Applying NUCAPS as a Predictive Tool to Mitigate Dangerous Firefighting Conditions

The Naval Research Laboratory - Marine Meteorology Division (NRL-MMD) is developing an
algorithm to predict pyrocumulonimbus (pyroCb) development over wildfires in North America.
These wildfire-generated thunderstorms present many hazards to firefighters in the form of
sudden wind shifts, enhanced spotting, lightning strikes, and even tornadic activity. Inputs are
derived from thermodynamic profiles obtained near active wildfires from the NOAA Unique
Combined Atmospheric Processing System (NUCAPS) onboard NOAA-20, including near-surface
lapse rate and vapor pressure deficit, mid-troposphere moisture content, and instability
parameters. NUCAPS profiles are skillful at identifying the atmospheric conditions known to
support pyroCb development, such as enhanced mid-level potential instability and the classic
low-level “inverted V” shape, with near dry adiabatic lapse rates. The algorithm is trained using
an inventory of 106 pyroCb events in North America during 2020 – 2022 and a similar dataset of
large wildfires that did not produce pyroCb activity (null cases). PyroCb probability of detection
is 78% and 85% when using K nearest neighbor (KNN) and logistical regression models,
respectively. Output is currently displayed as probability of pyroCb activity for clusters of active
fire pixels detected by GOES (example shown below). Initial real-time tests are provided for
several pyroCb events observed during the active Canadian fire season of 2023. The overall
objective is to provide pyroCb early warnings within a prototype pyroCb detection algorithm
framework developed by NRL-MMD to assist the firefighting community and Incident
Meteorology (IMET) teams in their daily operational schemes to combat wildfires.