Current browse context:
stat.ME
Change to browse by:
References & Citations
Statistics > Methodology
Title: A marginal modelling approach for predicting wildfire extremes across the contiguous United States
(Submitted on 31 Dec 2021 (v1), last revised 14 Mar 2023 (this version, v3))
Abstract: This paper details a methodology proposed for the EVA 2021 conference data challenge. The aim of this challenge was to predict the number and size of wildfires over the contiguous US between 1993 and 2015, with more importance placed on extreme events. In the data set provided, over 14\% of both wildfire count and burnt area observations are missing; the objective of the data challenge was to estimate a range of marginal probabilities from the distribution functions of these missing observations. To enable this prediction, we make the assumption that the marginal distribution of a missing observation can be informed using non-missing data from neighbouring locations. In our method, we select spatial neighbourhoods for each missing observation and fit marginal models to non-missing observations in these regions. For the wildfire counts, we assume the compiled data sets follow a zero-inflated negative binomial distribution, while for burnt area values, we model the bulk and tail of each compiled data set using non-parametric and parametric techniques, respectively. Cross validation is used to select tuning parameters, and the resulting predictions are shown to significantly outperform the benchmark method proposed in the challenge outline. We conclude with a discussion of our modelling framework, and evaluate ways in which it could be extended.
Submission history
From: Callum Murphy-Barltrop [view email][v1] Fri, 31 Dec 2021 10:23:42 GMT (4347kb,D)
[v2] Sun, 2 Oct 2022 12:31:00 GMT (703kb,D)
[v3] Tue, 14 Mar 2023 10:37:13 GMT (639kb,D)
Link back to: arXiv, form interface, contact.