This afternoon’s MODIS data sets were just posted:
I’ll zoom in on the new activity:
I think there are three classes of hot spots.
Those that are probably wildfire:
Those that are probably back burns:
And those that could be either:
The orange lines are the dozer fire breaks that were in as of 7/30, plus on that I added by hand. The pink lines are the ones that were planned as of that date. The hotspots near the dozer breaks are back burns. Some of the fire to the west of there may be as well, but it’s hard to tell from these satellite images.
How to read the MODIS/VIIRS heat indications: The size of the square represents the nominal margin of error. The fire could be anywhere in the square, not just at the center. Dark red squares were detected less than six hours before the data set was created. Light red squares were detected less than twelve hours before the data set was created. Orange squares were detected less than 24 hours before the data set was created. Yellow squares were detected less than six days before the data set was created. When the squares overlay other objects, the color of the square changes somewhat, but the color of the dot in the middle does not. So look at the dot if you’re uncertain what color the square should be. Since the fire is now over six days old, some of the early detections have dropped off the map. MODIS makes mistakes, sometimes missing outbreaks, and sometime misplacing them outside the nominal margin of error. It is also a snapshot of the activity at the time the satellite is overhead, and will definitely miss flare ups between passes. MODIS can’t tell the difference between wildfires and intentional back burns.
By the way, the more I work with the VIIRS high-resolution data sets, the more I think that they are far preferable to the MODIS data, when they are available. It appears to me that, not only are the uncertainty squares smaller, but that the actual errors are more likely to be within the squares.
Here’s an example of what I’m talking about. Let’s look at the southwestern part of the active fire, first with MODIS:
And then wit VIIRS hi-res:
Look at the relationship between where VIIRS put the hot spots that are outside of the black line that marks last night’s official fire boundary and and where MODIS put them. I should mention that the timing of the data sets are different.
Leave a Reply