We’re currently researching conventions and best practices for weather data feeds for a syndication project. I came across some good sources that I wanted to share and put in one place, in case others are working on similar projects.

There is no widely accepted common standard on the web for weather data. So while languages have rich ways of talking about the weather, it’s less clear how machines do that. We wanted to flush that out more here by posting the resources we’ve found so far:

• Yahoo has its own weather XML namespace for forecast RSS feeds, like what you see here. The namespace defines seven elements (location, units, wind, atmosphere, astronomy, condition, and forecast) that structure meteorological data.

• The NOAA’s National Digital Forecast Database (NDFD) offers an XML based web service that defines more than 40 aspects of weather conditions and forecasts.

• METAR is a format for weather reporting in aviation. The format goes back to the 1960’s, and that’s pretty apparent in how METAR communicates forecasts. Check out its report for Washington, DC’s National Airport: KDCA 091852Z 25006G18KT 10SM SCT055 03/M09 A3016 RMK AO2 SLP213 T00281089. You can try it yourself on the METAR Data Access website.

• Another one is WeatherML, however most of the information I found on it was through news stories from 2000 that claim it would be the global standard protocol for weather data.

We’re actively working on a weather data syndication project, so if anyone has good resource links please post them in the comments.