Marleen Cobben
Centre for Genetic Resources
The Netherlands
CGN, met goede opmaak

Postdoc at CGN
The effects of climate change on wild relatives of crops
important for European breeders

Starting with gap analysis
• Using gap analysis and GCM to assess future habitat
areas for CWR
• Here at CIAT for two weeks to learn this methodology
• Important tool to get an estimate of the vulnerability of
species to climate change
• Species with declining or disappearing habitat area are
considered vulnerable

Jarvis et al. AGEE 2008

Gap analysis
• I am currently however inclined to think that we should
prioritise the collection of species in populations we think
will go extinct soon.
• This is not necessarily the same
• The focus shifts from species to populations
• In the following I will explain why I feel this way
• And I like to hear your thoughts about this

Marleen Cobben René Smulders Jana Verboom Rolf Hoekstra Paul Opdam
Adapt, move or perish
The interaction of range shifts and
genetics under climate change

Collaboration within Wageningen-UR
Plant Breeding
Alterra
Land Use Planning
Genetics

Metapopulations
In fragmented landscapes species are often confined to
metapopulations

Metapopulations
In fragmented landscapes species are often confined to
metapopulations

Metapopulations
In fragmented landscapes species are often confined to
metapopulations

Metapopulations
In fragmented landscapes species are often confined to
metapopulations

Climate change
• Natural populations of species need to respond to
climate change. They may
– track suitable climate, and thus shift their range
– adapt to changed climate
• These responses may occur together and interact
• In my thesis I investigated both responses and their
interaction

Research question
How will the level and distribution of neutral genetic
diversity in metapopulations be affected by range shifts
which are induced by current climate change?

Climate scenarios
• Overall increased temperature
– Hadley Centre:
• 1 C warming by 2100
• 2 C warming by 2100
• 4 C warming by 2100
• Increased weather variability: more weather extremes

Simulation study
• METAPHOR: simulates metapopulation demography
• + shaking windows: simulates temperature increase and
weather variability
• + genetics: each individual has its own genome
100 genes, diploid
WHAT DOES IT LOOK LIKE?

North
METAPHOR Habitat patch
individuals of the species
chance to survive
2000 km
chance to breed
chance to disperse
South

North
METAPHOR
individuals of the species
chance to survive
2000 km
chance to breed
chance to disperse
South

North
METAPHOR
individuals of the species
chance to survive
2000 km
chance to breed
chance to disperse
South

North
METAPHOR
individuals of the species
chance to survive
2000 km
chance to breed
chance to disperse
South

North
+ shaking windows
climate translates to habitat Habitat becomes suitable
quality
chance to survive Shaking movement
chance to breed of bell shaped window
2000 km
chance to disperse
Optimal Habitat
each year varies randomly
around the optimum
but on average moves Habitat deteriorating
northwards:
1 C : 2 km/year
2 C : 4 km/year
4 C : 8 km/year
South

North
+ genetics
neutral genes
diploid inheritance
recombination
2000 km
mutations: 10-4 per generation
(microsatellite mutation rate)
numbers of alleles per locus
effective numbers of alleles per locus
spatial distribution of both
South

North
+ genetics
2000 km
number of alleles: 7
effective number of alleles: 5.33
South

Simulation result
climate optimum: 400 km
temperature speed: 2 km/year, so 1 C scenario
weather variability: 140 km

Concluding
Under 2 and 4 degrees temperature increase scenarios the
metapopulation goes extinct
All temperature increase scenarios show loss of neutral
genetic diversity as a combination of ‘allele surfing’ at the
leading edge and ‘allele wipe-out’ at the trailing edge
Cobben et al. 2011 Ecography

Maybe it’s the landscape?
2.5% 5% 10%

2.5% landscape area

10% landscape area

So
 Enhancing landscape connectivity may lead to a delayed
loss of genetic diversity in metapopulations under
climate change
 But additional measures are likely necessary to ensure
its long-term conservation
Cobben et al. 2012 Landscape Ecology

Well....
 Surely genetic variation that is selected for will not go
extinct...
 And adaptation will improve either the species’ tracking
capabilities or its local survival?

Modelling adaptive genetic diversity
• Up till now neutral genes: not affecting individual
performance
• Set of models allowing selection for traits or a
combination of traits
• Under temperature increase and increased weather
variability

Research question
What is the outcome of the interaction of local evolution
and range shifts when the central populations in the
species range differ genetically from the marginal
populations?

North
neutral model
2000 km
South

North
Central-marginal model
climate GENERALISTS versus
climate SPECIALISTS
2 traits involved:
- experienced maximum habitat
quality
chance to survive
chance to breed
chance to disperse
- thermal tolerance
South

North
Central-marginal model
climate GENERALISTS versus
climate SPECIALISTS
only 2 alleles coding for
generalist and specialist types,
heterozygotes are intermediates
South

temperature speed: 2 km/year, so 1 °C scenario
mutation rate 10e-6

Compare N in time:
metapopulation with specialists and generalists SG, and
2 single genotype metapopulations, G and S

So
• Increase of the generalist numbers is not local evolution
towards increased frequency of better-adapted genotype
but an effect of the range shift
• Range shift causes maladaptation of the species:
specialist and generalist genotypes are in the wrong
location
• This affects the metapopulation size
• Temperature increase can ultimately lead to extinction of
the specialist allele and of the metapopulation
Cobben et al. Global Change Biology, online

Evolution of dispersal
• Range shifts are known to lead to increased dispersal
capacity at range borders

Short wing vs
long wing
Long wing vs
extra-long wing
Thomas et al. 2001 Nature

Hmmm....
• I wonder if this pattern could be partly explained by
founder effects as a result of the range shift....

Research questions
Will range shifts lead to selection for increased dispersal
probability in the metapopulation?
Could this increase be caused by founder effects?
If so, are there adverse effects of this?

Dispersal probability model
Individuals in the model have different
chances of leaving their patch to disperse
6 alleles coding for 11 different levels of
dispersal probability
AA AB BB BC CC CD DD DE EE EF FF
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Initialisation of the model with alleles A-C
Mutations A-F

temperature speed: 2 km/year, so 1 °C scenario
mutation rate 10e-6
Dispersal probability
0.0
0.1
0.2

Dispersal probability
0.0
0.1
0.2

Dispersal probability
0.0
0.1
0.2

This is beneficial for the metapopulation
1 metapopulation 3 metapopulations
with all genotypes with single genotypes

However, availability of genetic
variation changes the pattern

So
• The pattern of genotypes is not caused solely by
selection pressure
• And is thus partly the result of the local availability of the
genotypes
• Under climate change, evolution towards increased
dispersal probability is therefore enhanced by the
founder effect

But this has a drawback when selection
pressure changes

So
• Under stable climate conditions the metapopulation
consisted of only 0.0, 0.1 and 0.2 dispersal probability
individuals
• Under temperature increase we saw selection for the 0.2
dispersal probability genotype
• The increase of this genotype was additionally enhanced
by the founder effect
• But when temperature stabilised the local lack of genetic
variation for dispersal probability caused a slow recovery
of the optimal distribution of genotypes

So this may be partly caused by a
founder effect
Short wing vs
long wing
Long wing vs
extra-long wing
Thomas et al. 2001 Nature

With short-term positive effect
Short wing vs
long wing
Long wing vs
extra-long wing
Thomas et al. 2001 Nature

But possibly a long-term negative effect
Short wing vs
long wing
Long wing vs
extra-long wing
Thomas et al. 2001 Nature

Overall conclusions
1. The founder effect is an important determinant of the
allele composition in newly established populations
under range shift across fragmented habitat.
2. The genetic impoverishment resulting from such founder
events requires considerable restoration time in fragmented
habitat and may consequently be a risk to species’ survival.

Implications
• We are investigating future suitable habitat areas for
species that are currently shifting their ranges
• And basing our estimate of their vulnerability on the
amount of habitat left
• But we don’t know whether these species will actually be
able to reach these new areas
• And if they do, they may look very different
• I therefore suggest that we further investigate whether to
prioritise the collection of genetic diversity in populations
assessed to go extinct soon

Thank you