Transcript Worldwide experiment: tree seedling growth in savannas

GEST
Global Experiment on Savanna Tree seedlings
Dr Frank van Langevelde
Dr Kyle Tomlinson
Prof Steven de Bie
Wageningen University
Shell Research Foundation
Global Experiment on Savanna Tree seedlings

Context

Research questions

Global experiment
Context

Savanna is vegetation where trees and grasses coexist

we focus on savannas where seasonality is controlled by rainfall
rather than temperature

However: large variation in vegetation structure and
composition around the world

Environmental characteristics:





water (resource: limited in dry season)
nutrients (resource: limited in wet season)
fire (disturbance)
herbivory (disturbance)
Heavy human interventions!
Theoretical issues

How can trees and grasses coexist (“the savanna
question”)?

Determinants of savanna tree cover?

…?
However: focus on Africa?
Applied issues

Increase in woody cover (wet regions, “bush encroachment”)

Recruitment limitations of trees (dry regions)

Loss of large savanna trees

…?
“The savanna question”

Existing theories (Sankaran et al. 2004)



Equilibrium: stable co-existence independent of rainfall variability
and disturbance (Walter’s 2-layer hypothesis, Walker et al. 1982)
Non-equilibrium: co-existence depends on recruitment bottlenecks
mediated by disturbance; independent of competitive interactions
(Higgins et al. 2000)
Disequilibrium: unstable co-existence depends on disturbance;
without disturbance → grassland or woodland (Jeltsch et al. 2000)
Comments on theories

Problem: no unified theory of how trees and grasses coexist
in savannas

2 types of arguments for coexistence:


Competitive interactions (equilibrium, disequilibrium)
Demographic bottlenecks on tree recruitment (climate, disturbances)
(non-equilibrium, disequilibrium)

Disequilibrium and non-equilibrium: tree seedling recruitment
is critical to long-term dynamics & disagree over relevance of
tree-grass competition

Research on tree seedling growth could help to resolve this
debate
Savanna tree seedling research to date
(Tomlinson et al. in prep)

Research on tree seedling growth and survival in
savannas is fragmentary and continentally-biased

Preponderance of demographic studies of recruitment


effects of fire, herbivory, drought
Scarcity of formal experiments to elucidate relative
importance of competitive suppression versus resource
limitation for seedling growth


mostly North American
single species considered
But if we look at savannas more broadly…

Not all responses of trees are adaptations to current
environment, they may also be explained by evolutionary
or biogeographic constraints

Major physiognomic differences within savannas around
the world



Assemblage structure (rainfall, soil type)
Plant characteristics (selection pressure, phylogeny)
There are major biogeographic patterns in tree distribution
between savannas around the world
Africa

Numerous genera present,
but local dominance by
single genus or family



Mimosoideae (eutrophic)
Combretaceae (dystrophic)
Caesalpinioideae (dystrophic)

Soil type and rainfall
generally define dominant
species abundance

High tree species diversity
Australia

Local dominance by single
genus or family




Eucalyptus (Myrtaceae)
Mimosoideae
Proteaceae
Soil type and rainfall
generally define dominant
genera abundances
South America

In Cerrado, co-dominance by species of numerous
families some of which have low representation in other
continental savannas
Asia

More extreme soil fertility
(both poor and rich)

Phylogenetic commonality
with Africa, but also Asian
genera and codominance by
genera of multiple families?

The most important:



Combretaceae
Mimosoideae
Dipterocarpaceae
North America

Warm temperate to tropical savannas

Small tree family pool, but some dominant families

Fagaceae, Pinaceae, Mimosoideae
What do we observe?

Tree characteristics of savannas are varied, both within
continents and across continents … Why?

There are differences in environmental limitations constraining
savanna tree seedling recruitment


rainfall, nutrients, temperature, light, fire, herbivory
Are different traits required for each of these environments?
What can we predict?

Current environmental factors determine tree seedling
growth and survival in the presence and absence of grass
in savannas around the world, irrespective of phylogeny.

Environmental adaptation




Abiotic: water & variability in water, nutrients, light
Biotic: competition with grasses (local species), defoliation (local
herbivores), N-fixation, VAM
Role of fire?
Phylogenetic origin


Evolutionary selection from local species pool
Invasion from external source? (Mimosoideae?)
Research questions

Do tree seedlings in savannas around the world differ in
their




response to resources (water, nutrients, light)?
ability to compete with grasses for different resources?
ability to tolerate defoliation?
Do the species’ seedlings show convergent traits for
environmental conditions, or are the differences related to
continent of origin, or to phylogeny?
Global experiment

Experiment comparing seedling growth of dominant (“matrix”)
tree species of tropical and warm temperate savannas
around the world

Runs for 2 years

Experiment will be conducted at two sites on each continent


high rainfall: MAP > 800 mm
low rainfall: MAP < 600 mm
Location of sites
Tree species choice

In wet and dry savannas, we identified eutrophic and
dystrophic savannas (total 4 savanna types)

dry+eutrophic, dry+dystrophic, wet+eutrophic, wet+dystrophic

We selected 3 tree species from each type

In the field experiment, we shall grow at least 2 species from
each type (total 8 species) = transplant experiment

12 species from each continent grown in a comparative pot
trial in the Netherlands (total 60 species) = common-garden
experiment
Site characteristics
1.
Fenced or no macroherbivores
2.
Shallow gradient
3.
Soils sandy-textured, well-drained
4.
No trees
Preferably
5. Access to tapped water
Proposed treatments
1.
Water: even rainfall & natural rainfall
• dry site: 400 mm/year, per week over 6 months
• wet site: 1000 mm/year, per week over 6 months
2.
Nutrients: no fertilizer added & fertilizer added
• NPK start + every month for first 3 months (4 applications), so that N =
4 g N m-2 per application (Kraaij & Ward 2006)
3.
Shade: full sun & 80% shading
4.
Grass: no grass & grass
• local abundant species
5.
Defoliation: no defoliation & defoliation
• Seedling (above 2nd internode) and grass (3 cm height) clipped twice
over 6 months (wet season)
Experimental design

Incomplete factorial design of 16 treatment combinations
L0 = full sun
W0 = natural rain
N0 = no nutrients
added
G0 = no grass
D = defoliation
Experimental design

All 8 species grown together in ±2.5 m2 treatment plots
Experimental design

4 seedlings of each species per plot: 2 harvested after 1st
year, 2 replanted after 1st year, 2+2 harvested after 2nd year
Experimental design

5 block reps, giving a total of 10 seedlings per species per
treatment
Measurements

Monitored development
1.
stem length & stem height (2 weeks)
2.
basal stem diameter (1 cm above ground, 2 weeks)
3.
number of leaves (bimonthly) & leaf area (bimonthly)
4.
mark 1 leaf per seedling per month
5.
number of live seedlings at begin of wet season
6.
Grass biomass (disc meter, monthly)

Harvested measurements (even: 1 week after last water supply)

aboveground dry mass of each seedling (separate leaves and
stem)

aboveground dry mass of grasses

seedling and grass leaves N and P and K and PSC
Harvested measurements (natural: same moment as even)

aboveground dry mass of stem of each seedling

aboveground dry mass of grasses

Greenhouse experiment

Rigorous comparison of tree species to test effect
environmental adaptation and phylogenetic constraints

Why?


Common-garden experiment & transplant experiment
Ecological significance of variation in life-history traits
Common-garden & transplant experiments

Transplant experiment: estimate local performance by
measuring individuals that have been moved between
environments


Common-garden experiment: estimate performance by
measuring species drawn from local environments within
common-garden (greenhouse) experiment


there is an effect of environmental variation
minimizes contribution of environmental variation
Careful design to make quantitative statements about the
importance of environmental variation (Nuismer & Gandon 2008)
Variation in life-history traits

Cause for variation: coping with disturbance and adapting to
fast versus slow growth conditions

Life-history invariants: traits that occur in same combination


for example: leaf mass per area – leaf life span
Greenhouse experiment allows
to measure this variation
Products
Several papers

Synthesis paper: “Evolution and ecology of savannas around the
world: differences and similarities” (co-authored by all participants)

Worldwide comparison of seedling growth under different
environmental conditions: results of field experiment (co-authored by
all participants)

Traits adapted to environment or constrained due to phylogeny:
results of greenhouse experiment (co-authored by all participants)

Experiments per continents (co-authored by local participants and
WUR team)

Possibly: a subsequent synthesis paper on strategies used by trees
in different savannas (co-authored by all participants)
Miscellaneous

GEST website




All presentations on the site (accessible for GEST members)
All data on the site
Suggestions or material (papers, pictures)?
GEST is opportunities for training


For example, Wageningen MSc students could join experiments
(Regular) GEST meeting for participating students? For example
to learn about statistics (EU-funding?)