Continental-scale empirical evidence for relationships between fire response strategies and fire frequency

Abstract

Introduction

Fire is a fundamental ecological process for many ecosystems on Earth (Bond & Keeley, 2005; Andela et al., 2019; McLauchlan et al., 2020), which shapes the evolution of traits that allow organisms to thrive in fire-prone environments (Keeley et al., 2011; He & Lamont, 2018; Pausas & Bond, 2019). A key focus of research has been the mechanisms that enable plants to persist in fire-prone environments. Past research has identified two main strategies adopted by plant species: resprouting from surviving tissues (hereafter ‘resprouting’) and postfire germination from seed (hereafter ‘postfire seeding’) (Lamont et al., 1991; Whelan, 1995; Bond & van Wilgen, 1996). While some species are capable of one mechanism alone, so-called ‘obligate’ resprouters and seeders, others are capable of both, called ‘facultative’ species. Theoretical and empirical research surrounding these mechanisms suggests that the two strategies exist at opposite ends of a spectrum of resource allocation (Iwasa & Kubo, 1997; Bell, 2001; Bowen & Pate, 2017), with varying costs and benefits, and depending on the fire regime. It follows that patterns in the relative proportion of these fire response strategies could vary across gradients of fire regime characteristics, such as fire frequency or severity. Although established in theory (Hilbert, 1987; Bellingham & Sparrow, 2000; Bond & Midgley, 2003; Pausas & Keeley, 2014b), these patterns have remained largely untested on a broad scale, mainly due to the lack of large-scale data on fire regimes and species regeneration mechanisms.

In fire-prone ecosystems, the success of different regeneration mechanisms is dependent on the fire regime. A fire regime captures the typical event- and frequency-driven characteristics of fires in a given place within an ecological time frame (Gill, 1975) and comprises multiple parameters, including fire frequency, intensity (rate of heat energy release), severity (biological impacts on above- and below-ground vegetation), type (ground, surface, crown, mixed), size and seasonality (McLauchlan et al., 2020). Fire regimes vary markedly between vegetation types, such as in the mesic forests of eastern Australia and boreal forests of North America where fires burn infrequently but often intensely (Gill & Catling, 2002; Keeley & Pausas, 2022), vs the tropical and subtropical savannas in southern Africa and South America which undergo frequent, low-intensity fires (Archibald et al., 2013; Lehmann et al., 2014). Two major drivers of selection on plant response strategies are fire frequency and severity. Fire frequency defines the length of time between fires in which plants can grow and maintain resprouting organs and/or reach reproductive maturity and produce a sufficient seed bank. Severity relates to the amount of biomass consumed and levels of plant mortality and reproductive success. Severity is influenced by the intensity of fire, which is in turn dependent on frequency, with frequent fire regimes constrained to mostly moderate-intensity fires and rare fire regimes capable of producing both low- and high-intensity fires (Archibald et al., 2013). Here, we focus on one major fire regime characteristic, fire frequency, which has been at the core of theory relating to plant–trait relationships (Hilbert, 1987; Bellingham & Sparrow, 2000; Bond & Midgley, 2003; Pausas et al., 2004). However, the interplay of fire frequency and severity necessitates their joint interpretation. Globally, ecosystems range from experiencing frequent fire, close to 100 fires per century, to very infrequent fire, less than one fire per century (Archibald et al., 2013).

Theory suggests that the frequency of resprouting and postfire seeding strategies of woody plants should vary with factors that modify the relative survivorship of adult and juvenile plants (Pausas & Keeley, 2014b). Pausas & Keeley (2014a) likened obligate resprouting and obligate seeding strategies to perennial and annual life history strategies, when fire events are considered equivalent to annual cycles. Obligate resprouters follow a longer-lived, ‘perennial’ life history as they live and reproduce through many fire intervals (iteroparity) (Pausas & Keeley, 2014a). Conversely, obligate seeding species follow a shorter-lived, ‘annual’ life history, with mature individuals suffering high mortality after severe fire and thus depending on a single reproductive event per generation (semelparity) (Bond & van Wilgen, 1996; Pausas & Keeley, 2014a). Pausas & Keeley (2014a) proposed that selection towards these opposing strategies depends on the ratio of parent to offspring survivorship; resprouting should be selected for when adult plants are more likely to survive a fire relative to juveniles. Resprouting is therefore expected to be more common under less severe fire regimes, and varying levels of fire severity determine the effectiveness of different types of resprouting (basal, epicormic and axillary bud) (Moreira et al., 2009; Kenefick et al., 2024). Facultative species exist between these extremes, resprouting and seeding being nonmutually exclusive strategies.

Along with fire severity, fire frequency is also a major driver of selection on resprouting and seeding strategies. Investment in resprouting capacity may be wasted when fires are very rare, but may also be inviable in environments where fires are both frequent and severe, as there is insufficient opportunity for recovery between disturbances (Iwasa & Kubo, 1997; Bellingham & Sparrow, 2000; Pausas & Keeley, 2014a). Bellingham & Sparrow (2000) hypothesised that the ability to resprout would increase as the frequency of severe disturbances increases, until a threshold where resprouting becomes less viable and so ability to resprout then declines. Similarly, postfire seeding is predicted to be more prevalent at intermediate fire frequencies (Hilbert, 1987; Lamont et al., 1991), as a balance between two competing risks. Seeders suffer an immaturity risk when fire return intervals (FRIs) are too short (Keeley et al., 1999), but can also experience a senescence risk when FRIs are longer than the longevity of the plants and seed bank combined (Keeley, 1986).

While discussion around plant–fire relationships has predominantly been centred around woody plants, herbaceous plants have been comparatively understudied, despite being an important group that makes up much of the world's most fire-prone ecosystems (Mouillot & Field, 2005; Keeley & Pausas, 2022). We define ‘woody’ as having a prominent aerial stem that lasts through time and changing environmental conditions (Zanne et al., 2014); in practice, this corresponds to a longer lifespan of the aboveground part of the plant as well as a taller potential height. Recently, Simpson et al. (2021) extended the model by Bellingham & Sparrow (2000) to grasses; however, a small difference in their predictions for grasses was that, although resprouting may decline at very high fire frequencies, it could still be a common strategy, as many grasses can resprout from protected underground buds or insulated leaf bases (Klimešová & Klimeš, 2003; Simpson et al., 2021). Additionally, herbaceous seeders were hypothesised to be less restricted by immaturity risk as their time to maturation is relatively short, supported by findings that the ratio of grass seeders to resprouters was higher at very high frequencies (Simpson et al., 2021).

To fulfill contrasting life histories, seeders and resprouters may also employ different strategies of resource allocation and growth. Seeders might allocate more resources to rapid aboveground growth and early reproduction, whereas resprouters might allocate more resources to storage organs and protective structures that improve survival and regrowth after fire (Iwasa & Kubo, 1997; Bell, 2001; Pausas et al., 2004; Bowen & Pate, 2017). These resource allocation strategies could also be manifested in leaf economics traits (Wright et al., 2004), such as leaf mass per area (LMA) and leaf nitrogen (N) content. Fast-growing seeders are expected to have ‘quick-return’ leaves, with low LMA, high leaf nutrients (including N), high leaf turnover and high rates of photosynthesis and respiration, whereas the opposite is expected for slower-growing resprouters (Wright et al., 2004). Past studies investigating leaf traits across fire response strategies have found varying results across climates and ecosystems (Ackerly, 2004; Pausas et al., 2004; Paula & Pausas, 2006; Saura-Mas & Lloret, 2007; Vivian & Cary, 2012), thus raising questions about how generally leaf traits relate to fire strategy. Moreover, the fundamental differences between woody and herbaceous growth forms may lead to different patterns of leaf traits across fire response strategies (e.g. in grasses; Simpson et al., 2021). Generally, it is known that herbaceous and woody species differ in their leaf traits (Towers et al., 2024), but it remains unclear whether there are further differences within herbaceous species with respect to fire response, and whether such differences are consistent across herbaceous and woody species. In addition to fire history, both leaf traits and the proportion of resprouters relate to plant lifespan, with very short lifespan species (i.e. annuals) thought typically not to resprout following fire. Thus, nonfire climatic factors that ecologically select for very short lifespan species may also lead to a greater proportion of nonresprouters. Whatever its origin, the inability of species in these climates to resprout following fire may still have important implications as fire moves into both new seasons and new parts of the world.

Despite an abundance of theory, the consequences of fire have rarely been evaluated at large biogeographic scales, with data to span hundreds to thousands of species. However, a flourishing of biodiversity resources, such as species distribution, fire occurrence and trait data, now enables such questions to be addressed. Recently, Simpson et al. (2021) tested the effects of fire frequency on the distribution of 734 grass species, globally. While this global analysis of grasses represented a significant advance in our understanding of plant–fire relationships for herbaceous species, the scope of the study was still constrained to just a single plant family with a particular adaptation: having a basal meristem. Here, we use large-scale empirical data to investigate the distribution of fire response strategies at an unprecedented scale, by quantifying characteristics for > 9500 species spread across the entire continent of Australia. Specifically, we ask the following questions:

1. What is the distribution of fire frequencies experienced by Australian plant taxa?
2. Do the fractions of plant species that are resprouting and seeding support the hypothesised hump-shaped response with fire frequency, in both woody and herbaceous plants?
3. Within woody and herbaceous species, do leaf traits (LMA and leaf N content) differ between resprouters and nonresprouters, or between postfire seeders and nonseeders?