Dynamics of sediment and nutrient fluxes from burnt forest catchments

During the South Eastern Australia 2003 Alpine wildfires, two long-term water quality research catchments were burnt. This gave researchers an opportunity to measure the impact of wildfires on water quality.

Previously this impact could only be estimated, due to the unpredictable nature of wildfires and therefore the lack of robust before and after experimental design. A two-year experiment with newly developed field instrumentation provided the best available parameters for ‘burnt high rainfall forest’ land use type for the current Australian water quality models. Concurrently, an investigation of the runoff and erosion processes in the catchment during the recovery period made it possible to evaluate the process based models typically used to predict how land use activities, such as fire, impact erosion and water quality. Results showed that discharge increased by around 70% and total suspended sediment, phosphorus and nitrogen exports increased by 5-10 fold. The study found existing models could not be used to predict infiltration and runoff generation characteristics. Using probabilistic-based modelling approaches to deal with the extreme variation in post-fire water quality impacts over large burnt areas is recommended.

Over three years of data on discharge, sediment and nutrient exports and hillslope processes have been collected at high resolution from two research catchments in the East Kiewa Vallley; Slippery Rock Creek (136 ha) and Springs Creek (244 ha). Both catchments had repairs or construction of gauging structures for measurement of discharge, and have been instrumented with pluviometers, turbidity probes, water autosamplers and bedload traps. Relationships between turbidity and total suspended solids (TSS) and between TSS and total phosphorous (TP) and total nitrogen (TN) have been developed, yielding estimation of these parameters at a 15 minute timestep. The processes of runoff, sediment and nutrient generation have been investigated through rainfall simulation and shear-stress experiments. Vegetation recovery has been monitored, and soil water repellency and nutrient status measured through time.

Project Objectives

• 1.Quantify the fluxes of sediment and nutrients from small forested catchments following bushfire, including the immediate post-fire "flush" and the longevity of impacts: * This entails the measurement of catchment discharge and sediment and nutrient concentrations.
• 2. Identify the physical processes and properties of the system driving changes in pollutant exports. These processes fall into three categories: hydrologic processes, sediment detachment and transport, and nutrient mobilisation and transport.
  • Hydrologic processes - quantifying the temporal changes in infiltration, hydrophobicity, canopy interception and evapotranspiration that will drive changes in sediments and nutrient exports.
  • Sediment detachment and transport - identify and quantify the hillslope and in-channel processes that generate sediments and nutrients, including the impact of changes in infiltration on runoff generation, rill and inter-rill erodibility and consequent mobilisation and transport of sediment and nutrients through time.
  • Nutrient mobilisation and transport - quantify nutrient mobilisation and immobilisation processes that contribute to increases or reductions in nutrient loads, identify the transport mechanism of nutrients by proportioning nutrient attachment to suspended sediment and to bedload, and quantify the production and export of both sediment transport phases at the hillslope and catchment scales. The transport mechanism is important in establishing in-stream residence times of pollutants and potential storages of nutrients.
• 3. Improve the representation of processes within, and provide parameters for hillslope erosion and catchment water quality models. Numerous water quality models exist for the prediction of land-use or disturbance on catchment exports of pollutants. These models range in spatial scale from hillslope to large catchment. Appropriate parameterisation is hindering the testing and application of these models. In particular, it is unknown whether existing models can adequately represent fire-induced pollutant generation and transport. The pollutant load and process data captured in this project, combined with water quality observations from larger burnt and unburnt catchments and fire severity and vegetation recovery mapping will allow model parameterisation at a hierarchy of spatial scales.
    

Outcomes

• Streamflow increased by around 70% following the fires, with a greater increase in the second year after the fire. Year 3 saw a lowering of increases at Springs Creek and no change at Slippery Rock Creek. The former is consistent with recovery of understorey and ground cover leaf area, and early E. delegatensis regrowth.
• Flow duration curve analysis indicates that flows over the whole flow regime have increased. This suggests runoff generating mechanisms have not changed markedly at the daily scale. The shape of the curves has been consistent over the 3 years.
• In the first post-fire water year, exports of TSS, TN and TP in suspended load increased by 6-10 fold, diminishing to a 1.5-3 fold increase in year 2. By year 3 the TSS and nutrient loads and had decreased to pre-fire levels at both catchments.
• Undisturbed values of TSS, TN and TP appear to be greater than those commonly reported in the literature
• Loads are dominated by the particulate component in first post-fire year, but dissolved load becomes more important as erosion declines. This has important implications for predicting nutrient loads.
• Bedload exports increased by around an order of magnitude in the first post-fire year, were still 2-3 times pre-fire estimates in year 2, and appear to be still elevated although further declining to pre-fire levels in year 3.
• Sediment concentrations generated on the hillslopes under simulated rainfall increased by around 20 fold immediately following the fire, and has subsequently declined to near unburnt values. This decline in sediment generation rates and the recovery in water quality occurs in concert with the recovery of vegetation cover.
• Natural seasonality of water repellency has been measured on unburnt sites, which showed only small differences in hillslope saturated hydraulic conductivity (Ksat) between burnt and unburnt slopes in summer. The unburnt sites then displayed a marked increase in Ksat in winter, while the water repellency persisted (although muted) on the burnt areas until summer 2005.
• Rill erodibility experiments showed that the soil is strongly resistant to scour by overland flow, and that the sediment delivery rate is not a simple function flow shear stress.
• Volume-to-breakthrough overland flow studies showed that the development of concentrated overland flow at the hillslope scale is unlikely in these catchments
• The results of the hillslope process experiments indicates the bulk of sediment delivered to the stream is from near-stream areas, rather than widespread hillslope sources
• Existing erosion models cannot adequately model the system. A new model is under development that will account for the spatial heterogeneity of runoff and sediment in these forests.
    

Publications and Resources

None listed

Citation

Land & Water Australia. 2009. Dynamics of sediment and nutrient fluxes from burnt forest catchments. [Online] (Updated June 24th, 2009)
Available at: http://lwa.gov.au/node/2874 [Accessed Tuesday 26th of April 2011 04:46:51 AM ].