Big disparity Carrot vs The Weather app

I was curious how the two apps compare in predicting the weather. Carrot is using Dark Sky, the weather app The Weather Company.

The Weather app

Carrot Weather

It’s not a comparison between Carrot and Apple Weather is it – it’s a comparison between Dark Sky and The Weather Company? I don’t see a “big disparity” in your report. I think we want to look at weather forecasts as received truth, when in reality forecasts are informed guesses and guidance: do I need to pack a parka and toque for Wednesday or a mackintosh?

1 Like

To really know how they fare in predicting the weather, we’d have to see what the weather actually was on the days they forecast. To my mind what you’re seeing is that the different programs are using different weather services and models. In the short range the predictions are similar, at the mid-range they start to differ.

I’ve wondered which services are being used by Carrot. I believe that Apple is using the weather channel, which is owned now by IBM. IBM has a newish short-term weather model they run on some serious hardware.

Living in New England, and this being the beginning of meteorological spring, this sort of variance is pretty common in my experience. The jet-stream is hard to model, and Atlantic Oscillation is flipping between positive and negative at the moment.

(Weather is one of my hobbies…)

1 Like

You are absolutely right, which is why I noted the source of the weather data in my screenshots. I should have been clearer, however.

Carrot weather does not have the option of selecting The Weather Company for its data. I found it interesting to see the contrast in rain predictions, and to a lesser extent temperature ranges, from different weather sources for the exact same location. Obviously, a weather forecast is just that, a forecast, a probability dependent on the model used—garbage in/garbage out.

For those who desire the most accurate prediction the questions is, “what weather source is the most accurate overtime for my location?” Based on the answer the next question is ”which weather apps use that source?” Once that is determined, one can then select the weather app using that source and the user’s preferred app design and features. This assumes that one places a higher priority on accuracy than design.

1 Like

The thing is, when I check that weather accuracy rating site, there isn’t a huge difference. Foreca is half a percentage better than Weather Channel for me. A point and half over Accuweather. None of them are perfect, and they are all in the ball park most days. At least the big names, the ones I haven’t heard of are mostly dismal.

I have pretty much seen every weather app tell me it’s sunny while it’s raining at some point though. I sometimes compare Forca to Weather Channel, and the reality is somewhere in the middle. None of them are perfect, but most are acceptable.

I am more to the point of using the weather app I like as long as it has common sources.

https://www.forecastadvisor.com

1 Like

@Leeabe51 You know, after reading your response, I realized that the differences are real but may in fact not be that significant. For example, when I check Forecast Advisor, I get this:

The Weather Channel has the highest short and annual accuracy but Foreca and AccuWeather are not that far behind.

Thanks for your response!

1 Like

Unfortunately the answer usually is “look at the sky and compare”
weather models vary from day to day, and tend to be inconsistent even for the same provider.
There are just too many variables to consider.

I usually take the average of 3 providers and that seems to work for me.
(Apple weather, Netatmo and my local weather station)

1 Like

Does anybody know a similar site for the EU?