Environmental Impacts of Wildfires

Shelly MillerMike Hannigan

Jana MilfordMike Kleeman

David HendersonMike Robert

CU Mechanical Engineering Graduate Student SeminarFall 2002

Fuelcompound class

% of dry

weight

amino acids 0.6

nucleic acids 0.2

proteins 7.6

carbohydrates 38

monosaccharides 1.9

sucrose 3.1

cellulose 21

hemicellulose 9.8

pectin 1.9

lipids 5.3

fatty acids 4.0

glycerols 0.3

other 1.0

lignin 23

phenolics 20

organic acids 3.5

minerals 1.5

Molecular composition of a pine tree.

The “glue” is a heterogeneous polymer termed lignin.

After a cell stops growing, the cell creates a secondary wall which gives the plant it’s rigid structure. To the right is a depiction of this secondary wall, minus the “glue” that fills the empty space. This “glue” gives wood it’s strength.

Lignin monomers

carbohydrate base

During the growth stage, all plant cells form only a primary cell wall.

The cell wall is composed exclusively of cellulose.

Cellulose is a linear polymer of glucose, which is a C6 carbohydrate (a.k.a., sugar).

litter

Litter is preferentially composed of leaves and needles, which have less cell wall per unit mass and more surface area per unit mass. Plant surfaces are composed of the cuticle, which is a heterogeneous polymer of fatty acids, phenolics, long chain alcohols, fatty acid derivatives and odd-numbered n-alkanes. In addition, the N:C ratio is higher in leaves and needles so the ground fuel will have a relative emission of N.

duff

Duff has a higher percentage of the more stable plant components, such as metals.

humus

Humus is more homogenous and has even higher ratios of metals.

complete combustion products

CO2

CO2

CO2

CO2

CO2

CO2

CO2

CO2

CO2

CO2 CO2

CO2

H2O

H2O

H2OH2O

H2O

H2O

H2O

heat

heat

heat

CO2, H2O, and heat

The heat does work by making the plume rise. Thus, the heat is lost

relatively quickly.

Heat has only local effect.H2O has local/regional effect.

H2O

H2O

H2O

H2O

In the plume, the H2O vapor cools as the heat

is lost. Some of this H2O will condense and

grow rain drops.

CO2 is removed by reaction with OH (very slow) or by

gradual partitioning into water drops that are rained out.

CO2 has global effect.

CO2

CO2

CO2

CO2

CO2

CO2

CO2

CO2

CO2CO2

CO2CO2

CO2 emission rates measured in the plumes of

wildfires range from1.3-1.8 kg CO2 per kg fuel.

Globally, CO2 emission from wildfires are 2-5 Pg/yr which is 30-80% of fossil

fuel CO2 emission estimates.

incomplete combustion products

trace gasesCO, CH4, hydrocarbons, NOx, NH3, HCN, organic N

VOCsacetaldehyde, acetone, methanol, vinyl acetate, …

particlessoot, minerals, organic mixtures

Cause:• Lack of oxygen.• Complex fuels.• Poor heat transfer.

CO emission rates vary with combustion condition, ranging from 5 to 25% of

CO2 emissions.

CO + CO2 emission account for 85 to 99% of wildfire carbon emissions.

CO and CO2 have similar removal mechanisms, except OH + CO is

much faster.CO/CO2 ratios decrease slowly as a

plume ages. Still, CO emissions have a large spatial scale effect.

Globally, CO emissions from wildfires account from

10 to 50% of total CO emissions.

COCO

CO

CO

CO

CO

CO

CO

CO

CO

CH4

CH4

CH4

CH4

CH4 is also a greenhouse gas, actually more efficient than CO2.Wildfire emissions account for

25% to almost all the total global CH4 emissions.

CH4 is very stable and only removed via oxidation by OH or O3.

CH4 is global.

HC

HC

HC

NOx

NOx

NOx

O3

O3

O3

Just as in urban areas,NOx and hydrocarbons (HC)

react to form ozone.Due to biomass burning, O3 in the tropics is 30% greater than

background

reduced N

reduced N

reduced N

Nitrogen is key component of the nutrient cycle.

Reduced nitrogen emitted from biomass burning is

causing global redistribution of nitrogen.(from mostly tropical to

global)

CO

othe

r in

com

plet

eco

mbu

stio

n pr

oduc

ts

Fine particles

Mor

talit

y ris

k

6 cities study

6 cities study looked at acute effect of exposure. More recently, researchers have

seen correlations with fine particle levels and mortality for chronic exposure.

flaming

smoldering

total particle mass emission rate

soot

mas

s em

issi

on r

ate

particles

incomplete combustion products

Water drops come from condensing water vapor onto particle. We call the particles cloud condensation nuclei (CCN). A particles ability to act as

CCN depends on the composition and size. The more CCN, then more cloud drops per cloud. The more cloud drops per cloud, the more the

cloud scatters light. This will change the earth’s energy budget.

Soot (black or light absorbing particles) are not excellent CCN; however, they still cause energy budget effects. They absorb incoming solar radiation, heating up the upper atmosphere but cooling the earth.

What matters:size and composition and amount

primary product

Reported emission rates of levoglucosan from biomass

combustion range from 2-12% of the particle mass.

Particle Composition

Other anhydrous sugars have been seen that correspond to

constituents of hemicellulose and pectin, but not to the same degree.

pyrolysis

Lignin pyrolysis products also abound.Prominent metals:

• Potassium• Sulfur• Chlorine

Terpenoids

Functionally and chemically diverse group of plant compounds.•Hormones•Oils and resins•Pigments•Sterols

Terpenoids are polymers with

commonality of the same basic unit – isoprene.

During wildfires,these terpenoids can be

pyrolyzed and the products emitted

orthey can be directly

emitted by volatilization.

0

10

20

30

40

50

50 250 450 650 850

Particle Diameter (nm)

Pe

rce

nt

of

PM

1.0

Ma

ss

(%

)

total masscellulose pyrolysis productslignin pyrolysis productsresin emissions

0

1

1

2

2

50 250 450 650 850

Particle Diameter (nm)

rati

o o

f c

om

po

nd

pe

rce

nt

to t

ota

l p

erc

en

t

total masscellulose pyrolysis productslignin pyrolysis productsresin emissions

COMPOUND CONCENTRATIONLignin Pyrolysis Products

0

50

100

150

200

250

300

50 250 450 650 850

particle diam eter (nm )

co

mp

ou

nd

co

nc

en

tra

tio

n (

ng

/mg

sa

mp

le)

vanillin guaiacyl acetone

coniferyl aldehyde syringaldehyde

1 10 100 1000 10000 100000 1000000

compound concentration (ng/mg)

resin emissions lignin pyrolysis products cellulose pyrolysis products

Size and Composition

Photo by: Bryan Day, Idaho, 2000

Research Objectives• Measure residential indoor

and outdoor PM2.5 (particulate matter smaller than 2.5 m) while wildfire smoke is present

• Determine the effectiveness of recommended indoor exposure mitigation measures– Keeping windows

closed– Using portable air

cleaners

History and Future of Fire• In the 1930s, 39 million acres burned

naturally per year• In the1980s 4.2 million acres burned per

year (NIFS 2002)• Cost of a prescribed burn ~$75/acre• Cost of a wildfire ~$1000/acre (USFS 2002)

Currently, the annual target for prescribed burns by 2005 is greater than 5 million acres (OAR 1998), compared to fewer than 700,000 acres burned from 1984–1994 (APCD 2000)

PM Measurements near Fires

• In Hoopa Valley, CA 1999 PM10 levels exceeded 350 g/m3 for over a week and exceeded the EPA’s hazardous levels of 425 g/m3 for two days (Mott 2001)

• In Hamilton, MT 2000 PM10 levels of 300–600 g/m3 were experienced during multiple days and during a 1-h period concentrations exceeded 999 g/m3 (Acheson, 2001; Ward and Smith, 2001)

• In Indonesia, 1997 a 2-mo period of uncontrolled wildfires produced total suspended particles levels up to 15 times the established limit of 260 g/m3 (WHO, 1999).

Health Effects

• Significant associations between outdoor PM concentrations during fires and health effects are:– increased hospitalization and visits to emergency room– increased respiratory symptoms– exacerbation of asthma– decreased lung function

• These impacts have been observed primarily in the elderly, the very young and in individuals with pre-existing respiratory and/or cardiovascular illness (WHO, 1999; Mott 2001)

Experimental Design

•Locate wildfire producing smoke that will impact local residents•Identify and recruit two homes to be studied•Install air cleaners in one of the 2 homes

•Test the effectiveness of keeping windows closed and air cleaner operation on the indoor air quality

•Monitor indoor and outdoor PM2.5 concentrations at both homes during fire

Recruitment Methods• Fire located by monitoring local news and then

contacting local forest service for hourly updates• When smoke would impact a populated location,

equipment quickly loaded and taken to the area• Local fire department contacted for 2 sets of volunteers

and CDPHE provided 2 sets • Homes of similar construction and age• Residents all nonsmokers and wood burning stoves not

used• Residents told to keep all windows and doors shut and

to record any activities which may introduce PM into the indoor environment

• 2-3 air cleaners placed in one of the residence

Measurements

• Indoor and outdoor 24-h average PM2.5 mass concentrations using Harvard impactors

• Real-time indoor and outdoor particle # concentrations <0.5 m for 24 h using Climet optical particle counters

• Air-exchange rate using CO2 tracer gas decay method

Air CleanersPRE-FILTER catches larger particlesIONIZING CELL electrically charges

particles when they pass through a powerful electric field

COLLECTOR PLATES immediately attract "charged" particles

ACTIVATED CARBON FILTER removes most common odors and fumes

• 3 tests performed on the air cleaners in a test chamber to measure Clean Air Delivery Rate (CADR)

• We measured an average CADR of 420 m3/ h

• Agrees with CADR of 325-370 m3/ h that is published by Association of Home Appliance Manufacturers

Friedrich C-90a Electrostatic Precipitator

The Fires•Polhemus prescribed burn Burned all October 2001, producing heavy smoke on several days. Monitoring took place 10/20/01 to 10/21/01 when 2500 acres were ignited. Houses 1 and 2 were located 24 and 27 km north of this fire

•Schonover wildfireStarted by lightning on 5/21/02 and quickly grew to 2000 acres. Houses 5 and 6 were both located 24 km north of the fire. Monitoring took place 5/22/02 to 5/23/02. Ultimately the fire consumed 3800 acres and cost 2.4 million dollars to fight

Schoonover fire 5/23/02 (NOAA)

The Fires 4/24/02 5:30 PM 1800 acres (NOAA)• The Snaking wildfire

Began 4/23/02 behind Platt Canon High School. By 4/28/02 the fire consumed 2590 acres. Monitoring took place 4/25-4/26 11 km east of the fire.

Snaking wildfire 11: 15 am- 3:30 pm (MST) 4/24/02 Snaking wildfire 11: 15 am- 3:30 pm (MST) 4/24/02 (NOAA)(NOAA)

The Biggest Wildfire in Colorado Recorded History!!!

The Hayman wildfire 6/10/02 (NOAA) 11am-5pm.

The FiresThe Hayman wildfire

Began 6/8/02, contained 7/2/02 Consumed 137,760 total acres (67,700 in first two days), 133 residences, 1 commercial building, 466 out-buildings Sampling during this fire was difficult because the smoke plume was energetic enough to rise to the upper atmosphere Sampling took place on 6/10/02 with no results A second trip was planned on 6/18/02. There was smoke impact the previous day in Denver. Similar atmospheric conditions were forecasted. Two houses were located in southern Denver located 47 km from the fire. This day was successful.

Colorado fires on 6/9/02 7:03 pm (NOAA)Colorado fires on 6/9/02 7:03 pm (NOAA)

Data Analysis

AHouse

BHouse

AHouse

cleanerairtodue

reductionPMPercent

OutsidePM

cleanerairowInsidePM

OutsidePM

cleanerairwInsidePM

5.2

5.2

5.2

5.2

5.2

/1

Impact of air cleaners:

Impact of windows closed:

windowsclosedtodue

reductionPMPercent

OutsidePM

InsidePM

OutsidePM

InsidePM

Background

Fire 5.2

5.2

5.2

5.2

5.2

1

Polhemus Prescribed Burn 24-h average PM2.5 (g/m3)

22.0022.00

7.22

3.472.22

1.762.96

37.99

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

45.00

House1 fire10/20/01

House 2 fire10/20/01

House 1 bgd11/19/01

House 2 bgd2/19/02

PM 2.5

(ug/m3)

inside

outside2 air cleaners

Cooking event

Concentration in house 2 was 83% lower than in house 1

I/O ratio in house 1 was 50% lower than background I/O ratio

Error bars= data rangeWindows closed

Polhemus Particle Counts

Model assumes: V=815 m3, =0.21 h-1

House 1 OPC, number of particles 0.5-5.0 /m3

0.00E+00

5.00E+07

1.00E+08

1.50E+08

2.00E+08

2.50E+08

10/20/024:48 PM

10/20/027:12 PM

10/20/029:36 PM

10/21/0212:00 AM

10/21/022:24 AM

10/21/024:48 AM

10/21/027:12 AM

10/21/029:36 AM

Mountain time

Co

nce

ntr

atio

n

(#/m

3)

Outside

inside

model

Snaking Wildfire

• Measurements were made on two consecutive days in houses 3 and 4, May 25 and 26 2002

• On May 25th air cleaners were installed in house 4

• PM2.5 concentration in house 4 was 80% lower than in house 3 on May 25th

• On May 26th air cleaners were moved to house 3

• On May 26th, house 3 was 62% lower than house 4

Snaking Wildfire 2-day Comparison

63

80

6975

0102030405060708090

100

House 3 to 4 House 4 to 3 House 3 to 3 House 4 to 4

% P

M2.

5 r

educ

tion

AHouse

BHouse

AHouse

cleanerairtodue

reductionPMPercent

OutsidePM

cleanerairowInsidePM

OutsidePM

cleanerairwInsidePM

5.2

5.2

5.2

5.2

5.2

/1

24.3

3.05.5 5.0

33.132.8

5.0 4.6

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

House 10 fire6/18/02

House9 fire6/18/02

House 9 bgd7/16/02

House 10 bgd7/16/02

PM

2.5

ug

/m3 inside

outside

Hayman Wildfire24-h average PM2.5 (g/m3)

Concentration in house 9 was 87% lower than in house 10

I/O ratio in house 10 was 21% lower than background I/O ratio

Error bars= data range

3 air cleaners installedWindows closed

0.00E+00

5.00E+07

1.00E+08

1.50E+08

2.00E+08

2.50E+08

3.00E+08

6/18/023:36 PM

6/18/026:43 PM

6/18/029:50 PM

6/19/0212:57 AM

6/19/024:04 AM

6/19/027:12 AM

6/19/0210:19 AM

6/19/021:26 PM

6/19/024:33 PM

mountain time

Co

nc

en

tra

tio

n

(#/m

3 )

outside

inside

model

Hayman Wildfire Particle Counts

Model assumes =0.17 h-1, V=510m3

3 air cleaners installed total CADR= 1260 m3h-1

House 9 OPC, number of particles 0.5-5.0 /m3

Indoor PM2.5 Increases during Fires in Homes without Air Cleaners

57

92

59

73

100

0

20

40

60

80

100

House 1 House 3 House 4 House 6 House 10

PM

2.5 (

Indo

or)/

(out

door

)%

Air Cleaners Reduce Indoor PM2.5 during Fires

83 80

62

88 87

0

10

20

30

40

50

60

70

80

90

100

House 2 House 4 House 3 House 5 House 9

% P

M2

.5 r

edu

ctio

n

Results are valid, because homes have similar air exchange rates and indoor/outdoor background measurements

50

19 18

37

21

0

10

20

30

40

50

60

House 1 House 3 House 4 House 6 House 10

%P

M2.

5 r

edu

ctio

n

Keeping Windows Closed during Fires Reduces PM2.5 in Homes

without Air Cleaners

Summary

• Wildfires and prescribed burns caused an increase in indoor PM2.5 – Indoor levels increase to 57%-100% of outdoor

concentrations when windows are closed

• Air cleaners reduced indoor air PM2.5 by an average of 80% when compared to homes without air cleaners

• Closed windows provided 18-50% reduction of indoor PM2.5 when compared to background

• Indoor and outdoor background measurements were all similar and range between 3-5 g/m3.

Acknowledgments• US EPA Region 8

– Funding

• US EPA Radiation and Indoor Environment Lab• Tri-county Health Department

– Sampling Equipment

• CDPHE– Gravimetric filter analysis– (Colleen Cambell) Information on weather, smoke

movement, and access to volunteers

• Boulder Fire Department, US Forest Service– Data on wildfires, and access to volunteers

Thank You!