Here in Charlotte a few years back, we were treated to an interesting thundersnow event. The primary cold front moved through in the very early morning hours and was working its way off the east coast. We got a mixture of light rain and snow during most of the early morning with surface temperatures running in the upper 30s to low 40s. But the real threat in terms of snow wasn't going to happen until much later in the day.
The upper-level trough axis was still out to the west. As cold air aloft moved in over a relative warm ground, this created a fair amount of instability in the first 15,000 feet of the ground as can be seen by this Skew-T log (p) diagram.
This isn't much instability compared to what might occur during the spring and summer, but with these cold temperatures and the assistance of the upper-level trough, you can get enough vertical motion to produce thundersnow. Here's the XM weather radar image showing the snow moving in along with lightning (yes, there are four different yellow lightning bolt symbols showing in the red circles).
Now onto the visibility issue. As the weather pushed to the east, Rock Hill airport (KUZA) that's southwest of Charlotte began to experience the push of cold air and snow. At 2054Z, the routine hourly observation showed a temperature of +4C and calm winds.
KUZA 162054Z AUTO 00000KT 10SM SCT023 SCT029 BKN060 04/M01 A2983 RMK AO2 SLP106
I have two customers that live down the road from this airport. They told me it was snowing hard, but Rock Hill still failed to report any precipitation (this is an ASOS). That's because when the precipitation identification (PI) sensor detects -SN, SN or +SN and the temperature is greater than 38°F (as it was in this case) the precipitation identification algorithm negates the snowfall and reports No Precip. So no SPECI is issued even though it's snowing heavy. It's not until 2132Z that lightning occurs within the terminal area and the ASOS finally reports light rain and thunderstorm (-TSRA) with an 8SM visibility. The thunderstorm event is what triggers the SPECI. But still no snow reported.
KUZA 162132Z AUTO 35006KT 8SM -TSRA BKN028 BKN039 OVC047 04/M01 A2984 RMK AO2 RAB33 TSB35
Well, 10 minutes after that observation the ASOS finally caught up with reality and reports a 1/2SM visibility and thundersnow (TSSN) given that the temperature falls to +2C which now meets the threshold for reporting snow.
KUZA 162142Z AUTO 01018G25KT 1/2SM TSSN FEW006 BKN022 OVC044 02/M01 A2986 RMK AO2 LTG DSNT NW RAB33E42SNB42 TSB35
KUZA 162147Z AUTO 35011G23KT M1/4SM VCTS FG BKN011 BKN023 OVC035 00/M01 A2986 RMK AO2 LTG DSNT SW AND NW RAB33E42SNB42SNEMM TSB35E50 P0002 PWINO FZRANO $
KUZA 162154Z AUTO 36012G23KT M1/4SM VCTS +SN FG BKN009 OVC023 00/M01 A2986 RMK AO2 LTG DSNT SW AND NW RAB33E42SNB42SNEMMB54 TSB35E50 SLP118 P0002 I1000 T00001006 $
KUZA 162202Z AUTO 35010G21KT 1/4SM +SN FG BKN003 OVC024 00/00 A2986 RMK AO2 P0001 T00000000 $
It took about 15 minutes for this weather to move over my house once it started in Rock Hill (I live 20 miles to the east of KUZA). It was just astonishing to watch (even for me) when the snow event started. What appears as a drop from 8SM to M1/4SM (less than 1/4SM) in just 15 minutes according to the ASOS likely happened in about 15 seconds. It wasn't more than a blink of an eye and the entire area was in total white-out. Visibility dropped so quickly that I could barely see my neighbor's house across the street.
It goes without saying that, the more uniform the weather the more accurate the automated visibility. But in this kind of rapidly changing environment the visibility sensor algorithm is not equipped to respond instantly. An ASOS employs an algorithm called the "harmonic" mean which is better than just a straight average. Lower visibility values have a greater impact than higher visibility values using a harmonic mean. Nevertheless, quick drops in visibility will still lag behind reality.
As an example, let's say the visibility drops suddenly (in one minute or less) from 8SM to 1SM, it takes about 3 minutes for the 10-minute harmonic mean values to report 3SM and transmit a SPECI. A total of 9 minutes will have elapsed before the algorithm lowers the visibility to 1SM. And it'll be several more minutes before it lowers the visibility to M1/4SM.
The ASOS is intentionally designed to raise surface visibility more slowly than to lower it. This design provides a margin of safety and buffers rapid changes when the visibility is wildly fluctuating over a short period. An ASOS relentlessly measures the weather and could inundate pilots with more frequent special observations than a human observer when the weather is changing rapidly. Thus, the system is purposely throttled to only provide SPECIs at 5-minute intervals to limit the number of observations that can be transmitted during the hour.
Another important element to consider is called the ASOS lockout period. Starting at 47:20 past the hour, the ASOS begins to make its routine hourly observation. By 53:20, the hourly observation has been prepared and should be ready for transmission. Many of the routine hourly METARs will carry an issuance time of 53, 54, 55 or 56 minutes past the hour.
During this lockout period, the ASOS is prevented from issuing any other reports including SPECIs no matter what weather may be occurring. The ASOS still continuously monitors and records the weather during the lockout period; however, it just can’t issue a surface observation. This does not affect the 1-minute weather you receive by calling the voice phone link or by the ground-to-air radio broadcasts, but it will affect any formal observations that get transmitted to flight crews, air traffic controllers and meteorologists at the Center Weather Service Units (CWSUs).
Granted, this thundersnow scenarios is an extreme example, but still shows how quickly the weather can deteriorate while the automated observations play catch-up.
Most pilots are weatherwise, but some are otherwise™
Scott Dennstaedt, PhD
Weather Systems Engineer
CFI & former NWS meteorologist