Improving predictions of fire ignitions and process-based simulations of fire spread to assess future wildfire regimes
Wildfires are an important disturbance agent in many ecosystems worldwide, with a wide range of ecological and evolutionary effects. Natural (weather/climate, forest structure) and anthropogenic (population, road network) factors have a large influence on fire regimes. Modeling is a helpful tool for studying the drivers and dynamics of fire regimes so as to increase our scientific process understanding and to make reliable future projections.
This project aims to increase the scientific understanding of selected aspects of forest fire regimes, emphasizing on predictions of ignition probability and our capability to simulate fire dynamics in forest landscape models. Specifically, this project has three main goals:
1. To achieve a better understanding of spatial ignition patterns based on cross-regional modelling in Europe.
2. To improve the modelling of wildfire ignition probability and spread in the forest landscape model LandClim.
3. To assess future landscape dynamics under climate change for case study areas in central Europe and in the Mediterranean.
Part 1: Cross-regional modelling of fire ignitions in Europe
Statistical models are common data-driven methods that are useful for investigating the driving forces of fire regimes or ignition probabilities. However, the application of such models should be restricted to the conditions for which they have been calibrated, but many of them are extrapolated beyond the calibration condition. Hence, our understanding of the spatio-temporal transferability and reliability of such models needs to be improved. Moreover, there are many statistical models focusing on specific regions, but a unified cross-regional model has not been developed. The first part of the project aims to address these problems to increase our understanding of spatial ignition patterns.
Part 2: Improving the capability of wildfire modelling in forest landscape models
Forest landscape models are useful to better understand landscape dynamics and derive scenarios of future landscape development. However, fire processes are rather poorly represented in these models. Therefore, the second part of the project will focus on improving the fire module in LandClim. This part of the project consists of two work package. The first one focuses on integrating the findings of the first part into LandClim. The second one focuses on enhancing the modeling of fire spread by adding a connection between fire spread probability and fuel amount in a cell.
Part 3: Assessing future landscape dynamics under climate change
Fire plays a major role in the Mediterranean area today, whereas most parts of Western Europe are rather wet and thus not very fire-prone. However, with global warming and regional drying particularly in summer, fire may become important even in Western Europe. To be able to model such a transition of the fire regime, modeling of non-fire-prone landscapes is required. Moreover, there is an increasing trend in disturbance damage due to intensifying disturbance regimes in Europe, and continuation of increasing disturbance frequency and severity is expected for the future. The impacts of such changes can possibly be mitigated by forest management.
The last part of the project aims to assess short, mid- and long-term landscape dynamics under climate change scenarios for fire prone and non-fire-prone case study areas in Western Europe and in the Mediterranean, taking into account a range of adaptive forest management strategies.
Funding
This project is funded by Swiss Government Excellence Scholarship and ETH.
Status of the project
The project runs from September 2016 to September 2019.
Contact
Please contact Ismail Bekar, Gunnar Petter or Harald Bugmann for more information.