Natural ignition usually refers to lightning activity, which is the dominant natural cause of fires and can be the main ignition mechanism in remote areas with low population density and poor accessibility. The primary type of lightning that can ignite wildfires is lightning strikes with a long continuing current (LCC), which refers to a discharge that lasts for more than 40 ms in a return stroke. Compared to short-duration lightning, LCC lightning can heat wildland fuels for a more extended period and transfer more energy, making it more effective at igniting fuels. Therefore, lightning-caused wildfires can potentially be assessed based on the duration of lightning’s discharge or the amount of energy it produces [1]. Another dimension is added to this by considering the polarity of the lightning strike: Latham and Williams (2001)[2] theorised that positive Cloud-to-Ground (CG) strikes are more likely to ignite wildfires. Interestingly, regions with the highest number of lightning flashes do not necessarily produce the most flashes containing LCC [3]. In other words, less frequently occurring lightning is more likely to contain LCC. It is essential to take into account the change in the global occurrence rate and spatial pattern of total lightning due to climate change [4].
The first step of the natural ignition model was the prediction of the lightning flashes themselves. Despite the availability of global lightning observing systems, their quality and data availability may be problematic in many parts of the world. However, modern weather prediction models provide a sufficiently detailed description of convective clouds and thunderstorms, allowing for quite accurate lightning flash density prediction for up to a few days forward [5]. The next step includes converting lightning flash density into fire ignitions. It involves numerous governing physical processes such as wind speed, precipitation occurrences and strength, lightning characteristics (e.g., presence of a long continuous stream of electrical current), and fuel readiness and availability to ignite. Therefore, statistical models seem more suitable for their description [6]. The FirEUrisk project utilises the Fire Forecasting Model developed by the Finnish Meteorological Institute, which employs a multi-step machine-learning procedure to construct a statistical model predicting Fire Radiative Power (FRP). The model uses various meteorological parameters and Fire Danger Indices as predictors for training and predictions computing their corresponding contributions to total FRP. Then, Cloud-to-ground lightning flash density contribution to FRP serves as a proxy for natural ignition probability that is integrated with ignitions caused by humans.
Return to Conceptual Framework Diagram
Zhang, H., Guo, P., Chen, H., Liu, N., Qiao, Y., Xu, M., & Zhang, L. (2023). Lightning-induced smoldering ignition of peat: Simulation experiments by an electric arc with long continuing current. Proceedings of the Combustion Institute. ↩︎
Latham, D., & Williams, E. (2001). Lightning and forest fires. In E.A. Johnson, & K. Miyanishi (Eds.), Forest fires. Behavior and ecological aspects (pp. 375-418). San Diego, California: Academic Press. ↩︎
Bitzer, P.M. (2017). Global distribution and properties of continuing current in lightning. Journal of Geophysical Research: Atmospheres, 122, 1033-1041. ↩︎
Pérez-Invernón, F.J., Gordillo-Vázquez, F.J., Huntrieser, H., & Jöckel, P. (2023). Variation of lightning-ignited wildfire patterns under climate change. Nature Communications, 14, 739. ↩︎
Lopez, P. (2016). A lightning parameterization for the ECMWF integrated forecasting system. Monthly Weather Review, 144, 3057-3075. ↩︎
Coughlan, R., Di Giuseppe, F., Vitolo, C., Barnard, C., Lopez, P., & Drusch, M. (2021). Using machine learning to predict fire-ignition occurrences from lightning forecasts. Meteorological Applications, 28, e1973. ↩︎