Abstract. Pyrogenic plants dominate many fire-prone ecosystems. Their prevalence suggests some advantage to their enhanced flammability, but researchers have had difficulty tying pyrogenicity to individual-level advantages. Based on our review, we propose that enhanced flammability in fire-prone ecosystems should protect the belowground organs and nearby propagules of certain individual plants during fires. We base this hypothesis on five points: (1) organs and propagules by which many fire-adapted plants survive fires are vulnerable to elevated soil temperatures during fires; (2) the degree to which burning plant fuels heat the soil depends mainly on residence times of fires and on fuel location relative to the soil; (3) fires and fire effects are locally heterogeneous, meaning that individual plants can affect local soil heating via their fuels; (4) how a plant burns can thus affect its fitness; and (5) in many cases, natural selection in fire-prone habitats should therefore favor plants that burn rapidly and retain fuels off the ground. We predict an advantage of enhanced flammability for plants whose fuels influence local fire characteristics and whose regenerative tissues or propagules are affected by local variation in fires. Our ''pyrogenicity as protection'' hypothesis has the potential to apply to a range of life histories. We discuss implications for ecological and evolutionary theory and suggest considerations for testing the hypothesis.

Summary 1.Fire strongly influences plant populations and communities around the world, making it an important agent of plant evolution. Fire influences vegetation through multiple pathways, both above-and belowground. Few studies have yet attempted to tie these pathways together in a mechanistic way through soil heating even though the importance of soil heating for plants in fire-prone ecosystems is increasingly recognized. 2. Here we combine an experimental approach with structural equation modelling (SEM) to simultaneously examine multiple pathways through which fire might influence herbaceous vegetation. In a high-diversity longleaf pine groundcover community in Louisiana, USA, we manipulated fine-fuel biomass and monitored the resulting fires with high-resolution thermocouples placed in vertical profile above-and belowground. 3. We predicted that vegetation response to burning would be inversely related to fuel load owing to relationships among fuels, fire temperature, duration and soil heating. 4. We found that fuel manipulations altered fire properties and vegetation responses, of which soil heating proved to be a highly accurate predictor. Fire duration acting through soil heating was important for vegetation response in our SEMs, whereas fire temperature was not. 5. Our results indicate that in this herbaceous plant community, fire duration is a good predictor of soil heating and therefore of vegetation response to fire. Soil heating may be the key determinant of vegetation response to fire in ecosystems wherein plants persist by resprouting or reseeding from soil-stored propagules. 6. Synthesis. Our SEMs demonstrate how the complex pathways through which fires influence plant community structure and dynamics can be examined simultaneously. Comparative studies of these pathways across different communities will provide important insights into the ecology, evolution and conservation of fire-prone ecosystems.

Organisms capable of rapid clonal growth sometimes monopolize newly freed space and resources. We hypothesize that sequential disturbances might change short-term clonal demography of these organisms in ways that promote formation of monotypic stands. We examined this hypothesis by studying the clonal response of Arundinaria gigantea (giant cane, a bamboo) to windstorm and fire. We studied giant cane growing in both a large tornado-blowdown gap and under forest canopy, in burned and unburned plots, using a split-block design. We measured density of giant cane ramets (culms) and calculated finite rates of increase (lamda) for populations of ramets over three years. Ramet density nearly doubled in stands subjected to both windstorm and fire; the high ramet densities that resulted could inhibit growth in other plants. In comparison, ramet density increased more slowly after windstorm alone, decreased after fire alone, and remained in stasis in controls. We predict that small, sparse stands of giant cane could spread and amalgamate to form dense, monotypic stands (called "canebrakes") that might influence fire return intervals and act as an alternative state to bottomland forest. Other clonal species may similarly form monotypic stands following successive disturbances via rapid clonal growth.

Ecological disturbances frequently control the occurrence and patterning of dominant plants in high-diversity communities like C4 grasslands and savannas. In such ecosystems disturbance-related processes can have important implications for species, and for whole communities when those species are dominant, yet mechanistic understanding of such processes remains fragmentary. Multiple bunchgrass species commonly co-dominate disturbance-dependent and species-rich pine savannas, where small-scale fuel heterogeneity may influence bunchgrass survival and growth following fires. We quantified how fire in locally varying fuel loads influenced dynamics of dominant C4 bunchgrasses in a species-rich pine savanna in southeastern Louisiana, USA. We focused on two congeneric, co-dominant species (Schizachyrium scoparium and S. tenerum) with similar growth forms, functional traits and reproductive strategies to highlight effects of fuel heterogeneity during fires. In experimental plots with either reduced or increased fuels versus controls with unmanipulated fuels, we compared: 1) bunchgrass damage and 2) mortality from fires; 3) subsequent growth and 4) flowering. Compared to controls, fire with increased fuels caused greater damage, mortality and subsequent flowering, but did not affect post-fire growth. Fire with reduced fuels had no effect on any of the four measures. The two species responded differently to fire with increased fuels – S. scoparium incurred measurably more damage and mortality than S. tenerum. Logistic regression indicated that the larger average size of S. tenerum tussocks made them resistant to more severe burning where fuels were increased. We speculate that locally increased fuel loading may be important in pine savannas for creating colonization sites because where fuels are light or moderate, dominant bunchgrasses persist through fires. Small-scale heterogeneity in fires, and differences in how species tolerate fire may together promote shared local dominance by different bunchgrasses.

Numerous bamboos are known to form extensive single-species stands, including species in the United States. Formerly prominent in the southeastern US, canebrakes are dense stands of the bamboos collectively called ''cane' ' [Arundinaria (Michx)]. Canebrakes are now a critically endangered component of the bottomland hardwood forest ecosystem. Cane still occurs in its historic range, primarily in small remnant patches. A poor understanding of the ecological processes that generated large canebrakes limits their restoration and management.We hypothesize that cane's spreading clonal structure enables these bamboos to persist beneath a forest canopy and then respond rapidly to large-scale wind disturbances. We quantified patterns of clonal growth in one cane species, ''giant cane'' [Arundinaria gigantea (Walt.) Muhl.], in a very large tornado-generated canopy gap and in surrounding bottomland hardwood forest in Louisiana. We tested these four hypotheses over a 12-month study period in the large canopy gap: (1) production of new culms should be greater, (2) clonal expansion should be greater, (3) culm damage rate should be reduced, and (4) culm size should be reduced compared to giant cane stands under forest canopy.We found that new culm production in tornado-blowdown plots was twice that in forest plots. Accordingly, culms were younger on average in the tornado blowdown than under forest. Rate of clonal expansion was similar between the two environments, suggesting clonal spread was not disturbance-dependent. With fewer branch-fall impacts, culms in the tornado blowdown were less often damaged. Culms were smaller in tornadoblowdown plots than in forest plots.Giant cane's clonal plasticity should enable it to persist in old-growth bottomland forests by responding to local light conditions. Genets should increase culm production in small gaps and senesce as gaps fill in. Giant cane stands could thereby shift location over time. Wind disturbance that opens forest canopy should trigger redevelopment of denser stands that could merge with other expanding stands into expansive canebrakes. Giant cane's clonal ecology may be a useful model for understanding spreading bamboos and other forest-growing clonal perennials. #

No abstract

No abstract

2017. Groundcover community assembly in high-diversity pine savannas: seed arrival and fire-generated environmental filtering. Ecosphere 8(3):Abstract. Environmental filtering-abiotic and biotic constraints on the demographic performance of individual organisms-is a widespread mechanism of selection in communities. A given individual is "filtered out" (i.e., selectively removed) when environmental conditions or disturbances like fires preclude its survival and reproduction. Although interactions between these filters and dispersal from the regional species pool are thought to determine much about species composition locally, there have been relatively few studies of dispersal 9 filtering interactions in species-rich communities and fewer still where fire is also a primary selective agent. We experimentally manipulated dispersal and filtering by fire (pre-fire fuel loads and post-fire ash) in species-rich groundcover communities of the longleaf pine ecosystem. We tested four predictions: (1) That species richness would increase with biologically realistic dispersal (seed addition); (2) that the immediate effect of increased fuels in burned communities would be to decrease species richness, whereas the longer-term effects of increased fuels would be to open recruitment opportunities in the groundcover, increase species richness, and increase individual performance (growth) of immigrating species; (3) that adding ash would increase species richness; and (4) that increased dispersal would generate larger increases in species richness in plots with increased fuels compared to plots with decreased fuels. We found that dispersal interacted with complex fire-generated filtering during and after fires. Dispersal increased species richness more in burned communities with increased and decreased fuels compared to burned controls. Moreover, individuals of immigrating species generally grew to larger sizes in burned communities with increased fuels compared to burned controls. In contrast to dispersal and fuels, ash had no effect on species richness directly or in combination with other treatments. We conclude that filtering occurs both during fires and in the post-fire environment and that these influences interact with dispersal such that the consequences are only fully revealed when all are considered in combination. Our experiment highlights the importance of considering the dynamic interplay of dispersal and selection in the assembly of species-rich communities.