Transcript Pot Experiment - Wageningen UR

Bush encroachment in African savannas
David Ward
How do we
go from this ?
to this ?
Namibia
India
Bush encroachment affects
between 12- 20 million hectares of
South Africa
This is a biodiversity problem
that is also
an agricultural problem
A multi-species grass sward is
transformed into an impenetrable and
unpalatable thicket dominated by a single
species of thorn tree
Heavy Grazing is often
considered to be the cause
of bush encroachment
• Walter’s (1939) two-layer model –
– grasses outcompete trees in open savannas by
growing fast and intercepting moisture from the upper
soil layers,
– trees are thereby prevented from gaining access to
moisture in the lower soil layers where their roots are
mostly found.
– when heavy grazing occurs, grasses are removed and
soil moisture then becomes available to the trees,
allowing them to recruit en masse.
Post hoc ergo propter hoc
• The fact that many bush-encroached
areas are heavily grazed means
neither that grazing causes
encroachment nor that Walter’s
model is correct
• Bush encroachment is widespread in
areas where there is a single soil
layer and where grazing is infrequent
and light
Magersfontein battlefield in 1899 and 2001
– it is now bush encroached in spite of an
absence of heavy grazing
Distribution of A. mellifera
Pniel study site
(nr. Kimberley)
Acacia mellifera
Resource allocation models of
plant community structure
David Tilman
Univ. of
Minnesota
In order to predict the outcome of
competition for a single limiting
resource, it is necessary to know:
• The resource level (=R*) at which the
net rate of population change for a
species is zero
• This occurs when vegetative growth
and reproduction balance the loss rate
the species experiences in a given
habitat
R* and loss or disturbance rates
• The loss rate of a population is caused by
numerous components, including
disturbance, seed predation, fire and
herbivory
• Independent of the causes of losses, the
number of species competing, or
competitive abilities of species in a
habitat, average (equilibrial) resource
levels (R*) will increase with the loss rate
Growth or Loss rate, dB/Bdt
R* will increase with the loss rate
Loss
R*
R, Resource level
A population can only be maintained in a
habitat if its growth rate > loss rate
Growth or Loss rate, dB/Bdt
Species A
LossA
Species B
LossB
Species C
LossC
R*C R*A R*B
R, Resource level
Species C will exclude the other 2 species
in competition because it has the lowest R*
Resource-dependent Growth
Isoclines
• When a species consumes 2 or more
resources, it is necessary to know the total
effects of the resources on the growth rate of
the species
• These effects can be summarized by the zero
net growth isocline (ZNGI)
• This isocline shows the levels of 2 or more
resources at which the growth rate per unit
biomass of a species balances its loss rate
Perfectly essential resources
y
R2
x
0
0
R1
If a habitat is at point x,
an increase in R1 will
not affect population
size. However, any
increase in R2 will
cause an increase in
population size (& vice
versa for habitat at y).
Population size decreases for resource
levels in the white region and increases in
the green region
Species A dominant
A
R2
B
Species B dominant
R1
When the ZNGI cross, each species will
have a range of R* for the 2 resources
where it will dominate
R2
Bc1
Bc
Bc2
•Thus far, we have
considered resource
availability
•Consumption also
needs to be considered
because it affects
subsequent availability
R1
The consumption vector, Bc, has 2
components: c1 = amount of resource 1
consumed per unit biomass per unit time
and c2 (~ for R2)
R2 (light)
Bc1
Bc
Bc2
The consumption
vectors are
determined in large
part by the plasticity
of plant growth
R1 (nutrient)
e.g. If R1 = a nutrient and R2 = light, the
plant must allocate resources to aboveground growth (towards the light) and to
below-ground growth (towards the
nutrients)
A wins
A
R2
A+B
Stably coexist
B
B wins
cB
cA
R1
When there are perfectly essential resources,
the optimal strategy for a plant is to grow
so that the 2 resources are consumed in a
way that they equally limit growth
Grasses
Soil Water
In South African savannas
Trees
win
Stably
coexist
+H2O
+N
Grasses win
Acacia
Grasses
Soil Nitrogen
How do grazing or fire affect
the isoclines ?
• Grazing/Fire increase the loss rate
for grasses
• Thus, R* for grasses is raised relative
to that of the Acacia trees
Acacia
Grasses
Soil Nitrogen
When ZNGIs do not cross,
Acacias always outcompete
grasses
Soil Water
Grasses
Soil Water
+H2O
+N
Either of these
scenarios is
possible
Soil Nitrogen
Global climate change models predict that
C3 trees will grow faster following climate
change than C4 grasses
C3 (trees)
30
C4 (grass)
20
10
Now
200
600
CO2 (ppm)
Predicted
1000
Increased atmospheric CO2 levels will
mean that:
•Net photosynthetic rates of C3 trees will
increase more than those of grasses
•Consequently, growth rates of trees will
increase, and…….
Because more carbon will be available:
• Acacia trees will be able to invest more
in carbon-based defences, such as
condensed tannins (see e.g. Lawler et al.
1997, Kanowski 2001, Mattson et al.
2004)
•Consequently, loss rates of Acacias are
likely to decline
Growth or Loss rate, dB/Bdt
Increased growth and decreased loss for
Acacias results in a lower R*
Growth –
after climate
change
Growthnow
Lossnow
R*predicted R*now
R, Resource level
Loss – after
climate
change
Soil Water
This resource allocation model predicts that
this will lead to bush encroachment because
the ZNGI of Acacias will be lower (closer to
the origin) than that of grasses on both axes
Soil Nitrogen
Grasses
Soil Water
Do we have any empirical support
for this model ?
Trees
win
Stably
coexist
+H2O
+N
Grasses win
Acacia
Grasses
Soil Nitrogen
Pot Experiment
• Treatments: rain, nutrients, grazing
• Completely crossed design
Mean # surviving plants (+SE)
Rainfall frequency
overwhelmingly more important
than other factors
80
R = rain
D = dry
60
N = nitrogen
40
O = no nitrogen
G = grazing
20
_ = no grazing
0
RN_
RO_
RNG ROG
DN_
DO_
DNG DOG
Field experiment - randomized block design
Treatments: rain, fire, nutrients, grazing
Rainfall addition
increased Acacia
germination &
survival
4
3
2
1
0
Rain Added
Control
Nitrogen addition
decreased Acacia
germination &
survival
No. Tree Seedlings
No. Tree Seedlings
5
7
6
5
4
3
2
1
0
Nitrogen
Added
Control
Jack Kambatuku, a PhD student of mine,
has shown that Δ15N is related to
competition with grass
15N
Natural Abundance
6
5
F(2, 165) = 93.9, p < 0.001
4
3
2
1
0
-1
Grass
No Grass
Competition
Dry Matter Production (g)
Jack has shown that dry matter
production is affected by competition
with grass
= Total D.M. Production
= AboveGround D.M. Prod.
= BelowGround D.M. Prod.
8
6
4
2
0
No Grass
Grass
Jack has also shown that free-growing trees have higher
nitrogen content than trees growing with grasses
Interaction effect (Rain*Seeds):
F=7.961, p=0.006
Vertical bars denote 0.95 confidence intervals
3
130 year Max. Rainfall
Natural Rainfall
2
1
0
-1
Added
Control
Seeds
Experimental results thus far
• Grazing and fire not important
• Rainfall far more important than other
factors
• Rainfall frequency more important than
rainfall amount
• Nutrients = second-most important factor
• More nutrients = competitive advantage to
grasses = tree suppression
• Thus, the resource allocation model
seems appropriate
Biomass
The relationship between
grass/tree biomass and rainfall
Without grazing
Open Savanna
Grass
Trees
Annual Rainfall
In areas prone to bush encroachment,
farmers should limit stock in WET years
With heavy grazing
Grass
Trees
Annual Rainfall
We are also using
Spatially-explicit Patch Dynamic
Models of Savanna Dynamics
Experiments show that mature trees are
competitively superior to grasses while
grasses tend to outcompete
immature trees
• This asymmetry in competitive effects implies
instability
• However, weakening the suppressive effect of
the grass layer on young trees in a patch of a
few hectares can lead to an open savanna
patch being converted to a tree-dominated
thicket (bush encroachment)
• Once established, the thicket may take decades
to revert to an open savanna
A
B
C
Honeycomb
rippling
model
D
E
F
Figures show a time series of hexagonal subsets of
a larger patch. Each (small) hexagonal represents a
bush, the relative sizes of the hexagonals represent
relative bush sizes
The predictions of the
honeycomb rippling model are
consistent with field data that
show that:
• Distances between trees increase
with age
• Trees become more evenly spaced
as they age
Distances
between
trees increase
as they age
Variability
in distances
between
trees
decreases
as they age
We showed experimentally that
there is significant competition
between trees
30
25
20
% Size 15
Increase 10
5
0
Neighbours
removed
Control
Summary of patch dynamic model results
We have shown that:
•Any process that weakens the suppressive
effect of grasses on young trees can convert
an open savanna patch into a treedominated thicket (= bush encroachment)
•Thicket may eventually revert to an open
savanna as a result of intra-specific
competition between trees (= cyclical
succession)
Viewed this way, bush encroachment may
be a natural stage in savanna dynamics
Another South African example of
cyclical succession – Karen Esler
One of our students, Jana Förster, has
shown that there may be strong
competition between two encroaching
species, Acacia mellifera and
Tarchonanthus camphoratus
Relative frequency
With A. mellifera removed,
T. camphoratus gets larger
and has recruitment
Uncut plots
Relative frequency
2
44
Cut plots
2
44
A. mellifera
T. camphoratus
a
86 120 168 210 260 292 >>
b
86 120 168 210 260 292 >>
Canopy diameter, cm
Overall Conclusions
• Heavy grazing is only one of several
sources of loss to plants that affect R*
and consequently competitive ability of
trees against grasses
• Rainfall frequency and nutrient availability
are important in initiating encroachment
• Resource allocation models are useful for
predicting changes in savanna dynamics
• Patch dynamic models can explain bush
encroachment as a natural stage in
savanna dynamics