Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

300 meters resolution SF Bay Area Forecast

sf.atmo.ai

Post reply on HN
Print view View original

81–90 of 119 posts

Re: 300 meters resolution SF Bay Area Forecast

#81

Post by milancurcic »

lispisok wrote: post #27

If I had to guess they are running the WRF model [1][2]. The AI part is post-processing the model output. With a fair amount of reading the manual anybody can run their own WRF. WRF scales from running on a laptop to supercomputers with 1000s of cores. [1] https://www.mmm.ucar.edu/models/wrf [2] https://github.com/wrf-model/WRF

Having worked with WRF for 13 years now, contributed patches to many releases, and built a SaaS business centered on WRF (https://cloudrun.co), I still discover new things about it and run into interesting scientific and engineering challenges. It's not the kind of software that's easily picked up and run by a non-expert. It's a large framework that's more niche, more obscure, and not as well documented as something like, say, Tensorflow. There's still a ton of value to derive from making WRF and similar models more usable by non-experts and without access to supercomputers.

Re: 300 meters resolution SF Bay Area Forecast

#82

Post by _yb2s »

hadlock wrote: post #51
Waterluvian wrote: post #12

A few thoughts from a geographer (be prepared to shoot me): - Basically every comment is wowed by this, but nobody questions what the accuracy is. I, too, can Krig interpolated surfaces to any resolution. - off-nadir view doesn't seem to offer much - We've been dealing with janky tile loading for like 20 years now. I really hope we'll get a much smoother approach for viewing these tiles as they load. The dissolve tra…

Few thoughts as a sailor in the SF Bay area. - Accuracy seems at least somewhat correct.That wedge shape you see in the late afternoon sailors call "the wind engine". Local sailing magazine Lattitude 38 has a special PDF that talks about doing a sailing trip around the bay accounting for this local wind phenomenon. Correct stuff: - The SF waterfront, out to the edge of the piers is mostly calm which is correct - Berk…

Also, as a person with experience sailing on the Bay, this model immediately seems unusually accurate. It shows, for example, the Angel Island wind shadow moving around correctly as it does during the day, which I have never seen in another model.

Could anyone with more understanding of meteorology (or OP) please explain what is different about this model vs say the ECMWF model that you can see in apps like Windy, that are supposedly great, but just don't seem to get these features right? Those models are incredibly bad when dealing with the unusual local geographic features on the bay. What resolution are they operating at?

Re: 300 meters resolution SF Bay Area Forecast

#84

Post by macNchz »

barbazoo wrote: post #78
dan-robertson wrote: post #76

Earlier quoted context omitted.

I hadn’t even thought of that until I read this. I had thought thunderstorms after hot weather were just a fact of life. I guess in places near German latitudes that get thunderstorms the hot weather is caused by high pressure systems but maybe that isn’t really the cause in California.

Coastal weather is often lacking the conditions that creates thunderstorms. I live near the coast now too and haven't experienced the kind of thunderstorm I know from Germany.

Depends on the coast I guess...NYC, for example, gets some real whopper summer thunderstorms https://www.youtube.com/watch?v=wk3Gz9o9yw4&t=12s

Re: 300 meters resolution SF Bay Area Forecast

#86

Post by bandyaboot »

The secret sauce to making accurate forecasts at 300 meter resolution has to be something more than “AI” right? I mean, there’s little reason to think AI should be meaningfully better at filling in details at 300m resolution than it would be at picking lottery numbers, no? Am I thinking about this wrong?

Re: 300 meters resolution SF Bay Area Forecast

#87

Post by namibj »

zsz wrote: post #75
johmathe wrote: post #43

Earlier quoted context omitted.

One of the interesting things the model captures at this resolution is the dynamics of the wind going in the bay through the golden gate. See for instance: https://sf.atmo.ai/wind@37.80911,-122.44543,11.68,36,0,16669...

The Global Forecast System (GFS), i.e. the model presently used at NCEP, has a grid resolution of 18 miles (28 km). It is (has been, for years, actually), the second best global forecast system, right behind the European ECMWF (sometimes outperforming it, but on average slightly underperforming it, in terms of accuracy). I don't know how the ECMWF model works, but even as someone who did not study meteorology (but st…

Germany actually has discrete-cloud (modeling radar-detected convective cells as polygons) forecasts covering the next few hours, combined with some further nowcasting aspects into SYMFONY for localized flash flood warnings (+2h precipitation predictions updated every 5 minutes): https://www.dwd.de/EN/research/researchprogramme/sinfony_iaf...

There's the ICON-D2 prediction system with native 2.2km grid, run every 3 hours, with reach of +27h (the 3am UTC run reaches +48h). Also available as an ensemble of 20 possibly futures: https://www.dwd.de/EN/ourservices/nwp_forecast_data/nwp_fore... (open data; feel free to check it out)

Re: 300 meters resolution SF Bay Area Forecast

#88

Post by johmathe »

_yb2s wrote: post #82
hadlock wrote: post #51

Earlier quoted context omitted.

Few thoughts as a sailor in the SF Bay area. - Accuracy seems at least somewhat correct.That wedge shape you see in the late afternoon sailors call "the wind engine". Local sailing magazine Lattitude 38 has a special PDF that talks about doing a sailing trip around the bay accounting for this local wind phenomenon. Correct stuff: - The SF waterfront, out to the edge of the piers is mostly calm which is correct - Berk…

Also, as a person with experience sailing on the Bay, this model immediately seems unusually accurate. It shows, for example, the Angel Island wind shadow moving around correctly as it does during the day, which I have never seen in another model. Could anyone with more understanding of meteorology (or OP) please explain what is different about this model vs say the ECMWF model that you can see in apps like Windy, th…

Thanks for the feedback :) The highest resolution models you can currently find on windy is around 3x3 km (HRRR) and this resolution is unfortunately too big to capture fine terrain and water features. 300x300 gives you 100 times more data points to work with.

Re: 300 meters resolution SF Bay Area Forecast

#89

Post by namibj »

londons_explore wrote: post #58

Weather forecasts are so hard for a user to evaluate... Are you going to check it every day and remember how many days it was right or wrong? Please can weather providers just publish a headline statistic of "Our rain/no rain one day ahead forecast is right 85% of the time. That is better than NOAA (80%), Met Office (72%) and weather.com (65%)."

On part that's what ensembles are about: they run like 20 or so simulations of the near future, randomizing the start values and simulation parameters within the range of uncertainty, and get one potential future per simulation out of the process.

A 30% rain chance every hour of the entire afternoon until sunset tends to be less actionable than "85% chance of a 50~80 minute rain shower during the afternoon, unclear when, but likely a bit longer if it'd start very late", as one can often adapt around such, for example by scheduling the homework to get done "wherever it's raining, at the latest so it'll be finished by dinner time", spending the dry time with outdoor physical activity that doesn't care about wet ground (but getting wet from above isn't nice).

Re: 300 meters resolution SF Bay Area Forecast

#90

Post by nikisweeting »

londons_explore wrote: post #58

Weather forecasts are so hard for a user to evaluate... Are you going to check it every day and remember how many days it was right or wrong? Please can weather providers just publish a headline statistic of "Our rain/no rain one day ahead forecast is right 85% of the time. That is better than NOAA (80%), Met Office (72%) and weather.com (65%)."

Atmo built their business on beating their competitors on a particular ranking metric that does exactly as you describe. I discussed it with the founder but am not sure how much of that convo was public info so I hope they post it here.
Post reply on HN