Weather Forecasting accuracy deteriorating

brcase

brcase

En-Route
PoA Supporter
Joined
Jun 11, 2008
Messages
3,267
Location
Boise, Idaho
Display Name
Brian
For Soaring we use a number for forecasting models. We have been noticing the accuracy of these models is not as good as it has been in the past.

In researching this loss of accuracy a friend sent me this...

"the US weather service has had a number of their weather balloon sites shut down, more noted today. The issue with Sky Site model not being very accurate is due to the reporting station info they use for NW Weather. Now, for us it is BOI, Salem, OR, Medford, Auburn, WA, Quillayute, WA I talked to them last week, and these sites do not launch twice a day every day due to staffing - DOGE and lack of helium They do not use sites east of us or south. The other issue is the time delays. They now rely more on ground observations which are lacking upper air info which is the big part of the weather forecast for them. This is becoming a big issue for flying safety and bad weather forecasting. "
https://www.weather.gov/upperair/nws_upper

1742833567179.png
Brian
 
as that risk of this thread continuing to spiral down the SZ path:

<deleted>

But a more important question is how people can get essential services restored if they've been cut due to whatever reason.
 
Helium shortage is more a cost issue (any supply shortage can be overcome with $$$). I paid $350 for a cylinder of 99.999% and delivery took a couple weeks.
 
I have noticed the models seem to be more erratic run from run. I wonder what level the models depend on atmospheric reading inputs, and if a reduction in those causes the models to not produce as accurate a forecast.
 
bonefishpete said:
Back in the day anyone remember landing at an airport and walking into a Flight Service Station to look at radar returns? We got by when they closed.
Click to expand...
Ummm.... the information was available to us directly. Back in the day you HAD to go to them.
 
455 Bravo Uniform said:
Helium shortage is more a cost issue (any supply shortage can be overcome with $$$). I paid $350 for a cylinder of 99.999% and delivery took a couple weeks.
Click to expand...
Always funny when I get old timers asking if I can heli-arc something. Most people dont know its just tig welding with helium instead of argon. I have a helium tank in the cage, but rarely use it Helium prices are nuts. Supposedly they found a massive supply in Minnesota.
 
bonefishpete said:
Back in the day anyone remember landing at an airport and walking into a Flight Service Station to look at radar returns? We got by when they closed.
Click to expand...
Yeah my dad did FSS before it went automated. A remember well the chatter of teletype machines, DF steer consoles, stacks of paper FLIPs and yes, they launched weather balloons as well. They also took weather observations with an old school ceilometer and a sling psychrometer…a lost art. He still types in all caps on FB today.
 
Funny thing is, I don't recall anyone talking about how good the weather forecasts were before now. What changed to make them seem better back then?
It reminds me of the saying "The older I get the better I was".
 
bonefishpete said:
Back in the day anyone remember landing at an airport and walking into a Flight Service Station to look at radar returns? We got by when they closed.
Click to expand...
True, but they didn't turn off the radar, just moved the access. Now they are not able to continue to operate some of the sources of data.

It's a little like folks saying "I don't need NOAA, I've got weather on my phone." Where does the phone get the weather?
 
Velocity173 said:
Yeah my dad did FSS before it went automated. A remember well the chatter of teletype machines, DF steer consoles, stacks of paper FLIPs and yes, they launched weather balloons as well. They also took weather observations with an old school ceilometer and a sling psychrometer…a lost art. He still types in all caps on FB today.
Click to expand...
DZ HE ABRVIAT THE HK OT OF EVRYTNG?
 
flyingcheesehead said:
DZ HE ABRVIAT THE HK OT OF EVRYTNG?
Click to expand...
Ha! Nah he doesn’t do that but he really does do all caps. Hurts my head to read his posts.

The other thing I always found interesting is the change in attire through the years. He started in the early 70s and retired in 98. I look at old pics and everyone had buttoned shirt, tie and slacks. By the time he retired it was just pollo shirt and cargo pants. My brother did AFSS for awhile and when he left in 2005 it was T-shirt, jeans and ball cap! Saw the same evolution with ATC.
 
JOhnH said:
Funny thing is, I don't recall anyone talking about how good the weather forecasts were before now. What changed to make them seem better back then?
It reminds me of the saying "The older I get the better I was".
Click to expand...
They weren't any better back then, 50yrs ago.

I remember getting a forecast for a night x-c for dinner. Walked out of the airport restaurant to see light flakes on the airplane but stars overhead. Walked back into the FAA Weather Office on the field and asked what changed in the forecast. Nothing, but there is a line of clouds off shore moving in, standard offshore fogbank. Oh.. Ok. Just over 50 miles to get home. Took off, about 15miles from landing, the city lights started disappearing, hmm... that cloud bank is inland. Descended to TPA and below the clouds. Had snow in the landing light on downwind, good visibility 10 miles down the river. Landed and put the plane away. Next morning had 3 inches of unforecast snow on the ground.

Another time, middle of May, no storms forecast for 3 days, Clear Blue! Flew 120 miles home for Mother's Day, figured to surprise our moms and fly back after sunset. About 10 miles from landing and talking to approach control, Delta arriving from the NW reported snow about 40 miles north. What? Landed uneventfully and tied down. By the time I drove 3 miles to get home it was snowing. Woke in the morning to 8in of fresh snow.
 
BillTIZ said:
Walked back into the FAA Weather Office on the field and asked what changed in the forecast. Nothing,
Click to expand...
I remember years ago taking some friends to Nantucket one Sunday. Good VFR forecasted all day and *some* clouds coming in the evening. When I walked out of the rest. after lunch I looked around and the Wx just didn't look right, even though it can be deceiving on an island. Called FSS and indeed the ceilings were coming down. Rushed to the airport and in that time it was starting to look bad, so called again and indeed it had really dropped all along the coastline from BOS to DC.

Ended up finally getting on a ferry to the mainland and then had to wait for a 3rd train to get back to NYC. ALL of the traffic had been shut down do to low fog and everyone trying to fly out of BOS was now on the train along with all the other weekend commuters trying to get home. I think we finally rolled in around 3:30 in the morning.
 
It was around the time we moved to 5G when weather forecasting got really bad. I don’t know if they were correlated.
 
JOhnH said:
Funny thing is, I don't recall anyone talking about how good the weather forecasts were before now. What changed to make them seem better back then?
It reminds me of the saying "The older I get the better I was".
Click to expand...
I've probably posted this before. I'd be curious to see what it looks like in another 10 or 15 years.

Until then, maybe some loitering HALE drones pooping out dropsondes twice a day?
170707-modelsForecast.jpg
 
scottd said:
Most radiosondes use hydrogen-filled balloons. July 1, 2024, the National Weather Service (NWS) indefinitely suspended radiosonde observations at Tallahassee, Florida, due to a global supply chain disruption of Helium, and the prohibitive cost of the Helium gas contracts. The NWS launches radiosondes from 100 Upper Air sites throughout the United States and the Caribbean, using Helium at 12 of these sites. Hydrogen is the preferred gas because it is cost-effective and a more reliable gas option. The NWS is converting many of these sites to Hydrogen as soon as practical and as funding becomes available. However, safety concerns prevent Hydrogen from being used as an option at the current Tallahassee Upper Air facility. Still today, no balloons are being launched at Tallahassee.

When I worked for the NWS many years ago, I worked in the modeling branch that developed many of the models we use today. I can tell you that losing radiosonde observations (RAOBs) is a huge issue for forecast accuracy. Of course, this affects all numerical weather prediction models, not just those from NOAA. Radiosondes are essential because they provide vital data from the upper atmosphere, that are ingested into weather prediction models that provide guidance to meteorologists that issue severe weather, aviation and marine forecasts, and “ground truth” for satellite data validation. The validation of space-based observations is quite important since satellite data is one of the ways numerical weather prediction models get weather information over data-sparse areas such as the vast oceans around the world. If the satellite wind data is not consistent with radiosonde observations of wind, ingesting them into the model can create significant forecast errors that impact weather guidance used by pilots. Therefore, meteorologists use radiosonde data to calibrate satellite instruments and correct for any systematic errors (biases) in the satellite data.
Click to expand...
Thanks, Scott.
 
scottd said:
Most radiosondes use hydrogen-filled balloons. July 1, 2024, the National Weather Service (NWS) indefinitely suspended radiosonde observations at Tallahassee, Florida, due to a global supply chain disruption of Helium, and the prohibitive cost of the Helium gas contracts. The NWS launches radiosondes from 100 Upper Air sites throughout the United States and the Caribbean, using Helium at 12 of these sites. Hydrogen is the preferred gas because it is cost-effective and a more reliable gas option.
Click to expand...
About 25 years ago, some of the participants in the America's Challenge and Gordon Bennett Cup gas balloon races began to convert to hydrogen from helium as shortages increased the cost of helium. The change to 100% hydrogen usage was complete a few years later.

This thread recalled my memories of the several helium tube trailers gathered around the Double Eagle II on the ground in Presque Isle, Maine in 1978 as the envelope filled. As the sun neared the western horizon and nightfall approached, the envelope began to take shape and lift from its spread out state across a huge field.

It was the beginning of a great adventure.
 
Last edited:
I realize hydrogen is flammable, but what are the safety concerns with small (like birthday) balloons?
 
scottd said:
Helium is the only element on the planet that is a completely nonrenewable resource.
Click to expand...
At the risk of sounding argumentative, is that true? Alpha decay produces helium. A lot (most? Almost all?) of our planet’s original helium was blown away by solar winds before the magnetosphere formed, I think. Someone much smarter than I can probably clarify the story of helium but I have the sense a good amount of helium we access today is from radiation.

Not at all saying it should be wasted; it is pretty rare
 
I’ve learned a lot more about helium, thank you all.

My comment was strictly an economic one, how low supply and high demand is increasing the cost and that whoever is willing to pay more/most gets the supply.

In the analytical chemistry realm, helium can be replaced with hydrogen if cost becomes a big enough issue.
 
scottd said:
Most radiosondes use hydrogen-filled balloons. July 1, 2024, the National Weather Service (NWS) indefinitely suspended radiosonde observations at Tallahassee, Florida, due to a global supply chain disruption of Helium, and the prohibitive cost of the Helium gas contracts. The NWS launches radiosondes from 100 Upper Air sites throughout the United States and the Caribbean, using Helium at 12 of these sites. Hydrogen is the preferred gas because it is cost-effective and a more reliable gas option. The NWS is converting many of these sites to Hydrogen as soon as practical and as funding becomes available. However, safety concerns prevent Hydrogen from being used as an option at the current Tallahassee Upper Air facility. Still today, no balloons are being launched at Tallahassee.

When I worked for the NWS many years ago, I worked in the modeling branch that developed many of the models we use today. I can tell you that losing radiosonde observations (RAOBs) is a huge issue for forecast accuracy. Of course, this affects all numerical weather prediction models, not just those from NOAA. Radiosondes are essential because they provide vital data from the upper atmosphere, that are ingested into weather prediction models that provide guidance to meteorologists that issue severe weather, aviation and marine forecasts, and “ground truth” for satellite data validation. The validation of space-based observations is quite important since satellite data is one of the ways numerical weather prediction models get weather information over data-sparse areas such as the vast oceans around the world. If the satellite wind data is not consistent with radiosonde observations of wind, ingesting them into the model can create significant forecast errors that impact weather guidance used by pilots. Therefore, meteorologists use radiosonde data to calibrate satellite instruments and correct for any systematic errors (biases) in the satellite data.
Click to expand...
Scott…. Curious if the modeling methods are making use of Kalman filtering of the various data sources?
 
Horrible forecasting recently. Thursday went from 30 % chance of showers to 100 % in the same day. Got one of our aircraft stuck because the storms blocked his way back to base. Then fog forms over night and the whole AO is IFR. None of that was forecasted. Makes the job more interesting though.
 
scottd said:
If we look at Chattanooga, keep in mind that this area was on the southern end of some convection well ahead of a cold front early in the day on Thursday. The heaviest precip was north into eastern KY. This was just the first "wave" of convection expected to move through.

View attachment 139936

This thin and weak line moved through around between 1545Z and 1630Z.

View attachment 139937

That line of convection weakened and moved well east of the area. It wasn't until about 20-21Z that the next round of convection began to move into the terminal area.

View attachment 139938

If we look at Chattanooga, the first TAF for KCHA (00Z routine issuance) issued for the Thursday (issued Wednesday evening) shows a 30% probability of thunderstorms and light rain from 16Z to 19Z.

KCHA 092334Z 1000/1024 18003KT P6SM SCT250
FM101500 22007KT P6SM BKN050
PROB30 1016/1019 4SM -TSRA BR OVC030CB=

If you read the area forecast discussions, the weather forecast office out of Knoxville/Tri Cities, suggested...

"CAMS are in agreement of some late morning/early afternoon convection across the northern and central parts of East TN and SW VA, but the severe potential comes in the mid to late afternoon in southern sections where surface heating will have greater effect to aid destabilization, and deep shear will be more favorable for a supercellular convective mode."

CAMS means convective allowing models. So this was written Wednesday, mid-afternoon and pretty much summed up what happened. Basically two waves of convection possible occurring on Thursday.

The next issuance (06Z routine issuance) on Thursday predicted showers in the vicinity (a forecast for convection) beginning at 14Z with a 30% probability of thunderstorms and light rain from 15Z to 21Z in the terminal area. Then at 21Z, they forecast thunderstorms in the vicinity beginning at 21Z with a 30% probability of thunderstorms and moderate rain in the terminal area from 21Z to 03Z.

KCHA 100520Z 1006/1106 19006KT P6SM FEW250
FM101400 19007KT P6SM VCSH BKN050
PROB30 1015/1021 20007KT P6SM -TSRA BKN050CB
FM102100 23005KT P6SM VCTS BKN050CB
PROB30 1021/1103 3SM TSRA BKN040CB
FM110300 24002KT P6SM -SHRA VCTS BKN060CB=

The next issuance (12Z routine issuance) on Thursday predicted showers in the vicinity (a forecast for convection) starting at 12Z, then thunderstorms in the vicinity beginning at 21Z with a 30% probability of thunderstorms and moderate rain in the terminal area from 21Z to 03Z.

KCHA 101127Z 1012/1112 18006KT P6SM VCSH SCT120 BKN200
FM102100 26007KT P6SM VCTS SCT080CB
PROB30 1021/1103 3SM TSRA BKN040CB
FM110700 23002KT P6SM SCT025=

Below are the surface observations from 12Z to 00Z. Essentially for KCHA, the TAFs were a little overzealous bringing in the convection early into the terminal area or its vicinity. There was that thin line of weak convection (especially weak to the south). By the time it hit the CHA terminal area, it was very weak. Nothing really transpired from a serious convective standpoint in the terminal area until 21Z where supercellular convection developed just ahead of a cold front.

KCHA 101153Z 14003KT 10SM FEW060 SCT140 BKN220 10/06 A3011 RMK AO2 SLP193 VIRGA OHD-SW-N T01000061 10111 20089 56007=
KCHA 101253Z 20003KT 10SM BKN080 BKN140 BKN200 12/06 A3012 RMK AO2 SLP196 T01170061=
KCHA 101353Z 18008KT 10SM BKN080 BKN150 13/04 A3012 RMK AO2 SLP196 T01330039=
KCHA 101453Z 17008KT 10SM SCT070 BKN110 BKN150 15/04 A3010 RMK AO2 SLP190 T01500039 58003=
KCHA 101553Z 16008KT 10SM SCT065 SCT120 BKN250 17/04 A3010 RMK AO2 SLP188 T01720039=
KCHA 101653Z 11006KT 10SM FEW035 SCT085 BKN120 BKN250 18/08 A3008 RMK AO2 RAB03E30 SLP182 P0000 T01830078=
KCHA 101753Z 23006KT 10SM FEW045 SCT085 SCT120 BKN250 21/07 A3004 RMK AO2 SLP170 60000 T02110072 10211 20100 58019=
KCHA 101853Z 23009KT 10SM FEW050 SCT100 SCT250 23/08 A3002 RMK AO2 SLP162 T02330078=
KCHA 101953Z 28012G23KT 10SM SCT065 SCT110 SCT250 24/03 A2998 RMK AO2 SLP148 TCU DSNT N T02440028=
KCHA 102053Z 27007KT 10SM SCT060 BKN110 BKN140 BKN250 23/03 A2995 RMK AO2 SLP139 CB DSNT NE AND W T02280028 58030=
KCHA 102117Z 25010G29KT 10SM -TSRA BKN065 BKN095 OVC250 21/04 A2996 RMK AO2 PK WND 30035/2106 RAB08 TSE16B17 CB DSNT SW-NW P0000 T02060039=
KCHA 102119Z 24010G23KT 10SM TS BKN065CB BKN095 BKN250 21/04 A2996 RMK AO2 PK WND 30035/2106 RAB08E19 TSE16B17 OCNL LTGICCC S-SW TS S-SW MOV SE P0000 T02060039=
KCHA 102153Z 21016KT 10SM -TSRA SCT065CB SCT080 BKN100 BKN250 17/08 A2997 RMK AO2 PK WND 30035/2106 SLP146 RAB08E19B43 TSB16 OCNL LTGIC SE-S TS SE-S MOV SE P0000 T01720083=
KCHA 102209Z 20011G25KT 10SM SCT050 SCT095 BKN120 BKN250 19/06 A2996 RMK AO2 RAE09 TSE09 TS MOV SE CB DSNT SE P0000 T01890061=
KCHA 102253Z 31003KT 10SM FEW055 SCT095 SCT110 BKN250 19/06 A2996 RMK AO2 RAE09 TSE09 SLP141 CB DSNT SE-S P0000 T01940056=
KCHA 102353Z 23010G21KT 10SM -RA SCT046 BKN075 BKN090 BKN250 17/08 A2994 RMK AO2 RAB35 SLP136 P0006 60006 T01720078 10250 20172 58004=
KCHA 110005Z 20008KT 10SM SCT046 SCT075 BKN090 BKN140 BKN250 16/10 A2995 RMK AO2 WSHFT 2345 RAE2356 P0000 T01610100=

KTYS TAFs were very similar (they are also issued by the Knoxville weather forecast office) with light rain moving into the terminal area around 16Z and more substantial convection moving into the area around 21-22Z.

TAF for Asheville below is issued by the Greenville-Spartanburg weather forecast office and issued a 30% probability for light rain showers with MVFR conditions in the terminal area (a convective forecast) from 17Z to 23Z.

KAVL 092330Z 1000/1024 14011KT P6SM SCT250
FM101200 17007KT P6SM BKN050
FM101600 18010G18KT P6SM OVC040
PROB30 1017/1023 4SM -SHRA BR OVC025=

Once it became a little more clear and the 00Z guidance from the CAMS was updated, the forecasters for the 06Z routine issuance added a 30% probability of thunderstorms and moderate rain from 17Z to 23Z and then showers in the vicinity followed by light rain showers after that.

KAVL 100532Z 1006/1106 17003KT P6SM SCT250
FM101400 18006KT P6SM SCT020 OVC060
PROB30 1017/1023 4SM TSRA BR OVC020CB
FM102300 26005KT P6SM VCSH BKN070
FM110400 00000KT 6SM -SHRA BR BKN050=

It's worth noting that the NWS TAFs only provide a forecast for 30% probability. You won't see a PROB40, PROB50, etc. For FM or TEMPO groups, the probability increases to 50% or greater for it to impact the terminal area. The terminal area is the region within 5 statute miles from the center of the airport's runway complex. That's essentially a point forecast. So they should never be used as a zone or area forecast.
Click to expand...
Yeah, I get it. There are challenges in forecasting these events. The conditions were changing rapidly, hence the TAF updates and SPECIs all over the place.

Simply saying when I briefed my crew in the morning (0700) on the chances of storms the prediction didn’t match the reality. Just like when my night relief briefed his crew (1900) the prediction for fog or lack there of, didn’t match the reality of IFR over night and into the morning.

It’s not a big deal. We get paid to adapt to the conditions. Still doesn’t change the fact we had an aircraft stuck in KGVL for 6 hrs because of extensive storms that the pilot didn’t foresee.
 
scottd said:
As I chat with more and more of my colleagues who work at the various local weather forecast offices, there have been impacts to the guidance they produce. Most have been keeping up, but they are starting to cut back on forecast discussion details and when severe weather occurs, TAF amendments may suffer since they don't have the staff to work the radar desk, send out warnings and keep the TAFs aligned with the latest weather and forecasts.
Click to expand...
I cant imagine it’s going to be easy to quantify the outcome. I hope what’s happening is being well documented for future study.
 
Doesn’t automatic reporting by airlines via airinc do much of what the balloons did?
 
scottd said:
Most of the data is at FL390. There is data around major airports, but it’s shaped like a salad bowl.
Click to expand...

Shame they didn’t incorporate that into ADSB, of course that’s not what ADSB tracking was about.
 
You must log in or register to reply here.
Back
Top