Thursday, January 20, 2011

Ice is nice

The Norman time series of weather data, specifically wind, was rather strange today. Just after 2:20am the wind began to slow down dropping to 3-4 mph. Funny thing is I remember it being windy at the site around 9am. This is the effect of freezing rain on an anemometer. The propeller didn't freeze altogether as I have seen in Iowa data but the extra mass made it harder to spin. The entire assembly didn't freeze either as the wind direction did change albeit slightly.

The airport in Norman first reported unknown precipitation at 2:11 am at a temperature of 30 and it dropped from there, while the winds stayed in the 15-20mph range throughout the night.

As far as good data goes, I think there is a sonic anemometer but it is made of metal and I wonder how good the wind data would be with a coat of ice on it. Therefore it would be difficult to correct. Human quality control could handle this but I doubt with station to station comparison this time series could be flagged as being questionable. Realize that the rain gauge at the site is not a typical heated tipping bucket so there was no precipitation recorded at the station.

This is not a criticism of the data in any way; rather it highlights the difficulty of making all-weather observations year round even on well funded networks.  This even happens in hurricanes when the power goes out or when water tops the instruments or penetrates the wiring.

Its important to maintain the observations that are good for future work either in weather or climate and hopefully this provides a gut-check. Some observations will sneak in that are bad and therefore it is important to realize that no observation system is perfect or error-free. Sometimes this is why "tricks" or manipulation must be performed because real data has issues and they must be sorted out, flagged, approved or removed.