All I want is an accurate temperature and wind chill factor
November 8, 2010 7:18 AM Subscribe
On the current Weather Underground "all weather stations" widget the variations look excessive. Can you help me find a HOWTO on quality control of weather observations and how to find the best data?
Temp is from 57 -71 degrees fehrenheit; wind from 0-12; oldest update is 8 minutes.
It is calm and perfectly clear outside and there is no topography in this geographical region (Houston, TX); the only conceivable earthly variable is nearness to large water bodies and large concrete bodies. If you just throw out the high and low numbers, they are still all over the place. It's almost enough to make me distrust all of it
Temp is from 57 -71 degrees fehrenheit; wind from 0-12; oldest update is 8 minutes.
It is calm and perfectly clear outside and there is no topography in this geographical region (Houston, TX); the only conceivable earthly variable is nearness to large water bodies and large concrete bodies. If you just throw out the high and low numbers, they are still all over the place. It's almost enough to make me distrust all of it
...and as a backup for NOAA, I tend to favor Accuweather. See also this great sidebarred comment re: weather forecasting.
posted by jquinby at 7:42 AM on November 8, 2010
posted by jquinby at 7:42 AM on November 8, 2010
There are two main factors here: accuracy of the weather station and age of the observation.
Official airport observations are the most accurate, but they can be up to an hour old. A quick way to spot these on wunderground is to look for a four-letter identifier, starting with "K" in the US, "C" in Canada, or something else in other countries.
If you are in an area with an official local mesonet, those stations will give good readings, too, and they tend to update every few minutes. These are a bit harder to identify, but sometimes they'll have their own tab on the wunderground weather stations page.
Personal weather stations can be all over the place - but after a few days you can start to get a feel for which ones are accurate and which ones need to be recalibrated.
posted by penguinicity at 7:43 AM on November 8, 2010 [2 favorites]
Official airport observations are the most accurate, but they can be up to an hour old. A quick way to spot these on wunderground is to look for a four-letter identifier, starting with "K" in the US, "C" in Canada, or something else in other countries.
If you are in an area with an official local mesonet, those stations will give good readings, too, and they tend to update every few minutes. These are a bit harder to identify, but sometimes they'll have their own tab on the wunderground weather stations page.
Personal weather stations can be all over the place - but after a few days you can start to get a feel for which ones are accurate and which ones need to be recalibrated.
posted by penguinicity at 7:43 AM on November 8, 2010 [2 favorites]
surfacestations.org has some interesting analysis of weather station data.
posted by alikins at 8:22 AM on November 8, 2010 [1 favorite]
posted by alikins at 8:22 AM on November 8, 2010 [1 favorite]
I should note, I have no idea if the surfacestations.org site has external motivations or not. I didn't see anything obvious to make me think so, but then I wasn't really looking. I just thought the basic idea sounded interesting in a nerdy "lets compulsively collect data about something" way.
posted by alikins at 8:41 AM on November 8, 2010
posted by alikins at 8:41 AM on November 8, 2010
Penguinicity is correct, the official airport observations will tend to be the most accurate but only report once an hour. You can call a phone number to get an automated voice telling you the conditions up to the minute like pilots do, but the temperature would be in Celsius.
And the conditions you describe (clear and calm) are the most likely to bring wide variations in surface temperature. Even small changes in elevation, or more likely in this case, differences in how high the temperature sensor is placed (airport sites are standardized to a certain height, but private sites are not) can make a difference. Proximity to water will have a huge effect, keeping nighttime readings warmer than other sites and daytime temps cooler. Concrete holds heat well both day and night, so any ob site that is near a bunch of it will tend to be warmer than others.
posted by weathergal at 8:46 PM on November 8, 2010
And the conditions you describe (clear and calm) are the most likely to bring wide variations in surface temperature. Even small changes in elevation, or more likely in this case, differences in how high the temperature sensor is placed (airport sites are standardized to a certain height, but private sites are not) can make a difference. Proximity to water will have a huge effect, keeping nighttime readings warmer than other sites and daytime temps cooler. Concrete holds heat well both day and night, so any ob site that is near a bunch of it will tend to be warmer than others.
posted by weathergal at 8:46 PM on November 8, 2010
This thread is closed to new comments.
posted by doctord at 7:27 AM on November 8, 2010