With respect to the weather, there's no doubt that your attention to details can make the difference between an uneventful flight and one that you regret for the rest of your life. And as you get closer to your departure time, you should rely less on forecasts and more on observations.
I thought this was an interesting case where none of the forecasts really picked up on a low IFR event until after the event was occurring. This event occurred on the evening of May 31, 2019 PDT in the SoCal area. For simplicity, let's look specifically at Van Nuys, CA (KVNY) here.
It's often not a bad idea to look at the previous days weather to see if a similar event occurred and when. Often when the weather is very homogeneous, persistence can be an excellent forecast. That is, you can expect the same event and timing will occur the following day. During the prior evening, the low IFR event didn't arrive until 0748Z when a SPECI was issued for a broken ceiling at 500 feet (BNK005) with a visibility of five statute miles as shown below (oldest observations are presented first). Notice the visibility is 10SM (or potentially better) throughout the entire morning and into the late afternoon and early evening on the previous day until the visibility started to drop around 0600Z.
KVNY 301851Z VRB04KT 10SM CLR 26/12 A2990 RMK AO2 SLP116 T02560117=
KVNY 301951Z 11006KT 10SM CLR 27/11 A2988 RMK AO2 SLP111 T02670111=
KVNY 302051Z 13008KT 090V160 10SM CLR 28/11 A2987 RMK AO2 SLP107 T02780111 58011=
KVNY 302151Z 12009KT 090V150 10SM CLR 28/10 A2986 RMK AO2 SLP102 T02780100=
KVNY 302251Z 14009KT 10SM CLR 28/09 A2984 RMK AO2 SLP097 T02830094=
KVNY 302351Z 14009KT 10SM CLR 28/09 A2983 RMK AO2 SLP092 T02780089 10283 20222 58014=
KVNY 310051Z 14008KT 10SM CLR 27/08 A2982 RMK AO2 SLP090 T02670083=
KVNY 310151Z 11007KT 10SM CLR 24/10 A2983 RMK AO2 SLP094 T02440100=
KVNY 310251Z 14007KT 10SM CLR 19/12 A2986 RMK AO2 SLP104 T01940117 53010=
KVNY 310351Z 10006KT 10SM CLR 17/12 A2988 RMK AO2 SLP111 T01670122=
KVNY 310451Z 11005KT 10SM CLR 16/12 A2990 RMK AO2 SLP119T01610122=
KVNY 310551Z AUTO 12005KT 9SM CLR 16/12 A2991 RMK AO2 SLP120 T01560122 10278 20156 51017=
KVNY 310651Z AUTO 14003KT 8SM CLR 15/12 A2991 RMK AO2 SLP120 T01500122=
KVNY 310736Z AUTO 10006KT 6SM BR SCT007 14/12 A2991 RMK AO2 T01440122=
KVNY 310748Z AUTO 12006KT 5SM BR BKN005 14/12 A2991 RMK AO2 CIG 004V008=
KVNY 310751Z AUTO 12005KT 4SM BR BKN005 14/12 A2991 RMK AO2 CIG 003V008 SLP121 T01440122 402830133=
KVNY 310812Z AUTO 13006KT 4SM BR OVC004 14/12 A2991 RMK AO2 T01390122=
KVNY 310851Z AUTO 15006KT 4SM BR OVC003 14/12 A2991 RMK AO2 SLP120 T01390122 51001=
KVNY 310933Z AUTO 15004KT 5SM BR OVC005 14/12 A2991 RMK AO2 T01390122=
KVNY 310951Z AUTO 10004KT 5SM BR OVC005 14/12 A2990 RMK AO2 SLP118 T01440122=
KVNY 311033Z AUTO 18005KT 4SM BR OVC004 14/12 A2990 RMK AO2 T01390122=
KVNY 311043Z AUTO 15004KT 4SM BR OVC005 14/12 A2990 RMK AO2 CIG 004V007 T01390122=
Here's the official amended forecast (TAF) for VNY for the following day leading up to the next low IFR event. The forecaster went with the belief that the ceiling would once again go down to an IFR flight category beginning at 0800Z with a forecast of an overcast sky at 900 feet (OVC009), but kept the visibility in the VFR category or better than six statute miles (P6SM).
KVNY 312024Z 3120/0118 15008KT P6SM SKC
FM010300 14005KT P6SM SKC
FM010800 VRB03KT P6SM OVC009
FM011700 14005KT 5SM HZ SCT012
In general terms, the forecast made sense to me given there wasn't any significant change to the air mass between the two days. And none of the MOS forecasts (LAMP, MAV, MET) seemed to pick up on this low IFR event - see below.
But what made me pause was the fact that the visibility struggled to recover from the previous night's IFR event discussed above. You can see below (oldest observations are presented first) that it did manage to peak at 10SM at 2200Z about when the high temperature was recorded for the afternoon, but then trended down once again ushering in the low IFR event much earlier than forecast. At about 0407Z, the ceiling was broken at 600 feet (BKN006) and the visibility was 7 statute miles. The event occurred about four hours earlier than the previous day's event.
KVNY 311551Z VRB03KT 4SM HZ BKN017 17/12 A2991 RMK AO2 SLP122 T01670122
KVNY 311651Z 19004KT 5SM HZ BKN017 19/12 A2990 RMK AO2 SLP118 T01890122
KVNY 311658Z 16003KT 4SM HZ CLR 19/12 A2989 RMK AO2 T01890117
KVNY 311751Z VRB03KT 6SM HZ CLR 21/12 A2990 RMK AO2 SLP118 T02110122 10217 20139 58002
KVNY 311751Z VRB03KT 6SM HZ CLR 21/12 A2990 RMK AO2 SLP118 T02110122 10217 20139 58002
KVNY 311851Z 14007KT 090V170 7SM CLR 22/12 A2989 RMK AO2 SLP115 T02220122
KVNY 311951Z 15007KT 8SM CLR 23/12 A2987 RMK AO2 SLP110 T02330122
KVNY 312051Z VRB05KT 8SM CLR 24/13 A2986 RMK AO2 SLP103 T02440128 58014
KVNY 312051Z VRB05KT 8SM CLR 24/13 A2986 RMK AO2 SLP103 T02440128 58014
KVNY 312151Z 14007KT 100V170 10SM CLR 25/13 A2984 RMK AO2 SLP097 T02500128
KVNY 312251Z 13010KT 9SM CLR 24/13 A2982 RMK AO2 SLP093 T02390128
KVNY 312351Z 13009KT 10SM CLR 23/13 A2982 RMK AO2 SLP092 T02280128 10256 20211 56012
KVNY 010051Z 13011KT 9SM CLR 21/12 A2981 RMK AO2 SLP090 T02060122
KVNY 010151Z 12008KT 8SM CLR 18/13 A2982 RMK AO2 SLP094 T01830128
KVNY 010251Z 15006KT 7SM CLR 16/12 A2985 RMK AO2 SLP101 T01560122 53008
KVNY 010351Z 14008KT 7SM FEW006 14/12 A2986 RMK AO2 SLP105 T01440122
KVNY 010407Z 13007KT 7SM BKN006 15/13 A2987 RMK AO2 T01500128
KVNY 010451Z 12007KT 6SM BR OVC007 15/13 A2988 RMK AO2 SLP110 T01500128
KVNY 010551Z AUTO 17006KT 6SM BR OVC006 14/13 A2988 RMK AO2 SLP113 T01440128 10228 20144 51013
KVNY 010651Z AUTO 18005KT 4SM BR OVC005 14/13 A2988 RMK AO2 SLP111 T01390128
KVNY 010710Z AUTO 20004KT 4SM BR OVC004 14/13 A2988 RMK AO2 T01440128
KVNY 010743Z AUTO 17005KT 4SM BR OVC005 14/13 A2988 RMK AO2 T01390128
KVNY 010751Z AUTO 17005KT 4SM BR OVC005 14/13 A2987 RMK AO2 SLP110 T01390128 402560139
KVNY 010851Z AUTO 18004KT 4SM BR OVC005 14/13 A2986 RMK AO2 SLP106 T01390128 58006
KVNY 010951Z AUTO 00000KT 4SM BR OVC006 14/13 A2986 RMK AO2 SLP103 T01440128
KVNY 011042Z AUTO 17004KT 5SM BR OVC010 14/12 A2985 RMK AO2 CIG 008V013 T01440122
KVNY 011051Z AUTO 15003KT 5SM HZ OVC012 15/12 A2985 RMK AO2 SLP100 T01500122
The fact that the visibility struggled to recover signaled more robust moisture near the surface. With the additional moisture remaining from the previous day's IFR event, the potential of an earlier onset was looming. These trends are key to raising the necessary red flags and avoiding an unexpected encounter with adverse weather. Despite the poor MOS forecasts, I had full confidence that a low IFR event would happen. The writing was on the walls based on the IFR event on the previous day. But sometimes it's all about timing.
The moral of the story is to always pay close attention to the observations once you are within a couple hours of your flight. Not that you need to ignore the forecast, but close scrutiny of the observational data will often uncover weaknesses (if any) in the forecast. It's all about minding the breadcrumbs left behind by Mother Nature. Noticing that the visibility never recovered was a key element here to "doubting" the accuracy of the terminal forecast.
Even the 00Z TAF issued at 2340Z (below) didn't catch this. In fact, the forecast of an overcast sky at 1,500 feet forming at 08Z, was moving the ceiling in the wrong direction.
KVNY 312340Z 0100/0124 13010KT P6SM SKC
FM010300 14006KT P6SM SCT250
FM010800 VRB03KT P6SM OVC015
FM011400 VRB03KT P6SM OVC020
FM012100 17006KT P6SM SCT025 SCT250=
It wasn't until the IFR ceiling actually developed after 0400Z that the forecaster amended the TAF at 0430Z.
KVNY 010430Z 0105/0124 13010KT P6SM BKN006
FM010700 VRB03KT P6SM OVC012
FM011400 VRB03KT P6SM OVC020
FM012100 17006KT P6SM SCT025 SCT250=
Given that the TAF didn't do so well, how about the automated forecasts? Unfortunately, none of the automated forecasts provided useful guidance either. Both the NAM MOS (MET), GFS MOS (MAV) or LAMP MOS (LAV) began to forecast the potential for this event only after it began to occur.
All of these automated MOS tools show a categorical forecast for ceiling and visibility. For example, a forecast ceiling of 3 does not mean 300 feet, it means it's in the IFR flight category of 500 to 900 feet. Below is the secret decoder ring to translate categorical forecasts.
Up first is the MOS bulletin from the NAM (MET). The NAM MOS is only issued twice a day at 00Z and 12Z. Here's the 12Z issuance for Van Nuys. The red box shows the forecasts valid at 03Z, 06Z and 09Z. As mentioned above, actual IFR conditions began around 04Z with a broken ceiling of 600 feet. The NAM was forecasting a clear below 12,000 feet condition for 03Z and 06Z and a marginal VFR ceiling (overcast) of 1,000 to 1,900 feet at 09Z. Visibility was forecast to be greater than 6 statute miles (categorical forecast of 7) during all three times.
The GFS (MAV) MOS is updated 4 times a day and below is the 18Z issuance for Van Nuys. This forecast is a little better than the NAM with a categorical forecast of a MVFR ceiling of 5 (2,000 to 3,000 feet) at 06Z and categorical MVFR ceiling of 4 (1,000 to 1,900 feet) at 09Z. Similar to the NAM, the GFS forecast for visibility is greater than 6 statute miles during all three times.
Next is the guidance from the LAMP. This forecast is issued hourly and uses the GFS MOS as once input source. In addition to its hourly refresh (it actually refreshes every 15 minutes for some elements such as ceiling and visibility), it is designed to improve on the GFS MOS for the first 24 hours and has a higher temporal resolution as well. Here is the 22Z issuance.
In this case, it had the worst performance of the three MOS bulletins, forecasting ceilings no worse than a VFR category of 6 (3,100 to 6,500 feet). The first indication of MVFR ceilings were not until 11Z.
I reached out to the folks at the Meteorological Development Laboratory (MDL) to see if their new version of LAMP (still experimental) would have picked up on this IFR ceiling much sooner. It did provide a better forecast than the operational version of LAMP with the onset of MVFR conditions occurring at 07Z, but was showing clear below 12,000 feet at 04Z, 05Z and 06Z.
Nevertheless, none of the automated forecasts shown above provide good guidance to pilots. The 18Z TAF was much better at predicting an IFR ceiling, but not until after 08Z. Even then, the next update of the TAF at 00Z didn't catch this either and made it worse by being in better alignment of the MOS forecasts. There's no way to know if the MOS guidance swayed the forecaster (or subsequent forecaster) to change the forecast from IFR to MVFR.
Most pilots are weatherwise, but some are otherwise™
Scott Dennstaedt
Weather Systems Engineer
CFI & former NWS meteorologist
