95th ESA Annual Meeting (August 1 -- 6, 2010)

COS 2-6 - Decadal stability of ecosystem structure and process over a tropical rain forest landscape

Monday, August 2, 2010: 3:20 PM
329, David L Lawrence Convention Center
David B. Clark, Depts. of Biology & Environmental Sciences, University of Missouri-St. Louis & University of Virginia, La Selva, Costa Rica and Deborah A. Clark, Departments of Biology & Environmental Sciences, University of Missouri-St. Louis & University of Virginia, St. Louis, MO
Background/Question/Methods: The responses of tropical rain forest (TRF) landscapes to changing global climate will have major impacts on global carbon cycling and biodiversity conservation.  Meta-analyses of TRF forest inventory plots in old growth have suggested the following changes are already underway: increased rates of mortality and recruitment, leading to increased turnover, driven by higher tree growth rates due to increasing atmospheric CO2 concentrations; increasing above-ground biomass; and directional floristic shifts including increasing abundance of lianas.  We examined each of these predictions using data from 13 consecutive annual censuses (1997-2009) of 18 0.5 ha plots sited with a stratified random design across an old-growth TRF landscape at the La Selva Biological Station in Costa Rica.
Results/Conclusions:   The major factor affecting annual rates of mortality, recruitment and turnover over the 12 year period was the 1997-1998 Mega-El Niño, which caused high mortality followed by high recruitment.  There was no suggestion of the hypothesized increase in these rates over the interval.  Diameter growth was depressed by the Mega-El Niño but quickly recovered in the following year; even ignoring the El Niño-depression, there was no suggestion of the hypothesized increase in growth rates over time.  Landscape-scale stem density was stable over the interval with ca. 0.5 SEM variation among the 13 years, with equal stability shown by the proportions of trees, palms and lianas in the sample.  Liana density declined non-significantly, in contrast to the hypothesized increase.  Stand biomass increased non-significantly, by an average total of 3.4 ± 3.0 SEM Mg/0.5 ha plot over the 12 yr, in contrast to the predicted increase of 7.4 Mg.  Given the large within-landscape variation in annual biomass estimates, we cannot definitively reject the biomass increase hypothesis.  Overall, our results suggest that over the last two decades the old-growth TRF landscape at this site has been stable in most aspects of stand structure and dynamics.  We discuss a variety of reasons this TRF landscape does not support most of the hypothesized global trends.  While meta-analyses are useful tools for summarizing disparate studies, there is an urgent need for site-based annual-interval studies with statistical designs that permits calculation of confidence intervals around estimates of ecological parameters.  Such studies can provide robust assessments of the landscape-scale structure and dynamics of tropical rain forests, as well as a powerful analysis framework for detecting ecosystem responses to on-going climate change.