Wednesday, September 4, 2013

Critical Aircraft Weather Data Unused By NOAA

Knowing the structure of the lower atmosphere is critical for weather forecasting, but  NOAA (National Oceanic and Atmospheric Administration, which includes the National Weather Service and the NOAA labs) is not using one of the most valuable sources of data: from landing and departing commuter and short-haul aircraft.

A number of weather research programs (e.g.,the International H2O Experiment, IHOP) have documented that a better description of low-level atmospheric structures substantially improves forecasts of thunderstorms and other important weather features.  There is no debate about this.   So it is surprising that NOAA is not acquiring an extraordinarily dense real-time data source of such information, a data stream created by private sector investment and government research dollars:  TAMDAR, weather data collected by mainly short-haul and commuter aircraft (for the curious TAMDAR stands for Tropospheric Airborne Meteorological Data Reporting).

 External part of TAMDAR sensor on an aircraft

This blog will tell the story of this unfortunate and counterproductive situation and how the acquisition of TAMDAR data could substantially improve weather prediction over the U.S.

Upper-air observations over the U.S.

Since roughly the end of World War II, the core upper air observations over the United States have been from balloon-launched weather instruments, called radiosondes.  But as shown by the map below, this network is quite sparse, with only around 100 radiosonde locations over the lower 48 states (here in Washington State there are only two:  Quillayute and Spokan).  Even worse, the balloons are launched only twice a day, making this network relatively useless for keeping a tab on local or regional weather features.

 
During the 1970s, World Meteorological Organization (WMO) established the AMDAR program(Aircraft Meteorological DAta Relay) in which wide body planes collected temperature, pressure, and wind data and transmitted it in real-time for weather analysis and forecasting.   This program now provides thousands of observations per day, including ascents and descents from major airports (e.g., Sea-Tac, Portland) as well as information collected while the planes are flying at cruise altitude.  Every time such aircraft land or takes off from an airport, it secures a vertical sounding (similar to a single radiosonde launch).  To get an idea of how much data is produced by AMDAR flights over the U.S., here is sample, for the entire U.S. (colors give you the elevations) for a three-hour period during 1 Sept 2013.  Pretty impressive!

Put look closely, say over the NW (see graphic below), and you see a major issue...the aircraft soundings are only occurring a few major airports:  useful but limited.   And the rest of the time the planes are flying at 30,000 to 40,000 ft, which doesn't help much  in getting critical low-level structures.  Also problematic is the lack of humidity 

information, which is very important for convection forecasting.  Another issue is that the U.S AMDAR data is not made public for general use due to lack of funding by the U.S. government (the airlines are footing most of the bill and don't want to allow everyone to have it until they get more money from the Feds).

With some of  deficiencies in mind, NASA, in partnership with the FAA, NOAA, and private industry, sponsored the development and evaluation of a multi-function in-situ atmospheric sensor system for aircraft called TAMDAR, that not only provides wind and temperature, but humidity, icing, and turbulence as well.  And the sensors were designed to fit on small planes.  

A private sector company, AIRDAT, secured venture capital funding and installed the TAMDAR equipment on a number of short-haul and commuter fleets across the country (and in locations around the world, including Mexico, which lacks a decent radiosonde system).  And they started collecting some magnificent data, including more critical parameters (like humidity, icing, and turbulence) at many more airports (like Yakima, Bellingham, Walla Walla, etc) than AMDAR.   And these smaller planes spend most of the time in the lower atmosphere where we needed data:  a seeming weather home run!

TAMDAR is still expanding, but here is an example of the data over the Northwest for  a few-hour period.  You will notice the aircraft are arriving and leaving from a lot more airports!
One of the biggest problems for weather prediction in the U.S. is a lack of radiosonde soundings over Mexico:  TAMDAR has plenty (see graphic).

For a few years, AIRDAT was giving the TAMDAR data for free to NOAA, which tested the impacts on the U.S. Rapid Refresh (RUC) prediction system (the main short-term prediction system used for convection and aviation).  The results, described in a published paper by NOAA scientists, were emphatic: TAMDAR made a very positive contribution to improved forecast skill:

"TAMDAR’s impacts on 3-h RUC forecasts of temperature, relative humidity, and wind are found to be positive and, for temperature and relative humidity, substantial in the region, altitude, and time range over which TAMDAR- equipped aircraft operated during the studied period of analysis."

TAMDAR's benefits go beyond numerical weather prediction: they help National Weather Service forecasters do a better job, particularly for predicting severe thunderstorms.   This was demonstrated during the period in which the NWS had access to TAMDAR information and is documented here.

Is NOAA using ANY TAMDAR today?  The answer is no.   
Hard to believe.

AIRDAT, like any private sector firm, needed income to pay for the TAMDAR network.  At first, NOAA was willing to pay for a small amount of TAMDAR data (only a few percent of the total) and last year they dropped even that small amount.   The cost of TAMDAR data would be several million dollars a year, but that would be paid back ten or hundred times over by a few improved forecasts of major weather events.  We are talking about an amazing cost-benefit ratio.

 TAMDAR over Florida

With a lack of NOAA support, AIRDAT was in financial trouble until it was purchased by Panasonic Avionics Corporation, headquartered in Bothell, Washington.  So the TAMDAR system is still alive, for the moment.  

A number of Republican Congressmen have talked a great deal about privatizing satellite data acquisition (check out the hearings of the House Environment Subcommittee to hear plenty of complaints that NOAA is not using private satellite vendors more).  But here we have a private sector firm that has paid for much of the development using private funds, one that has established a reliable, effective system of great value for weather prediction, and neither Congress or NOAA seem to be able to find the modest funds needed to acquire the resulting data.  


NOAA management has short-changed weather prediction for a long time, including lack of computer power, lack of funding for maintaining weather buoys, insufficient funding for supporting NOAA personnel, zeroing out the important COSMIC-2 weather satellites, and lack of extramural research support.  Lately, under some pressure, they have started to deal with some of the problems (including a new weather computer and more funding for the CSTAR academic research program).   But the lack of support of TAMDAR is another serious error that is greatly undermining weather forecasting in the U.S. and it needs to be fixed.  Quickly.

4 comments:

Michael Gersh said...

This is easy to understand. The government will spend literally unlimited funds on climate, but bare minimum funds for weather. We are living in a time of the ascendance of politics and propaganda, and apparently the execution of good governance has lost quite a bit of importance coincidentally with the growth of bureaucratic regulation and political correctness.

dbostrom said...

Infinite dollars spent on climate research, only surpassed by the amount of hyperbole found on the Internet.

Real-world details such as sequestration, turning the budget into a political football and hence causing massive uncertainty about the ability to fund such things as TAMDAR are important.

For my part, I'd like to see NOAA figure out how to maintain an Earth observation satellite fleet in a more sane way, staring with avoiding chasing perfection to the exclusion of the good enough. Imaginary more perfect satellites are not as good as ones that actually exist, such as GOES 15. Why not make plans for continuous coverage that include continuing to exploit existing, useful designs a checkpoint for entering into new and better plans? In other words, any proposal for improved platforms must show how the observations we currently can make will continue to be available.

Buying and storing a few more copies of platforms will cost a bit more money but seems better than accidentally planning in massive, dangerous coverage gaps.

Simfish InquilineKea said...

What about installing sensors on birds - sort of similar to the way we put CTD sensors onto elephant seals?

Bryce Ford said...

Let’s be fair now. In addition to all the aircraft data NOAA gets from the MDCRS program, they are also acquiring aircraft Obs from the very active Water Vapor Sensing System program, WVSS-II. Currently there are 112 aircraft equipped with WVSS-II in the U.S. That is through the partnership NOAA has with ARINC, UPS, and Southwest Airlines. It enables equal or better data quality with the lowest cost/profile. At last count, they were producing as much as 2000 profiles per day. So there is no need for them to buy the other data you reference if they don’t need it. It doesn’t help the Weather Enterprise to bash our fiends at NOAA for doing a good job in spending our taxpayer funds wisely.
If you are not familiar with WVSS-II, you can easily find out all about it if you try. There is lots of independent research on its performance, and it is growing in use around the world by WMO AMDAR participants. So please, be sure you independently research all the facts before assuming the worst.