2017 Conference on Fire Prediction Across Scales





The 2017 Conference on Fire Prediction Across Scales took place from the morning of October 23rd through the late afternoon of October 25th 2017 on Columbia University’s Morningside Campus.

Realistic models of fire activity and behavior are necessary for operational fire management, and to understand past and future changes in fire activity. Developing such models, however requires taking into account vegetation cover, land use practices, fire management capacity, extreme weather, and climate variability. To help foster knowledge exchange among the diverse fields of expertise involved in fire prediction, the Columbia University Initiative on Extreme Weather and Climate is hosting a 2.75-day conference, organized around the themes of “Fire Prediction” and “Fire Management and Impacts”.


Outcomes of the Conference

Understanding the roles of fuels, climate and people in predicting fire: taking the long view

IGACnews February/March 2018

NSF Project Reporting


Confirmed Speakers

Dominique Bachelet, Oregon State University

David Bowman, University of Tasmania

Matthew Butler, United States Forest Service

Ruth DeFries, Columbia University

Ed Delgado, Bureau of Land Management

Francesca Di Giuseppe, ECMWF

Mark Finney, United States Forest Service

Jennifer Marlon, Yale University

James Randerson, University of California Irvine

Andy Robertson, Columbia University

Maria Uriarte, Columbia University

Leroy Westerling, University of California Merced

Please view their bios here.


Conference Program

Please view the Fire Prediction Conference Updated Program

Please view the abstracts here.


Conference Themes

Fire Prediction

Synoptic scales: How well can fire occurrence in the next week be predicted? Once a fire has started, how well can its behavior be predicted over the course of hours and days? Are such predictions making use of the state of the art in fuel and terrain mapping, weather forecasting (including lightning), characterizations of human activity, and for model initialization, modern satellite-based fire detections?

Sub-seasonal and seasonal scales: Are there regions where fire potential can be reliably assessed at sub-seasonal (up to 60 days) and seasonal (more than 2 months) time-scales? What are the hard predictability limits for fire-prone tropical and extratropical regions? What are the roles of statistical and dynamical climate forecast ensembles?

Interannual, decadal and centennial time scales: What are the main interannual, decadal and centennial controls on regional fire activity, including natural vegetation changes, land use and fire management capacity, and large-scale climatic variations? How do these differ between statistical and process-based estimates? What is the potential value of including higher-resolution fire behavior models in Earth System models as both become more sophisticated?

Model benchmarking: How well do models used for projections of future fire activity perform for the observed period? What are appropriate observational benchmarks for models at different scales, including paleoclimate reconstructions, operational records from fire management agencies, and satellite estimates of active fires, burned area, land cover change, and emissions of trace gases and aerosols? What potential do modern data assimilation techniques hold for constraining estimates of fire activity and behavior?

Fire Management & Impacts

Fire management: What advances can be made in fire prescription, prevention, suppression, and mitigation using the state of the art in fire prediction, at the wildland/urban interface in particular? How does fire management policy make use of future estimates of fire activity and associated impacts? From what other sectors can we learn about how to use weather and climate predictions?

Smoke: Is there a role for advanced fire prediction in chemical weather forecasting and climate-chemistry interactions, particularly as they relate to air quality impacts and climatic effects of smoke? What is the role of assimilated trace gas and aerosol satellite retrievals for making such predictions?

Ecological and carbon cycle impacts: Do advanced fire prediction tools have a role in ecological and impacts research, particularly in the context of Earth System models used to project future states of the terrestrial carbon cycle?

Conference Goals

The goals of the conference are to synthesize cutting-edge research across the above topics and to identify areas where important advances in fire prediction at any temporal or spatial scale can be made over the next decade. The conference is intended for all in academia, government, and the private sector with an interest in latest science behind fire prediction. Through a small set of invited talks, contributed posters, and discussion sessions the conference will showcase the latest research on fire prediction and provide opportunities for networking and unstructured discussion.


Accommodations in NYC

We encourage you to book your accommodations ASAP.

For your convenience, we have placed hotel blocks for these dates at the following hotels:

Aloft Harlem

2296 Frederick Douglass Blvd.

New York, NY 10027


Rates: $239 per night

The Hotel Newton

2528 Broadway

New York, NY 10025


Rates: $230 for one queen bed, $250 for two queen beds

Additionally, here is a link to the Columbia hotels with preferred rates.

We also encourage you to check options on Airbnb.

Furthermore there are a few affordable options located on the Columbia Campus with information provided on their websites linked below:

Teachers College Guest Housing

International House Guest Housing

Hostel International New York


Organizing committee:  Robert Field, Kátia Fernandes, Park Williams, Ruth DeFries, Suzana Camargo, Adam Sobel, Jaclyn Rabinowitz


Contact:  For questions or further information, please contact Jaclyn Rabinowitz, jr3357@columbia.edu.