Bark thickness and fire regime: another twist.

Bark is the outermost covering of stems in woody plants and plays a fundamental protective role. Recently I hypothesized that ‘at the global scale, a significant proportion of the variability in bark thickness is explained by the variability in fire regimes’, and specifically predicted that frequent low intensity fires select for thick bark (Pausas, 2015). In addition, I suggested that differentiating between inner and outer bark thickness would help us gain a better understanding of the functional role of bark, especially in nonfire prone ecosystems. Based on an understanding of the selection pressure by fire and on other plant traits, I showed that some fire regimes select for thick bark at the base of the trunk, others select thick bark on the whole plant (stem, branches, twigs), while other fire regimes do not select for thick bark – and thus relatively thin barks are the more likely to be observed (Table 1). However, the paucity of available data at a global scale limited an empirical demonstration of the proposed framework. A new paper has now provided evidence for the fire hypothesis of bark thickness at a global scale. Rosell (2016) sampled bark thickness in woody species from 18 sites in different climates and fire regimes, and has demonstrated that fire regime was the main environmental factor explaining variability in bark thickness (after accounting for plant size; Fig. 1). But perhaps the most valuable contribution ofRosell is that, in addition to total bark thickness, she accurately measured inner and outer bark thickness, and showed that they behave differently: the role of fire is especially relevant for explaining outer bark thickness (Fig. 1; see Schafer et al., 2015, for similar results), while inner bark does not seem to provide protection for the cambium fromheat. This provides a step forward in our understanding of the ecology of bark. These results were found despite the relatively simple estimation of fire regime (semiquantitative fire frequency, 1–5), compared with alternative parameters (climate) that were more precise and variable. This simple estimation of fire regime is understandable as fire history is not as available as climate data (remotely-sensed fire activity could be a possible source; Pausas & Ribeiro, 2013). Below I take the opportunity to comment on how we could advance our understanding of the role of fire in shaping bark thickness by considering more detailed fire regime information coupledwith some plant lifehistory traits. Fire regime is the complex combination of fire characteristics that prevails in a given area, and includes frequency, intensity, seasonality and type of fuels consumed (Keeley et al., 2012). For its relevance in relation to the bark, I will focus on first fire frequency, defined as the fire return interval in relation to plant longevity; and second on the fire intensity, defined as the flame height in relation to the height of canopy fuels. The latter variable defines two very contrasted fire regimes: understory (or surface) fires (i.e. when the flame height is lower than the overstory; crowns are not consumed by the fire) and crown fires (otherwise). This distinction is important because the predictions of bark thickness differ in these two fire regimes (Table 1), and the linear expectation of bark thickness and fire frequency may not apply when mixing understory and crown fire regimes. In fact, Rosell showed that Mediterranean ecosystems have thin barks despite frequent fires; this is exactly the prediction for those ecosystems as they are subject to frequent crown fires (Table 1; Fig. 3 in Pausas, 2015). Stating that fire regimes explain an important part of the variability in bark thickness does not mean that bark thickness and fire frequency should show a strong positive correlation; for example, the high frequency of understory fires selects for thick basal bark in trees but not in the coexisting understory plants; and the high frequency of crown fires does not select for thick bark at all (Keeley & Zedler, 1998; Pausas, 2015). In fact, I expect bark thickness to be related to the frequency of low intensity fires (Pausas, 2015). When Rosell analysed her data separating short (< 2 m) and tall (> 2 m) species, the effect of fire on bark thickness disappears in the former and increases in the latter (Fig. 1). This is probably because many of the short species grow in shrublands subject to crown fires or are in the understorey of forests; in such cases thin barks are expected