Sedgwick Fire

Below: Comparison of 7/14 (left) versus 7/19 (right) Sentinel-2 images. Spectral change is minor. Possibly some fading of spectral signature of fire retardant.

Below: Spectral difference (true color, center ; false color SWIR/NIR/Vis, right) versus reflectance (false color, SWIR/NIR/Vis, left) from 7/14 to 7/19.

• Areas of fire retardant are fading by becoming greener at visible waelengths

• SWIR-bright areas are losing SWIR reflectance (reduced smoldering?) 

• many low albedo areas are brightening

• many high albedo areas are darkening

• greenest areas in west of image (outside fire scar) are reddening

BRDF effects are possible (7/4 and 7/14 were S2B ; 7/19 is S2A). Should be better constrained in 7/24 image.

Below: WorldView-2 quicklook. July 9th. 8-band VNIR multispectral. 2m native resolution, 0.5 m panchromatic. Data ordered.

Below: Sentinel-2 True Color (left) and SWIR/NIR/Visible False Color (right). July 14th.

Below: False color composite (SWIR/NIR/Vis) of (after - before) spectral change. Bright areas indicate high albedo ash and dark areas have low albedo ash and char.

Below: Before (7/4; left), During (7/9; center), and After (7/14; right) fire images from Sentinel-2. False color composite, SWIR/NIR/Vis.

Below: Reflectance spectra before (left) & after (right) fire projected onto the same (after fire) Principal Component basis. Change in scatterplot position implies burning.

Below: UMAP (NN = 50) achieves complete stastical separation of burned vs unburned pixel spectra.

Below: UMAP of stacked Sentinel-2 pixel spectra identifies statistically distinct burn trajectories.

Below: EMIT image from June 24, 2024. True color (left) and false color SWIR/NIR/Visible.

Below: Histograms of aerosol optical depth (AOD; unitless; yellow) and water vapor (cyan; g/cm2) for the full 6/24 EMIT scene (right) and Sedgwick subset (left):