Augment predictions with realtime data
Can predictions be updated with realtime information? Or at least let us know when the prediction was made. For example, right now 6/7/2020, I am looking at Windy data for a marine location near New London, CT. There happens to be an NOAA weather station here with wind data as well, so I can compare the Windy data to what is happening right now. So Windy says right now at 9:46AM EDT near the New London Ledge Light weather station that the wind speed is between 5Knts and 13 knts depending on the prediction model used. However the New London Ledge Light station shows the wind speed at 18knts at 9:25am (last reported) with gusts at 22knts.
Moreover if you look at the general prediction for this area today in Windy, it looks like very light air. The 1-hour increment trend for the day shows no model above 10knts at any time of day (ignoring gusts for now), which is weird cause if I just click the general area I did get that 13knt data.
So 10knts and under is pretty different than the 18knts and above we are seeing from NOAA data. Graph of data (this will show current data not when this post was made) https://www.ndbc.noaa.gov/show_plot.php?station=ldlc3&meas=wdpr&uom=E&time_diff=-4&time_label=EDT
How can windy adjust its models based on more realtime information?
Can predictions be updated with realtime information? Or at least let us know when the prediction was made.
Click the small clock in the low right corner of your screen to see when the model was updated.
Obviously Windy CANNOT adjust its models with more realtime information. These models are provided by meteorological centers including the NOAA which update their models themselves, GFS and NAM for the NOAA. Windy visualize only the data that the NOAA and others offices have predicted.
The weather station that you mention is at the top of a lighthouse, at 20m instead of 10m as usual. Not so surprising that it shows a wind slightly stronger.
On Windy if you select Reported wind, you get the observed wind speed near your location from weather stations (You can forget MADIS weather stations i.e. Personal weather station, which are not so good for wind measurements specially in town)
I suggest that you compare forecast and observation at Block Island Airport. That seems not too bad for ECMWF showing 10kt for 12kt reported. And you can choose another model.
Thanks for the information @idefix37 , the model time information is very helpful as well as the reported wind feature. Didn't know about either of these.
Yes I agree the 20m height for LDLC3 lighthouse can account for some higher wind than Windy. However, looking back over that day I still don't think it accounts for all the differences I saw (not that I expect predictions to be exact). Would the extra 10m account for a doubling of the wind speed? I suppose it can, you might know better than me.
While Windy may not use realtime information today to update the models, someone is going to do this. It would be great to start some experimentation on this. It would give you huge advantage if it panned out. But it would definitely be an investment to hire some people with that expertise.
Yes you are right 20m high instead of 10m do not make a so big difference for wind speed measurement.
The idea of using real-time reported data would be a tremendous challenge. Today global weather models update 4 times a day, and for these updates they are fed with huge quantity of observations throughout the world from weather stations, planes, ships and mainly satellites.
With very limited-area models it is much more possible, it’s called now casting.
And I would say that this is really the role of the model providers to add any additional data inputs, rather than that of Windy.
Because Windy does not create any forecast data but instead only visualises forecast and actual data received from various third party providers.