I seen the screen shots that’s why I was looking into it to see if I could replicate to possibly help narrow it down but not in Europe so that doesn’t help you.
I appreciate the effort.
You can add a new place and just tap somewhere in the are on the map below, asumin screenshot works now , then (assuming that you are have a gold prescription) you’ll find HIRLAM in the list of data sources.
Peace and chears!
@Ohan I can add but I already have a lot of info being downloaded lol but I guess I could if it will help narrow things down
@Ohan ok added Westminster England loaded HIRLAM data no blue/purple spot and yes single hindcast day shows in graphs only thing other than a glitch is the data is unavailable.
The blue an purpul “flash” is only just at the beginning of the 1 day hindcast data, and only if more than 1 day hindcast is set. It’s only “one frame each” so one must swipe verry slowly to see it.
@Ohan thanks for added info the flash happens at the point where data ends on hindcast
It used to be several flashes and it used to be several fake bumps in the graph for rain in the hindcast of the HIRLAM data. See the screenshot below, in the HIRLAM graph there are two small spikes in dark blue, claiming that it rained, when no other data said rain.
And those small spikes, was always present in the HIRLAM hindcast, two of them of displaying earlier during current day, and several more with exactly same shape, hight and distance between them if displaying an additionall day of hindcast.
And they where always only in hindcast, never in the forecast. It seamd as if the forecast was good, and the hindcast got corrupted by the same pattern all the time.
Maybe @duane chose to hide the corrupted data until the issue could be solved?
I’m not sure exactly when it changed, but I noticed the new behavior first time when radar was available.
@BrianLY-38 se I’ve added screenshots and more text to previous post.
I just checked and all graphs show a chance of perception for Westminster except CMC GDPS. I’ve had it show precipitation and not rain because it’s showing the radar return there is moisture in the air but doesn’t mean it’s going to rain.
Yes, I’ve fixed the HIRLAM flash problem. This is fixed on the server, not in the app, so it’ll roll out to all versions of Flowx including past version.
HRRR Smoke data is nearly all fixed too.
@duane Thanks for the Update looking forward to HRRR again, as Always Keep up the Great Work
@duane I’d hate to see the amount of data on your server
I think I process 50-100GB per day.
I think this is an instance of this happening before (1th July 2019):
And this is an earlier instance of a similar problem (30th May 2019):
I think that in May the “fake rain” was reacuring every 6 h in the HIRLAM hindcast (see graph above), and this would then correspond to the update rate of HIRLAM.
@Ohan, I was wondering if anyone would remember. In this case, the HIRLAM forecast would release data for 00hr, 03hr, 06hr… for all data except precipitation which would start from 03hr, 06hr… My script would copy the prediction for 00hr from the last forecast but the copy process had a bug and it didn’t copy, which is why you saw the spike every 6 hours, the time between forecasts.
Since then, I think this is the first case, but this is a whole forecast missing.
I know how to add checks and overcome this but the bug doesn’t happen often enough to warrant attention. I think the radar and widget refresh bugs are higher priority. Oh, the the growth bug, I need to fix the growth bug. It a catch-22 if I get growth I can afford help, but if there are bugs instead of marketing, growth is low.
Yeah, sorry about that. I broke the processing for HIRLAM but it’s fixed now. Unfortunately, it’ll take 3 days to recover the hindcast (fill in the blank days)
Good that you are aware of it, and that it is fixed now!