CSIRO PUBLISHING

International Journal of Wildland Fire 2008, 17, 724–741

www.publish.csiro.au/journals/ijwf

The historical range of variability of fires in the Andean–Patagonian Nothofagus forest region Thomas T. VeblenA,E , Thomas KitzbergerB , Estela RaffaeleB , Mónica MermozC , Mauro E. GonzálezD , Jason S. SiboldA and Andrés HolzA A Department

of Geography, University of Colorado, Boulder, CO 80309, USA. Nacional de Investigaciones Científicas y Técnicas de Argentina and Laboratorio Ecotono (CONICET)-Universidad Nacional del Comahue, Quintral 1250, Bariloche, 8400, Argentina. CAdministración de Parques Nacionales, Vicealmirante O’Conner 1188, Bariloche, 8400, Argentina. D Facultad de Ciencias Forestales, Universidad Austral de Chile, Valdivia, Casilla 567, Chile. E Corresponding author. Email: [email protected] B Consejo

Abstract. The present synthesis addresses key questions about several extreme fire events that occurred in the Nothofagus forest region of southern Argentina and Chile in the late 1990s and early 2000s: (1) are there historical precedents for the extent and severity of these recent wildfires? (2) To what extent can large, severe fires be attributed to influences from modern humans, either indirectly through land-use practices or directly through ignition? (3) What are the relationships of these fire events to interannual climatic variability and trends? (4) What are the medium-term ecological consequences of these fire events, particularly in terms of the resiliency of the burned ecosystems? Historic fire regimes vary greatly across the different ecosystem types in the southern Andean region, and the tree-ring record shows that before the 20th century, large severe fires also played a significant ecological role in shaping even the wettest forests. Recent severe droughts at an annual time scale have been facilitated by a trend towards higher temperatures since the mid-1970s. In large parts of the region, the risk of wildfire ignition and spread has been exacerbated by increases in lightning associated with higher temperatures, increased ignitions associated with exurban development, and conversion of less flammable native vegetation to more flammable plantations of exotic conifers. Additional keywords: Argentina, Chile, fire ecology, fire history. Introduction Reports of large wildfires in many parts of the world since the early 1980s, including in temperate and tropical ecosystems, have received substantial attention in the popular media and contributed to a widespread view among resource managers and ecological researchers that these large, severe fires are symptomatic of fire regimes that have been altered either directly or indirectly by modern humans (Myers 2006; Shlisky et al. 2007). Public concern about large fire events in recent decades is driven by the perception that recent wildfires in many regions have been unusually extensive or burned with unusual severity due to either land-use practices (e.g. fire suppression or logging) or climate change. Initial steps to create a knowledge base for developing locally appropriate integrated fire management (as in Myers 2006) are to assess the ecological role of wildfire in the area of interest and to determine if historical fire regimes have been altered by land-use practices and climate change (Shlisky et al. 2007). Thus, our objective in the present review is to synthesise existing research to assess the precedents and consequences of extreme fire events in the Nothofagus forest region of the southern Andes (south of ∼38◦ S). © IAWF 2008

A key element in examining the issues above, and in informing fire policy discussions, is a sound understanding of the historical range of variability of local fire regimes and associated ecosystem conditions. In the modern paradigm of ‘ecosystem management’(Christensen et al. 1996), the focus of management is on sustaining ecosystems, while also sustaining yields of particular ecosystem products and services. For resource managers, it is important to know the range of critical ecological processes and conditions that have characterised particular ecosystems over specified time periods and under varying degrees of human influences. This range of critical processes and ecosystem conditions has been termed natural range of variability or historical range of variability (HRV). HRV is viewed as a coarse-filter guide in ecosystem management because it identifies the processes and conditions that in the recent past (i.e. the past several centuries) have sustained ecosystem function, and helps managers set limits to desired future conditions (Landres et al. 1999). In the present synthesis, the concept of HRV is used to examine the role played by fire historically and in modern times (i.e. after 1900) in shaping ecosystem structures in the Nothofagus forest region of southern South America. 10.1071/WF07152

1049-8001/08/060724

Fire in the Andean–Patagonian Nothofagus forest region

In the Nothofagus forest region of the mid-latitude southern Andes, in the late 1990s and early 2000s, wildfires occurred that were more extensive and more severe than any fires in recent memory. For example, in Chile during the austral summer (December–February) of 2001–2002 large, severe fires burned in the region of Araucaria forests (∼38◦ to 39◦ 30 S) that were believed to be the most extensive fires in this region for at least the past 50 years (González and Veblen 2007). Likewise, nearby in Argentina in the 710 000 ha Nahuel Huapi National Park (NP) (∼40◦ S), more area burned in 1998–99 (14 399 ha) than in any previous fire season in the 50-year documentary record (Administración de Parques Nacionales 2006, unpubl. data). Among resource managers and the general public in Chile and Argentina, these extreme fire events triggered numerous questions about the contributions of past land management and global climate change to these fire events as well as about the ecological consequences of these events. The current paper addresses the following questions related to the extreme wildfires of the late 1990s and early 2000s in the Nothofagus forests of the southern Andean region. (1) Are there historical precedents for the extent and severity of these recent wildfires? (2) To what extent can large, severe fires be attributed to direct influences from modern humans (e.g. effects of fire suppression on fuel accumulation, increased anthropogenic ignitions, or land-use practices such as livestock production and plantation forestry)? (3) What are the relationships of these fire events to climatic variability and trends? (4) What are the medium-term (i.e. a few years to a few decades) ecological consequences of these fire events, particularly in terms of the resiliency of the burned ecosystems? We address these questions for the two areas within the Nothofagus forest region for which fire history and fire ecology have been most thoroughly researched: (1) an extensive area in the Argentine Lake District from ∼38◦ to 43◦ S encompassing wet Andean forests eastwards to xeric woodlands, and (2) a smaller area of Araucaria–Nothofagus forests at 38◦ to 39◦ 30 S in the Andes in the Chilean region known as La Araucanía. We then place these two case studies into broader context by briefly reviewing the fragmentary evidence on potential influences of modern humans and climate variation for the larger area of Nothofagus forests extending from ∼38◦ to Tierra del Fuego at ∼55◦ S. Climate and vegetation patterns Climate For the purpose of the present synthesis, the climate and vegetation patterns are briefly described for the mid-latitude (i.e. 38◦ to 43◦ S) and high-latitude (south of ∼44◦ S) Andean region (see Veblen 2007 for greater detail). The dominant climate drivers of the southern Andes are the persistent mid-latitude westerlies, the seasonally shifting subtropical anticyclone of the south-eastern Pacific region, and the topographic influences of the coastal and Andean mountains (Garreaud and Aceituno 2007). Westerly flowing air is orographically uplifted by the mountain barriers to result in mean annual precipitations of 3000 to over 5000 mm on the windward slopes of mountains from ∼37◦ S to Tierra del Fuego (∼55◦ S) whereas precipitation steeply declines leeward of the Andes. Along the west coast as far south as ∼42◦ S, Mediterranean-type precipitation seasonality is associated with

Int. J. Wildland Fire

725

the summer influence of a subtropical high-pressure cell in the south-eastern Pacific (Garreaud and Aceituno 2007). Southwards of ∼47◦ S, the seasonal distribution of precipitation is relatively uniform with stormy conditions prevailing year-round. In the far south at ∼52◦ to 55◦ S, the influence of the circumAntarctic low pressure trough becomes more evident, and cool, windy conditions prevail for most of the year. Regional vegetation patterns The Nothofagus forest region south of ∼37◦ 45 S includes rainforests mostly west of the Andes as well as the cool temperate Nothofagus forests and woodlands mostly east of the Andes (Veblen et al. 1996). The Valdivian rainforest occurs from ∼37◦ 45 to 43◦ 20 S, the North Patagonian rainforest from ∼43◦ 20 to 47◦ 30 S, and the Magellanic rainforest south of ∼47◦ 30 S (Fig. 1). The Valdivian rainforest district is characterised mainly by evergreen broadleaved trees and evergreen conifers (Table 1), but also includes the deciduous Nothofagus obliqua and N. alpina (synonym = N. nervosa) in its northern, drier extent. In contrast, the North Patagonian and Magellanic rainforests are purely evergreen (mainly broadleaved trees with a small conifer component). The giant evergreen conifer Fitzroya cupressoides (Cupressaceae) occurs in both the southern part of the Valdivian rainforest and the northern extent of the North Patagonian rainforest. Another important evergreen conifer is Pilgerodendron uviferum (Cupressaceae), which occurs from 39◦ 30 S in the Valdivian rainforest to 55◦ 30 S in Magellanic rainforest and bogs on Tierra del Fuego. Common dominants of the North Patagonian rainforests are the evergreen Nothofagus (N. dombeyi, N. nitida, and N. betuloides), which typically occur in stands associated with fewer than five or six other angiosperm or conifer tree species. The most characteristic tree of the Magellanic rainforests is Nothofagus betuloides, which often forms monotypic stands or co-occurs with just a few other tree species (Table 1). The deciduous N. pumilio and N. antarctica also occur within all three rainforest districts in areas transitional to subalpine forests (N. pumilio) or on poorly drained sites (N. antarctica). Bamboos (Chusquea spp.) typically dominate the rainforest understoreys in the north but are absent south of ∼48◦ S. On the eastern side of the Andes, parallel to the strong west-to-east decline in precipitation is a gradient from temperate rainforests, through cool temperate Nothofagus forests and xeric woodlands to the Patagonian steppe of bunchgrasses and shrubs (Fig. 1). Cool temperate Nothofagus forests and woodlands occur from ∼37◦ 30 S southwards to ∼55◦ S on Tierra del Fuego and include subalpineAndean forests as well as drier forests eastward in the rain shadow of the Andes (Veblen et al. 1996). In the northern part of the extensive cool temperate Nothofagus forests, the evergreen conifers Araucaria araucana (north of ∼40◦ 20 S) and Austrocedrus chilensis (north of ∼44◦ S) occur along the mesic forest to xeric woodland portion of this trans-Andean gradient (Veblen et al. 1995). South of ∼47◦ S, the strong west-to-east precipitation gradient is also strongly reflected by the vegetation gradient. The far western district (west of the Magellanic rainforest) is known as Magellanic moorland and is characterised by scattered Nothofagus betuloides, bog, and ericaceous heath. East of the Magellanic moorland, the west-to-east zonation of vegetation runs from Magellanic forest dominated by N. betuloides through a broad

726

Int. J. Wildland Fire

T. T. Veblen et al.

N 40°

200 km

Grasslands and scrub Warm temperate deciduous Nothofagus forest

PACIFIC

Valdivian and North Patagonian rainforest Patagonian Nothofagus forests

45°

Cool temperate Nothofagus forest and woodland Magellanic rainforest Magellanic moorland Andean tundra Glacier

OCEAN

50°

ATLANTIC

OCEAN

55°

80°

75°

70°

Fig. 1.

65°

60°

55°

Major vegetation zones of southern South America.

zone of deciduous Nothofagus forests most typically dominated by N. pumilio, and then the extensive Patagonian steppe dominated by bunchgrasses and small shrubs (Fig. 1).At high latitudes conifers are absent from the xeric eastern habitats near the steppe ecotone. Fire in the Argentine Lake District (∼38–43◦ S) What are the historical precedents of the recent large, severe fire events? We focus on the Argentine Lake District (∼38–43◦ S) because it has been the most thoroughly studied region in relation to the

four questions addressed in the present synthesis. According to NP records covering most of the Andean area of these latitudes, human-set and lightning-ignited fires burned similar-sized areas from 1938 to 2005 (Administración de Parques Nacionales 2006, unpubl. data). Nevertheless, there is a high degree of interannual variation in the presence of lightning storms, and in most years no lightning-ignited fires are recorded. In the large Nahuel Huapi NP that covers ∼710 000 ha from ∼40◦ 10 to 41◦ 35 S, more area burned in the years 1996 and 1999 (11 252 and 14 399 ha respectively) than in all the other years combined since the Park’s records began in 1938 (Fig. 2). In adjacent Lanín NP, covering

Fire in the Andean–Patagonian Nothofagus forest region

Int. J. Wildland Fire

727

Table 1. Common tree species of the major forest zones of southern Chile and south-western Argentina Deciduous broadleaved species are indicated by a superscript H and evergreen conifers by a superscript hash (#). All others are evergreen broadleaved trees

Nothofagus alpinaH Nothofagus obliquaH Laurelia sempervirens Amomyrtus meli Crinodendron hookerianum Eucryphia cordifolia Persea lingue Aextoxicon punctatum Gevuina avellana Fitzroya cupressoides# Luma apiculata Podocarpus saligna# Nothofagus nitida Laureliopsis philippiana Amomyrtus luma Pseudopanax laetevirens Chusquea bamboos Caldcluvia paniculata Saxegothaea conspicua# Weinmannia trichosperma Nothofagus dombeyi Podocarpus nubigena# Maytenus magellanica Drimys winteri Pilgerodendron uviferum# Embothrium coccineum Nothofagus betuloides Nothofagus antarcticaH Nothofagus pumilioH Araucaria araucana# Austrocedrus chilensis# Lomatia hirsuta

Valdivian rainforest

North Patagonian rainforest

X X X X X X X X X X X X X X X X X X X X X X X X X X X X X

X X X X X X X X X X X X X X X X X

412 000 ha from ∼39◦ 10 to 40◦ 10 S, the 1980s and 1990s had twice as many years of significant area burned (i.e. >0.01% of the Park’s total area) as in the previous 20 years, but there is no trend towards an increase in total area burned (Fig. 2). When all four NPs in the Argentine Lake District are considered, there does not appear to be a trend towards increasing area burned since the records began in the late 1930s. For example, in Los Alerces NP (263 000 ha; 42◦ 34 to 43◦ 20 S) and the small Lago Puelo NP (27 674 ha; 42◦ 06 to 42◦ 16 S), the peak year of surface area burned was 1944 when 13.7 and 30.7%, respectively, of each park burned (Fig. 2). However, interpreting the NP records is complicated by the fact that fire suppression activities have been better supported inside than outside the NPs (i.e. on private property and properties of the provinces), which may be reflected in greater areas burned in areas adjacent to NPs in some recent fire events. For example, in three large fire events that occurred in 1987, 1996, and 2001 in the area of Lago Puelo NP, a total area of 13 111 ha burned but only 12% of the burned area was within NP boundaries (Administración de Parques Nacionales 2006, unpubl. data). Owing to the absence of long-term fire records for

Magellanic rainforest

Cool temperate Nothofagus forests and woodlands

X X X X X X X X

X X X X X X X

administrative units outside of NPs, it is not possible to determine if trends or lack of trends in NP records are representative of patterns for the larger region. Nevertheless, qualitative historical observations of years of exceptionally large fire events are helpful in placing the 1996 and 1999 fire events into a broader regional perspective. Extensive fires in the early 1940s in the Andean area of the provinces of Neuquén, Rio Negro and Chubut from ∼36◦ to 46◦ S motivated an extensive report on the extent, causes and consequences of those fires commissioned by the Argentine Ministry of Agriculture and executed by Tortorelli (1947). Tortorelli (1947) estimated that 275 000 ha burned in 1943–44 in the province of Chubut primarily in areas outside of NPs. Likewise, he documented extensive fires in the northern part of Neuquén, north of Lanín NP, in 1943 (Tortorelli 1947). In earlier resource surveys conducted in 1912–15 covering the Andean area from ∼39◦ to 44◦ S, large areas were mapped as ‘recently burned’(Willis 1914; Rothkugel 1916). For example, in a survey covering 137 760 ha of tall forest and 33 080 ha of Nothofagus antarctica shrubland in the southern part of what is today Nahuel Huapi NP, Willis (1914)

728

Int. J. Wildland Fire

T. T. Veblen et al.

Lanín NP (412 000 ha)

2

1

Percentage of park surface burned

0 Nahuel Huapi NP (710 000 ha) 1

0 30 Lago Puelo NP (27 674 ha)

20 10 0 15

Los Alerces NP (263 000 ha)

12

1 0 1940

1950

1960

1970 Year

1980

1990

2000

Fig. 2. Area burned per year as percentages of total areas of four national parks in the Argentine Lake District.Years listed indicate the year of January in periods from October to March (spring to summer). Data from Administración de Parques Nacionales, Bariloche, Argentina, 2006.

mapped 52 060 ha (23%) as ‘recently burned’. According to reports of scientific explorers, these fires occurred mostly in the 1890s-to-1914 period that coincided with intentional burning by Euro-American settlers as well as drought (Moreno 1897; Steffen 1909; Willis 1914; Rothkugel 1916). For the Provinces of Neuquén, Rio Negro and Chubut, Rothkugel (1916) mapped 692 000 ha (37% of the total forest area) as having burned during the European settlement period from the 1890s to 1914. Despite the lack of continuous record-keeping of areas burned in the Lake District, these fragmentary observations indicate that the large fire events of the late 1990s to early 2000s were not more extreme than events recorded in the early 1940s and 1890s–1910s. Tree-ring records from northern Patagonia clearly document years of synchronous and widespread fires numerous times both during the late 19th century period of European colonization as well as before permanent European settlement (Kitzberger et al. 1997; Veblen et al. 1999). For example, of 41 fire history sample sites extending from 39◦ to 43◦ S, 61% recorded fire in 1827 and 52% in 1897 (Veblen et al. 1999). As discussed below, these years of widespread fire coincided with exceptional droughts, but the temporal and spatial pattern of all

fires, including small fires that did not spread greatly, presumably owing to less extreme drought, imply that Native Americans significantly influenced fire frequency, especially in the ecotone between the steppe and Austrocedrus woodlands. For example, areas described by 19th-century explorers as important Native American hunting grounds recorded more fires during the mid19th century than ecologically similar areas less frequented by hunters (Veblen et al. 1999). Fire spread in these relatively dry habitats of steppe andAustrocedrus woodland is not as dependent on exceptional drought as fire spread is in the western rainforests where human-set fires can spread only under extreme drought conditions. Qualitative historical descriptions from the 18th and 19th centuries as well as historical photographs from the late 19th and early 20th centuries indicate that extensive, severe fires occurred in all the woodland and forest types (Veblen et al. 2003). Even in the wettest forests of the region, 18th and 19th century explorers reported seeing large areas of burned forest. In crossing the Andes from west to east in 1787 via the famed camino de Vuriloche south-west of Lake Nahuel Huapi, Padre Francisco Menéndez mentions seeing extensive burns, including

Fire in the Andean–Patagonian Nothofagus forest region

one so large that ‘. . .it extends as far as one could see. . .’ (Fonck 1896). In 1856, the explorers Francisco Fonck and Fernando Hess retraced this portion of the route travelled by Menéndez and reported seeing large burns in the same area and also at Lago Moreno (on the south side of Lake Nahuel Huapi) and Lago Gutiérrez (Fonck 1900). For the region from Lake Nahuel Huapi southwards to Rio Puelo (42◦ S) and Rio Palena (43◦ 40 S), Fonck and other explorers observed extensive fires in the mesic forests that are described (Fonck 1896, pp. 74–75) as: ‘The phenomenon [fires] is notable due to its almost general character and for having such vast proportions . . .; the strong winds and relative drought of the land and forest make these fires propagate with great rapidity’. Tree-ring dating of fire scars further documents the occurrence of numerous fires in the 18th to 19th century and earlier from even the wettest forests containing Fitzroya cupressoides at ∼41◦ to 42◦ S (Veblen et al. 1999). Likewise, at millennial time scales, there is abundant sedimentary charcoal evidence of past burning across the full range of ecosystem types throughout the Holocene in the area from 41◦ to 42◦ 30 S (Whitlock et al. 2006). Most sites show increases in charcoal coinciding with European settlement or the spread of livestock into the region during the 17th to 19th centuries (Whitlock et al. 2006). Overall, Argentine NP records beginning in 1938 as well as qualitative historical observations from the 18th and 19th centuries corroborated by tree-ring dating of past fires indicate that the extent of the fires of the late 1990s to early 2000s are not without precedent at time scales of the past 60 years as well as during the past several centuries. Quantitatively precise comparisons are not feasible, but multiple sources of evidence imply that in the Argentine Lake District, episodes of burning similar or greater in both spatial extent and range of forest types affected as in the 1990s–2000s occurred in the 18th and 19th centuries as well as in the 1890s–1910s and early 1940s. How have modern land-use practices contributed to recent fire events? The fire suppression–fuels accumulation model In various parts of the world, historical fire regimes (either natural or affected by aboriginal populations) of some semiarid ecosystems were characterised by relatively frequent lowseverity fires before the effects of modern fire suppression or fuels reduction through livestock grazing. For some of these ecosystems, there is evidence that following a reduction in the frequency of low-severity grass fires due either to livestock grazing or modern fire suppression, accumulations of woody fuels have created the potential for high-severity crown fires (Covington and Moore 1994; Schoennagel et al. 2004; Bond et al. 2005). To what extent could fire suppression and fuels accumulation contribute to the large, severe fires of the late 1990s–2000s in the Argentine Lake District? The fire suppression–fuels buildup model is not applicable to the wet Nothofagus-dominated forests (i.e. >2000 mm annual precipitation), where fire spread is clearly not limited by lack of woody fuels. Nor is there clear evidence that suppression has significantly altered the historical fire regime, which consisted of relatively long intervals (usually a century or more) between fires that were associated with extreme drought and ignited either by

Int. J. Wildland Fire

729

lightning or humans (Veblen et al. 1999; Kitzberger and Veblen 2003). The wet Nothofagus forests are located in areas of difficult access where before the use of aerial fire suppression tactics in the 1980s (which still are quite limited), the infrastructure for combating fires was not effective. In the tall shrublands dominated by Nothofagus antarctica, the bamboo Chusquea culeou and numerous tall shrub species (e.g. Schinus, Embothrium, Lomatia), transition of a surface fire into a crown fire is not limited by insufficient quantity or vertical continuity of woody fuels. All the dominant woody species of this ecosystem type resprout quickly after fire. After fewer than 5 years, there is sufficient fuel, including live and dead foliage and small twigs, near the ground surface to allow surface fires to become intense crown fires, given appropriate fire weather.The effectiveness of the fire suppression policy since the 1930s is difficult to assess because the crown fires typical of this ecosystem type destroy most tree-ring evidence of the pre-1930s fire regime. However, even if fire frequency is assumed to have declined in this ecosystem type during the last ∼60 years of fire suppression policy, the result would not have been a shift from surface to crown fire, given the rapid recovery of woody fuels due to the resprouting of the dominant shrubs. Rather, fire exclusion over decades leading to continued accumulation of highly flammable live and dead fuels may be expected to result in more intense crown fires affecting larger areas. Thus, at a landscape scale, if fire exclusion leads to greater fuel accumulation in tall shrublands, it is logical to predict greater potential for fire spread to other vegetation types such as the higher-elevation subalpine forests. However, more research is required to determine the ecological consequences, especially at a landscape scale, of fire suppression policy of the past ∼60 years in these tall shrublands. In contrast, for xeric Austrocedrus woodlands at the ecotone where the dominant fuels are bunchgrasses and small shrubs, the fire suppression–fuels accumulation model applies. The history of changes in fire frequency and fire severity in Austrocedrus woodlands fits a globally common pattern for semi-arid areas where grass fuels and low-severity surface fires formerly predominated but under many decades of fire exclusion and grazing by livestock, woody fuels accumulate sufficiently to create the potential of intense crown fires (Bond et al. 2005). The Austrocedrus woodland is the only ecosystem type in the Argentine Lake District that formerly had a fire regime of relatively frequent but low-severity fires (Kitzberger et al. 1997; Veblen et al. 1999). Fire occurrence and spread also are limited by availability of fine fuels as indicated by the lack of spread of modern fires into areas with patchy, discontinuous grass fuels, and by increases in fire extent one to a few years following periods of above-average moisture that favour production of fine fuels (Kitzberger et al. 1997; Veblen et al. 1999). The tree-ring record of fire in Austrocedrus woodlands near the steppe ecotone shows a marked reduction in fire occurrence after ∼1900 (Fig. 3). Survival of the fire-recording trees as well as eye-witness accounts from the 19th century indicate that previous fires were low-severity fires that mainly burned grass and small shrub fuels (Cox 1863; Musters 1871). Fire frequencies declined dramatically at these sites after ∼1900 due to a combination of factors: (1) fewer fires were set by the Native American population that declined dramatically around ∼1900; (2) growth of the livestock population reduced the quantity of grass fuels;

730

Int. J. Wildland Fire

1450

1500

T. T. Veblen et al.

1550

1600

1650

1700

1750

1800

1850

1900

1950

RAHUE HUECHU SN PED COLLUN TROMPH CALEUF E TRAF N LIM1 N LIM2 S LIM E LIM HUALA C VIRG EPUYEN L RIVA R GRAN 2000

Year Fig. 3. Composite fire scar records indicating years in which at least 10% of the recorder trees (minimum two scars) recorded fire at sample sites in Austrocedrus woodlands in the Argentine Lake District. Each horizontal line represents a different site indicated by the codes to the right, for which dates of fire scars are indicated by short vertical lines. Sites are arranged from north to south, and site codes are defined in Veblen et al. (1999).

and (3) fires were actively suppressed, especially after the 1940s (Veblen et al. 2003). The decline in fire frequency coincides with a substantial increase in the density of woody vegetation in this ecosystem (Fig. 4), which mainly reflects enhanced juvenile tree survival permitted by longer fire-free intervals (Veblen and Lorenz 1988). However, livestock-induced releases from competition from grasses or exposure of bare mineral soil as well as episodes of climate favourable for tree establishment also contributed to the sometimes dramatic increase in densities of Austrocedrus woodlands (Villalba and Veblen 1997; Kitzberger et al. 2000). Repeated historical photographs show that by the late 20th century, there were extensive areas of relatively dense Austrocedrus stands that could support crown fires where 80 to 100 years earlier the fuels would support only low-severity surface fires (Fig. 4). In these areas of formerly open vegetation types near the ecotone with the steppe, woody encroachment during the fire suppression period appears to have greatly elevated the risk of crown fires. However, during the record fire season of 1998–99 in Nahuel Huapi NP, these Austrocedrus woodlands accounted for a relatively small proportion of the area burned relative to more mesic forest types (J. Salguero et al., unpubl. data). Although woody encroachment associated with fire suppression and other factors near the steppe ecotone have increased woody fuel loads over an extensive area, a much larger area does not fit the scenario under which fire suppression would have resulted in an increase in woody fuels. Effects of other land uses and anthropogenic ignitions Although the fire events of the late 1990s and early 2000s in the Argentine Lake District are not unprecedented and are strongly linked to climatic variation (see below), several aspects of modern land-use appear to be creating either increased hazard

of fire spread or increased probability of ignition. Livestock populations in the steppe and ecotone with Austrocedrus woodlands peaked in the early to mid-20th century, and have declined sharply, especially since the early 1990s in the NPs (Veblen et al. 2003). Reduction of livestock in turn would have resulted in increased accumulation of grass fuels in the steppe–woodland ecotone, which has created a greater potential for fire spread. Livestock populations have also been reduced in forested areas within the NPs, favouring the development of taller and denser understoreys of palatable shrubs and bamboos (Veblen et al. 1992; Blackhall et al. 2008) in Nothofagus stands that in some cases formerly had sparse, open understoreys. Thus, reduction of livestock use has increased availability of understorey fuels in some forest types, potentially raising the probability of surface fires transitioning into crown fires. However, in the mesic Nothofagus forests, these changes in fuel type and fuel amount are likely to have less influence on the chance of surface fires becoming crown fires than would the occurrence of extreme drought. Since the early 1980s, large areas of the steppe and Austrocedrus woodlands, both inside and outside of NPs, have been afforested with introduced conifers, including ponderosa pine (Pinus ponderosa), lodgepole pine (Pinus contorta), and Douglas-fir (Pseudotsuga menziesii). In northern Patagonia (i.e. Neuquén, Rio Negro and Chubut provinces), the rate of establishment of plantation of fast-growing exotic conifers in the late 1990s was estimated at 1000 to 1500 ha per year, and it has been suggested that there are 700 000 to 2 000 000 ha of land (mostly at the ecotone with the steppe) that could be converted to plantations (Schlichter and Laclau 1998). Ecotonal areas, where formerly there was a mosaic of patches of shrubs and extensive grasslands, have been converted to areas of continuous fuels of flammable conifers. The fuel-defined hazard of crown fire in

Fire in the Andean–Patagonian Nothofagus forest region

Int. J. Wildland Fire

731

(a)

(b)

Fig. 4. Matched historical and modern photographs at the steppe and Austrocedrus woodland ecotone in Nahuel Huapi NP in (a) 1896 and (b) 1985. Woody encroachment into the bunchgrass- and small shrub-dominated steppe has been by the native conifer Austrocedrus chilensis and small trees and shrubs Schinus patagonicus and Discaria articulata. Photographs: F. P. Moreno 1896 and T. T. Veblen 1985.

these plantations is high, and in the recent fires in Nahuel Huapi NP, conifer plantations were among the vegetation types that burned (J. Salguero et al., unpubl. data). Given that international trading of global carbon credits is likely to promote more conifer plantations in the region (Sedjo 1999), this source of increased hazard of crown fire is likely to increase. The other important change in land use in northern Patagonia contributing to increased fire risk is rapid exurban development

and associated recreational activities. Many areas that were formerly lightly settled, again especially near the steppe and woodland ecotone, have experienced dramatic residential growth related both to international and national trends. The sharp increase in the permanent and transient population of exurban areas both within parks (but limited to multiple-use zones) and adjacent to the parks raises the question of the relative importance of human-set fires v. natural ignitions from lightning. In

732

Int. J. Wildland Fire

T. T. Veblen et al.

Bariloche 2

Spring (Oct.–Dec.) temperature

Departures (SD) from the mean

0 ⫺2 4

Spring (Oct.–Dec.) precipitation

2 0 ⫺2 ⫺1.5 Spring (Oct.–Dec.) aridity index ⫺1.2 ⫺0.9 1900

1920

1940

1960

1980

2000

Year Fig. 5. Spring temperature, precipitation, and aridity index departures (standard deviations) from mean monthly values for the Bariloche climate station. The curve in the upper two panels is an 11-year moving average of the respective value. The aridity index (plotted inversely) is P/(T + 10), where P is mean spring precipitation in mm and T is mean spring temperature in (◦ C) (De Martonne 1926). The dots indicate the record fire seasons of 1995–96 and 1998–99.

the case of the record fire seasons of 1995–96 and 1998–99 in Nahuel Huapi NP, the origins of 22 fires that accounted for 48% of the total of 25 651 ha burned could be determined with certainty (Administración de Parques Nacionales 2006, unpubl. data); 77% of the fires of known origin were human-set and burned 33% of the area burned by fires of known origin. Over the 1938–2005 record of fires in Nahuel Huapi NP, of fires of known origin (64%), human ignitions account for 56% of the area burned (Administración de Parques Nacionales 2006, unpubl. data). Over the period from 1985 to 1999, the area burned within 10 km of the major regional urban centre (Bariloche) was 3.5 times greater than the area expected to burn based on the area that burned in the same vegetation types over a total study area of 780 000 ha (Mermoz et al. 2005). As exurban development continues at several foci in the region, patterns of increased fire occurrence similar to that documented for the Bariloche area are likely to be more widespread. Although residential development is greatest towards the steppe–woodland ecotone, mesic and wet forests have also experienced increased use both for residence and for recreation especially since the early 1990s. Recent development has occurred in mesic Nothofagus forests that were burned during the late 19th–early 20th century (Veblen et al. 2003). Burning of these tall forests converted them to either much shorter secondary forests of the same species or in some instances to tall shrublands; in both cases, the result is a transformation of

relatively fire-resistant tall forests to structural types that have greater vertical fuel continuity and increased fire hazard as determined by the fuels complex. Today, these areas are subjected to increased chance of fire ignition because of greater human presence. What are the relationships of recent fire events to climatic variability and trends? The 1998–99 fire season (December through March), which was the peak year in area burned in Nahuel Huapi NP, followed the driest calendar year (1998) and the warmest spring (October– November 1998) in the ∼90-year instrumental climate record of Bariloche, located at the south-east edge of the Park (Fig. 5). The 1996 fire season, the second-ranked year of widespread burning, also followed one of the driest springs on record (Fig. 5), and the mean temperature of December was the warmest on record. Interannual variability in winter–spring precipitation and summer temperatures has been shown to strongly influence fire occurrence in theArgentine Lake District (Kitzberger et al. 1997; Veblen et al. 2003).Although the drought conditions at a seasonal and annual scale were extreme in 1995–96 and 1998–99 relative to the 20th century instrumental record, a 1599–1988 tree-ring reconstruction of precipitation based on Austrocedrus sampled throughout the Argentine Lake District indicates numerous dry years over the past several centuries associated with widespread

Fire in the Andean–Patagonian Nothofagus forest region

733

Summer temperature (11-year moving average)

15

14

13

12 1940

Number of lightning fires (11-year moving average)

fires (Villalba et al. 1998; Veblen et al. 1999). A direct quantitative comparison of the pre-20th century drought years with the drought of 1998–99 is not feasible, but it is likely that at an annual scale, some of the droughts associated with widespread fire in the previous two centuries were similar in severity to those of the late 1990s. Although years of severe drought and widespread fire similar to those of the 1990s are not unprecedented, the current warming trend (especially since the mid-1970s; Villalba et al. 2003) leads to the possibility that the combination of natural variability in rainfall with unprecedented warm summers is creating conditions favourable to widespread fire more frequently than in the past several centuries. Analyses of fire–climate relationships based on tree-ring reconstructions for the period 1740–1995 indicate the primacy of interannual variability over multidecadal-scale climatic influences on fire occurrence in the Argentine Lake District (Veblen et al. 1999). Thus, effects of decadal-scale climatic trends on the fire regime may be difficult to detect because years of widespread fire may often be associated with only 1 or 2 years of extreme climate within a decade. Nevertheless, climate trends over the last several decades of the 20th century appear to be more conducive to widespread fire in the Argentine Lake District. At a decadal scale, since 1976, spring (October–December) precipitation has been below average and temperature has been above average at the Bariloche climate station (Fig. 5). The shift towards warmer temperatures in the Argentine Lake District is also reflected by increased tree growth at subalpine sites in the region (Villalba et al. 1997; Daniels and Veblen 2004). Annual variations in the regional climate and fire occurrence in the Argentine Lake District are strongly associated with variations in El Niño– Southern Oscillations (ENSO) (Kitzberger et al. 1997; Veblen et al. 1999; Kitzberger and Veblen 2003). The post-1976 trend towards warmer temperatures in the Argentine Lake District is a hemispheric trend linked to warmer sea-surface temperatures in the Pacific and Atlantic and to more frequent strong to very strong El Niño events since 1976 (Daniels and Veblen 2001; Giese et al. 2002; Villalba et al. 2003). The strong shift in temperature regime of the tropical Pacific after 1976 is reflected in both warmer summers and increased numbers of lightning-ignited fires in the Argentine Lake District (Fig. 6). In the four NPs in the Argentine Lake District, the mean number of lightning-ignited fires shifted sharply from an average of less than 10 to well over 25 after the late 1970s. The increase in lightning ignitions is associated with greater convective activity under the warmer conditions. Thus, the sharp shift towards warmer and drier conditions since the late 1970s in combination with increased lightning events appears to have resulted in a qualitative shift towards increased fire risk. The past record of fire implies that an increase in lightning can have a major influence on fire occurrence. For example, over the period from 1938 to 2005, lightning-ignited fires were recorded in only 17 years, implying that in most years, fire occurrence is at least partially limited by lack of lightning ignitions (Administración de Parques Nacionales 2006, unpubl. data). The number of lightning ignitions recorded since 1938 is strongly correlated with warmer summer temperatures that are essential for allowing fire spread in the wetter forest types (Kitzberger and Veblen 2003). Lightning is known to have ignited some of the largest fires that have burned the usually fire-resistant wet Nothofagus

Int. J. Wildland Fire

1950

1960

1970

1980

1990

2000

1950

1960

1970

1980

1990

2000

60

40

20

0 1940

Fig. 6. Mean summer (December–March) temperature (◦ C) of the Bariloche climate station and number of lightning-ignited fires in Lanín, Nahuel Huapi, Lago Puelo, and Los Alerces National Parks. Data from Administración de Parques Nacionales, Bariloche, Argentina, 2006.

forests in the western part of the region (Administración de Parques Nacionales 2006, unpubl. data). Furthermore, the tree-ring record of fire scars implies that some of the fires in the wettest forests dominated by Fitzroya were ignited by lightning because of the remote locations in areas unlikely to have been transited by Native Americans (Veblen et al. 1999). What are the medium-term ecological consequences of recent fire events? The ecological consequences of recent major fire events are sitespecific and highly contingent on the reproductive modes of the dominant plant species, post-fire weather, and post-fire herbivory. Analyses of fire–vegetation relationships in a 780 000 ha study area from 1985 to 1999 showed that shrublands, dominated by Nothofagus antarctica and Chusquea bamboos and other resprouters, are proportionally more affected by fire than adjacent tall mesic forests (Mermoz et al. 2005). Tall shrublands dominated mainly by N. antarctica and Chusquea bamboos show a positive association with burnt areas over this large area, and analyses at a fine spatial scale show that once shrubland patches are ignited, they tend to burn completely until a topographic break or a less-flammable vegetation type is encountered. In contrast, tall forests dominated by N. dombeyi or N. pumilio burn less extensively than expected on the basis of the area occupied by each forest type. Patches of the subalpine N. pumilio forests often tend to serve as natural fire breaks, except under the most severe

734

Int. J. Wildland Fire

drought conditions, when they also burn (Mermoz et al. 2005). This is consistent with matched historical (∼1900) and modern photographs showing sharp fire boundaries between burned shrublands and unburnt forests (Veblen and Lorenz 1988). Several factors may contribute to the higher flammability of the shrublands (Veblen et al. 2003): (1) high density of multistemmed small trees and shrubs; (2) rapid fuel recovery due to resprouting; (3) heavy loads of flammable epiphytes (e.g. Usnea lichens) and vines (e.g. Mutisia spp.) that provide fuel ladders; (4) rapid accumulation of dead biomass from partial crown dieback of the shrub and tree species, and from annual foliage replacement of vines; (5) location on steep, north-facing mid-slopes where fuels are prone to desiccation; and (6) dominance of shrub species with high contents of volatile substances. In contrast, tall Nothofagus forests have coarser, moister fuels, and lack vertical continuity between the understorey and canopy. The differential susceptibility of shrublands and mesic forests to fire creates important feedbacks that facilitate long-term shifts in the relative abundance of these landscape components. The life history characteristics of the dominant tree and shrub species and the effects of introduced herbivores in this landscape play key roles in determining post-fire vegetation patterns. As a broad generalisation, in the Argentine Lake District, vegetation types characterised by resprouting tall shrubs and small trees have expanded at the expense of tall tree species that are all obligate seed reproducers. At many sites, in the long-term absence of fire, these shrublands can be successionally replaced by forests, and can be considered quickly recovering, fire-prone systems (Veblen et al. 2003; Kitzberger et al. 2005a, 2005b). In contrast, the tall forests are dominated by tree species (mainly Nothofagus dombeyi, N. pumilio, and Austrocedrus chilensis) that depend exclusively on reproduction by seed, which is a slower and less certain mechanism of post-fire recovery (Veblen et al. 2003; Kitzberger et al. 2005a, 2005b). Under favourable conditions of seed availability and post-fire weather, these obligate seed reproducers form dense post-fire populations (Veblen and Lorenz 1987; Villalba and Veblen 1997). However, their regeneration may be slow or fail entirely owing to exceptionally severe fires, fire-induced edaphic changes, lack of seed sources, post-fire herbivory, or unfavourable weather (Veblen et al. 2003; Kitzberger et al. 2005a; Tercero-Bucardo et al. 2007). Post-fire herbivory can be particularly important to long-term conversion of tall forests to shrublands because the resprouting shrubs resist browsing better than the seed-producing tall tree species (Raffaele and Veblen 1998, 2001; Kitzberger et al. 2005b). When fires are followed by intense herbivory from introduced mammals, tall forests are converted to shrublands that appear to be self-replacing because of the positive feedback between fire and shrub cover. Thus, the tall forests are slowly recovering, fire-sensitive systems. The more fire-prone shrublands occur both in a zone of drier conditions and as patches on edaphically unfavourable sites surrounded by mesic forests. As the connectivity of shrublands increases throughout the landscape, the positive feedback on fire occurrence and spread is expected to increase. Overall, under a continued trend towards warmer and drier conditions, increased fire ignitions by humans and by lightning are likely to result in relatively permanent displacement of tall forest by shrublands due to the positive

T. T. Veblen et al.

feedback of shrubland fuels and fire, which is likely to be exacerbated by livestock.

Fire in the Araucanía region of Chile (38◦ to 39◦ 30 S) What are the historical precedents of the recent large, severe fire events? In Chile’s Araucanía (IXth Region; ∼37◦ 30 to 39◦ 30 S), ∼20 000 ha burned in December to March 2001–02. Thirty per cent of the area burned was in the Araucaria–Nothofagus cover type, and most of the burned area was within Chilean NPs (González and Veblen 2007). One of the fires burned over 60% of the total area of Tolhuaca NP (38◦ 10 S), which was created primarily to protect Araucaria forests (González 2005). Given the importance of Araucaria araucana as a national symbol of Chilean conservation areas, the large, severe fires of 2001–02 triggered substantial public concern and raised questions about appropriate policy and management responses (Echeverría 2002; CONAF 2002). In addition, the fact that the most extensive fires were ignited by lightning contradicted the widespread belief that fires in the Araucaria–Nothofagus cover type of this region were exclusively of human origin (CONAF 2002). Unaware of any role for lightning-ignited fires in this forest type, resource managers and others in the forestry industry initially blamed the ignitions on the native Mapuche people (González 2005). Indeed, the historical fire statistics reported by the Chilean Forest and Park Service do not list lightning as a potential cause; instead, all fires are assumed to be anthropogenic (CONAF 2002). For the Araucanía region (IXth Region) as a whole, the area burned in the 2001–02 fire season was a record since the beginning of reliable record-keeping in 1971 but was not much greater than the extent of burning in 1978–79 and 1987–88 (Fig. 7). Although data on past fire extent by vegetation type are not available, the extent of burning in the Araucaria–Nothofagus cover type was believed to be unprecedented over the previous 50 years (CONAF 2002). For the period before the modern documentary record of fires (i.e. before the 1970s), however, there is ample evidence of widespread, severe fires in Araucaria–Nothofagus forests in Chile and in the same forest type in adjacent areas in Argentina. Sedimentary records of charcoal in conjunction with fossil pollen document the presence of fire generally in the northern Lake District to over 45 000 years before present (BP) and commonly throughout the Holocene (Heusser 1994). Specifically, for the modern Araucaria–Nothofagus forest type, abundant charcoal dates from 3000 years BP from a site in Argentina 50 km east of the Araucaria areas burned in Chile in 2001–02 (Heusser et al. 1988). The extent and severity (i.e. low-severity surface fires v. high-severity crown fires) cannot be determined from these sedimentary charcoal records, but it is clear that fire has long played an ecological role in Araucaria– Nothofagus forests. Likewise, the origin of these fires – either natural or human-set – cannot be determined from the charcoal records, despite the assertion by Heusser (1994) that presence of charcoal in fossil records from the Lake District could be taken as evidence of paleo-human activity. Although infrequent in theAraucaria–Nothofagus region of both Chile andArgentina, lightning-ignited fires such as those of the 2001–02 fire season do occur in this cover type.

Fire in the Andean–Patagonian Nothofagus forest region

Int. J. Wildland Fire

735

40 000 Region IX: Araucanía 20 000

0 40 000 Region X: Los Lagos

Hectares

20 000

0 Region XI: Aysén 20 000

0 20 000 Region XII: Magallanes 10 000

0 1970

1980

1990

2000

Year Fig. 7. Area burned per year for the four administrative regions of southern Chile. Regions are: Araucanía (IXth Region, ∼37◦ 30 to 39◦ 30 S); Los Lagos (Xth Region, ∼37◦ 30 to 39◦ 30 S); Aysén (XIth Region, ∼43◦ 30 to 48◦ 40 S); and Magallanes (XIIth Region, ∼48◦ 40 to 55◦ S). Data from Corporación Nacional Forestal, Santiago, Chile, 2006.

Even though it is not possible to determine the number of prehistoric fires ignited by lightning v. Native Americans, it is known that the native population of Pehuenche and Mapuche used fire to hunt guanaco (Cox 1863; Musters 1871; Fonck 1896) and to clear undergrowth to facilitate collection of Araucaria seeds, one of their staple foods (Montaldo 1974; Aagesen 1998). Although administrative control and permanent Chilean settlement of the northern Lake District did not occur until ∼1883, by the 17th century the indigenous population had adopted livestock introduced by the Spaniards (Bengoa 2000), which would have motivated burning of vegetation to improve pasture. The sharp increase in tree-ring evidence of fire in Tolhuaca NP and Villarrica NP (39◦ 35 S) in the late 19th century implies a substantial increase in burning associated with Euro-Chilean settlement after ∼1883 (Fig. 8). Extensive intentional burning of Araucaria–Nothofagus forests is reported in association with the settlement of the region by Euro-Chileans in the 1880s (Elizalde 1958; Donoso and Lara 1996; Lara et al. 1996). Fire was a preferred tool for opening forest areas to create pasture areas. Fire scars indicate relatively frequent fires through the 1950s, followed by a decline in fires with the establishment of NPs in the areas studied (Fig. 8). Based on a fire history derived from tree rings inVillarrica NP, the extent and severity of the 2002 fire event were not unique over

the past several centuries (González et al. 2005). High-severity widespread events are indicated both by the locations of firescarred remnant trees and post-fire cohorts for the years 1827, 1909, and 1944. Overall, the historic (i.e. pre-1900) fire regime of this cover type appears to be a mixed-severity regime in which both low-severity and less frequent high-severity fires occur. Likewise, tree-ring data from the area burned in 2002 inTolhuaca NP indicate there has been at least one other fire event (∼1762) of comparable extent to the 2002 event (González 2005). In general, the severe and widespread fires that burned Araucaria– Nothofagus forests of this region in 2002 are within the range of the historic fire regimes that have shaped this forested landscape.

How have modern land-use practices contributed to recent fire events? The fire suppression–fuels accumulation model Given that the 2002 fires were mainly ignited by lightning and that there were no recent forestry activities (either logging or plantations of exotic conifers) in the NPs burned, the only relevant land-use hypothesis to consider is that fire suppression or release from livestock grazing significantly increased fuel quantities during the late 20th century. Since infrequent, severe fires

736

Int. J. Wildland Fire

T. T. Veblen et al.

Araucaria

Tolhuaca

Villarrica

1660 1680 1700 1720 1740 1760 1780 1800 1820 1840 1860 1880 1900 1920 1940 1960 1980 2000 2020

Year Fig. 8. Composite fire-scar records for Araucaria forests in Tolhuaca and Villarrica National Parks. Fire dates before 1650 are not included. Dates of fire scars are indicated by short vertical lines. After: González and Veblen (2007).

before the 1950s played a key role in shaping the Araucaria– Nothofagus forest structures (González et al. 2005), it is unlikely that fire suppression practised mostly since the 1960s has significantly altered the natural pattern of fuel accumulation. Fire scars and stand age data indicate that in the Villarrica NP area, most Araucaria stands originated after a relatively few widespread high-severity fire events (González et al. 2005). The area affected by the 2001–02 fires consisted of a mosaic of tall forest of Araucaria–Nothofagus forests, Nothofagus antarctica shrublands, and grasslands. Although it is plausible that heavy grazing by livestock at one time could have reduced fuels and affected fire behaviour, any effects of release of fuels from grazing pressure on fire behaviour in 2001–02 appears minor compared with the effects of climatic variability (discussed below). It is unlikely that any increase in grass fuels associated with removal of livestock in NPs and reduction in grazing since the 1960s had a major influence on fire extent or severity in 2001–02. In fact, tree-ring evidence indicates that relatively widespread and high-severity fires occurred in 1909 and 1944 in the Araucaria area of Villarrica NP when the area was heavily utilised by livestock (González et al. 2005). What are the relationships of recent fire events to climatic variability and trends? Drought was clearly the primary factor responsible for the extent and severity of the 2001–02 fire events. Based on the nearest long-term climate station, which is Temuco, located ∼100 km to the south-west of the area burned, the 2001–02 fire season was preceded by 4 consecutive years of spring (October–December) drought (Fig. 9). Two of those previous springs (1998 and 1999) were the third and second driest springs, respectively, in the 1963–2006 record. December of the 2001–02 fire season recorded the third highest mean temperature and third lowest precipitation for December in the 1963–2006 record. Four successive years of drought may have had a cumulative effect on

moisture content of coarse woody fuels, and dry conditions in December 2001 clearly predisposed the vegetation to burning during the lightning storms of 2002. Tree-ring fire dates from Araucaria–Nothofagus forests in Villarrica NP and regional climate records from the coast of Chile to the eastern foothills of the Andes in Argentina have been used to characterise the association of fire with climatic variability over the period 1899 to 1997 (González and Veblen 2006). Over this period, fire years were associated with aboveaverage temperatures of the same fire year, and with a tendency towards below-average spring–summer precipitation for fire season and the 2 preceding years (González and Veblen 2006). At the spatial scale of the entire Araucanía region, 1987, the year of second most extensive burning, is also associated with protracted drought from 1980 through 1985 (Fig. 9). Based on fire history from 1690 to 2000 in Araucaria– Nothofagus forests in Villarrica NP and a tree-ring proxy record of spring–summer moisture availability, years of widespread fire are associated with low moisture availability resulting from a combination of above-average temperature and belowaverage precipitation in late spring–summer (González and Veblen 2006). Analogous to the pattern described for the Argentine Lake District, Pacific sea surface temperatures as manifested by variations in ENSO are important in promoting these warm, dry conditions and associated fire. Years of high fire activity coincide with warm and dry summers following El Niño events (González and Veblen 2006). The time of transition from a La Niña event to an El Niño event is favourable for fire because of the dry winter–spring conditions associated with the former and the warm summers associated with the latter (Montecinos and Aceituno 2003). Similar to the pattern for the Argentine Lake District, the strong connection of years of widespread fire in the Chilean Araucanía region to variations in Pacific sea-surface temperatures implies that fire in this region will be affected by future hemispheric-scale trends in temperature.

Fire in the Andean–Patagonian Nothofagus forest region

Int. J. Wildland Fire

737

Temuco Spring (Oct.–Dec.) temperature

2 1 0 ⫺1 Departures (SD) from the mean

⫺2 3

Spring (Oct.–Dec.) precipitation

2 1 0 ⫺1 ⫺2

⫺1.0

Spring (Oct.–Dec.) aridity index

⫺0.5

0.0 1950

1960

1970

1980

1990

2000

Year Fig. 9. Spring temperature, precipitation, and aridity index departures (standard deviations) from mean values for the Temuco climate station. The curve in the upper two panels is an 11-year moving average of the respective value. The aridity index (plotted inversely) is De Martonne’s (1926). The dot indicates the record fire season of 2001–02 in the Araucanía region of Chile.

What are the medium-term ecological consequences of these fire events? As a species, Araucaria araucana can both resist damage from fire and recover after being severely burned. Mature trees have thick bark that protects them from surface fires, and a crown shape, which places foliage out of reach of surface fires (Veblen et al. 1995). Also, they possess basal epicormic buds that may sprout after fire, and protected terminal buds on branches allowing continued growth after crown fire (Tortorelli 1947; Montaldo 1974; Veblen et al. 1995). Seedlings produce long tap roots that may be five to six times as long as the shoot, and lateral root length in mature trees can be as great as 30 m (Veblen et al. 1995). Araucaria araucana can occasionally reproduce asexually by root suckering (Veblen et al. 1995). By these means, mature A. araucana frequently survive fire, although small individuals typically do not (Veblen et al. 1995; González et al. 2005; Peñaloza 2006). Where severe fires kill most trees in an Araucaria stand, they commonly regenerate along with Nothofagus species as long as seed sources are present (Veblen et al. 1995; González et al. 2005). Three years after the severe fire in February 2002 in the Araucaria–Nothofagus forest of Tohuaca NP, post-fire vegetation was sampled in a severely and a moderately burned site (González and Veblen 2007). At the severely burned sites, 100% of the canopy trees of both Araucaria and Nothofagus spp. were

dead, and at the sites of moderate burn severity, high percentages (e.g. 30 to over 50%) were alive and in many cases showed no visible signs of fire damage in the crown (González and Veblen 2007). Within a period of 3 years, these sites showed substantial recovery of vegetation in terms of plant cover and species richness by means of new establishment from seed and vegetative reproduction from rhizomes, roots, bulbs and other subterranean organs.The bamboo Chusquea culeou was the most dominant species, but all the dominant tree species (A. araucana, N. dombeyi, N. pumilio, and N. nervosa) were also present as seedlings. Overall, initial data collection and observations 3 to 5 years following the 2002 Tolhuaca fire indicates post-fire recovery to forest conditions similar to those that existed pre-fire (González and Veblen 2007). Potentially significant to the initial establishment of tree seedlings of the dominant tree species is the fact that during the 2 years following the fire, spring moisture condition was not below average (Fig. 9). Fire in other areas of Nothofagus forests in the southern Andean region The two case studies presented above are the only areas for which existing knowledge of fire history and fire ecology is sufficient to address the four central questions of the present synthesis. Although less complete, preliminary research from other areas

738

Int. J. Wildland Fire

in the Nothofagus region of southern South America will be briefly presented in this section. Were the areas burned in the late 1990s–2000s unusually large? Over the period from 1971 through 2005, the 40 722 ha burned in the Los Lagos region of Chile (Xth Region; ∼37◦ 30 to 43◦ 30 S) in the 1997–98 season is many times greater than the average and more than 65% greater than any other year of widespread fire (Fig. 7). The same year was also the peak in area burned for the Aysén District (∼43◦ 30 to 48◦ 40 S) since 1971. In the Magallanes region (XIIth Region; ∼48◦ 40 to 55◦ S), the 2004– 05 season was a near-record year exceeded only by the 1985–86 fire season. Even though single-year records or near records for area burned occurred in the late 1990s to 2005 in all four administrative regions, this time period does not stand out as a time of unusually widespread burning in comparison with the previous three decades (Fig. 7). Qualitative descriptions of earlier episodes of widespread burning especially in the 1940s and late-1800s to early 1900s in all four of the Chilean regions also suggest that the areas burned since the mid-1990s are not unusually large in comparison with the previous 100 to 150 years (Elizalde 1958; Pérez 1958; Donoso and Lara 1996). Furthermore, sedimentary records of charcoal document at least low levels of charcoal in all vegetation types of the region south of 38◦ S throughout the Holocene, including even the wet coastal rainforests (Heusser 1994; Haberle et al. 2000). In some cases, sharp increases in charcoal appear to indicate increased human activities, either during the pre-European period or in association with the arrival of Europeans or their livestock as early as the 1600s (Huber and Markgraf 2003; Huber et al. 2004). Tree-ring studies for the large area of Chile south of ∼40◦ S are in an incipient stage but document the long-term presence of fire in even the wettest rainforests. Initial tree-ring data on fire history from the long-lived Fitzroya cupressoides in the coastal mountains of the Los Lagos District document fire occurrence at relatively long intervals (on average >100 years) since at least 1300 (Lara et al. 2003). These rare fire events appear to be associated with infrequent droughts affecting the coastal rainforests at ∼40◦ S (Lara et al. 2003). On-going research in this same area based on large fire-scar datasets from F. cupressoides should further elucidate the role of fire in these forests over the past millennium (J. Sibold, unpubl. data). Tree-ring fire histories from Pilgerodendron uviferum, the other long-lived conifer of the southern rainforests, also show substantial promise in elucidating the roles of anthropogenic v. natural fires and their relationships to climatic variation in forests at 43◦ to 48◦ S over the past several hundred years (A. Holz, unpubl. data). Have modern land-use practices increased the risk of large, severe fires? To what extent have modern (i.e. post-1970) land-use practices such as fire suppression or planting of exotic trees altered the potential for high-severity fire in the Nothofagus forest region in general? For the rainforest areas, it is unlikely that fire suppression has significantly altered fuel accumulation and consequently the potential for more severe fires. In these wet forests,

T. T. Veblen et al.

fuel quantity does not limit fire ignition or spread; exceptional drought is the main limitation to fire occurrence in these usually wet forests. The fire suppression–fuels buildup hypothesis is more plausible in xeric Nothofagus woodlands and steppe that are widespread in the eastern Aysén and Magallanes Districts, where it is likely that low-severity surface fires played some role in shaping these ecosystems. However, the historic frequency of fire (i.e. pre-1950s) in these ecosystems is not known, and the effectiveness of modern fire suppression dating in most areas from the 1960s–70s is not known. A land-use that is more likely altering the potential for fire spread is the widespread planting of exotic trees. Throughout the IXth and Xth Regions of Chile, large percentages of native forests and agricultural lands were converted to plantations of exotic conifers (especially Pinus radiata) since the 1970s (Lara et al. 1996). Likewise, in the XIth and XIIth Regions, vast areas of steppe and sites formerly occupied by Nothofagus pumilio forests have been planted with exotic pines such as P. ponderosa (Lara et al. 1996; Donoso and Lara 1996). Conversion of native Nothofagus forests and grass-dominated steppe to homogeneous plantations of highly flammable conifers has substantially increased the risk of large, crown fires in these regions. Exurban development is occurring in southern Chile, especially in the Los Lagos and Araucanía regions where vacation homes and resorts have grown spectacularly since the early 1990s. Likewise, in the Aysén and Magallanes areas, vast areas that in the 1970s were sparsely settled are now popular national and international destinations for tourism. The increasing human presence throughout the Nothofagus region of the southern Andes is likely to increase the number of accidental and arson fires. Such appears to have been the case in 2005, when a fire ignited by a careless tourist burned 11 685 ha in Torres del Paine NP, making 2005 the year of greatest fire extent in the Magallanes region since 1985. Are recent climatic trends increasing the risk of large, severe fires? In the Nothofagus forest region, both short-term studies based on instrumental records as well as tree-ring studies of fire history and climatic variation clearly show the dependence of years of widespread fire on exceptional drought. Years of widespread fire in different parts of this extensive region have been shown to be dependent on drought at monthly, seasonal, annual, and supra-annual time scales that can be driven by either decreased precipitation or increased temperatures (Kitzberger and Veblen 2003; Lara et al. 2003; González and Veblen 2006). At a subcontinental scale, temperature trends over the past ∼100 years follow the warming pattern established for the southern hemisphere (Villalba et al. 2003). Based on instrumental climate records, the mean hemisphere temperature rose by between 0.3◦ and 0.6◦ C since the late 19th century, and by ∼0.2◦ to 0.3◦ C since the mid1970s (Villalba et al. 2003). Tree-ring records from 37◦ to 55◦ S indicate that 20th-century temperatures through 1989 rose by 0.5◦ to 0.9◦ C above the 1640–1899 means (Villalba et al. 2003). Furthermore, the temporal pattern of temperature fluctuations since 1856 is strongly linked to global sea surface temperatures anomalies and is likely to continue to follow the warming hemispheric and global warming trend (Villalba et al. 2003). The

Fire in the Andean–Patagonian Nothofagus forest region

relatively sharp increase in spring temperatures observed after 1976 in the Bariloche and Temuco climate records (Figs 5 and 9) is representative of a general post-1976 warming trend for the entire region from latitudes 37◦ to 55◦ S (Villalba et al. 2003). In addition to this recent warming trend, at least in the Argentine Lake District there has been a documented increase in the number of lightning-ignited fires since the mid-1970s (Kitzberger and Veblen 2003). Furthermore, a recent analysis of data from numerous climate stations on both sides of the Andes showed a declining trend in mean annual precipitation for latitudes 40◦ to 47◦ S for the 1950–2000 period (Aravena 2007). Finally, general circulation models predict a decline in precipitation for most of this region during the next century (Vera et al. 2006; Christensen et al. 2007), which will further promote fire occurrence and, for some forest types, will impede post-fire regeneration (Tercero-Bucardo et al. 2007). Conclusions When events occur at relatively long intervals, such as years of widespread fires in mesic forests, it is difficult to detect temporal trends in the absence of precise long-term records. Such is the case for many of the areas affected by severe, widespread fires in the late 1990s and early 2000s in the Nothofagus forest region of southern South America. Documentary records of fire are either short (i.e. reliable Chilean records start only in 1970s) or are limited in spatial coverage (i.e. Argentine NPs). However, these records support the tentative conclusion that widespread, severe fires of the 1990s–2000s were not uniquely large in comparison with the previous 30 to 60 years. Likewise, qualitative historical descriptions imply that fire events of comparable scale occurred in the same forest ecosystems at least several times over the previous 100 to 150 years. Although the scale of recent fires appears to be within the historic range of variability of the past several centuries in the Nothofagus forest region, the cumulative effects on the landscape of land-use changes and past episodes of drought-induced fires appear to be creating conditions outside the HRV of the past several centuries. This is clearly the case for large parts of the Argentine Lake District, where positive feedbacks between fire and more flammable fuel structures have contributed to conversion of significant areas of mesic forests to tall shrublands. The inhibitory effects of livestock on post-fire recovery of forests further add to this positive feedback, resulting in the potential for more widespread fires in the future. At a regional scale in both Argentina and Chile, the potential for greater fire spread has been increased by massive conversion of relatively fire-resistant native forests to plantations of fire-prone exotic conifers. At the same time, potential for anthropogenic ignitions has increased as a result of rapid growth of exurban development and recreational use of some parts of this region. Overall, the most important driver of large, severe fire events in the Nothofagus region in the 1990s–2000s was interannual climatic variability similar to that documented from tree rings for the past several centuries. There has been a trend towards higher temperatures and increased drought that is evident throughout the region for the post-1976 period. In at least parts of the region, there also has been a substantial increase in the frequency of lightning storms and lightning-ignited fires in the

Int. J. Wildland Fire

739

last few decades. The current multidecadal trend towards higher temperatures evident over most of southern South America and linked to global anthropogenic warming is likely to continue to increase the frequency and severity of extreme droughts during the 21st century. These climate trends in combination with the noted land-use changes imply that the large fire events of the late 1990s–2000s may be the initiation of a longer-lasting trend towards increasingly widespread fire with accumulating changes in landscape structure. Acknowledgements Research was supported by grants from the US National Science Foundation (Award No. 0117366), the National Geographic Society, the Council for Research and Creative Work of the Graduate School of the University of Colorado, International Foundation for Science (IFS, D/3124–2), DID-Universidad Austral de Chile (S-2005–08), Núcleo Científico Milenio (FORECOS-MIDEPLAN PO4–065-F, Chile), Fundación Andes-Chile (C-13860), and the Agencia Nacional de Promoción Científica y Tecnológica (ANPCyT, Argentina) PICT 97–01–2268. T.K. and E.R. are researchers of Consejo Nacional de Investigaciones Científicas y Tecnológicas (CONICET).

References Aagesen D (1998) Indigenous resource rights and conservation of the monkey-puzzle tree (Araucaria araucana, Araucariaceae): a case study from Southern Chile. Economic Botany 52, 146–160. Aravena JC (2007) Reconstructing climate variability using tree rings and glacier fluctuations in the southern Chilean Andes. PhD dissertation, University of Western Ontario, Canada. Bengoa J (2000) ‘Historia del Pueblo Mapuche: Siglo XIX y XX.’ (Editorial Lom: Santiago, Chile) Blackhall M, Raffaele E, Veblen TT (2008) Cattle affect early post-fire regeneration in a Nothofagus dombeyi–Austrocedrus chilensis mixed forest in northern Patagonia, Argentina. Biological Conservation 141, 2251–2261. Bond WJ, Woodward FI, Midgley GF (2005) The global distribution of ecosystems in a world without fire. The New Phytologist 165, 525–538. doi:10.1111/J.1469-8137.2004.01252.X Christensen NL, Bartuska AM, Brown JH, Carpenter S, D’Antonio C, Francis R, Franklin JF, MacMahon JA, Noss RF, Parsons DJ, Peterson CH, Turner MG, Woodmansee RG (1996) The report of the Ecological Society of America committee on the scientific basis for ecosystem management. Ecological Applications 6, 665–691. doi:10.2307/ 2269460 Christensen JH, Hewitson B, Busuioc A, Chen A, Gao X, Held I, Jones R, Kolli RK, Kwon TW, Laprise R, Magaña Rueda V, Mearns L, Menéndez CG, Räisänen J, Rinke A, Sarr A, Whetton P (2007) Regional climate projections. In ‘Climate Change 2007: the Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change’. (Eds S Solomon, D Qin, M Manning, Z Chen, M Marquis, KB Averyt, M Tignor, HL Miller) pp. 1–94. (Cambridge University Press: New York) CONAF (2002) ‘Antecedentes sobre impacto de incendios forestales en la IX región. Documento de visita a la zona afectada por incendios.’ (Corporación Nacional Forestal de Chile: Temuco, Chile) Covington WW, Moore MM (1994) South-western ponderosa forest structure: changes since Euro-American settlement. Journal of Forestry 92, 39–47. Cox G (1863) Viajes a las regiones septentrionales de Patagonia 1862–1863. Anales de la Universidad de Chile 23, 3–239 and 437–509. Daniels LD, Veblen TT (2001) ENSO effects on temperature and precipitation of the Patagonian–Andean region: implications for biogeography. Physical Geography 21, 223–243.

740

Int. J. Wildland Fire

Daniels LD, Veblen TT (2004) Spatio-temporal influences of climate on altitudinal treeline in Northern Patagonia. Ecology 85, 1284–1296. doi:10.1890/03-0092 De Martonne E (1926) Une nouvelle fonction climatologique: l’indice d’aridité. Meteorologie 2, 449–458. Donoso C, LaraA (1996) Utilización de los bosques nativos en Chile: pasado, presente y futuro. In ‘Ecología de los Bosques Nativos de Chile’. (Eds JJ Armesto, CVillagran, MKArroyo) pp. 363–387. (Editorial Universitaria: Santiago, Chile) Echeverría C (2002) Ecological disaster in southern Chile. New York Botanical Garden Plant Talk 28, 14 Elizalde R (1958) ‘La sobrevivencia de Chile.’ (Departamento de Conservación y Administración de Recursos Agrícolas y Forestales, Ministerio de Agricultura: Santiago, Chile) Fonck F (1896) ‘Libro de los Diarios de Fray Francisco Menéndez.’ (CF Niemeyer: Valparaiso, Chile) Fonck F (1900) ‘Viajes de Fray Francisco Menéndez a Nahuelhuapi Publicados i Comentados por Francisco Fonck.’ (CF Niemeyer: Valparaiso, Chile) Garreaud RD, Aceituno P (2007) Atmospheric circulation and climatic variability. In ‘Physical Geography of South America’. (Eds TT Veblen, A Orme, K Young) pp. 45–59. (Oxford University Press: New York) Giese BS, Urizar C, Fuckar NS (2002) Southern Hemisphere origins of the 1976 climate shift. Geophysical Research Letters 29, 1014. doi:10.1029/2001GL013268 González ME (2005) Fire history data as reference information in ecological restoration. Dendrochronologia 22, 149–154. doi:10.1016/J.DENDRO. 2005.04.001 González ME, Veblen TT (2006) Climatic influences on fire in Araucaria araucana–Nothofagus forests in the Andean cordillera of south central Chile. Ecoscience 13, 342–350. doi:10.2980/I1195-6860-13-3-342.1 González ME, Veblen TT (2007) Incendios en bosques de Araucaria araucana y consideraciones ecológicas al madereo de aprovechamiento en áreas recientemente quemadas. Revista Chilena de Historia Natural (Valparaiso, Chile : 1983) 80, 243–253. González ME, Veblen TT, Sibold JS (2005) Fire history of Araucaria– Nothofagus forest in Villarrica National Park, Chile. Journal of Biogeography 32, 1187–1202. doi:10.1111/J.1365-2699.2005.01262.X Haberle SG, Szeicz JM, Bennett KD (2000) Late Holocene vegetation dynamics and lake geochemistry at Laguna Miranda, XI Region, Chile. Revista Chilena de Historia Natural (Valparaiso, Chile : 1983) 73, 655–669. Heusser CJ (1994) Paleoindians and fire during the late Quaternary in southern South America. Revista Chilena de Historia Natural (Valparaiso, Chile : 1983) 67, 435–442. Heusser CJ, Rabassa J, Brandani A, Stuckenrath R (1988) Late-Holocene vegetation of the Andean Araucaria region, Province of Neuquén, Argentina. Mountain Research and Development 8, 53–63. doi:10.2307/ 3673406 Huber UM, Markgraf V (2003) European impact on fire regimes and vegetation dynamics at the steppe–forest ecotone of southern Patagonia. The Holocene 13, 567–579. doi:10.1191/0959683603HL647RP Huber UM, Markgraf V, Schabitz F (2004) Geographical and temporal trends in Late Quaternary fire histories of Fuego-Patagonia, South America. Quaternary Science Reviews 23, 1079–1097. doi:10.1016/ J.QUASCIREV.2003.11.002 Kitzberger T, Veblen TT (2003) Influences of climate on fire in northern Patagonia, Argentina. In ‘Fire and Climatic Changes in Temperate Ecosystems of the Western Americas’. (Eds TT Veblen, W Baker, G Montenegro, TW Swetnam) pp. 290–315. (Springer Verlag: NewYork) Kitzberger T, Veblen TT, Villalba R (1997) Climatic influences on fire regimes along a rain forest-to-xeric woodland gradient in northern Patagonia, Argentina. Journal of Biogeography 24, 35–47. doi:10.1111/ J.1365-2699.1997.TB00048.X

T. T. Veblen et al.

Kitzberger T, Steinaker DF, Veblen TT (2000) Effects of climatic variability on facilitation of tree establishment in northern Patagonia. Ecology 81, 1914–1924. Kitzberger T, Raffaele E, Heinemann K, Mazzarino J (2005a) Effects of fire severity in a north Patagonian subalpine forest. Journal of Vegetation Science 16, 5–12. doi:10.1658/1100-9233(2005)016[0005: EOFSIA]2.0.CO;2 Kitzberger T, Raffaele E, Veblen TT (2005b) Variable community responses to herbivory in fire-altered landscapes of northern Patagonia, Argentina. African Journal of Range & Forage Science 22, 85–91. Landres P, Morgan P, Swanson F (1999) Overview of the use of natural variability concepts in managing ecological systems. Ecological Applications 9, 1179–1188. Lara A, Donoso C, Aravena JC (1996) La conservación de los bosques nativos de Chile: problemas y desafíos. In ‘Ecología de los Bosques Nativos de Chile’. (Eds JJ Armesto, C Villagrán, MK Arroyo) pp. 335–362. (Editorial Universitaria: Santiago, Chile) LaraA, Wolodarsky-FrankeA,Aravena JC, Cortés M, Fraver S, Silla F (2003) Fire regimes and forest dynamics in the Lake Region of south-central Chile. In ‘Fire and Climatic Change in Temperate Ecosystems of the Western Americas’, Ecological Studies 160. (Eds TT Veblen, W Baker, G Montenegro, TW Swetnam) pp. 322–342. (Springer Verlag: NewYork) Mermoz M, Kitzberger T, Veblen TT (2005) Landscape influences on occurrence and spread of wildfires in Patagonian forests and shrublands. Ecology 86, 2705–2715. doi:10.1890/04-1850 Montaldo P (1974) La bio-ecología de Araucaria araucana (Mol.) Koch. Boletín Instituto Forestal Latino-Americano de Investigación y Capacitación 46/48, 3–55. Montecinos A, Aceituno P (2003) Seasonality of the ENSO-related rainfall variability in central Chile and associated circulation anomalies. Journal of Climate 16, 281–296. doi:10.1175/1520-0442(2003)016 <0281:SOTERR>2.0.CO;2 Moreno FP (1897) Reconocimiento de la región andina de la República Argentina. Apuntes preliminares sobre una excursión a los Territorios de Neuquén, Rio Negro, Chubut y Santa Cruz. Revista del Museo de La Plata 8, 1–180. Musters GC (1871) ‘At Home with Patagonians: a Year’s Wandering over Untrodden Ground from the Straits of Magellan to the Rio Negro.’ (John Murray: London, UK) Myers R (2006) ‘Living with Fire – Sustaining Ecosystems and Livelihoods through Integrated Fire Management.’ (The Nature Conservancy Global Fire Initiative: Tallahassee, FL) Peñaloza R (2006) Zonificación de severidades de incendio y determinación de la potencial restauración natural de Araucaria araucana (Mol) K. Koch en el Parque Nacional Tolhuaca, Novena Región, Chile. BSc (Forest Engineering) thesis, Universidad Austral de Chile, Valdivia. Pérez R (1958) ‘V. Recuerdos del Pasado: 1814–1860.’ (Carlos de Vidts: Santiago) Raffaele E, Veblen TT (1998) Facilitation by nurse shrubs on resprouting behavior in a post-fire shrubland in northern Patagonia, Argentina. Journal of Vegetation Science 9, 693–698. doi:10.2307/3237287 Raffaele E, Veblen TT (2001) Effects of cattle grazing on early post-fire regeneration of matorral in north-west Patagonia, Argentina. Natural Areas Journal 21, 243–249. Rothkugel M (1916) ‘Los Bosques Patagónicos.’ (Ministerio de Agricultura: Buenos Aires, Argentina) Schlichter T, Laclau P (1998) Ecotono estepa–bosque y plantaciones forestales en la Patagonia norte. Ecología Austral 8, 285–296. Schoennagel TL, Veblen TT, Romme WH (2004) The interaction of fire, fuels and climate across Rocky Mountain forests. Bioscience 54, 661–676. doi:10.1641/0006-3568(2004)054[0661:TIOFFA]2.0.CO;2 Sedjo RA (1999) Potential for carbon forest plantations in marginal timber forests: the case of Patagonia, Argentina. Discussion paper 99–27. (Resources for the Future: Washington, DC)

Fire in the Andean–Patagonian Nothofagus forest region

Shlisky A, Waugh J, Gonzalez P, González ME, Manta M, Santoso H, Nuruddin AA, Rodríguez-Trejo DA, Swaty R, Schmidt D, Kaufmann M, Myers R, Alencar A, Kearns F, Johnson D, Smith J, Zollner D (2007) ‘Fire, Ecosystems and People: Threats, and Strategies for Global Biodiversity Conservation – Report of the Global Fire Initiative.’ (The Nature Conservancy: Boulder, CO) Steffen H (1909) ‘Viajes de Exploración: Estudio en la Patagonia Occidental.’ (Imprenta Cervantes: Santiago, Chile) Tercero-Bucardo N, Kitzberger T, Veblen TT, Raffaele E (2007) A field experiment on climatic and herbivore impacts on post-fire tree regeneration in north-western Patagonia. Journal of Ecology 95, 771–779. doi:10.1111/J.1365-2745.2007.01249.X Tortorelli LA (1947) ‘Los Incendios de Bosques en la Argentina.’(Ministerio de Agricultura: Buenos Aires) Veblen TT (2007) Temperate forests of the Southern Andean Region. In ‘Physical Geography of South America’. (Eds TT Veblen, A Orme, K Young) pp. 217–231. (Oxford University Press: New York) Veblen TT, Lorenz DC (1987) Post-fire stand development of Austrocedrus– Nothofagus forests in Patagonia. Vegetatio 73, 113–126. Veblen TT, Lorenz DC (1988) Recent vegetation changes along the forest/ steppe ecotone in northern Patagonia. Annals of the Association of American Geographers. Association of American Geographers 78, 93–111. doi:10.1111/J.1467-8306.1988.TB00193.X Veblen TT, Kitzberger T, Lara A (1992) Disturbance and forest dynamics along a transect fromAndean rain forest to Patagonian shrubland. Journal of Vegetation Science 3, 507–520. doi:10.2307/3235807 Veblen TT, Burns BR, Kitzberger T, Lara A, Villalba R (1995) The ecology of the conifer of southern South America. In ‘Ecology of the Southern Conifers’. (Eds NJ Enright, RS Hill) pp. 120–155. (Melbourne University Press: Melbourne) Veblen TT, Donoso C, Kitzberger T, Rebertus AJ (1996) Ecology of southern Chilean and Argentinean Nothofagus forests. In ‘The Ecology and Biogeography of Nothofagus Forests’. (Eds TT Veblen, RS Hill, J Read) pp. 293–353. (Yale University Press: New Haven, CT)

Int. J. Wildland Fire

741

Veblen TT, Kitzberger T, Villalba R, Donnegan J (1999) Fire history in northern Patagonia: the roles of humans and climatic variation. Ecological Monographs 69, 47–67. Veblen TT, Kitzberger T, Raffaele E, Lorenz D (2003) Fire history and vegetation changes in northern Patagonia, Argentina. In ‘Fire and Climatic Change in Temperate Ecosystems of the Western Americas’, Ecological Studies 160. (Eds TT Veblen, W Baker, G Montenegro, TW Swetnam) pp. 265–295. (Springer Verlag: New York) Vera C, Silvestri G, Liebmann B, Gonzalez P (2006) Climate change scenarios for seasonal precipitation in South America. Geophysical Research Letters 33. doi:10.1029/2006GL025759 Villalba R, Veblen TT (1997) Regional patterns of tree population age structures in northern Patagonia: climatic and disturbance influences. Journal of Ecology 85, 113–124. doi:10.2307/2960643 Villalba R, Boninsegna JA, Veblen TT, Schmelter A, Rubulis S (1997) Recent trends in tree-ring records from high elevation sites in the Andes of northern Patagonia. Climatic Change 36, 425–454. doi:10.1023/A:1005366317996 Villalba R, Cook ER, Jacoby GC, D’Arrigo RD, Veblen TT, Jones PD (1998) Tree-ring based reconstructions of northern Patagonia precipitation since AD 1600. The Holocene 8, 659–674. doi:10.1191/095968398669095576 Villalba R, Lara A, Boninsegna JA, Masiokas M, Delgado S, Aravena JC, Roig FA, Schemelter A, Wolodarsky A, Ripalta A (2003) Large-scale temperature changes across the southern Andes: 20th-century variations in the context of the past 400 years. Climatic Change 59, 177–232. doi:10.1023/A:1024452701153 Whitlock C, Bianchi MM, Bartlein PJ, Markgraf V, Marlon J, Walsh M, McCoy N (2006) Postglacial vegetation, climate, and fire history along the east side of the Andes (lat. 41–42.5◦ S), Argentina. Quaternary Research 66, 187–201. doi:10.1016/J.YQRES.2006.04.004 Willis B (1914) ‘El Norte de la Patagonia.’(Dirección de Parques Nacionales: Buenos Aires) Manuscript received 13 November 2007, accepted 24 October 2008

http://www.publish.csiro.au/journals/ijwf