The worst airline disaster in Brazil’s history occurred on the evening of 17 July 2007 as a TAM Airlines Airbus A320 was attempting to land at the Congonhas airport in Sao Paulo, Brazil. Early media reports (CNN) indicated that the plane was landing during a “driving rainstorm”, which led us to take a look at GOES-10 satellite imagery to examine the meteorological conditions leading up to the crash. However, GOES-10 10.7µm InfraRed (IR) imagery (above; Java animation) suggests that only light (to possibly moderate) rain might have been falling from the comparatively warm cloud top brightness temperatures (-30 to -40º C, dark blue to green enhancement) seen in the vicinity of Sao Paulo (station identifier SBSP) — the closest area of significantly cold cloud top temperatures (-60 to -70º C, red to black enhancement) indicative of heavy to severe convective rainfall was still far to the west of SBSP over interior southern Brazil at that time.

Indeed, a time series of surface observations or “meteorogram” from SBSP (below) showed only light rain which was reducing surface visibility to 4-7 miles during the hours leading up to the crash at 21:50 UTC (18:50 local time). The rainfall was likely a factor in contributing to this particular tragedy, but it is important to note that the relatively short runway that was used (which had just been resurfaced in June 2007) had not yet been grooved to facilitate water run-off and prevent hydroplaning — other media reports (BBC) also stated that two additional smaller planes had skidded off that same runway only a day before the 17 July accident.

The extensive cloud cover across much of southern Brazil prevented GOES-10 sounder retrievals necessary for the generation of Total Precipitable Water (TPW) Derived Product Imagery (below), which may have offered an additional clue as to the precipitation potential of any convective activity in the Sao Paulo region that day.

View only this post Read Less

The 2007 tropical cyclone season has been rather boring quiet in both the Atlantic and East Pacific basins so far…but on 16 July 2007, Cosme became the first East Pacific hurricane of the season (about 1600 miles east-southeast of Hawaii). It is interesting to note that only 4 previous East Pacific tropical cyclone seasons have had their first hurricane form later than Cosme. An animation of GOES-11 visible imagery (above) showed some hints of an eye from 19:00-21:00 UTC, but subsequent convective bursts around the core of the storm masked the presence of an eye after that time. The CIMSS “TCTrak” analysis tool (below) indicated that Hurricane Cosme existed in an environment of relatively weak deep-layer (200-850 hPa) wind shear, which was a factor that aided in the slow intensification to hurricane strength during the day. The “TCTrak” tool is a feature on the newly-revised CIMSS Tropical Cyclones website that allows the user to select a variety of satellite views and meteorological product overlays.

View only this post Read Less

As much as we enjoy evangelizing satellite imagery and demonstrating all its useful applications, there are situations when satellite imagery simply is not as helpful as we would hope it might be. On 11 July 2007, a lightning strike started a fire at the Sunoco Oil Refinery near West Deptford, New Jersey (just south of Philadelphia, Pennsylvania, station identifier KPHL). An AWIPS image of the GOES-12 10.7µm IR channel with cloud-to-ground lightning strike data overlaid (above) was rather unremarkable in it’s presentation of the convection that was producing the lighting just prior to the time that the fire reportedly started. While there were some cloud top brightness temperatures as cold as -62º C (dark red enhancement) located a bit farther to the north between KPHL and KPNE, there was no other IR signature (such as an “enhanced-v”) that would suggest a severe convective potential; in fact, IR brightness temperatures in the area of the southernmost cluster of oil refinery lightning strikes were as “warm” as about -48º C (light green enhancement). In addition, the storm was producing just over 100 cloud to ground lightning strikes every 15 minutes, but that number is by no means extraordinary. Finally, cloud cover in the hours following the start of the fire prevented a “hot spot” from being detected on the GOES-12 3.9µm IR channel imagery.

View only this post Read Less

A GOES-12 imager “decontamination procedure” was performed on 02 July 2007, where certain internal optical components were heated in an attempt to drive off contaminants that had been accumulating for several years. One obvious problem associated with this internal contamination was a “rollover” of warm 3.9µm IR brightness temperatures, such that they were incorrectly displayed as very cold temperatures. This problem made it impossible to determine the hottest brightness temperatures associated with areas of intense fire activity (see recent examples from the Lake Tahoe, Lake Okeechobee, New Jersey, and Georgia fires).

Two sequences of daily GOES-12 3.9µm IR images show regions of sun glint off the Pacific coast of Mexico at 20:45 UTC (above), and also off the coast of Washington / British Columbia at 03:45 UTC (below). Prior to the decontamination procedure at 12:34 UTC on 02 July, portions of the sun glint that should have exhibited very warm IR brightness temperatures (black enhancement) were instead displayed as very cold areas (yellow enhancement); following the decontamination, the regions of daily sun glint were correctly displayed as having warm IR brightness temperatures, indicating that the procedure was successful.

View only this post Read Less