Alternative fire-driven vegetation states

Traditionally, most vegetation patterns were thought to be driven by environmental conditions (Holdridge 1947; Whittaker 1975). The role of animals in determining patterns of vegetation has also long been recognized (Hairston et al. 1960; Terborgh et al. 2006), given that herbivores consume vegetation and, in turn, are controlled by carnivores and parasites. There is an increasing volume of literature suggesting that fire also plays an important role in determining vegetation patterns in many ecosystems (Bond et al. 2005; Keeley et al. 2012). This is because fire can consume large amounts of plant biomass (Bond & Keeley 2005; Pausas & Keeley 2014a), affect vast areas in short periods and thus play an overwhelming role in some ecosystems (fire-prone ecosystems). One of the clearest pieces of evidence for the role of fire in shaping vegetation is the occurrence of alternative vegetation types maintained by different fire regimes in a given climate. The different flammability of alternative communities generates different fire feedback processes that maintain contrasted vegetation types with clear boundaries in a given environment; and fire exclusion blurs this structure. This has been well documented in tropical landscapes (Bond et al. 2005) that are oftenmosaics of two alternative stable states – savannas and forests – with distinct structures and functions and sharp boundaries. While savannas are subject to frequent grass-fuelled fires (fire-prone, flammable or pyrophilic ecosystems), forests rarely burn (firefree, non-flammable or pyrophobic ecosystems). This strong contrast is because savannas are open ecosystems dominated by highly flammable grasses, while closed forests inhibit grass fuels and generate microclimatic conditions less conducive to fire. Topography may also determine different vegetation types, but topographic differences are not necessary because micro-scale spatial differences, or temporal windows with longer fire intervals at local scales, may enhance the growth of trees and start the negative feedback process of inhibiting fire conditions – and thus abruptly cause a switch from an open to a closed system. Similarly, some rare events can favour a light fire in a forest and feed back to increase flammability and causes a switch to a savanna (Hoffmann et al. 2012;Murphy & Bowman 2012; Dantas et al. 2013). Alternative stable vegetation states with sharp boundaries driven by different fire regimes can be observed in other warm ecosystems (App. S1). In this issue, Paritsis et al. (2015) document the existence of alternative stable vegetation states in a cold temperate ecosystem – the beech (Nothofagus) forests of southern Patagonia. After studying community composition, stands, fuel structures and micro-environmental conditions in burned and unburned Patagonian forests and shrublands, Paritsis et al. suggest that these two plant communities behave as alternative stable states maintained by distinct fire regimes. Forests are dominated by the non-resprouting, thin-barked tree Nothofagus pumilo and rarely burn because this deciduous, broad-leaved tree generates a shady, cool and moist environment that inhibits the growth of light-demanding flammable plants and so reduces fire spread conditions. In contrast, nearby shrublands burn more frequently and are dominated by resprouting shrubs (including the congeneric N. antartica); these communities have more continuous fine fuel (Paritsis et al. 2015) and generate microclimatic conditions more conducive to fire than forests (namely, high insolation and thus higher temperatures, lower humidity and greater exposure to wind). In addition, recurrent fires, as well as livestock grazing, increase plant flammability (Blackhall et al. 2015) and thus enhance vegetation–fire feedback processes. Note that the post-fire regeneration strategy of N. pumilo is better described as non-resprouter (compared with the resprouter N. antartica) rather than obligate seeder (as termed by Paritsis et al. 2015) because the species does not accumulate a persistent seed bank (see Pausas & Keeley 2014b for definitions). Also, Paritsis et al. named N. pumilo forests as