Understanding a flammable planet – climate, fire and global vegetation patterns
2005; Wiley; Volume: 165; Issue: 2 Linguagem: Inglês
10.1111/j.1469-8137.2004.01301.x
ISSN1469-8137
Autores Tópico(s)Coastal wetland ecosystem dynamics
ResumoThe extraordinary intellectual achievement of the 19th century German botanist Andreas Schimper was his book Plant-Geography upon a Physiological Basis (Schimper, 1903). Through sheer force of imagination and by drawing on numerous written observations from around the world, he described the correspondence between global climate and vegetation zones. Such 19th century global ecological syntheses were superseded in the 20th century because attention was directed to specific questions using the hypothetico-deductive approach. However, growing concern over global environmental change and the advent of powerful space-age and computer technologies has seen the pendulum swing away from narrowly focused analyses back towards global synthesis. ‘A world without fire has fundamentally different forest zones than occur on our real, highly flammable planet’ Rather than being based on observation and induction, the 21st century syntheses are powered by models that enable the mechanistic integration of data collected across spatial scales and disciplinary boundaries. The predictions possible are sufficiently accurate to be testable against independent observations and meta-analyses of existing datasets. The paper in this issue by Bond et al. (pp. 525–538) is an exemplar of this intellectual moment. They have provided the first evidence that global vegetation patterns are shaped by landscape fire. Their study is based on the disciplined marshalling of relevant field and satellite observations and strategic application of existing mechanistic Dynamic Global Vegetation models (DGVMs) based upon physiological processes. Their approach demonstrates a new way of ecological thinking that provides profound insights into global ecological processes and the evolution of the biosphere. Despite the fact that satellite sensors were not originally designed to map landscape fire, an unexpected spin-off of global remote sensing was the demonstration of the ubiquity of landscape fire on every vegetated continent (Cochrane, 2003; Justice et al., 2003). The effect of landscape fire on global vegetation patterns is implicit in several DGVMs because they include ‘fire modules’ that introduced frequent disturbances to modelled vegetation patterns and processes. Bond et al. (2005) asked a beguilingly simple question – what happens if these fire modules are switched off? They found that a world without fire has fundamentally different forest zones than occur on our real, highly flammable planet. Without fire, the extent of forests with >80% tree cover doubled from 26.9% to 56.4% of the vegetated surface of the Earth. Further, more than half (52.3%) of the current global distribution of C4 grasslands was transformed to angiosperm-dominated forest. Of the 41% of C3 grassland that was replaced by forests, 53% were dominated by gymnosperms, 34% by angiosperms and 13% by a mixture of both these taxa. The analysis of Bond et al. (2005) was unable to capture postfire secondary successional sequences within their broad vegetation formations. If they had done so, there is no doubt they would have provided even more startling evidence of landscape fire upsetting the vegetation–climate equilibrium. This would be particularly so for pyrophytic forests such as those dominated by Eucalyptus in Australia and Pinus in the northern hemisphere. Landscape fire has not been a central concern in ecology. Indeed, only in the past decade have books been published outlining the general principles of fire ecology (Whelan, 1995; Bond & van Wilgen, 1996); most knowledge has been regionally focused. Fire ecologists working in specific flammable biotas, such as tropical savannas, mediterranean shrublands and pyrophytic forests, have long appreciated that landscape fire decouples the tight interrelationship between vegetation and climate: the achievement of Bond et al. (2005) has been to unite these disparate finding into a single global perspective. A common inference from regional landscape-scale studies has been that the juxtaposition of patches of forests within a highly flammable matrix is the work of recurrent fires (Fig. 1). Bond et al. (2005) argue that such patterns reflect the evolutionary divergence of fire-adapted and fire-tolerant taxa, a process that has occurred independently on all vegetated continents. Evidence for this evolutionary dichotomisation is largely circumstantial, based mainly on field correlation and fire exclusion. The recent study by Fensham et al. (2003) is a notable exception. They demonstrated that recurrent fires caused the differential survival of evergreen tree species characteristic of fire-prone savannas compared with evergreen tree species from ‘rainforests’ on fire-protected sites (Fig. 2). However, the underlying mechanism that causes this differential response remains unexplained. Savanna fire burning around a topographically fire-protected patch of rainforest. Such a pattern of ‘islands’ of fire-sensitive vegetation in a ‘sea’ of fire-tolerant vegetation is evidence that landscape burning disturbs the tight interrelationship between vegetation type and climate. The juxtaposition of floras with contrasting fire tolerances is also interpreted as representing the evolutionary dichotomisation driven by landscape burning. However, there are few hard data about physiological and morphological bases of fire tolerance or how these traits evolved. (Photographer: David Hancock). Declining mean percentage (±se) of the survival of resprouting species in response to five successive fires in a previously fire-protected savanna fragment in north-east Queensland. Closed circles: rainforest species; open circles: savanna species. The physiological and morphological base of the differential survival between rain forest and savanna species to recurrent fire remains to be elucidated. Significance of Mann–Whitney u-tests are where: NS, P > 0.05; **P < 0.01; ***P < 0.001. (Modified from Fensham et al. (2003) with additional data from R. J. Fensham, unpublished). There are very few examples that demonstrate the evolution of specific features that enable plants to survive fire. Burrows (2002) showed that epicormic buds were situated on the inside rather than the outside of the cambium in some south-east Australian eucalypts and related taxa. Burrows (2002) interpreted this unique anatomical arrangement as an adaptation to recover vegetatively following fire damage. Research by Prior et al. (2003, 2004) has pointed to whole-plant differences between fire-tolerant and fire-sensitive taxa. Eucalypts that dominate vast tracts of fire-prone savanna were found to have photosynthetically less efficient leaves and slower stem growth rates than rainforest tree species that are restricted to small fire-protected sites (Fig. 1). Such whole plant differences associated with fire tolerance may account for some of the variation between climate parameters and leaf functional attributes documented by Wright et al. (2004). The physiological basis for fire tolerance, particularly whole-tree carbon allocation, is a fertile area for research. Whereas some recent studies have provided some microevolutionary insights into plant strategies to survive fire (e.g. Schwilk & Ackerly, 2001), the deeper question of the convergent evolution of fire tolerance between continental floras remains open. Interrogation of the fossil record to determine when flammable vegetation evolved is stymied by the absence of unambiguous morphological features to survive fire. An alternative approach to advance the question of the evolution of flammable floras may be sought by using molecular phylogenies to trace the evolution of unambiguously fire-adaptive traits such as transposed epicormic bud strands (Burrows, 2002). Palaeoecology has proved that landscape fire occurred for millions of years before the advent of fire-wielding hominids. Bond et al. (2005) suggest falling CO2 levels may have stimulated the development of fire-prone C4 grassland that, in turn, greatly increased the frequency of landscape fire. Keeley and Rundel (2003) argue that the development of monsoon climates may be as an important driver as low atmospheric concentrations of CO2. This is because the dry seasons characteristic of monsoon climates are concluded by intense convective storm activity that produce high densities of lightning strikes. The integration of global lightning activity (Fig. 3) in DGVMs would provide far more realistic probability distributions of ignitions than the unrealistic assumption that ignition is not limiting. It would also be instructive to discover the degree of congruence between predicted ‘hot spots’ of natural fire activity and the diversity of fire adapted biotas. Such an analysis may help advance the timing of the evolution of flammable biotas on Earth and to gauge the evolutionary effect of anthropogenic burning. Global lightning activity (number of flashes per km2 per year). These data include both cloud-to-cloud and cloud-to-ground strikes. There is a general concordance between high lightning activity in seasonally dry climates and those areas identified by Bond et al. (2005) as susceptible to vegetation change when fire is ‘switched off’ in Dynamic Global Vegetation Models. Further, there appears to be association between high lightning activity and regions with fire-adapted floras such as the Florida Peninsula, California coast, southern Africa and northern Australia. The v1.0 gridded satellite lightning data were produced by the NASA LIS/OTD Science Team (Principal Investigator, Dr H. J. Christian, NASA/Marshall Space Flight Center) and are available from the Global Hydrology Resource Center (http://ghrc.msfc.nasa.gov). Although it is accepted that indigenous people have moulded landscapes through the use of fire, understanding the extent of this impact is difficult given uncertainty about the background rate of fire activity from lightning. For example, in North American forests it is widely regarded that the impact of Native American burning was negligible because stand-replacing fires are under the control of long-term drought cycles (e.g. Grissino-Mayer et al., 2004). Conversely, it is widely assumed that Aboriginal landscape burning caused a continental-wide transformation of the Australian flora and fauna (Bowman, 1998; Miller et al., 1999). Following the same logic as Bond et al. (2005), a comparison of actual global vegetation patterns with those produced under lightning ignitions alone would help resolve the effect of anthropogenic ignitions, both historically and prehistorically, on changing global vegetation patterns (Fig. 3). The positive feedback between smoke plumes and cloud-to-ground lightning strikes (Lyons et al., 1998), however, may confound a simple causal relationship between the apparent concordance of the spatial distribution of current observations of high lightning activity and fire-tolerant floras. There can be no escaping the increasing global impact of contemporary anthropogenic landscape burning. The increased spatial scale of landscape burning in fire-prone environments reflects failed attempts to totally suppress fires (e.g. Grissino-Mayer et al., 2004) or the breakdown of skilful indigenous fire management (e.g. Bowman et al., 2004). Fire is being used indiscriminately to clear tropical rain forests. An ensemble of positive feedbacks greatly increases the risk of subsequent fires above the extremely low background rate (Cochrane et al., 1999; Cochrane, 2003). Recurrent burning can therefore trigger a landscape-level transformation of tropical rainforests into flammable scrub and savanna. The transformation of tropical rain forest by fire provides insights into the evolution and spread of flammable floras worldwide. Clearly, much remains to be done to bring fire to the same footing as climate variables as a factor driving biogeographic patterns and biogeochemical processes. Discovering the causes of the evolution of flammable vegetation is of great importance in understanding and managing landscape fire, particularly given the accelerating rate of global environmental change. Of prime interest are the effects of climatic variation, atmospheric CO2 concentrations and prehistoric anthropogenic fire use relative to the background rate of lightning ignitions. Bond et al. (2005) provide a vital jolt in developing such global perspective and evolutionary thinking about landscape fire.
Referência(s)