Saturday, February 8, 2020

Smartphone Weather Apps. Can You Trust Them?

If you picked up the Seattle Times on Wednesday, you would have seen at the very top of the front page a reference to an article "Why your weather apps isn't very good"
______________________________________________________________________


__________________________________________________________________________

Heading to the article, you learn you can't necessarily trust your smartphone weather app, that one should go "straight to the source" and use the National Weather Service forecast, and that weather apps “ don’t factor in that we are sandwiched in between mountains and the very warm Lake Washington and the warm Pacific.”
________________________________________________________________________
__________________________________________________________________________

Unfortunately, most of the information in the Seattle Times article was simply not true.  

 The truth is that many of the leading weather apps available on your smartphone (or through your web browser) are VERY GOOD, in fact usually much better than the National Weather Service forecast.

The Real Situation

Fortunately for all of us, there is website that provides verification of weather apps as well as the National Weather Service: https://www.forecastadvisor.com/

Let's take a look at their evaluation of forecast accuracy (1-3 day forecasts, temperature and precipitation) for Seattle, Washington (see below).

The Weather Channel (weather.com, weather channel smartphone app) is the leader, both last month and last year, with others like AccuWeather, WeatherUnderground and Foreca, right behind).

The National Weather Service (NWS Digital Forecast) is lagging the leaders.   These results are not a fluke: I look at this website every week and the results during the past few years were very similar.  I have verified some of the result myself quantitatively and they were consistent with the above.

Other cities?   Pretty much the same story.  Here is San Francisco.


The Seattle Times article has a quote that these apps don't know about local terrain effects.  This is incorrect.    Weather apps like the Weather Channel use local observations and the output from high resolution models.

I know many of you have the same question:  why are these weather apps so good and why are they often better than the local meteorologists of the National Weather Service?   

This can be easily explained.  All the best weather apps follow a similar approach, but let me talk about one I know well, the technology behind the Weather Channel app, which is actually owned and developed by IBM (through one of its companies, WSI).

This underlying forecasting system brings in a wide range of observations and a big collection of forecast models.  Such forecast models are not limited to the American models (e.g., GFS) and includes highly skillful international ones (like the European Center).  In addition, IBM runs a high resolution model (WRF) over important areas (like the U.S.) and now runs its own global model (MPAS), with extraordinary resolution over the U.S. and Europe.

But the Weather Channel system doesn't stop there.  It uses a sophisticated statistical system to combine all these inputs, with weights dependent on past performance.  Such technology was based on a system first developed at the National Center for Atmospheric Research (DiCast) and is known as GRAF (Global High-Resolution Atmospheric Forecasting System), with a description here: https://newsroom.ibm.com/graf-democratizing-the-worlds-weather-data)

For most forecast situations, it is very, very difficult for human beings at the National Weather Service (NWS) to beat such a system--that is why NWS skill is generally less than that of the leading apps.  IBM is also heavily using technology, models, and science developed by the university community (e.g., NCAR and the universities).

Let me be clear: human forecasters at the National Weather Service are still very valuable.  They can spot the rare situations when the automated systems are failing and intervene, they can communicate the forecasts and their uncertainties to the public and key users, and they can monitor the local observation assets, among many other tasks.

But these smartphone weather apps and their online cousins are very, very good,  and with improving models, increasing numbers of observations, and the growth in use of machine learning/artificial intelligence for combining all the very information sources, they are only going to get better.

38 comments:

  1. Something Cliff does not mention but that I've observed is the apps' display of weather radar is not as good as the one at the UW Atmospheric Sciences web site (https://atmos.uw.edu/current-weather/northwest-radar/). I rely on the radar to time my bike commutes, and I have not found an app whose display is nearly as accurate as the UW's, which is weird since they must all be using the same data. I compared a bunch of them with the UW site during a recent heavy rain, and I observed two common issues: 1) the apps all used a 5-minute increment, while the UW uses a 3-minute increment; 2) the apps' rendering of the images were all coarser-resolution than the UW's, which meant that they "smoothed out" and missed fine features and light rain toward the edges of showers. The difference was plenty big enough to get me wet on my commute, so I've given up on all of them and just use the UW site on my phone even though it is not very mobile-friendly.

    ReplyDelete
    Replies
    1. The raw radar doesn't update at 3 minute intervals so that's why other sites use 5 minutes. Most non-meteorologists prefer the 'smoothed out' view too...

      Delete
    2. The UW Atmospheric website radar isn't any better than other radars. They all use the same soucre which is the Doppler Radar located on Camanon Island and on the coast. I prefer my radar source from the NWS because you can change the time frame on it. I also like that NWS gives you full access to the GOES satellites and all the different bands. UW only gives you access to one band of the GOES. If your using a forecast to plan your day then UW is probably fine, but if you're a weather geek like I am then the data you get off the NWS website is better suited.

      Delete
  2. One thing to clarify - they are very good WHEN THERE IS NO ACTIVE WEATHER. On a day-to-day basis, temps and general conditions are often better on those apps. But when there is snow or storms or anything else that would be considered "real" weather, the apps crash and burn. (But the public doesn't realize it, and that's where this article failed.)

    ReplyDelete
  3. I generally use the Weather Underground app myself mainly because I have two consumer level weather stations, an Acurite 5 in 1 and a brand new Ambient weather unit that both transmit my data as a PWS location accessible by the app to WU. My Acurite also transmits to CWOP that passes it along to MADIS. In both cases my weather is being used to help model the local areas weather so as best as I can tell being located on the Kitsap Peninsula, the information i see on the weather services for the locale, and the data from both of my PWS's are pretty well close enough to not be an issue with the apps accuracy.

    ReplyDelete
  4. The Weather Channel app sucks for NE Oregon. Weather Underground is closer, but we're in the middle of the black hole and there's a huge difference between what happens in the Grande Ronde Valley and the Wallowa Valley...or even the upper and lower Wallowa valley.

    ReplyDelete
  5. Thanks Cliff for "clearing the air" I wonder why ST decided to do a hit piece on these apps. Would it be possible for you to come in on the weather radar apps. I have turned to one to help time my bike commute between deluges, but I'm not sure how to interpret the moving blobs especially when they seem to appear out of know where. Thanks again for insights.

    ReplyDelete
  6. Hello Dr Masss,

    I go directly to the U W Atmospheric Sciences weather models 1.33 grid spacing. It doe not get any better than this! Kudos to your department.

    ReplyDelete
  7. My take is to not use only one app/service. Better read the forecasts from different forecasting services, weather channel, NWS, etc. It is interesting when they forecast the same for the next 72 hours. Then, you know there is a high probability it will be correct. But, for a user it's often difficult to tell which one is more correct especially for the 5-day outlook because the forecasts keep being updated.

    ReplyDelete
  8. I see no mention of Windy.com phone app. That's the one I use and I really like it. It seems to be very accurate compared to the other top apps since I've tried them before.

    Or maybe Windy.com app got a very low score?

    ReplyDelete
    Replies
    1. Windy.com app uses raw GFS/ECMWF/NAM output.

      Delete
  9. The Weather reporting Rock-on-string in BOW WA is never wrong in its reporting. Dry, wet, white, warm, cold, moving, not moving. But it is not much use if I'm trying to ride home from Edison.

    ReplyDelete
  10. The forecast quality of WEATHER UNDERGROUND has gone done the toilet bowl in the last month as shown in the table. The Weather Underground mobile site used to be my go to weather app.....then IBM purchased it and destroyed the quality of the app. On Google Play Store it went from 4.5 stars **** to something like 2.8 ** the last I looked. Thousands of users wrote in negative comments and dropped the app in disgust. I am still looking for a replacement mobile weather app....ANY ADVICE?

    ReplyDelete
    Replies
    1. Well you could always read the article and go for one of the higher-accuracy apps clearly listeds.

      Delete
    2. I've noticed the decrease in Weather Underground accuracy as well in recent months.

      If both Weather Underground and The Weather Channel are owned by IBM, does anyone know why they have a big difference in accuracy?

      Delete
    3. Poopie...it's not necessary to be condescending when replying to a comment. My comment was not about accuracy, but about the reduced quality of the display and included content of the changed Weather Underground app. There is enough arrogant and disrespectful activity on the web. We don't need it here on this website as well.

      Delete
  11. MyRadar app on my iPhone is fantastic and it’s a must have for those who spend time outdoors and it’s free

    ReplyDelete
  12. The forecast discussions the NWS puts out are great - I read them every day.

    The Seattle Times once again shows its stuff - it's the best bird cage liner available anywhere.

    ReplyDelete
    Replies
    1. I agree re: the NWS discussions - they tend to be very comprehensive & as spot-on as weather forecasts can be.

      I ALSO agree 1000% re: the Seattle Times usefulness = tiz one of several factors in our decision to cancel our subscription !

      Delete
  13. When I looked at the "forecast adviser" more detailed accuracy reporting, there appeared to be an erroneous zero in the "icon" column for WUnderground which is what took it's % way down below The Weather Channel... aren't they owned by the same company and using the same forecasts?

    ReplyDelete
  14. Whatever happened to the UWx app? It doesn't seem to work anymore.

    ReplyDelete
  15. I can only guess, but it would seem that the volume of real-time (or near real-time) PWS sensor data employed by those apps gives them an edge. How much does NWS and-or NOAA look to folks' data? I've looked at lists of the stations used for climate "data" in this county, and some of those locations seem like poor choices. There's a trio up here I've started calling "the Sahara triangle"; their precip is routinely far less (like half) of the watershed's on average. If the NWS and other models rely on those spots...Oye!

    ReplyDelete
  16. Now's Day android weather apps are more accurate. But i can't trust them because it's shows false weather details dua to some reasons or bug. But i can check it my self. Google Find My Device app क्या है?

    ReplyDelete
  17. Here in Central Washington Moses lake the radar is very inaccurate. Often when it shows rain it's not raining and vice versa. The Noaa office in Spokane is not near as accurate forecasting Moses lake weather, as the Seattle Noaa is forecasting Seattle weather. I have tried every big weather app out there and nothing is really very good. Honestly I can usually forecast when their forecast is going to be wrong. It's pretty useless for the farmers around here. I have written the Spokane office about the radar problem in the basin and they tell me it's $3 million to update properly and it probably will not happen anytime soon. Sucks for us.

    ReplyDelete
  18. Cliff,

    I'm sticking with the UW Atmospheric Sciences site! It's very interesting to learn about what it provides and how to best use the information in your daily life.

    ReplyDelete
  19. FYI Cliff, there's a typo in your copy of the article title but not in the source. The screenshot directly below the quote shows "Why your weather app isn't very good" rather than "Why your weather apps (sic) isn't very good".

    ReplyDelete
  20. Startup Climacell is another app that I find wonderfully accurate. They use cell tower info and AI. https://www.climacell.co/ Also along Cliff's wonderful analysis, I use the UW/NOAA site. Very, very useful and I really enjoy learning about PNW weather. It's fascinating. (And I have no connection to Climacell, other than using their app)

    ReplyDelete
  21. Look out your window... check your thermometer...most times the weather folks can't predict what it did the day before much less a week out...I use any of their guesses as just that..a guess for severe weather...at least if its severe enough the computers will at least give me a heads up...but in any event the weather here isn't normally severe enough to change my plans...I do however want to send my thoughts to those affected by the latest floods..I always wanted to be a weather guy...still would now...

    ReplyDelete
  22. Most weather apps are not even internally consistent, i.e. if it says the high and low temperature for a given day, clicking to the hourly forecast shows different highs and lows. But without clicking to the days, the little icons of sun and cloud don't mean very much. Only the discussions on the NWS station pages can really do justice to the details and uncertainties inherent in forecasting weather.

    ReplyDelete
  23. Funny thing is the majority of people I know use their phone as a phone not as a pretend computer. They just call people or text people maybe do casual shopping and that's it.

    ReplyDelete
  24. I've used Weather Underground for years, and since they were purchased by the Weather Channel their quality has gone down hill. Do you think the Weather Channel is sabotaging Wunderground?

    ReplyDelete
  25. Windy.com and Weather.us are both outstanding for users who want visualization tools for the top forecast models. And I have a paid subscription to Radarscope Pro, which is in a different league than the mass of radar apps that target casual/consumer users.

    All three of these services require self-instruction to use properly, and they're likely to frustrate and/or overwhelm casual users. But for enthusiasts they're head and shoulders above the usual fare.

    One fact of life here is that precip will, in many cases, appear "out of nowhere" or suddenly vanish off the radar. That's the result of fast developing features like a PSCZ, orographic effects, and sudden, drastic changes in low level wind patterns. If you're trying to predict the exact timing, placement, and intensity of a convergence zone, for example, just accept that no forecast can (or probably ever will) give you useful results. There have been plenty of days when I'm out with my camera, and a PSCZ turns a clear and sunny morning into a torrential downpour or a flash hailstorm in the space of 15 minutes. I carry really good raingear and never change my plans -- i just deal with whatever the weather throws at me and marvel at how cool it is to experience one of the world's most unpredictable and dynamic climates right outside my front door.

    ReplyDelete
  26. Can someone here tell me what is considered an accurate forecast of temperature low and high? I can speaking of a short-term forecast, for example, checking the temperature forecast at 7:00 pm for the overnight low. It is not unusual for the temperature for the Olympia Airport to be five or more degrees colder than forecast. Is that considered accurate? For deciding whether to protect may plants I used to assume a low of four degrees lower than the forecasted low but I have had some scares where lows forecasted to be in the upper 30s have instead been below freezing. Thanks for your help.

    ReplyDelete
  27. I'm confused - why don't the human forecasters at the National Weather Service just use one of the apps instead of their own models if the apps are usually more accurate?

    ReplyDelete
  28. isn't overnight low (unless warm front comes in) generally right before sunrise? Definitely not at 7 PM anywhere I've lived.

    ReplyDelete
  29. IBM purchased wunderground and already owned weather.com. so natch they'll kill one, looks like wunderground. Monopolies kill competition very effectively.

    ReplyDelete
  30. I still think the smartphone app is very accurate in weather forecast, and it is updated in a timely manner, which is very helpful for my travel. Moreover, if in different places, my weather will change, of course, I need to refresh. So even if I come to China from America, I won't worry about the weather information.

    ReplyDelete
  31. The issue I have isn't about the accuracy but rather the radar images just not loading. This has been plaguing my radar apps for ten years now, across many different apps and two operating systems (Android and then iOS) and different internet providers. Right now there's a massive storm racing across the midwest where I am, and the app I'm using this time (Dark Sky) is only loading the radar on the left half of the screen, west of Lake Michigan. The right half of the screen (where I am, Michigan) is blank. And there's a solid vertical edge where the image didn't load. And that's exactly what I've seen over and over across all these apps over the years. Radar images simply not loading or only part of them loading. (And it happens whether I'm using Verizon data over the air or my local internet connection.) I just switched to Accuweather right now as I typed this, and now that one is finally loading (it wasn't earlier) and I can see how close the storm is. But I don't understand why the images just don't load or will only partially load.

    ReplyDelete