine root standing crop
Biomass of understory (non-pine) roots was less in gaps
than in the uncut savanna (F
1,411
= 5.45, p = 0.0201) approx
-
© 2001 NRC Canada
770 Can. J. For. Res. Vol. 31, 2001
Fig. 4. Annual net ammonification (a), nitrification (b), and net
N mineralization (c) within canopy gap treatments and the uncut
savanna (±1 SE). Significant differences between gap edge and
center are shown with asterisks: *, p < 0.05; **, p < 0.005.
I:\cjfr\cjfr31\cjfr-05\X01-003.vp
Tuesday, April 17, 2001 3:16:14 PM
Color profile: Disabled
Composite Default screen
imately 1 year (July 1997) after overstory removal; however,
the following season there was no difference in non-pine
root biomass (Fig. 5). Overstory tree (pine) roots were
greater under the uncut savanna than in gaps at every sample
period (F
8,411
= 60.08, p < 0.0001) as was total root biomass
(July 1997: F
1,411
= 35.22, p < 0.0001; Feb. 1998: F
1,411
=
13.15, p=0.0003; Aug. 1998: F
1,411
= 8.52, p = 0.0037).
Between gaps, 0.11-ha gaps consistently had more pine roots
than the 1.63-ha gaps (July 1997: F
1,411
= 17.09, p < 0.0001;
Feb. 1998: F
1,411
= 6.52, p = 0.0110; Aug. 1998: F
1,411
=
9.60, p = 0.0021). However, pine root biomass decreased af
-
ter the first growing season largely in the 0.11-ha gap treat
-
ment (F
1,411
= 4.04, p = 0.0450). Pine roots did not vary
between growing seasons under the uncut savanna or 1.63-
ha gaps.
Understory (non-pine) root length did not differ between
gap edges or centers or among the gaps and the uncut sa
-
vanna 2 years after overstory removal (Fig. 6). For the 0.11-
ha gaps, however, overstory (pine) root length decreased rap
-
idly from the gap edge to9m(F
1,75.7
= 8.0, p=0.0060)
(Fig. 6a). Pine roots were only found along the gap edge in
the 1.63-ha gap (Fig. 6b). Despite the decrease in pine root
length, total root length did not differ from gap edges to cen
-
ters within gaps (Fig. 6).
Aboveground understory standing crop
Understory biomass increased in response to gaps (Fig. 7).
This response, however, varied between the first and second
growing season. Following the first growing season, 0.11-ha
gaps (F
1,14.7
= 5.09, p = 0.0398) and 1.63-ha gaps (F
1,9.73
=
7.01, p = 0.0249) had more understory biomass than that of
the uncut savanna (0.41-ha gap not measured) but did not
differ from one another (Fig. 7). Following the second grow
-
ing season, understory biomass increased systematically with
gap size. Compared with the uncut savanna, understory bio-
mass was 35% greater in 0.41-ha gaps (F
1,11.4
= 3.81, p =
0.0760) and 50% greater in 1.63-ha gaps (F
1,9.73
= 8.3, p =
0.0168). Aboveground understory biomass in 0.11-ha gaps
did not differ from the uncut savanna (Fig. 7). Spatial trends
of understory biomass within gaps were only evident after
the second growing season, where biomass increased from
gap edge to center in 0.41-ha gaps (F
1,65.2
= 4.91, p = 0.0301)
and in 1.63-ha gaps (F
1,65.2
= 3.18, p = 0.0792) (Fig. 8).
Seedling response
Seedling survival was lower at the end of the first growing
season (October 1997) in the gaps relative to the uncut sa
-
vanna (0.11-ha gaps: t
1568
= –5.79, p < 0.0001; 0.41-ha gaps:
t
1568
= –4.39, p = 0.0007; and 1.63-ha gaps: t
1568
= –17.48,
p < 0.0001). By the end of the second growing season, how
-
ever, survival in the uncut savanna did not differ from that in
any gap and was approximately 10% in both gaps and under
the uncut savanna (Fig. 9). Spatial trends in seedling survival
were inconsistent across time and appeared not to be medi
-
ated by location from the gap edge (data not shown).
Average root-collar diameter (RCD) was larger within gap
openings than within the uncut savanna (F
1,8.16
= 31.03, p=
0.0005). However, RCD did not differ with gap size (Ta
-
ble 3). RCD averaged around 12 mm for gap treatments and
9 mm for the uncut control. Compared with seedlings in the
uncut savanna, longleaf pine shoots were larger in all gap
sizes (F
1,6.24
= 38.13, p < 0.0007). Average seedling shoot
size did not differ among gaps (Table 3). Seedling root bio
-
mass was larger in gaps than that of the uncut savanna
(F
1,6.32
= 27.77, p < 0.0016) but did not differ among gaps
(Table 3).
© 2001 NRC Canada
McGuire et al. 771
Fig. 5. Overstory and understory plant fine root biomass (±1 SE) for two growing seasons following gap formation. Groupings of sta
-
tistical tests are as follows: overstory (pine) root (lowercase letters), p < 0.001; understory (non-pine) root (lowercase letters enclosed
in boxes), p < 0.05; and total root biomass (uppercase letters above error bars), p < 0.05.
I:\cjfr\cjfr31\cjfr-05\X01-003.vp
Tuesday, April 17, 2001 3:16:14 PM
Color profile: Disabled
Composite Default screen
Since seedling response was similar among the three gap
sizes, data were averaged across treatments and compared
against the uncut savanna (Fig. 10). Seedling RCD increased
linearly with distance from the gap edge until 18 m, when
RCD became asymptotic (F
1,15.1
= 42.62, p < 0.0001)
(Fig. 10a). Seedling shoot biomass also increased linearly
with distance from the gap edge and became asymptotic
18 m from the gap edge (F
1,12.1
= 45.32, p < 0.0001)
(Fig. 10b). Seedling root biomass increased linearly from the
gap edge to the threshold distance between 18 and 36 m
(F
1,8.87
= 48.27, p < 0.0001). After 36 m, seedling root bio
-
mass became asymptotic (F
1,145
= 0.27, p = 0.6045)
(Fig. 10c). The estimated distances from the gap edge in
which seedlings reached 75% of maximum growth observed
in this study were 9.6 m (RCD), 13.8 m (shoot biomass),
and 23.2 m (root biomass).
Resource availability and plant response to
experimental gaps
Light availability explained the most variation in both un
-
derstory biomass and pine seedling biomass. Understory bio
-
mass was correlated with light availability in the first (r
2
=
0.30, p = 0.0003) (Fig. 11a) and second (r
2
= 0.33, p <
0.0001) (Fig. 11b) growing seasons. Total seedling size was
also significantly and positively correlated to light availabil
-
ity (r
2
= 0.44, p < 0.0001) (Fig. 12a). Net N mineralization
explained less variation in understory (r
2
= 0.10, p = 0.0515)
(Fig. 11c) and seedling (r
2
= 0.08, p = 0.0569) (Fig. 12b)
biomass after the second growing season. Nitrogen mineral
-
ization was not related to herbaceous growth in the second
growing season. Neither seedling size nor understory bio
-
mass in the first or second growing season were related to
soil moisture content.
Discussion
Experimental gaps cut in a longleaf pine savanna strongly
influenced the resource environment and vegetation re-
sponse. Light reaching the understory of the uncut longleaf
pine matrix averaged more than 45% of full sunlight. The
amount of light is high particularly in relation to undis-
turbed, closed-canopy forests, where light values are often
less than 1% of full sun (Minckler et al. 1973; March and
Skeen 1976; Canham et al. 1990, 1994; Dirzo et al. 1992;
Easter and Spies 1994). Despite the high light availability in
uncut longleaf pine savannas, understory light increased as a
function of distance from the edge when gaps were cut.
Light was also influenced along a north to south bearing,
with less light in the southern aspect of gap openings. In the
largest gap treatment, the mean light availability increased
rapidly from the gap edge and was maximized between 18
and 36 m from the gap edge. The difference in light between
north and south aspects of gaps was attenuated as gap size
increased.
Unlike closed-canopy forests, open-canopy systems such
as longleaf pine savannas do not have distinct boundaries be
-
tween gaps and canopy-dominated areas (Brockway and
Outcalt 1998); thus, light environments are highly variable
temporally and spatially. Because of this inherent variability,
the approaches used to measure light may explain the dia
-
metrically opposed light data reported for disturbed and in
-
tact longleaf pine savannas. Palik et al. (1997) found a
curvilinear increase in light availability as overstory density
decreased within natural canopy disturbances of longleaf pine
forests (average gap size 0.15 ha). In contrast, Brockway and
Outcalt (1998) reported that “solar radiation was uniformly
distributed across [longleaf pine] canopy gaps” because of
the ability of light to reach the forest floor through “numer
-
ous interstitial spaces in the sparse pine overstory” (gap
sizes 0.1–0.2 ha). The numerous spaces allow light to reach
the understory but in far from uniform ways. Open-canopy
longleaf pine savannas can be thought of as a matrix of sun
-
lit understory interrupted by shade from the overstory can
-
opy that moves across the forest floor as sun angle changes
© 2001 NRC Canada
772 Can. J. For. Res. Vol. 31, 2001
Fig. 6. Fine root length (±1 SE) for overstory trees and understory
plants within gap treatments for August 1998 (2 years after har
-
vest): (a) 0.11-ha gap, (b) 1.63-ha gap, and (c) uncut savanna.
Groupings of statistical tests are as follows: overstory (pine) root
(shaded bars and lowercase letters), p < 0.001; understory (non-
pine) root (open bars and lowercase letters), p < 0.05; and total
root biomass (capital letters above error bars), p < 0.05.
I:\cjfr\cjfr31\cjfr-05\X01-003.vp
Tuesday, April 17, 2001 3:16:15 PM
Color profile: Disabled
Composite Default screen
daily and seasonally. Accounting for this temporal and spa-
tial variability is critical to establishing any relationship among
overstory structure, light availability, and seedling response.
Hemispherical photos can account for variation by simulat-
ing the sun angle throughout a day and season (Anderson
1964; Chazdon and Field 1987; Canham 1989; Whitmore et
al. 1993; Easter and Spies 1994). Some studies have charac-
terized light environments by measuring light 2 h before and
after solar noon under clear sky conditions (Morgan et al.
1985; Messier et al. 1989; Brockway and Outcalt 1998).
However, this approach can give highly questionable results
because of high variability (Gay et al. 1971; Rich et al.
1993) caused by the interaction of solar angle, gap size, and
location in the overstory (Comeau et al. 1998).
In the study reported here, not only did the cutting of gaps
increase light, but understory vegetation growth was posi
-
tively correlated with light reaching the understory. Warm-
season C
4
grasses, particularly wiregrass (Aristida stricta
Michx.), dominated the understory vegetation at our study
sites. Mitchell et al. (1999) found that broomsedge (Andro
-
pogon virginicus L.), a warm-season C
4
grass, showed growth
increases up to nearly full sunlight, similar to responses of
the wiregrass-dominated understory reported here. In this
study, however, harvesting trees disturbed the understory such
that the relationship between light availability and under
-
story biomass was three times greater in the second year
(r
2
= 0.33, p > 0.0001) than the first year after harvest (r
2
=
0.10, p > 0.005). The understory vegetation in longleaf pine
woodlands has evolved with fire and is resilient to natural,
aboveground disturbances. The extent that managed distur
-
bances, such as timber harvesting, are similar to or differ
from more natural overstory disturbances, e.g., lightning and
wind, with respect to understory response is not known and
should be investigated.
Longleaf pine showed positive growth responses to the in
-
creased light reaching the understory of gaps and, to a much
smaller extent, increased soil N levels that were found in the
© 2001 NRC Canada
McGuire et al. 773
Fig. 7. Understory plant biomass (±1 SE) for the first two growing seasons following overstory harvest. Bars with different letters are
significantly different (p < 0.05).
Fig. 8. Spatial responses of the understory plant community (±1
SE) after the first two growing seasons following gap formation:
(a) November 1997; (b) September 1998. Significant differences
between gap edge and center are shown with asterisks: *, p =
0.0792, **, p = 0.0301. The 0.41-ha gap treatment was not sam-
pled in November 1997.
I:\cjfr\cjfr31\cjfr-05\X01-003.vp
Tuesday, April 17, 2001 3:16:16 PM
Color profile: Disabled
Composite Default screen
experimental gaps. Longleaf pine response to both light and
N was reported by Palik et al. (1997), while Brockway and
Outcalt (1998) reported no relationship between light and
longleaf pine seedling growth. Multiple limitations (particu
-
larly light and N) to pine growth in field competition studies
have previously been reported (Mitchell et al. 1999). Long
-
leaf pine is classified as shade intolerant. These data and
those of Palik et al. (1997) suggest that the growth response
of longleaf pine seedlings is not maximized until nearly full
sunlight, certainly a characteristic of a shade-intolerant spe
-
cies. However, gaps as small as 0.1 ha were large enough
for seedling growth to reach a maximum. Furthermore, 20%
of the 0.11-ha gaps contained seedlings growing at least
75% of the maximum observed in this study.
The increase in herbaceous understory biomass with in
-
creasing light within gaps may have attenuated N increases
in gaps. In closed-canopy forests, increases in N tend to be
observed only in larger gaps (Parsons et al. 1994). Palik et
al. (1997) found that, when the understory community was
removed (application of foliar herbicide) in a longleaf pine
savanna, N increased in areas of overstory disturbance but
only in the center of large gaps (0.1 ha). This may be due to
overlapping root domains of the remaining overstory trees
that quickly colonize soil after small-scale overstory distur
-
bance (Parson et al. 1994). The remaining live trees are then
able to forage for any N liberated by the reduced demand
associated with small-scale tree mortality. In the study re-
ported here, increased herbaceous understory growth may
rapidly fill root gaps created by overstory mortality, thus re-
ducing the overall impacts of gaps on soil resources.
Facilitation of survival, i.e., greater survival in the uncut
savanna than in the gaps, was observed the first year after
gap creation but not the second. Both years, however, experi-
enced significant drought. Only a few studies suggest that
facilitation of longleaf pine seedling survival in the shade of
overstory may be a factor in regeneration (Allen 1954,
1955), while many references can be found reporting long
-
leaf pine’s sensitivity to competition (with respect to growth)
during the grass stage (Boyer 1963; Bruce 1958; Palik et al.
1997; Brockway and Outcalt 1998). Often different mea
-
sures of fitness, e.g., growth and survival, will present dia
-
metrically opposed views of plant–plant interactions
(Goldberg 1990). De Steven (1991a, 1991b) reported that
old-field loblolly pine (Pinus taeda L.) establishment was fa
-
cilitated in the presence of old-field herbs, but pine growth
was diminished through competition with old-field, herba
-
ceous plant communities.
Overall, survival was strongly impacted in this study by
drought in both years. Even though drought influenced sur
-
vival, no patterns in soil moisture or seedling survival were
found withingaps. This was true even though pine growth re
-
sponse was strongly influenced by gaps. Brockway and
Outcalt (1998) speculate that reduced competition for soil
moisture in gaps should increase growth and survival of
longleaf pine seedlings; however, our data is contrary to
those suggestions. They report a 12- to 16-m seedling exclu
-
sion zone from adult longleaf pine resulting from competi
-
tion for soil moisture with overstory trees. We can find no
evidence for such interactions. In fact, we frequently find
naturally regenerated longleaf pine seedlings within 3–4 m
of adult pine (R.J. Mitchell, personal observation). However,
it should be noted that the Brockway and Outcalt (1998)
study assessed establishment patterns of naturally regener
-
© 2001 NRC Canada
774 Can. J. For. Res. Vol. 31, 2001
Seedling Survival (%)
0
10
20
30
40
50
60
70
80
90
100
0.11-ha gap
0.41-ha gap
1.63-ha gap
Uncut Savanna
AB
A
A
B
C
C
C
D
E
E
E
E
March 1997
October 1997
September 1998
Fig. 9. Longleaf pine seedling survival (±1 SE) under canopy gaps and the uncut savanna in the first 2 years following overstory har
-
vest. Bars with the same letters are not significantly different among sample dates (p < 0.0001) and across time (p < 0.001).
RCD (mm)
Shoot
biomass (g)
Root
biomass (g)
Uncut savanna 9.10 (0.17)b 1.74 (0.17)b 1.52 (0.06)b
0.11-ha gap 12.04 (0.40)a 6.49 (0.90)a 3.30 (0.34)a
0.41-ha gap 12.45 (0.34)a 8.26 (0.68)a 4.16 (0.29)a
1.63-ha gap 11.68 (0.24)a 6.93 (0.52)a 3.37 (0.18)a
Note: Values are mean with SE given in parentheses. Values with
different letters are significantly different (p < 0.008).
Table 3. Longleaf pine seedling root-collar diameter (RCD),
shoot biomass, and root biomass averaged by experimental
treatment.
I:\cjfr\cjfr31\cjfr-05\X01-003.vp
Friday, April 20, 2001 3:05:10 PM
Color profile: Disabled
Composite Default screen
ated longleaf pine seedlings that differed in origin (estab
-
lished from seed) and age from the containerized seedlings
in the study reported here. Survival from seed is influenced
by elements that were not as important to the containerized
seedlings planted in this study, i.e., susceptibility to fire. Re
-
search into differences of seedling establishment from seed
within gaps could help elucidate the differences noted be
-
tween the two studies.
Fire and competition for light may result in aggregating
naturally regenerated seedlings toward the center of gaps
(Platt et al. 1988). Grace and Platt (1995) reported that seed
-
lings in areas of high density of mature longleaf pine were
smaller than those found in areas of low density and that
seedling survival after fire was strongly reduced in areas of
high pine basal area. Smaller seedlings are more susceptible
to fire (Boyer 1974; Grace and Platt 1995), and fires with
higher temperatures are more lethal to seedlings. The greater
needlefall near mature pines has been associated with more
intense fires (Williamson and Black 1981; Rebertus et al.
1989; Grace and Platt 1995). Higher-intensity fires com
-
bined with smaller seedlings due to resource competition
spatially segregates seedlings from mature longleaf pine. Since
wiregrass foliage is an important fuel in longleaf pine –
wiregrass savannas, the increase in standing crop of a
wiregrass-dominated understory in gaps as reported here
may also influence fire behavior and seedling dynamics.
This possibility has not been studied.
© 2001 NRC Canada
McGuire et al. 775
Fig. 10. Longleaf pine seedling root-collar diameter (a) and bio
-
mass (b and c) within canopy openings and the uncut savanna
(±1 SE) after two growing seasons. Size and biomass values in
cut treatments were pooled to yield an average within gap; con
-
trol is presented with a 95% confidence boundary. Significant
differences between gap edge and 18 m are shown with aster
-
isks: *, p < 0.0001.
Fig. 11. Understory plant biomass response to light availability and
nitrogen mineralization for the first and second growing seasons fol
-
lowing harvest. Equation for light for the first growing season (a)is
y = –48.1 + 614.3(x/(9.04 + x)) (r
2
= 0.29, F = 18.13, p < 0.0001)
and the second growing season (b)isy = 24.42x + 300.9 (r
2
=
0.33, p < 0.0001). Equation for understory biomass response to soil
N(c)isy = 59.08x + 1721 (r
2
= 0.10 p < 0.005 15).
I:\cjfr\cjfr31\cjfr-05\X01-003.vp
Tuesday, April 17, 2001 3:16:18 PM
Color profile: Disabled
Composite Default screen
Conculsion and management implications
We found that light reaching the understory was increased
in gaps cut from longleaf pine savannas. Longleaf pine seed
-
ling growth and understory standing crop response was sig
-
nificantly and positively correlated to light availability. Soil
resources were not strongly influenced by gaps, perhaps be
-
cause of herbaceous plants rapidly colonizing any root gaps
caused by overstory disturbance.
Gaps as small as 0.11 ha were large enough to elicit maxi
-
mum growth of longleaf pine seedlings. In fact, 75% of
maximum seedling growth could be found approximately
10 m into gap openings. Thus, even though longleaf pine has
some characteristics of a shade-intolerant tree, the high light
levels in the understory of uncut savannas and the increase
in light from the cutting of small gaps allows for sufficient
resources to sustain a new cohort of seedlings and encourage
an uneven-aged stand structure. If properly applied, uneven-
aged management can more closely emulate natural distur
-
bance patterns in longleaf pine savannas (Palik et al. 2001)
and may result in a greater ability to maintain important
components of the biodiversity of this system. For example,
red-cockaded woodpecker (Picoides borealis Vieillot) habi
-
tat is dependent upon older trees remaining in southern pine
stands (Engstrom et al. 1996). Some argue that the intolerant
nature of southern pines biologically restricts the use of some
uneven-aged management approaches, particularly single-tree
selection (Boyer 1993; Rudolph and Conner 1996; Brockway
and Outcalt 1998). Data presented here, however, clearly
show that frequently burned, uncut, longleaf pine savannas
exhibit high levels of light reaching the understory, and need
only small gaps to promote regeneration. These small gaps
can frequently be achieved by removal of one to few adult
trees, encouraging longleaf pine regeneration. Thus, uneven-
aged management, including single-tree selection, may be ide
-
ally suited to longleaf pine stands with objectives that in
-
clude both timber and conservation goals (Engstrom et al.
1996).
Acknowledgments
Funding for this study was provided by the Robert W.
Woodruff Foundation, International Paper, and Auburn Uni
-
versity’s School of Forestry. We thank numerous employees
of the Joseph W. Jones Ecological Research Center who
aided in this project. In particular, L. Forrester, S. Glickauf,
B.J. Harris, T. Hay, S. Hurst, S. McGee, D. O’Connor, and
A.M. Velez were vital for their field and laboratory assis
-
tance. Thank you to D. Coates, K. Kirkman, and three anon-
ymous reviewers for providing useful critiques.
References
Allen, R.M. 1954. Shade may improve longleaf survival. USDA
For. Serv. Res. Note SO-90.
Allen, R.M. 1955. Growth of longleaf pine on cutover lands and
old fields. J. For. 53: 587.
Anderson, M.C. 1964. Studies of the woodland light climate. II.
Seasonal variation in the light climate. J. Ecol. 52: 643–663.
Belsky, A.J. 1994. Influences of trees on savanna productivity: tests
of shade, nutrients, and tree–grass competition. Ecology, 75:
922–932.
Belsky, A.J., and Amundson, R.G. 1992. Effects of trees on under
-
story vegetation and soils at forest/savanna boundaries in East
Africa. In Nature and dynamics of forest–savanna boundaries.
Chapman & Hall, New York. pp. 353–366.
Belsky, A.J., and Canham, C.D. 1994. Forest gaps and isolated sa
-
vanna trees: an application of patch dynamics in two ecosys
-
tems. Bioscience, 44: 77–83.
Belsky, A.J., Amundson, R.G., Duxbury, J.M., Riha, S.J., Ali, A.R.,
and Mwonga, S.M. 1989. The effects of trees on their physical,
chemical, and biological environments in a semi-arid savanna in
Kenya. J. Appl. Ecol. 26: 1005–1024.
Belsky, A.J., Mwonga, S.M., Amundson, R.G., Ali, A.R., and
Duxbury, J.M. 1993. Relative effects of trees on their understory
environments in high-rainfall versus low-rainfall amounts. J.
Appl. Ecol. 30: 143–155.
Bernhard-Reversat, F. 1982. Biogeochemical cycle of nitrogen in a
semi-arid savanna. Oikos, 3: 321–332.
Bernhard-Reversat, F. 1988. Soil nitrogen mineralization under a
Eucalyptus plantation and a natural Acacia forest in Senegal.
For. Ecol. Manage. 23: 233–244.
Boyer, W.D. 1963. Development of longleaf pine seedlings under
parent trees. USDA For. Serv. Res. Pap. SO-4.
Boyer, W.D. 1974. Impact of prescribed fires on mortality of re
-
leased and unreleased longleaf pine seedlings. USDA For. Serv.
Res. Note SO-128.
© 2001 NRC Canada
776 Can. J. For. Res. Vol. 31, 2001
Fig. 12. Longleaf pine seedling size as a function of light avail
-
ability (a) and cumulative net nitrogen mineralization (b). Equa
-
tions for light availability and nitrogen mineralization are y =
0.02x – 6.71 (r
2
= 0.44, p < 0.0001) and y = 0.387x +4.52(r
2
=
0.08, p = 0.0569), respectively.
I:\cjfr\cjfr31\cjfr-05\X01-003.vp
Tuesday, April 17, 2001 3:16:19 PM
Color profile: Disabled
Composite Default screen
© 2001 NRC Canada
McGuire et al. 777
Boyer, W.D. 1993. Long-term development of regeneration under
longleaf pine seedtree and shelterwood stands. South. J. Appl.
For. 17: 10–15.
Brockway, D.G., and Outcalt, K.W. 1998. Gap-phase regeneration
in longleaf pine – wiregrass ecosystems. For. Ecol. Manage.
106: 125–139.
Bruce, D. 1958. Effect of low competition on longleaf pine seed
-
ling growth. 57th Proc. Soc. Am. For. pp. 151–153.
Buckner, J.L., Landers, J.L., Barker, J.A., and Perkins, C.J. 1979.
Wildlife food plants following preparation of longleaf pine sites
in southwest Georgia. South. J. Appl. For. 3: 56–59.
Canham, C.D. 1988. Growth and canopy architecture of shade-
tolerant trees: response to canopy gaps. Ecology, 6: 786–795.
Canham, C.D. 1989. Different responses to gaps among shade-
tolerant tree species. Ecology, 70: 548–550.
Canham, C.D., Denslow, J.S., Platt, W.J., Runkle, J.R., Spies, T.A.,
and White, P.S. 1990. Light regimes beneath closed canopies
and tree-fall gaps in temperate and tropical forests. Can. J. For.
Res. 20: 620–631.
Canham, C.D., Finzi, A.C., Pacala, S.W., and Burbank, D.H. 1994.
Causes and consequences of resource heterogeneity in forests:
interspecific variation in light transmission by canopy trees. Can.
J. For. Res. 24: 337–349.
Chazdon, R.L., and Fetcher, N. 1984. Photosynthetic light environ
-
ments in a lowland tropical rainforest in Costa Rica. J. Ecol. 72:
553–564.
Chazdon, R.L., and Field, C.B. 1987. Photographic estimation of
photosynthetically active radiation: evaluation of a computerized
technique. Oecologia, 73: 525–532.
Comeau, P.G., Gendron, F., and Letchford, T. 1998. A comparison
of several methods for estimating light under a paper birch mixed
wood stand. Can. J. For. Res. 28: 1843–1850.
Denslow, J.S. 1987. Tropical rainforest gaps and tree species diver-
sity. Annu. Rev. Ecol. Syst. 18: 431–451.
De Steven, D. 1991a. Experiments on mechanisms of tree estab-
lishment in old-field succession: seedling emergence. Ecology,
72: 1066–1075.
De Steven, D. 1991b. Experiments on mechanisms of tree estab
-
lishment in old-field succession: seedling survival and growth.
Ecology, 72: 1076–1088.
Dirzo, R., Horvitz, C.C., Quevedo, H., and Lopez, M.A. 1992. The
effects of gap size and age on the understory herb community of
a tropical rain forest. J. Ecol. 80: 809–822.
Easter, M.J., and Spies, T.A. 1994. Using hemispherical photogra
-
phy for estimating photosynthetic photon flux density under can
-
opies and in gaps in Douglas-fir forests of the Pacific Northwest.
Can. J. For. Res. 24: 2050–2058.
Endler, J.A. 1993. The color of light in forests and its implications.
Ecol. Monogr. 63: 1–27.
Eno, C. 1960. Nitrate production in the field by incubating the soil
in polyethylene bags. Soil Sci. Soc. Am. J. 24: 277–279.
Engstrom, R.T., Brennan, L.A., Neel, W.L., Farrar, R.M., Linde
-
man, S.T., Moser, W.K., and Hermann, S.M. 1996. Silvicultural
practices and red-cockaded woodpecker management: a reply to
Rudolph and Conner. Wildl. Soc. Bull. 24: 334–338.
Franklin, J.F. 1985. Tree death as an ecological process: the causes,
consequences, and variability of tree mortality. BioScience, 35:
550–556.
Franklin, J.F., Shugart, H.H, and Harmon, M.E. 1987. Tree death
as an ecological process. BioScience, 37: 550–556.
Gardner, W.H. 1986. Methods of soil analysis. Part 1. Physical and
mineralogical methods. American Society of Agronomy, Inc,
Madison, Wis. pp. 493–544.
Gay, L.W., Knoerr, K.R., and Braaten, M.O. 1971. Solar radiation
variability on the floor of a pine plantation. Agric. Meteorol. 8:
39–50.
Georgiadis, N.J. 1989. Microhabitat variation in an African sa
-
vanna: effect of woody cover and herbivores in Kenya. J. Trop.
Ecol. 5: 93–108.
Goldberg, D.E. 1990. Components of resource competition in plant
communities In Perspectives in competition. Edited by J.B. Grace
and D. Tilman. Academic Press, New York. pp. 27–50.
Grace, S.L., and Platt, W.J. 1995. Effects of adult tree density and
fire on the demography of pregrass stage juvenile longleaf pine
(Pinus palustris Mill.). J. Ecol. 83: 75–86.
Hart, S.C., Stark, J.M., Davidson, E.A., and Firestone, M.K. 1994.
Nitrogen mineralization, immobilization, and nitrification. In
Methods of soil analysis. Part 2: microbiological and biochemi
-
cal properties. Soil Science Society of America, Madison, Wis.
Book Ser. 5.
Jackson, L.E., Strauss, R.B., Firestone, M.K., and Bartolome, J.W.
1990. Influence of tree canopies on grassland productivity and
nitrogen dynamics in deciduous oak savanna. Agric. Ecosyst.
Environ. 32: 89–105.
Joffre, R., and Rambal, S. 1988. Soil water improvement by trees
in the rangelands of southern Spain. Acta Oecol. Oecol. Plant.
9: 405–422.
Kellman, M. 1979. Soil enrichment by neotropical savanna trees. J.
Ecol. 67: 565–577.
Kennard, D.G., and Walker, B.H. 1973. Relationships between tree
canopy cover and Panicum maximum in the vicinity of Fort Vic-
toria. Rhod. J. Agric. Res. 11: 145–153.
Kenward, M.G., and Roger, J.H. 1997. Small sample inference for
fixed effects from restricted maximum likelihood. Biometrics,
53: 983–997.
Littell, R.C., Milliden, G.A., Stroup, W.W., and Wolfinger, R.D.
1996. SAS system for mixed models. SAS Institute Inc., Cary,
N.C.
March, W.J., and Skeen, J.N. 1976. Global radiation beneath the
canopy and in a clearing of a suburban hardwood forest. Agric.
Meteorol. 16: 321–327.
McPherson, G.R. 1997. Ecology and management of North Ameri
-
can savannas. University of Arizona Press, Tucson.
Messier, C., Honer, T., and Kimmins, J.P. 1989. Photosynthetic
photon flux density, red:far-red ratio, and minimum light re
-
quirement for survival of Gaultheria shallon in western red
cedar – western hemlock stands in coastal British Columbia.
Can. J. For. Res. 19: 1470–1477.
Minckler, L.S., Woerheide, J.D., and Schlesinger, R.C. 1973. Light,
soil moisture, and tree reproduction in hardwood forest openings.
USDA For. Serv. Res. Pap. NC-89.
Mitchell, P.L., and Whitmore, T.C. 1993. Use of hemispherical pho
-
tographs in forest ecology. Oxford Forestry Institute, Department
of Plant Sciences, University of Oxford, Oxford, U.K. Publ. 44.
Mitchell, R.J., Zutter, B.R., Gjerstad, D.H., Glover, G.R., and Wood,
C.W. 1999. Competition among secondary-successional pine com
-
munities: a field study of effects and responses. Ecology, 80: 857–
872.
Mladenoff, D.J. 1987. Dynamics of nitrogen mineralization and ni
-
trification in hemlock and hardwood treefall gaps. Ecology, 68:
1171–1180.
Morgan, D.C., Warrington, I.J., and Rook, D.A. 1985. Some obser
-
vations on the spectral distribution characteristics of short-wave
radiation within Pinus radiata D. Don canopies. Plant Cell En
-
viron. 8: 201–206.
Obot, E.A. 1988. Estimating the optimum tree density for maxi
-
mum herbaceous production in the Guinea Savanna of Nigeria.
J. Arid Environ. 14: 267–273.
I:\cjfr\cjfr31\cjfr-05\X01-003.vp
Tuesday, April 17, 2001 3:16:19 PM
Color profile: Disabled
Composite Default screen
© 2001 NRC Canada
778 Can. J. For. Res. Vol. 31, 2001
Pacala, S.W., Canham, C.D., and Silander, J.A., Jr. 1993. Forest
models defined by field measurements I. The design of a north
-
eastern forest simulator. Can. J. For. Res. 23: 1980–1988.
Palik, B.J., Mitchell, R.J., Houseal, G., and Pederson, N. 1997. Ef
-
fects of canopy structure on resource availability and seedling
responses in a longleaf pine ecosystem. Can. J. For. Res. 27:
1458–1464.
Palik, B.J., Mitchell, R.J., and J.K. Hiers. 2001. Modeling silvi
-
culture after natural disturbances to sustain biodiversity in the
longleaf pine (Pinus palustris) ecosystem: balancing complexity
and implementation. For. Ecol. Manage. In press.
Parker, V.T., and Muller, C.H. 1982. Vegetational and environmen
-
tal changes beneath isolated live oak trees (Quercus agrifolia)in
a California annual grassland. Am. Midl. Nat. 107: 69–81.
Parsons, W.F.J., Knight, D.H., and Miller, S.L. 1994. Root-gap dy
-
namics in a lodgepole pine forest: nitrogen transformations in
gaps of different size. Ecol. Appl. 4: 354–362.
Platt, W.J., Evan, G.W., and Rathbun, S. 1988. The population dy
-
namics of a long-lived conifer (Pinus palustris Mill.). Am. Mid.
Nat. 131: 491–525.
Radwanski, S.A., and Wickens, G.E. 1967. The ecology of Acacia
albida on mantle soils in Zalingei Jebbel Marra, Sudan. J. Appl.
Ecol. 4: 569–579.
Rebertus, A.J., Williamson, G.B, and Moser, E.B. 1989. Longleaf
pyrogenicity and turkey oak mortality in Florida xeric sandhills.
Ecology, 70: 60–70.
Reich, P.B., Peterson, D.A., Wrage, K., and Wedin, D. 2001. Fire
and vegetation effects on productivity and nitrogen cycling across
a forest–grassland continuum. Ecology, 82. In press.
Rich, P.M., Clark, D.B., Clark, D.A., and Oberbauer, S.F. 1993.
Long-term study of solar radiation regimes in a tropical wet forest
using quantum sensors and hemispherical photography. Agric.
For. Meteorol. 65: 107–127.
Rudolph, D.C., and Conner, R.N. 1996. Red-cockaded woodpeck-
ers and silvicultural practice: is uneven-aged silviculture prefer
-
able to even-aged? Wild. Soc. Bull. 24: 330–333.
SAS Institute Inc. 2000. SAS/STAT® and base SAS®, version 8.0
release edition. SAS Institute Inc., Cary, N.C.
Scholes, R.J., and Archer, S.R. 1997. Tree–grass interactions in sa
-
vannas. Annu. Rev. Ecol. Syst. 28: 517–544.
Seber, G.A.F., and Wild, C.J. 1989. Nonlinear regression. John
Wiley & Sons, New York.
Snedecor, G.W., and Cochran, W.G. 1989. Statistical methods. 8th
ed. Iowa State University Press, Ames.
U.S. Department of Agriculture (USDA). 1986. Soil survey of Baker
and Mitchell counties. Soil Conservation Service, Albany, Ga.
Van Pelt, R., and Franklin, J.F. 1999. Response of understory trees
to experimental gaps in old-growth Douglas-fir forests. Ecol.
Appl. 9: 504–512.
Vetaas, O.R. 1992. Micro-site effects of trees and shrubs in dry sa
-
vannas. J. Veg. Sci. 3: 337–344.
Vitousek, P.M., and Denslow, J.S. 1986. Nitrogen and phosphorus
availability in treefall gaps of a lowland tropical rainforest. J.
Ecol. 74: 1167–1178.
Vitousek, P.M., and Walker, L.R. 1989. Biological invasion by Myrica
faya in Hawaii: plant demography, nitrogen fixation, ecosystem ef
-
fects. Ecol. Monogr. 59: 247–265.
Weltzin, J.F., and Coughenour, M.B. 1990. Savanna tree influence on
understory vegetation and soil nutrients in northwestern Kenya. J.
Veg. Sci. 1: 325–332.
Whitmore, T.C., Brown, N.D., Swaine, M.D., Kennedy, D., Goodwin-
Bailey, C.I., and Gong, W.K. 1993. Use of hemispherical photo-
graphs in forest ecology: measurement of gap size and radiation to-
tals in a Bornean tropical rain forest. J. Trop. Ecol. 9: 131–151.
Williamson, G.B., and Black, E.M. 1981. High temperature of for-
est fires under pines as a selective advantage over oaks. Nature
(London), 293: 643–644.
I:\cjfr\cjfr31\cjfr-05\X01-003.vp
Tuesday, April 17, 2001 3:16:20 PM
Color profile: Disabled
Composite Default screen