Windthrows control biomass patterns and functional composition of Amazon forests.

Amazon forests account for ~25% of global land biomass and tropical tree species. In these forests, windthrows (i.e., snapped and uprooted trees) are a major natural disturbance, but the rates and mechanisms of recovery are not known. To provide a predictive framework for understanding the effects of windthrows on forest structure and functional composition (DBH ≥10 cm), we quantified biomass recovery as a function of windthrow severity (i.e., fraction of windthrow tree mortality on Landsat pixels, ranging from 0%-70%) and time since disturbance for terra-firme forests in the Central Amazon. Forest monitoring allowed insights into the processes and mechanisms driving the net biomass change (i.e., increment minus loss) and shifts in functional composition. Windthrown areas recovering for between 4-27 years had biomass stocks as low as 65.2-91.7 Mg/ha or 23%-38% of those in nearby undisturbed forests (~255.6 Mg/ha, all sites). Even low windthrow severities (4%-20% tree mortality) caused decadal changes in biomass stocks and structure. While rates of biomass increment in recovering vegetation were nearly double (6.3 ± 1.4 Mg ha-1  year-1 ) those of undisturbed forests (~3.7 Mg ha-1  year-1 ), biomass loss due to post-windthrow mortality was high (up to -7.5 ± 8.7 Mg ha-1  year-1 , 8.5 years since disturbance) and unpredictable. Consequently, recovery to 90% of "pre-disturbance" biomass takes up to 40 years. Resprouting trees contributed little to biomass recovery. Instead, light-demanding, low-density genera (e.g., Cecropia, Inga, Miconia, Pourouma, Tachigali, and Tapirira) were favored, resulting in substantial post-windthrow species turnover. Shifts in functional composition demonstrate that windthrows affect the resilience of live tree biomass by favoring soft-wooded species with shorter life spans that are more vulnerable to future disturbances. As the time required for forests to recover biomass is likely similar to the recurrence interval of windthrows triggering succession, windthrows have the potential to control landscape biomass/carbon dynamics and functional composition in Amazon forests.

Pre-Columbian floristic legacies in modern homegardens of Central Amazonia.

Historical ecologists have demonstrated legacy effects in apparently wild landscapes in Europe, North America, Mesoamerica, Amazonia, Africa and Oceania. People live and farm in archaeological sites today in many parts of the world, but nobody has looked for the legacies of past human occupations in the most dynamic areas in these sites: homegardens. Here we show that the useful flora of modern homegardens is partially a legacy of pre-Columbian occupations in Central Amazonia: the more complex the archaeological context, the more variable the floristic composition of useful native plants in homegardens cultivated there today. Species diversity was 10% higher in homegardens situated in multi-occupational archaeological contexts compared with homegardens situated in single-occupational ones. Species heterogeneity (ß-diversity) among archaeological contexts was similar for the whole set of species, but markedly different when only native Amazonian species were included, suggesting the influence of pre-conquest indigenous occupations on current homegarden species composition. Our findings show that the legacy of pre-Columbian occupations is visible in the most dynamic of all agroecosystems, adding another dimension to the human footprint in the Amazonian landscape.