Error in cloud depiction for SKMD Meteogram in early morning
-
For the past several weeks I have noticed a bug that was not present before.
If I look at the meteogram forecast for SKMD(UTC -5)the night before(9:03pm local)it shows the information as expected and usual.
As can be observed the darknesss dissipates shortly after sunrise close to 8am as expected.The problem arises when the information is updated in the middle of the night, where now the darkness delays to close to mid day which is not realistic or the case.
At this time (5:45am local) the day es quite clear as seen below
When the information was updated about 7:45am local time the shown information reflected was the expected and correct again with the darkness dissipating close to 8am local.
At 7:45 local time this is the picture, nothing close to the middle of the night forecast.
As a summary when the information is updated in the middle of the night, close to 3am local it always shows a darkness forecast in the Meteogram for SKMD that goes all the way to 12pm(mid day) which is never the case. The previous update the night before around 9pm local or the same day forecast in the mornings around 8am shows the correct information once again. There is a big error forecast for the information of midnight.
Hoping for a prompt solution. Thanks !!!
-
@jadiez
If you see this cloud near the ground late in the morning, it means the ECMWF model forecasts are not providing the correct Relative Humidity (the clouds in Meteogram are based on this parameter). Try another model, GFS or ICON.
But Windy cannot modify forecast data from weather models. It would be useful that you give the reference time of the model (00z, 06z, 12z or 18z) if there is a link between this type of forecast and the model reference time as you say.
-
@idefix37
Thanks for your comments and suggestions.
The Cloud near the ground in a dark bar like form usually shows up until about two hours after sunrise in the ECMWF model when it is correct. This was the case for all model reference times in the past. It is still the case for reference time 06Z, 12Z and 18Z. Just in the recent weeks the ECMWF model is showing an error (bug) for the 00Z reference time, where the dark bar extends all the way to about 6 hours after sunrise. After your suggestion I checked models GFS and ICON; they show the correct information even at 00Z.
I like the model ECMWF better, because it seems to be more accurate. In addition since I am a pilot and when I look for airfields(aerodromes) the default model is the ECMWF. This is why I mentioned SKMD. How can we get windy to look in to this issue, so they can correct the forecast for the 00Z reference time in the ECMWF. -
@jadiez
Thank you for your clear information.
It's really strange that the ECMWF 00z run has such a bias every day. I don't know if anyone at Windy can look into this. I'll try to check it out in the next few days. -
@idefix37
Thanks for your help!! -
Here are some screenshots from ECMWF for last Sunday forecast, all for Olaya Herrera airport.
And for today, Monday
I’ll try to catch a complete cycle of 4 runs.
You may find some explanation in this document. See page 15
It is about 2m humidity but the low stratus over this location may show the same bias. -
@idefix37
Thanks for your explanation. I read page 15 of the document and eventhough it is above my knowledge, it is clear to me that the 2m bias issue is geographic, time and season related and applies for the tropic throught the year where SKMD is located.
What astonishes me is that this Bug at 00z as I call it, is new and only for the past weeks as I mentioned before. In the previous 2 years I never saw this condition happening for the 00z forecast in the ECMWF. Is there a possibility of there being something additional causing this condition? -
@jadiez said in Error in cloud depiction for SKMD Meteogram in early morning:
@idefix37
this Bug at 00z as I call it, is new and only for the past weeks as I mentioned beforeIt is not linked to a major change in the model. IFS cycle 49r1 was launched last autumn and the future cycle 50r1 will be launched next autumn.
I checked these forecast during a complete update series.
The 00z shows that the low stratus is staying up to midday. But 18z is worse confirming my last screenshot !
-
Thanks for your feedback. I am traveling with limited access to internet. You are right in that the 18z is even more weird. Since I check with the 00z and 06z information I had not notice the 18z biased.
When you say the r50 will be updated next autumn, do you refer to sometime between September to November 2025? Will this issue be correct? -
I was referring to IFS cycle 50r1.
-
@jadiez
I don’t know precisely. Just an information I found in ECMWF news.
