What is the Extended Attack Assessment Index (EAA)?
EAA quantifies the fire activity potential using observed satellite-detected hotspots and variables that can explain it. EAA expands the actual used of NFDRS by applying Machine Learning techniques to observed fire behavior and using variables that help to identify extreme conditions. EAA analyzes the next 7 days fire potential in 3 hours intervals. It is produced daily, calculated at different h3 resolutions up to hexagons 1,275 ac and 1.75 mi resolution approximately. One of the main advantages of this index is that it was calibrated with real fires (2012 to 2022) using VIIRS hotspots (NASA-Firms) as a proxy of fire activity. EAA comprises several variables including fuels, topography, short and long-term drought, as well as atmosphere stability and probability of convective fire behavior (Figure 1).
Input variables
Fuel Model Complex (FMComplex)
Fuel model complex is a measure of the fuel model loads and age present in the territory. We plan to include pest impacts in the next update too.
This metric weights the type of structure, its load and the time elapsed since the last fire to quantify how the fuel model may affect fire behavior, fire type and fire suppression difficulty. It is scaled between values 0 and 1 and the nominal resolution is 30 m. We used the Technosylva fuel maps when available and LandFire in those States where Technosylva has not developed high resolution fuel maps yet.
Terrain Difficulty Index (TDI)
TDI includes the Accessibility, Penetrability and Fireline Opening Indexes. It is scaled between unique values 1 and 5 and the nominal resolution is 30 m.
Evaporative Demand Drought Index (EDDI 6m)
EDDI is calculated from observations of the atmosphere near the land surface: temperature, humidity, windspeed, and solar radiation. Nominal resolution is 12 km, with a 5 days lag. It is computed with different time windows. After an evaluation of different time configurations, from short term to longer term, we found the best results with the EDDI 6 months product. Data source: NOAA-PSL.
Vapor Pressure Deficit (VPD)
The vapor pressure deficit is directly related to the plants transpiration rates, turning it an indicator of the vegetation moisture status. It is also correlated with dead fuel moisture content. As a guideline, the following classification used in agriculture can be used to quantify the effect of VPD and to assess water losses through evapotranspiration:
-
High VPD. High transpiration with values above 25 hPa.
-
Optimal VPD. Optimal transpiration levels, values between 7 hPa and 25 hPa.
-
Low VPD. Low transpiration levels, values below 7 hPa.
Wind Speed (WS)
Wind is one of the main factors affecting fire behavior, both in summer and winter. Its contribution to the model is key to identify wind driven fires. We used RTMA data 2011-2021 for wind speed and VPD and National Blend of Models (NBM) from July 2020- September 2022 for both variables.
Lifted Index (LI)
The lifted index (LI) is the temperature difference between the environment Te(p) and an air parcel lifted adiabatically Tp(p) at a given pressure height in the troposphere (lowest layer where most weather occurs) of the atmosphere, usually 500 hPa (mb). The temperature is measured in Celsius. When the value is positive, the atmosphere (at the respective height) is stable and when the value is negative, the atmosphere is unstable. With LI we can estimate the occurrence of storms, their severity, and analyze lighting fires.
Convective Flag (CFG)
Recently, a procedure for identifying atmospheric conditions that favor deep, moist plume growth in wildfire smoke plumes was proposed, which estimates a theoretical minimum heat flux (or firepower) required for such plume growth, named the Pyrocumulonimbus Firepower Threshold (PFT, e.g., Tory 2019).
We included CFC in the EAA model to consider the potential of pyroconvection. Lower CFG numbers indicate more unstable conditions and potential development of convective fires.