Understand Daily Fire Suppression Resource Ordering and Assignment Patterns by Unsupervised Learning

Wei, Yu; Thompson, Matthew P.; Belval, Erin J.; Calkin, David E.; Bayham, Jude

doi:10.3390/make3010002

Cited by 5 publications

(2 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The IMSR is a valuable data source for both fire managers and researchers 25 , and we have seen efforts to obtain and use several pieces of IMSR information in research including the PL 26 , 27 and suppression resource use 28 – 30 . However, we are unaware of other efforts to generate a research-ready version of the IMSR archive, which would support broader use of these data.…”

Section: Background and Summarymentioning

confidence: 99%

Dataset of United States Incident Management Situation Reports from 2007 to 2021

Nguyen,

Belval,

Wei

et al. 2024

Sci Data

Self Cite

View full text Add to dashboard Cite

This paper presents a unique 15-year dataset of Incident Management Situation Reports (IMSR), which document daily wildland fire situations across ten geographical regions in the United States. The IMSR dataset includes summaries for each reported day on national and regional wildfire activities, wildfire-specific activities, and committed fire suppression resources (i.e., personnel and equipment). This dataset is distinct from other wildfire data sources as it provides daily information on national fire suppression resource utilization, national and regional preparedness levels, and management priority for each region and fire. We developed an open-source Java program, IMSR-Tool, to process 3,124 IMSR reports available from 2007 to 2021 to generate this structured IMSR dataset, which can be updated when future reports become available. The dataset presented here and its future extension enable researchers and practitioners to study historical wildfire activity and resource use across regions and time, examine fire management perceptions, evaluate strategies for fire prioritization and fire resource allocation, and exploit other broader usage to improve wildfire management and response in the United States.

show abstract

Section: Background and Summarymentioning

confidence: 99%

Dataset of United States Incident Management Situation Reports from 2007 to 2021

Nguyen,

Belval,

Wei

et al. 2024

Sci Data

Self Cite

View full text Add to dashboard Cite

show abstract

“…Resource use on large fires (e.g., Hand et al, 2017) and comparisons between resource use on a geographical scale (e.g., Wei et al, 2020) has also been examined; but there remains a need for further understanding of the decision-making concerning airtanker IA dispatch because resource allocation during this stage can be costly but can contribute to rapid containment. Taylor and Nadeem (2022) recently examined factors which resulted in the dispatch of airtankers on IA on fires in British Columbia, Canada, and found that the average projected FWI over the next three days was a good indicator of such airtanker use.…”

Section: Introductionmentioning

confidence: 99%

Modelling decisions concerning the dispatch of airtankers for initial attack on forest fires in Ontario, Canada

et al. 2023

View full text Add to dashboard Cite

Airtankers are commonly used for initial attack (IA) to reduce the likelihood of wildland fires escaping containment efforts. We examined IA airtanker dispatch decisions for forest fires in Ontario, Canada through an analysis of historical fire records from 2001 to 2019. A hurdle modelling approach that predicts the probability of airtanker(s) being dispatched and then the number of airtankers sent was used. Two different hurdle models were considered depending on the timing of the information available to the decision maker: a ‘fire report model’ based upon the information available when a fire is first reported, and an ‘initial attack model’ using information available at the time IA action began. Both models indicated that the most influential covariates for airtanker dispatch are: the Fire Weather Index, fuel volatility, observed fire rate of spread, fire size at IA, and cause of ignition. When evaluating the predictive ability of the models on a validation dataset, the ‘initial attack model’ performed better than the ‘fire report model’. Our models generally perform well when predicting none, one, or two airtankers on IA, but they generally underpredict when more than two airtankers are dispatched, which suggests risk-averse decision making in fire management resource dispatching.

show abstract

Characterising ignition precursors associated with high levels of deployment of wildland fire personnel

Cullen,

Goldgeier,

Belval

et al. 2024

Int. J. Wildland Fire

View full text Add to dashboard Cite

Background As fire seasons in the Western US intensify and lengthen, fire managers have been grappling with increases in simultaneous, significant incidents that compete for response resources and strain capacity of the current system. Aims To address this challenge, we explore a key research question: what precursors are associated with ignitions that evolve into incidents requiring high levels of response personnel? Methods We develop statistical models linking human, fire weather and fuels related factors with cumulative and peak personnel deployed. Key results Our analysis generates statistically significant models for personnel deployment based on precursors observable at the time and place of ignition. Conclusions We find that significant precursors for fire suppression resource deployment are location, fire weather, canopy cover, Wildland–Urban Interface category, and history of past fire. These results align partially with, but are distinct from, results of earlier research modelling expenditures related to suppression which include precursors such as total burned area which become observable only after an incident. Implications Understanding factors associated with both the natural system and the human system of decision-making that accompany high deployment fires supports holistic risk management given increasing simultaneity of ignitions and competition for resources for both fuel treatment and wildfire response.

show abstract

Understand Daily Fire Suppression Resource Ordering and Assignment Patterns by Unsupervised Learning

Cited by 5 publications

References 31 publications

Dataset of United States Incident Management Situation Reports from 2007 to 2021

Dataset of United States Incident Management Situation Reports from 2007 to 2021

Modelling decisions concerning the dispatch of airtankers for initial attack on forest fires in Ontario, Canada

Characterising ignition precursors associated with high levels of deployment of wildland fire personnel

Contact Info

Product

Resources

About