Understanding the data

There are many data types that can be labeled and used with SIT_FUSE. In this example, we will take a look at scenes of a wildfire taken by various satellite instruments and compare.

The William's Flat's Fire was well documented through the 2019 FIRE-X AQ field campaign by multiple different airborne and satellite instruments, including AirMSPI, eMAS, GOES, etc.

Through each instrument, we are able to see different aspects of the fire due to the abundance of spectral bands. Different bands allow us to see key elements of the fire such as the smoke, the active fire, and the remaining burn scar.

For example, eMAS data, originally comprised of HDF files but since converted to GeoTiff files (preprocessing step), has 38 spectral bands. We can look at these spectral band in QGIS. We will use an eMAS image for our demonstration (Figure 1).

1. Open QGIS.

2. Open your converted GeoTiff file (.tif).

Figure 1. eMAS data of William's Flat's Fire (Multiband color)

1. In the Layer Styling panel, select Multiband color and then from the drop down that appears, select Singleband gray. This will cause the data to be in grayscale (Figure 2).

Layer Styling Panel

Figure 2. Grayscale (Singleband gray)

1. Now, right under the Singleband gray button, we see Gray band and Band 01 (Gray). Select the latter and you should see a drop down of all of the different bands. Of course the number of bands will vary based on what instrument the data was sourced from, but for eMAS we see 38.

eMAS bands

1. For example, using the eMAS image, let's select band 25, which appears much darker except in areas where there is an active fire (Figure 3). Now, we can use QGIS to label this fire.

Figure 3. Active fire (band 25)

Labeler's Note: For best practices, look through all of the bands when using data from different instruments and make note of which bands from which instruments you can see the feature you are looking for the best. In addition, you should check documentation of the bands from the instrument website itself. Here's an example.

Every instrument will have different bands where certain features appear most prominently. For example, looking at the active fire on data from two instruments, band 25 on eMAS data shows us the active fire while band 30 on MASTER data also shows us the active fire. But if we look at band 30 on the eMAS data, it doesn't show us the active fire at all. For our example, band 1 is good for seeing the smoke plume, band 9 (vegetation band) is good for seeing the burnscar, and bands in the 20's (depending on the instrument) are good for seeing the active fire. But again, this varies heavily, so make sure to familiarize yourself with your instrument and the features you are looking for.