Release 3.220 (6 Dec 2019)
-
16-days of GFS (FV3) weather forecast in all pro versions.
-
Is 16-days forecast useful? Check out my new blog (projectink.ink) which talks about the value of this data.
3 Likes
One one more point for your project Inc is that a 16 day forecast gives you another 6 days of Trends. Trends can be very important and very useful.
1 Like
Is a 16-day forecast useful? I dunno.
If you had asked if it were interesting, that’s a much easier question to answer. More data is always more interesting. 
Here’s my local forecast going out 7 days:
So in 7 days, the models diverge and the temperature will be someplace between -10F and 30F. That range seems to be accurate. 
The wind speeds and direction, similarly diverges:
It appears that the wind will be blowing in some direction in 7 days, and it will be between 0 and 30 mph.
(These are from Meteoblue: https://www.meteoblue.com/en/weather/forecast/multimodel)
That said, more data is a good problem to have. Generally, the forecast for the pressure is more consistent between models. Seeing that can help determine how much weight you give the forecast.
Thanks for the continued updates!!
1 Like
There is no doubt that the accuracy of wind, temperature, etc get worse with time especially with a low pressure system in the region.
But if you were to consider pressure systems, that is definitely useful. High pressure systems are pretty stable and predictable. So if there was a high pressure system in the region, then you can be more confident of the forecast. This is extremely valuable information.
On the other hand, low pressure systems are unstable and difficult to predict. Just look at the spaghetti lines for hurricane tracks. If you have a low pressure system in the region, then you know you have less reliable forecasts. This too is valuable information. For example, if you planned an event and there was a low pressure system one weekend and a high pressure system the next, you can delay the event to the weekend with the high pressure system.
I wonder if this fixation on accuracy has blinded us to the actual value and usefulness of forecast data. Why would NOAA spend CPU cycles to solve 16-days of weather forecast if it’s not useful??
3 Likes
Incidentally, with the wind direction graph, you only have to a low pressure system pass south of a point, instead of north, and you’ll get the opposite direction for wind. So it’s understandable that you can get wildly varying values.
This is why it’s important to understand the path of low pressure systems instead of just looking at graphs. Watching the low pressure system and understanding where it can possibly go will indicate the reliability of forecasts. Or put another way, the possibility of different forecasts. This is where ensembles models come in to play.
Many people think forecasts are absolute when it’s in fact the most likely outcome out of many possibilities. I hope to write more about this in the future.
3 Likes
I like the 16 day long view ability , even with the inevitable fog of war that far out.
A 16 day forecast does highlight the need for another view option.
I have four suggestions that would help the way I use flow:
-
Provide a full screen view of the graph(s). One without a map, that goes full width in landscape orientation.
-
Provide a shortcut to change look ahead. I use a 3 day widget graph, 6 day for home, and 10 for the town I work in. This sort of provides me a views to: plan my day, see decent forecasts, and hints at what may come. I like the idea of 16 days, as that is how I currently use 10. However I can’t even read 10 without glasses, so I revert to 6.
-
Scrolling would be another option. Have a preferance for days visible and then allow past and future to scroll.
-
Indicate weeks (maybe by weekend days) on the graph.
If adding the look ahead shortcut involved changing the top line I would love to see more obvious separation of location and data source. Instinctively I first click on the location name to change it. Gets me every time, so always 4 clicks to get to my 10 day work location view, and 4 more to return to my default home 6 day view.
Given this long middle of the night rant you would think I don’t like Flowx. Truth is that I love what it does. The only true deficit it has was no current overlay, which rain viewer addressed. Thanks for the hard work and stability of this app, it is appreciated.
1 Like