A Spatially Explicit Assessment of Climate
Change Genetic Risk to 200 Forest Tree Species
Kevin M. Potter
William W. Hargrove
Frank H. Koch
Genetic Resource Management
Climate Change Workshop
March 3, 2010
Overview
1) Outline the potential genetic impacts of climate
change on forest tree species
2) Describe a US Forest Service project to identify species
and populations most genetically at risk from climate
change
3) Explain the range modeling tools we’re using to
identify such species and populations
4) Present preliminary results from Eastern and Western
species
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Plant responses to climate change
1) Toleration/adaptation
2) Shifting range
3) Population extirpation
All could have negative
genetic consequences
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Possible genetic consequences
1) Toleration/adaptation

Strong selection could reduce genetic variation
2) Shifting range

Founder effects, loss of trailing edge populations
3) Population extirpation

Potential loss of unique genes and novel gene
combinations
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Why do we care about genetics?
 Genetic variation =
evolutionary potential to
adapt to changing
environmental
conditions
 Genetic degradation may
increase susceptibility to
other stressors (pests,
pathogens, changing
climate, etc.)
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Project background

2008: the Forest Health Monitoring (FHM)
program identified climate change as a priority
area
 FHM mission: determine status, changes and trends
in indicators of forest health

2009: FHM sponsored an assessment of the
genetic risk of climate change on North
American forest tree species
 Cooperative agreement between NC State University
and the Eastern Forest Environmental Health
Assessment Center (EFETAC)
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Project objectives
1) Forecast location and quality of habitat for ~200 North
American forest tree species under multiple climate change
scenarios
2) Identify existing populations most at risk from climate
change, based on distance to future habitat
3) Assess susceptibility of species and their populations to
genetic degradation, based on climate change impacts and
species biology
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Project applications
1) Climate change monitoring efforts
(species and populations to watch)
2) On-site and off-site gene
conservation
3) Restoration efforts
4) Future refugia for conservation
Collecting Fraser fir cones at Mount
Rogers, Va., for ex situ gene
conservation
5) Molecular marker studies for atrisk species
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
1) Future habitat prediction
1) Predict changes in appropriate habitat for a large
number of species (~200 across North America)
2) Flexible data inputs on species occurrences (Forest
Inventory and Analysis and other sources)
3) High-resolution (applicable at population level)
4) Global (doesn’t end at the border)
5) Incorporates pertinent environmental variables for
determining plant distributions
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Multivariate Spatio-Temporal Clustering

MSTC developed by Bill
Hargrove of US Forest
Service and Forrest
Hoffman of ORNL

Clusters each 4 km2 pixel
across the globe into
one of 30,000 unique
“ecoregions” based on
16 spatial environmental
variables
(Hoffman et al. 2002)
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Climate change predictions

MSTC generates maps of current
conditions, and expected
conditions in 2050 and 2100



Four scenarios: Hadley and PCM, high
and low emissions
Using tree occurrences from FIA data
(though can use other data)
Predicts current ecoregions
suitable for species, then
compares future ecoregions to
current ecoregions
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Environmental variables

Soils






Plant-available water capacity
Bulk density of soil
Kjeldahl soil nitrogen
Organic matter in soil
Precipitation





Temperature

In the coldest quarter

In the warmest quarter


Diurnal temperature difference
Biotemperature

Solar insolation

Ratio of precipitation to potential
evapotranspiration
Topography


In the driest quarter
In the wettest quarter
In the warmest quarter
In the coldest quarter
Compound topographic index
Growing season

Length in integer months
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Current
ecoregions
2050 ecoregions
under Hadley B1
scenario
2100 ecoregions
under Hadley B1
scenario
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Global range modeling
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Forest Inventory and Analysis data
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Forest Inventory and Analysis data
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Shumard oak
Shumard oak (Quercus shumardii)
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Shumard oak
Current prediction
Fundamental
niche
Realized
niche
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Hadley B1, 2100
Table Mountain pine
Table Mountain pine (Pinus pungens)
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Table Mountain pine
Current prediction
Fundamental
niche
Realized
niche
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Hadley B1, 2100
2) Population extirpation risk

What portions of tree species ranges are most at risk
of extirpation because of their distance to expected
future habitat?


Assumption: greater distance to the nearest future refuge =
greater risk of population-level extirpation
Measuring the straight-line Minimum Required
Movement (MRM) distance

From each 4 km2 pixel in current species niche to the nearest
suitable habitat in the future
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
American sycamore
American sycamore
(Platanus occidentalis)
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Range comparison over time
American sycamore
Current prediction
Hadley B1, 2100
New habitat in 2100
Habitat overlap, now and 2100
Current habitat gone in 2100
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Distance to future suitable habitat
American sycamore
Current prediction
Hadley B1, 2100
Overlapping and new habitat
Short distance to 2100 habitat
Moderate distance to 2100 habitat
Long distance to 2100 habitat
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Tulip-poplar
Tulip-poplar
(Liriodendron tulipifera)
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Distance to future suitable habitat
Tulip-poplar
Current prediction
Hadley B1, 2100
Overlapping and new habitat
Short distance to 2100 habitat
Moderate distance to 2100 habitat
Long distance to 2100 habitat
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Sweetgum
Sweetgum
(Liquidambar styraciflua)
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Distance to future suitable habitat
Sweetgum
Current prediction
Hadley B1, 2100
Overlapping and new habitat
Short distance to 2100 habitat
Moderate distance to 2100 habitat
Long distance to 2100 habitat
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Loblolly pine
Loblolly pine
(Pinus taeda)
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Distance to future suitable habitat
Loblolly pine
Current prediction
Hadley B1, 2100
Overlapping and new habitat
Short distance to 2100 habitat
Moderate distance to 2100 habitat
Long distance to 2100 habitat
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
American chestnut
American chestnut (Castanea
dentata), Pilot Mountain, N.C.
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Distance to future suitable habitat
American chestnut
Current prediction
Hadley B1, 2100
Overlapping habitat
Short distance to 2100 habitat
Moderate distance to 2100 habitat
Long distance to 2100 habitat
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Canyon live oak
Canyon live oak
(Quercus chrysolepis)
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Distance to future suitable habitat
Canyon live oak
Current prediction
Hadley B1, 2100
Overlapping and new habitat
Short distance to 2100 habitat
Moderate distance to 2100 habitat
Long distance to 2100 habitat
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Sitka spruce
Sitka spruce
(Picea sitchensis)
www.conifers.org
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Distance to future suitable habitat
Sitka spruce
Current prediction
Hadley B1, 2100
Overlapping and new habitat
Short distance to 2100 habitat
Moderate distance to 2100 habitat
Long distance to 2100 habitat
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Mountain hemlock
Mountain hemlock
(Tsuga mertensiana)
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Distance to future suitable habitat
Mountain hemlock
Current prediction
Hadley B1, 2100
Overlapping and new habitat
Short distance to 2100 habitat
Moderate distance to 2100 habitat
Long distance to 2100 habitat
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Availability of maps
Niche maps, global
Niche maps, continental U.S.
Migration distance, continental U.S.
Migration distance, global
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Still to come…

Identifying “lifeboat” areas that
preserve multiple species from
broad regions

Quantifying the suddenness of range
shifting over time

Identifying corridors of possible
movement from current to future
habitat

Incorporating barriers to movement
(urban areas, lakes, etc.)

Quantifying uncertainty: consensus
within and across species under
different scenarios
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
+
3) Genetic risk assessments

Risk of genetic degradation,
given climate change and
the biology of a species


Population level
Synthesize maps with
knowledge about species
biology, population
structure, existing threats

How much do we worry
about at-risk areas on maps?
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Assessing the genetic implications
 Bayesian Belief Network
approach incorporating
expert opinion
 Population geneticists,
ecologists, species specialists
 Case studies: ponderosa pine,
eastern hemlock
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
We need your help…
 Suggesting tree species at
climate change risk
 Identifying coordinate data for
rare/non-U.S. locations
 Assessing the genetic
implications of the results
 Understanding relationships
among biological attributes and
risk of genetic integrity loss
 Assessing results for individual
species
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.
Thoughts? Please contact me:
kevinpotter@fs.fed.us
(919) 549-4071
Thanks to:

EFETAC: Danny Lee, Bill
Bechtold, Kurt Riitters

USDA Forest Service Forest
Health Monitoring Program:
Borys Tkacz

NC State: Barb Conkling, Fred
Cubbage, Mark Ambrose

ORNL: Forrest Hoffman
Forest Health Monitoring Research Group,
Research Triangle Park, N.C.