Skip to content
Thoughtful, detailed coverage of everything Apple for 34 years
and the TidBITS Content Network for Apple professionals
80 comments

Dark Sky Fading; iOS 16’s Weather Brightens

A few years ago, Apple bought its way to better weather forecasts, as we covered in “Apple Acquires Dark Sky Weather Service and App” (31 March 2020). Apple removed Dark Sky’s Android and Wear OS apps shortly after on 1 August 2020. It took a little longer than anticipated for Dark Sky’s core features to migrate to Apple’s bundled Weather app, but at least some of them did last year (see “Weather in iOS 15 Gains Precipitation Notifications (But Do They Work Reliably?),” 17 November 2021).

Now Apple has announced that the Dark Sky iOS app will cease to be available and will no longer provide weather data as of 1 January 2023. The Dark Sky API and website will hold on a few more months, until 31 March 2023, possibly to allow all of Dark Sky’s contracts to sunset. Apple’s new WeatherKit API allows developers to integrate Apple Weather forecast data into their apps.

Dark Sky ending

This news should come as a shock to no one—Apple has been quite clear that Dark Sky was on borrowed time—but it does mean that Dark Sky fans will need to find an alternative by the end of the year. The obvious choice is Apple’s Weather app, which received several key new features in iOS 16.

For those like me who are interested in hyperlocal weather patterns to see if you’ll get wet on your run or ride, Weather has added to the existing 12-hour animated forecast map (below left). The new next-hour forecast animation (below right) is far more detailed and likely to help you plan your outing. Switch by tapping the forecast name at the bottom, just to the right of the pause button.

Weather's Next-Hour Radar

The other change is even more significant. When you tap the hourly forecast or a day in the 10-day forecast, Weather now displays a daily time-based graph of one of eight metrics: temperature, UV index, wind, precipitation, “feels like” temperature, humidity, visibility, and pressure. Switch by tapping the menu on the right side of the screen. You can also tap a day at the top or swipe left and right to navigate to other days in the 10-day forecast.

Weather's Daily Graphs

My gripe with the daily graph screen is that it defaults to temperature rather than remembering what you had previously viewed. I’m less interested in temperature than precipitation most of the time, so I’d have to switch to the precipitation forecast manually every time. The workaround is to scroll down to the display of a particular metric on the main screen and tap that to jump directly to that metric’s graph screen. Better, but still fussy.

Welcome as these features are, I don’t see myself using them. Apple’s Weather app may have integrated Dark Sky’s capabilities, but I don’t love its interface, particularly with the individual metric cards, which are difficult to parse quickly. Ever since it became clear that Dark Sky was living on borrowed time, I’ve become fond of CARROT Weather and its interface construction set approach, which lets you tweak all the possible display options to your preferences (see “CARROT Weather Predicts Cloudy with a Chance of Snark,” 22 January 2018). Since CARROT also lets you pick from numerous weather services, including Dark Sky until next year and Apple Weather now, there’s no liability when it comes to data quality. (I admit to sometimes switching between services to see if I can shake loose a better forecast.) But of course, Weather is free, whereas CARROT’s advanced features—which also include Home and Lock screen widgets—require a Premium membership. Well worth it for me since CARROT is among the apps Tonya and I use most frequently.

Weather vs CARROT Weather
Left: Apple’s Weather app; Right: CARROT Weather

Is Weather now your main weather app, and if not, what do you prefer?

Subscribe today so you don’t miss any TidBITS articles!

Every week you’ll get tech tips, in-depth reviews, and insightful news analysis for discerning Apple users. For over 33 years, we’ve published professional, member-supported tech journalism that makes you smarter.

Registration confirmation will be emailed to you.

This site is protected by reCAPTCHA. The Google Privacy Policy and Terms of Service apply.

Comments About Dark Sky Fading; iOS 16’s Weather Brightens

Notable Replies

  1. I am back to using Apple weather again, I wasn’t sure what source to choose on Carrot weather, most seemed to be inaccurate for me in N.Ireland.

    Dark Sky was accurate’ish.

    Not one of them was any good for Hungary when I travelled to Budapest for F1, we resorted to looking at the Sky and feeling the wind direction to see how wet we might get.

  2. What has me a bit concerned is that for the same location, macOS weather widget, iOS Weather, and the DarkSky website do not agree on forecast.

    Not just a degree here or there, but sunshine vs. overcast skies or rain vs. dry.

    Since all three belong to Apple, I would have hoped that they can determine which one of their data sources is most accurate and then stick to that across the board.

    It appears that has not (yet) happened. So I know at least one of my 3 sources of weather information is not as good as it could be. But unfortunately, without knowing conclusively which one of the three it is, I don’t know which one to drop in favor of another.

  3. I’ve been using Forecastbar and Hello Weather on iPhone, and Forecastbar on the Mac. I’ve switched to Apple’s Weather app on the iPhone, and may do the same in the Mac, depending on its implementation. I hope it has a menu bar drop-down for a quick glance at weather conditions.

  4. You can check ForecastAdvisor to see which services are best for your area.

  5. There has been much discussion on this topic in meteorological forums in the past few years and the consensus is that the methodology used by ForecastAdvisor is suspect. For example, in my city they compare the forecast and verification at different locations that are about 20 miles apart and at different elevations. Some locations are worse; most are probably better.

  6. I notice this site is only US

  7. Drat, that’s a bummer, @david_blanchard, and yes, sorry @heronlen, I do now see it’s US only.

    I was going to add that Forecast Advisor has a Further Analysis button at the bottom of the screen that breaks out high and low temperature, icon precipitation and text precipitation. Dark Sky was weak on temperature but the best on precipitation for Ithaca; I care more about precipitation than temperature, and this somewhat confirms my belief that Dark Sky is better than others in this area. I never know what to make of temperature anyway, since it can vary so much by location around here.

  8. I am sold on Dark Sky the way it is. I have used the paid edition for years and run it alongside the Apple Weather app. With both set to use current location, Dark Sky is often more accurate in terms of humidity, dew point, and temperature when compared to Weather Underground at the same location. So I do not understand how Apple is getting their information from Dark Sky; they vary as much as 5% often.

    My favorite feature in Dark Sky is simply to touch the daily temperature at the top the screen. Every reading for current conditions appear on one screen – no scrolling required: Temp, Feels like, Wind, Cloud Cover, Visibility, Humidity, Dew Point, Pressure, Precipitation Probability, UV index, Sunset, Moon.

    I am anxious to see iOS 16. I’ve been trying to get it on my Xs since September 12, but can only receive version 15.7. The press release said it was available for iPhone 8 and newer. So, I’ won’t make too much noise until I see what they truly use from Dark Sky. Believing for the best!

  9. When comparing weather apps, location permissions come into play:

    If you want it to be able to issue warnings, you need to set it to be able to ‘Allow Location Access’ flag to ‘Always’ and to turn the ‘Precise Location’ switch on. That should give the best information that the app is capable of.

  10. Yes, weather is my main app. And it’s utterly unreliable lately. It said it’s cloudy. It says it’s going to be cloudy all day. But it’s obviously raining outside. Or it says it will remain sunny all do so I go out without an umbrella and 10 minutes later there’s a rainstorm.

    It obviously can’t cope with Tokyo weather this season. It’s hard to believe any weather app. Weather Bug is no better.

    This was from yesterday - in the middle of a rainstorm.

    IMG_9566

  11. Is the forecast in Weather hyperlocal like Dark Sky or does it just use the nearest official station? It does show my home’s location on the map but it isn’t clear from looking at the weather values that it is here or where since it just shows my location as North Fort Myers.

  12. I’m about to go out. The weather app says it’s raining. But it’s bright outside. Taking an umbrella just in case.

    IMG_9572

  13. When it comes to precipitation notifications, DarkSky really hasn’t been all that reliable for me in New England. This evening has been a perfect example.

    What I did love about DarkSky compared with other third party apps was the one-time payment for the app with no annual subscription, but I don’t really fault the weather apps for charging a subscription fee - weather data isn’t free. And, for what it’s worth, the stock weather app is good enough. In fact, its precipitation notifications have been more reliable than DarkSky since I turned them on. I may decide not to stop the renewal of Carrot before my subscription expires and just go with the stock app and see how that goes for a few months.

  14. One feature seemingly missing from Apple’s Weather app is DarkSky’s Time Machine, where you can look up the past weather at your location. I agree with Adam’s assessment of the cards being difficult to parse. I decided, eventually, that they were adequate, but I still prefer DarkSky’s interface on all counts.

    For my area, Forecast Advisor ranks The Weather Channel first, but when Apple’s Weather app was using it as their primary data source, it often felt incorrect whereas DarkSky was generally more on the money. Guess time will tell.

  15. I’ve been a fan of Wunderground since it first went live. I’m still subscribing to their premium service. I’ve also been using DarkSky for several years to get its hyper-local forecasts. I was saddened/happied when I saw that Apple had bought them. I’m happier now that more DarkSky features are appearing in Apple Weather. I’m not pleased that the DarkSky UI is more informative and simpler to use than Apple’s. I suppose Apple has to assert its UI superiority, which reminds me of a Heinlein quote (Jubal Harshaw) that story editors like the flavor better after they’ve peed in it.

    Being in wildfire territory (Northern California), I suggested to DarkSky that they should try modeling and forecasting air quality. They responded positively, but then got acquired by Tim & Co.

    I’m trying to replace my DarkSky habit with Apple Weather, and continue to use Wunderground for longer range forecasts (they do 2 weeks, FWIW) and it’s information-dense charts.

  16. Actually, weather data in the United States is free. The National Weather Service provides access to all surface, upper air, radar, satellite, and numerical weather prediction data for free to anyone that wants it.

  17. Whoever is charging you for National Weather Service data is ripping you off!

  18. Well, I guess the rain was from one of those “mostly cloudy” clouds that parked over you! :grinning:
    I’ve had it pouring rain over my house, yet a couple of streets over is enjoying a sunny blue sky and vice versa. Weather does weird things all of the time; as the saying goes: “Don’t like the weather? Wait a few minutes.”

  19. I guess the fact a typhoon is passing by adds to it.

  20. I rely mostly on NOAA. There is a huge amount of data available, including hourly histories, radar, etc. I can pinpoint my exact location and forecast and current temps/winds are adjusted accurately. I also rely heavily on radar apps (Weather Radar and MyRadar) to track rain/snow activity — I can tell instantly if rain is imminent at my house and for how long. That’s important living in the mountainous boonies where access in/out is affected by weather. Best of all, as noted, the NOAA data is free. The radar apps are nominal.

  21. Hi Adam. I am in one of those tricky weather areas in Scotland where the changes over five or six miles are significant and Apple Weather was never reliable. I use weather apps for hill climbing and for construction sites. Dark Sky was better but I find I need to use two apps to gauge what’s really going on. These are Yr.no ‎Yr.no on the App Store and WeatherPro but I do not subscribe to a two week forecast as the reliability rarely extends beyond three or four days. If you like to have free or low cost apps, these work well here. atb Drew

  22. Long been a fan of Carrot, although I can only deal with mildl snarky lol :stuck_out_tongue_closed_eyes:

  23. In the UK, Apple’s Weather app is pretty useless. It doesn’t seem to be granular enough for our country of a thousand micro-climates. My recommendations for here are the free BBC Weather or Met Office apps. On my occasional visits to Switzerland, I have found the same and recommend the free MeteoSwiss app. Maybe it is expecting too much for one weather app to work well all over the world?

  24. There’s essentially two sorts of weather forecasting app - those that serve forecasts, generally using data from national agencies worldwide, and those that supplement those forecasts with rain radar prediction to provide what Adam calls hyperlocal forecasting. The latter are very good at alerting you to rain within a couple of kilometres or so, but can only serve this up for around 3-6 hours ahead.
    There’s hundreds of examples of both varieties - testament to the availability of the data, so it really comes down to an interface and presentation of data that works for you.
    Personally, I use Weatherplaza, which I find generally very reliable and gives rain alerts about an hour ahead.
    But remember all weather apps are only as good as the data they access and the algorithms interpreting those data. None are 100% accurate.

  25. Apple’s Weather includes a “Feels Like” prediction, but what exactly is that? Is it dew point? I’ve always found dew point to be the most reliable indicator of how I’m going to feel. But digging around out there I could find no info on how “Feels Like” is calculated.

  26. So what are the privacy track records of these weather apps? I seem to remember a minor furor over just how much data some of these apps were selling to third parties. I guess I trust Apple on this more than I trust, say, Weather Channel, but I’m not sure even Apple’s hands are clean on this.

  27. I use MyRader Pro as my goto weather app. I loved Dark Sky but alas, no more I guess.
    Anyhow, MyRadar and looks much like the Carrot interface with all the color enhancement. Carrot is a little expensive, and I don’t think they have learned how to share their license across iPhone, and Apple TV yet. I love weather apps, but I don’t want to pay by platform. The new Apple Weather is chock full of features, but looks a little drab to me. My 2 cents.

  28. Feels like should be a combination of humidity, temp and dew point.

    Diane

  29. But dew point already is a combination of humidity and temperature.

  30. Dew point is temp at which air (at same barometric pressure) becomes saturated with water vapor (i.e. 100% humidity). Assuming pressure constant, it can be calculated by current temp and humidity.

    Feels like temp (“Heat Index”) is an attempt to adjust temp to account for the fact that we feel hotter in humid weather due to our evaporative cooling (perspiration). Note, perspiration has more to do with the wet-bulb temperature than dew point actually. And wet bulb, in a sense, is the counterpart of dew point. Instead of asking how low temps could get before humidity condensates, you’re asking how low something can be cooled by evaporating moisture into ambient air. If humidity reaches 100%, you can no longer evaporate and hence you can’t cool below ambient.

  31. FWIW, I’ve always liked the forecasts from AccuWeather.

    I deleted their mobile app year ago because they were tracking all kinds of data unrelated to weather (e.g. browsing history), probably for their advertisers. Apple’s privacy report indicates that they are no longer doing this, but I haven’t bothered reinstalling it, since the experience I get by just bookmarking the web site is good enough for me.

    In other words, when the temperature drops below the dew point, water will condense out of the air. In other words, dew (or frost, if it’s cold) will form on surfaces. Which is a good way to remember what the term means (and probably how it was coined).

  32. It’s amusing how the different weather apps disagree over forecasts for rural areas such as the one where I live. I find that for local weather the most accurate one is Weather Underground, which is free. You can search and find local (amateur) electronic weather stations, within a mile or three no matter where you are, that have locked into the weather underground network, and that report temperature, humidity, rainfall, wind, etc. to a degree of accuracy unmatched by other apps. Their forecasts are borrowed from The Weather Channel, and are more accurate than most. For storms and hyperlocal forecasts, Storm Radar is the most accurate in my rural area of New England. I’ll try Apple’s new Weather app to see if it’s any better than the others. Until now I haven’t much used it.

  33. Dew point is reported on the humidity card. If you tap the card to see the forecast graph and pick any point on it, the dew point associated with that point will also be reported.

    It would be nice if dew point got its own card, or if it appeared as a 2nd scale and line on the humidity graph.

  34. Excellent – good to know! Thanks for the tip.

  35. @heronlen: Do you have an App Store link for this? I can’t seem to track it down in the US store. Lots of other BBC apps, but haven’t found Weather.

    Thanks.

  36. liz

    I feel like the Apple Weather app in iOS 16 is missing the “Percentage chance of precipitation” by hour that the Dark Sky one has. Am I missing something? You used to be able to see “oh, ok 75% chance of light rain at 2 PM and 5% chance at 10 AM. Guess I’ll go for hike at 10.” But in Weather now, it shows you a graph representing amount of rain at a particular hour and you can swipe around to see it go from 0" to .05" to .25" etc. There’s no “likelihood” or “percentage” of that rainfall happening. Does that mean it’s 100% likely we’ll get .05" rain at 2 PM? Maybe I am missing the % likelihood somewhere in the new interface? HOPING SO!

    Also now we have to know if a certain amount of rain is a lot? Dark Sky tells you “light rain” or “heavy rain” etc so you don’t have to be a weather nerd to know if .05 is a lot or not.

  37. This is exactly why I’ve not acquired some of recommended apps (e.g. Met Office, Wundergound, Weather Channel, etc.). Weather apps are one of the few that can accurately track you everywhere you go, especially if you have Always and Precise for location settings.

    I’m with you in that I guess I trust Apple more than most developers, but I’ve looked at the privacy policy on most of these and have passed on them.

    I’m currently using Yr.no, Hello Weather and Forecast Bar because I think their privacy policies are reasonable.

    I’m fully willing to pay a subscription fee to the developer for access to ad-free weather, but, in return, my data (location, etc.) is not for them to sell. I suspect that some of the issue regarding privacy is that “race to the bottom” for subscriptions to be as cheap as possible. A developer needs to meet their costs and selling location and other data is one way to do that if the subscription doesn’t cover them.

  38. You do get the chance in the hourly forecast. When you expand the precipitation box, you get a graph showing the amount expected in various time segments.

  39. This is what I am seeing in the weather app right now for my location (I’ve cropped out the location itself just above.) That top bar scrolls to the right for a full 24 hours.

    If you mean that after you tap that bar there is no % chance of precipitation - that’s correct. But it’s right there on the opening screen.

    I’d say the height of the bars in the rainfall part of the graph show how heavy the rain will be. The higher the bar, the heavier the rain. But, again, where I live, those predictions are generally useless. Will it rain or not - pretty useful. Will it rain heavily at 4 pm and lighter at 4:15 - not so much.

  40. Typically that’s what feels like means…sort of a combo of wet bulb, dry bulb, and wind chill but it isn’t much different than dew point most of the time I elieve.

  41. You’re showing Apple Watch complication screenshots—does the iPhone app agree with them? I could imagine the complications not updating for a variety of reasons.

    Some of them are really problematic. I stopped using WeatherBug for that reason; I don’t know if they’ve improved since.

    CARROT Premium has a more expensive family subscription for sharing to other people in your family. I don’t think there’s any connection to the Apple TV.

  42. liz

    Thanks. Thing is that in Dark Sky, it actually gives you percentage chance of precipitation across the hours. While it might show you 30% chance of rain in the Weather app as you show in the screenshot, Doug, that’s an average across the day. In Dark Sky it will show you that it’s 10% at 9 AM and 90% at 4 PM etc.

  43. liz

    Thank you for this. It’s helpful. Is there a way to see the chance for rain on a following day? For example, mine shows in NYC on this Thursday 50% chance of rain. But when I go in and look at that day, I can’t see if that’s predicted in the AM or PM etc. I can on Dark Sky. So maybe I just don’t know how to find it.

  44. After you click on the day, you’ll see a temperature graph for the day, In the upper right, you’ll see a thermometer icon with the carat indicating a drop-down menu. Tap that and then the Precipitation item. That should show you a graph indicating the relative amount of rain expected during each hour of the day.

  45. Dew point is never greater than the temperature, but it can be much lower. In hot, humid summer weather, “Feels Like” is higher than the temperature.

  46. I think this very much depends on what “feels like” means or rather what has been implemented. If it’s just heat index, you would expect temps above the dry bulb. If it’s wind chill, however, you would expect temps below the dry bulb. I would expect that a reasonable feels like display incorporates both, whenever conditions are met: that is HI when dry bulb >80F or humidity >40%, wind chill when dry bulb <50F or wind speed >3 mph.

  47. No, from that photo it’s 30% at 5 pm and 30% at 6pm. (Looking right now it isn’t listing a chance for “Now”, 30% at 5pm [which I assume is the hour from 5 to 6, as I am minutes away from 5 pm right now], 80% at 6pm, nothing at 7 pm, 30% at 8 pm. The rest if the hours as I scroll across show no percentage chance.)

    Here are a couple of photos as I scroll a bit to the right in the second photo, taken just now. I was in a slightly different location earlier today BTW.

    (Not sure why I’m hiding the location, as people could probably figure it out close enough based on the sunset time…)

  48. I tapped the day for Thursday and then changed the drop down control (I’ve added the red arrow to show you where it is) to rainfall. That shows the hours when it’s expected to rain and how much rain is expected at that time on that day.

  49. Hi Adam. The Apple Watch complications don’t always agree, but in these two cases it did. The weather was highly changeable though the last few days because of the passing typhoon. And I suspect the Weather app just can’t figure out my location closely enough, or the closest data center is too separated from where I actually am.

    Yesterday it’s good I decided to take an umbrella with me because the weather turned torrential suddenly!

  50. Untrue. When the temperature is at at or below the dew point, you get DEW.

  51. Dewpoint is never greater than the temperature. If the temperature falls and reaches the dewpoint then both will fall in tandem – with dew or frost forming.

    With respect to Wikipedia, their explanation is poorly written and can lead to confusion.

    I’m a retired meteorologist – trust me. :grinning:

    Edit: part of my response was incorrectly typed in the quoted material.

  52. Following up on my privacy post…there’s a new weather app which apparently collects nothing: Mercury Weather. It was linked on Daring Fireball this morning along with Forecast Advisor.

    As a devotee of the dear departed WeatherLine, I like the look of Mercury Weather. It does require iOS 16.

  53. I once used Weather Underground until I heard that IBM was using this to track people for advertising purposes after purchasing the service. And if you look at the App Privacy report for the app - it looks pretty ugly compared with some other apps, like Carrot, Hello Weather, etc. Sometimes I hear about new weather apps, but I always check the app privacy report before I even think of installing it.

    As for Apple, this is what they say about weather and privacy: Legal - Weather & Privacy - Apple

  54. On my watch I have Apple weather in Celsius and Dark Sky in Fahrenheit. After upgrading my iphone to ios16 both show a range. Apple weather shows the temperature but dark sky just hows a range. (the range low - high is new since iOS 16). Dark sky now longer shows the temperature. I can’t find another temperature compilation for my watch. Any ideas?

  55. Do you have any links to these discussions? Gruber and @mjtsai mentioned it after it came up here, so I’d be curious to hear what the criticisms were.

  56. I’m really going to miss Dark Sky. It’s been my go-to weather app for a long time. The interface is perfect for me. The iOS 16 Weather app is fine, I guess, but it’s just a little harder for me to use than Dark Sky.
    I’ve also tried WeatherBug, which can give you a lot of useful info, but the interface isn’t great.
    And I’ve tried Hello Weather, and may go back and give that another try soon.


  57. This information density of Weatherstrip can’t be beat. As a bicyclist, I am concerned about wind speed and direction, cloud cover, UV, humidity, sunrise/sunset, as well as temp and precipitation. Weatherstrip gives you all of this as an hourly forecast that you can side-scroll through for up to a week.
  58. This is really interesting. I did download it but the wind can’t be depicted on that image unless you scroll I think?

    I feel like us cyclists have a set of requirements that many others don’t! My most important one is wind, especially as we head into winter.

    Diane

  59. Now that I’m a recreational cyclist I’m interested in the same things, Diane. :smiling_face: But when I rode to work here in Boston I pretty much ignored all of them, except when snow, snow melt and freezing temperatures iced up the bike paths. When that began I took the winter off, usually until mid-March.

    That didn’t mean I didn’t try to avoid getting wet, at least until I’d warmed up a bit, and the only thing that reliably showed me how was the animated weather radar I pulled up on my desk PC just before I went out the door. No app, no matter how “localized” can do that, and we shouldn’t expect them to.

  60. In the settings for Weather Strip, you have to change the windsocks to arrows.

  61. Accuweather has what they call “Minute Cast”, which is essentially a graphic version of local radar. (That I usually call them “Inaccuweather” is an indication of how reliable I find it.) There’s a ConnectIQ widget for both Garmin Edges and Garmin watches produced by Accuweather that can show you the Minute Cast on your device, provide you have some connection to the Internet via your phone. There’s also a ConnectIQ field called “Rain & Wind” that I use which shows you sort of real time temperature and relative wind (takes into account your direction of motion relative to the wind) which seems to update the wind vector and temperature every 20 minutes or so (again through your phone). Several times this year Accuweather has said “No precipitation for the next 120 minutes” only to have that prove very wrong.

  62. As I type, Minute Cast thinks it’s raining in my area code. It probably is, somewhere. At the same time Wunderground’s radar feed accurately shows, in real time, the small dry lobe over my house.

  63. elf

    I use Windy. It’s pretty exhaustive. I sometimes find I have to check it against Sailflow, which gives me more detailed wind info I need when on the water. Weather doesn’t seem to say anything about moon behavior, and I need that for photography. I also use a Tide app for detailed info about local waters and TPE for precise info about sun/moon behavior. I haven’t used the apple Map since my first iPhone, use Google, and TPE, zooming in on Windy gives me a map as well. Windy also gives me tide info with a handy little graphic on the map which links to tide tables. Its 7-10 day weather plays with a tap of the arrow and I can watch Ian approach Englewood, FL as the weather vid moves the wind.

  64. I will have to check out Windy. I downloaded Carrot and Weather Strip, and last night found Motorcycle Weather.

    Motorcycle Weather only shows one day at a time for details in the free version but it’s pretty much what I wanted for wind. I can see a summary in a list for a week at a glance. i.e. tomorrow says 0% chance of ran with 7mph winds. If I click on it I get the hour by hour also with wind direction. It’s a really simple chart but easy to read at a glance.

    Between Carrot and Weather Strip, Carrot is kind of winning with the wind display.

    I have a tides app too, I use it to check to make sure one of my routes isn’t under water in the spring.

    I will wait until I upgrade to iOS 16 before making a monetary commitment.

    Diane

  65. For those like me who are interested in hyperlocal weather patterns

    imho, this ain’t terribly hyperlocal …

    so far haven’t found anything to toggle that’ll make it hyperlocal or even vaguely close. it does have my location so that’s a win …

  66. Already in iOS 15 I noticed that if you zoom in on those precipitation forecasts their time range becomes very limited. And if I want a 12-hr forecast, I need to zoom to a level that exceeds all of CA. :man_facepalming: I was hoping that would improve in iOS 16 weather considering all that DarkSky knowhow Apple bought.

  67. Do you mean ios16? Yes, I noticed that there was something wrong with the precipitation forecasts that day (it was limited to only the current day), but it seems to be fine for me now.

  68. No. 15.7 at the moment. The issue of short forecast timespan when looking at a more limited geographical area has certainly been observed here since 15.0 though.

  69. Yep, that’s different in ios16. The hourly precip forecast seems to go for ~10 days. As well as wind, temp, uv index, feels-like temp, humidity, visibility, and pressure.

  70. The use of ‘percentage chance of precipitation’ was one of the reasons I could never use Dark Sky for anything more than the next hour precipitation graph. Like @haver I use Yr.no as it’s the most accurate in Scotland, but I also really like the way it displays predicted quantity of precipitation (similar to what Apple’s Weather app is now doing). I always find percentage a bit useless for precipitation because I don’t understand what I’m supposed to take away from it:

    • Does 30% chance mean that there’s a 30% of any rain in the hour, even if it lasts for 15 seconds?
    • Or does every minute have an independent 30% chance (in which case you’d expect an average of 18 min rain in an hour)?
    • And is it very light so you’re ok for the 18 minutes (or 15s?) or a heavy downpour so you need full gear even for 5 minutes?

    Dark Sky’s ‘next hour’ graph is brilliant at showing the quantity of precipitation. And a list of mm rain per hour tells me a lot more than a list of percentages per hour in terms of how I should dress to be outside.

    I say all this not because anyone is ‘wrong’ to want the percentages, but to give a perspective that might help in using the precipitation quantity display. And also in case anyone can enlighten me on what the percentages actually mean!

  71. There is an official answer to this.

    https://www.weather.gov/lmk/pops

    Another description that’s been given by the meteorologists on TV here is that they take the percentage confidence that there will be precipitation of at least 0.01” in the region at all (say, 50%) during that hour (and it’s at some point that hour, not the full hour) and the chance it will be in that exact location if it does precipitate (say, 80%) and multiply them to get the chance for that area (so, 40% chance of at least 0.01” during the time period they are making a prediction for).

    The stock weather app does the same thing really, just in a different sort of graph. Here’s what the app says for my location for tomorrow.

    At the bottom it tells me that it’s expecting 0.15” for the full day, with graphs above showing me that it’s going to be no more that 0.03” for any hour (and I can look at the X axis for the time and the y axis for the amount for that hour, and the bars show the relative amount. But honestly the X axis is poorly designed - which hour is which exactly? Hard to say.)

    Here’s Dark Sky for the same day. (Ive tapped on precip rate in inches/hr). Not quite the same, but pretty close. (To me telling me that you think that it will rain 0.15” or 0.4” for a full day is basically telling me the same thing - it’s going to be very light rain.)

    I do think that Dark Sky was pretty well designed to pack a lot of information into one screen.

  72. Rain probability doesn’t mean any of those necessarily…but then it never has. It’s just a general guide between 0% and 100% for whatever area it’s being forecast for, which in Dark Sky’s case is your individual location. I’ve never figured out exactly how it figures out the probability at my house as opposed to a mile down the road, but I have checked here and then when I was a mile down the road and it’s not identical so there’s some sort of localization thing going on. Wikipedia says that rain probability is the likelihood of > 0.01 inches will fall in a single spot averaged over the entire forecast area…and with the size of Dark Sky’s forecast area unknown it’s really just a crapshoot…which is why I think of it as a general guide sort of thing.

    As a recovering engineer…I’ve always found the nebulous nature of rain probability only somewhat useful.

  73. It’s fascinating to read this thread, and the see obvious interest expressed in the various nuances of weather reporting. It makes me wonder, though, about accuracy. Do people find the increasingly granular predictions of precipitation and wind actually reliable and useful? My curiosity/ignorance is I’m sure fueled by the fact that I live in Southern California, where all we’ve really needed was the blanket forecast issued by the one-hit-wonder Albert Hammond in 1972.

    :smiley:

  74. I’ve found that down here in SW FL you can be just a few miles apart and have different weather. I can’t count the number of times that the weather guessers forecast winds and torrential rain from SW (that’s the prevailing wind direction here) and we get a bit of wind and no rain while 3 or 4 miles north and south it rains heavily. Something about the way the coastline or islands or whatever affects the flow I guess…but you can clearly see it on the radar and Dark Sky has always been pretty good at it.

  75. So somewhat on topic, this supposedly true story relayed to me recently:
    When a meteorologist was asked how they calculated the likelihood of precipitation, they responded that there are five meteorologists in the office and if three of them think it will rain, then that’s a 60% chance of rain.

    Not sure how the apps do the calculation, but if they are using multiple sources of data, maybe it’s the same?

    Cheers,
    Jon

Join the discussion in the TidBITS Discourse forum

Participants

Avatar for ace Avatar for Simon Avatar for joeholmes Avatar for aforkosh Avatar for schinder Avatar for neil1 Avatar for ddmiller Avatar for steved Avatar for jonathan2 Avatar for dianed143 Avatar for gwhiesterman Avatar for doug2 Avatar for fischej Avatar for jajvj1 Avatar for plossl Avatar for heronlen Avatar for jtummins Avatar for mmcneil Avatar for Shamino Avatar for David_L Avatar for jonathan4 Avatar for liz2 Avatar for Ratsima Avatar for david_blanchard