Update: I've recently been using Ventusky to compare NBM and HRRR results, and there are divergences and I think it does add something. I posted some screenshots but I think the board rejected their file size.
Posts made by danfarina
-
RE: Model Request: ConUS NOAA NBM Model
-
RE: Model Request: ConUS NOAA NBM Model
@idefix37 Sort of...but the fact that NBM is principally HRRR, and updated every hour at a high resolution with a brief forecast depth, satisfies a very different use case. The closest thing to a duplicate is HRRR itself.
The main thing that could be said to be the same about Meteoblue and NBM is they are ensembles, but not much more.
-
Model Request: ConUS NOAA NBM Model
I was clicking around on the info tab on models at Ventusky.com, and saw they had an interesting model listed: "NBM," which I looked up and is written up here:
https://theweatherguy.net/blog/basic-info-about-the-nbm-and-icon-models/
and:
https://vlab.noaa.gov/web/mdl/nbm
It can be considered an ensemble derivative of HRRR that includes inputs from a pretty extensive list of weather models from all sorts of organizations to tweak the output. The most common thing I'd be doing is comparing the two looking for divergences to assess model sensitivity, with one model as my default...probably NBM, assuming it lives up to its promise of improving on HRRR accuracy.
-
RE: We are launching Windy Premium - even better forecast for just price of a lunch
I've been a premium subscriber for a while, even when the benefits were marginal to me, and I like Windy a lot. But, watch your premium segmentation with care, ventusky.com is free and addresses a number of models without ads [it appears to be something of a side project?].
Once people are more alert about it, segmenting certain features into Premium may be a bit pushing on a string. For what it's worth, the availability of hourly forecasting data with HRRR/NBM is why I suggest it to people not invested enough to pay for Windy to use Ventusky.
So here's what keeps me paying for Windy without complaint:
- Higher resolution ECMWF
- The denser presentation of forecast data across hours and days
- More obscure map layers, like soil moisture
Ventusky could fix #2, in many ways the mobile application and web site are a bit more polished than Windy, but perhaps some of the polish makes it difficult to make dense and functional presentation.
They have a number of layers, so I think they could fix #3 in all practical respects for widening use cases in time.
But with their side-project-ish (non) revenue, it seems unlikely they could approach #1, assuming high resolution ECMWF costs a decent amount of money.
While I don't mean to suggest that Windy should be looking to "compete" with Ventusky for market share or anything like that, the choice of what to put behind premium and advertise as reasons to buy premium exists in equilibrium with other things available, and I guess I'm affectionate enough about Windy that I wanted to give my impression of reasons why I feel satisfied in continuing to pay for Premium, as I have for some time. Consider it a bit of unsolicited user or market research.
-
RE: Would like to see the HRRR model added
@filip_k Picked up another one just now:
-
RE: Increase web app zoom level where weather overlays are displayed
@korina I suppose I simply disagree: it's disruptive to be aware to not zoom the map to look at San Francisco or even the Bay Area (of which Brentwood probably not a part) without losing track of what the contours of the weather forecasting model looks like and squeeze me into 8cm * 8cm of the screen.
-
RE: Increase web app zoom level where weather overlays are displayed
@korina The problem is the user interface: I'm stuck dragging the weather picker around a couple of inches on my screen, being very conscious at all times to not look at San Francisco the normal way. Weather resolution does still matter: this is less important for 13KM models where there's not much point looking at the north-east city or the south-west city beyond generalities.
Besides that, of what utility is the street map at that resolution? Doesn't seem to have much comparative advantage.
And finally: mobile gets it right, in terms of not squeezing my use into a tiny sliver of the screen. Are you saying you should make it a street map at this zoom level?
-
RE: Increase web app zoom level where weather overlays are displayed
@idefix37 It's easier to orient on and move the weather picker around the map. It's disruptive to lose the gradient when looking at the most common view of the city of San Francisco, I have to focus a great deal on looking at the Bay Area a certain way. The reasoning you give cuts both ways, since there's little point in having street-by-street information about the weather and dragging the picker around.
Like, for what is a weather mapping application, what's the point of showing a street map without weather gradient here, so you can maintain your orientation about where the interfaces (e.g. in cloudcover) are in the model?
To show how this feels in the browser viewport, I have scaled these images so they can fit the forums.
Natural view where it's reasonable to want gradients maintained:
The highest zoom level I can use with gradients, and I must be very conscious at all times to never zoom:
-
RE: Would like to see the HRRR model added
@korina Alright, well, I guess that's all the encouragement I needed, I'll post another one if I see it. Let me know if it's just noise and you have monitoring on data delay and just need time to fix it.
-
Increase web app zoom level where weather overlays are displayed
This is the maximum zoom level before the Windy web application switches to street maps. I find myself often squinting at this to figure out how my activities relate to predicted cloudcover and temperature gradients. The cropped image probably does not reproduce my subjective experience, because my screen is a high-DPI one. It's even more pronounced when applied against the (wonderful!) HRRR forecast availability, as it has such high resolution.I have been working around this since, well, forever, by either moving the weather picker around when using the street map (for simple checks) or using the browser font-increase function (when I'm checking forecasts over time or evaluating a more complex interface), at a lower zoom level, but it's been annoying me and I'm not sure of what use street maps are at such a low zoom level.
Blessedly, this frustration is not replicated on mobile, where gradients stay active at much higher zooms.
In a related consideration regarding zoom levels, I do find when I type in the name of a favorite or feature in the Windy search that it loses my zoom level and zooms out way, way too far. Both issues seem to relate to a design with much lower zoom levels in mind.
-
RE: Would like to see the HRRR model added
Huh here's another one, also active right now. I'll stop reporting these unless you ask, seems like they might trip an automated system at 4h or 9h of delay.
-
RE: Would like to see the HRRR model added
Not sure about that one, but here's a live one active right now: (4:45 PM in America/Los_Angeles)
-
RE: Would like to see the HRRR model added
@tz I've been using it extensively for my daily planning, I've more or less switched to Windy in Safari to be able to use it. One thing I think seems a little odd is I think I saw an update (the only time I took a look) two hours apart? I was under the impression that HRRR is 24 updates a day.
-
RE: Would like to see the HRRR model added
@korina I just saw HRRR show up the page. This is wonderful!
-
Would like to see the HRRR model added
I'm not the first to ask for HRRR, and won't be the last, but seeing as there's no very recent thread I thought I'd start another. More frequent, higher resolution, useful for places with steep weather gradients like the American West Coast where driving fifteen minutes is consequential.
Thanks for Windy, it's good.