CHANGES IN FOREST DYNAMICS ALONG A 2.5 KM ELEVATION GRADIENT ON THE SOUTHEASTERN FLANK OF THE PERUVIAN ANDES

BY

WILLIAM FARFAN RIOS

A Thesis Submitted to the Graduate Faculty of

WAKE FOREST UNIVERSITY GRADUATE SCHOOL OF ARTS AND SCIENCES

in Partial Fulfillment of the Requirements

for the Degree of

MASTER OF SCIENCE

Biology

May 2011

Winston-Salem, North Carolina

Approved By

Miles R. Silman, Ph.D., Advisor

Kathleen A. Kron, Ph.D., Chair

William K. Smith, Ph.D.

ACKNOWLEDGEMENTS

I would like to sincerely thank my advisor, Miles R. Silman, for introducing me to

tropical community ecology, giving me the opportunity to work and travel to remote

areas of Peru, for his patience, support, and for being a mentor and friend to me over the

past seven years. I thank my committee members, Dr. Kathleen A. Kron and Dr. William

K. Smith, for their time, helpful ideas and comments in this research.

Special thanks to Norma Salinas, my undergraduate professor for introducing me to

Botany in the Peruvian Andes, and also my special gratitude to Karina Garcia for her

leadership and support in the plots establishments and data processing, and also for her

friendship and companionship.

I thank the team that accompanied me during my fieldwork, foremost to Natividad

Raurau, Alberto Gibaja, Marlene Mamani, Vicky Huaman and Tatiana Boza, not only for

their help during the data collection process but also for their friendship and support.

Without the Peruvian students I would not have been able to complete this research, my

special thanks to Darcy Galiano, Walter Huaraca, Javier Silva, Angela R. Brian V., Israel

Cuba, Jhoel Delgado, Percy Chambi, Erickson Urquiaga, Alex Nina, Jhoel Salazar, Flor

Zamora, Kilmenia Luna, Karina Luna, Luis Mansilla, Janet Mamani, Jesus Castaneda,

Rosalvina B. and Yolvi V., for being with me in the field collecting data in the first and

second plot census.

ii

I thank my veteran friend Luis Imunda G., the master of the “machete” for his

inexhaustible strength in our campaigns from the Andes to the Amazon. I also thank Josh

Rapp for all of the help with data analysis in “R”, and to Kenneth Feeley, Noah Yavit,

Rachel Hillyer, Becky Dickson and Sarah Maveety for their help and support.

Finally, I would like to thank my family for their unconditional support, especially to

Julio F., Visitacion R., Maikol F. and Henry F.

This study was funded by Andes Biodiversity and Ecosystem Research Group (ABERG),

Wake Forest University, the National Science Foundation, the Gordon and Betty Moore

Foundation and the Blue Moon Foundation. I thank the Servicio Nacional de Areas

Naturales Protegidas por el Estado (SERNANP), Amazon Conservation Association,

Cock-of-the-Rock Lodge and Manu National Park for the opportunity to perform this

research.

iii

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ……………………………………………………… ii

LIST OF TABLES ……………………………………………………………….. vi

LIST OF FIGURES ……………………………………………………………… vii

ABSTRACT ……………………………………………………………………… viii

INTRODUCTION ……………………………………………………………….. 1

Forest dynamics in the lowland tropics …………………………………. 1

Vital Rates in Forest Tree Communities …………………………………. 2

Dynamics in Montane Forests …………………………………………… 4

METHODS ………………………………………………………………………. 7

Study Site ………………………………………………………….……… 7

The Plots …………………………………………………………………. 7

Environmental Characteristics …………………………………….…….. 8

Calculation of Stand-level Rates ………………………………………… 8

Mode of Death …………………………………………………………… 10

Wood Density data ……………………………………………………….. 10

RESULTS ………………………………………………………………………... 12

Montane Forest Structure and Elevation ………………………………… 12

Montane Forest Dynamics ………………….……………………………. 12

Mode of Death in Montane Forest ……………………………………….. 13

Stands Dynamics and Wood Density …………………………………….. 14

iv

DISCUSSION ……………………………………………………………………. 16

Trends in Montane Forest Structure ……………………………………... 16

Stand Forest Dynamics …………………………………………………... 17

How do stems die? ……………………………………………………….. 19

Montane forest dynamics as a function of wood density ………………… 20

CONCLUSIONS …………………………………………………………….…… 22

LITERATURE CITED ….……………………………………………………….. 23

TABLE. …………………………………………………………………….….…. 33

FIGURES ………………………………………………………………………… 34

CURRICULUM VITAE …………………………………………………….…… 53

v

LIST OF TABLES

Table 1. Description of the one hectare permanent plots included in this research. 33

vi

LIST OF FIGURES

Figure 1. Study site ………………………………………………………………. 38

Figure 2. Stem density and multi-stem production across elevation …………….. 39

Figure 3. Relationship of stem basal area and elevation …………………………. 40

Figure 4. Boxplot of mortality and recruitment rates …………………………….. 41

Figure 5. Frequency distribution of death stems across elevation ……………….. 42

Figure 6. Mortality and recruitment rates across elevation .…………………..….. 43

Figure 7. Turnover rates across elevation ………………………………………... 44

Figure 8. Population change rates across elevation ……………………………… 45

Figure 9. Histogram of population change for the entire community ……….…… 46

Figure 10. Mode of death for all stems ………………………………………..…. 47

Figure 11. Mode of death across elevation ………………………………………. 48

Figure 12. Mean species wood density across elevation ………………………… 49

Figure 13. Wood density weighted by species and stems across elevation ……… 50

Figure 14. Mortality, recruitment and turnover vs. wood density ……………….. 51

Figure 15. Mode of death vs. wood density ……………………………………… 52

vii

ABSTRACT

Tropical montane cloud forests are poorly studied botanically, and nearly unstudied

ecologically, even for basic stand-level measurements critical for understanding questions

ranging from maintenance of diversity to ecosystem services (Budd et al. 2004). The aim

of this study was to describe changes in forest structure, dynamics and mode of tree death

across 2.5 km elevation gradient in a tropical montane cloud forest in the Andes of

southeastern Peru. In 16 1-ha permanent tree plots over a four year interval all stems (≥

10 cm DBH) were tagged, identified and measured initially, then recensused 2-4 yrs later.

Results showed that stem mortality decreased significantly with elevation while

recruitment did not, resulting in a significant decrease in stem turnover with elevation.

Across all elevations the dominant mode of death was by being uprooted. Deaths due to

stem snapping were more frequent at lower elevations while uprooting was greater at

higher elevations. Trees dying standing had a non-linear relationship with elevation. This

study also gave insight into the effects of elevation on the species trait of wood density,

and further tested the hypothesis that wood density was an important predictor of how

stems died. Results showed that species wood density increased with elevation and was

associated with decreasing stem mortality; in particular the frequency of snapped stems.

Overall stem density declined in the four year period due the high rates of stem mortality

that were not balanced by recruitment. The high dynamism in low elevations can

probably be attributed to changes in species composition and species trait spectra, as well

as soil-nutrient relationships. The patterns found in this study lay the groundwork for

mechanistic studies of factors controlling forest dynamics in tropical montane sites.

viii

ix

Keywords: Tropical montane cloud forest, Andes, elevation gradient, forest dynamics,

permanent tree plots, mortality, recruitment, turnover, wood density.

INTRODUCTION

Stand dynamics, such as stem mortality, recruitment and turnover rates, have important

implications for forest ecology. Tree diversity, composition, species coexistence

(Denslow 1987, Manokaran and Kochummen 1987, Condit 1992, Phillips et al. 1994,

Rees et al. 2001, Laurance et al. 2004), the evolution of tree life histories (Clark and

Clark 1992), the generation of habitat for other organisms, the control of terrestrial

nutrient cycling (Laurance et al. 2010) and even a major component of earth's carbon

balance depend fundamentally on tree growth, recruitment and mortality at broad

temporal and spatial scales (Phillips et al. 1998, Whittaker et al. 1998, Clark 2002, Chave

et al. 2003, Baker et al. 2004, Phillips et al. 2004, Rice et al. 2004). While studies of

forest dynamics are relatively common in lowland tropical forests, tropical montane

forests are poorly studied (Budd et al. 2004). This study attempts to addresses this gap in

knowledge by describing forest dynamics across a 2.5 km elevation gradient in the

Peruvian Andes.

Forest dynamics in the lowland tropics

Studies of tropical forest dynamics have been conducted mainly in lowland rain forest,

and show contrasting results at different geographic and spatial scales. Across the

Neotropics, forests appear to have become more dynamic, with stand-level measurements

such as mortality, recruitment, and turnover of individuals increasing over the past

several decades (Phillips and Gentry 1994, Phillips et al. 1994). Paleotropical sites,

though, do not show this pattern (Phillips et al. 1998). At the landscape scale, Amazonian

1

forests appear to be more dynamic in terms of both mortality and recruitment rates since

the 1970s (Phillips and Gentry 1994, Susan et al. 2009), while Central American and

Malaysian forests have shown periods of both increasing and decreasing rates, with

mortality and recruitment decoupled in time and both dependent on large-scale climate

forces (Condit et al. 1999).

The cause(s) of widespread acceleration in Neotropical forest dynamics are still in

question. At the regional scale, forest dynamics can change due to environmental

gradients in both space and time while temporal trends appear affected by inter-annual

climatic fluctuations, El Niño events (Condit et al. 2004), unusual droughts (Condit et al.

1995, Phillips et al. 2010) and local squalls (Negron-Juarez et al. 2010). All of these can

increase tree mortality rates and consequently shift tree species composition, especially if

the tree species recruiting differ in relative abundance from those dying, and forest

structure via changes in diameter and height distributions (Laurance et al. 2004, Phillips

et al. 2010). In the Amazon Basin, variations in forest dynamics at regional spatial scales

have been attributed to soil fertility (Condit et al. 1999, Phillips et al. 2004). For example

tree dynamism was greater in richer soils of southern and western Amazonian than in the

poor soils of eastern and central Amazonia (Phillips et al. 2004).

Vital Rates in Forest Tree Communities

Tree mortality and recruitment estimates are fundamental descriptors of Pantropical

forest tree populations (Lewis et al. 2004). Tree recruitment is a manifestation of the

fecundity of the species and the growth and survivorship of juveniles in the population

2

through different sizes class , e.g. seedling, saplings, poles and mature trees (Swaine et al.

1987), which can vary as the successional status of the forest changes. Tree mortality

rates have been correlated with maximum longevity and size class distribution

(Lieberman and Lieberman 1987, Swaine et al. 1987), but these correlations are weak,

with a large component of tree mortality due to damage from random events. Measured

rates of mortality across the tropics are 1-2% y-1 and are independent of size class for

trees (≥ 10 cm DBH) reaching the forest canopy (Lieberman and Lieberman 1987,

Swaine et al. 1987).

Tree recruitment is fundamentally linked to openings in the forest canopy (gap phase

dynamics); (Watt 1947, Denslow 1987). Because tree death creates canopy gaps, how

trees die is important. Different modes of death create a variety of tree regeneration

opportunities due to the diverse environments created in disturbances of different scales

(Gale and Barfod 1999, Arriaga 2000). Trees that die by being uprooted or snapped

cause small-to-large (0.01 to 1 ha) light gaps in the forest, with these light gaps being a

major mode of secondary succession (Denslow et al. 1998). Conversely, when trees die

standing, very little change occurs in light availability in the forest and most of the

change that does occur happens in the canopy layers of the stand rather than at ground

level. Death by uprooting is due to the interplay of disturbance agents such as wind and

rainfall with site properties such as soil anchorage and steep slopes (Gale and Barfod

1999, Arriaga 2000). Snapping involves the same physical agents as uprooting, but biotic

interactions (Gale and Barfod 1999, Arriaga 2000) and wood properties (Putz et al. 1983,

Chao et al. 2009) can be important as well, which means that there can be large

3

interspecific differences in the susceptibility and recovery from snapping. In contrast,

standing death may indicate intrinsic loss of vigor, senescence, and the effects of biotic

interactions such as pathogens, insect attack and competition with other trees or lianas

(Carey et al. 1994, Midgley et al. 1995, Phillips et al. 2005, Ingwell et al. 2010). The

dominant mode of stem death differs at regional scales. For example, snapping is more

common in the north-western Amazonia in contrast with the north-eastern Amazonia

where the standing mode is more common (Putz et al. 1983, Chao et al. 2009). Mode of

death also differs between functional groups, e.g. dicotyledonous trees predominantly die

uprooted and snapped, while arborescent palms mostly die standing and snapped (Gale

and Barfod 1999).

The size of disturbance is also important in forest dynamics, and this can vary from small

scale differences, such as whether trees die standing, in single tree-fall gaps, or are tied

together by lianas into multiple tree-fall gaps (scales of 0.01 to 0.5 ha). Larger scale

disturbances are driven by microbursts (1-10 hectares), local squalls (10-1000 ha) or

regional frontal systems and hurricanes (1000-100000 ha), and can kill much larger

numbers of trees (Negron-Juarez et al. 2010).

Dynamics in Montane Forests

Major gradients in forest diversity, species composition and ecosystem function are seen

along elevational gradients in tropical mountains. Elevational gradients are important for

the study of vegetation–environment relationships because significant environmental

variations are compressed into short distances, with rapid changes in tree communities.

4

They may also be excellent natural laboratories for understanding the effects of biotic-

abiotic interactions in plant communities and processes of ecosystem function (Kitayama

1992, Malhi et al. 2010). Patterns observed across forested tropical elevational gradients

include tree diversity decreasing with greater elevation (Gentry 1988, Kitayama 1992,

Madsen and Ollgaard 1994, Lieberman et al. 1996, Homeier et al. 2010) and turnover in

species composition decreasing at higher elevations (Lieberman et al. 1996, Vazquez and

Givnish 1998). In terms of forest structure, both increases (Heaney and Proctor 1990,

Kitayama 1992, Vazquez and Givnish 1998, Aiba and Kitayama 1999) and decreases

(Homeier et al. 2010) in basal area and stem density have been observed with elevation,

while above ground biomass and net primary productivity decline with increasing

elevation (Girardin et al. 2010, Homeier et al. 2010).

Little is known about forest dynamics in tropical montane forests across elevational

gradients. The few studies in tropical montane forest that have analyzed mortality and

recruitment rates have focused on differences between slope and ridge rather than

elevation per se. In the montane forest of Australia tree turnover is greater on upper

slopes or near the ridge-crest in comparison to downslope areas (Herwitz and Young

1994), whereas in Jamaica mortality can be higher on slopes than on ridges in areas

affected periodically by hurricanes (Bellingham and Tanner 2000).

Even though forest dynamics are important determinants of species composition, species

diversity, functional traits, and ecosystem function, they are little studied on tropical

mountains. Thus, the purpose of this study is to provide the first analysis of stands

5

dynamics in a New World montane forest. The study uses 16 1-ha permanent tree plots

along a 2.5 km elevational gradient in a primary tropical montane cloud forest (TMCF) in

Manu National Park in the southeastern Peruvian Andes. The general goal of this study

was to detect the spatial and temporal patterns of tree communities in TMCF across the

elevational gradient. The specific goals of the study were to: (1) explore the trends in

forest structure across the elevational gradient; (2) explore patterns in tree dynamism

(mortality, recruitment, turnover and population change rates) along the gradient; (3)

investigate trends of mode of death across the gradient, and (4) ask whether this

dynamism and mode of death can be explained by the wood properties in tree

communities. I hypothesized that tree dynamism will decrease with increasing elevation

and predicted that the changes would be explained by wood mechanical properties (wood

density as a proxy), I also hypothesized that the snapped and standing modes of death

would predominate at lower elevations, with uprooted trees being more common at

higher elevations. Finally, I predicted that snapped trees would be stems with low wood

density.

6

METHODS

Study Site

This study was carried out in the tropical montane cloud forest (TMCF) in Manu National

Park (MNP) on the eastern slope of the Peruvian Andes (Figure 1).

The Plots

A network of 16 1-ha permanent plots was established by the Andes Biodiversity and

Ecosystem Research Group (ABERG 2010) along a 2.5 km elevational gradient in a

primary montane forest in MNP. Thirteen plots were established along the transect from

the localities of Trocha Union through Tono in the Kosñipata Valley between 3450 -

1000 m of elevation, with plots located approximately every 250 m in elevation.

Elevational replicate plots were established at the Wayquecha (3000m) and Callanga

(1250m and 1500m) sites (Figure 1). Eleven plots were located on ridges, four were

adjacent to the ridge top, and the last plot was established the in relatively flat ground at

the foot of the mountains. The basic description of the plots is presented in Table 1. For a

detailed description of tree species composition, diversity, structure and additional

permanents plots from ABERG (2010) group see Silman (in prep).

The first census in Trocha Union and Wayquecha was in 2003, for Callanga and Tono in

2004, San Isidro in 2005 and San Pedro in 2006; after 2-4 years the second census for

each tree plot was undertaken. Plot establishment dates and census intervals are found in

Table 1. All living wood stems ≥ 10 cm in diameter at breast height (DBH) were

7

measured, identified, tagged and mapped. Life forms included true trees, tree ferns,

palms and lianas ≥ 10 cm diameter within 2 m of the ground. For individuals with

multiple stems the DBH of each stem was measured. The diameter of each stem was

measured at breast height (1.3 m above ground) unless the trunk presented irregularities

in which case it was measured at the nearest point where the stem was cylindrical and the

most uniform.

Environmental Characteristics

Most of the plots (12 plots) are on Ordivician shales and slates, with exception of the

plots in Trocha Union (2000 m, 1800 m) and Callanga (1500 m, 1250 m) which are on

Permian granite. Temperature decreases linearly with elevation with a lapse rate of 5.2

oC/km and ranges from a mean annual temperature of ~ 20 oC at 1000 m to 7.7 oC at the

top of the transect at 3500 m (Rapp 2010). Annual rainfall is high across the gradient

with an average exceeding 5000 mm at the base of the transect decreasing to ~ 2600 mm

at the highest elevation plot (3450 m), with significant inter-annual variability throughout.

Only at the highest plot does potential evapotranspiration (PET) ever exceed rainfall, and

then only occasionally in month of June. There is a distinct seasonality in rainfall, with

highest rains in January and February and the lowest in June and July. Winds have little

seasonal variation, although higher wind speeds occur at lower elevation. Strong wind

gusts were rare (2007/8), affecting the lower elevations more strongly during austral

summer (Rapp 2010).

Calculation of Stand-level Rates

8

In the final census, all the new (recruits) stems ≥ 10 cm (DBH) were measured, including

individuals with multiple stems. All were marked with spray paint at 1.3 m above ground

or at the closest cylindrical point if there were irregularities in the trunk. To evaluate

mortality, all dead and missing stems were recorded. Stems that were snapped or

uprooted but had sprouts or bark showing fresh tissue were considered alive. To calculate

mortality, recruitment, turnover and population change rates standard methodology was

used (Phillips et al. 1994, Sheil et al. 1995, Condit et al. 1999). In the following equations,

N0 is the number of initial stems, Nt number of final stems, Ns the initial stems surviving

at the final inventory, t is the time interval, and ln N is the natural logarithm of N.

Mortality (m), recruitment (r) and population change (λ) rates were calculated as:

For mortality

tNN

m slnln 0 −= (1)

For recruitment

tNN

r st lnln −= (2)

And for population size change

tNNt 0lnln −

=λ (3)

Turnover rates were calculated as the mean of the mortality (1) and recruitment (2) rates

(Phillips et al. 1994).

9

Mode of Death

Mode of stem death was recorded in the second census (2007/09) for all the ABERG

(2010) plots. Definitions for the three main modes of death are according to Gale and

Barfod (1999). Standing were those dead stems with most of the crown branches on their

standing stems or those with the debris scattered around the trunk. Snapped were stems

broken close to the ground level, characterized by a broken trunk and part of the stem

fallen with the crown and branches attached. Uprooted stems were dead trunks showing

the roots with soils attached. Stems where the mode of death could not be identified were

assigned as unclassified and included in the stems that were not found in the second

census. For this study was not classify tree death in terms of gap size (i.e. single versus

multiple tree-falls) because the goal was to know how the stems died rather than the

spatial distribution of the mode of death. Linear regression was used to determine the

relationship between elevation and the mode of death.

Wood Density data

Wood density is defined as the ratio of the wood dry mass divided by its green volume

(Chave et al. 2006). Wood samples were collected (2009/10) using an increment borer in

individuals of species in the same transect in the Kosñipata Valley (n=185), but outside

the study plots examined here. The samples were taken only from live trees and multiple

individuals from the dominant species were cored in the surroundings of each permanent

plot, but not within the plots. Additional data of wood density was used from published

compilations (Chave et al. 2006, Chave et al. 2009, Zanne et al. 2009). For the arboreal

species where wood density information was not available, trait values were estimated by

10

using the mean values at genus or family level, assuming that 74 % of the variation at the

species level is explained by genus and 34 % by family (Chave et al. 2006). Linear

regression was used to determine the relationship between elevation and wood density,

and also correlations were used to determine the association of wood density with

measures of stand dynamics and wood density with different modes of death.

11

RESULTS

Montane Forest Structure and Elevation

In the Kosñipata Valley, Callanga and Wayquecha sites, a total of 12, 612 stems (≥ 10 cm

DBH) were measured in the 2007/09 census across the elevation gradient. There was not

significant relationship between stem density and elevation (R2 = 0.03, p = 0.51). Stem

density increased from the lowest plot (1000 m) to the middle elevations, reaching a peak

between 1800 - 2500 m, and then decreasing to the highest plot at 3450 m (Figure 2a). In

addition, the percentage of multi-stemmed trees increased significantly (R2 = 0.6, p <

0.001) with increasing elevation (Figure 2b). Stem basal area varied among the plots and

was not significant related to elevation (R2 = 0.14, p = 0.15; Figure 3). Mean stem basal

area for the entire community across the gradient showed no significant increase between

censuses [32.3 m2 ha-1 ± 3.6 (95 % CI), to 32.8 m2 ha-1 ±3.6 (95 % CI) in the second

census].

Montane Forest Dynamics

Over the study period the number of stems in the community across all elevations showed

a statistically significant decline from 12,850 to 12,612 stems (including recruitment) for

all the 16 1-ha permanent tree plots with an annual decrease of 0.52 % y-1 for the entire

community (Paired t-test; p < 0.05). The distribution of mortality and recruitment rates

for all of the plots across the elevations were skewed with mortality exceeding

recruitment rates (Wilcoxon signed rank test, p < 0.001, Figure 4).

12

Stem mortality had a significant effect on stem density (Paired t-test; p < 0.001), with a

decline from 12, 850 to 11, 871 stems for the entire community. The maximum mortality

rate was 4.3 % y-1 in the plot at 1500 m, the minimum was 0.8 % y-1 in the plot at 3000 m

of elevation, and mean mortality rate was 2.4 % y-1 ± 0.5 % (95 %CI), (Figure 6a).

Annualized stem mortality rates decreased significantly with elevation (R2 = 0.4, p <

0.05), with mortality peaking between 1500 - 2250 m, and decreasing above (Figure 6a);

mortality rate was higher for smaller diameter stems (≥ 10 cm DBH, Figure 5) with 72 %

of dead stems found in the 10-20 cm DBH size class. Mean stem recruitment was 1.9 %

y-1 ± 0.35 % (95 % CI) for the entire community, stem recruitment was higher in the

lower plot at 1000 m with an increment of 2.9 % y-1, while the lowest rate was at 1500 m

with 0.9 % y-1. Overall stem recruitment was unrelated to elevation (R2 = 0.01, p = 0.78;

Figure 6b). Although there was not a consistent relationship of recruitment across

elevations, the mortality and recruitment rates taken together showed a significant

decrease in stand turnover (R2 = 0.3, p < 0.05) with increasing elevation (Figure 7). Mean

stand turnover was 2.16 % ± 0.3 % (95 % CI) for all plots combined. Rates of population

growth were negative for all of the plots below 2750 m except for the plot at 1000 m

(Figure 8). Across elevations population growth became less negative with increasing

elevation (R2 = 0.3, p < 0.05). Overall, there was a decline in the density of stems in the

tree community for this four year period in the tropical montane forest at Manu National

Park (Figure 9).

Mode of Death in Montane Forest

13

For all dead stems across the elevational gradient where information was available, the

uprooted mode of death was predominant (43.6 %), followed by snapped (36.7 %) and

standing (19.7 %), (Figure 10). The three types of modes of death showed different

trends across the elevation, the proportion of stems that died snapped decreased

significantly (R2 = 0.4, p < 0.05) with elevation (Figure 11a), standing proportions had no

relationship (R2 = 0.02, p = 0.59) with elevation (Figure 11b) and finally uprooted

increased significantly (R2 = 0.4, p < 0.05) with elevation (Figure 11c).

Stands Dynamics and Wood Density

Across the elevational gradient wood density values for 395 tree species (120 species

ABERG group (2010) and 275 between Chave (2009) and Zanne (2009) data set) were

collected and compiled. Species mean wood density for the study site was 0.61 g/cm3 ±

0.02 (95 % CI) and the median was 0.64 g/cm3. The maximum was 1.12 g/cm3 for

Machaerium acutifolium (Fabaceae) at 1250 m, and the minimum was 0.1 g/cm3 for

Erythrina ulei (Fabaceae) at 1500 m. Across the gradient, values of mean species wood

density at the plot level increased significantly with elevation (R2 = 0.6, p < 0.001; Figure

12).

The association of forest dynamics (mortality, recruitment, turnover rates) with species

wood density varied across the study area. Mortality rate had a tendency to increase with

species with low wood density (Figure 14a), however the coefficient of determination

was not significant (R2 = 0.2, p = 0.14). The relationship between recruitment and species

wood density was not significant (R2 = 0.004, p = 0.81, Figure 14b). Overall, forest

14

turnover had a slight relationship with species wood density (Figure 14c). Nonetheless

the relationship was not significant (N = 16, R2 = 0.2, p = 0.14). Frequency of snapped

stems had a significant association with wood density (N = 16, R2 = 0.4, p < 0.05; Figure

15a), but standing (Figure 15b) showed no association (N = 16, R2 = 0.002, p = 0.87).

Uprooted showed marginal association with species wood density (Figure15c), there was

no relationship (N = 16, R2 = 0.2, p=0.06).

15

DISCUSSION

This study provides insight into forest dynamics across a 2.5 km elevational gradient on

the eastern flank of the Peruvian Andes, ranging from the base of the Andes to tree line.

Across the gradient tropical montane cloud forest dynamics changed markedly with

elevation. This study found a significant effect of elevation in overall stem turnover and

mode of death, and a significant decline in stem density of the entire community.

Trends in Montane Forest Structure

The tendency of stem density and basal area along elevational gradients in the tropics is

to increase with elevation (Heaney and Proctor 1990, Kitayama 1992, Vazquez and

Givnish 1998, Aiba and Kitayama 1999). The results of this research showed a non-linear

relationship of stem density across the elevation from 1000 m to 3450 m, with a peak at

middle elevations (Figure 2a). This peak at mid-elevation could be due to the high

density of tree ferns in the cloud zone (Lieberman et al. 1996). The increase of multi-

stemmed trees was consistent across the elevation (Figure 2b) and the high density of

multi-stemmed trees at higher elevations support the findings in the Borneo transects

(Aiba and Kitayama 1999). The increase of frequency of multi-stemmed trees with

elevation could be due to the low stem diameter increment in the montane tropical forest

(1.8 mm y-1 ± 0.4 95% CI). It could also be explained by nutrient deficiency at high

elevations (Aiba and Kitayama 1999) in this case the high production of multi-stemmed

trees in poor soils promotes persistence of individuals even if some stems die

(Bellingham and Sparrow 2009). Thus, the strategy changes for tree species, from rapid

16

growth and higher individual mortality at low elevations to slower growth with lower

individual mortality at high elevations. Across the transect, the ability to produce multiple

stems was observed as a trait for species restricted to higher elevations with narrow

ranges (250 m range, e.g. Miconia setulosa Cogn., Weinmannia microphylla Kunth) and

also for species with a wide elevational range, though these produced multi-stemmed

individuals mostly at higher elevations (1250 m range, e.g. Prunus integrifolia (C. Presl)

Walp.).

Basal area was independent of elevation (Figure 3), which could be due to the presence of

large diameter trees close to the tree line (68.1 cm max DBH) where the trees have nearly

the same diameter, but are shorter. The lack of relationship of stem density and basal area

with elevation in this study is consistent even if the elevation range is expanded to the

lowland Amazonian forest (between 300 and 1000 m) with an average stem density of

600 h-1 and average of basal area of 38 m2 ha-1 (Pitman et al. 2001, Silman et al. 2006).

Stand Forest Dynamics

Across the 2.5 km elevational gradient, for the entire community in the tropical montane

cloud forest, there was a marked linear decrease of stem mortality with elevation (Figure

6a). Stand recruitment had a non-linear relationship with elevation (Figure 6b). A

decrease of recruitment with elevation was expected because stem diameter increment

decreases with elevation (Rapp 2010). A plausible explanation for the observed non-

linearity could be the internal heterogeneity of each permanent plot caused by local

disturbance (e.g. landslides, tree falls) effects on stand structure (stand diameter and

17

height) and species composition. Among hectares, variation can also be associated at

landscape-level patch dynamics in montane forests (Myster and Fernandez 1995), but this

has yet to be studied in the Andes.

The paucity of studies of stand dynamics in TMCF across elevational gradients makes it

difficult to compare results across tropical montane regions and hinders a better

understanding of patterns of forest dynamics. There are a few focal studies at only one

elevation examining tree mortality, recruitment and turnover in TMCF. Mean mortality

rates for the present study was 2.4 % y-1, which was almost the same value for TMCF at

the Jamaican site (2.5 % y-1), even though there was a high frequency of disturbance by

hurricanes (Bellingham and Tanner 2000). Mean stand recruitment rates were lower in

the present study in contrast to Jamaican stand recruitment rates, with values of 1.9 y-1

and 2.8 y-1 respectively. Overall, the mean stand turnover rate from this study was

2.16 % y-1, which is high in comparison to forest in Australia (1582 m), with a 1.7 % y-1

turnover rate (Herwitz and Young 1994), and was ~ 15 % lower than TMCF in Jamaica

(1300-1920 m) with a turnover increase of 2.6 %y-1 (Bellingham and Tanner 2000).

Across the gradient there was a marked negative association of stem turnover with

elevation. This significant linear relationship may be explained by the increase in species

turnover at lower elevations with the presence of fast growing species with short life

spans. An alternative explanation of the high turnover at lower elevations could be due to

the effect of soil nutrients along a fertility gradient (Phillips et al. 2004, Madelaine et al.

2007, Susan et al. 2009) in which high and low turnover rates are related to rich and poor

18

soils, respectively. In the study area across the elevational gradient, high nutrient soils are

often at lower elevations with progressively poorer soils with increasing elevation

(Zimmermann et al. 2009, M. Silman, unpublished data).

The significant decline of stems over the census interval for the entire community in the

eastern Peruvian TMCF could be explained by environmental factors that may play an

important role in the abundance and distribution of tree communities. High wind speeds

occurs frequently at lower elevations, and the strong gusts documented during 2007/8 wet

season (Rapp 2010) which fell between the two censuses (2003 to 2009) could have

increased the risk of mortality at lower elevations. An alternative explanation for the

decline in stem numbers could be the drought that affected the Amazonian Basin during

the austral winter in 2005 (Marengo et al. 2008) which lead to an increase in tree

mortality across the Amazon due to hydraulic limitations (Phillips et al. 2010). This event

could increase the mortality of stands with the largest impact expected at cloud base

(between 1500 -2000 m), which is consistent with the observed areas where population

growth was the most negative (Figure 8).

How do stems die?

The mode of tree death changed in a consistent manner across the gradient. Uprooting

was the most common mode of death along the gradient in contrast with snapped which

is the predominant mode of death in lowland Amazonian forest (Gale and Barfod 1999,

Chao et al. 2009). The significant increase of death by uprooting with elevation could be

explained by a weak tree anchorage (Putz et al. 1983) combined with shallow soils at

19

higher elevations. Snapped trees are related to biomechanical features and also to soil

depth that controls tree anchorage (Gale and Barfod 1999). These factors may explain the

significant linear relationship with more individuals dying by being snapped at lower

elevations. There may be fewer snapped tress simply due to forest structure. Snapped

trees often result from damage by other falling trees. Smaller trees at higher elevations

may not be able to generate the force required to snap trees. An alternative explanation

could be the difference in species composition along the gradient, where species (e.g.

Ficus spp, Sloanea spp.) at lower elevations have adaptations such as aerial roots and

buttressing for better anchorage. The Standing mode of death is related to slow growth

(Chao et al. 2009) and biotic interactions such as insect attack and fungal pathogens.

Across the gradient, standing had not significant relationship with elevation and it was

the least predominant mode of death that occurred in the lowlands.

Montane forest dynamics as a function of wood density

Wood density is an important functional trait of tropical trees and is also essential for

biomass calculations (Baker et al. 2004, Chave et al. 2006). The results here improve the

knowledge of wood density of tropical trees across an elevational gradient. The

significant increase of species wood density with elevation was not consistent with

findings in other Neotropical forests (Chave et al. 2006) where wood density decreases

with elevation. In this study, mean species mean wood density was 0.61 g/cm3 similar to

0.60 g/cm3 in Chave et al. (2006). Dividing the gradient into three major elevations (from

1000 to 1500 m, 1750 to 2500 m, 2750 to 3450 m), and weighting the wood density by

the number of species (Figures 13 a, b, c) shows an increase of wood density with

20

elevation with mean values of 0.57, 0.61, 0.62 respectively for each elevation. This trend

is even stronger if wood density is weighted by number of stems (Figures 13 d, e, f) with

mean values of 0.54, 0.64, 0.65 g/cm3 respectively for the three elevations. The observed

difference in the results of this study could be a result of the range of elevation sampled

from 1000 to 3450 m for this study, compared to ~ 50 to 2500 m, (Chave 2006).

Ultimately, more samples across elevational gradients and geographical regions are

needed to have a comprehensive view of elevation effects on wood density.

In the tropics, high wood density is associated with decreasing tree mortality, increasing

the probability of death in low wood density species (Chao et al. 2008, Chave et al. 2009).

In the Amazonian forest, fast growing species with low wood density tend to die snapped

(Chao et al. 2008). The results of this research confirm patterns of previous studies

showing a significant association of species with low wood density with dying by being

snapped. This is not a strong test to relate stand mortality and recruitment with wood

density at the plot level because it does not take into account factors such as tree size, tree

height and growth rate. Another way to model mortality, recruitment and the mode of

death across the gradient would be to calculate the individual-level probabilities of death

and recruitment.

21

CONCLUSIONS

This study illustrates the relationship in forest structure with respect to stem density and

basal area across the elevation. The results suggest that in these forests stem size does not

change markedly with elevation. Increasing the production of multi-stemmed individuals

at higher elevations could be a tradeoff of slow stem growth and low nutrients in the soil.

During a four year period of time, there was a decline in stem density due to higher rates

of stand mortality over stand recruitment for the entire community of 16 1-ha tree plots.

Possible explanations include severe wind events and the 2005 drought both of which

may have influenced the decline in stem density. The significant linear relationship of

stem turnover and elevation suggests that the species in this community die and grow

faster at lower elevations than at higher elevations and this could be attributed to both

higher diversity and an increased range of life history strategies coupled with a strong soil

fertility gradient. Species wood density has a strong relationship with elevation and in

turn appears to be related to mortality and it is correlated with snapped as the most

common mode of death at lower elevations. Forest dynamics are a consequence of

multiple factors, and the process of understanding which factors predict stem mortality

and recruitment across elevation is a continuing challenge. More research at larger

temporal and spatial scales is needed to fully understand the changes of dynamics in

montane forest. Studies of small tree size classes (e.g. seedlings, saplings and poles) and

future censuses of the permanent plot network will help to elucidate the life history and

demography of tree communities across elevational gradients in tropical regions.

22

LITERATURE CITED

ABERG. 2010. Permanent tree plot network. http://darwin.winston.wfu.edu/andes/.

Aiba, S., and K. Kitayama. 1999. Structure, composition and species diversity in an

altitude-substrate matrix of rain forest tree communities on Mount Kinabalu,

Borneo. Plant Ecology 140:139-157.

Arriaga, L. 2000. Types and causes of tree mortality in a tropical montane cloud forest of

Tamaulipas, Mexico. Journal of Tropical Ecology 16:623-636.

Baker, T. R., O. L. Phillips, Y. Malhi, S. Almeida, L. Arroyo, A. Di Fiore, T. Erwin, T. J.

Killeen, S. G. Laurance, W. F. Laurance, S. L. Lewis, J. Lloyd, A. Monteagudo,

D. A. Neill, S. Patino, N. C. A. Pitman, J. N. M. Silva, and R. V. Martinez. 2004.

Variation in wood density determines spatial patterns in Amazonian forest

biomass. Global Change Biology 10:545-562.

Bellingham, P. J., and A. D. Sparrow. 2009. Multi-stemmed trees in montane rain forests:

their frequency and demography in relation to elevation, soil nutrients and

disturbance. Journal of Ecology 97:472-483.

Bellingham, P. J., and E. V. J. Tanner. 2000. The influence of topography on tree growth,

mortality, and recruitment in a tropical montane forest. Biotropica 32:378-384.

Budd, P., I. May, L. Miles, and J. Sayer. 2004. Cloud Forest Agenda. United Nations

Environment - Programme-World Conservation Monitoring Centre, Cambridge,

UK.

23

Carey, E. V., S. Brown, A. J. R. Gillespie, and A. E. Lugo. 1994. Tree Mortality in

Mature Lowland Tropical Moist and Tropical Lower Montane Moist Forests of

Venezuela. Biotropica 26:255-265.

Chao, K. J., O. L. Phillips, E. Gloor, A. Monteagudo, A. Torres-Lezama, and R. V.

Martinez. 2008. Growth and wood density predict tree mortality in Amazon

forests. Journal of Ecology 96:281-292.

Chao, K. J., O. L. Phillips, A. Monteagudo, A. Torres-Lezama, and R. V. Martinez. 2009.

How do trees die? Mode of death in northern Amazonia. Journal of Vegetation

Science 20:260-268.

Chave, J., R. Condit, S. Lao, J. P. Caspersen, R. B. Foster, and S. P. Hubbell. 2003.

Spatial and temporal variation of biomass in a tropical forest: results from a large

census plot in Panama. Journal of Ecology 91:240-252.

Chave, J., D. Coomes, S. Jansen, S. L. Lewis, N. G. Swenson, and A. E. Zanne. 2009.

Towards a worldwide wood economics spectrum. Ecology Letters 12:351-366.

Chave, J., H. C. Muller-Landau, T. R. Baker, T. A. Easdale, H. Ter Steege, and C. O.

Webb. 2006. Regional and phylogenetic variation of wood density across 2456

neotropical tree species. Ecological Applications 16:2356-2367.

Clark, D. A. 2002. Are tropical forests an important carbon sink? Reanalysis of the long-

term plot data. Ecological Applications 12:3-7.

Clark, D. A., and D. B. Clark. 1992. Life-History Diversity of Canopy and Emergent

Trees in a Neotropical Rain-Forest. Ecological Monographs 62:315-344.

24

Condit, R., S. Aguilar, A. Hernandez, R. Perez, S. Lao, G. Angehr, S. P. Hubbell, and R.

B. Foster. 2004. Tropical forest dynamics across a rainfall gradient and the impact

of an El Nino dry season. Journal of Tropical Ecology 20:51-72.

Condit, R., P. S. Ashton, N. Manokaran, J. V. LaFrankie, S. P. Hubbell, and R. B. Foster.

1999. Dynamics of the forest communities at Pasoh and Barro Colorado:

comparing two 50-ha plots. Philosophical Transactions of the Royal Society of

London Series B-Biological Sciences 354:1739-1748.

Condit, R., S. P. Hubbell, and R. B. Foster. 1995. Mortality-Rates of 205 Neotropical

Tree and Shrub Species and the Impact of a Severe Drought. Ecological

Monographs 65:419-439.

Condit, R., Hubbell, S.P., Foster R.B. 1992. Short-Term Dynamics of a Neotropical

Forest. BioScience 42:822-828.

Denslow, J. S. 1987. Tropical Rain-Forest Gaps and Tree Species-Diversity. Annual

Review of Ecology and Systematics 18:431-451.

Denslow, J. S., A. M. Ellison, and R. E. Sanford. 1998. Treefall gap size effects on

above- and below-ground processes in a tropical wet forest. Journal of Ecology

86:597-609.

Gale, N., and A. S. Barfod. 1999. Canopy tree mode of death in a western Ecuadorian

rain forest. Journal of Tropical Ecology 15:415-436.

Gentry, A. H. 1988. Changes in Plant Community Diversity and Floristic Composition on

Environmental and Geographical Gradients. Annals of the Missouri Botanical

Garden 75:1-34.

25

Girardin, C. A. J., Y. Malhi, L. E. O. C. AragÃO, M. Mamani, W. Huaraca Huasco, L.

Durand, K. J. Feeley, J. Rapp, J. E. Silva-Espejo, M. Silman, N. Salinas, and R. J.

Whittaker. 2010. Net primary productivity allocation and cycling of carbon along

a tropical forest elevational transect in the Peruvian Andes. Global Change

Biology:no-no.

Heaney, A., and J. Proctor. 1990. Preliminary Studies on Forest Structure and Floristics

on Volcan Barva, Costa Rica. Journal of Tropical Ecology 6:307-320.

Herwitz, S. R., and S. S. Young. 1994. Mortality, Recruitment, and Growth-Rates of

Montane Tropical Rain-Forest Canopy Trees on Mount Bellenden-Ker, Northeast

Queensland, Australia. Biotropica 26:350-361.

Homeier, J., S. W. Breckle, S. Gunter, R. T. Rollenbeck, and C. Leuschner. 2010. Tree

Diversity, Forest Structure and Productivity along Altitudinal and Topographical

Gradients in a Species-Rich Ecuadorian Montane Rain Forest. Biotropica 42:140-

148.

Ingwell, L. L., S. J. Wright, K. K. Becklund, S. P. Hubbell, and S. A. Schnitzer. 2010.

The impact of lianas on 10 years of tree growth and mortality on Barro Colorado

Island, Panama. Journal of Ecology 98:879-887.

Kitayama, K. 1992. An Altitudinal Transect Study of the Vegetation on Mount Kinabalu,

Borneo. Vegetatio 102:149-171.

Laurance, S. G. W., W. F. Laurance, A. Andrade, P. M. Fearnside, K. E. Harms, A.

Vicentini, and R. C. C. Luizao. 2010. Influence of soils and topography on

Amazonian tree diversity: a landscape-scale study. Journal of Vegetation Science

21:96-106.

26

Laurance, W. F., A. A. Oliveira, S. G. Laurance, R. Condit, H. E. M. Nascimento, A. C.

Sanchez-Thorin, T. E. Lovejoy, A. Andrade, S. D'Angelo, J. E. Ribeiro, and C. W.

Dick. 2004. Pervasive alteration of tree communities in undisturbed Amazonian

forests. Nature 428:171-175.

Lewis, S. L., O. L. Phillips, D. Sheil, B. Vinceti, T. R. Baker, S. Brown, A. W. Graham,

N. Higuchi, D. W. Hilbert, W. F. Laurance, J. Lejoly, Y. Malhi, A. Monteagudo,

P. N. Vargas, B. Sonke, N. Supardi, J. W. Terborgh, and R. V. Martinez. 2004.

Tropical forest tree mortality, recruitment and turnover rates: calculation,

interpretation and comparison when census intervals vary. Journal of Ecology

92:929-944.

Lieberman, D., and M. Lieberman. 1987. Forest Tree Growth and Dynamics at La-Selva,

Costa-Rica (1969-1982). Journal of Tropical Ecology 3:347-358.

Lieberman, D., M. Lieberman, R. Peralta, and G. S. Hartshorn. 1996. Tropical forest

structure and composition on a large-scale altitudinal gradient in Costa Rica.

Journal of Ecology 84:137-152.

Madelaine, C., R. Pelissier, G. Vincent, J. F. Molino, D. Sabatier, M. F. Prevost, and C.

de Namur. 2007. Mortality and recruitment in a lowland tropical rain forest of

French Guiana: effects of soil type and species guild. Journal of Tropical Ecology

23:277-287.

Madsen, J. E., and B. Ollgaard. 1994. Floristic Composition, Structure, and Dynamics of

an Upper Montane Rain-Forest in Southern Ecuador. Nordic Journal of Botany

14:403-423.

27

Malhi, Y., M. Silman, N. Salinas, M. Bush, P. Meir, and S. Saatchi. 2010. Introduction:

Elevation gradients in the tropics: laboratories for ecosystem ecology and global

change research. Global Change Biology 16:3171-3175.

Manokaran, N., and K. M. Kochummen. 1987. Recruitment, Growth and Mortality of

Tree Species in a Lowland Dipterocarp Forest in Peninsular Malaysia. Journal of

Tropical Ecology 3:315-330.

Marengo, J. A., C. A. Nobre, J. Tomasella, M. F. Cardoso, and M. D. Oyama. 2008.

Hydro-climatic and ecological behaviour of the drought of Amazonia in 2005.

Philosophical Transactions of the Royal Society B-Biological Sciences 363:1773-

1778.

Midgley, J. J., M. C. Cameron, and W. J. Bond. 1995. Gap Characteristics and

Replacement Patterns in the Knysna Forest, South-Africa. Journal of Vegetation

Science 6:29-36.

Myster, R. W., and D. S. Fernandez. 1995. Spatial Gradients and Patch Structure on Two

Puerto Rican Landslides. Biotropica 27:149-159.

Negron-Juarez, R. I., J. Q. Chambers, G. Guimaraes, H. C. Zeng, C. F. M. Raupp, D. M.

Marra, G. Ribeiro, S. S. Saatchi, B. W. Nelson, and N. Higuchi. 2010.

Widespread Amazon forest tree mortality from a single cross-basin squall line

event. Geophysical Research Letters 37.

Phillips, O. L., T. R. Baker, L. Arroyo, N. Higuchi, T. J. Killeen, W. F. Laurance, S. L.

Lewis, J. Lloyd, Y. Malhi, A. Monteagudo, D. A. Neill, P. N. Vargas, J. N. M.

Silva, J. Terborgh, R. V. Martinez, M. Alexiades, S. Almeida, S. Brown, J. Chave,

J. A. Comiskey, C. I. Czimczik, A. Di Fiore, T. Erwin, C. Kuebler, S. G.

28

Laurance, H. E. M. Nascimento, J. Olivier, W. Palacios, S. Patino, N. C. A.

Pitman, C. A. Quesada, M. Salidas, A. T. Lezama, and B. Vinceti. 2004. Pattern

and process in Amazon tree turnover, 1976-2001. Philosophical Transactions of

the Royal Society of London Series B-Biological Sciences 359:381-407.

Phillips, O. L., and A. H. Gentry. 1994. Increasing Turnover through Time in Tropical

Forests. Science 263:954-958.

Phillips, O. L., P. Hall, A. H. Gentry, S. A. Sawyer, and R. Vasquez. 1994. Dynamics and

Species Richness of Tropical Rain-Forests. Proceedings of the National Academy

of Sciences of the United States of America 91:2805-2809.

Phillips, O. L., Y. Malhi, N. Higuchi, W. F. Laurance, P. V. Nunez, R. M. Vasquez, S. G.

Laurance, L. V. Ferreira, M. Stern, S. Brown, and J. Grace. 1998. Changes in the

carbon balance of tropical forests: Evidence from long-term plots. Science

282:439-442.

Phillips, O. L., R. V. Martinez, A. M. Mendoza, T. R. Baker, and P. N. Vargas. 2005.

Large lianas as hyperdynamic elements of the tropical forest canopy. Ecology

86:1250-1258.

Phillips, O. L., G. van der Heijden, S. L. Lewis, G. Lopez-Gonzalez, L. Aragao, J. Lloyd,

Y. Malhi, A. Monteagudo, S. Almeida, E. A. Davila, I. Amaral, S. Andelman, A.

Andrade, L. Arroyo, G. Aymard, T. R. Baker, L. Blanc, D. Bonal, A. C. A. de

Oliveira, K. J. Chao, N. D. Cardozo, L. da Costa, T. R. Feldpausch, J. B. Fisher,

N. M. Fyllas, M. A. Freitas, D. Galbraith, E. Gloor, N. Higuchi, E. Honorio, E.

Jimenez, H. Keeling, T. J. Killeen, J. C. Lovett, P. Meir, C. Mendoza, A. Morel, P.

N. Vargas, S. Patino, K. S. H. Peh, A. P. Cruz, A. Prieto, C. A. Quesada, F.

29

Ramirez, H. Ramirez, A. Rudas, R. Salamao, M. Schwarz, J. Silva, M. Silveira, J.

W. F. Slik, B. Sonke, A. S. Thomas, J. Stropp, J. R. D. Taplin, R. Vasquez, and E.

Vilanova. 2010. Drought-mortality relationships for tropical forests. New

Phytologist 187:631-646.

Pitman, N. C. A., J. W. Terborgh, M. R. Silman, P. Nunez, D. A. Neill, C. E. Ceron, W.

A. Palacios, and M. Aulestia. 2001. Dominance and distribution of tree species in

upper Amazonian terra firme forests. Ecology 82:2101-2117.

Putz, F. E., P. D. Coley, K. Lu, A. Montalvo, and A. Aiello. 1983. Uprooting and

Snapping of Trees - Structural Determinants and Ecological Consequences.

Canadian Journal of Forest Research-Revue Canadienne De Recherche Forestiere

13:1011-1020.

Rapp, J. 2010. Climate control on plant performance across an Andean altitudinal

gradient. Dissertation, Wake Forest University, Winston Salem, North Carolina,

USA.

Rees, M., R. Condit, M. Crawley, S. Pacala, and D. Tilman. 2001. Long-term studies of

vegetation dynamics. Science 293:650-655.

Rice, A. H., E. H. Pyle, S. R. Saleska, L. Hutyra, M. Palace, M. Keller, P. B. de Camargo,

K. Portilho, D. F. Marques, and S. C. Wofsy. 2004. Carbon balance and

vegetation dynamics in an old-growth Amazonian forest. Ecological Applications

14:S55-S71.

Sheil, D., D. Burslem, and D. Alder. 1995. The Interpretation and Misinterpretation of

Mortality-Rate Measures. Journal of Ecology 83:331-333.

30

Silman, M., A. Araujo Murakami, H. Pariamo, M. Bush, and D. Urrego. 2006. Changes

in tree community structure at the southern limits of Amazonia: Manu and Madidi.

Ecologia en Bolivia 40:443-452.

Susan, G. W. L., W. F. Laurance, E. M. N. Henrique, A. Andrade, P. M. Fearnside, R. G.

R. Expedito, and R. Condit. 2009. Long-term variation in Amazon forest

dynamics. Journal of Vegetation Science 20:323-333.

Swaine, M. D., D. Lieberman, and F. E. Putz. 1987. The Dynamics of Tree Populations

in Tropical Forest - a Review. Journal of Tropical Ecology 3:359-366.

Vazquez, J. A., and T. J. Givnish. 1998. Altitudinal gradients in tropical forest

composition, structure, and diversity in the Sierra de Manantlan. Journal of

Ecology 86:999-1020.

Watt, A. S. 1947. Pattern and Process in the Plant Community. Journal of Ecology 35:1-

22.

Whittaker, R. J., S. F. Schmitt, S. H. Jones, T. Partomihardjo, and M. B. Bush. 1998.

Stand biomass and tree mortality from permanent forest plots on Krakatau,

Indonesia, 1989-1995. Biotropica 30:519-529.

Zanne, A. E., G. Lopez-Gonzalez, D. A. Coomes, J. Ilic, S. Jansen, S. L. Lewis, R. B.

Miller, N. G. Swenson, M. C. Wiemann, and J. Chave. 2009. Data from: Towards

a worldwide wood economics spectrum. in Ecology Letters. Dryad Data

Repository.

Zimmermann, M., P. Meir, M. Silman, A. Fedders, A. Gibbon, Y. Malhi, D. Urrego, M.

Bush, K. Feeley, K. Garcia, G. Dargie, W. Farfan, B. Goetz, W. Johnson, K.

31

32

Kline, A. Modi, N. Rurau, B. Staudt, and F. Zamora. 2009. No Differences in Soil

Carbon Stocks Across the Tree Line in the Peruvian Andes. Ecosystems 13:62-74.

Table 1. Basic description of the 16 1-ha permanent tree plots along an elevational

gradient in the Manu National Park, Peru.

Location Plot Elevation (m) UTM E UTM N 1°Census 2°Census Substrate

Tono TO_1000 1000 225215 8567314 2004 2007 Shale

San Isidro SI_1250 1245 222870 8562054 2005 2009 Shale

Callanga CA_1250 1250 197947 8582756 2004 2008 Granite

Callanga CA_1500 1500 198534 8582816 2004 2008 Granite

San Pedro SP_1500 1500 224863 8556052 2006 2008 Shale

San Isidro SI_1500 1521 222305 8561620 2005 2009 Shale

San Pedro SP_1750 1750 224258 8556260 2006 2008 Shale

Trocha Union TU_1800 1800 222887 8553630 2003 2007 Granite

Trocha Union TU_2000 2000 222622 8553538 2003 2007 Granite

Trocha Union TU_2250 2250 221737 8552556 2003 2007 Shale

Trocha Union TU_2500 2500 220816 8551070 2003 2007 Shale

Trocha Union TU_2750 2750 219193 8549686 2003 2007 Shale

Trocha Union TU_3000 3000 218992 8549454 2003 2007 Shale

Wayquecha WA_3000 3000 219595 8540370 2003 2007 Shale

Trocha Union TU_3250 3250 217594 8549222 2003 2007 Shale

Trocha Union TU_3450 3450 217309 8548834 2003 2007 Shale

33

FIGURE LEGENDS

Figure 1. Map of the 16 1-ha ABERG network permanent tree plots in Manu National

Park across an elevational gradient. Each dotted circle represents a 1-ha tree plot.

Figure 2. (a) Variability of stem density (ha-1) across elevation in 16 permanent plots

(Linear regression; R=0.18, p=0.52, DF=14, n=16; R2=0.03, p=0.51), error bars depict

the 95 % confidence intervals. (b) Multi-stem production across elevation (Linear

regression; R=0.77, p < 0.001, DF=14, n=16; R2=0.6, p < 0.001), each point represents

the percentage of multi-stem individuals per hectare of stems ≥ 10 cm DBH below 1.3 m

height, error bars depict the bootstrapped 95 % confidence intervals. In both cases solid

lines are a nonparametric regression fit using a LOESS function.

Figure 3. Relationship of stem basal area (m2 ha -1) across elevation (Linear regression;

R=-0.38, p=0.15, DF=14, n=16; R2=0.14, p=0.15). Error bars depict bootstrapped 95 %

confidence intervals.

Figure 4. Difference in mortality and recruitment rates for all 16 1-ha plots across the

elevational gradient (Wilcoxon signed rank test, p < 0.001).

Figure 5. Frequency distribution for all dead stems along an elevational gradient as a

function of DBH (≥ 10 cm DBH).

34

Figure 6. Rates of forest dynamics across the elevation gradient for all stems ≥ 10 cm

DBH for: (a) Mortality using equation (1), points show annualized rates as a function of

elevation (Linear regression; R=-0.61, p < 0.05, DF=14, n=16; R2=0.4, p < 0.05). (b)

Recruitment using equation (2), points show annualized rates as function of elevation

(Linear regression; R=0.08, p=0.78, DF=14, n=16, R2=0.01, p=0.78). Error bars depict

the bootstrapped 95 % confidence intervals.

Figure 7. Turnover rates (mean of mortality and recruitment rates) for all stems ≥ 10 cm

DBH. Points show annualized rates as a function of elevation (Linear regression; R=-0.50,

p < 0.05, DF=14, n=16; R2=0.3, p < 0.05). Error bars depict the bootstrapped 95 %

confidence intervals.

Figure 8. Population change rates for all stems ≥ 10 cm DBH across the elevational

gradient. Rates for each permanent plot were calculated from equation (3), (Linear

regression; R=0.56, p < 0.05, DF=14, n=16; R2=0.3, p < 0.05), dashed lines indicate no

change. Error bars depict the bootstrapped 95 % confidence intervals. Solid line is a

nonparametric regression fit using a LOESS function.

Figure 9. Histogram of rates of population change for all stems ≥ 10 cm DBH across the

gradient. Rates for each permanent plot were calculated from equation (3). Dashed lines

depict no change.

35

Figure 10. Overall mode of death for all dead stems across the elevation gradient

(Percentage ha -1).

Figure 11. Variability of the mode of death for all of the dead stems ≥ 10 cm DBH

across the elevational gradient for (a) snapped stems (Linear regression; R=-0.62, p <

0.05, DF=14, n=16; R2=0.4, p < 0.05); (b) standing stems (Linear regression; R=-0.14,

p=0.59, DF=14, n=16; R2=0.02, p=0.59) and (c) uprooted stems (Linear regression;

R=0.60, p < 0.05, DF=14, n=16; R2=0.4 p < 0.05). Error bars depict the bootstrapped

95 % confidence intervals.

Figure 12. Relationship of mean species wood density for the 1 ha tree plots (weighted by

species number for each plot) across the elevational gradient (Linear regression; R=0.75,

p < 0.001, DF=14, n=16; R2=0.6, p < 0.001). Error bars depict the bootstrapped 95 %

confidence intervals. Solid line is a nonparametric regression fit using a LOESS function.

Figure 13. Frequency of wood density weighted by number of species in three elevational

groups from (a) 1000 – 1500 m, (b) 1750 – 2500 m and (c) 2750 – 3450 m. Wood

density weighted by number of stems for (d) 1000 – 1500 m, (e) 1750 – 2500 m and (f)

2750 – 3450 m. Dashed lines indicate the mean wood density for each elevational group.

Figure 14. Forest dynamics as a function of species wood density. For (a) mortality

(Linear regression; R=-0.38, p=0.141, DF=14, n=16; R2=0.2, p=0.14), (b) recruitment

36

(Linear regression; R=-0.06, p=0.81, DF=14, n=16; R2=0.004, p=0.81), and (c) turnover

(Linear regression; p=0.14, DF=14, n=16; R2=0.2, p=0.14).

Figure 15. Mode of death as a function of species wood density. For (a) snapped (Linear

regression; R=-0.61, p=0.01, DF=14, n=16; R2=0.4, p < 0.05), (b) standing (Linear

regression; R=-0.01, p=0.87, DF=14, n=16; R2=0.002, p=0.87), and (c) uprooted stems

(Linear regression; R= 0.43, p=0.06, DF=14, n=16; R2=0.2, p=0.06).

37

Figure 1.

Tono

San Isidro

San Pedro

Callanga

Wayquecha

Trocha Union

Andes

Amazon

38

Figure 2

1000 1500 2000 2500 3000 3500

400

600

800

1000

1200

Elevation (m)

Ste

ms

(ha-

1)

(a)

1000 1500 2000 2500 3000 3500

02

46

810

1214

Elevation (m)

Per

cent

age

(ha-

1)

(b)

39

Figure 3

1000 1500 2000 2500 3000 3500

2030

4050

Elevation (m)

Bas

al a

rea

(m2

ha-1

)

40

Figure 4.

plots$mort.rate.stem plots$recruit.rate.stem Recruitment rates Mortality rates

0.01

00.

015

0.02

00.

025

0.03

00.

035

0.04

0

41

Figure 5

Stem diameter (cm)

Freq

uenc

y of

dea

d st

ems

20 40 60 80

010

020

030

040

050

0

42

Figure 6

(a)

0.01

0.02

0.03

0.04

0.05

0.06

1000 1500 2000 2500 3000 3500

Elevation (m)

Mor

talit

y ra

te (h

a -1

)

(b)

1000 1500 2000 2500 3000 3500

0.01

0.02

0.03

0.04

Elevation (m)

Rec

ruitm

ent r

ate

(ha

-1)

43

Figure 7

1000 1500 2000 2500 3000 3500

0.01

0.02

0.03

0.04

Elevation (m)

Turn

over

rate

(ha

-1)

44

Figure 8

1000 1500 2000 2500 3000 3500

-0.0

3-0

.02

-0.0

10.

000.

010.

020.

03

p g

Elevation (m)

% c

hang

e in

ste

ms

(yr

-1)

45

Figure 9

-0.04 -0.02 0.00 0.02

05

1015

2025

% change in stems (yr -1)

Frac

tion

of s

tem

s

46

Figure 10

0.0

20.0

40.0

60.0

80.0

100.0

Standing Snapped Uprooted

Perc

enta

ge o

f dea

d st

ems

(ha-

1)

Mode of death

47

Figure 11

(a) (b)

020

4060

1000 1500 2000 2500 3000 3500

2040

6080

Elevation (m)

Per

cent

age

of s

napp

ed s

tem

s (h

a -1

)

(c)

1000 1500 2000 2500 3000 3500

020

4060

8010

0

Elevation (m)

Per

cent

age

of u

proo

ted

stem

s (h

a -1

)

1000 1500 2000 2500 3000 3500

Elevation (m)

Per

cent

age

of s

tand

ing

stem

s (h

a -1

)

48

Figure 12

1000 1500 2000 2500 3000 3500

0.55

0.60

0.65

0.70

Elevation (m)

Mea

n sp

ecie

s w

ood

dens

ity (g

/cm

3 h

a-1)

49

Figure 13

Fre

quen

cy o

f ste

ms

0.2 0.4 0.6 0.8 1.0 1.2

020

040

060

080

010

0012

00

(d)

Fre

quen

cy o

f spe

cies

0.2 0.4 0.6 0.8 1.0 1.2

050

100

150

200 (a)

Fre

quen

cy o

f ste

ms

0.2 0.4 0.6 0.8 1.0

050

010

0015

0020

0025

00

(e)

Fre

quen

cy o

f ste

ms

Fre

quen

cy o

f spe

cies

0.2 0.4 0.6 0.8 1.0

020

4060

8010

012

0

(b)

(c) (f)

0.2 0.4 0.6 0.8 1.0

050

010

0015

0020

0025

0030

00

Fre

quen

cy o

f spe

cies

020

4060

80

0.2 0.4 0.6 0.8 1.0

Wood density g/cm3

50

Figure 14

(a) (b)

(b)

0.56 0.58 0.60 0.62 0.64

0.01

00.

015

0.02

00.

025

0.03

00.

035

0.04

0

Mean species wood density (g/cm3 ha-1)

Mor

talit

y ra

te (h

a -1

)

0.56 0.58 0.60 0.62 0.64

0.01

00.

015

0.02

00.

025

0.03

0

Mean species wood density (g/cm3 ha-1)

Rec

ruitm

ent r

ate

(ha

-1)

0.56 0.58 0.60 0.62 0.64

0.01

50.

020

0.02

50.

030

0.03

5

Mean species wood density (g/cm3 ha-1)

Turn

over

rate

(ha

-1)

51

Figure 15

010

2030

4050

(a) (b)

(c)

0.56 0.58 0.60 0.62 0.64

2030

4050

Mean species wood desnity (g/cm3 ha-1)

% s

napp

ed (h

a -1

)

0.56 0.58 0.60 0.62 0.64

3040

5060

70

Mean species wood desnity (g/cm3 ha-1)

% u

proo

ted

(ha

-1)

0.56 0.58 0.60 0.62 0.64

Mean species wood desnity (g/cm3 ha-1)

% s

tand

ing

(ha

-1)

52

CURRICULUM VITAE

WILLIAM FARFAN RIOS EDUCATION B.S. Biology, Universidad Nacional de San Antonio Abad del Cusco (Cusco-Peru). RESEARCH EXPERIENCE • Researcher, Andes Biodiversity and Ecosystem research Group, 2007 – present.

Project: Conservation implications of climate change and fire in the Eastern Andes: Impacts on plant distributions and montane ecosystems” in southern Peru, Location: Manu (Peru)

• Researcher, Conservacion Internacional, 2007 - 2009 Project: Expanding the permanent plot network into the montane forests of the Andes Location: Bolivia, Colombia, Ecuador, Peru

• Researcher, Wake Forest University, 2003 - 2006 Project: Vegetation and climate along an altitudinal gradient in Manu National Park. Location: Manu (Peru)

• Research assistant, Centro Huaman Poma de Ayala, 2004. Project: Diagnostico de Recursos Naturales del Valle de Cusco Location: Cusco (Peru)

• Responsible botanist, O.N.G. Jardín Botánico de Missouri, 2003 - 2004 Location: Manu (Peru)

• Research assistant, Erick A. Christenson, 2002 Project: Orchid of the South of Peru Location: Calca, Cusco, Manu (Peru)

• Research assistant, Fondo Nacional Para Áreas Protegidas Por el Estado (PROFONAMPE), 2001-2002 Project: Orquídeas del Santuario Histórico de Machu Picchu Location: Machu Picchu (Peru)

PUBLICATIONS • Feeley K.J., M.R. Silman , M. Bush, W. Farfan, K. Garcia Cabrera, Y. Malhi, P. Meir,

N. Salinas Revilla, M. N.Raurau Quisiyupanqui, S. Saatchi. 2011. Journal of Biogeography 38:783–791.

• Gibbon, A., M. R. Silman, Y. Malhi, J. B. Fisher, P. Meir, M. Zimmermann, G. C. Dargie, W. R. Farfan, and K. C. Garcia. 2010. Ecosystem Carbon Storage Across the Grassland-Forest Transition in the High Andes of Manu National Park, Peru. Ecosystems 13:1097-1111.

• Zimmermann, M., P. Meir, M. R. Silman, A. Fedders, A. Gibbon, Y. Malhi, D. H. Urrego, M. B. Bush, K. J. Feeley, K. C. Garcia, G. C. Dargie, W. R. Farfan, B. P. Goetz, W. T. Johnson, K. M. Kline, A. T. Modi, N. M. Q. Rurau, B. T. Staudt, and F.

53

Zamora. 2010. No Differences in Soil Carbon Stocks Across the Tree Line in the Peruvian Andes. Ecosystems 13:62-74.

• Farfan, W., K. Feeley. 2009. Deforestación y el mercado de carbono en los bosques tropicales. Xilema 26:11-16.

• Girardin C.A.J., W. Farfan, K. Garcia, Y. Malhi, T. Killeen, K.J. Feeley, M.R. Silman, C. Reinel, D. Niell, P. Jorgensen, M. Serrano, J. Caballero, M.A. De la Torre Cuadrada, M. Macía. 2009. RAINFOR-Andes Expanding the Amazon Forest Inventory Network to the montane forests of the Andes. Technical Report, Conservation International.

• García K., W. Farfan, N. Salinas, B. León. 2005. Helechos Arbóreos de Trocha Unión. Environmental & Conservation Programs, The Field Museum, Chicago, USA. Rapid color guide Nº 182 version 1, 2.

• Collaboration in the field work for the elaboration of the book “AMANECER EN EL BAJO HUATANAY”, Diagnostico de Recursos Naturales del Valle de Cusco”, published by the Centro Huaman Poma de Ayala, Cusco 2004.

• Farfan W. and E.A. Christenson. 2003. Lankesterella, A GENUS NEW TO CUSCO, PERU. THE ORCHID REVIEW, Volume 111, Number 1251, pp 150-151.

• Moretz C.C. and W. Farfán. 2003. A NEW Telipogon FROM SOUTHERN PERÚ. THE ORCHID REVIEW, Volume 111, Number 1252, pp. 239-241.

• Salinas Revilla N., W. Farfán. 2003. Dos Nuevas Especies de Orquídeas para el Santuario Histórico de Machupicchu. CANTUA Revista de Ciencias Biológicas. Nº 11. pp. 31-34.

IN REVIEW • Jankowski J., C. Merkord, W. Farfan Rios, K. Garcia Cabrera, N. Salinas Revilla,

M.R. Silman. Links between Andean bird and tree communities: Using current associations to anticipate effects of climate change. Journal of Biogeography.

• Gurdak D.J., L.E.O.C. Aragão, A. Rozas-Dávila, W. Huaraca Huasco, W. Farfan Rios, K. Garcia Cabrera, C.A.J. Girardin, M.R. Silman, D.B. Metcalfe, J.E. Silva Espejo, N. Salinas Revilla, Y. Malhi. Tropical Necromass - Dynamics along an Elevational Gradient of Mature Forest in the Peruvian Andes. Ecosystems.

FELLOWSHIPS AND GRANTS • Andes Biodiversity and Ecosystem Research Group (ABERG). Fellowship

2009-2011 • Wake Forest University. Vecellio grant 2010 • Wake Forest University. Vecellio grant 2009 • Fondo Verde. Fellowship 2007 • Pan – Amazonia. Fellowship 2006 • Red Latinoamericana de Botánica. Travel grant 2005 • Missouri Botanical Garden. Fellowship 2003

54

55

WORKSHOPS and ADDITIONAL TRAINING • Trained “Diplomado de Experto en desarrollo sustentable Universidad Nacional del

Centro” – Facultad de pedagogía y humanidades, Institución Fondo Verde. 2007 - 2008.

• II Workshop “Internacional de Treinamento em Analises e Interpretacao de Datos em Ecosistemas Florestais”, Bolivia, July 2006.

• II Curso de Ecología de Bosques Nublados, Bolivia. April 2005. • Curso - Procesamiento y Análisis de datos con el Paquete SPSS 11.5, November 2004. • Curso - Pasantía Preprofesional del Programa de Capacitación en Botánica y Conservación auspiciado por Christensen Fund y Missouri Botanical Garden, Ecuador. February–March 2003. SERVICE Member of the scholarship selection committee for “Asociacion para la Conservacion de la Cuenca Amazonica” (ACCA). Peru. Jun – July 2010.