Increased aridity drives post-fire recovery of Mediterranean forests towards open shrublands

Corrigendum to New Phytologist 225 (2020), 1500–1515, doi: 10.1111/nph.16252.

Since its publication, the authors of Baudena et al. (2020) have identified an error for the set of parameter values representing flammability in Table 2. In this correction, the authors would also like to report that, when using the flammability values as originally published in Baudena et al. (2020; i.e. a factor 2 larger than those actually used in the simulations), the main results do not change qualitatively (see Supporting Information Figs S1, S2 to this correction).

Namely, when increased aridity was simulated as negatively affecting oak post-fire recovery and colonization rate, while positively affecting the community flammability, the authors observed that the forest state was resilient to the separate impact of fires and increased aridity. Yet, water stress could convert forests into open shrublands by hampering post-fire recovery and at the same time either increasing flammability or decreasing the oak forest colonization rate (or both). A tipping point (emerging from bistability of the open shrubland and forest state) was detected at intermediate levels of aridity (Fig. S1). In the ‘short-term’ run, that is a century, the authors observed again that the probability of a mixed successional community becoming an oak forest after 100 yr decreased drastically with increasing aridity (moving from bottom left to top right in Fig. S2, e.g. with flammability equal to 1.5 times the baseline value as published in table 2 in Baudena et al., 2020). The main differences between the two parameter sets were that the effects of aridity were more dramatic in Figs S1 and S2, as their baseline flammability (given in table 2 in Baudena et al., 2020) was twice as high as the baseline flammability that we actually used in figs 3 and 4 in Baudena et al. (2020) (as reported here in Table 2).

We apologize to our readers for this mistake.

The authors would like to kindly acknowledge Matilde Torrassa for finding the error in the original version of the paper.