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

WeatherNext 2: Our most advanced weather forecasting model

blog.google

Post reply on HN
Print view View original

41–50 of 140 posts

Re: WeatherNext 2: Our most advanced weather forecasting model

#41

Post by sails »

binsquare wrote: post #4

I find it interesting that they quantify the improvement on speed and number of forecast-ed scenarios but lack details on how it results in improved accuracy of the forecast per: ``` WeatherNext 2 can generate forecasts 8x faster and with resolution up to 1-hour. This breakthrough is enabled by a new model that can provide hundreds of possible scenarios. ``` As an end user, all I care is that there's one accurate for…

As others have explained, ensembles are useful.

As a layperson, what _is_ useful is to look at the difference between models. My long range favourite is to compare ECMWF and GFS27 and if the deviation is high (windy app has this) then you can bet that at least one of them is likely wrong

Re: WeatherNext 2: Our most advanced weather forecasting model

#42

Post by jakewins »

mlmonkey wrote: post #27
jakewins wrote: post #7

Is anyone aware of good sources of higher resolution models? Hourly resolution like this model provides doesn’t help much now that energy markets have moved to 15-min and 5-min resolution.

How does one use weather data in an energy market, if you don't mind my asking?

Yeah exactly like hackitup7 says, it has a huge impact on both sides of the supply and demand equation. It both drives house heating and cooling, which has a massive consumption impact, and it drives solar and wind production.

But knowing "there will be a massive drop in temperature between 1pm->2pm" doesn't help much anymore, you need to know which 15-minute or 5-minute block all those heat pumps will kick on in, to align with markets moving to 15-min and 5-min contracts.

Major forecasts like ECMWF don't have anything like that resolution; they model the planet at 3 hour time scale, with a 1 hour "reanalysis" model called ERA5.. hoping to find good info on what's available at higher resolution.

Re: WeatherNext 2: Our most advanced weather forecasting model

#43

Post by edbaskerville »

bitdivision wrote: post #35
mlmonkey wrote: post #27

Earlier quoted context omitted.

How does one use weather data in an energy market, if you don't mind my asking?

Seems like it would be pretty useful to forecast the supply of renewables (wind, solar, maybe some hydro).

Indeed. In the not-too-distant future where renewables are the vast majority of generation (sooner in China than in the U.S. at current rates of progress), the weather matters more and more.

Re: WeatherNext 2: Our most advanced weather forecasting model

#44

Post by jachee »

bilsbie wrote: post #38

Where can I use this? I’ve been trying to find hyperlocal forecasts like darksky used to be.

Apple integrated the hyperlocal darksky stuff into their native Weather app. It had a few growing pains, but it's as good as it ever was, imho.

Re: WeatherNext 2: Our most advanced weather forecasting model

#45

Post by jadbox »

paxys wrote: post #39
bilsbie wrote: post #38

Where can I use this? I’ve been trying to find hyperlocal forecasts like darksky used to be.

> We're now taking our research out of the lab and putting it into the hands of users. WeatherNext 2's forecast data is now available in Earth Engine and BigQuery. We’re also launching an early access program on Google Cloud’s Vertex AI platform for custom model inference. > By incorporating WeatherNext technology, we’ve now upgraded weather forecasts in Search, Gemini, Pixel Weather and Google Maps Platform’s Weathe…

Google Maps has... weather predictions?

Re: WeatherNext 2: Our most advanced weather forecasting model

#46

Post by batrat »

jadbox wrote: post #45
paxys wrote: post #39

Earlier quoted context omitted.

> We're now taking our research out of the lab and putting it into the hands of users. WeatherNext 2's forecast data is now available in Earth Engine and BigQuery. We’re also launching an early access program on Google Cloud’s Vertex AI platform for custom model inference. > By incorporating WeatherNext technology, we’ve now upgraded weather forecasts in Search, Gemini, Pixel Weather and Google Maps Platform’s Weathe…

Google Maps has... weather predictions?

if you search for a city usually it shows the current weather, but I've seen in some cities there is also a 7 day forecast

Re: WeatherNext 2: Our most advanced weather forecasting model

#47

Post by baxtr »

jachee wrote: post #44
bilsbie wrote: post #38

Where can I use this? I’ve been trying to find hyperlocal forecasts like darksky used to be.

Apple integrated the hyperlocal darksky stuff into their native Weather app. It had a few growing pains, but it's as good as it ever was, imho.

Agreed.

The one thing I’d like them to improve are the precipitation maps though. They just feel awkward and unreliable.

Re: WeatherNext 2: Our most advanced weather forecasting model

#48

Post by LogicFailsMe »

politelemon wrote: post #8
LogicFailsMe wrote: post #3

This year, the wild variance in hourly weather reports on my phone has really been something. I attributed it to likely budget cuts as a result of DOGE, but if those forecasts came from Google itself the whole time, all is clear now.

I find that unlikely, my forecasts for much of Europe and East Asia have been consistently accurate.

Yeah... So you know that's not the United States right? Though judging by the down votes, it's quite triggering for some and I can't say which side when I pivot from blaming DOGE to blaming bad AI. Curious(tm)...

And I say that as a huge fan of AI, but being vocally self-critical is an important attribute for professional success in AI and elsewhere.

Re: WeatherNext 2: Our most advanced weather forecasting model

#49

Post by cleak »

lysecret wrote: post #30

Im pretty deep into this topic and what might be interesting to an outsider is that the leading models like neuralgcm/weathernext 1 before as well as this model now are all trained with a "crps" objective which I haven't seen at all outside of ml weather prediction. Essentially you add random noise to the inputs and train by minimizing the regular loss (like l1) and at the same time maximizing the difference between…

That’s pretty neat. It reminds me of how VAEs work: https://en.wikipedia.org/wiki/Variational_autoencoder
Post reply on HN