10 November 2007

A fast-moving NT fire

An large and very quick-burning bushfire was seen in late-October in eastern NT. The NAFI website does not record this fire in the real-time fire scar mapper. The fire occurred in a very remote area, about midway between Lajamanu and Yuendumu, roughly 100 km SSE of the AWS at Rabbit Flat. As a very rough estimate,made by estimating the position and shape of the burnt area in NAFI from the satellite imagery, the total area burnt was on 100 000 – 150 000 ha. All told, the fire only burned for 3-4 days and overwhelming majority of the spread of the fire occurred in a roughly 24-h period.

I didn't actually catch this in real-time, as I was still traveling when it occurred. Rather, I have had to track the all the information down approximately two weeks later. AS it is in such a remote area, there are no media reports to work off of.

Below is a set of daily images culled from the “Australia2” MODIS subset image archives. I have excerpted the small area around the fire. The area shown in the various images is the same in all of them. The pixel size is 1 km and the image size is 520 x 240 pixels. The odd-shaped white spot in the southwest corner is Lake Mackay, which straddles the WA/NT border. For a scale, the longest length of the lake, along the diagonal, is about 100 km. The images speak for themselves; I will only add a few comments.


25 October – no fire apparent

26 October – fire begins. Terra image (not shown) is clear, shows small
hotspots. Aqua has clouds, pyrocumulus suggested rising from location of hotspot. Also a small hotspot to SE is seen on the Terra image.

27 October – some small fire spread. Two sets of hotspots in Terra image (not shown). In Aqua image, the eastern hotspot is obscured by pyrocumulus and smoke.

28 October – fire explodes. Huge smoke plume extending to the ESE. The fire to the SE also expands, but not to the extent of the northern area.

29 October – still some hotspots, but no expansion of area. Aqua image (not shown) suggests plume is now traveling to the NW (i.e. From the SE). Terra image shows full extent of fire. Terra image time uncertain, as fire region is between two swaths of data.

30 October (no image)– extensive clouds, but no hotspots or smoke are apparent in breaks. Fire apparently out.


The blowup of the fire between the 27th and 28th was apparently associated with the passage of a deep low in the Bight. Surface charts suggest W to NW winds, which gradually increased from the 26th to the 28th, peaking on the last day. Maximum temperatures were above 40C at both Lajamanu and Rabbit Flat, 2-4 degrees above normal. Fire dangers were approaching extreme (at a minimum!) with fully cured grass and a even a moderate wind. Quite possibly a fire ban was issued for the region. There was one ongoing in SA that day...The fires were possibly started by lightning. The satellite imagery for both the 25th and 26th (archived here) does show some suggestion on convective activity in the general vicinity. Yuendumu reported 20 mm of rain in the 24 hrs to 9am on the 26th, likely of convective origin.

No comments: