EDMG530Wk6

November 200 1 - Journal of Forestry 9

David T. Butry, D. Evan Mercer, Jeffrey P. Prestemon, John M. Pye, and Thomas P . Holmes

We modeled and analyzed the economic impacts of the six weeks of large, catastrophic wild- fires in northeastern Florida in June and July 1998, among Florida’s most devastating in recent history. The result of the unusually strong El Niño-Southern Oscillation (ENSO) in 1998, the Florida wildfires produced economic impacts of at least $600 million, similar in scale to recent category-2 hurricanes. Improved understanding of the interactions between management, wildfire, and its costs may yield large payoffs to society by identifying optimal intervention activities.

Keywords: economics;fire management; Florida; policy

ldfires engender far-reach- such as timber and capital. However, ing costs and losses to soci- wildfires also can produce gains, for ex- ety. Government agencies ample, by sustaining fire-dependent

and private landowners spend money forested ecosystems. for presuppression and suppression ac- The increasing costs and damages tivities as wildfires damage the physical associated with wildfires during the stocks and productivity of resources, 1990s suggest a need to analyze the

amount we should spend to prevent and suppress wildfires. Using the least- cost-plus-net-value-change criteria, Gorte and Gorte (1979) explained that the economically optimal amount of presuppression effort or expenditures is the point at which they minimize the wildfire’s cost, including both presup- pression and suppression costs, plus the net value change of the affected re- sources (the sum of all wildfire-induced gains and losses).

Clearly, minimizing the economic impacts of catastrophic wildfires re-

Above: Fire plumes in Florida as seen from the NASA GOES-8 satellite, June 6, 1998.

figure I. The locations of the June and July 1998 wildfires in Florida are shown in red. The St. John’s River Water Management District is outlined in black.

Average fire size Total fire area (acres) ( t h o u s a n d a c r e s )

2 5 0‘C 2 0 0

- Average fire size

- Total fire area

1 500 4 5 0

4 0 0

3 5 0

3 0 0

2 5 0

2 0 0

1 5 0

F i g u r e z . Average fire size and total fire area for Florida in June and July, all fuel types (federal lands excluded), 1981-99.

quires careful calculation of all associ- sion expenditures averaged $500 mil- ated costs, losses, and gains. Neverthe- lion per year for the period 1994-99. less, we know of no organization in the Such totals, however, shed no light on United States that systematically and suppression expenditures for one fire or empirically quantifies economic im- set of fires to enable optimal suppres- pacts of wildfires. The National Intera- sion and prevention pol icies . Although gency Fire Center (2000) reported that a few analysts have expanded wildfire the federal portion of wildfire suppres- cost accounting to include other eco-

nomic components of costs and net value change (e.g., Bellinger et al. 1983; Mills and Flowers 1985; Price- w a t e r h o u s e C o o p e r s 1998), i n m o s t studies net value change usually refers only to timber losses (Gorte and Gorte 1979).

Wildfires create both short- and long-run economic impacts. Where human populations are high and wild- fires are catastrophic, unique near-term losses associated with large-scale local economic disruptions are often gener- ated. Long-term costs and losses are in- curred in vegetation management, rou- tine wildfire monitoring, and impacts on timber and other forest values. The summer of 2000, when wildfires burned more than 8 million acres in the United States, is a testament to the potential impact of wildfires on people and their economic systems. As human populations increase, and as the num- ber of structures in and near wildlands increases over time, the costs and eco- nomic losses from occasional cata- strophic wildfires are l ikely to increase. Detailed analyses of the potential im- pacts of catastrophic wildfires are needed to help the public and policy- makers better understand the potential return to alternative presuppression and suppression strategies to reduce risks and impacts of wildfires.

Wildfires of 1998 The fire season of 1998 was the

most devastat ing in Florida’s recent his- tory. Approximately 500,000 acres burned, mostly in the forests of 18 northeastern Florida counties (fig. I) that compose the St. John’s River Water Management District. Although the actual number of fires in Florida (federal lands excluded) for 1998 was below average (4,9 16 fires versus the 5,720 average for the period 1981-99), more than twice the usual number of wildfires occurred in the summer months of June and July 1998 (2,255 versus an average of 935). The 1998 fires also were unusually large, averag- ing 200 acres in 1998 compared to 59 acres in an average year, propelling total area burned to more than twice the long-term average between 198 1 and 1999 (Jig. 2).

The 1998 catastrophic wildfires

10 Journal of Forestry * November 2001

were caused in large part by the weather patterns associated with an un- usually strong (“super”) El Nifio- Southern Oscillation (ENSO), the central Pacific Ocean temperature fluc- tuation that affects weather patterns worldwide. ENS0 can be divided into three recognizable phases-El Niiio, La Niña, and neutral (Green et al. 1997)--all of which affect weather in Florida. The El Niño phase is charac- terized by heavier-than-normal precip- itation in Florida, whereas the La Nifia phase usually produces reduced levels of precipitation or drought. The neu- tral phase, following an El Niiio-La Nifia oscillation, is typified by normal precipitation patterns in Florida.

The ENS0 pattern of 1997-98 (fig. 3) was atypical compared with other ENSOs because of the extraordinarily high rainfall in Florida during the wet El Niño portion of the cycle. The quick return to the La Nifia phase was also atypical and created a severe drought. During the El Nifio phase (1997 and early 1998), vegetation ac- cumulated in northeastern Florida. By April 1998, when the drought took hold, the vegetation dried up and be- came a large and hazardous fuel source. In late spring and early summer 1998, the Keetch-Bryam drought index, a standard measure of wildfire danger, flirted with the maximum possible value of 800 for several weeks. Indeed, these conditions were so extreme that bald cypress swamps and other lowland forests burned more often than upland types (Mercer et al. 2000).

In addition, the weather patterns present under La Nifia and the result- ing fuel buildup changed the mix of ig- nition sources for wildfire. Typically, arson is the most common source of wildfires in Florida. During the spring and summer of 1998, however, light- ning was the most prevalent ignition source, with total strikes exceeding 60,000 for the St. John’s River Water Management District from June to mid-July 1998.

Economic Effects We examine seven major categories

of costs and losses associated with cata- s trophic wildf ires: presuppression costs , suppression costs , disaster rel ief expen-

N i b 3 sea su r face t e m p e r a t u r e a n o m a l y

(“(2

D e p a r t u r e f r o m n o r m a l m c n t h l y (1950-2ooo)

Fkdda rainfall l i nches~

El Nlfio n n

3

1 . 5

0

-1.5

- 1 9 9 7 1998

Figure 3. Nifio 3 sea surface temperature anomaly and departure from normal rainfall for Florida by month from May 1997 to December 1998. The El Nifio and La Niña phases are depicted in green and orange, respectively.

ditures, timber losses, property dam- age, tourism-related losses, and human health effects. This list is by no means exhaustive, as other potential costs and losses may exist (e.g., lost wages, de- creased quality of life, higher long-run firefighting expenditures, landscape re- habilitation, environmental degrada- tion). But examining these seven cate- gories al lows us to focus on data that is readily available to produce a conserva- tive, lower-bound estimate.

Unlike much of the wildfire-vulner- able West, northeastern Florida is heav- ily populated, and private land owner- ship predominates. The wildfires of 1998, therefore, required evacuations of hundreds of thousands of people and destroyed private and public tim- ber resources. Because Florida, like the West, is a popular tourist destination, the 1998 wildfires also disrupted tourism to popular attractions such as beaches, Disney World, and a major stock-car race. The smoke from the wildfires also posed a potential health risk to Florida’s large and vulnerable el- derly population. In the following sec- tion, we examine the economic scale of each of these impacts in Florida.

Timber losses. Natural catastrophes such as wildfires can have short- and long-term effects on timber markets.

Short-term effects (one to two years) include the immediate destruction of valuable standing timber and eco- nomic disequilibrium associated with the flooding of markets with salvaged timber. The glut of salvaged timber dri- ves prices downward temporarily, which affects owners of the killed tim- ber, owners of undamaged timber, and timber consumers. Holmes’s 1991 analysis of a major southern pine bee- tle outbreak in Texas and Louisiana, and Prestemon and Holmes’s 2000 study of the effects of Hurricane Hugo in South Carolina document the short- run price drops caused by gluts of sal- vage material. Long-run effects on tim- ber markets can arise from the loss of a large portion of standing inventory, a loss that tends to drive prices upward for extended periods and produce a windfall for owners of undamaged tim- ber. They also can create conditions fa- vorable for a contraction in timber de- mand. Prestemon and Holmes (2000) found that the elimination of more than 10 percent of the inventory in a region could drive up long-run prices, assuming no contraction in demand. Therefore, large-scale catastrophes often redistribute wealth among pro- ducers and consumers and cause a net economic loss .

November 2 0 0 1 l Journal of Forestry 11

Timber Market Effects of the 1998 Wildfires in Florida

We quantified economic effects of the 1998 wildfires in northern Florida’s southern pine sawtimber and pulpwood markets (data limitations prohibit analyses of hardwood mar- kets impacts). The affected region corresponded closely to Timber Mart-South Area 1 (Norris Foundation 1998), so base p r i ce es t imates were based on the T imber Mar t -Sou th report for Area 1 for the second quarter of 1998. We as- sumed that northern Florida timber prices moved in a way similar to price changes identified by Prestemon and Holmes (2000) for sawtimber and pulpwood stumpage in South Car- olina following Hurricane Hugo: a salvage glut period, in which prices were lower than they would have been without the catastrophe, and a price rise after the salvage period was completed. The price rise after Hugo followed the loss of 16 percent of southern pine inventory in the coastal plain and ‘lower Piedmont ‘of South Carolina. A similar scale of inventory loss was experienced in northern Florida-14 percent of saw- timber inventory and 19 per- cent of pulpwood inventory (Mercer e t a l . 2000)--so this price path seems justi- fied. Of the sawtimber and pulpwood inventories killed in Florida in 1998, landown- ers salvaged approximately 24 percent and 23 percent of killed pine and hardwood volume, respectively. These sa lvage quant i t ies amounted to more than twice the aver- age annual removals for the region, implying that both pine and hardwood timber

P I P O

PI

and harvest quantities, we calculated the economic surplus (consumer and producer surplus effects) of the fires on the northern Florida economy. This approach was partial equilib- rium, ignoring any feedback effects that may occur between the timber sector and other sectors.

markets faced gluts of salvage materials for several months after the fires ended. Communications with experts in the field revealed that the salvage period was over by mid- 1999.

A statistical analysis of the price effects of the Florida wildfires in the manner of Prestemon and Holmes (2000) was not possible, because we lacked a long time series of observed prices following the wildfires. As an alternative, we estimated the timber price changes by applying a structural economic model: using supply-and-demand relationships available in the literature and data from Mercer et al. (2000) on inventory losses and salvage volumes. The price effects were estimated by assuming a path of salvage volume and regular harvest volumes in northern Florida and then finding the pure effects of the salvage glut and inventory reductions caused by the fire on prices. In other words, we calculated how far observed prices in northern Florida should have de- parted from actual prices in the quarters following the fires. Using those price departure estimates, salvage volumes,

Methods The figure below is a graphical representation of the

northern Florida timber market, showing the shift in supply and the glut of salvage of either sawtimber or pulpwood s t u m p a g e offered to the market. Before the 1998 wildfires, equilibrium was at point a, the intersection of the supply curve based on the original available inventory, So(P,@, and demand, D , establishing a price P O and a quantity Qo. During the first salvage period after the fires, a glut of damaged ma-

terial, V, entered the pulp- wood and sawtimber mar- kets. This material was com- bined with the undamaged timber ( f & I offered from the new supply curve, S 7 P , ! 1 1 ) . The new supply of timber was based on a smaller in- ventory (1,) and intersected the demand curve at point b, establishing a short-run price ( 4 ) and quantity ( & I . The sa lvage vo lume, V, g radual ly shifted toward the price axis as the salvage material was exhaus ted . We assume tha t four-sevenths of the sal- vaged volume was recov- ered in the first period after the fires (the third quarter of 1998), two-sevenths in the

fourth quarter of 1998, and one-seventh in the first quarter of 1999. (Although other allocations across quarters could have been used, such as a linear decrease over time, the es- timates of economic surplus changes would have been only trivially different from our assumed allocations. The salvage allocations over time also better fit price paths observed during the salvage period for Hurricane Hugo [Prestemon and Ho lmes 20001. )

After the salvage was exhausted, the new equilibrium, based on the new supply curve, was point c , a price of P I and a volume per quarter of 91.

Supply-and-demand equations were specified as expo- nential functions, and changes in surplus were based on changes in the areas between the price axis and the curves (Just et al. [1982] provided an overview of this technique for ca lcu la t ing we l fa re changes . )

Each quarter after the fires, an estimated price change was calculated, based on a salvage wood quality adjustment

12 Journal of Forestry * November 2001

Losses and gains in producer and consumer surplus caused by 1998 wildfires in the St. John’s River Water Management District, Florida ($ million 1999), assuming a 6 percent real discount rate.

P r i c e e l a s t i c i t i e s

Change in producer surplus, undamaged market Change in consumer surplus, entire market Change in producer surplus due to lost inventory Change in producer surplus from salvage revenues

Total change in consumer and producer surplus

NOTES: Supply elasticities with respect to inventory were set at 1.0 for both sawtimber and pulpwood markets. Demand elasticities with respect to price for both products were -0.5. Supply elasticities with respect to price for both products were 0.3. These were the same as those used by Abt et al. (2000). Negative surplus changes imp/y net economic losses, and positive changes imp/y net economic gains.

c ; ; : \., ‘ , . : : r1 . * _ I; : . .; ‘ i

(_ %

& e m b e r 2 0 0 1 * ~ ~ ~ ~ ~ ~ ~ ~ ~ $ 9 : : .̂

Table 1. Wildfire impact on summer tourism for counties in the St. John’s River Water Management District, Florida, June-August 1998.

Change in summer County h o t e l r e v e n u e

Alachua $ (403,181) Baker -

Brevard (4,599,488) C l a y ( 3 5 , 7 4 4 )

Duval 1,033,772 F l a g l e r ( 4 4 2 , 1 4 8 )

Indian River (302,288) L a k e ( 3 3 0 , 6 7 6 )

Marion - Nassau ( 2 9 9 , 3 4 7 )

Okeechobee 28,883 O r a n g e ( 48 ,835 ,556 )

Osceola 9,169,125 P o l k ( 6 6 6 , 8 4 6 )

Putnam 129,424 S t . J o h n ’ s (7 ,823 ,967)

Seminole 1,169,430 V o l u s i a (8 ,790 ,112)

T o t a l $ ( 60 ,998 ,679 )

Change in summer tourist spending T o t a l c h a n g e

$ (510,522) ^ $ ( 9 1 3 , 7 0 3 ) - -

(5,823,976) (10,423,424) ( 4 5 , 2 6 0 ) ( 8 1 , 0 0 4 )

1,308,997 2,342,769

( 5 5 9 , 8 6 2 ) (1,002,010)

(382,768) (685,056) ( 4 1 8 , 7 1 3 ) ( 7 4 9 , 3 8 9 )

( 3 7 9 , 0 4 4 ) ( 6 7 8 , 3 9 1 )

36,573 65,456 ( 61 ,837 ,230 ) ( 1 1 0 , 6 7 2 , 7 8 6 )

1 1 , 6 1 0 , 2 5 6 2 0 , 7 7 9 , 3 8 1 ( 8 4 4 , 3 8 4 ) (1 ,511 ,230)

163,881 293,395 (9 ,906 ,972) (17 ,730 ,939 )

i , 4 8 6 , 7 7 1 2 , 6 5 0 , 2 0 1 (11 ,130 ,336 ) (19 ,920 ,448 )

$ ( 77 ,238 ,589 ) $ ( 1 3 8 , 2 3 7 , 2 6 8 )

fires produced between $322 and $509 We calculated that the 1998 wild-

million in net losses to the Florida pine timber market (for details, see “Timber Market Effects of the 1998 Wildfires in Florida,” p. 12). These losses exclude the hardwood timber market, where ef- fects are probably smaller but likely amount to tens of millions of dollars. Because of data limitations, we did not estimate the economic effects of the wildfires on hardwood markets, but hardwood t imber inventory losses were similar in scale to softwood inventory losses (Mercer et al. 2000). Different segments of society were unequally af- fected by the wildfires, and these dif- ferences were economically significant: Consumers lost $400 million to $1.5 billion, owners of damaged timber lost $360 to $460 million, and owners of undamaged timber gained $400 mil- lion to $1.4 billion.

fires in June, about $9.4 million (cov- lion annually. Before the major wild-

ering 377,936 acres) had been spent on prescribed fire in Florida in 1998. Total presuppression and vegetation management expenditures may easily surpass $12.7 million annually, as we do not have data for other manage- ment activities aimed at wildfire risk reduction, such as thinning and herbi- cide treatments.

Costs of suppression and disaster relief: The Federal Emergency Management Agency (FEMA) approved wildfire sup- pression reimbursement and disaster re- lief aid to Florida in the wake of the 1998 wildfires. FEMA originally autho- rized $50 million for the Fire Suppres- sion Assistance Program (FSAP), al- though the final tally was expected to top $100 million (this discussion draws from Suiter and Copenhaver 1999). FSAP reimbursed state and local gov- ernments for 100 percent of their fire- f ighting costs , including wages for fire- fighting crews and the costs of their gear, and the leasing of multiple fixed- wing aircraft, helicopters, fire engines, and bulldozers. FEMA disaster relief expenditures were expected to total $20 to $25 million, with 25 percent repre-

Costs ofpresuppress ion andprescribed burning. Each year in Florida, approx- imately 506,000 acres of land are sub- jected to prescribed fire to reduce the risk of wildfire. Assuming a treatment cost of $25 per acre, prescribed burn- ing in Florida costs federal, state, and private landowners about $12.7 mil-

senting state and local government dis- aster relief expenditures. Disaster relief covered the expenses of debris removal, relief centers, temporary housing, and cris is counsel ing.

Property losses. The 1998 wildfires damaged or destroyed 340 homes, 33 businesses, and several cars and boats. These losses amounted to between $10 and $12 million (Saunders 1998). Be- cause of data constraints , this est imate only includes losses to insured prop- erty. Homes accounted for the major- i ty of the total property losses .

Tourism and trade losses. Compared to previous years, tourism expenditures in northeastern Florida fell consider- ably below expected levels during the summer months of 1998. While na- tionwide media coverage of the effects of the 1998 wildfires-including re- ports on mandatory evacuations, exces- sive smoke, and road closures-dis- couraged travel to the state, the unusu- ally hot, dry conditions may also have reduced the attractiveness of Florida to tourists . As we were unable to separate out the independent economic effect, if any, of merely bad weather on tourism, we assumed that al l the reduc- tions in tourism were the result of the wildf i res .

We stat ist ical ly tested for changes in tourism during the 1998 wildfire sea- son for the 18 counties lying at least partially within the St. John’s River Water Management District. The sta- tist ical analysis of hotel revenue effects used data from the Florida Depart- ment of Revenue’s transient rental (hotel) tax receipts, where revenues equal the hotel tax divided by the cor- responding county hotel tax rate (Baker and Marion Counties were ex- cluded because they do not report hotel taxes). The months of June, July, and August 1998 were analyzed, with August included to account for any lagged effects the fires may have had on tourism.

The change in tourism expenditures was estimated by comparing predicted (without fire) with observed expendi- tures. The predicted level was deter- mined by multiplying 1997 tourism revenues, as recorded by each included county, by 1 plus the mean annual per- cent change occurring over the previ-

1 4 Journal of Forestry - November 2001

ous lo-year period. Calculations were performed separately for June, July, and August 1998. To determine whether the change in gross sales from 1997 to 1998 was statistically signifi- cant, the percent change was compared to the 95 percent confidence interval for the average percentage change oc- curring in the previous 10 years for each county. If the percent change from 1997-98 fell outside the confi- dence interval, it was determined to be statistically significant.

The difference in tourism expendi- tures in June and July 1998 compared to the previous 10 years was not statis- tically significant. August tourism losses (792,19 1 hotel-room nights), however, were statistically significant compared to the predicted revenues for August 1998 (p < 0.05). Reduced tourism resulted in an estimated $61 million gross loss in hotel revenues and an estimated $77.2 million gross loss in non-hotel-related tourist spending. Most of these losses were experienced by the hotels and businesses of Orange, Volusia, St. Johns, and Brevard Coun- ties

Health care costs. Air pollution emis- sions were estimated for the 1998 sum- mer wildfires by county (table 2), and these emissions formed the basis for es- timating some of the economic effects of the wildfires on health. Brevard, Fla- gler, and Volusia Counties accounted for over 82 percent of the wildfire- related pollution. The Volusia County Health Department and the Florida Department of Health examined the frequency of hospital visits for asthma, bronchitis, and other respiratory con- ditions between June 1 and July 6, 1998. Sorenson et al. (1999) surveyed seven hospitals in Volusia County and one in Flagler County in 1998 and compared admissions and emergency room visits for each county for the same period in 1997. They found that emergency room visits increased by 9 I percent for asthma and by 132 percent for bronchitis with acute exacerbation during the 1998 fire season.

Usually, this type of fine-scale health data is not recorded or easily accessible. Longer time series data would, among other things, allow hypotheses to be tested, such as whether and to what de-

Table 2. Wildfire emissions by county for the St. John’s River Water Management District, Florida, June-July 1998.

County Acres C a r b o n V o l a t i l e burned’ organics*

N i t r o g e n Particulates m o n o x i d e o x i d e

. . . . . &“S . . . . . . . . . . . . I . . . .

Alachua 6 , 7 8 4 5 1 9 4 , 2 7 4 7 3 3 1 2 2

Baker 6 6 5 5 1 4 1 9 7 2 1 2

Brevard 5 7 , 0 4 2 4 , 3 6 4 3 5 , 9 3 7 6 , 1 6 1 1 , 0 2 7

C l a y 6 , 9 6 2 5 3 3 4 , 3 8 6 7 5 2 1 2 5

Duval 8 , 4 3 8 6 4 6 5 , 3 1 6 9 1 1 1 5 2

F l a g l e r 9 5 , 8 0 9 7 , 3 2 9 6 0 , 3 6 0 1 0 , 3 4 7 1 , 7 2 5

I n d i a n R i v e r 3 0 2 1 9 3 1

Lake 8 8 2 6 7 5 5 6 9 5 1 6

Marion 9 7 3 7 4 6 1 3 1 0 5 1 8

Nassau 2 , 2 1 9 1 7 0 1 , 3 9 8 2 4 0 4 0

Okeechobee 1 , 2 7 4 9 7 a03 138 2 3

O r a n g e 6 , 7 0 4 5 1 3 4 , 2 2 3 7 2 4 1 2 1

Osceola 8 , 2 0 1 6 2 7 5 , 1 6 6 a 8 6 1 4 8

P o l k 7 , 0 9 4 5 4 3 4 , 4 6 9 7 6 6 1 2 8

P u t n a m 4 , 5 3 2 3 4 7 2 , 8 5 5 4 8 9 8 2

St. John’s 1 1 , 8 4 9 9 0 6 7 , 4 6 5 1 , 2 8 0 2 1 3

S e m i n o l e 2 , 0 4 6 1 5 7 1 , 2 8 9 2 2 1 3 7

V o l u s i a 1 6 1 , 3 2 9 1 2 , 3 4 2 1 0 1 , 6 3 7 1 7 , 4 2 3 2 , 9 0 4

T o t a l 3 8 2 , 8 3 3 2 9 , 2 8 7 2 4 1 , 1 8 5 4 1 , 3 4 6 6 , 8 9 4

‘ G r a s s y f u e l s e x c l u d e d . ‘Expressed as methane.

gree wildfire size, intensity, or severity affects the incidence of respiratory ail- ments. A complication may be that res- piratory ailments also are affected by other factors associated with the kind o f w e a t h e r experienced during droughts, such as increased levels of dust, pollen, and fungal spores. Disen- tangling the specific effects of fire-cre- ated pollutants using empirical analy- ses is challenging, but respiratory prob- lems represent a tangible cost that can exceed thousands of dollars per patient, and these ailments traditionally target sensitive populations such as children and the elderly.

We applied the findings of Sorenson et al. (1999) to calculate the economic effects of the 1998 wildfires on health. We assumed that the observed in- creases in treatment for asthma and bronchitis were the direct result of the wildfires and that these increases oc- curred at the same rate in the remain- ing counties in the St. John’s River Water Management District (normal- ized for population differences). The additional asthma-related health care

expenditures due to wildfires ranged from $325,000 to $700,000. These ex- penditures included costs associated with emergency room visits, inpatient care, outpatient care, and doctor’s of- fice visits. The majority of the expendi- tures stemmed from inpatient and emergency room care (because of data limitations we examined asthma only). Cost per treatment was calculated by dividing total expenditures by the number of visits, from Weiss et al. (1992) and Smith et al. (1997), for in- patient care and emergency room vis- its. Weiss et al. and Smith et al. exam- ined a sample of asthma occurrences in 1985 and 1987, respectively, and ex- trapolated to determine nationwide in- cidence and cost. To calculate the range of expenditures from outpatient and doctor’s office visits, we assume that the ratio between inpatient and emer- gency room visits to outpatient and of- fice visits that occurred nationwide in 1985 and 1987 also held for 1998. We estimated that the wildfires caused the inpatient admission of 59 more people than normal, at a cost of $2,200 to

November 2001 . Journal of Forestry 1 5

$6,300 per visit of roughly five days, and that an additional 686 patients were treated in emergency rooms, at a cost of $110 to $290 per visit. The fires also produced 80 to 800 additional doctor’s office visi ts , at a cost of about $45 per visit, and 20 to 90 additional incidents of outpatient care, at a cost of $80 to $400 per occurrence. Because of data limitations, these numbers do not include costs associated with t reat- ment for other conditions such as bronchitis, shortness of breath, wheez- ing, or other respiratory ailments that have been shown to increase during the six-week extreme wildfire season. Fi- nally, these estimates do not include expenses paid for medications or indi- rect costs from lost wages.

Discuss ion Based on this analysis, we predict

that the 1998 wildfires will cause net damages of $300 to $500 million to the pine timber market in Florida in the long run. These are conservative es- timates, as they ignore both the hard- wood timber market and the general equilibrium effects that changes in the timber market have on the larger econ- omy. The additional $280 million in estimated nontimber economic losses are also conservative estimates. Pre- scribed burning costs around $12.7 million annually, but how much more is being spent on other presuppression methods? FEMA considered the $100 million in fire suppression and related support to be a minimum value. Wild- fire-related property losses of $10 to $12 million also provide a lower bound to actual losses, given that they do not include uninsured property losses, nor do they account for residents’ t ime and effort spent cleaning up and repairing damaged property. Tourism losses may have l ingered well past August and may have become more severe as cleanup continued and media attention per- sisted. Gross tourism revenue losses in August were almost as large as the com- bined losses of June and July.

Asthma-related healthcare costs, proxied as the cost of asthma treat- ment, were small compared to other economic losses. However, three im- portant points should be made before we discard them for economic insignif-

icance. First, it is possible that socially and politically important distributional effects exist, as sensitive populations are harmed first. Weiss et al. (1992) found that almost half of all emergency room visits for asthma were to treat children. Second, these health effects were only for the short-run costs of treatment and do not include potential longer-term health impacts. Third, they do not account for the real costs associated with pain and discomfort experienced by all people who were subject to smoky conditions.

Even given the conservative nature of our cost accounting for the 1998 wildfires, the season’s total estimated impact, at $600 to $800 million, rivals the economic effects associated with category-2 hurricanes. The size of this impact underscores the importance of a complete and accurate accounting and economic analysis of such cata- strophic events. In future studies of catastrophic losses from wildfire, we recommend a more detailed account- ing framework than that used currently by government agencies:

1. Because more than half of the total costs of the 1998 wildfires in Florida were in timber losses, we sug- gest a more sophisticated evaluation of timber market impacts. This evalua- tion should include a more detailed tally of inventory losses, salvage vol- umes by species, and salvage material quality degradation, and it should in- clude an analysis of the long-term ef- fects on the forest sector.

2. We believe that an established system for monitoring wildfire smoke emissions in the vicinity of vulnerable populations and in conjunction with careful monitoring of hospital admis- sion rates would enable analysts to more accurately link smoke to health impacts .

3. A refined accounting framework should include a system for performing a complete accounting of property losses, including both insured and uninsured. This is particularly impor- tant for evaluating the effects of such events in heavily populated areas and in areas where people have limited eco- nomic means.

The magnitude of our impact esti- mate highlights the need for mitigating

16 Journal of Forestry l November 2001

fire hazards, especially in areas where population is dense and in proximity to fire-prone forest ecosystems. Because relatively large expenses are incurred by wildfire suppression crews and emer- gency personnel to prevent the kinds of economic losses detailed in this article, a better understanding of actual losses can inform decisionmakers on the ap- propriate amount of suppression ex- penditures and presuppression expen- ditures (e.g., vegetation management activities). Enlightened wildfire policy and land-use decisions are unlikely without more complete and accurate accounting of the costs and losses to society of wildfires.

literature Cited Aiii, R.C., EW. Cwii\c;i-, dnd (;. l’A(:HI;CO. 2000.

Southern forest iebonrce assessment using the Subre- gional Timber Supply (SRTS) model. Forest l’roductr /otrn?a/50(4):25-33.

Bt:l, INGFR, M.D., H.F. K,\I\FI:, and H.A. HAI?I~I\ON. 19X3. Economic efficiency of fire managemenr on inonfederal forest and range lands. Journal r~fFormry 8 1(6):373-75.

BIIOWN, M.J. 1996. Form rturirticrjr Florida, 1995. Re- source Bulletin SRS-6. Asheville, NC: USDA Forest Service.

GOKI t., J.K., and R.W. GOR-I F. 1979. Applictltion ofeco- nomic techniqm to,& mmagmmt: A mm review ml cdu7tion. General Technical Report INT-53. Ogden. U’I‘: USDA Forest Service.

(kthv, PM., D.M. LI:GI.I.K, C.J. MIKANIIA V, and J.J. O’BKIIN. 1997. 7‘hp North American clirn~~t~~pnttcmnr naociatrd zurth thr El NiAo-Southern Orrillution. COAI’S l’wjecr Reporr Series 97- 1 . T&~hassee: Florid‘~ Start Uniwsiry, Center for OceawAtmos- pheric Prediction Studies. Available online ar www.co~~ps.iiu.ed~lilibibookler/. Last accessed by sraff August LOO I.

HOI MI s, II! I 99 I, Price and welfare effects of cata- strophic forest damage from southern pine bcrtle epi- dcmio. i~omt Sciuncc .37(2):500-5 16.

JL>~I; R.E., l).L. H~JI.II+, and A.S. S C twiIr/. 1982. Ap- plied w$m e~or~ov~ic~ and public policy. Englewood Clif&, NJ: Prentice Hall.

Ml.ilc,t II, D.E., J.M. PYI., J.P l’lli~\I’I~MON, D.‘I: B~IKY, and 711’ 1 IOI hf1.5. 2000. &~:;&r &tr of’c/lt~- atrophic u~ildfim. Research ‘liiangle Park, NC: USDA Foreat Servicr, Southern Research Station, Forestry Sciences I.,~boratory. AvGlahle online at www.rtp.srs. fs.fed.us/ecoi~ii~lih\idemOO1.hrm. L.ast accesed by staff Augntr 200 I.

Joumnl ofAgricultural Economics X2( 1): 145-60. Pl~lctw~ri-r~riousi~c~~ol~~Rs. 1998. Econconomic /meis-

merit IJ~ 1998 Florida f;res, final report of contract #EDOO24268000024-12. Washington, DC: US Department of Commerce, Economic Development Administration.

SAUNIXXS, J. 1998. Feds to absorb $145 million of fire- fighting cm. Florida &e~Union. Available online at www.jacksonville.com/tu-online/srories/O~O99~/ met-2Alwildf.html. Last accessed by staff August 2001.

SMITH, D.H., D.C. MAIONE, K.A. L~~WSON, L.J. OKAMOTO, C. BXITIS~A, and W.B. SAUNLXRS. 1997. A national estimate of the economic costs of asrhma. American Journal of Respiratory Critical Cure Medicine 156:787-93.

SOKENSON, B., M. Fuss, Z. MULLA, W. BIGLER, S. WIERSMA, and R. HOPKINS. 1999. Surveillance of

morbidity during wildfires-central Florida 1998. Morbidity and Mortaliq Weekly Report48(4):78-79.

Sun IX, L., and COP~NHAVFK, J. 1999. Statements made ar dn appearance before rhe US House of Representa- tives Committee on Resources, Subcommirtee on Forests and Foresr Health and Commitwe on Trans- portation and Infrastructure, Subcommittee on Over- sight, Investigations, and Emergency Management. Avnilable online x www.fema.govilibraryilib1Oc.hrm. Last accessed by staff Augusr 2001.

Wtxs, K.B., l?J. GIKXN, 2ndT.A. HOIXXON. 1992. An economic evaluarion of asthma in the United Stares. New England Journal ofMedicine 326~862-66.

David 7: Butry ([email protected]) is economist, D. Evan Mercer is research economist, Jefiey I? Prestemon is research forester, John M. Pye is ecologist, and Thomas I? Holmes is research forester, Disturbance Economics Research Team, USDA Forest Service, Southern Research Stat ion, IQ Box 12254, Research Trian- g/e Park, NC 27709. This research was partially $nded by a 1999 Joint Fire Science Program grant t i t led “Ecological and Economic Consequences of the 1998 Florida WiUjres”and by a National Fire Plan-funded study titled ‘Ympact of Wildjre on Local Economies. ”

Noven~ber 2001 . Journal of l%restry 1 7