Slashdot Mirror
Weather Radar Goes Miniature
quackking writes "As reported today in the Boston Globe, the NSF has committed at least $17M to build out a new network of miniature (at least in comparison with today's monsters) weather radars. This is to radar what Beowulf clusters are to the mainframe; the scientists at U Mass Amherst project that eventually a weather radar node will be deployable for under $20K! Now to figure out how to get real-time access to this mesh of sensors and create a really cool screensaver..."
I would imagine that the array would work quite well for several reasons.
First of all, the amount of energy you have to use to send a signal decreases with distance squared. So covering an area with with several small radar stations should keep the working signal in the relatively strong range with less power consumption.
Also, there will be a lot of overlap near the edges of the stations' zones, reducing anomolous readings through error checking. (This would have to be implemented. It wouldn't happen on its own.)
And it also opens the door for the possibility of trasmitting signals from one station to the next instead of always waiting for the ping to come back. That could allow for new methods and better results and more accuracy.
Slashdot Syndrome: the sudden, extreme urge to correct someone in order to validate one's self.
Wow, pretty insightful considering that's basically what's already happening. :) Although the reporting stations aren't every mile. That would be kinda overkill I think. When I was a weatherman (10 year ago), there were somewhere in the neighborhood of 550-600 reporting stations that did hourly weather observations syncronized with ZULU time, aka UTC, aka GMT. Some weather reporting stations are completely automated, but they are limited. There are some things like skycover and accurately representing highly variable conditions that an automated weather station simply can't do. Other than the hourly observations that are taken and disseminated, there are "special" observations that are taken when special conditions are met. I don't have an FMH-B handy, and I can't recall them all off the top of my head, but it's for things like radical changes in wind direction, speed, ceiling height, visibility, thunderstorm activity, etc. Most weather reporting stations are near airports because weather is very important for forcasting flight weather condition. That and a majority of weather stations are USAF or other military. All that data ends up in a system called AWDS (Automated Weather Distribution System) that has 3 super computing "hubs". If I'm not mistaken, two of them are here in the US, and one is in England. Those numbers are then turned into NGM's and GSM's and other Nested Gridded Models that are still not perfect and need corrected slightly by a good forecaster. With that data the 6 hourly forecasts are generated and issued so that local TV Weatherman can steal them and use them. I can remember one time in Deleware, we intentionally put a forecasted high temp for the day 6 degrees too high and watched 3 of the local channels quote it.
For every annoying gentoo user, are three even more annoying anti-gentoo crybabies. Take Yosh from #Gimp for example.
The article is not clear, but it sounds like they may be using a new radar technology called Phased Array. It works like doppler, except that instead of the mechanical motor needed to move the dish to scan the area, it scans the area electronically, completing an entire scan in one minute to Doppler's 5 minutes or 8 minutes.
Another benefit of phased radar array is that the scanning speed can be upgraded with better computer equipment with higher bandwidth and number crunching capabilities.
-Cyc
/.'s 10 Millionth
NOAA provides a great deal of information for free already --- as they are a public service anyway, this only makes sense. It may not contain the precise details that you wish to research, but you can find information about the various forms of data, including ftp-accessible satellite data, at either of the following two sites: NWS Telecommunications Operations Center, or the National Climatic Data Center.
Most likely the #1 Unfunny Meta/Moderator on
This is already being done, at the moment not all states participating have made the data accessible. Here are a few that have.
Oklahome Mesonet
West Texas Mesonet
MesoWest
Note: The Texas Mesonets are particularly interesting during landfall of tropical cyclones!
$17M divided by $20K = 850 sensors. Area of US is 9,629,091 sq km (according to the CIA). That's 7,578,834 sq mi. That's one sensor per 8,916 sq mi (11,328 sq km). That's just over the area of New Jersey.
Now, the sensors you propose, at $200 each, could get you 85,000 sensors for $17M. That's one sensor per 892 sq mi -- a bit smaller than Ocean County, NJ.
Not quite "every couple of miles", but not too bad. Still, I'd have to think the radar might return more data points anyway, sweeping across the landscape as they do. Plus, it couldn't cost much to add a package of temp/humidity/wind/etc. to each radar, which, while not 85,000 locations, wouldn't be bad either. And I'd have to guess those radars can cover at least a New Jersey each.
"A great democracy must be progressive or it will soon cease to be a great democracy." --Theodore Roosevelt
This radar array sounds nice but I think there is a lot of hype in the article. A more low tech solution to collecting rainfall data and other weather data is to use a community of volunteers. There is in Colorado such a community. The Community Collaborative Rain and Hail Studyi (and Snow too) provides an unique way to study weather in Colorado.
The data could be made available to the public through a Web Map Server following the Open GIS Consortium specs OpenGIS.org.
This data could then be incorporated into your own personal Web enabled apps utilizing an open source product like U of MN Mapserver
You can get raw weather data from here:
C .r adar/
http://weather.noaa.gov/pub/SL.us008001/DF.of/D
Different products, different sites...all updated as soon as the data comes in from the radar site. Then you can use the Geo::Nexrad perl code from sourceforge to parse at least the radial products.
In a few months, NOAA will get further into testing of their new DVB satellite delivery and you'll be able to get all of the NOAA data & imagery with a small satellite dish and a DVB card compatible with Linux.