Supervisors: Dr David Fuchs (firstname.lastname@example.org), Prof Steven Sherwood (email@example.com)
Errors in identifying the effective fire scar, that is, the portion of a long-lasting bushfire that is actively burning or smouldering, are by far the largest errors in forecasting emissions from these events. For example, in the last bushfire season, the inclusion of a full fire scar into air-quality forecast operations in NSW/DPIE Climate and Atmospheric Science (CAS) branch caused PM2.5 forecasts that reached 2000 g/m^3, while observations showed around 200 g/m^3.
This type of errors can be solved in a manner that is both advanced and a low hanging fruit. In fact, part of the approach that is proposed here is already implemented in CAS Airflow and our internally developed smoke model.
The proposed approach will develop and analyse fire progression tracking based on the NSW Rural Fire Services spatial layers that are being pulled into the Climate and Atmospheric Science (CAS) branch at 30-minute intervals. The student will develop tools to read the spatial data and derive from it time-based scar differences. This will be tested on data from the 2019-2020 bushfire season.
Possible additions to that are:
- Testing approaches to estimate smouldering vs flaming fire scars.
- Estimating fire progression and growth.
- Estimating the gross emissions forecasted using this approach and the projected improvement in the forecast of emissions from bushfires using this approach.
Prerequisites: The student should have moderate python/GIS orientation and would receive further guidance throughout the project (missing skills should be factored into the timeline).