COMMENTARY

Paradise burnt: How colonizing transform landscapes with fire

David M. J. S. Bowmana,1 and Simon G. Haberleb aSchool of Plant Science, University of , , Tasmania 7001, ; and bDepartment of Archaeology and Natural History, College of Asia and the Pacific, Australian National University, Canberra, Australian Capital Territory 0200, Australia

striking feature of Southern forests are highly responsive to slight Hemisphere landscapes is the changes in fire regime. occurrence of grasslands in McWethy et al. are able to disclose the A  regions that are climatically impact of Maori burning on the South suitable for forests (Fig. 1). Ecologists Island by analyzing variation in macro- and biogeographers working in these scopic charcoal data, pollen records, lim- southern lands have developed a range of nobiota (diatom and chironomid taxa), theories to account for the biogeographic and sediment geochemistry from 16 small anomaly of grassland–forest mosaics lakes in watersheds spanning the domi- (1–7). Broadly speaking, these theories nant east–west precipitation gradient (8). divide into those that privilege the im- They find that in the decades immediately portance of an ensemble of environmental after Maori colonization there was a factors, including fire, or those that stress burst of severe fire events. This “Initial the legacy of landscape burning. Burning Period” was destructive of forest The report by McWethy et al. in PNAS (8) cover and an initiator of soil erosion and provides incontrovertible evidence that associated nutrient influx in all but the anthropogenic burning transformed tem- wettest and topographically fire-sheltered perate forested landscapes on the South watersheds. After the Initial Burning Pe- Island of . They show that riod there was less frequent and less severe Polynesian (Maori) firing commenced burning, which enabled forests in higher- shortly after colonization around A.D. rainfall areas to recover. McWethy et al. 1280 and transformed 40% of the original are able to reject the hypothesis that the forest cover of the island to grassland burst of fire activity after Maori settlement and fern-shrubland. There is little room for was associated with anomalous doubting their findings given the elegant patterns given that the severe fires were integration of a range of paleoecological neither synchronous nor associated with methodologies, very precise dating, and anomalous summer temperatures inferred a high level of replication across the island. from silver pine ( colensoi) This report will spark renewed interest in tree-ring chronologies calibrated against the relative importance of fire, humans, instrumental data (8). The importance and in shaping forest–grassland of human ignition in forest destruction landscape mosaics worldwide (9). was further underscored by a surge of New Zealand is a superb model system burning coincident with mid-19th century for understanding the effect of human European settlement that also caused fire usage on temperate forests for the grassland expansion, soil erosion, and following reasons. The north–south consequent changes to the chemistry of mountain chain, which forms a spine lake sediments. across the South Island, causes a pro- Fig. 1. Grassland forest mosaics are a feature of McWethy et al. recognize that the nounced rainfall, elevation, and many landscapes: (A)South conversion of forest to grassland and gradient from dry [<650 mm mean annual West Tasmania, (B) the interior of , associated soil changes in drier sites due precipitation (map)] low-elevation water- (C) Drakenburg Mountains, South , and (D)the  South Island of New Zealand. Various theories have to Maori and then European landscape sheds [<300 m above sea level (masl)] burning is a clear example of a fire- > been proposed to explain these biogeographically in the east to wet ( 1600 mm map) mon- puzzling vegetation patterns that in general em- initiated ecological feedback process. tane (>800 masl) watersheds in the west. phasize the importance of anthropogenic landscape Indeed such precise chronicling of the Across this gradient there is a high density burning or climatically driven fire regimes (1–7, 9, 17, development of grassland–forest mosaic of small lakes that trap sediments, char- 18). The report by McWethy et al. (8), based on an is important in developing the theory coal, and pollen. The comparatively recent extremely robust paleoecological database, leaves of alternate stable states that posits that colonization by Maori some 800 y ago no doubt that the grasslands in the South Island extreme disturbance, such as recurrent are derived from a short burst of intense landscape and Europeans in the mid 19th century landscape firing, can cause ecological sys- 14 burning associated with the colonization of the island means that C dating can be used to de- by Maori 800 y ago. This Initial Burning Period was tems to rapidly switch from one state to velop robust chronologies to chart envi- not associated with any anomalous climatic con- another and that ecological feedbacks ronmental change and fire activity. Before ditions. (Photo credits: A and C, David Bowman; B, human settlement fire activity was re- William Bond; D, Matt McGlone). markably low, with charcoal records Author contributions: D.M.J.S.B. and S.G.H. wrote the paper. indicating fires once or twice per millen- The authors declare no conflict of interest. fi nium. In marked contrast to Australian no adaptations to survive re or any de- See companion article 10.1073/pnas.1011801107. fi eucalypt-dominated temperate forest pendency on re to regenerate and are 1To whom correspondence should be addressed. E-mail: flora, New Zealand forest trees have easily killed by burning (10). Thus, these [email protected].

www.pnas.org/cgi/doi/10.1073/pnas.1016393108 PNAS Early Edition | 1of2 Downloaded by guest on October 1, 2021 provide strong resistance to a return to the posed global generalization of the over- argued that the expanses of subtropical predisturbance state (11). The McWethy whelming importance of climate in driving grasslands in the drier areas of Mada- et al. study contributes to tackling one of fire activity before the industrial revolu- gascar are not derived from human fire use the key challenges presented by alternative tion (A.D. 1750), based on a pioneering after colonization some 2000 y B.P., in state theory: discrimination of the feed- analysis of a compilation of charcoal contrast to a well-established theory de- back processes that maintain the alterna- records from around the world (13). veloped by several generations of French tive state from the disturbance regime that scientists (3, 17). Conversely, on the basis caused the switch from one state to an- of pollen and ecological evidence, Fletcher other (e.g., forest to grassland) (12). In the Anthropogenic burning and Thomas (18) claim that that the New Zealand case there is little doubt transformed temperate sedgelands that dominate more than 40% the driver of the switch was a burst of of the humid temperate western Tasma- severe burning by humans. Yet recovery forested landscapes nian landscape are the legacy of sustained to forest is clearly facilitated by high- Aboriginal burning, which effectively rainfall but inhibited by low-rainfall envi- on the South Island preserved a Pleistocene landscape these ronments. Such climate–anthropogenic colonists found some 35,000 y ago. fire interactions are of great significance of New Zealand. Contrasts of fire activity in the Southern in developing predictive capacity for the Hemisphere clearly can provide crucial consequences of climate change (13). insight into how human burning has Settling the issue of the origin of grass- It remains to be seen whether, if the high- modified landscape, the relative impor- land–forest mosaics also has immediate resolution paleoecological approach used tance of biotic history, and the interplay conservation policy implications (14, 15). by McWethy et al. were to be applied with climate variation (18). Such a project, For example, Bond and Parr (16) suggest to other flammable landscapes, it will of which the McWethy et al. study (8) that the perception, which they reject, challenge this paradigm that before can be considered the first step, will also that tropical grasslands are anthropogeni- industrialization climate, not humans, was bring to international attention the in- cally derived reduces their conservation the primary driver of both fire activity sights and data that have been accumu- value and makes them prime targets for and vegetation patterns. lated by generations of Southern Hemi- agriculture and “reaforestation” programs The time is certainly ripe to undertake sphere ecologists, archaeologists, and designed to provide wood products and comparative analyses of southern hemi- biogeographers who have been studying capture carbon. sphere island systems given the current the impacts of fire on Southern Hemi- The report in PNAS (8) highlights the contrasting perspectives of the importance sphere landscapes and islands, often in importance of combining a range of pa- of climate and anthropogenic burning in intellectual isolation from each other (1–5, leoecological techniques, sampling across the Southern Hemisphere. For example, 17). We believe that, when combined climate gradients, having high levels of researchers working on the islands of carefully, the findings from such “mini- replication, and using consistent method- Tasmania and Madagascar have reached worlds” that serve as model systems hold ologies to effectively disaggregate the opposite conclusions regarding climati- the answer to globally significant questions effects of human fire use and natural fire cally anomalous treeless vegetation. On about the impact of anthropogenic land- regimes. This robust study provides an the strength of ecological and historical scape burning, climate change, and their important exception to the recently pro- biogeographic evidence, Bond et al. have interactions on forest cover (19, 20).

1. Ellis RC (1985) The relationships among eucalypt forest, ics, ed Cochrane MA (Springer-Praxis, Heidelberg), pp 14. Motzkin G, Foster DR (2002) Grasslands, heathlands grassland and rainforest in a highland area in north- 87–111. and shrublands in coastal New England: Historical in- eastern Tasmania. Aust J Ecol 10:297–314. 8. McWethy DB, et al. (2010) Rapid landscape transformation terpretations and approaches to conservation. J Bio- — in South Island, New Zealand following initial Polynesian 2. Jackson WD (1968) Fire, air, water and an ele- geogr 29:1569–1590. settlement. Proc Natl Acad Sci, 10.1073/pnas.1011801107. mental . Proc Ecol Soc Aust 3:9–16. 15. Willis KJ, Gillson L, Brncic TM (2004) Ecology. How ‘vir- 9. Bond WJ, Woodward FI, Midgley GF (2005) The global 3. Bond WJ, Silander JA Jr., Ranaivonasy J, Ratsirarson J gin’ is virgin rainforest? Science 304:402–403. distribution of ecosystems in a world without fire. New (2008) The antiquity of Madagascar’s grasslands and 16. Bond WJ, Parr CL (2010) Beyond the forest edge: Ecol- Phytol 165:525–537. the rise of C4 grassy biomes. J Biogeogr 35:1743–1758. 10. Burrows G, et al. (2010) A wide diversity of epicormic ogy, diversity and conservation of the grassy biomes. 4. McGlone MS (2001) The origin of the indigenous grass- – structures is present in Myrtaceae species in the north- Biol Conserv 143:2395 2404. lands of southeastern South Island in relation to pre- ern Australian savanna biome—implications for adap- 17. Willis KJ, Gillson L, Virah-Sawmy M (2008) Nature or human woody ecosystems. N Z J Ecol 25:1–15. tation to fire. Aust J Bot 58:493–507. nurture: The ambiguity of C4 grasslands in Madagas- 5. Webb LJ (1964) An historical interpretation of the grass 11. Scheffer M, Carpenter S, Foley JA, Folke C, Walker B car. J Biogeogr 35:1741–1742. balds of the Bunya Mountains, South Queensland. (2001) Catastrophic shifts in ecosystems. Nature 413: 18. Fletcher M-S, Thomas I (2010) The origin and temporal Ecology 45:159–162. 591–596. development of an ancient cultural landscape. J Bio- 6. Fensham RJ, Fairfax RJ (2006) Can burning restrict euca- 12. Petraitis PS, Latham RE (1999) The importance of scale geogr 37:2183–2196. lypt invasion on grassy balds? Austral Ecol 31:317–325. in testing the origins of alternative community states. 19. Cochrane MA (2003) Fire science for rainforests. Nature 7. Bowman DMJS, Wood SW (2009) Fire driven land cover – Ecology 80:429 442. 421:913–919. change in Australia and W.D. Jackson’s theory of the 13. Marlon JR, et al. (2008) Climate and human influences 20. Bowman DMJS, et al. (2009) Fire in the Earth system. fire ecology of southwest Tasmania. Tropical Fire Ecol- on global biomass burning over the past two millen- Science 324:481–484. ogy: Climate Change, Land Use and Ecosystem Dynam- nia. Nat Geosci 1:697–702.

2of2 | www.pnas.org/cgi/doi/10.1073/pnas.1016393108 Bowman and Haberle Downloaded by guest on October 1, 2021