Constructing an improved time series of satellite-retrieved fire radiative power during large wildfires

Detailed analysis of fire behavior and fire-weather interactions requires information on fire characteristics with a high temporal frequency. For large wildfires that have an active suppression response, frequently updated information is sometimes available, but it is often unsystematic and challenging to apply quantitatively. Sensors on geostationary satellite platforms offer consistent systematic observations and repeat cycles of 30 minutes or less, with the potential for observations every minute. However, all satellite observations offer relatively coarse information on fire properties, which are subject to both random and systematic sources of uncertainty. The Naval Research Laboratory has characterized these uncertainties and developed a method for processing satellite fire observations into time series of Fire Radiative Power (FRP) that can be systematically compared across dates, across fires, and across multiple fire seasons. This method includes corrections for systematic biases in the observations, normalization for variations in observation sensitivity, and quantitative estimation of uncertainty based on observing conditions, fire size, and observation schedule. The utility of this corrected FRP time series is demonstrated using several large wildfire case studies occurring in western North America during the summer of 2013.