417

Creating fire-smart forestsand landscapes

by Paulo M. FERNANDES

Introduction

Land use mosaics and the intensity of biomass use in theMediterranean Basin have constrained fire incidence in the past.Forests have expanded in the last decades and in parallel their man-agement has generally decreased, increasing stand-level fuel accumu-lation and landscape-scale fuel connectivity. Contemporary fire man-agement policies rely heavily on fire suppression and do not sufficientlyaddress the root of the problem, i.e. the socio-economical and land man-agement issues behind the inception and spread of fires. The effective-ness of fire fighting operations is greatly reduced when unfavourableweather conditions coincide with fuel accumulation (Figure 1). In fact,because successful fire suppression implies fuel build up, it can con-tribute to larger and more severe fires in the future.It is now recognized that short-term and reactive fire control policies

should be replaced by “longer-term policies aimed at acting on thestructural causes of fires and integrating fire and forest managementstrategies” (EFI 2010). In order to support integrated fire management,a stronger research effort is required in regards to landscape-scale firespread, mitigation of immediate fire effects (fire severity) in foreststands, and the resilience of different forest types in relation to varia-tion in the fire regime. Climate change projections make these topicseven more relevant, because the expected increase in fire danger willraise burned area and CO2 emissions (THONICKE et al. 2010).

Mediterranean Forest Week of Antalya

forêt méditerranéenne t. XXXI, n° 4, décembre 2010

Mediterranean forests will adapt to climatechange with difficulty and their protectionfrom wildfire will be important, includinglarge-scale fuel management through pre-scribed burning (PARRY et al. 2007).Proactive forest management towards

higher resistance to fire spread andincreased fire resilience, i.e. the achievementof fire-smart forests and landscapes (HIRSCHet al. 2004), comprehends two complemen-tary approaches, respectively the treatmentof fuels in fire prone vegetation types andvegetation type conversion. This paperoverviews the state of the art on these sub-jects as it relates to Mediterranean Europe.

Goals, strategiesand know-how

The ultimate goal of fire management is tomodify the fire regime, which results fromthe interaction between ignitions and thefire environment, i.e. topography, weatherand fuels. By modifying fuels, fire-smart for-est management is ultimately expected todelay fire spread and expand the weatherscenarios under which wildfire control is pos-sible, hence leading to less area burned. Anadditional goal — often more relevant orrealistic (REINHARDT et al. 2008) — is toincrease tree resistance to fire and decreasefire severity, thus diminishing damage and

increasing the ability to recover from the dis-turbance. Differentiation between fuel man-agement strategies considers fuel isolation(fuel-breaks), fuel modification and fuel typeconversion (PYNE et al. 1996). Fuel isolationand fuel modification can be viewed respec-tively as linear and area-wide options fortreating fuels and presuppose differentexpectations, respectively fire containmentand fire modification. Consequently, lineartreatments will be of limited (or null) value ifthey fail to hinder fire spread, whereas areatreatments will be beneficial as long as fireseverity mitigation is apparent.Current guidelines and practices to man-

age stands to decrease fire hazard — usuallytermed “preventive silviculture” — are quan-titatively incipient. Some published recom-mendations even oppose empirical evidenceand inference from fire behaviour models,especially in regards to stand density.Learning with wildfires, experimenting withfire or using simulation tools are the threebasic approaches that are available to refinefire-smart prescriptions. However, the capac-ity to develop sound prescriptions remainslimited: wildfire case studies are scarce andprovide anecdotal data, fire experimentationinvolving modification of the tree canopy ispractically non-existent, and fire modellingis not reliable enough, e.g. CRUZ &ALEXANDER (2010).

Assessing the effectivenessof fuel treatments

The relative role of fuel and weather inshaping the fire regime differs by vegetationtype. If the role of fuel in controlling wildfireincidence is minor then the rationale forinvesting in fuel management programs isweak. Weather is generally viewed as theprevailing driver of the high-intensity fireregimes that characterize Mediterraneanenvironments (e.g. KEELEY & ZEDLER 2009).Fire frequency analysis for Portugal(FERNANDES et al. 2010a) indicates a rela-tively short fire-free interval (12-16 years)but fire hazard, the probability of reburn,grows exponentially with time since fire, asthe aging of fuels results in fuel accumula-tion and higher flammability. Furthermore itseems that this time-dependency of fire inci-dence is only marginally affected by extreme

418

Mediterranean Forest Week of Antalya

Fig. 1:Size of burn patches

in Portugal (1998-2008)as a function of the fuel

age mosaic: maximumobserved (circles) and

coefficient of variation(crosses).

Drawn from datain FERNANDES et al.

(2010a).

419

weather, increasing the likelihood of effec-tive fuel treatment performance underunfavourable weather scenarios. Fire sizeand maximum fire size tend respectively tobe more variable and higher in older fuels(Fig. 1). Hence, the control of fuels over land-scape fire spread occurs on a relatively short-term scale but is effective, which lends sup-port to a prominent role of fuel treatments infire management. The more fragmented andhuman-influenced landscape might beinvolved in explaining the more pronouncedrole of fuel in burn probability in comparisonwith other shrub-dominated Mediterraneanregions.

Linear fuel treatments are the most com-mon option in Mediterranean Europe, buttheir performance in the face of fire is uncer-tain. In their analysis of the 2003 wildfiresin southern France, PERCHAT & RIGOLOT(2005) found out that most fuel breaks werecrossed or transposed by high-intensity fire.Still, they note that headfire growth wasdelayed and that lateral (flank) fire spreadwas generally restrained. The width, place-ment and maintenance of fuel breaks,together with the potential for spotting andthe resources available for fire fighting arecritical factors in the success of a fuel man-agement strategy based on isolation.

Fire-smart silviculture modifies the fireenvironment in ways that can frustrate thetreatment objective (GRAHAM et al. 2004).Removing or modifying the fuels resultingfrom pruning and thinning is mandatory, orthe decrease in crown fire potential will beoutweighed by the increase in surface fireintensity. Although research in this topic issurprisingly scarce, raising the tree canopyand decreasing its density creates a drierand windier environment. In NW Spain,RUIZ (2007) measured a 2-3% absolutedecrease in dead fuel moisture content whencomparing unthinned (36 m2 ha-1) andthinned (22 m2 ha-1) Pinus pinaster stands.

Fire modelling allows simulation of firecharacteristics for different fuel and standmanagement scenarios (e.g. CRUZ et al.2008), as well as landscape-level analysis offire-spread potential in response to variationin fuels and other factors (e.g. LOUREIRO etal. 2006). Expert knowledge can be analyzedto relate fire hazard with stand and fuelstructure (GONZALES et al. 2007). However,evidence of differences in fire behaviour andseverity between alternative fuel treatments

or in treated versus untreated stands can beobtained only by actually observing fires andtheir effects. Although valuable — e.g.McARTHUR (1962) reported a decrease by afactor of 3 in fire spread rate from unprunedPinus radiata to pruned P. pinaster stands— the conclusions that can be drawn fromwildfire data are usually limited in scope.Sound guidelines for treatments are morelikely to be inferred from fire-resistant foreststands, i.e. where fire-induced tree mortalityor fire severity is mitigated to some degree.Abundant documentation exists on the inter-action between fire severity and stand struc-ture in North-American continental andmediterranean conifer forests (e.g. AGEE &SKINNER 2005), and similar patterns seem tooccur in the Iberian Peninsula, wheremature and uneven-aged Pinus nigra (FULÉet al. 2008) and P. pinaster (VEGA 2000)stands persist under a regime of low to mod-erate fire severity. Fire-resilient P. pinasterpatches in northern Portugal (Picture 1) areopen, vertically discontinuous and coincidewith frequent low-intensity fires (VEGA etal. 2010).Experimental studies of fire behaviour and

effects in relation to fuel treatments havebeen extremely scarce worldwide. In SWAustralia, GOULD et al. (2007) related firebehaviour in eucalypt forest with time sinceprescribed burning. In Portugal, a drasticchange in fire behaviour — from crowning torelatively mild surface fire — was observed

Mediterranean Forest Week of Antalya

Picture 1:Fire-resistant Pinuspinaster stand nearMurça, NE Portugal.Tree density = 250 ha-1,basal area = 11 m2 ha-1

and median fire returninterval = 6 yearsPhoto P.F.

when an experimental summer fire movedfrom an untreated 28-year old stand to areasthat had been prescribed burnt 2-3 yearsbefore (FERNANDES et al. 2004). Differences infire characteristics between 13- and 28-yearold fuels could not be proven, but in a relatedstudy (FERNANDES 2009a) surface fire inten-sity was lower in prescribed burnt plots forat least 10 years after treatment.

Assessing how different foresttypes burn and recover fromfire

Forests that differ in their specific compo-sition can represent distinct fire potentials,due to differences in the nature, quantityand arrangement of fuels, which providesthe rationale for cover type conversion.Conventional wisdom assumes that someforest types, namely deciduous broadleaves,are effective at modifying fire behaviour anddisrupting landscape fire spread. Fire model-ling (FERNANDES 2009b) and fire selectivity(MOREIRA et al. 2009) studies support suchhypothesis. In NE Spain, DIAZ-DELGADO etal. (2004) report less fire incidence from pineto evergreen broadleaved to deciduousbroadleaved forests, and GONZALES et al.(2006) found that hardwoods (Quercus robur,

Q. ilex) and short-needled mountain pineswere less fire prone than the more flamma-ble pine species. The fire behaviour gradientcorresponding to the transition of one vege-tation type to another, e.g. from shrubland toQuercus rotundifolia (AZEVEDO et al. 2009),can be modelled by taking into account thespatial variation in fuels and stand struc-ture. Local weather (fuel moisture, windspeed) and the fuel-complex are both affectedby stand structure. Consequently, standcharacteristics can minimize or offset thecover type effect, as in the simulation studyof FERNANDES (2009b), where the range infire hazard was similar between and withinforest types.The fire severity implications of changes in

cover type are expected to correlate with fireincidence but have been poorly quantified. Innorthern Portugal, FERNANDES et al. (2010b)compared fire severity between adjacentstands of P. pinaster and of other species(deciduous and evergreen broadleaves andshort-needled conifers). Fire intensity washighest in P. pinaster, followed by deciduousbroadleaved and short-needled conifer forest.In addition to cover type, fire severity wasexplained by stand characteristics (height,density, basal area), terrain aspect, firespread pattern and distance to the edgebetween P. pinaster and the contiguous covertype. A faster decline in fire severity wasobserved in deciduous broadleaves (Picture2), and fire severity tended to decrease withstand maturity and in moister aspects.Implicit in these results is the fact that dif-ferent cover types will not be different just intheir fuel complexes. Simultaneous measure-ments of micrometeorological variables andfuel moisture contents should highlightweather-related differences in the fire envi-ronment between forest types, provided thatthe stands are contiguous and do not differin aspect and slope.Fire resilience is determined by the inter-

action between fire severity and speciestraits related with post fire response.Consequently, research on post fire tree mor-tality patterns is an important supplementto fire severity studies. The description andprediction of fire-induced mortality to south-ern Europe tree species has recently gainedmomentum, covering the entire fire severityrange and addressing both conifers (P.pinaster, P. nigra) and broadleaves (Quercusspp., Castanea sativa, Eucalyptus globulus)(MOREIRA et al. 2007, FERNANDES et al. 2008,

420

Mediterranean Forest Week of Antalya

Picture 2:Fire self-extinction

in a Betula alba stand,Mezio, NW Portugal.

Photo P.F.

421

CATRY et al. 2010, VEGA et al. 2010). Themost fire-resilient types are those thatrecover quickly from high-intensity fire —species able to sprout from the crown, i.e.Quercus suber (Picture 3) and Pinuscanariensis — and those associated to lowflammability environments (deciduousbroadleaves and mountain conifers), pro-vided that their fire-resistance traits(namely bark thickness) are sufficientlydeveloped to assure tree survival.

Conclusion

Fire policies in Mediterranean countriesare centred on fire suppression, which makesthem unsustainable and often counterpro-ductive. Fuel management, including theplanned use of fire, deserves a more promi-nent role in fire management. Furthermore,and as the Mediterranean environmentbecomes more fire prone, the management ofunplanned fires will have to be considered,especially in more remote areas, and both asa fuel treatment and an ecological process.Fire-smart landscapes are obtained by

area-wide fuel treatments and by fuel typeconversion, rather than by fuel isolation. Thespatial features of fuel management are crit-ical, as random patterns can locally mitigatethe effects of wildfire but have no impact onits growth. Proactive management shouldconcentrate on expanding (i) less flammableforest types, and (ii) vegetation types thatare resilient regardless of flammability, thelater being the preferred option in a climatechange context (STEPHENS et al. 2010). Bothwill require minimal treatment, in contrastto highly flammable forest plantations infire-prone regions where costly fuel treat-ments are mandatory. However, it is impor-tant to note that climate change will likelyreduce the prospects for type conversion intomore mesic forests, and will favour open dryforests, where resistance and resilience tofire can be promoted through relativelyundemanding fuel modifications.

P. M.F.

ReferencesAgee, J., & C. Skinner. 2005. Basic principles offorest fuel reduction treatments. Forest Ecologyand Management, 211, 83-96.Azevedo, J., A. Possacos, R. Dias, A. Saraiva, C.Loureiro, & P. Fernandes. 2009. Survival of holmoak woodlands in fire prone landscapes in north-eastern Portugal. In Proc. Latin American IALEConference, Campos do Jordão, Brasil.Catry, F., F. Rego, F. Moreira, P. Fernandes, & J.Pausas. 2010. Post-fire tree mortality in mixedforests of central Portugal. Forest Ecology andManagement (in press).Cruz, M., & M. Alexander. 2010. Assessing crownfire potential in coniferous forests of westernNorth America: a critique of current approachesand recent simulation studies. InternationalJournal of Wildland Fire, 19, 377–398.Cruz, M., M. Alexander, & P. Fernandes. 2008.Development of a model system to predict wild-fire behaviour in pine plantations. AustralianForestry, 71, 113-121.EFI. 2010. A Mediterranean Forest ResearchAgenda – MFRA. European Forest Institute,Joensuu.Fernandes, P. 2009a. Examining fuel treatmentlongevity through experimental and simulatedsurface fire behaviour: a maritime pine casestudy. Canadian Journal of Forest Research, 39,2529-2535.Fernandes, P. 2009b. Combining forest structuredata and fuel modelling to assess fire hazard inPortugal. Annals of Forest Science, 66, 415p1-415p9.Fernandes, P., C. Loureiro, & H. Botelho. 2004.Fire behaviour and severity in a maritime pinestand under differing fuel conditions. Annals ofForest Science, 61, 537-544.

Mediterranean Forest Week of Antalya

Picture 3:Flammable butfire-resilient: Quercussuber woodland 11 yearspost fire in Romeu,NE Portugal.Photo P.F.

Paulo M. FERNANDESCentro deInvestigação e deTecnologias Agro-Ambientais eBiológicas (CITAB)Universidade de Trás-os-Montes e AltoDouroApartado 1013, 5001-801, PortugalTel.: +351 259 350885Fax: +351 259 350480.E-mail: pfern@utad.pt

Fernandes, P., J. Vega, E. Jiménez, & E. Rigolot.2008. Fire resistance of European pines. ForestEcology and Management, 256, 246-255.Fernandes, P., C. Loureiro, M. Magalhães & P.Ferreira. 2010a. Testing the fire paradox: is fireincidence in Portugal affected by fuel age? InProc. IUFRO Landscape Ecology InternationalConference, Bragança, Portugal (in press).Fernandes, P., A. Luz & C. Loureiro. 2010b.Changes in wildfire severity from maritime pinewoodland to contiguous forest types in the moun-tains of northwestern Portugal. Forest Ecologyand Management, 260, 883-892.Fulé, P., M. Ribas, E. Gutiérrez, R. Vallejo & M.Kaye. 2008. Forest structure and fire history inan old Pinus nigra forest, eastern Spain. ForestEcology and Management, 255, 1234-1242.González, J., M. Palahi, A. Trasobares & T.Pukalla. 2006. A fire probability model for foreststands in Catalonia (north-east Spain). Annals ofForest Science, 63, 169-176.González, J., O. Kolehmainen & T. Pukalla. 2007.Using expert knowledge to model forest standvulnerability to fire. Computers and Electronicsin Agriculture, 55, 107-114.Gould, J., L. McCaw, P. Cheney, P. Ellis, I. Knight& A. Sullivan. 2007. Project Vesta – Fire in dryeucalypt forest: fuel structure, fuel dynamics andfire behaviour. Ensis-CSIRO and Department ofEnvironment and Conservation, Canberra, ACT,and Perth, WA.Graham, R., S. McCaffrey & T. Jain (tech. eds.).2004. Science Basis for Changing ForestStructure to Modify Wildfire Behavior andSeverity. Gen. Tech. Rep. RMRS-GTR-120. FortCollins: USDA Forest Service.Hirsch, K., V. Kafka & B. Todd. 2004. Using for-est management techniques to alter forest fuelsand reduce wildfire size: an exploratory analy-sis. Pp. 175–184 In R.T. Engstrom, K.E.M.Galley & W.J. de Groot (eds.). Proceedings ofthe 22nd Tall Timbers Fire Ecology Conference:Fire in Temperate, Boreal, and MontaneEcosystems. Tall Timbers Research Station,Tallahassee, FL.Keeley, J. & P. Zedler. 2009. Large, high-intensityfire events in southern California shrublands:debunking the fine-grain age patch model.Ecological Applications, 19, 69-94.Loureiro, C., P. Fernandes, H. Botelho & P.Mateus. 2006. A simulation-based test of a land-scape fuel management project in the Marãorange of northern Portugal. In D.X. Viegas (ed.).Proc. 5th Int. Conf. Forest Fire Research,Elsevier B.V., Amsterdam. CD-ROM.McArthur, A. 1962. Fire behaviour characteristicsof the Longford fire. Leaflet No. 91, O.D.C. 43,Forestry and Timber Bureau, Department ofNational Development, Commonwealth ofAustrália.

Moreira, F., L. Duarte, F. Catry & V. Acácio. 2007.Cork extraction as a key factor determining post-fire cork oak survival in a mountain region ofsouthern Portugal. Forest Ecology andManagement, 253, 30-37.Moreira, F., P. Vaz, F. Catry & J. Silva. 2009.Regional variations in wildfire susceptibility of land-cover types in Portugal: implications for landscapemanagement to minimize fire hazard. InternationalJournal of Wildland Fire, 18, 563–574.Parry, M., O. Canziani, J. Palutikof, P. van derLinden & C. Hanson. (eds). 2007. Contribution ofWorking Group II to the Fourth AssessmentReport of the Intergovernmental Panel on ClimateChange. Cambridge University Press, Cambridge,United Kingdom and New York, NY, USA.Perchat, S. & E. Rigolot. 2005. Comportement aufeu et utilisation par les forces de lutte descoupures de combustible touches par les grandsincendies de la saison 2003. Morières: Ed. De laCardère Morières.Pyne, S., P. Andrews & R. Laven. 1996.Introduction to Wildland Fire. 2nd ed. JohnWiley & Sons, New York.Reinhardt, E., R. Kean, D. Calkin & J. Cohen.2008. Objectives and considerations for wildlandfuel treatment in forested ecosystems of the inte-rior western United States. Forest Ecology andManagement, 256, 1997-2006.Ruiz, A. 2007. Efecto de las claras sobre lahumedad de los combustibles muertos. In Proc.Wildfire 2007 - 4th International Wildland FireConference, 13-17 May, Seville, Spain.Thonicke, K., A. Rammig & M. Gumpenberger.2010. Changes in managed fires and wildfiresunder climate and land use change and the roleof prescribed burning to reduce fire hazard underfuture climate conditions. Deliverable D4.2-1c /D4.2-4 of the Integrated project “Fire Paradox”,Project no. FP6-018505, European Commission.Vega, J. 2000. Resistencia vegetativa ante el fuego através de la historia de los incendios. Pp. 4.66-4.85In La Defensa Contra Incendios Forestales:Fundamentos y Experiencias, McGraw-Hill,Madrid.Vega, J., P. Fernandes, G. Defossé, M. Conedera,S. Bravo, N. Cassagne, J-L. Dupuy, M.Fernandes, E. Jiménez, L. Lucini, M., Leiva, E.Rigolot, C. Loureiro, C. Kunst, J-C. Valette, G.Pezzatti, H. Botelho, D. Portier, J. Pérez, P.Petit, R. Ledesma, J. Maréchal, J. Godoy, F.Pimont & V. Navarrete. 2010. Tree resistance tofire: final results. Deliverable D3.2-6 of theIntegrated project “Fire Paradox”, Project no.FP6-018505, European Commission.Stephens, S., C. Millar & B. Collins. 2010.Operational approaches to managing forests ofthe future in Mediterranean regions within acontext of changing climates. EnvironmentalResearch Letters, 5, 024003.

422

Mediterranean Forest Week of Antalya

forêt méditerranéenne t. XXXI, n° 4, décembre 2010