ABSTRACT
Ability to predict the risk of damaging events (e.g. wildfires) is crucial in helping emergency services in their decision making processes, to mitigate and reduce the impact of such events. Today, wildfire rating systems have been in operation extensively in many countries around the world to estimate the danger of wildfires. In this paper we propose a data-driven approach to predict wildfire risk using weather data. We show how we address the inherent challenge arising due to the temporal dynamicity of weather data. Weather observations naturally change in time, with finer-scale variation (e.g. stationary day or night) or large variations (nonstationary day or night), and this determines a temporal variation of the predicted wildfire danger. We show how our dynamic wildfire danger prediction model addresses the aforementioned challenge using context-based anomaly detection techniques. We call our predictive model a Context-Based Fire Risk (CBFR) model. The advantage of our model is that it maintains multiple historical models for different temporal variations (e.g. day versus night), and uses ensemble learning techniques to predict wildfire risk with high accuracy. In addition, it is completely unsupervised and does not rely on expert knowledge, which makes it flexible and easily applied to any region of interest. Our CBFR model is also scalable and can potentially be parallelised to speed up computation. We have considered multiple wildfire locations in the Blue Mountains, Australia as a case study, and compared the results of our system with the existing well-established Australian wildfire rating system. The experimental results show that our predictive model has a substantially higher accuracy in predicting wildfire risk, which makes it an effective model to supplement the operational Australian wildfire rating system.
Supplemental Material
- Australian bureau of meteorology weather stations. http://www.bom.gov.au/climate/cdo/about/sites.shtml.Google Scholar
- Australian bureau of meteorology weather stations. http://www.bom.gov.au/vic/forecasts/fire-map.shtml.Google Scholar
- Australian emergency management knowledge hub. https://www.emknowledge.gov.au/resource/4781/2013/ bushfire-new-south-wales-2013.Google Scholar
- Australian fire danger ratings. http://www.esa.act.gov.au/wp-content/uploads/fire-danger-ratings.pdf.Google Scholar
- Black saturday bushfires. https://en.wikipedia.org/wiki/black saturday bushfires.Google Scholar
- Emergency management victoria strategic action plan. https://www.emv.vic.gov.au/plans/strategic-action-plan/.Google Scholar
- Google maps. https://www.google.com.au/maps.Google Scholar
- The human cost of natural disasters 2015: a global perspective, http://reliefweb.int/report/world/human-cost-natural-disasters-2015-global-perspective.Google Scholar
- Natural resources canada. http://cwfis.cfs.nrcan.gc.ca/background/summary/fwi.Google Scholar
- The united nations office for disaster risk reduction, http://www.unisdr.org/archive/42814.Google Scholar
- Wildland fire assessment system. fire danger rating. http://www.wfas.net/index.php/fire-danger-rating-fire-potential--danger-32.Google Scholar
- P. L. Andrews, D. O. Loftsgaarden, and L. S. Bradshaw. Evaluation of fire danger rating indexes using logistic regression and percentile analysis. International Journal of Wildland Fire, 12(2):213--226, 2003.Google ScholarCross Ref
- E. Cohen and M. Strauss. Maintaining time-decaying stream aggregates. In ACM SIGMOD symposium on Principles of database systems, pages 223--233, 2003. Google ScholarDigital Library
- D. Dee, S. Uppala, A. Simmons, P. Berrisford, P. Poli, S. Kobayashi, U. Andrae, M. Balmaseda, G. Balsamo, P. Bauer, et al. The era-interim reanalysis: Configuration and performance of the data assimilation system. Quarterly Journal of the Royal Meteorological Society, 137(656):553--597, 2011.Google ScholarCross Ref
- C. A. Farris, C. Pezeshki, and L. F. Neuenschwander. A comparison of fire probability maps derived from gis modeling and direct simulation techniques. In Joint Fire Science Conference and Workshop, pages 131--138, 1999.Google Scholar
- M. A. Finney. The challenge of quantitative risk analysis for wildland fire. Forest Ecology and Management, 211(1):97--108, 2005.Google ScholarCross Ref
- A. G. McArthur. Fire behaviour in eucalypt forests. 1967.Google Scholar
- C. Miller and A. A. Ager. A review of recent advances in risk analysis for wildfire management. International journal of wildland fire, 22(1):1--14, 2013.Google ScholarCross Ref
- M. Moshtaghi, T. C. Havens, J. C. Bezdek, L. Park, C. Leckie, S. Rajasegarar, J. M. Keller, and M. Palaniswami. Clustering ellipses for anomaly detection. Pattern Recognition, 44(1):55--69, 2011. Google ScholarDigital Library
- M. Moshtaghi, S. Rajasegarar, C. Leckie, and S. Karunasekera. An efficient hyperellipsoidal clustering algorithm for resource-constrained environments. Pattern Recognition, 44(9):2197--2209, 2011. Google ScholarDigital Library
- I. Noble, A. Gill, and G. Bary. Mcarthur's fire-danger meters expressed as equations. Australian Journal of Ecology, 5(2):201--203, 1980.Google ScholarCross Ref
- B. Saglam, E. Bilgili, B. Dincdurmaz, A. I. Kadiogulari, and Ö. Küçük. Spatio-temporal analysis of forest fire risk and danger using landsat imagery. Sensors, 8(6):3970--3987, 2008.Google ScholarCross Ref
- M. Salehi, C. A. Leckie, M. Moshtaghi, and T. Vaithianathan. A relevance weighted ensemble model for anomaly detection in switching data streams. In Advances in Knowledge Discovery and Data Mining, pages 461--473. 2014.Google ScholarCross Ref
- C. Vasilakos, K. Kalabokidis, J. Hatzopoulos, and I. Matsinos. Identifying wildland fire ignition factors through sensitivity analysis of a neural network. Natural hazards, 50(1):125--143, 2009.Google ScholarCross Ref
- B. Vidakovic. Statistics for bioengineering sciences: with MATLAB and WinBUGS support. Springer Science & Business Media, 2011.Google ScholarCross Ref
- L. Yu, N. Wang, and X. Meng. Real-time forest fire detection with wireless sensor networks. In International Conference on Wireless Communications, Networking and Mobile Computing, volume 2, pages 1214--1217, 2005.Google Scholar
Index Terms
- Dynamic and Robust Wildfire Risk Prediction System: An Unsupervised Approach
Recommendations
A Dynamic Pipeline for Spatio-Temporal Fire Risk Prediction
KDD '18: Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data MiningRecent high-profile fire incidents in cities around the world have highlighted gaps in fire risk reduction efforts, as cities grapple with fewer resources and more properties to safeguard. To address this resource gap, prior work has developed machine ...
Development and application of a system for dynamic wildfire risk assessment in Italy
In this paper, the architecture and the application of a system designed for the assessment of the distribution of dynamic wildland fire risk over the whole Italian territory are presented. Such an assessment takes place on the basis of static ...
Human-Sensors & Physics Aware Machine Learning for Wildfire Detection and Nowcasting
Computational Science – ICCS 2023AbstractThis paper proposes a wildfire prediction model, using machine learning, social media and geophysical data sources to predict wildfire instances and characteristics with high accuracy. We use social media as a predictor of wildfire ignition, and a ...
Comments