Animation of 2015 Indonesian fires by Francesca Di Giuseppe (ECMWF) made with the estimated aerosol optical depth from fires detected from MODIS sensors. The simulation is performed using a concatenation of 24-hour forecasts with the ECMWF model in the configuration used for the Copernicus Atmosphere Monitoring Service.
The NOAA CPO Modeling, Analysis, Predictions, and Projections (MAPP) program hosted a webinar on the topic Fire: Modeling and Prediction Issues (Part 1) on Thursday, April 19, 2018, 10 – 11:30 a.m. ET
Date/Time | Title & Presenters |
April 19, 2018 10:00 AM – 11:30 AM ET
|
Fire: Modeling and Prediction Issues (Part 1) |
Speakers and Topics |
Francesca Di Giuseppe (ECMWF) Fire and weather: How well can we predict fire from weather? how much is weather modified by fires? The European Centre for Medium-Range Weather Forecasts is a leading institution in numerical weather prediction. In the last years, thanks to its crucial role in the management of some of the European Copernicus programs, ECMWF has been particularly active in demonstrating the capability of its weather forecasts to support sectoral applications. This effort has invested all time scales from the medium range (up to 10 days forecast) to the seasonal scale (up to 7 months ), including the the subseasonal to seasonal (S2S) range as well. As of today the ECMWF provides several datasets from three different fire danger rating systems; an historical reanalysis dataset, a daily medium range forecast and an extended range forecast. Following the Copernicus general data policy, all data are freely available to any user both public and commercial. The predictability of fire danger from ECMWF forecasts will be revised for few large fires which occurred in the last years Given the impact that fire emissions from large fires have in modifying the surface radiative budget there is also an interest in including these phenomena into weather forecast. The longer range forecasts is the most likely time scale being affected by fire emissions and the subsequent smoke aerosols transport. Sub-seasonal to seasonal simulations performed prescribing observed fire emissions have already highlighted how the inclusion of this missing component can improve forecast scores up to 4 weeks.In its current setup, ECMWF model does not forecast emissions from fires while allowing these to be prescribed. However the challenge remains to design and implement a fully dynamical fire model which could allow to ignite and extinguish fires as required by long range simulations. In this short presentation I will also present some results from the ultimate challenge of including interactive fires into ECMWF numerical weather prediction system —– Keren Mezuman (NASA GISS) —– Sam Rabin (Karlsruhe Institute of Technology) Etienne Tourigny (Barcelona Supercomputing Center) The recent extreme wildfire events that occurred during the fall of 2017 in Northern and Southern California made world headlines due to their environmental and economic impacts as well as dramatic and catastrophic images. According to the National Centers for Environmental Information (NCEI), the 2017 fall wildfires in California and the Western U.S. generated financial losses estimated at $18 billion, making the 2017 fire season the most destructive in U.S. history. The factors thought to create such dramatic wildfires at the Wildland-Urban Interface (WUI) in California are numerous: a wetter than average winter of 2016 allowed for vegetation to grow abundantly, followed by the warmest summer in recorded history, which dried the excessive fuel, culminating to hot, dry and windy events known as Santa Ana winds in the South and Diablo winds in the North, which allowed for rapid and uncontrolled fire spread. As daily data sources we use the ERA-Interim and NARR re-analyses, gridded products covering an extensive period, and burned area is obtained through the MCD64 global burned area product. This allows the study of the temporal and spatial evolution of fire danger, compared to observed burned area, focusing on extreme events such as those of 2017. The study of the variability of FWI and its input data allow to separate the different physical controls on fire occurrence and understand the relative importance of seasonal climate and weather events. We will also present a framework for seasonal prediction of fire risk based on FWI computed from operational seasonal products. |
Watch Webcast |