DIVERSITY AND ABUNDANCE OF ORCHIDS IN A PERUVIAN CLOUD FOREST

2010·6.9 MB·English

by  Rebecca Repasky Luke

AND ABUNDANCE OF ORCHIDS A PERUVIAN CLOUD FOREST
IN
DIVERSITY
Luke John Janovec Christenson
Repasky 1 Eric
Rebecca
P.
mica! Research Institute of Text
m
500
East 4th
Christian University St.
USA
TCU 298930 Worth, Texas 76102-4060,
Box
USA
[email protected]
Texas 76129,
North,
McNew
Barfield
Keri
Department ofBiology Botanical Research Institute ofTexas
Texas Christian University
USA
TCU Box 298930 Fort Worth, Texas 76102-4060,
USA
Fort Worth, Texas 76129,
ABSTRACT
RESUMEN
*
Gravendeel
Gentry 1988;
(Stebbins 1981;
plants
Orchidaccae flowering
is one of the largest families of
The
ah 1996).
» Peru alone (Ibhsh et
in
occurring
2004) with species
mostly epiphytic
at least 1,000
Andean cloud forests has
gh orchids in
Peruv epiphytic
ian that of
species diversity of orchids, especially
J‘
Peruvia
from the
orchid
n
Journal of the Botanical Research
318 Institute of Texas
4(1)
Hiram Bingham Machu
was characterized in 1760 (Rolfe 1916). In 1911, described the Picchu
first site and
commented on the orchids occurring there (Christenson 2003). This high orchid diversity occurs
due
to
a mixture of conditions including a cool, moist climate favorable to epiphytes, (2) the diversification
(1) of
isolated populations on separate mountain ranges (Tremblay et al. 2005), and (3) the ability of orchid specks
hybridize (Tremblay 2005).
to easily et
al.
Although this diversity has been recognized, there is limited quantitative information on the species
richness and spatial distribution of orchids across sites and along elevation gradients in the Peruvian Andes.
known how
More about these factors affect epiphytes in general than orchids in particular. These montane
is
&
forests harbor a high diversity and abundance of epiphytic plant species (Grubb et al. 1963; Comelissen
maximum
Ter Steege 1989) with species richness and endemism occurring in wet aseasonal forests on
fertile
&
soils at elevations of 1500 to 2000 masl (Gentry Dodson 1987; Iblish et al. 1996; Kessler, 2002; Kuperet
&
m
2004; van der Werff Cosiglio 2004; Kromer 2005). Species richness declines above 2000 aslbut
al. et al.
whether this decline indicates some climatic optimum or merely reflects the decline in available land mass
above this elevation remains unclear (Kromer et 2005). Orchid diversity generally follows the trends for
al.
may among
all epiphytes (Kessler 2002; Kuper et al. 2004; Kromer et al. 2005), but the pattern vary orchid
genera (Kessler 2002).
two
Detailed studies of orchid diversity and distribution at cloud forest elevations are limited to studies
One Machu
in southeastern Peru. was a 12-month study in a 143.5-hectare section on the Picchu Historical
Sanctuary (MPHS) near the Winay-Wayna structures (WW; Zambrano et 2003a). The other was a 12-month
al.
Manu
inventory of 16 100-m2 plots in the National Park (MNP; Zambrano et 2003b). Both studies report
al.
numerous
more than 170 orchid species, with limited overlap with the extensive species compiled by
list
MPHS
composed
investigations at the (Christenson 2003). These two studies suggest a diverse orchid flora
numerous may document high
of species that have limited spatial distributions. Although these studies
species diversity at each site and high site-to-site variability in species composition, neither study assesses
compositions
the relative abundances of the various orchid species. Comparison of the variability in species
among resolve
to
these studies is also complicated by potential problems in nomenclature that are difficult
because of the lack of access to reference specimens, imagery, or detailed collection data.
digital
The existed fat
present extent of our understanding of Andean orchid diversity similar to those that
is
on
western Amazonian abundances. Based early
tree floras before the advent of quantitative studies of species
compilations mosaics of relatively
of species lists at different sites, the forests could be depicted as small-scale
recommendations
unpredictable species composition. As were performed following the
quantitative studies
commo
of Phillips and Raven (1996) and others, patterns of structure involving some widely dispersed, y
&
2005)-
occurring, and abundant Macia Svenning
relatively taxa could be recognized (Pitman 2001;
et al.
comm
Although some studies have not provided strong support the existence of such widespread
for
Tuomist0
formation
ity, of a species community, of abundant 2003;
tree species Phillips et al.
(e.g.,
o
2003; Vormisto et al. 2004), they have demonstrated that testing for the presence, or the absence,
communality
abundance
requires as well as species composition be documented.
This study had two main orchid diversity
goals. First, to initiate quantitative assessments of
com ar*
abundance in cloud forests similar to those employed Amazonian and second to P
for tree flora,
poss
determine
results to other detailed studies of cloud forest orchid diversity in southeastern Peru to
spec®
orchid
communalities in abundant The determine
taxa. specific objectives of the study were
to: (1)
o
abundant
richness; (2) estimate the relative abundances of species; and question whether the
(3)
species at this site were also abundant, or at least present, other research sites in the region,
at
and heterogeneity
abundance.
sity If elevations and habitats have large they could contribute to
effects,
species compositions and
abundances.
and abundance of Orchids
Diversity
Lukeetal.,
METHODS
The design of this study emulates the general protocols recommended by Phillips and Raven (1996) for
assessing tree diversity in the Neotropics. These protocols share procedures designed to quantify diversity
on a local scale by obtaining: (1) a more complete inventory of plant species composition; an assess-
(2)
and
ment of the spatial distribution of species; basic phenological information. These objectives are
(3)
accomplished by conducting surveys to document the species composition and by monitoring of perma-
nent plots for species abundances. For orchids, these objectives were addressed between September, 2005
and August, 2006 through monthly, systematic collections of the orchid species present and the monthly
monitoring of permanent plots to measure diversity, distribution, and abundance of the species present.
new
Because the study involves the application of protocols to quantify orchid abundance and diversity,
an evaluation of these protocols with recommendations for potential modifications are also discussed.
—
Study Research was conducted 560-ha Wayqecha Cloud Research (WCFR;
site. at the Forest Station
km
13°10'40"S, 71°36'20"W) located 60 northeast of Cusco in southeastern Peru (Fig. The station borders
1).
Manu National Park and owned and operated by the Amazon Conservation Association (ACA) of Wash-
is
DC
ington, and its sister organization in Peru, the Asociacibn para la Conservacibn de la Cuenca Amazbnica
WCFR
(ACCA). Elevations at range from 2200 to 3200 m, and the primary natural vegetation is upper
&
montane forest (Young Lebn 2000). The forest continually saturated with rain and fog. Temperatures
is
mm
1PC
< months and
average with seasonal variation. Precipitation ranges from 10 in the of June
little
mm
July to > 100 in the months of January, February, and March.
WCFR,
Systematic plant collection.—To document the orchid species occurring at approximately
10 days of each month were devoted photographing, and preserving specimens of flowering
to collecting,
orchids. Every orchid with a unique vegetative or reproductive morphology was considered to be a poten-
new
tial morphospecies unless could be positively identified as belonging to a previously collected type.
it
number
Although minimized
this practice resulted in replicate collections of some orchid species, it the of
missed
species.
One more new morphospecies. At one specimen
or specimens were each potentially least
collected for
was (USM)
Lima,
^posited in the herbarium Museo Nacional Mayor de Historia Natural in Peru.
of the
Another specimen was deposited in the herbarium of the Botanical Research Institute of Texas (BRIT) in
Fort Wonh, Flowers
Texas, and additional specimens were dispersed to orchid experts at other herbaria.
were
also collected and 80% and 20% glycerine and deposited at BRIT.
preserved in alcohol
Each number, and pertinent data were re-
potential morphospecies was assigned an identification
corded Garmin Map76C positioning
including: determined by a global
(1) date of collection; (2) location as
Sysiem
5 being the color
unit (GPS, nominal accuracy < 15 m); relative flower size (1 to 5, largest); (4)
(3)
°f the flower or epiphytic). The habitat of the
(particularly and column and habit (terrestrial
lip colors); (5)
individual was with heights > 15 m; short cloud forest with
also cloud tree (2)
classified forest
as: (1) tall
tree hei8hts < 15 m; and grassy areas with no trees or few trees with heights < 3 m. Grassy areas likely
(3)
&
Lebn
(Young
from over the past 60 years
or
disturbances, such road construction, fire
as logging,
°°
0; Lozan *8 mega-pixel digital photographs were taken for
<> et al. 2006). To aid in species identification,
^
Atrium
through the
Potential photographs are accessible
morphospecies. Collection data and digital
iversity
Information System BRIT (http://atrium.andesamazon.org).
at
u
compared
pon morphospecies were in
completion specimens of potential
of the
field collections,
*
^barium with another collected
be conspecific
and judged be unique, valid taxon, or to
to either a
^ospecies. specimens were combined
For those morphospecies judged to be conspecific, the collected
“
t0
confirm
a single were reviewed by taxonomic experts to
morphospecies. morphospecies collections
All
^morphospecies and description.
species identification
and process of
status begin the
to
on
and differences orchid
Quantitative of elevation habitat
data collection.—To analyze the effects
,
m
rs
‘iy, were established across the elevation
'^ abundance, and permanent, 5 x 5 plots
distribution, 47
Wkn
t wi[hin WCFR. and dense vegetation, the plots were established at ran-
Because gradients
of steep
BRAZIL
dom habitats
locations along and grassland
pre-existing passing through
trails the characteristic forest
were classified
The
the area. latitude, longitude, and elevation of each plot were measured by GPS. Habitats
as cloud
tall forest, short cloud forest, or grassy areas, as described above.
^
,
^
From and the
September 2005 9th
through August 2006 the plots were monitored between the
1
"
ote
each month ofa P
for the occurrence occurrence
of potential morphospecies in flower. For the first
“
c0
morphospecies general
in a plot, the species was catalogued using number from the
identification
its
occu
and number was
the orchid
of individuals with number and whether the
flowers, total of flowers,
^
encounte
as a was
terrestrial or epiphytic form was recorded. a potential new morphospecies first
If
edt
the was col
plots, added procedures to
it to the general collection using the previously described
phoiograp^
able specimens
located outside of the plot. If the orchid only occurred in the plot, only digital
es
were morphospec‘
taken
to assist in species identification. For subsequent monthly observations of a
number Genera
plot, the
of individuals with and number were recorded.
flowers the total of flowers
.
of elevation and habitat type on the species richness and density of orchids were
the effects
of
model tests
&
Cody
using the Statistical Analysis System (SAS; Smith 2006).
performed
and
and conditions prevented each individual plant from being tagged,
resources field it is
Limited
individuals counted as flowering in one month could also be counted as flowering in subse-
that
possible
minimum
months. For this reason, the term density as used in subsequent discussions is a estimate
quent
maximum
number morphospecies observed flowering in that
taken as the of individuals of that
density
of
month.
any one
in
plot
—
how
Rarefaction analysis. To evaluate well the plot sampling could estimate the species richness
curves Colwell 2001) were computed using
by systematic collections, rarefaction (Cotelli
determined
Chao (Chao and Chao (Chao 1984, 1987). These
non-parametric species richness estimators 1984) 2
1
the
number
use the occurrence of rarer species in the samples to adjust the observed of species for
procedures
The Chao procedure defines rareness as
number of species that were likely missed in the sampling. 1
the
The Chao procedure defines rareness as being observed in
being represented by only a few individuals. 2
Chaol and Chao and 95% Cl were computed from 10,000 randomiza-
few 2 estimators their
only a plots.
were
Other richness estimators
sequences using the software Estimates 8.0 (Colwell 2006).
tions of plot
RESULTS
239 unique morphospecies
?m morphospecies collected,
Systematic collections.— the 341 potential
were from gen Over 134 morphospecies have been identified as previously described spe-
identified 49
.
&
Mera The
Galan de 2008).
e (Chri
lining 102 morphospecies r identified as previ islynat
:
morphospecies
new unidentified taxa reported here are
to science. It is important to note that the as yet
However,
based on which may not be supported by subsequent genetic analysis.
structural characteristics
orchids into separate
may apparently similar
separate
it is also possible that subsequent genetic analyses
<2
32 other genera.
species in
species. The number of morphospecies per genus ranged from 41 for Stelis to
were
and
Pleurothallis
The number common genera Epidendrum, Maxillaria,
of morphospecies in the other
Appendix
SI
A by genus presented in
three, six and organized is
nine, respectively. species list
most species flowered
orchid diversity as
Monthly assessing
collections were important for adequately
fors3 phenology data from the plot stud-
by flowering
months. Monthly the
importance best illustrated
is
be
observed
among 65 species to
ies where only September 2005 were the
six of the 17 species flowering in
were among the 23 species
February
flowering flowering in
in February, 2006. Only six of the 65 species
^g
Do
in August,
2°°6.
d
k equal
^ -
t
All three habitats
National each interval (Fig. 2).
intervals of 175 m, there were similar numbers of plots in
Were
Present
in each
interval
he
recorded in of plots
Orchid minimum number of individuals all
«
densities and distributions.-The
month
any
maximum number recorded in
sum
the
>n of the study was 2565, which the of
is
with elevation
0r showed variation
a species in each Minimum numbers individuals per plot little
of
plot.
SD) densittes^ for
Mean Standard Delation;
habitat type = Appendix (±
(F = P > S2).
0.33; df 0.10;
41;
5,
m
per 2 per 100
0.38 individuals
of
a rate
|*7’ Actively. with increasing elevations at
Densities declined
normally
®°f were approximately
Densities
from
elevation, 0.
but was different
this rate not significantly
Abated
Conover 1971).
(Kohnogorov-Smirnov normality;
of
test
were observed in plots
239 species
distribution among plois.-Only 128 of .he
of orchid
species
**"* (lowered durtng the
“* observed in plots
individuals
12-month orchid
study. Because almost all
heterogeneity of the orchtd
£*• ,h“ sampling of the spatial
nnderrepresentation was due inadequate
to
with median of nine species per
a
J'W' »»mber from three to twenty-five
of species per plot ranged
Mosl
of epiphytes.
orchid occurred as
species (72%)
pb*
per
For were number species
terrestrial species, there significant differences in the of terrestrial
the
among P <
habitats (F = df = P < and = df = 43; 0.01).
6.15; 2, 43; 0.01) at different elevations (F 14.09; 1,
mean number was significantly
(± SD) of terrestrial species per plot for tall cloud forest was 1.8 (± 2.5). This
pet
(P< species
0.05) less than the means of 4.0 (± 2.7) species per plot in short cloud forest and 3.4 (± 2.6)
m
increa*
100
plot in grassland. There was = each
a decrease of 0.81 (SE 0.22) terrestrial species per plot for
secundum
in elevation. The two Epidendrum
terrestrial orchids with the highest frequency of occurrence,
Elleanthus occurred
sp., in 17 of the 47
plots.
among
For epiphytic species, there were number of species per plot
also significant differences in the
The tin®'
habitats (F = 7.29; df = 2, 43; P < 0.01) and at different elevations (F = 19.12; df = 1, 43; P < 0.01).
for
ber of epiphytic species per plot for short cloud forest (10.1 ± 4.5) was significantly greater than that
m
there
cloud ± and elevation,
tall forest (6.3 3.7) the grass areas ± For each 100 increase in
(6.2 4.0).
frequency
a decrease of 1.4 (SE = 0.3) epiphytic species per plot. The epiphytic orchid with the highest
occurrence was
Pleurothallis acuminata which occurred in 20
plots.
T
ere
There was
among occurrence.
considerable variation species in abundances and frequencies of
majority
The
27 species with an abundance of one individual and 12 with an abundance of two.
species
m
m These
orchid had mean 2 ^
species > per
densities < 2 per 2 Only nine species had mean densities 2 .
.
brenrac
listed in order of decreasing abundance, Stehs ^
Pachyphyllum
vestigipetala,
sp.l, Pleurothallis aff.
acummuW^
Elleanthus
sp.
1, Stelis grandibracteata, Maxillaria Epidenrum secundium, Pleurothallis
alpestris,
species
Pachyphyllum
pectinatum. Most (77%) species were observed < Forty-nine
in 5 plots (Fig. 3).
m
>
found in only one and observed
plot, 22 species occurred in only two Only species were
plots. six
and abundance of Orchids 323
Diversity
Lukeetal.,
NUMBER OF PLOTS
were Pachyphyllum Pleu-
°f the 10 plots. These, in order of decreasing frequencies of occurrence, sp. 1
listed ,
and Pachyphyllum pectinatum.
r
°thdlis acuminata, Epidendrum secundum, Elleanthus sp. 1,
breviracema,
Stelis
underestimated the
Chao procedures consistently
Rarefaction Analysis.— Chao and
Both 2
the
1
which
^cies The Chao procedure, defines
2
richness of 239 observed in the systematic collections (Fig. 4).
which
Chao procedure,
r»eness than the
based on was consistently greater 1
incidences of occurrence,
Chao procedure, based
defines The estimate for the 2
rareness based on the abundance of individuals. final
240
00 (95% Cl) of 155 to
95% Confidence Interval
randomizations with
of 47 plots was 183 species a
all
Species The was 158 species with a 95% Cl of 140 to 206 species.
Chao procedure
- final estimate for the
1
13%
ranged from in
were flower
Flowering in
that
Phenology.-The percent of species in the plots
among types or
Member habitat
percents
in
51% was differences
to in February and there little
(Fig.
5), m
2
minimum per
SCross ranged from a of 0.65 (± 1.1)
Ovations. The mean SD) density of flowers also
(±
months
m occurred during
September Peak flowering
maximum 2 February.
to a per in
of 25.7 (± 37.6)
mm
The me-
mm
lh > months with < 5 of rainfall.
100
minimum occurred in
of and flowering
rainfall,
The
months.
aian than six
lowering flowered for less
duration was months and most (80%) species
three
no
and were
there
j*°P°mon months was S 0.40 (Fig.
6),
of flowering between successive
species shared
synchronous
jounced exhibiting
of species
major subsets
peaks indicated
in the proportions shared that
Journal of the Botanical Research
Institute ofTexas
4(1)
DISCUSSION
WCFR
<*
The morphospecies,
cloud forests at harbor a rich diversity of orchids with 130 species and 109
wp
*
239 orchids
unique taxa, collected and documented during the 12-month sampling period. Most of the
occunw
Over 75% them or only
relatively rare. of were either found only during the systematic collections
Only nffl*
in s 2 of the 47 plots. Those that did occur in plots were often represented by < 3 individuals.
of
one-third
of the species could be described as being abundant due occurrence in more than
to their
m
Epidendnu*
mean
plots or their densities being > 2 individuals per 2 These included the identified species
.
secundum, vesti^*T
Maxillaria alpestris aff., Pachyphyllum pectinatum, Pleurothallis acuminata, Pleurothallis
Pachyphy™*
and
aff., Stelis breviracema, and Stelis grandibracteata and the morphospecies Elleanthus sp. 1
m
and The Pachyphylh‘
sp. 1. Elleanthus sp. 1 Epidendrum secundum were primarily species.
terrestrial
with
si
Pleurothallis species were primarily epiphytes. The remaining three abundant species occurred
^
abundances in both and
terrestrial epiphytic growth forms.
have
Andes
Similar intensive, 12-month assessments of orchid species richness in the Peruvian
k®
100
made near the Winay-Wayna (WW; Zambrano MPHS which about
ruins 2003a) is
et at
al.
WCFR
northe
of and in a small portion of Manu National Park (MNP; Zambrano 2003b) to the
et al.
WCFR. compiled a
In addition has
to these intensive analyses of smaller Christenson (2003)
sites,
MPHS Comparing the
list for the 35,952-ha based on collections since that discovery in 1911.
site’s
WW
nclude
WCFR MNP
compositions of these and and studies i
sites complicated because: the
is .
(1)
undesc
chid from or
species elevations above and below number unidentified
the cloud a of
forest;
(2)
morphospecies and classifications
still exist; (3) variation in the use of nomenclature and taxonomic
and
study and richness
sites researchers. Despite these complications, three general trends in species
^
position are apparent.
species-
the orchid
First, intensive studies of small areas have found comparably numbers of
large
and abundanc
Diversity
Luke et al„
60
FMAMJJA
SONDJ
MONTHS
WCFR
(Christenson 2003). The number of species at each of these <
been
reported for of MPHS. As many 50 species of epiphytic o
all as
forest elevations (Kromer
2005).
et
al.
Epidendrum
genera
belong to the large
of the species o ied sites
it
WW
MNP most numerous
Epidendrum the most numerous genus the and sites, but Stelis is the
is at
W
•T at CFR- WCFR, but less diverse than the
Epidendrum the second most diverse genus at Stelis is
is
WW
°
ther MNP
three
genera
at the and
sites.
Third, although the important genera appear to be similar across sites, the species representing these
WW
®tnera differ among No but Zambrano et al. (2003b) listed the
sites. species provided for the site,
list is
63
remaining 149
species MNP known MPHS. Apparently, the
that occurred at which are also to occur at
MNP
Sp MNP common and MPHS,
reies
at MPHS Of 63 species to
could not be matched to species. those
MNP MPHS members
28 common and were of
r* occurred WCFR.
at Thirty-six of the 63 species to
Vwndrum, WCFR.
occurred
Maxillaha, and but only 17 of these at
Pleurothallis Stelis,
^are
WCFRand MPHS. Of the 79 genera reported
other between
compositions
differences in species
r
WCFR
both represented
and MPHS both and 25 of these are
(Christenson 2003) only 43 occur at sites,
CFR WCFR,
were
hy numerous there
only by species at
a single species. For those genera represented
MPHS.
Wdendrum from
not reported
species that are
species, 16 Maxillaria species, and six Pleurothallis
made on
accurate
"t®***1 species WCFR MPHS may be discovered as progress is
that occur but not
at at
WCFR
Rations number
a of species
In addition,
of the numerous morphospecies in the difficult Stelis.
&
Repasky 2008).
(Christenson
additions undescribed species
to the flora of Peru or previously
Journal of the Botanical Research
326 Institute
ofTexas4(1)
SPECIES
SHARED
OF
PROPORTION
MONTHS
Fis. 6. The proportions of species flowering in one month that were also flowering in the following month.
^
earl
MPHS the
At present, the species lists from and the sites of intense study are analogous to
stru
unpredictable
species lists for the Amazonian rainforests. They suggest a small-scaled and as yet
^
commonality
of the orchid community. The only quantitative data on abundance which might imply
amony
WCFR, communality
structure are from and two do simple
aspects of these data not suggest a
sites. First, only four of the seven named abundant WCFR, Epidendrum secundum,
species
at
c0
are
that
alpestris, Pachyphyllum pectinatum, and acuminata, were among the 28 species
Pleurothallis
spec*’
WCFR, MNP abundant
to and MPHS. Second, Pachyphyllum sp. and Elleanthus sp. two of the
1 1,
WCFR, other s
were abundant at
at not
readily-identifiable or familiar species. they were similarly
If
may specie*
be expected that they would be among commonly-known and readily-recognized
the
u
a
the
This lack of correspondence abundant among variation in " ^
of species implies local
sites
a
a
kilometers P
orchid
species. Large only
variation in orchid composition between
species sites
presen
been reported by Kuper et al. (2004) where 70 % of the species occurring at one site are not