Modeling the impacts of drought-fire relationship in carbon stocks in the Xingu basin

Background/Question/Methods

The forests of the Amazon Basin are an important component of the global carbon cycle. Their importance are compounded by the vulnerability of carbon stocks in these forests to drought, fire and other disturbances. During recent years understory fires have affected large areas of Amazon forest, especially for the southeastern Amazon forest, during severe droughts. However, modeling fire in tropical forests is still at an early stage. There are only a few models of fire for the Amazon, and most of them attempt to describe the risk of fire rather than quantify fire impact on carbon stocks by forest degradation (increased tree mortality, reduced tree growth). The potential responses of vegetation in the Xingu basin, under different scenarios of drought and fire, were evaluated using the CARLUC ecosystem model (Hirsch, 2004). CARLUC was calibrated to the forests of Xingu by adopting the fire mortality rates measured in field experiments at Tanguro (Soares 2012). The effects of drought-fire interactions on environmental variables derived from carbon stocks was evaluated: above ground biomass (AGB), leaf area index (LAI) and leaf litter (LL).

Results/Conclusions

We performed three experimental simulation runs of a 36-month period under three vapor pressure deficit (vpd) scenarios (<1.5; 2; >2.5) with a fire event prescribed in each of the first two years. After the fire events AGB and LAI decreased (4;27;49%), (-6;16;37%), respectively for three vpd scenarios (<1.5; 2; >2.5). The LL increase (20-37%) depended on the level of dryness. In the >2.5 scenario the rates were substantially higher than field measurements, where in the drought year VPD reached 2.3KPa in the forest interior. This anomalously dry and hot microclimate condition suggests high severity fires consequently high tree mortality rates as measured in forest edge. Another characteristics for occurrence of high-intensity fires are the fine fuel loads as leaf litter. The field measurements indicated an increase of 30% in LL for drought years, close to the simulated range 20-37%, but our simulation showed abrupt increases if drought conditions remained for more months. The vpd2 scenario, in general is in agreement with field measurements (Brando et. al., 2014), with vpd varying 1-2.3 KPa, in forest interior, AGB decreased by 12-30% and LAI by 23-30%. Although, CARLUC does not represent some intrinsic processes in vegetation dynamics that are captured in field-based studies, this new calibration provides an exceptional opportunity to investigate fire impact and postfire recovery across a wide range of fire conditions.