Open Collections

Smoothing with application to stochastic fire growth modelling

University of British Columbia

Modelling wildland fire spread stochastically is an important way to incorporate the uncertainty associated with this phenomenon. Fitting such a model to data from remote-sensed images could be used to provide accurate fire spread risk maps. We study a particular model from this perspective. One objective of this thesis is to verify the model on data collected under experimentally controlled conditions. We present the analysis of data from small-scale experimental fires that were digitally recorded. Data extraction and processing methods and issues are discussed, along with an estimation methodology. A critical part of the estimation methodology revolves around the smoothing of observed counts of burning and burnt out pixels as functions of elapsed time. We employ nonparametric regression for this purpose and consider two bias reduction strategies as possible ways to obtain more accurate estimates of the parameters underlying the stochastic fire spread model. An argument for partial validation of the model is also provided.

Text

2017-08-04

Attribution-NonCommercial-NoDerivatives 4.0 International

http://hdl.handle.net/2429/62501

Mathematics

University of British Columbia

UBCO

http://creativecommons.org/licenses/by-nc-nd/4.0/