Challenges of satellite-dervied icing retrievals: false alarms/overestimation in the FIT

While the FIT performed very well over the Great Lakes this morning, there were also several areas of false alarms that were noted. The first was over eastern Kentucky and western North and South Carolina.

20160222 1445 UTC Flight Icing Threat index with PIREPs overlaid

Fairly high FIT index values were noted in this area, however the few icing reports seen were only in the light category. After further examining the parameters that are used as input to the FIT, it was shown that high values of the estimated supercooled liquid droplets were shown in the same area. The FIT algorithm places a heavy weight on SLD thresholds and as such, significantly upped the severity indicated in the index. For this reason, this region was deemed to have been overestimated.

Another area that appeared to have been overestimated or had false alarms was in the Northern Plains. A fairly broad area of heavy and MOG likely (reds and pinks) was seen in eastern North Dakota, western Minnesota, and the nearby Canadian region. Additionally, it was fairly consistent throughout the morning hours. However as with the above-mentioned area, the few PIREPs only reported light icing. In this case however, it was surface conditions that caused the overestimation. This particular region is cold and snow-packed at this time of year, and the subsequent very cold surface temperatures are known to cause false alarms or overestimation in the FIT this time of year.

So, the lesson learned here? While satellite-derived icing retrievals have come a long way over the years, there will -as with any satellite product- always be challenges. Next generation satellite technology such as GOES-R will help improve these retrievals and potentially reduce some of these issues and allow for more accurate data!