300 meters resolution SF Bay Area Forecast
sf.atmo.ai
sf.atmo.ai
Then I moved to the Bay Area, and weather does not only change quickly, it may also be vastly, vastly different just a short distance away. Temperature differentials of 10°C or more within just 40 miles are interesting enough that it's a frequent topic of conversation with friends in Germany.
Everything suddenly made sense. The weather widgets. The hoodies: Easy to put on or off.
Sunny weather was a bit like building up a sort of pressure, that must release violently.
Took a while to let go of that feeling in California. It basically rains in winter, and does not rain in summer. Like, at all.
(Note that I moved away about a decade ago, climate may have changed in the meantime.)
You needn't have moved so far from southern Germany to experience that! Lack of rain in the summer is a defining characteristic of the Mediterranean climate. You would have experienced the same a few hundred miles south. :-)
When I first got here I was stunned by how noticeable nicer the weather was when driving from Santa Clara to Palo Alto, and more than once have I forgotten to bring a sweater to SF.
On the bright side, a hoodie is often all I ever need, all year long.
Even if you stay right next to the ocean/river during the whole trip by avoiding the highway you'll notice a big difference most days.
[1] https://www.mmm.ucar.edu/models/wrf [2] https://github.com/wrf-model/WRF
Its not clear to me that this is either.
I don't know how the ECMWF model works, but even as someone who did not study meteorology (but studied electrical engineering, which forms the theoretical basis of weather forecasting via the Kalman filter), I can say the following (having spent a number of years working at NCEP): 1. Initial conditions/parameters are fundamental in setting up a model run. 2. Forecasts have for a long time relied on ensembles, which are repeat model runs with slightly varying parameters. The idea of ensembles is, if you run enough of them, you will frequently notice one or more convergence(s) that various sets of parameters produce, e.g. where some sets of parameters predict one movement pattern for a hurricane, while others produce a different movement pattern. Historically, such discrepancies were resolved by actual forecasters, who decided based on their knowledge and experience which one was more likely. In addition, they also had meetings every morning between scientists (developing the model) and forecasters (who relied more on general knowledge and experience) and involved occasionally heated discussions between the groups. But I digress. 3. Considering it involves a chaotic system, I cannot say how much value something like deep learning might bring to the table that produces consistent value above and beyond what's already obtained by using ensembles of Kalman predictive filtering. It is however noteworthy to point out that if the grid resolution is 28,000 meters, then it may not make much sense to set the resolution of the model itself substantially lower (like 300 meters), because any resulting data is more likely to be an artifact of the model itself, rather than reflective of real life information. Luckily, this issue has been and is being addressed through the development of rigorous testing standards, which inform of the inherent quality of forecasts produced by a particular model (this is how they can assign an objective rank to e.g. the GFS and the ECMWF, when forecast quality is generally very close and the model producing the most accurate prediction varies between the two). To put it plainly, the degree to which the website mentioned above has any value is based not on its best predictions, but on the overall variance (i.e. how close predicted data comes to actual measurements of the same, which is necessarily retrospective). 4. That said, it's worthwhile to point out that just because it doesn't involve a government agency with something like a thousand employees, hundreds of scientists (in the case of NCEP alone), and very powerful supercomputers, does not necessarily mean it's bunk (even if it frequently does). For example, I do recall Panasonic (IIRC) showing up out of the blue, with its own forecasting system, which was shown to be competitive after requisite, rigorous testing. I don't remember many details and this was years ago—and its disappearance alone is suspect, but it's worth adding for completeness.
Kalman filtering is only one part of the process, and plays a critical role during the data assimilation part. Classical Kalman filtering is optimal for Gaussian-distributed linear dynamical systems, but needs tweaks for non Gaussian distributions and non linear systems.
Classical NWP models for instance will integrate the primitive partial differential equations in time and space and run various parameterizations (which can be in some cases even more expensive than integrating the primitive equations). ECMWF on their end use IFS, which is a spectral method for solving the PDEs.
The whole process of solving these models accurately has definitely been some of the most fascinating science and engineering I’ve had the pleasure to work with. It’s extremely humbling :)
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)
Numerical weather prediction is a _very_ well established field. In fact, large tranches of modern computer science and computing in general owe their existence in direct ways to the importance of numerical weather prediction, since this was one of the original applications of digital computers! Modern weather forecasting models are extraordinarily sophisticated scientific and engineering achievements. It's not obvious that AI actually offers any significant, immediate benefit over these tools save for niche, simplified forecasts (e.g. precipitation nowcasts) - certainly, given the prowess of modern NWP, the ROI is likely to be very low for research investments into general purpose AI weather forecasts.
One might then argue that perhaps AI can be useful to help refine or post-process these existing forecast systems? But of course - we've been doing just that since the 1970's. In fact, even the basic weather forecast that you might get from your national weather service these days is based on a sophisticated statistically post-processed machinery applied to not one but dozens of weather forecasts.
Weather prediction is unlikely to be a field where AI practitioners stumble across a significant improvement to the status quo. It would be far wiser to work closely with meteorology experts to solve practical and _useful_ weather forecast problems - like, is that thunderstorm I see on radar likely to produce a tornado in the next 45 minutes?
zero interpolation, zero forecasting. just the real data from the radar feed in 10 meter resolution at the current time (plus 3 past snapshots at 1 hour intervals):
and the app versions:
https://play.google.com/store/apps/details?id=net.conceptual...
https://apps.apple.com/ca/app/truweather/id1537614881
i find it useful because every other weather app is wrong in some way ;) instead, it enables you to form your own mental models
I don't see that reflected in this map at all fwiw. https://sf.atmo.ai/temperature@37.83200,-122.51075,13.53,20,...
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%)."
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).
With so much heat map display, why is there no key? For wind, red/purple are counterintuitively lower speed than orange and yellow?
Also, who do so many applications force oblique views?
I'm able to adjust it by dragging with two fingers on my trackpad, which I think is the standard behavior for that (albeit hard to discover). But I do agree, it's weird for that to be the default.
There's also probably no key because the colors are mostly transparent, so it would be hard to make a key easy to understand. Labelling the contour lines seems like a reasonable approach imo.
- 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 transition hides it a bit, but makes the data uncomfortable to view when playing the timeseries.
- I'm deeply curious about the Picnic data layer. Can someone share the ArcGIS/QGIS model for that one? =D
Unrelated: The site breaks my back button in Chrome, which is an unforgivable UI sin.
Weather predictions seem to be accepted quite uncritically. Perhaps people have a lot of confidence in the smart people that built these predictions (a bit like how AI predictions can sometimes be accepted uncritically).
My grad thesis advisor encouraged me to actually get the Environment Canada models and learn how to run them (they're in FORTRAN). I could never make them spit out data consistent with what EC publishes. That's probably on me, but it was a real eye-opener to this whole domain's complexity.
I’m always excited to see new forecast products, generally. If I were to guess (as an above comment did) it looks like they are applying some dynamic downscaling on top of either a custom WRF model (expensive and complicated) or more likely already available weather model data like the HRRR, which still would represent a 10x resolution increase.
I’m more curious what the refresh rate is. Anyone can get a super accurate forecast for the next 3 hours that takes 10 hours to run, but at that point it’s no longer a forecast by the time the data is available.
I still think that windy has set the standards as far as modern weather visualization goes. Not saying everything has to be particles but other things (like the inclusion of isobars) is really clean and not trivial to execute.
Either way this has definitely piqued my interests and I will be keeping an eye on it, their advisory board looks legit (at least in the meteorology end)
It would be interesting to see how this behaves for longer prediction times and across a range of difficult forcing conditions off the ocean in the BA.
With regards to the sfbay specifically I used to work with a fairly high resolution wind model for the bay (this was a more traditional dynamic based simulation) and it worked pretty well overall, but every time a storm blew through it would crash. This ultimately had to do with the relatively steep terrain in the bay specifically (and the physics configurations we were using in the actual model).
Even if they are using DL they still need initial and boundary conditions. As I said there are a ton of weather stations around so I could imagine a DL type approach that looked at terrain elevation, and recent + historical observations to initialize a forecast, but I still imagine that boundary conditions would have to be provided by nesting this in a larger model somehow. Then again, I'm not a DL expert at all so there are probably some newer stuff in this field that I'm just out of date on.
Its really expensive to run your own dynamic forecast model, at a refresh rate acceptable for an actual forecast, at this resolution. That's why I suspected its taking existing weather models and downscaling them with DL techniques, but I can't really know just by looking.
This chapter on Numerical Weather Predictions [0] is great, especially the section on "Forecast Quality and Verification" (p777). The eye-opener for me was "Binary/Categorical Event". An example of a binary event is rain, one model could predict rain correctly but a second model might not predict the rain at all. This doesn't mean the second model was completely wrong, it still predicted the rain but it predicted the rain passing further to the south.
[0] https://www.eoas.ubc.ca/books/Practical_Meteorology/mse3/Ch2...
I've also noticed some model are better than other at predicting one phenomena while other models might be better in certain regions. For example, many people report that Canada's GDPS is better at higher latitudes whereas NOAA's GFS is better at equatorial regions.
One final note, just because someone is solving an WRF model without verifying the results, doesn't mean it's wrong. Many numerical techniques and physical models within WRF have been validated analytical and experimental models. But it is also true that someone can naively setup a WRF model that gives bad results.
I use a 900m WRF model that predicts the wind shadow around an island and we use it to find the best beach for a picnic - and it works. But this same model predicts the general pattern of rain but it doesn't get the start and stop time of rain correct.
People get fixated on accuracy as a single thing and use it as a single basis for argument but to take a quote from the chapter [0] above "One of the least useful measures of quality is forecast accuracy" (ref. p777, Forecast Quality and Verification, third paragraph).
The US Navy's COAMPS model is good for littoral regions.
I find global models like GFS are great for understanding the large scale weather systems. The regional high-resolution models, which are usually nested in a global model, give better definition of local weather phenomena like wind shadows or cooler temperatures in valleys.
Dues to averaging, weather simulations usually have a bias error in temperature predictions. These errors are corrected using statistics (look up Model-Output-Statistics) but is hyper-local, i.e., you loose the big picture. This is probably what you're looking at with Meteoblue.
The convective cell tracking for nowcasting seems finer, IMO reasonable as it's about predicting watersheds down to <10km² area flash-flooding and causing the local creek to swell to actually dangerous levels/requiring partial evacuation of a valley.
- 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
- Berkely, Oakland, Emeryville getting blasted late afternoon is correct
- Back side of treasure island, immediately to the east is much lower than the west side, particularly near clipper cove
- Vast majority of alameda estuary is dead calm, that's correct for this time of year
- There's a big blast of wind between Daly City and OAK international where there's a gap in the mountains
Weird stuff:
- Most noticable, is the wind is still strong up to and south of the bay bridge. The bay bridge has been described by many as "a wall" when it comes to the wind. There's a drop off but it's not in line with the bay bridge. At all. at least 45 degrees off from true wind speed.
- There's a very windy patch between golden gate coast guard station and belvedere, it's usually really patchy wind here but I guess if the wind direction is just right it'll blow there
- Pointe Bonita (lighthouse on the west side of gg bridge about 2 miles, north coast) they are modeling the gap in the rocks there and you can see it funnel through which is neat
It's a cool visualization though, gives you a great idea of where the wind is, and more importantly where it won't be. There are a bunch of races that start in the bay and head south towards Santa Cruz and Monterey so it's nice to better visualize where the wind just dies off on the coast as it skips over the mountains.
Anyone who wants to see what the wind is like in the bay I recommend reaching out to YRA.org they can put you in touch with a boat who needs crew most likely. There are races 4-5 days a week through november all around the bay. It is modeling a distinct drop off of wind speed on the south side of the bay
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?
ECMWF's model is 9km spatial resolution, so Angel Island probably doesn't even show up in the model domain.
I am very skeptical. Does the San Mateo bridge really block 10 knot winds for the entire south bay? Similarly, the land temperatures all seem the same close to sea level.
I did briefly see something that looked like a map, but then it smeared into an abstract art looking blob.
I see no sign that forecasting has improved at all in my part of the world. Auckland, New Zealand.
What does the picnic data show?
Pedantic but suggest title rename to say “SF Bay Area” or “inner SF Bay Area”. SF is just a small portion of the coverage area.
All those US customary units should never be the default in 2022. It's time to move on people. The whole world and the majority of the industry in the USA share the same system of measure. Let's get the American population there too.