• Biome;
  • dry forest;
  • fire;
  • savanna;
  • vegetation model


Aim  This study aims to improve the formulation and results of the Brazilian Center for Weather Forecasting and Climate Studies Potential Vegetation Model (CPTEC-PVM) by developing a new parameterization for the long-term occurrence of fire in regions of potential savannas in the tropics. Compared with the relatively slow processes of carbon uptake and growth in vegetation, fast mortality and biomass consumption by fires may favour grasses and reduce tree coverage.

Location  The tropics.

Methods  For finding large-scale relationships between fires and other environmental factors, we made two main simplifying assumptions. First, lightning is the most important source of ignition for natural fires. Second, over continental areas in the tropics, lightning is mainly related to the zonal flux of moisture transport.

Results  The parameterization of fire occurrence was built based on a simple empirical relationship, combining information on mean and intra-annual variance of the zonal wind.

Main conclusions  The implementation of this new relationship improved the formulation and the results of the CPTEC-PVM. As a result of this new parameter, the accuracy of the model in allocating the correct vegetation (seasonal forests) instead of savannas for large regions in India and Southeast Asia is now substantially higher than in previous studies.