-4

When using a cool PHP Library I find to get NOAA weather (https://github.com/amwhalen/noaa), I found that the results it returned differed from what I found on weather.gov. I found the weather.gov XML by simply clicking the button on the the page for my location.

They differ by a few degrees, have different date formats, and one shows the LOW before the HIGH (as opposed to HIGH before LOW). Which is more accurate: NOAA API or Weather.gov? Should I use a different library? I would use The Weather Channel (also Weather Underground according to their API), but forecasts more than 3 days costs money.

URLs:

  • PHP NOAA (ndfd): 
    http://graphical.weather.gov/xml/sample_products/browser_interface/ndfdBrowserClientByDay.php?lat=LATITUDEHERE&lon=LONGITUDEHERE&startDate=2013-10-21&numDays=7&format=24+hourly&Unit=e
  • Weather.gov ("MapClick"): 
    http://forecast.weather.gov/MapClick.php?lat=LATITUDEHERE&lon=LONGITUDEHERE&unit=0&lg=english&FcstType=dwml

UPDATE: Thank you Sammitch and Jeroen! What you say would have to be true. The results will vary. It doesn't make sense how they are both technically NOAA but have different results, but whatever (you would think that they would be using the same predictions and algorithms). I'm going to go with the direct NOAA for now since there are quite a few libraries and lots of documentation. I found another awesome PHP implementation here: Simple NWS.

DaAwesomeP
  • 585
  • 5
  • 18
  • 2
    You have to decide which one you want to use. Asking which is more accurate is like asking us if you should eat turkey or chicken. – Steve Oct 21 '13 at 22:50
  • 3
    I'd go for the chicken. – phil-lavin Oct 21 '13 at 22:57
  • 2
    The two organizations most likely use different models for predicting the weather, and weather prediction is most definitely *not* an exact science. Just assume they're equally accurate *within tolerance*, flip a coin, and go with it. – Sammitch Oct 21 '13 at 23:00
  • I'm not asking for an opinion. **Which one should I trust for the most accurate data?** That's what I want to know. In some cases I saw it differ by 5 degrees (Fahrenheit). – DaAwesomeP Oct 21 '13 at 23:04

1 Answers1

3

Weather predication is not an exact science. Depending on the model used, it can yield different results. One time one will be correct, the other time the other.

Best thing is to just to pick the one that is easiest to implement and which has the best documentation. It will save you time and frustration in the future.

Jeroen
  • 982
  • 1
  • 6
  • 14
  • 1
    What confused me is how the two sites have different results. I looked at the results the library was giving and then NOAA (which redirected to weather.gov). They were not often the same so I looked at the XML that was cached and the one on weather.gov and saw huge differences (of course the URLs were different too). I thought that they were two completely different sites (i.e. CNN and ABC News). Now I see that I should just stick with what I have. If one says it's raining dogs and cats, and the other not, I'll stick to the one that's true. – DaAwesomeP Oct 22 '13 at 03:30