CIMSS Satellite Blog

Wildfire activity began to increase across the northern portions of Saskatchewan and Manitoba in Canada on 30 June 2008. GOES-11 visible and 3.9 µm “shortwave IR” images (above) showed a number of thick smoke plumes (lighter gray features on the visible images) drifting southeastward from a large cluster of active fire “hot spots” (darker black pixels on the IR images). GOES-11 was placed into Rapid Scan Operations (RSO) during the afternoon hours, so images near the end of the animation were available at 5-7 minute intervals.

The largest fire was located between Pelican Narrows and Sandy Bay in Saskatchewan, as seen in an AVHRR false color image (below, viewed using Google Earth). A close-up view reveals that the fire had actually jumped the only road that was in that area (the seasonal road which connects Sandy Bay and Pelican Narrows). The GOES-11 shortwave IR brightness temperatures associated with this particular fire were as high as 341º K (68º C, 158º F), which is the saturation temperature of the 3.9 µm detectors on the GOES-11 satellite. Note that some small pyrocumulus clouds could be seen developing over this large and very hot fire on the GOES-11 visible imagery (above) as well as on the AVHRR false color image (below).

A closer view of the largest fire using AWIPS images of the 1-km resolution MODIS 3.7 µm and the 4-km resolution GOES-12 3.9 µm IR channels (below) shows the advantage of higher spatial resolution for displaying the shape and coverage of not only the largest fire cluster (located near the center of the image), but also the smaller fires in outlying areas. Many of the pixels were so hot that the IR brightness temperatures exceeded the 54.5º C upper threshold for AWIPS display, and showed up as black pixels (registered as “NO DATA“) on the imagery. The smoke from this fire was restricting surface visibility to 1 mile at Flin Flon (CYFO) and 3 miles at The Pas (CYQD) in Manitoba, even though those 2 sites were not in the direct path of the thickest portion of the smoke plume.

Some clues as to the locations of the hottest portion of the fire — which happened to be located within the eastern half of the active fire area, where the black “NO DATA” pixels were seen on the MODIS shortwave IR image — could be found by examining other MODIS images and products: note the darker black pixels on the 11.0 µm “IR Window” channel, the brighter white pixels on the 2.1 µm near-IR “Snow/Ice” channel, and the darker red pixels on the Land Surface Temperature (LST) product (below). AWIPS cursor sampling indicated that the hottest pixel on the IR Window image was 52º C (126º F), while the hottest pixel on the LST image was significantly warmer at 145º F (63º C).

Today’s CIMSS Satellite Blog entry will take a step back and focus on the larger scale. We’ll begin with an animation of GOES-12 visible images from 25 June 2008 (above). Several items of interest are apparent in addition to the normal diurnal development of thunderstorms across parts of North and Central America: (1) the very large area of haziness that occupies the lower right quarter of the images is due to airborne dust from Africa; (2) thick smoke from wildfires is evident over much of northern California and Nevada; (3) a good deal of Hudson Bay in Canada is still frozen. Such is the diversity of meteorological phenomena that can be seen on satellite imagery in late June!

(1) To confirm that the haziness seen on the visible imagery above is due to African dust, we examine the Meteosat-9 Saharan Air Layer (SAL) tracking product (above). The westward progression of a large Saharan dust outbreak (orange to red color enhancement) can be clearly seen during the 21-25 June period. The areal coverage of the airborne dust over the Atlantic Ocean could also be seen on a composite of AVHRR false-color imagery from 23 June (below, displayed using Google Earth).

(2) To explore the impacts of the thick smoke over northern California, it is interesting to view an AWIPS image comparison of the MODIS Land Surface Temperature (LST) product and the MODIS visible channel (below). Note that the LST values seemed to be about 20º F cooler under the areas of thickest smoke (in the 100-110º F range, orange to light red colors) compared to areas farther to the north and to the south of the smoke (where LST values were in the 120-130º F range, darker red colors). The actual air temperatures were not affected by such a drastic amount, although air temperatures in the Sacramento Valley were in the mid-upper 80s F under the thickest smoke (compared to low to mid 90s F farther to the south in the San Joaquin Valley. At Red Bluff in northern California (KRBL) the maximum temperature on 25 June was 87º F (the surface visibility was 2.5 miles or less the entire day due to smoke), several degrees below the daytime maximum temperatures of 98º F, 95º F, and 96º F on the previous 3 days (KRBL 96-hour meteorogram).

(3) Finally, to confirm that ice still remained in much of Hudson Bay, we next turn to MODIS true color and false color images from the SSEC MODIS Direct Broadcast site (below). The majority of the bright features seen over Hudson Bay on the true color image are indeed ice, which appears as a darker red color on the false color image (in contrast to supercooled water droplet clouds, which appear as varying shades of white).

Numerous wildfires were started by lightning activity across Northern California on 21 June 2008, and GOES-11 visible images from 22 June (above) showed a large number of smoke plumes drifting across that region. A MODIS true color image from the SSEC MODIS Today site (below) showed that areas of thick smoke remained across much of northern California on 23 June, with some of the smoke still trapped in the valleys of the North Coast Range. The thick smoke drifting eastward on 22 June reduced surface visibility to 2 miles at Redding (station identifier KRDD), and was causing major air quality problems over a good deal of that area.

An AWIPS image comparison of the 1-km resolution MODIS 3.7 µm and the 4-km resolution GOES-11 3.9 µm shortwave IR channels from around 06 UTC on 23 June (11 PM local time on 22 June) demonstrated the improved ability to detect many of the smaller fires using higher spatial resolution data (below) . In addition, the actual locations of the larger fires were more correctly depicted on the 1-km MODIS image; the comparatively large 4-km GOES IR field of view (and to a lesser extent, the large geostationary satellite viewing angle) tends to diminish the accuracy of such small-scale image details. Improved fire monitoring will eventually be possible using the Advanced Baseline Imager (ABI) instrument on the upcoming GOES-R satellite (scheduled to be launched in 2014), which will offer similar IR imagery at a 2-km spatial resolution (at 5-minute intervals on a routine basis).

UPDATE: A MODIS true color image from 24 June (below, displayed using Google Earth) showed that the thick smoke had increased in areal coverage across much of northern California. MODIS Aerosol Optical Depth values as high as 1.0 were seen across much of the Sacramento Valley region of northern California; surface visibilities on that day at Redding and Red Bluff were as low as 1 mile due to smoke.