Skip to main content
Dryad logo

Why forest fails to recover after repeated wildfires in Amazonian floodplains? Experimental evidence on tree recruitment limitation


Flores, Bernardo; Holmgren, Milena (2021), Why forest fails to recover after repeated wildfires in Amazonian floodplains? Experimental evidence on tree recruitment limitation, Dryad, Dataset,


Amazonian floodplain forests are particularly vulnerable to wildfires that spread during extreme droughts, causing large scale forest dieback. After a second wildfire, these forests persist trapped with low tree cover and empty seed banks, yet the mechanisms that could explain this arrested succession remain unknown. Here we use a four-year field experiment to test whether tree recruitment failures in burnt floodplain forests are caused by environmental filtering, limiting early seedling emergence and establishment. We sowed seeds and planted seedlings of six floodplain trees with contrasting life strategies, and tested the roles of environmental filters by comparing tree seedling performances under different habitats (i.e. unburned forest, forest edge with burnt site, forest burnt once and forest burnt twice), and by manipulating soil root mats and herbaceous cover. Our results show that seedling emergence was around 15 % across all habitats. In general, seedlings performed best in burnt forests. Seedling growth was highest in forests burnt once, possibly because of high nutrient availability after fire. In forests burnt twice, tree seedlings grew relatively less, as nutrients become limiting due to flood erosion; yet, seedlings survived longer, possibly because of lower competition with sparse, naturally recruiting trees. We found similar patterns for seedlings that emerged in the field from sowed seeds. Synthesis: Our experimental evidence suggests that environmental filtering related to soil nutrient limitations may slow down forest recovery after repeated wildfires. Yet, our findings showing that floodplain trees are able to germinate from seeds and establish successfully in twice burnt forests suggest that seed limitation may be the reason why forest recovery fails persistently. A corollary to the problem is that repeatedly burnt forests seem to be trapped by a self-reinforcing feedback, in which low tree cover reduces seed dispersal and consequently seed availability, keeping tree cover low. Overall, our findings indicate that active restoration initiatives based on seeding native tree species may help accelerating the recovery of degraded floodplain forests after repeated wildfires.