Flight Icing Threat evaluation over the Midwest

The Flight Icing Threat index and associated products have been the highlight of the GOES-R Proving Ground evaluation at the Winter Experiment this week. This morning we saw the good and the not so good, resulting in very good discussion and in detail examination of the product.

The good...

By around 1445 UTC the FIT was showing consistently high index values of MOG likely and some heavy in west central Iowa.

20160223 1445 UTC Flight Icing Threat index with PIREPs overlaid

Shortly after 1500 UTC, a small Cessna aircraft reported moderate to severe icing. This information was passed on to the CAWS folks as they debated whether to issue something for that region. The area with high FIT values was fairly small and isolated, and further examination of icing layer bases/tops showed that it was also fairly thin, on the order of ~3,000 feet or less. Soundings around the region at the same time corroborated this. More PIREPs occurred fairly quickly after that, but none that reached above light intensity.

Some debate was sparked as to whether CAWS (or SIGMETs) are issued for instantaneous icing or accumulated icing. Given the thin layer in this case, it wouldn't take much ascent or descent for an aircraft to exit this layer. For this reason, it was theorized that the pilot may have stayed in this layer for some time during this case, causing the accumulation.

Regardless, the FIT was a useful tool in further identifying which cloud was causing the icing and where it may be in the future. In this particular case a CAWS was not issued at that time given the thin layer and variable environment. Additionally, the already issued G-AIRMET would have sufficiently covered the threat occurring in that area. Even so, the area was continually monitored throughout the remainder of the day.

The not so good...

There were also some challenges noted in the FIT today, particularly over the Great Lakes. Note the area of high index values over southern WI and south central MI. This area remained consistent throughout the morning, but the yellow bordering around it hinted at something suspect. To determine what was going on, the supercooled liquid droplet imagery, one of the inputs used in the FIT imagery was examined.

20160223 1445 UTC supercooled liquid droplet imagery. Warm colors indicate areas of SLDs and cool colors indicate areas of SLDs where snow may be contaminating the retrieval

The area over Wisconsin and Lake Michigan did in fact indicate large supercooled liquid droplets (red), and subsequently explained the high FIT values. However, it still appeared suspect, as no reports of icing were recorded in the area all morning and visible imagery also showed high, thick clouds over the same region. Next the cloud phase generated by LaRC was examined.

20160223 1415 UTC NASA LaRC cloud phase retrieval

The cyan colors over the same area of interest (southern WI and Lake Michigan) were identified as liquid. In reality those clouds were actually all thick ice clouds. However, because they were identified as liquid, the algorithm then generated SLD and FIT values from those clouds.

Interestingly, the cloud phase algorithm generated using inputs from the GOES-R simulated ABI imagery showed quite different results:

20160223 1445 UTC cloud phase using GOES-R ABI inputs. Orange=ice, dark green=mixed phase, neon green=supercooled, and blue=liquid

This version of the cloud phase identified all of the clouds in the area of interest as composed entirely of thick ice, and would have -if used- helped to reduce the false alarms caused by high clouds. This is one situation in which the benefits of the Advanced Baseline Imager on GOES-R is really highlighted.

In the future, the GOES-R inputs from the ABI will be integrated into LaRC's icing algorithm. It will aid in improving the cloud property retrievals utilized by the Flight Icing Threat index, such as cloud phase, supercooled liquid droplets, etc., and subsequently help to improve the algorithm itself.