A question for NOAA and/or V2 people: Radiosonde launches, standard and auxillary

[Darren Addy]

I was hoping that someone "in the know" might reply to this set of questions regarding radiosonde launches in the U.S. Partly, I suspect that some of the information that I am finding via Google may not be up-to-date, so I was wondering if anyone could clarify a few things.

First I'm wondering about the radiosondes themselves.

The current radiosonde tracking systems are 1950's vintage and the data processing computer is a 1980's IBM PC/XT. These systems are obsolete and are increasingly difficult to maintain. NWS has begun a program to replace the ground systems at all NWS stations with a new GPS radiosonde system.

- source: http://www.ua.nws.noaa.gov/factsheet.htm

Is this replacement process completed or still ongoing? Is the GPS used only in mapping the data received, or is it also possibly useful in the recovery of the radiosonde itself? (If helpful in recovery, would it be possible to crowd-source the recovery and reuse of radiosondes by making the process available to the public?) Is the data from a GPS-equipped radiosonde superior in some way to that from the older radiosonde's?

Currently, 70 radiosonde stations are distributed across the continental United States. Radiosondes are launched from these stations twice daily, just prior to 0000 and 1200 UTC.

Is the number of stations still correct?

I notice that sometimes, local weather station(s) will launch an additional radiosonde - usually in mid-afternoon on a potentially volatile day. I was wondering about the procedure by which that is done: Is it at the request of the SPC, or can it be done solely at the discretion of the local NWS? Is there a particular set of information that is being sought, or a particular set of circumstances that leads to these "auxillary" radiosonde launches?

During Vortex 2, I noticed that the SPC would note recent information derived from a V2 launched radiosonde. Obviously, a V2-launched radiosonde was not at a standard launch station or at the standard launch time, but presumably is still providing valuable information to Day 1 forecasters.

In addition, while reading papers like Thomas M. Graziano and Toby N. Carlson's 1987 paper on "A Statistical Evaluation of Lid Strength on Deep Convection" I noticed that (like radar locations) the radiosonde locations can have an effect on the selection of storms being studied for a particular inquiry. For example, the paper above states:

In order to assure a reliable correlation with radiosonde data, only those events lying within a radius of 2.5% latitude (275 km) of a radiosonde station and within 2.5 h of 0000 or 1200 UTC were tabulated.

I'm wondering if someone could describe the relative value of real sounding data, as opposed to model sounding data and whether model sounding data might be improved by a "higher resolution" set of radiosonde data.

Would it not be true, that if there were more radiosonde launching stations, or more radiosonde launches per day that there would be a larger set of events from which to draw datapoints in research? (fewer gaps in the location of events that could be included in a particular study)?

I realize that the 0000 and 1200 UTC launches are used worldwide to get a synchronous snapshot of the upper air, but would there be local forecasting or research benefits to (for example) doing a set of 1800 UTC launch of radiosondes across the part of the country where severe storms were a possibility (for example, in the Spring?). Or might there be a local forecasting or research benefit to creating a mobile radiosonde unit, similar to a mobile mesonet (but obviously more spread out) that could be deployed around any particular geographical area of interest that would supplement the "nailed down" station data?

In a time when budget cuts are being considered, I realize that very little of this will be possible with existing funding, but it seems to me that if grants can be given to huge projects like Vortex 2, that science grants might also be written/given to improve the resolution of our upper air data, at least during "Prime Time" tornado season.

If higher resolution (either in space or in time) upper air data is not worthwhile or necessary, I'm wondering why it is not.

 

[Chris Nuttall]

Yes, the transition to a GPS-tracking system for the UA flights is the replacement system you are hearing about. It's officially called the Radiosonde Replacement System (RRS). I'm not sure, but I don't think the transition is complete, yet. A few sites are still running the older system. The GPS tracking is used to calculate the wind data. Typically battery life is only 3-4 hours at most, and a typical balloon flight in about 2 hours in duration, so the GPS is not used to recover the instrument. Approx. 70 stations in the CONUS (mostly NWS) perform the twice-a-day UA obs. There are some special sites (including military), but I'm not counting those. Including OCONUS sites, I think 92 out of the 122 NWS offices conduct UA observations.

Sometimes special observations are taken, usually for severe weather purposes, when it's determined that updated data are needed. These requests often come from NCEP, SPC, and NHC. When a tropical system is going to make landfall, it's common for stations in the region to perform UA flights every 6 hours for a couple of days. When Dolly and Ike made landfall a few years ago, I remember having to do those flights in AMA because NHC needed to see the strength of the upper ridge to forecasting the tracks. Sometimes the decision is made within the office for severe weather or winter weather purposes when an outside request hasn't already been made. Special flights are also used for testing purposes on rare occasions.

As far as the value of real sounding to model sounding data, an actual observation is better than a simulated/derived one based on model output. Think of a tornado-producing supercell along a dryline just east of the state line in W OK. AMA is behind the dryline and in a much different environment. OUN might be ahead of the dryline, but is capped and clouded over by cirrus. A sounding taken just in front of that dryline would be much more representative of the environment that the supercell developed in. Would more locations improve the quality of model soundings. Perhaps. Model soundings are just another form of model output. It's just like looking at a forecast of 500mb heights. If more real-time data could be ingested into the model for its initialization, then the output should be improved. These effects can sometimes be observed in the initializations from the 06z/18z model runs that do not ingest the newest UA obs like the 00z/12z runs.

It's always good to have more data, but it all comes down to resources. Yes, there is the money factor. Equipment, manpower for performing the obs, manpower for keeping the equipment operating....these are all things that cost money. Availability of resources is also a factor. Assuming one station performs 2 flights every day in a year, that's 730 flights at that station. Just for the stations in the CONUS, that's about 50,000 flights/year.

There was a really interesting presentation at a severe weather conference in Lubbock in early 2010 that talked about a low-cost, deployable UA system that could be moved around the country to supplement the current observing sites. Some examples were also shown about how a weather system could be "followed" coast-to-coast with these supplemental observations. I wish I could remember who gave the talk. You might be able to Google it and find something.