This document summarizes a workshop on climate change impacts on biodiversity in the San Francisco Bay Area. Key points include: - Climate models project the Bay Area climate to warm significantly by the late 21st century, increasing temperatures, drought conditions, and wildfire risk. - Multiple vegetation models predict future climates will favor shrub and grasslands over forests as some tree species approach the limits of their climate tolerances. - Vegetation transitions are expected to be patchy across the landscape and depend on factors like local propagule sources and disturbance regimes. - Maintaining a diversity of habitats and vegetation types can help support species' ability to shift ranges under climate change.

1. Climate change and biodiversity:
implications for Bay Area conservation
• Bay Area climate: historical patterns and future changes
• Climate impacts on Bay Area vegetation
• Climate heterogeneity and biodiversity
• Management in the face of change

2. Bay Area Climate Change and Protected Areas Workshop
‘The Pepperwood Meeting’
July 19-21, 2010
Left to right: Miguel Fernandez, Jim Thorne, Mary Lee Hannah, Alicia Torregrosa, Stu Weiss, Mike Hamilton, Meg
Krawchuk, Will Cornwell, Nicole Heller, Al Flint, David Ackerly, Lorrie Flint, Ryan Branciforte, Scott Loarie, Dave
Conklin, Jason Kreitler, Sam Veloz, Lisa Micheli, Healy Hamilton, Max Moritz, Morgan Kennedy, Beth Sabo, Jim
Johnstone
Missing: Kirk Klausmeyer, Lee Hannah, Diana Stalberg, Phil Duffy, Karen Gaffney, Adina Merenlender

3. , 3,000 Native Plant Species

4. Biodiversity hotspots in the United States
from Precious Heritage, 2000, Nature Conservancy and NatureServe

5. Global CO2 emissions – IPCC 4th assessment
Raupach et al. 2007 PNAS
A2
B1: stabilizing population,
rapid technology conversion
growing population, high
carbon energy sources

6. Figure 10.4
IPCC 2007, Fig. 10.4
Projections of future temperature – IPCC 4th assessment
A2
B1

7. Bay Area climate
summer max
temperature
precipitation
water deficit
winter min
temperature
PRISM climate layers downscaled to 270 m by Al and Lorrie Flint, USS

8. Temperature increase, averaged over Bay Area
1.6 – 4°C = 3 – 7 °F
Winter minimum temperatures Summer maximum temperatures
Year (end of 30 year periods)
+1.7 °C
+3.9 °C
+1.6 °C
+3.8 °C
preliminary analyses – please do not distribute without permission

9. Climatic Water Deficit:
excess evaporative demand relative to available water
PET depends on
temperature and
insolation
Water availability
depends on
precipitation, soil
storage and runoff
CWD
2001
<775
775 - 800
800 - 825
825 - 850
850 - 875
875 - 900
900 - 925
925 - 950
950 - 975
975 - 1000
!mm/yr"
Climatic
Water Deficit
Annual evaporative
demand
that exceeds
available water
Potential ‒ Actual
Evapotranspiration
courtesy: Al and Lorrie Flint, USGS
see Stephenson 1998 J. Biogeog.

10. Year (end of 30 year periods)
Climaticwaterdeficit(mm)
Water deficits are projected to increase due to evaporative
demand (whether precip goes up or down)
historical
preliminary analyses – please do not distribute without permission

11. The future is expected to be warmer and drier – the
uncertainty is about how fast these changes occur
Summer maximum temperatures (°C)
Climaticwaterdeficit(mm)
historical
preliminary analyses – please do not distribute without permission

12. Distance from the ocean is the primary influence on regional
temperature patterns
Summer max temperatures
Distance to coast or bay
4°C per 10 km
1°C per 10 km
Winter minima:
0.35°C colder per 10 km
Summer maxima
preliminaryanalyses–pleasedonotdistributewithoutpermission

13. B
A
Summer and winter temperatures are negatively correlated
across the Bay Area
Historical
1971-2000
C
A
B
C

14. A
Due to the coastal-inland pattern, rising temperatures
create novel climates throughout the Bay Area
Historical
1971-2000
GFDL A2
2041-2070
A
B
C
B
A
C
preliminaryanalyses–pleasedonotdistributewithoutpermission

15. 1900 1950 2000 2050 2100
17181920212223
!
!
!
!
!
!
!
!
!
!
!
!
!!
!
!
!
!
!
!
!
!
!
!!
!
!
!
!
!!
!
!
!
!
1970-2000 vs 2011-2040
GFDL SRES A2 GFDL SRES B1
GFDL SRES A2 GFDL SRES B1
1970-2000 vs 2011-2040
1970-2000 vs 2041-2070 1970-2000 vs 2041-2070
1970-2000 vs 2071-2100 1970-2000 vs 2071-2100
4 - 5
5.1 - 9.9
10 - 14
4 - 5
5.1 - 9.9
10 - 14
SED
0.2 - 2
2.1 - 3.9
4 - 5
5.1 - 9.9
10 - 14
SED
0.2 - 2
2.1 - 3.9
4 - 5
5.1 - 9.9
10 - 14
SED
0.2 - 2
2.1 - 3.9
4 - 5
5.1 - 9.9
10 - 14
SED
0.2 - 2
2.1 - 3.9
4 - 5
5.1 - 9.9
10 - 14
1970-2000 vs 2011-2040
GFDL SRES A2 GFDL SRES B1
GFDL SRES A2 GFDL SRES B1
1970-2000 vs 2011-2040
1970-2000 vs 2041-2070 1970-2000 vs 2041-2070
1970-2000 vs 2071-2100 1970-2000 vs 2071-2100
4 - 5
5.1 - 9.9
10 - 14
4 - 5
5.1 - 9.9
10 - 14
SED
0.2 - 2
2.1 - 3.9
4 - 5
5.1 - 9.9
10 - 14
SED
0.2 - 2
2.1 - 3.9
4 - 5
5.1 - 9.9
10 - 14
SED
0.2 - 2
2.1 - 3.9
4 - 5
5.1 - 9.9
10 - 14
SED
0.2 - 2
2.1 - 3.9
4 - 5
5.1 - 9.9
10 - 14
B1 A2
Mid-century
End-century
1970-2000 vs 2011-2040
GFDL SRES A2 G
GFDL SRES A2 GF
1970
1970-2000 vs 2041-2070 1970
1970-2000 vs 2071-2100 1970
4 - 5
5.1 - 9.9
10 - 14
SED
0.2 - 2
2.1 - 3.9
4 - 5
5.1 - 9.9
10 - 14
S
SED
0.2 - 2
2.1 - 3.9
4 - 5
5.1 - 9.9
10 - 14
S
1970-2000 vs 2011-2040
GFDL SRES A2 G
GFDL SRES A2 GF
1970
1970-2000 vs 2041-2070 1970
1970-2000 vs 2071-2100 1970
4 - 5
5.1 - 9.9
10 - 14
SED
0.2 - 2
2.1 - 3.9
4 - 5
5.1 - 9.9
10 - 14
SED
0.2 - 2
2.1 - 3.9
4 - 5
5.1 - 9.9
10 - 14
S
Future climates will rapidly
exceed the range of recent
historical variability
analysis and figure: Sam Veloz, PRBO
GFDL SRES A2 GFDL SRES B1
1970-2000 vs 2011-2040
GFDL SRES A2 GFDL SRES B1
GFDL SRES A2 GFDL SRES B1
1970-2000 vs 2011-2040
1970-2000 vs 2041-2070 1970-2000 vs 2041-2070
SED
0.2 - 2
2.1 - 3.9
4 - 5
5.1 - 9.9
10 - 14
SED
0.2 - 2
2.1 - 3.9
4 - 5
5.1 - 9.9
10 - 14
SED
0.2 - 2
2.1 - 3.9
4 - 5
5.1 - 9.9
10 - 14
SED
0.2 - 2
2.1 - 3.9
4 - 5
5.1 - 9.9
10 - 14
SED
0.2 - 2
2.1 - 3.9
4 - 5
5.1 - 9.9
SED
0.2 - 2
2.1 - 3.9
4 - 5
5.1 - 9.9
Year
AverageTmax(°C)
1971 –
2000
±1 s.d. interannual
variability
Berkeley CA
small symbols: annual
large symbols: 30 yr means
preliminary analyses – please do not distribute without permission

16. figure: Johnstone and Dawson 2010 PNAS
Year
map: Loarie and Johnstone unpubl.
please do not distribute without permission
Fog frequency, 2000-2010
Modis satellite imagery

17. Impacts on biodiversity:
observation, experiments, models

18. Several, independent approaches to vegetation modeling agree:
future climates favor shrub and grassland at the expense of forest
High resolution run of the MC1 vegetation model for the Bay Area
Analysis and figures: Dave Conklin, Conservation Biology Institute
present climate MIROC A2, end-century
conifer
hardwood forest
hardwood woodland
grassland
shrubland
please do not distribute without permission

19. Diana Stalberg et al. 2010 PLoS ONE (PRBO) Will Cornwell et al. in prep. (UC Berkeley)
Several, independent approaches to vegetation modeling agree:
future climates favor shrub and grassland at the expense of forest
‘Random forest’ model of CalVeg types
800 m resolution, UCSC regional climate model
Predictive vegetation modeling of Bay Area vegetation
270 m downscaled climate, GFDL mid-century future
forest remaining
forest  woodland
forest  shrubland
please do not distribute without permission

20. source: Bay Area Open Space Council, Ryan Branciforte & Stu Weiss
Bay Area Vegetation Map
Upland Habitat Goals Project
60 cover types
51 natural/semi-natural
30 m mapping units
Remote imagery +
vegetation plots +
expert opinion

21. Altamont Pass, Livermore CA
Vegetation 30m
Spatial downscaling to model vegetation-climate relationships
! "#$ $%#"$
&'()*+,+-.
/!01-203+2)45667893:02('*;,+
:+;40;44<;(0,+*=+-;,<-+
)>+-0%?/01-203+2)450@A:
!"#$%#$$&#'%(")*
+,-.%/0123% %
4567.%/3188% % 1
PRISM climate 800m
0 2.5 51.25
Kilometers
30 Arc Second Grid
Over 1/9 Arc Second DEM
1/9 sec DEM Mask
Meters
High : 691
Low : 94
Elevation 3m

22. 800 m
Topoclimate influence on vegetation

23. Modeling Bay Area Vegetation
Desired features:
1) small grain model with a realistic representation of
topography (30 m)
2) simultaneous model of all vegetation types
3) comparison with documented vegetation transitions

24. Predictive layers
1) Seasonal water deficit (270 m)
Al and Lorrie Flint (USGS)
2) Potential annual insolation
(annual, 30 m)
3) Min Temp (270 m downscaled
from PRISM)
4) Max Temp (270 m downscaled
from PRISM)
5) Wind (100 m)
6) Soil Depth (STATSGO)
Modeling Bay Area Vegetation
!"##$%& '()*+$%&"#,
DEDDEFDEGDEHDEIJED
Vector of probabilities for
each veg type in each
pixel
redwood
Doug fir
Montane hardwood
Baccharis
California bay
live oak
Other
multinomial logistic
regression
W. Cornwell et al. in prep.please do not distribute without permission

25. Relative probability of
vegetation transition
(GFDL A2, mid-century vs.
present)
The vulnerability of vegetation types is very patchy:
high probabilities of change occur where vegetation patches are
near the edge of their climate envelope
W. Cornwell et al. in prep.please do not distribute without permission

26. Regional and topographic shifts in vegetation types
Blue oak example
0 20 40 60 80
0.000.050.100.150.20
Distance from coast or bay (km)
FrequencyofBlueOak
050001000015000
Insolation
north-facing south-facing
Distance from coast or bay (km)
present
A2 mid-century
present
A2 mid-century
please do not distribute without permission

27. !"#$
%&'(
)&"*+,-.*&
!&/(
)0&1*2$
),&23&"
45-#6
),&23(
7*8$(
%&'
9,#:
;*-(
<-&22"&.=
>$=?,,=(
@&""$/(
%&'
A,-30
B&23
4,#30
C$.*.2#"&
DEFGEFEEFGEFHEFI
)0&.:$(*.(&-$&(,55#J*$=
(KJ-,J,-3*,.(,+(-$:*,.L
(
(
GFDL A2 2041-2070 versus present
pleasedonotdistributewithoutpermission

28. Native vegetation transitions vs. alien invasions
vegetation transitions depend on:
1) mortality of existing mature plants
2) propagule sources for new species
source: Larry Workman QIN, Panoramio.com
?

29. Agents of mortality: Fire
Historical probability of fire
1950-2003
(climate-driven model) 2010-2039 (A2) 2070-2099 (A2)
16 GCM ensemble (A2 scenario):
change relative to historical period
Figures: courtesy Meg Krawchuck and Max Mortiz, UC Berkeley
Historical: Parisien and Moritz 2009 Ecol. Monogr.
Futures: Moritz et al. in reviewplease do not distribute without permission

30. Agents of mortality: Disease
source: UC Davis; http://www.sciencedaily.com/releases/2007/08/070815145316.htm
Sudden oak death
source: Center for Invasive Species Research
UC Riverside

31. Agents of mortality: Drought and pests
piñon pine mortality
credit: Craig Allen, USGS

32. Documented vegetation transitions
(coastal CA)
Cornwell, Sandel and Ackerly, unpublishedplease do not distribute without permission

33. Documented vegetation transitions
vs. projected transitions
Cornwell, Sandel and Ackerly, in prep.please do not distribute without permission

34. Local diversity provides seed
sources for vegetation shifts

35. Heterogeneous landscapes support a greater diversity of
vegetation types
analysis and figures: Jason Kreitler, USGS and
Nicole Heller, Climate Central
UHG Vegetation Layer # of veg classes/12km cell Climatic water deficit
Legend
12km
<VALUE>
0 - 150
151 - 400
401 - 500
501 - 600
601 - 700
701 - 800
801 - 900
901 - 1,000
1,001 - 1,100
1,101 - 1,257
Legend
Variety
Value
High : 26
Low : 1
N 0 25 5012.5 Kilometers
CWD
# veg types
per 12km cell
Veg map
100 300 500 700
0510152025
RANGE of CWD
NumberofVegTypes
please do not distribute without permission

36. Loarie et al. 2009 Nature
Velocity of climate change:
how fast will populations need to move to offset rising temperature?
rate of climate change (°C/yr) ÷ spatial climate gradient (°C/km) = velocity (km/yr)

37. Eradication of invasives is
more important than ever in
the face of changing climates!
sources: nps.gov, cal-ipc.org

38. thermal refugia
Implications for conservation and management
Large, climatically heterogeneous reserves are critical to
maintain diverse local species pools as propagule sources
for potential vegetation transitions
In restoration and revegetation projects, a diverse pool of
species and genotypes may enhance success in the face of
uncertain future climate
Implications for strategic acquisition priorities....
Stu and Ryan – next talk!