UC BERKELEY GEOGRAPHY SEMINAR DECEMBER 10, 2008

EFFECTS OF CLIMATE CHANGE ON FOREST FIRES OVER NORTH AMERICA AND IMPACT ON U.S.

OZONE AIR QUALITY AND VISIBILITY

UC BERKELEY GEOGRAPHY SEMINAR

DECEMBER 10, 2008

Rynda Hudman 1,2, Dominick Spracklen 1,3, Jennifer Logan1, Loretta J. Mickley1, Maria Val Martin1,4, Shiliang Wu1,5, Rose Yevich1, Alan Cantin6, Mike Flannigan6, Tony Westerling7

Affiliations: 1 School of Engineering, Harvard 2 Now at UC Berkeley 3 Now at University of Leeds 4 Now at Barcelona Supercomputing Center 5 Now at Michigan Tech

6 Canadian Forest Service 7 UC Merced

• Releases 1-4 Pg C / yr (~30-50% of the fossil fuel source) • Accounts for 2/3 of the variability in CO2 growth rate between 1997 and 2001 • 20-60% of the global organic carbon aerosol (particulate) emission, 30% of the black carbon (soot) emission• Potential for climate feedbacks

•Impacts ozone/aerosol air quality, visibility, human health

WHY DO ATMOSPHERIC SCIENTIST CARE ABOUT WILDFIRES?

[Bowman et al., 2009]

TROPICS DOMINATE FIRE ACTIVITY BUT NORTH AMERICAN RECORD PUNCTUATED BY LARGE FIRE YEARS

Mean area burned:

~3 million hectares

2x size of Connecticut

Large fire years increase emissions by X10

NORTH AMERICAN FIRES AFFECT ATMOSPHERIC COMPOSITION ON A HEMISPHERIC SCALE

http://asl.umbc.edu/pub/mcmillan/www/index_INTEXA.html

In 2004, a blocking ridge set up over Canada and Alaska creating one of the largest fire seasons on record.

56000 ha, June 8-22, 2002 30 miles from Denver and Colorado Springs EPA 24-hr standard = 35 µg/m3, and annual standard = 15 µg/m3.

Colorado Department of Public Health and EnvironmentVedal et al., Env Res, 2006

June 8, 2002 June 9, 2002 PM10 = 372 μg/m3

PM2.5 = 200 μg/m3

PM10 = 40 μg/m3

PM2.5 = 10 μg/m3

Hayman fire caused worst air quality ever in Denver

LOCAL EFFECTS OF WILDFIRE EMISSIONS

WILDFIRE DRIVES INTERANNUAL VARIABILITY IN ORGANIC CARBON AEROSOL IN THE SUMMER

[Spracklen et al., 2007]

Model gives same variability as observed OC in summer at IMPROVE sites in the West

OC contribution to total fine aerosol: 40% in low fire years 55% in high fire years

same fires every year

[Spracklen et al., 2007]

PRESENT DAY FIRE IMPACTS ON OZONE

Ozone enhancement from NA biomass burning 0-2 km

Simulated July 2004 mean Max enhancement during July 15-24 2004

8-hr max ozone air quality standard in the United States = 75 ppbv

[Hudman et al., 2009]

CLIMATE DRIVES FIRE ACTIVITY OVER NORTH AMERICA

Canadian Fire Weather Index Model

TemperatureRainfallWind speed Relative Humidity

Other factors:Large Scale circulation

Fuel availability

Ignition Source

Fire Suppression

OBSERVED INCREASE IN WILDFIRE ACTIVITY OVER NORTH AMERICA DUE TO CLIMATE CHANGE?

[Gillett et al., 2004]

5 year means

Area burned in Canada has increased since the 1960s, correlated with temp. increase.

[Westerling et al., 2007]

Increased fire frequency over western U.S. in recent decades – related to warmer temp., earlier snow melt.

FUTURE AREA BURNED

OBSERVED AREA BURNED

CHEMICAL TRANSPORT MODEL

Climate ModelOutput

OBS WEATHER & FUEL MOISTURE/ FIRE SEVERITY

Yearly Area Burned = C1X1 + C2X2 + … + C0

Emissions

PREDICTING THE IMPACT OF FUTURE CLIMATE CHANGE ON WILDFIRE AND AIR QUALITY

1. DEVELOP RELATIONSHIPS BTWN CLIMATE AND ANNUAL AREA BURNED

2. CLIMATE MODEL OUTPUT PREDICT FUTURE AB 3. PRED. FUTURE AIR QUALITY

Combine ecoregions of similar vegetation and topography

Use observed meteorology from surface weather stations (USFS) FWI

I. WESTERN U.S. ECOREGIONS & MET USED IN REGRESSION

(Spracklen et al., 2009)

WHERE ARE THE FIRES IN THE WESTERN U.S.?

Mean area burned (1º x 1º grid) in 1980-2004 (Westerling, UC Merced) Mean fuel consumed

Large areas burned in CA and the southwest, but fuel burned is greater in forest than in shrub ecosystems

The Pacific North West and Rocky Mountain Forests are most important for biomass consumption and emissions.

(Spracklen et al., 2009)

52%

24%49%

37%

48%57%

R2 of Area Burned regressions

Regressions ‘capture’ 24 – 57% of the interannual variability in area burned over western US. Temperature contributes 80-90% of the regression in forested regions.

Are

a B

urn

ed (

ha)

Year

Are

a B

urn

ed (

ha)

PREDICTING WILDFIRE OVER THE WESTERN U.S.

(Spracklen et al., 2009)

Temperature

Predicted met Changes Temp. +1-3ºC across West Rainfall and RH increase slightly Wind speed decreases slightly

CHANGES IN MAY-SEPT TERMPERATURE (2000 – 2050)

GISS GCM3 A1B Scenario - CO2 concentrations reach 522ppm

(Spracklen et al., 2009)

Pacific Northwest US

Rocky Mountain Forests

78% increase

175% increase

Observed area burned Predicted area burned

Predicted area burned for 1995-2004 does not match observed areas on a yearly basis, as it is based on GCM output, but 10 year mean is the same.

PREDICTED INCREASE IN AREA BURNED

(Spracklen et al., 2009)

Pacific Northwest US

Rocky Mountain Forests

Predicted area burned for 1995-2004 does not match observed areas on a yearly basis, as it is based on GCM output, but 10 year mean is the same.

PREDICTED INCREASE IN AREA BURNED

(Spracklen et al., 2009)

+ 1-3K

Climate change projected to cause a 90% increase in biomass consumed and 40% increase in OC concentrations by 2050.

Change in wildfire biomass consumption

Change in surface OC aerosol (Jun-Aug)

Emissions

Chemical Transport Model

Δbiomass consumption = + 90%

Δsurface OC aerosol = + 40%

FUTURE WILDFIRE AND PARTICULATE AIR QUALITY

(Spracklen et al., 2009)

AB + FUEL

Present day fires in black, 1996-2000Future fires in red, 2046-2050

OC increases by 40%, EC increases by 20% (not shown).

For OC, 75% of increase is from fire emissions, 25% from higher biogenic emissions in a warmer climate.

FUTURE WILDFIRE AND PARTICULATE AIR QUALITY

(Spracklen et al., 2009)

(Hudman et al., in prep)

Consistent with these results, recent observational estimates of regional enhancements of 2 ppbv for each 1 million acres burned [Jaffe et al., 2008]

5 Years Future (2046-2050) vs. 5 Years Present (1996-2000)

PREDICTED JULY MEAN MAXIMUM 8-HR OZONEperturbation from fires doubles

•Regressions capture much of the variability in annual area burned over the western U.S. (24-57%). Temperature is the key predictor.

• 2050 climate change (A1B) is predicted to increase annual mean area burned over western U.S. (+54%) 90% increase in biomass consumed relative to the present-day driven by 1-3K increase in temperature.

• Future fires drive a 40% increase in organic carbon aerosol over the western US and a 1-3 ppbv enhancement (doubling fire enhancement) in summertime afternoon ozone.

SUMMARY WESTERN U.S.

II. BOREAL ECOREGIONS & MET USED IN REGRESSION

[Stocks et al., 1999][French et al., 2003]

Combine ecoregions of similar vegetation and topography (12 ecoregions)

Alaska wx stations (USFS) & Canadian wx stations (CFS)(Hudman et al., in prep)

[105 ha]

Largest Area Burned over Plain

regions

Jul 1 – Aug 15 2004 Anomaly

Strong Alaskan Ridge record fires

+60

SUMMER 2004: 500hPa GEOPOTENTIAL HEIGHT

Height of pressure level above mean sea level

Strong ridges are accompanied by warm and dry weather conditions at the sfc

(Hudman et al., in prep)

CANADIAN FIRE WEATHER INDEX MODEL

2/3 day 15 day 52 dayDrying time

Severity Rating

Severity Rating is a combination of drought and fire spread potential

REGRESSIONS CAPTURE VARIABILITY IN REGIONS WITH LARGEST AREA BURNED (15-62%)

GPH was chosen over temperature

ALASKA/CANADA SUMMARY: 2-3 predictors chosen per region

Most Common Predictors: •Monthly/Seas. 500 mb GPH Anomaly (Max contributor 7/12 ecoregions)•Max/Mon./Seasonal Severity Rating (Max contributor in 3/12 ecoregions)

More influenced by fire suppression and human caused fires

(Hudman et al., in prep)

Regressions capture 71% of the variability in Canada and Alaska

About as good a non-linear regression which use many more variables

PREDICTING WILDFIRE OVER CANADA AND ALASKA

- - - national totals for Canada (not included in regression) + Alaska

(Hudman et al., in prep)

DOES RAIN OFFSET TEMPERATURE/GPH INCREASE?GISS simulated May – August 2046-2055 vs. 1996-2005

(Hudman et al., in prep)

June 500mb anomaly over Fairbanks, Alaska (1940 – 2006)

GISS Mean 1999-2008 : -14 m 2045-2054 : 5 m

[Fairbanks GPH Courtesy of Sharon Alder, BLM]

DOES RAIN OFFSET TEMPERATURE INCREASE?GISS simulated May – August 2046-2055 vs. 1996-2005

(Hudman et al., in prep)

Seasonal Severity Rating

Rain

Dry spell length important…GISS suggests decreased dry spell length, likely very model dependent

MOST GCMS PREDICT INCREASED SUMMERTIME PRECIPITATION

Dry spell length important…GISS suggests decreased dry spell length, likely very model dependent

(IPCC, 2007, Ch 11)

A1B 1980-1999 vs. 2080-2099

Predicted Summer ppt Change # of models showing increased ppt

34% increase over Alaska, 8% (-34 to +118%) increase in Canada. Large regional variability. Seems consistent with recent study by Meg Krawchuck (UCB)

2000-2050 change in area burned

PREDICTED CHANGE IN AREA BURNED

DSR dominates

GPH dominates

Combination

(Hudman et al., in prep)

(Hudman et al., in prep)

• Distribute annual area burned by month ( ha/month)

• Randomly place AB w/in ecosystem into 1°x1° (based on current fire size stats)

• Combine with fuel consumption which varies throughout season based on fuel moisture + make assump. severity (kg DM/ha)

• Combine with emission factors ( g species/kgDM)

• Assume 20% of emissions in FT (Maria Val Martin MISR work)

• Input into GEOS-Chem CTM (w/ GISS met) future air quaity

PREDICTING FUTURE AIR QUALITY

(Hudman et al., in prep)

PRESENT DAY SURFACE OZONE ENHANCMENT JUL-AUG

Fires predicted to enhance 8-hr max ozone by 3-10 ppbv, 1-4 ppbv reaching Midwest U.S.

(Hudman et al., in prep)

CHANGE IN SURFACE OZONE ENHANCMENT JUL-AUG

Doubling of enhancement over Alaska, 1-2ppbv increase over populated Quebec cities and Midwest (20-40% increase)

A decrease of ozone toward the Arctic

PERCENT CHANGE IN SURFACE OC/EC JUL-AUG Preliminary Result

(Hudman et al., in prep)

[%]

Transport of Black Carbon aerosol to the Arctic decreases by 40%

FUTURE WORK

• Examine change in extreme events using current simulations and Regional modeling (U. Houston)

• Implement plume rise model into GEOS-Chem (Maria Val Martin)

• Improve regressions of desert southwest using PDSI (Harvard)

• Update Canada/Alaska regressions LFDB when avail.

•Do an envelope study of GCM response to Canada/Alaska regressions to look at variability in response (Harvard)

•Impacts of new understanding of NOx emission factors on ozone response (Harvard, Anna Mebust UCB)

Thanks for your attention!

•Regressions capture much of the variability in annual area burned over Alaska (53-57%), and Canada (15-62%). Key predictors : 500 mb GPH anomaly & severity rating.

• 2050 climate change (A1B) increases annual mean area burned: Alaska (+34%) relative to the present-day, but unlike most previous studies little change over Canada as a whole (8%), but varies regionally (-34 - + 118%) due to increases in GCM precipitation vs. temperature (scenario/GCM dependent).

•Present day ozone enhancements due to wildfire 3-10 ppbv over Canada and Alaska. Future fire increases range from -2 - +4 ppbv. Large decreases of BC toward the Arctic.

SUMMARY CANADA/ALASKA

1. WILDFIRE PREDICTION MODEL

Daily forest moisture/fire danger

parameters

Area burned database

Aggregate area burned to

ecosystem

Canadian Fire Weather Index

System

Predictors of Area Burned

Linear stepwise regression

Observed daily Temperature, Wind

speed, RH, Rainfall,

500hpa GPH anom. (Canada/Alaska)

Stepwise linear regression between meteorological/forest moisture variables & area burned[Flannigan et al. 2005]

IMPLICATION OF RISING OZONE BACKGROUND FOR MEETING AIR QUALITY STANDARDS

0 20 40 60 80 100 120 ppb

Europe AQS(seasonal)

U.S. AQS(8-h avg.)

U.S. AQS(1-h avg.)

Preindustrialozone

background

Present-day ozone background at

northern midlatitudes

Europe AQS (8-h avg.)

EPA policy-relevant background (PRB) : U.S. surface ozone concentrations that would be present in absence of North American anthropogenic emissions

PRB is not directly observable and must be estimated from global models

GEOS-Chem GLOBAL MODEL OF TROPOSPHERIC CHEMISTRY

• Driven by NASA/GEOS assimilated meteorological data with 6-h temporal resolution (3-h for surface quantities)

• Horizontal resolution of 1ox1o, 2ox2.5o, or 4ox5o; 48-72 levels in vertical

• Detailed ozone-NOx-VOC-PM chemical mechanism

• Applied by over 30 research groups in U.S. and elsewhere to a wide range of problems in atmospheric chemistry

• Extensively evaluated with observations for ozone and other species (~200 papers in journal literature)

http://www.as.harvard.edu/chemistry/trop/geos

Mean Asian surface pollution enhancement (GEOS-Chem)

Global Carbon Emissions

49% Africa

13% South America

11% equatorial Asia

9% boreal forests

6% Australia

Short-lived Pollutants Affect Climate and Air Quality

Regulations of short-lived species that improve air quality and warm the planet (BC) present a “win-win” situation, while regulations of short-lived species that reduce cooling and improve air quality (SO2) present a “win-lose” situation.

[IPCC, 2007]

ACCOUNTING FOR DRIZZLY GCM

ObservationsGISS Present DayGISS Future

-------- Corrected (GISS – 1.5 mm)_____ Uncorrected

Freq

uenc

y

Dryspell Length (days)

An increase from current conditions (red) is indicated by a PΔ greater than unity, little or no change (yellow) is indicated by a PΔ around unit, and a decrease (green) is indicated by a PΔ less than unity. Panels show the mean PΔ for the ensemble of ten FIRENPP (A–C) and FIREnoNPP (D–F) sub-models. Climate projections include 2010–2039 (A, D), 2040–2069 (B, E) and 2070–2099 (C, F).