Rainfall figures are very different between 3 hourly and hourly displays
-
For the same day and model ECMWF on the browser app:
For 3-hourly there's a total of 1.5mm rain expected throughout the daylight time
For 1-hourly there's a total of 0.7mm rain expected and just around lunchtime
Why are these so different?
Shouldn't they mostly agree except for rounding issues?
If it's an issue of resolution, could the hourly use 0.1mm levels? -
@chis
These differences are only due to the fact that Windy does not display values below below 0.1mm.
Note that 0.1 mm resolution is a WMO standard for rain gauges. Rainfall traces of 0.1mm are almost nothing, so in a forecast what does it matter?
May be values below 0.1 mm should not be considered in the 3-hour cumulative to show both panels consistent. -
@idefix37 0.1mm seems worth including, then I wonder whether you could have a smoothing algorithm that spreads the rain across the same duration and achieving the same total, just that some periods with less than 0.1mm wouldn't get a value and others would, which would appear as sporadic rain which is really what it's like in these situations anyway.
Having 1.5mm and 0.7mm makes a bit of a difference if you're in the outdoors. I'd always rather know the total expected and over what duration.
-
@chis said in Rainfall figures are very different between 3 hourly and hourly displays:
@idefix37 0.1mm seems worth including, then I wonder whether you could have a smoothing algorithm that spreads the rain across the same duration and achieving the same total, just that some periods with less than 0.1mm wouldn't get a value and others would, which would appear as sporadic rain which is really what it's like in these situations anyway.
I don’t understand. You were complaining about not seeing values below 0.1 mm of rain for one hour, while you see more than 0.1 mm for three hours. Am I right ?
Having 1.5mm and 0.7mm makes a bit of a difference if you're in the outdoors. I'd always rather know the total expected and over what duration.
Yes but here we are talking about less than 0.1mm over 1 hour.
-
@idefix37 I'd be happy if the 1hr totals were close to the 3hr ones. At present they can be wildly different so I need to switch to the 3hr version so see the rainfall then switch back to the 1hr to get temp and wind by hour.
It's not an issue for me if the 3hr value was 0.2m and the 1hr equivalent was 0.1, 0, 0.1 or even if 0.1 isn't available then 0, 0.2, 0
In the above:
0401-0700, 3hr shows 0.4mm and the 1hr 0mm => be good if 1hr shows total of 0.4mm, such as 0.1, 0.1,0.2 or 0,0.2,0.2 or 0,0,0.4
0701-1000, 3hr shows 0.3mm and 1hr 0mm => be good if 1hr showed total of 0.3mm
1001-1300, 3hr shows 0,.6mm and 1hr 0.7mm => 0.1 difference doesn't matter
1301-1600, 3hr shows 0.2mm and 1hr 0mm => be good if 1hr shows total of 0.2mm -
@chis
I do not agree with you.
You are proposing to add wrong values just to show an infinitesimal amount of rain spread over one hour !
If you get 0.2mm for 3h, it may be the addition of 0.05 + 0.04 + 0.07 = 0.16 rounded to 0.2.
If Windy introduce wrong values like in your proposal, many users would complain that the total of the 1h-values is not equal to the 3h-value.In addition when a rain amount is displayed in the 3h-table, there is a “rain” pictogram and this pictogram is kept in the 1h-table even without displayed rain amount. So check pictograms too.
-
1hr values that are close as reasonable to the 3hr totals is what I've been talking about.
I agree that for the 22nd up to 1600 the amount of rain is very low and you certainly shouldn't aim for numbers that disagree. I still think having one or more values in the 1hr that total to the same as the 3hr would be helpful. Needing to scan the pictograph as well as the numbers prevents a quick scan to see if any non-trivial rain is likely.
Other than that I go back to the images on my first post. The two versions of the time show quite different amounts of rain and there's nothing on the 1hr version that indicates there could be some between 0400 and 1000
-
I idefix37 referenced this topic