255.full

7/30/2019 255.full

1/7

255

American Journal of Botany 90(2): 255261. 2003.

EFFECTS OF TREE CROWN STRUCTURE ON BIOMASS OF

THE EPIPHYTIC FERN POLYPODIUM SCOULERI

(POLYPODIACEAE) IN REDWOOD FORESTS1

STEPHEN C. SILLETT2 AND MARK G. BAILEY

Department of Biological Sciences, Humboldt State University, Arcata, California 95521 USA

Redwood forests contain some of the largest and most structurally complex trees on Earth. The most abundant vascular epiphyte in

these forests is the fern Polypodium scouleri (Polypodiaceae). We measured dimensions of all 765 P. scouleri mats on 32 trees (27

Sequoia sempervirens, 5 Picea sitchensis). Eighteen P. scouleri mats from 11 trees were randomly selected for removal and dissection

in the laboratory. The total fern mat mass consisted of live fronds (3.3%), dead fronds (2.4%), live rhizomes (4.2%), dead rhizomes

(8.9%), roots (34.4%), humus (28.0%), and debris (18.8%). We used multiple regression analysis to develop equations for estimation

of fern masses, and we applied these equations to undisturbed fern mats on the 32 trees. Individual trees supported up to 742 kg dry

mass of P. scouleri mats. These are the highest whole-tree epiphyte masses ever reported. We also quantified crown structure and

counted the number of vascular plant species occurring as epiphytes on each tree. Very large, complex trees had more fern mat mass

and higher vascular epiphyte species richness than smaller, simpler trees. Desiccation-sensitive organisms dependent on water stored

in fern mats may be unable to survive in managed redwood forests lacking large, complex trees with abundant P. scouleri.

Key words: epiphyte biomass; Polypodiaceae; Polypodium scouleri; Sequoia sempervirens; temperate rain forest; tree structure;

vascular epiphytes.

Rain forest canopies are celebrated for their vascular epi-phytes, whose ecology has been the focus of considerable re-search (see Benzing, 1990, 1995). Large tree crowns providemyriad epiphytic habitats, and plants are distributed along gra-dients of substrate character and microclimate (ter Steege andCornelissen, 1989; van Leerdam, Zagt, and Veneklaas, 1990;Wolf, 1994; Freiberg, 1996; Sillett and Neitlich, 1996). Largebranches and crotches of the inner and lower crown often sup-port canopy soils that are exploited by desiccation-sensitivespecies, including accidental epiphytes whose primary habitatis the forest floor. Small branches of the outer and upper crown

are often occupied by drought-adapted species capable ofwithstanding frequent and severe desiccation.

The majority of studies on vascular epiphytes have focusedon tropical rain forests, presumably because temperate rainforests harbor far fewer species. Most epiphyte studies in tem-perate rain forests have focused on lichens and bryophytes(e.g., Pike et al., 1975; Sillett, 1995; Clement and Shaw,1999). Some temperate rain forests, however, harbor abundantvascular epiphytes. Ferns, shrubs, and trees regularly occur asepiphytes in Sequoia sempervirens (D. Don.) Endl. (hereafterredwood) forests.

The most abundant vascular epiphyte in redwood forests isthe evergreen fern, Polypodium scouleri Hook. & Grev. (Sil-lett, 1999). The geographic range of P. scouleri is strictlycoastal, extending from Alaska to Baja California, where itgrows epiphytically in foggy forests and terrestrially in salt-spray zones (Whitmore, 1993). In forests, this species formssprawling mats on branches and trunks with coriaceous frondsup to a meter long and succulent rhizomes embedded in can-opy soils. Redwood is also a coastal species, but it occurs

1 Manuscript received 12 October 2001; revision accepted 15 August 2002.The authors thank the following Humboldt State University students for

assistance with fieldwork: Anthony Ambrose, Sunny Bennett, Nolan Bowman,Billy Ellyson, Clint Jones, Brett Lovelace, Jim Spickler, Susan Stearns, andCameron Williams. Funding was provided by a grant from Global Forest (GF-18-1999-48).

2 Author for reprint requests (e-mail: [email protected]).

farther inland than P. scouleri. Some of the tallest redwoodforests, which occur on sheltered alluvial terraces 30 km fromthe ocean, have no P. scouleri in their canopies. Indeed, vascular epiphytes are abundant only in old-growth redwood for-ests within 10 km of the ocean. These rain forests containsome of the largest and most structurally complex redwoods(Van Pelt, 2001) as well as the tallest known Picea sitchensis(Bong.) Carr. trees (hereafter Sitka spruce).

While the great height of redwoods has long been recog-nized, the extreme complexity of their crowns has only recently been appreciated. Canopy studies since 1996 have re-

vealed that large redwoods have individualized crowns withmultiple, reiterated trunks arising from other trunks andbranches (Sillett, 1999; Sillett and Van Pelt, 2000, 2001). Reiterated trunks in redwood are orthotropic stems with their ownsystems of plagiotropic branches (Sillett, 1999). They are indistinguishable from free-standing trees except for their originswithin the crown of a larger tree. Fusions among reiteratedtrunks and their branches are common. Large redwood crownsalso retain considerable quantities of rotting wood in brokentrunks, dead branches, and fire cavities.

This study focused on P. scouleri in redwood forest canopies. Using rope-based methods, we climbed 32 trees of twospecies, redwood and Sitka spruce, to quantify both epiphyticferns and crown structure. We had three specific objectives

(1) to develop a nondestructive method to estimate masses ofP. scouleri mats, (2) to estimate whole-tree masses of P. scouleri on both redwood and Sitka spruce, and (3) to assess howtree structure may affect the abundance of vascular epiphytesin redwood rain forests.

MATERIALS AND METHODS

Study areaThis study occurred in the old-growth redwood forests o

Humboldt and Del Norte Counties, California, USA. Mean annual rainfall i

1.7 to 2.5 m, with mean summer temperature ranges from 4 to 29C and

mean winter temperature ranges from 1 to 16C (Redwood National and

7/30/2019 255.full

2/7

256 [Vol. 90AMERICAN JOURNAL OF BOTANY

TABLE 1. Locations, dimensions, and vascular epiphytes of 27 redwoods and five Sitka spruce trees included in this study. For each creek, treesare listed in descending order by total fern mat mass. Fern masses are given as estimated totals 1 SE. Overall complexity (PC1) numbersrange from 1, most complex, to 0, least complex (see text). DBH main trunk diameter at breast height. Epiphyte richness is the total numberof vascular plant species occurring as epiphytes on each tree. Asterisks next to tree names indicate members of the top ten largest known livingredwoods (Van Pelt, 2001).

Tree name Species Creek Overall

complexity Height (m) DBH (m)Main trunk

volume (m3)Reiteration

volume (m3)

No. ofreiterated

trunks

Diameterof largest

branch(m)

No. offern mats

Epi-phyte

richnessTotal fern

mat mass (kg)

AtlasPrometheusPleiades 2RheaKronosPleiades 1BallantineIluvatar*Bell

redwoodredwoodredwoodredwoodredwoodredwoodredwoodredwoodredwood

BoyesBoyesBoyesBoyesBoyesBoyesBoyesBoyesBoyes

0.750.460.420.050.460.190.441.000.39

90.097.490.595.591.694.497.291.595.7

7.15.65.34.14.34.85.26.16.0

652.4618.1584.3359.1339.2373.0548.8905.8633.2

97.348.226.7

1.127.6

6.417.1

127.011.7

258833174951

112134

85

1.50.62.10.42.30.81.12.00.6

653637513911163820

655232255

677 33236 29229 30207 32177 26148 16124 17119 23

66 3ZeusBroken TopDemeterFloating RaftVantageTrapdoorYang

Pleasant SurpriseYinSunny DelightLunarHorntailClean BarrelTwin TowerFangorn

redwoodredwoodredwoodredwoodredwoodredwoodredwood

redwoodredwoodsprucesprucesprucesprucespruceredwood

BoyesBoyesBoyesGodwoodGodwoodGodwoodGodwood

GodwoodGodwoodGodwoodGodwoodGodwoodGodwoodGodwoodMill

0.210.140.240.170.210.000.25

0.000.12

0.56

101.080.997.582.186.777.095.8

86.997.087.089.582.587.292.276.6

5.24.24.33.55.03.53.3

3.33.71.51.62.01.52.24.9

608.4322.0400.5290.5312.8174.8234.8

227.9260.3

82.580.1

100.875.3

153.1580.0

6.88.2

20.216.6

6.61.1

20.8

3.04.00.00.00.00.00.0

36.4

2530556255

561

1154

00000

23

0.90.70.50.90.80.71.6

0.30.60.20.20.20.20.32.1

1424

93417

810

59

4912

7101135

2218313

14222226

62 1862 2333 12

101 467 863 225 12

13 112 243 516 712 211 210 2

742 34Lost Monarch*El Viejo del Norte*StalagmightAldebaranAragornDel Norte Titan*Terex Titan*Adventure*

redwoodredwoodredwoodredwoodredwoodredwoodredwoodredwood

MillMillMillMillMillMillPrairiePrairie

0.510.860.360.280.390.800.540.37

98.098.786.793.085.793.780.2

101.3

7.76.75.03.94.27.26.64.9

977.9871.7471.1370.5349.9945.6873.5910.1

11.554.715.810.535.399.157.324.1

4345261223434241

1.02.81.61.81.61.41.21.3

3747193741381719

56544684

454 32383 31

74 1669 1956 1648 24

117 1939 12

State Parks). These coastal forests experience a summer dry season that is

ameliorated by persistent fog (Dawson, 1998).

Study areas included the redwood-dominated forests along Boyes, God-

wood, and Prairie Creeks in Prairie Creek Redwoods State Park and Mill

Creek in Jedediah Smith Redwoods State Park. Thirty-two trees, including

several of the largest known living redwoods, were selected from these forests

for detailed study (Table 1). All trees are rooted below 100 m a.s.l. and stand

less than 7 km from the Pacific Ocean. Along Godwood Creek, one redwood

and one Sitka spruce were chosen from each of five 0.1-ha reference stands.

The mean basal area of these stands is 402 m2/ha with redwood and Sitka

spruce accounting for 75.9% and 20.1% of this total, respectively. Other tree

species include Tsuga heterophylla (Raf.) Sarg., Chamaecyparis lawsoniana

(A. Murr.) Parl., and Rhamnus purshiana DC. Understory vegetation is dom-

inated by Polystichum munitum (Kaulf.) Presl., Vaccinium ovatum Pursh., andOxalis oregana Nutt. The redwoods along Boyes Creek occur within a 1.0-

ha reference stand. This stand has a basal area of 394 m 2/ha, 93.7% of which

is redwood. Other tree species in the stand include Acer macrophyllum Pursh.,

R. purshiana, Pseudotsuga menziesii (Mirb.) Franco, T. heterophylla, and Um-

bellularia californica Nutt. Like the forest along Godwood Creek, understory

vegetation is dominated by P. munitum, V. ovatum, and O. oregana. The

forests along Prairie and Mill Creeks are similar in floristic composition to

those along Boyes and Godwood Creeks.

Canopy accessLarge trees with full crowns were selected from each for-

est for climbing. Excessively leaning trees were avoided. We accessed tree

crowns by shooting rubber-tipped arrows trailing fishing filament over sturdy

branches with a compound bow mounted to a spinning reel. A nylon cord,

followed by a static kernmantle rope, was then hauled over the branches. One

end of the rope was anchored at ground level, and the other end was climbed

via mechanical ascenders. Arborists techniques (see Jepson, 2000) were used

to access progressively higher branches and to move laterally through tree

crowns.

Fern samplingLocations and dimensions of all epiphytic P. scouleri mats

on each tree were recorded. We defined mats as assemblages of ferns with

no rhizomatous connections to other ferns. The height and diameter of sup-

porting branch (if any) of each fern mat were recorded. We also noted loca-

tions of fern mats by recording whether they occurred on branches, in crotch-

es, or on trunks. Five dimensions of each fern mat were measured: number

of live fronds, surface area (i.e., [mat length/2] [mat width/2]), meandepth, maximum frond length, and maximum number of pinnae per frond.

Depth was measured with a thin metal probe graduated at 1-cm intervals.

Five depth measurements were usually taken across the area of the mat, but

large mats required more measurements to obtain an average representative

of the entire mat. We measured depths of all fern mats in trees along Godwood

Creek but only the largest fern mats in trees along Boyes, Mill, and Prairie

Creeks. When depth data were available, the product of mean depth and sur-

face area was used to estimate fern mat volume.

We randomly selected 17 of the 158 P. scouleri mats occurring on five

redwoods and five Sitka spruces along Godwood Creek for destructive sam-

pling. The mats were first stratified into size classes according to the number

of live fronds counted in the field. Since redwoods along Boyes, Mill, and

7/30/2019 255.full

3/7

February 2003] 257SILLETT AND BAILEYEPIPHYTIC FERNS IN REDWOOD FORESTS

TABLE 2. Product-moment correlations (r) between dimensions andcomponent masses (in kilograms) of 18 Polypodium scouleri maton 11 trees in old-growth redwood forests in Humboldt CountyCalifornia, USA.

Componentmasses

No. oflive fronds

Maximumfrond length

Maximumno. of pinnae Volume

Total mass

Live frondsLive rhizomesRootsDead frondsDead rhizomesHumusDebris

0.893

0.8570.9630.9870.7400.9690.9780.912

0.649

0.6310.6990.5930.5620.6510.6680.800

0.715

0.7060.5780.4890.6650.5080.5390.721

0.996

0.9920.8960.9870.9630.9120.9440.892

Prairie Creeks occasionally have fern mats larger than the largest size class

found in the canopy along Godwood Creek, an additional very large fern mat

was randomly selected from the Boyes Creek reference stand. In the trees,

fern mats were cut into pieces and stripped down to bare bark. A tarp was

suspended beneath large mats to catch material that broke loose during the

removal process. A total of 18 fern mats were thus removed from the canopy

and transported to the laboratory.

In the laboratory, fern mats were individually dissected and separated into

the following components: live fronds, dead fronds, live rhizomes, dead rhi-

zomes, roots, humus, and debris (e.g., leaf litter, bark, twigs, cones). The high

volume and interwoven character of non-green material (hereafter gross hu-

mus) made it impractical to completely separate individual components of

large fern mats. Roots were particularly difficult to separate from humus dur-

ing the initial sorting. Consequently, roots that could be easily separated at

this point were removed, and the remaining clumps of gross humus were

marked for subsampling. Each component, including gross humus, was

bagged, dried at 60C until a stable dry mass was obtained, and weighed to

the nearest 0.1 g.

One-tenth of the gross humus was wet-sieved through a 0.5-mm screen.

All material passing through the screen was considered humus. This material

was washed down the drain and lost during processing. The remaining ma-

terial within a subsample consisted of rhizomes easily removed by hand (i.e.,

those few pieces that escaped notice during initial sorting) as well as roots

and debris that were separated by using forceps. After sorting, componentswere dried at 60C for 24 h and weighed to the nearest 0.1 g. Proportions of

subsample masses attributable to humus, rhizomes, roots, and debris were then

determined and used to calculate masses of these components in entire gross

humus samples. Gross humus component masses were then added to the com-

ponent masses obtained during initial sorting.

Tree mappingWe quantified tree crown structure by measuring dimen-

sions of the main trunk and all reiterated trunks over 5 cm basal diameter.

Trunk diameters were measured directly with a tape, and total tree height was

measured by lowering a tape from topmost foliage to average ground level.

Tags were attached to the main trunk at 5-m intervals for use as benchmarks

in height measurements of reiterated trunks, branches, and epiphytes. The

following data were recorded for each trunk: top height, height of origin, basal

diameter, and diameter at 5-m intervals along the length of the trunk. For

trunks arising from branches, branch basal diameter was also measured. Dur-

ing mapping, we also recorded the number of vascular plant species occurring

as epiphytes on each tree.

Data analysesWe used regression analysis to estimate epiphytic fern

masses from nondestructive measurements of fern mat dimensions. Model

equations based on data collected from the eighteen destructively sampled

fern mats were developed via stepwise multiple regression (SAS, 2001). Fern

mat mass was the dependent variable. Fern mat dimensions (i.e., number of

live fronds, maximum frond length, maximum number of pinnae, and mat

volume) were used as independent variables. The resulting equations were

used to estimate masses of each fern mat described on all 32 trees surveyed

in this study. Whole-tree estimates of fern mat masses were obtained by sum-

ming estimated values for all fern mats on a tree.

Mapping data were used to generate the following 11 tree structure vari-ables for analysis: tree height, main trunk diameter at breast height (DBH),

main trunk volume, volume of reiterated trunks arising from other trunks,

volume of reiterated trunks arising from branches, percentage of total trunk

volume in reiterations, percentage of total reiterated trunk volume on branch-

es, total number of reiterated trunks, number of reiterated trunks 1 m basal

diameter, number of branches 0.5 m basal diameter, and diameter of the

trees largest branch. We calculated trunk volumes by applying the equation

for a regular conic frustum (i.e., volume length /3 [lower radius2

lower radius upper radius upper radius2]) to the trunk diameter data.

Since there was strong multicollinearity among the tree structure variables,

we used principal components analysis (PCA) to derive orthogonal linear

combinations, or components, of these variables (McCune and Mefford,

1999). Interpretation of principal components was based on correlations (r

between the original variables and the principal components themselves.

We explored potential relationships between epiphytes and tree structur

by examining correlations among tree-level variables. Epiphyte variables in

cluded the number of fern mats per tree, total fern mat mass, mass of the

trees largest fern mat, and number of epiphytic vascular plant species pe

tree. Tree structure variables included the 11 original variables and their sig-nificant principal components.

RESULTS

Model equationsCorrelations among fern mat dimensionsand component masses were strong (Table 2), and model equa-tions for fern mat masses had high predictive power. The bestequation for estimating fern mat mass was the following: totamass 372 0.326 volume 33.9 maximum frondlength (adjusted R2 0.995, P 0.00001). Since volumes ofmany fern mats in trees along Boyes, Mill, and Prairie Creekswere not measured, fern mat masses often had to be estimatedwith equations lacking a volume term. The following equationwas used to estimate masses of these fern mats: total mass

1470

57.3

number of live fronds (adjusted R2

0.785P 0.0001).

Fern mat dimensions and compositionA total of 765 Pscouleri mats occurred on the 32 trees surveyed in this studyThese mats had up to 192 kg of dry mass, 2280 live fronds21 m2 of surface area, 1.0 m of humus depth, and they con-tained fronds up to 103 cm long with up to 41 pinnae. Totafern mat mass consisted mostly of roots (34.4%), humu(28.0%), and debris (18.8%). The remainder consisted of livefronds (3.3%), dead fronds (2.4%), live rhizomes (4.2%), anddead rhizomes (8.9%). Fern mat mass was positively correlat-ed with branch diameter (r 0.46, P 0.0001) and uncorrelated with height above ground, although maximum frondlength and number of pinnae were negatively correlated with

height (r 0.38 and 0.28 respectively, P 0.001).Along Godwood Creek, redwoods supported larger P. scou

leri mats than adjacent Sitka spruces (mean fern mat mass 3.6kg vs. 0.9 kg, P 0.0002). Over one-fifth of the mats on theredwoods exceeded 5 kg, but only one of the fern mats on theSitka spruces was this large. Despite these differences in in-dividual fern mat sizes, the composition of fern mats on thetwo tree species was nearly identical. Mat size differencewere partly attributable to differences in supporting branchdiameters; redwoods had thicker branches than Sitka spruces(Table 1). Redwoods also had far more complex crowns thanadjacent Sitka spruces. For example, none of the Sitka spruces

7/30/2019 255.full

4/7


TABLE 3. Coefficients of determination (R2) for correlations between tree structure and whole-tree epiphyte variables (N 27 redwoods). Asterisksindicate significant correlations (P 0.05). All significant correlations are positive. Overall complexity numbers were derived from principalcomponents analysis (see Table 4). DBH main trunk diameter at breast height.

Tree structure variablesNo. of

fern matsTotal fern

mat mass (kg)Mass of largest

fern mat (kg)Epiphyterichness

Overall complexity (PC1)No. of reiterated trunks 1 m diameter

DBH (m)Volume of main trunk (m3)Volume of reiterated trunks arising from other trunks (m3)Total volume in reiterations (%)No. of branches 0.5 m diameterDiameter of largest branch (m)Volume of reiterated trunks arising from branches (m3)Total number of reiterated trunksHeight (m)Reiteration volume on branches (%)

0.34*0.14

0.21*0.140.23*0.30*0.16*0.19*0.16*0.000.000.00

0.22*0.10

0.23*0.120.060.100.16*0.24*0.20*0.020.020.01

0.31*0.31*

0.31*0.19*0.120.130.19*0.16*0.18*0.000.010.00

0.38*0.25*

0.28*0.31*0.23*0.24*0.18*0.20*0.140.020.030.01

TABLE 4. Correlation coefficients (r) between 11 tree structure variables and their significant principal components (N 27 redwoods). Asterisksindicate significant correlations (P 0.001). DBH main trunk diameter at breast height.

Tree structure variables PC1 PC2 PC3

No. of reiterated trunks 1 m diameterDBHVolume of main trunk (m3)Volume of reiterated trunks arising from other trunks (m3)Total volume in reiterations (%)No. of branches 0.5 m diameterDiameter of largest branch (m)Volume of reiterated trunks arising from branches (m3)Total number of reiterated trunksHeight (m)Reiteration volume on branches (%)

0.87*0.80*0.79*0.79*0.74*0.70*0.68*0.64*0.420.150.06

0.350.130.110.530.30

0.520.56*0.69*

0.250.02

0.91*

0.200.330.38

0.210.45

0.170.310.17

0.410.81*0.10

in this study had any reiterations (Table 1). Further analysesfocused on redwood crown structure and its potential effectson epiphytic ferns.

Tree crown structureThe redwoods sampled in this studyhad complex crowns with reiterated trunks accounting for upto 13% of their total trunk volume (Table 1). Many of theredwoods had multiple trunks and branches over 1 m diameter.These massive crowns supported up to eight species of vas-cular epiphytes, including enormous quantities of P. scouleriand associated material (Table 1). Thirteen redwoods had over100 kg of fern mats. The highest estimated fern accumulationon a single tree was 742 kg.

The largest fern mats occurred at the bases of huge branch-es, in crotches at the bases of large reiterations, and on broadplatforms at the tops of broken trunks. Of the total fern masson redwoods, 79.9% occurred on branches. The remainder oc-curred in crotches at the bases of reiterations (11.5%) or on

trunks (8.5%). Fern mats differed in frond dimensions de-pending on crown location. For comparably sized mats, frondswere significantly (P 0.01) larger in crotches than onbranches or trunks.

There were a number of significant positive correlationsamong the tree structure and epiphyte variables (Table 3). Mul-ticollinearity among the tree structure variables was eliminatedby PCA, which extracted three significant components (eigen-values 4.76, 2.51, 1.78) accounting for 79% of the totalvariation in the 11 original variables (Table 4). Principal com-ponent 1 represented the overall size and complexity of a tree

and its reiterated trunks. Very large, complex trees supportedmore and larger fern mats and had higher vascular epiphytespecies richness and than smaller, simpler trees (Table 3). Nei-ther PC2 nor PC3 were correlated with tree-level epiphyteabundance.

DISCUSSION

Within-tree fern distributionVery thick branches supportlarger fern mats than thinner branches simply because thickerbranches have broader platforms for debris accumulation andtheir fern mats have developed over the long life of the branch.This positive correlation between branch diameter and epi-phyte abundance has also been observed in other forests (e.g.,Ingram and Nadkarni, 1993; Clement and Shaw, 1999). Agedifferences among fern mats may account for the positive cor-relation between fern mat mass and maximum frond dimen-sions. The first P. scouleri frond emerging from a fertilized

gametophyte consists of a single pinna, and as the sporophytedevelops into a fern mat, frond lengths and numbers of pinnaeincrease. Small (0.5 kg) P. scouleri mats rarely have frondsover 30 cm long with more than 11 pinnae, while large (5kg) mats frequently have fronds over 70 cm long with morethan 23 pinnae.

Fern mats of the same size, however, may differ in fronddimensions depending on their locations within a tree crown.Crotches are better sites for accumulation than branches ortrunks, which will tend to shed material more easily from theirsteep sides. Fern mats in crotches receive more stemflow (An-

7/30/2019 255.full

5/7


drade and Nobel, 1997) and store more water than fern matson branches (Ambrose, 2003). A favorable moisture regimemay permit fern mats to develop more quickly in crotches,thus explaining why these mats have larger fronds than moredesiccation-prone mats on branches and trunks.

Coriaceous leaves and succulent rhizomes confer substantialdesiccation resistance on some epiphytic ferns (Hietz and Bri-

ones, 1998). Despite these adaptations, P. scouleri cannotwithstand severe droughts and is restricted to coastal environ-ments throughout its range (Whitmore, 1993). Even withinredwood rainforest canopies, vertical microclimatic gradientsappear to affect its morphology. Regardless of overall fern matsize, maximum frond dimensions become progressively small-er with increasing height. During rainless periods, the upperforest canopy is brighter, warmer, windier, and less humid thanthe lower canopy (Parker, 1995). Thus, P. scouleri is subjectedto more frequent and severe periods of desiccation in the uppercanopy than in the lower canopy. Thick cuticles contribute tovery low rates of uncontrolled water loss from vascular epi-phyte leaves (Helbsing, Reiderer, and Zotz, 2000), but a largerfrond loses more water than a smaller frond due to higher leafarea. Upper canopy fern mats may therefore produce smaller

fronds than lower canopy mats because rapid desiccation re-stricts frond growth.

The succulent rhizomes of P. scouleri store water that canbe tapped by transpiring fronds, but live rhizomes representless than 5% of fern mat mass, and their supply of stored wateris limited. Roots supply the bulk of water lost during evapo-transpiration. Like other vascular epiphytes, P. scouleri investsa large proportion of its fixed carbon in root production inorder to efficiently extract water and nutrients from desicca-tion-prone canopy soils (Vance and Nadkarni, 1992; Bohlman,Matelson, and Nadkarni, 1995).

The ability of fern roots to extract moisture from organicmatter in large mats depends upon the depth to which theypenetrate humus layers. Unfortunately, this cannot be fully

evaluated with our data because we were unable to distinguishliving from dead roots in our samples. Material sloughing fromthe bottoms of thick mats was often poorly consolidated andapparently lacking viable fern roots. It remains unclear wheth-er or not P. scouleri can extract water stored in the deepesthumus layers of very large mats. Even if it is unavailable toepiphytic ferns, other organisms in the canopy utilize thisstored water.

Ecological importanceDebris from tree litterfall, espe-cially foliage, is a major component of fern mats, but most ofthe dead organic matter in the mats comes from P. scouleriitself, especially humus derived from its roots. Furthermore,as mats develop, they occupy progressively larger surface ar-eas that can more effectively capture falling debris. Therefore,

P. scouleri is directly responsible for much, if not most, of theretention and accumulation of dead organic matter in large treecrowns in redwood rain forests.

The water stored by arboreal dead organic matter enablesterrestrial organisms to flourish high above the ground. A widevariety of vascular plants, including ferns (e.g., Athyrium filix-femina (L.) Roth, Polystichum munitum), shrubs (e.g., Gaul-theria shallon Pursh., Ribes spp., Sambucus callicarpa Greene,Vaccinium spp.), and trees (e.g., Lithocarpus densiflorus (H.& A.) Rehd., Rhamnus purshiana, Picea sitchensis, Pseudo-tsuga menziesii, Tsuga heterophylla, Umbellularia californica)occur as epiphytes amidst P. scouleri mats (Sillett, 1999). Ter-

restrial animals, including salamanders, mollusks, earthwormsand a plethora of arthropods, also frequently inhabit thesemats. Many of these organisms are highly sensitive to desic-cation, and it is unlikely that they could survive in redwoodcanopies without substantial quantities of P. scouleri and associated material.

The trees also directly exploit the water and nutrients stored

in their epiphytic fern mats. We occasionally find well-devel-oped adventitious tree roots amidst P. scouleri in both redwood and Sitka spruce. These roots are indistinguishable fromtheir counterparts on the forest floor; both possess well-devel-oped mycorrhizal associations. A similar phenomenon hasbeen observed in a wide variety of other temperate and tropicarain forest canopies (Nadkarni, 1981, 1994). In redwoods, adventitious roots are most frequently encountered at the basesof large reiterated trunks. The extent to which these trunkrely upon water and nutrients absorbed from fern mats remainsto be determined. The size and complexity of the roots systemswe have observed suggest that their contributions are not triv-ial.

Effects of tree structureBy measuring both epiphytes and

tree structure, our results permit a quantitative assessment ofcrown-level complexity and its potential effects on the abundance of vascular epiphytes. Very large, complex redwoodsupport more fern mass and have higher epiphyte species rich-ness than smaller redwoods. This positive correlation betweentree size and vascular epiphyte abundance has been observedin other forests (e.g., Hietz and Hietz-Seifert, 1995) and imost likely a consequence of tree age. Unfortunately, the ageof large redwoods is nearly impossible to determine withoutcausing unnecessary damage to the trees. The largest redwoods in this study may be over 2000 yr old (Sawyer et al.1999), while the smallest are probably less than 1000 yr oldOverall size and complexity (PC1) are undoubtedly correlatedwith redwood age, but the correlation between tree size and

age is often weak in shade-tolerant conifers (see Lyons, Nadkarni, and North, 2000).A number of factors other than tree size and age contribute

to crown-level complexity and epiphyte accumulation in red-woods. Similar-sized trees standing side by side in the samestand can support very different quantities of P. scouleri (e.g.Demeter and Rhea), some of the largest, most complex treessupport relatively small quantities of P. scouleri (e.g., DeNorte Titan), and some relatively small trees with simplecrowns support relatively large quantities of P. scouleri (e.g.Rhea). The sources of such variation remain unclear, but dif-ferences in crown-level microclimates may be importantSome of the deepest accumulations of organic matter, whichoccur in crotches in the lower canopy where light levels arevery low (Ambrose, 2003), contain very little if any P. scou

leri. Redwoods with very dense crowns may have a dispro-portionate amount of poorly illuminated substrates that arecurrently unsuitable for P. scouleri even though this speciesmay have thrived on these sites and contributed much of theorganic matter at earlier times when conditions were more favorable for fern growth.

The disturbance history of individual trees and forest standsalso influences epiphytic ferns. Fires, crown failures, and adjacent treefalls can lead to sudden, often dramatic changes intree structure, microclimate, and epiphyte abundance. Firescars, snapped trunks, and broken branches are widespread inlarge redwoods, whose crowns are highly individualized and

7/30/2019 255.full

6/7


asymmetrical (Sillett, 1999). The mass of epiphytes on anylarge redwood is thus a dynamic quantity affected by both theseverity and timing of past disturbance events. For example,a powerful storm in March 1998 dislodged a large fern matfrom one redwood (Trapdoor) when an adjacent tree fell intothe side of its crown (see Taylor, 2001). The total fern masson this tree was instantly diminished by over 40%.

Implications of whole-tree estimatesVery few studiessupply epiphyte mass estimates for whole trees. The highestpreviously reported estimates are 115.0 and 141.9 kg for singletrees in tropical cloud forests (Nadkarni, 1984; Hofstede, Wolf,and Benzing, 1993). By comparison, the Sitka spruces wesampled along Godwood Creek support a mean of 177 kg ofepiphytic material, including lichens, bryophytes, ferns, andassociated organic matter (Ellyson and Sillett, in press). Theredwoods we sampled along Godwood, Boyes, and MillCreeks support means of 47, 178, and 261 kg of P. scoulerimats, respectively. Many other epiphytes also occur on theseredwoods. Several trees (e.g., El Viejo del Norte, FloatingRaft, Terex Titan) support dozens of full-size shrubs and a fewstunted epiphytic trees. Organic matter also accumulates on

branches and in crotches not associated with ferns. Thus, ourresults greatly underestimate these trees total masses of epi-phytic material. Even so, they suggest that epiphyte masses onlarge trees in redwood forests rival those on trees in any otherforest previously studied.

Coastal forests containing trees almost as large as those in-cluded in this study no longer exist outside Redwood Nationaland State Parks, but historical records indicate that British Co-lumbia, Washington, and Oregon once had such forests (e.g.,see data in Carder [1995]). The few remaining old-growth red-wood forests provide a rare glimpse of conditions that wereprobably common in the vast low elevation temperate rainforests that once extended from Alaska to California, namelygiant trees supporting dozens to hundreds of kilograms of epi-

phytic material. Mature second-growth redwood forests arealso rare, and, aside from scattered remnant trees, the few thatremain do not contain large trees with complex crowns. Con-sequently, vascular epiphytes, including P. scouleri, are nowsparse or absent in all redwood forests except old-growth for-ests.

New silvicultural practices in the redwood region raise thepossibility of accelerating the development of old-growth char-acteristics in managed forests (Thornburgh et al., 1999). Ma-nipulations of canopy structure and microclimate are possiblevia thinning, green tree retention, and controlled burning, butdevelopment of forests containing redwoods with hugebranches and highly reiterated crowns will require several cen-turies or more. The development of large P. scouleri mats isalso likely to require centuries. Thus, the diversity of desic-

cation-sensitive species inhabiting old-growth redwood rainforest canopies may not be able to colonize managed forestcanopies into the foreseeable future. Conservation of old-growth redwood rain forests is therefore critical not only fortheir magnificent trees but also for their rich communities ofepiphytes and associated organisms thriving high above theground.

LITERATURE CITED

AMBROSE, A. R. 2003. Effects of arboreal humus accumulations on canopymicroclimates in an old-growth redwood forest. Masters thesis, Hum-boldt State University, Arcata, California, USA.

ANDRADE, J. L., AND P. S. NOBEL. 1997. Microhabitats and water relationsof epiphytic cacti and ferns in a lowland neotropical forest. Biotropica29: 261270.

BENZING, D. H. 1990. Vascular epiphytes: general biology and related biota.Cambridge University Press, New York, New York, USA.

BENZING, D. H. 1995. Vascular epiphytes. In M. D. Lowman and N. M.Nadkarni [eds.], Forest canopies, 225254. Academic Press, New York,

New York, USA.

BOHLMAN, S. A., T. J. MATELSON, AND N. M. NADKARNI. 1995. Moistureand temperature patterns of canopy humus and forest floor soil of a

montane cloud forest, Costa Rica. Biotropica 27: 1319.

CARDER, A. C. 1995. Forest giants of the world, past and present. Fitzhenry

and Whiteside, Markham, Ontario, Canada.CLEMENT, J. P., AND D. C. SHAW. 1999. Crown structure and the distribution

of epiphyte functional group biomass in old-growth Pseudotsuga men-

ziesii trees. Ecoscience 6: 243254.

DAWSON, T. E. 1998. Fog in California redwood forests: ecosystem inputs

and use by plants. Oecologia 117: 476485.ELLYSON, W. J. T. AND S. C. SILLETT. 2003. Epiphyte communities on Sitka

spruce in an old-growth redwood forest. Bryologist, in press.

FREIBERG, M. 1996. Spatial distribution of vascular epiphytes on three emer-

gent canopy trees in French Guiana. Biotropica 28: 345355.

HELBSING, S., M. REIDERER, AND G. ZOTZ. 2000. Cuticles of vascular epi-phytes: efficient barriers for water loss after stomatal closure? Annals of

Botany 86: 765769.

HIETZ, P., AND O. BRIONES. 1998. Correlation between water relations andwithin-canopy distribution of epiphytic ferns in a Mexican cloud forest.

Oecologia 114: 305316.

HIETZ, P., AND U. HIETZ-SEIFERT. 1995. Structure and ecology of epiphyte

communities of a cloud forest in central Veracruz, Mexico. Journal of

Vegetation Science 6: 719728.

HOFSTEDE, R. G. M., J. H. D. WOLF, AND D. H. BENZING. 1993. Epiphyticbiomass and nutrient status of a Colombian upper montane rain forest.

Selbyana 14: 3745.

INGRAM, S. W., AND N. M. NADKARNI. 1993. Composition and distribution

of epiphytic organic matter in a neotropical cloud forest, Costa Rica.

Biotropica 25: 370383.JEPSON, J. 2000. The tree climbers companion. Beaver Tree Publishing,

Longville, Minnesota, USA.

LYONS, B., N. M. NADKARNI, AND M. P. NORTH. 2000. Spatial distribution

and succession of epiphytes on Tsuga heterophylla in an old-growth

Douglas-fir forest. Canadian Journal of Botany 78: 957968.MCCUNE, B., AND M. J. MEFFORD. 1999. PC-ORD. Multivariate analysis of

ecological data, version 4. MjM Software Design, Gleneden Beach,

Oregon, USA.

NADKARNI, N. M. 1981. Canopy roots: convergent evolution in rainforest

nutrient cycles. Science 214: 10231024.NADKARNI, N. M. 1984. Epiphyte biomass and nutrient capital of a neotrop-

ical elfin forest. Biotropica 16: 249256.

NADKARNI, N. M. 1994. Factors affecting the initiation and growth of above-

ground adventitious roots in a tropical cloud forest tree: an experimental

approach. Oecologia 100: 9497.PARKER, G. G. 1995. Structure and microclimate of forest canopies. In M.

D. Lowman and N. M. Nadkarni [eds.], Forest canopies, 73106. Aca-

demic Press, New York, New York, USA.

PIKE, L. H., W. C. DENISON, D. M. TRACY, M. A. SHERWOOD, AND F. M.

RHOADES. 1975. Floristic survey of epiphytic lichens and bryophytesgrowing on old-growth conifers in western Oregon. Bryologist78: 389

402.SAS. 2001. JMP 4.0. SAS Institute, Cary, North Carolina, USA.

SAWYER, J. O., S. C. SILLETT, W. J. LIBBY, T. E. DAWSON, J. H. POPENOE,

D. L. LARGENT, R. VAN PELT, S. D. VEIRS, R. F. NOSS, D. A. THORN-BURGH, AND P. DEL TREDICI. 1999. Redwood trees, communities, and

ecosystems: a closer look. In R. F. Noss [ed.], The redwood forest: his-

tory, ecology, and conservation of the coast redwoods, 81118. Island

Press, Covelo, California, USA.

SILLETT, S. C. 1995. Branch epiphyte assemblages in the forest interior andon the clearcut edge of a 700-year-old Douglas-fir canopy in western

Oregon. Bryologist 98: 301312.

SILLETT, S. C. 1999. Tree crown structure and vascular epiphyte distribution

in Sequoia sempervirens rain forest canopies. Selbyana 20: 7697.

SILLETT, S. C., AND P. N. NEITLICH. 1996. Emerging themes in epiphyte

7/30/2019 255.full

7/7


research in westside forests with special reference to the cyanolichens.

Northwest Science 70, Special Issue: 5460.

SILLETT, S. C., AND R. VAN PELT. 2000. A redwood tree whose crown is a

forest canopy. Northwest Science 74: 3443.

SILLETT, S. C., AND R. VAN PELT. 2001. A redwood tree whose crown may

be the most complex on Earth. In M. Labrecque [ed.], LArbre 2000, 11

18. Isabelle Quentin, Montreal, Quebec, Canada.

TAYLOR, P. L. 2001. Sitting on top of the world. In Science at the extreme,

92117. McGraw-Hill, New York, New York, USA.TER STEEGE, H., AND J. H. C. CORNELISSEN. 1989. Distribution and ecology

of vascular epiphytes in lowland rain forest of Guyana. Biotropica 21:

331339.

THORNBURGH, D. A., R. F. NOSS, D. P. ANGELIDES, C. M. OLSON, F. EU-

PHRAT, AND H. H. WELSH JR. 1999. Managing redwoods. In R. F. Nos[ed.], The redwood forest: history, ecology, and conservation of the coastredwoods, 229261. Island Press, Covelo, California, USA.

VAN LEERDAM, A., R. J. ZAGT, AND E. J. VENEKLAAS. 1990. The distributionof epiphyte growth-forms in the canopy of a Colombian cloud-forestVegetatio 87: 5971.

VAN PELT, R. 2001. Forest giants of the Pacific coast. University of Washington and Global Forest Press, Seattle, Washington, USA.

VANCE, E. D., AND N. M. NADKARNI. 1992. Root biomass distribution in a

moist tropical montane forest. Plant and Soil 142: 3139.WHITMORE, S. 1993. Polypodium. In J. C. Hickman [ed.], The Jepson manual

100101. University of California Press, Berkeley, California, USA.WOLF, J. H. D. 1994. Factors controlling the distribution of vascular and

non-vascular epiphytes in the northern Andes. Vegetatio 112: 1528.

255.full

Documents

Transcript of 255.full