Comparing calibrated statistical and machine learning methods for wildland fire occurrence prediction: a case study of human-caused fires in Lac La Biche, Alberta, Canada

Abstract

Wildland fire occurrence prediction (FOP) modelling supports fire management decisions, such as suppression resource pre-positioning and the routeing of detection patrols. Common empirical modelling methods for FOP include both model-based (statistical modelling) and algorithmic-based (machine learning) approaches. However, it was recently shown that many machine learning models in FOP literature are not suitable for fire management operations because of overprediction if not properly calibrated to output true probabilities. We present methods for properly calibrating statistical and machine learning models for fine-scale, spatially explicit daily FOP followed by a case-study comparison of human-caused FOP modelling in the Lac La Biche region of Alberta, Canada, using data from 1996 to 2016. Calibrated bagged classification trees, random forests, neural networks, logistic regression models and logistic generalised additive models (GAMs) are compared in order to assess the pros and cons of these approaches when properly calibrated. Results suggest that logistic GAMs can have similar performance to machine learning models for FOP. Hence, we advocate that the pros and cons of different modelling approaches should be discussed with fire management practitioners when determining which models to use operationally because statistical methods are commonly viewed as more interpretable than machine learning methods.

Improving Wildfire Probability Modeling by Integrating Dynamic-Step Weather Variables over Northwestern Sichuan, China

“…Machine learning techniques have been widely used to predict wildfire probability, due to their convenience in integrating multisource data and higher accuracy than traditional statistical methods and fire weather or drought indices approach (Goldarag et al 2016;Cao et al 2017;Jaafari, Mafi-Gholami et al 2019;Zhang et al 2019;Phelps and Woolford 2021). One of the critical problems in wildfire probability modeling is the elimination of multicollinearity in the input variables.…”

An Integrated Grassland Fire-Danger-Assessment System for a Mountainous National Park Using Geospatial Modelling Techniques

“…Logistic regression (LR) is a modelling technique that is part of the family of generalized linear models [132] and is one of the popular statistical models utilized effectively to predict fire occurrence and examine the driving factors for fire ignition and propagation. In a regression model, an equation for predicting the value of the dependant variable based on one or more independent predictor variables is developed [133].…”

Characterizing Topographic Influences of Bushfire Severity Using Machine Learning Models: A Case Study in a Hilly Terrain of Victoria, Australia

“…A few of them are related to bushfire severity quantification. Most of the ML-based severity research utilises RF model in bushfire severity pattern mapping [23], [42], [45], influencing driver analysis and severity predictive assessment [3], [4], [8], [33], [46]- [49]. In addition, different other models including GB and SVM and their ensembles have demonstrated their capability in explaining complex variable relationships in bushfires applications [50]- [53].…”