B21K-04:
The Influence of Atmospheric CO2 Concentration and Climate Variability on Amazon Tropical Forest
Tuesday, 16 December 2014: 8:45 AM
Andrea D de Almeida Castanho1,2, David Galbraith3, Ke Zhang4,5, Michael Thomas Coe2, Marcos Heil Costa6 and Paul R Moorcroft5, (1)UFC Federal University of Ceará, Fortaleza, Brazil, (2)The Woods Hole Research Center, Falmouth, MA, United States, (3)University of Leeds, School of Geography, Leeds, LS2, United Kingdom, (4)University of Oklahoma Norman Campus, Norman, OK, United States, (5)Harvard University, Cambridge, MA, United States, (6)UFV Federal University of Vicosa, Vicosa, Brazil
Abstract:
Tropical forests are important regulators of atmospheric CO2 concentration and any change in tropical forest C balance will directly affect global climate. Long term studies from undisturbed old-growth forest monitoring sites distributed across Amazonia have presented an overall increase in aboveground biomass in the last decades, and the increase in atmospheric CO2 concentrations is considered the main driver for this observed carbon sink. The main goal of this work was to use simulations from dynamic global vegetation models (DGVM) to explore how much of the observed historical (1970-2008) increase in biomass in undisturbed tropical forest in Amazonia could be attributed to the CO2 fertilization effect or associated to climate change. We compared simulated biomass and productivity from three DGVMs (IBIS, ED2 and JULES) with observations from forest plots (RAINFOR). The analyses helped clarify the variability of historical and potential future simulations.The analyses showed that models shared similar results and deficiencies. The three models represented the two major model types: conventional dynamic global vegetation models that simulate community dynamics and competition between plant functional types (PFTs) using an aggregated ‘big-leaf’ representation (IBIS and Jules), and a size-and-age structured terrestrial ecosystem model that captures individual scale dynamics and competition (ED2). In general, the ED2 model results were more sensitive to climate, but all models greatly underestimate the impact of extreme climatic events (e.g. drought) compared to field data.
All the DGVM’s studied tend to simulate the average biomass well and to overestimate productivity of vegetation under current conditions. All the models presented very low spatial variability compared to field observation. The lack of spatial variability of biomass and productivity is attributed to the lack of nutrient and residence time spatial heterogeneity. All of the DGVMs results suggest that forests have gained AGB in the last decades, consistent with the observations. The CO2 fertilization effect is the strongest factor contributing to the increase in biomass followed by climate. However, models failed to simulate observed biomass dynamics at individual sites when compared to individual field sites.
