Fire and plant evolution


MEDECOS Special Session on ‘Fire as an evolutionary pressure shaping plant traits’, Los Angeles, CA, USA, September 2011

Throughout much of the history of ecological science, wildfires were viewed as disasters that destroyed ecosystems. Currently, however, most scientists and managers have overcome this short-sighted view and now consider fires as ecological processes that influence structure and function in many ecosystems worldwide. Despite this, there remain several outstanding controversies regarding the role of fire. One area of research and controversy is the potential of plant traits to result in feedback effects on fire regime (e.g. Bond & Midgley, 1995; Schwilk & Caprio, 2011). There is increasing recognition that wildfires are old phenomena on the Earth as there is charcoal evidence during the whole history of terrestrial plants (Glasspool et al., 2004). In addition, past periods with high atmospheric oxygen concentration made vegetation highly flammable even in wet conditions (Belcher et al., 2010; Glasspool & Scott, 2010). Consequently, vegetation–fire feedbacks may have had major effects on past vegetation and even on biome shifts. For instance, fire may have favored the expansion of angiosperms during the Cretaceous (Bond & Scott, 2010), a time with high concentrations of O2 (possibly exceeding current values by 25%). Similarly, the expansion of C4 grasses during the Miocene was favored by fire in extensive seasonally dry areas (Keeley & Rundel, 2005). These studies point to the increasing recognition that fire has been an ancient process on the Earth (Pausas & Keeley, 2009). However, the role of fire as an evolutionary force shaping plant adaptations has remained controversial (Keeley et al., 2011).

‘… in many fire-prone ecosystems alternative vegetation types with quite different flammabilities co-exist on landscapes’

During the recent Mediterranean Ecosystems (MEDECOS) XII conference a special session focused on compiling evidence of fire as an evolutionary process shaping plant traits took place (i.e. fire adaptations). The session benefited from an initial presentation by William Bond (University of Cape Town, South Africa) who provided an overview on the role of fire in biome distribution, with new evidence on the link between fire regime and the Cretaceous expansion of angiosperms (Bond & Scott, 2010). Bond emphasized the feedback role of vegetation on fire regime and pointed out that in many fire-prone ecosystems alternative vegetation types with quite different flammabilities co-exist on landscapes (‘open’ and flammable vs ‘closed’ nonflammable communities). These community types often represent groups of taxa with distinct life-histories and phylogenetic origins. Bond pointed out that the size and dynamics of these co-existing patches has been changing during the history in response to changes in fire regime and O2 and CO2 atmospheric concentration (with low CO2 and high fire activity favoring open ecosystems and with high CO2 and low fire favoring closed ecosystems). Consequently, vegetation has been shaped during the evolutionary history by complex atmospheric processes modifying resource allocation and flammability.

Micro-evolutionary approaches

The role of fire as an evolutionary pressure can be elucidated using both macro- and micro-evolutionary approaches (Fig. 1). Susana Gómez-González (Universidad Bío-Bío, Concepción, Chile) provided a clear example of the micro-evolutionary approach to demonstrate natural selection driven by fire. She showed that populations of an annual species growing under different recent fire histories have different seed traits, and these traits are proven to be heritable and to increase fitness in frequently burnt ecosystems (Gómez-González et al., 2011). Another micro-evolutionary study was presented by Juli G. Pausas (CIDE-CSIC, Spain) suggesting that individuals of Ulex parviflorus (Fabaceae; a shrub of the Mediterranean Basin) growing in populations recurrently burnt were more flammable than individuals of the same species growing in populations arising from old-field colonization (i.e. with fire-independent recruitment). Because U. parviflorus does not resprout but recruits from a soil seed bank after fire (a post-fire seeder species), high flammability increases the chance of opening space for offspring recruitment after fire (Pausas et al., 2012). Thus, if this flammability is heritable, repeated fires would select for individuals with higher flammability, and thus drive trait divergence among different fire regimes. These results provide some field support for the ‘kill thy neighbour’ hypothesis (Bond & Midgley, 1995), but they also highlighted the need for heritability studies to unambiguously demonstrate natural selection driven by fire.

Figure 1.

Approaches to elucidate the role of fire as evolutionary pressure. (a) The macro-evolutionary approach is based on understanding changes in the trait in relation to the different evolutionary pressures (top: different lines are different possible scenarios; different line shades represent periods of different evolutionary pressures; see Keeley et al., 2011), based on the ancestral reconstruction of trait values in dated phylogenies (bottom: symbols of different size and color represent different trait values). (b) The micro-evolutionary approach is based on searching for trait divergences along a current fire activity gradient (dark line suggests local adaptation, while the gray line represents a trait independent of the fire regime). Myr, million years.

Macro-evolutionary approaches

While the micro-evolutionary approach searches for trait divergences in different current selective environments, the macro-evolutionary approach uses dated phylogenies to trace the evolution of traits over long time scales (million years) (Fig. 1). Previous uses of phylogenies for understanding trait evolution in fire-prone ecosystems were able to find evolutionary correlations among traits as well as patterns in the order of appearance of different fire-related traits (Schwilk & Ackerly, 2001; Pausas & Verdú, 2005); however, these studies were limited by the lack of molecular data for dating the phylogenies. The recent blooming of such data is allowing researchers to generate time calibrated phylogenies that can be used to date the origin and evolution of fire-related traits. Byron B. Lamont (Curtin University, Perth, Australia) presented an example of this macro-evolutionary approach for Australian and South African species. He presented recent results on the origin of several fire traits, including resprouting traits (e.g. lignotubers, epicormic buds, rhizomes; Lamont et al., 2011) and serotiny, in the Proteaceae. For example, the origin of serotiny and dead floret retention around the cones (enhanced flammability around serotinous cones) was traced back to 61 million years ago in Banksia (He et al., 2011), a similar age to the origin of epicormic resprouting in Eucalyptus (Crisp et al., 2011). The simultaneous and independent origin of fire traits in these two Australian families suggests that fire was an effective agent of selection as far back as the early Paleocene. The use of older clades could provide an opportunity to trace the origin of fire as a selective force even earlier (Midgley & Bond, 2011), as there is evidence of fire persistence and regeneration traits in old lineages such as conifers and ferns.

Evolutionary role of some fire-related traits

In Mediterranean-climate shrublands, alternative life-histories co-exist, including species that resprout after fire but lack fire-cued seed germination (obligate resprouters), species that resprout and also have fire-cued seed germination (facultative seeders), and those that are killed by fire and are dependent upon post-fire seedling establishment (obligate seeders). Brandon R. Pratt (California State University, CA, USA) focused on physiological trade-offs between facultative and obligate resprouters from Californian chaparral growing in a common garden experiment. He presented evidence suggesting that facultative species have higher growth and lower root to shoot ratios than obligate resprouters (Pratt et al., 2009), a result in the same direction as those reported for the Mediterranean Basin (Paula & Pausas, 2006, 2011). Jon E. Keeley (US Geological Survey, CA, USA) proposed a model of evolution from obligate resprouters (ancestral state) to facultative resprouters, and then to obligate seeders. This model is consistent with phylogenetic studies (Pausas & Verdú, 2005), and it is based on two evolutionary steps driven by different environmental conditions. Resprouting is hypothesized to contribute to persistence on a wide variety of fire-prone landscapes. With increases in aridity and the predictability of high intensity fires capable of creating gaps for seedling recruitment, selection has led to delayed reproduction and a single post-fire pulse of seedling establishment (evolutionary step 1). Further changes in fire regimes that placed a greater premium on seedling recruitment and limited the value of resprouting have led to some lineages abandoning the resprouting habit and evolution of the obligate seeding life history (evolutionary step 2) (Keeley et al., 2012).

Two researchers presented information on two less-studied fire related traits. Byron B. Lamont presented a review of fire-stimulation flowering. He compiled data on 550 species in Australia and South Africa for which there is evidence of fire-stimulated flowering and pointed to the fitness benefits of this trait and its long evolutionary history (Lamont & Downes, 2012). Parker V. Thomas (San Francisco State University, CA, USA) provided evidence that the caching activities of rodents may be a fire adaptation by which Arctostaphylos species place their seeds in safe and protected sites for post-fire germination. However, it remains to understand why rodents would cache Arctostaphylos fruits given that apparently they do not retrieve and consume them (due to the little reward offered by Arctostaphylos fruits). One possible explanation is that caching activities are an anachronism from a time when similar plant species provided larger and palatable fruits, and thus current Arctostaphylos species benefit from what appears to be maladaptive animal behavior.


This session highlighted several key aspects on the role of fire in plant evolution. First, there is good evidence for vegetation-fire regime feedbacks at different spatial and temporal scales, in such a way that plant flammability is a major driver of plant evolution and vegetation distribution. Second, the evidence that fire acts as a selective force is apparent on both micro- and macro-evolutionary scales, suggesting that fire shapes plant traits and generates fire adaptations. And third, that fire is a complex selective pressure – plants adapt to (and, in turn, influence) particular fire regimes rather than fire in the abstract. This is an exciting time for fire ecologists, as fire is now recognized as fundamental for many ecological and evolutionary processes; the coming macro- and micro-evolutionary studies will certainly reinforce many of the ideas drawn during the meeting.