Skip to Navigation

Dynamics of sediment and nutrient fluxes from burnt forest catchments

Final Report for Land and Water Australia Project DSE1, and Victorian Department of Sustainability and Environment Project

Chris Sherwin, Gaby Szegedy, Gary Sheridan, John Costenaro, Patrick Lane, Phillip McKEnna, Phillip Noske

  • Report
  • Innovation
  • Product ID PN22320
  • PublishedFebruary 2009

Product Information

The magnitude of the impact of wildfire on water quality in SE Australia has rarely been estimated because the location and timing of wildfires is unpredictable, usually precluding the establishment of robust before and after impact experimental designs.  However the 2003 Alpine fires presented a rare opportunity to overcome this constraint when two long-term water quality research catchments were burnt by wildfire. The combination of a sound experimental design and newly developed field instrumentation provided a certainty to the results that had been lacking in the (limited) literature to date.  The resulting data set provides the best-available parameterisation for “burnt high rainfall forest” land use type for the dominant water quality models currently used in Australia.  A parallel investigation of the runoff and erosion processes active in the catchment during the recovery period enabled evaluation of the process based models typically used to predict erosion and water quality impacts of land use activities, including fire.

The key findings were:

  1. Discharge increased by around 70%, after the fire, and persisted for at least three years
  2. There was no apparent change in the runoff processes delivering water to the stream network
  3. Suspended and coarse sediment fluxes increased by 8-9 times in the first year post-fire, but relaxed to pre-fire levels by the end of the second year
  4. Phosphorus and nitrogen fluxes increased by perhaps 5-6 times, and showed the same recovery rate as sediment, with the majority of both P and N transported in the particulate form
  5. Water quality recovery was a function of the ground cover recovery
  6. Hillslope process experiments revealed the importance of soil water repellency and the spatial arrangement of saturated hydraulic conductivity in pollutant pathway length, and these data suggested near-stream areas to be the pollutant source areas
  7. These experiments demonstrated that existing erosion process models do not work in these environments
  8. The results from this study are germane to a wet eucalypt environment under “average” rainfall conditions and good vegetation recovery
  9. A probabilistic approach to modeling is recommended to deal with extreme variation in post-fire rainfall and vegetation recovery over large burnt areas.

Erosion process studies showed that many of the pre-conceptions of how sediment and other constituents are delivered to streams after fire do not hold in this environment.  Many of these incorrect preconceptions are imbedded, either implicitly or explicitly, within infiltration, erosion and runoff models.  In particular, it was found that the combination of highly macroporous forest soils combined with strong water repellence resulted in infiltration and runoff generation characteristics that could not be predicted by any existing models, not only with respect to magnitude, but even the direction of change.  The practical implication of these findings is that, in these systems, only the near stream area delivers sediment to the stream, with the remainder of the catchment being “disconnected” during most storm events.   

When the results from this knowledge-intensive, yet spatially limited, study are interpreted in the wider spatial and temporal context, two key considerations emerge.  Firstly, it is likely that the study catchments fall at the “low water quality impact” end of the spectrum of SE Australian forests.  This is because the high rainfall (>1500mm) results in deep, high infiltration capacity soils, and rapid vegetation recovery.  Secondly, the magnitude of water quality impacts across Victoria following the 2003 fires was characterised by extremely highly variability.  This observation signals that a shift in our conceptual representation of fire related water quality impacts is required, from traditional deterministic to more probabilistic-based approaches.  Ongoing research is focused in this direction, with probabilistic post-fire modelling methods being explored at both the hillslope and catchment scales.

PN22320.pdf4.53 MB

Product Data

Chris Sherwin, Gaby Szegedy, Gary Sheridan, John Costenaro, Patrick Lane, Phillip McKEnna, Phillip Noske
PublishedFebruary 2009
Product Type: 
Product Format: 
Land and Water Australia
Custom Identifier: 

Category Information


id: 2873 / created: 18 February, 2009 / last updated: 18 June, 2009