loading page

Response and recovery of tropical forests after cyclone disturbance
  • +2
  • Barbara Bomfim,
  • William McDowell,
  • Jess Zimmerman,
  • Anthony Walker,
  • Lara Kueppers
Barbara Bomfim
Lawrence Berkeley National Laboratory

Corresponding Author:bbomfim@lbl.gov

Author Profile
William McDowell
University of New Hampshire
Author Profile
Jess Zimmerman
University of Puerto Rico
Author Profile
Anthony Walker
Oak Ridge National Laboratory
Author Profile
Lara Kueppers
Lawrence Berkeley National Laboratory
Author Profile


Tropical cyclones dominate the disturbance regime experienced by forest ecosystems in many parts of the world. Interactions between cyclone disturbance regimes and nutrient availability strongly influence forest ecosystem dynamics. However, uncertainty exists over the importance of soil fertility properties (i.e., total soil phosphorus-P concentration) in mediating forest resistance and recovery from cyclone disturbance. We hypothesized that forests on soils with low total P (e.g., developed on limited-P parent material) have a higher resistance to but a slower recovery from cyclone disturbance than forests on high P soils. We investigated cyclone impacts on litterfall, an essential conduit for nutrient recycling in forest ecosystems. We compiled site-level forest litterfall data from 53 studies and datasets associated with 15 naturally-occurring one simulated tropical cyclone in 23 sites within five regions (Taiwan, Australia, Mexico, Hawaii, and the Caribbean)and four cyclone basins. We calculated the effect sizes of cyclone disturbance on the litterfall mass and nutrient (P and nitrogen-N) concentrations and fluxes during the first (< five) years post-disturbance across a total soil P gradient. We also assessed the effect of 20 covariates on the degree of cyclone impact on litterfall. Total litterfall mass flux increased by 4820%following cyclone disturbance. Such an initial increase in litterfall mass reflects the magnitude of cyclone-derived plant material input to the forest floor, which was highest in the Caribbean and lowest in Taiwan. Among 20 covariates, soil P and region were the best predictors of cyclone effects on total litterfall mass, explaining 80% of the variance. The effect sizes increased linearly with soil P and region, from significantly lower in Taiwan (low-P) to largest in the Caribbean (high-P). Total litterfall P and N fluxes increased significantly post-cyclone, whereas the increase in leaf P flux was twice as that in Nflux. Results highlight the importance of understanding the interactions between disturbance and nutrient gradients in forest ecosystems to understand forest responses to altered cyclone regimes expected under climate change.