Get access

Frequent fires reduce tree growth in northern Australian savannas: implications for tree demography and carbon sequestration

Authors

  • BRETT P. MURPHY,

    1. Bushfires NT, Northern Territory Government, Winnellie, NT, Australia,
    2. School of Plant Science, University of Tasmania, Hobart, TAS, Australia,
    Search for more papers by this author
  • JEREMY RUSSELL-SMITH,

    1. Bushfires NT, Northern Territory Government, Winnellie, NT, Australia,
    2. Tropical Savannas Management Cooperative Research Centre, Charles Darwin University, Darwin, NT, Australia,
    Search for more papers by this author
  • LYNDA D. PRIOR

    1. School of Plant Science, University of Tasmania, Hobart, TAS, Australia,
    2. School for Environmental Research, Charles Darwin University, Darwin, NT, Australia
    Search for more papers by this author

Brett P. Murphy, PMB 44, Winnellie, NT 0822, Australia, e-mail: brettpatrickmurphy@hotmail.com

Abstract

Tropical savannas are typically highly productive yet fire-prone ecosystems, and it has been suggested that reducing fire frequency in savannas could substantially increase the size of the global carbon sink. However, the long-term demographic consequences of modifying fire regimes in savannas are difficult to predict, with the effects of fire on many parameters, such as tree growth rates, poorly understood. Over 10 years, we examined the effects of fire frequency on the growth rates (annual increment of diameter at breast height) of 3075 tagged trees, at 137 locations throughout the mesic savannas of Kakadu, Nitmiluk and Litchfield National Parks, in northern Australia. Frequent fires substantially reduced tree growth rates, with the magnitude of the effect markedly increasing with fire severity. The highest observed frequencies of mild, moderate and severe fires (1.0, 0.8 and 0.4 fires yr−1, respectively) reduced tree growth by 24%, 40% and 66% respectively, relative to unburnt areas. These reductions in tree growth imply reductions in the net primary productivity of trees by between 0.19 t C ha−1 yr−1, in the case of mild fires, and 0.51 t C ha−1 yr−1, in the case of severe fires. Such reductions are relatively large, given that net biome productivity (carbon sequestration potential) of these savannas is estimated to be just 1–2 t C ha−1 yr−1. Our results suggest that current models of savanna tree demography, that do not account for a relationship between severe fire frequency and tree growth rate, are likely to underestimate the long-term negative effects of frequent severe fires on tree populations. Additionally, the negative impact of frequent severe fires on carbon sequestration rates may have been underestimated; reducing fire frequencies in savannas may increase carbon sequestration to a greater extent than previously thought.

Ancillary