MAPR Wind Profiler Data

This readme discusses MAPR wind profiler data from the T-REX project.

NCAR/EOL operated three 915 MHz boundary layer wind profiler radars at ISS (Integrated Sounding System) sites at T-REX. These radars detect scattering from clear-air refractivity gradients (such as those due to turbulence and inversions) and from precipitation. This scattering is used to make a measurement of the wind profile above the radar, typically from about 200 meters through the boundary layer to 2 to 5 km. The strength of the scattering (reflectivity) and thus ability to measure wind is a complicated function of temperature, humidity, turbulence, precipitation, and the presence of unwanted signals (radio interference, clutter echoes from trees, power lines, birds, etc). These radars can also measure the virtual temperature profile (up to 500 to 800 meters) using the Radio Acoustic Sounding System (RASS) technique.

MAPR (Cohn et al 2001) is our research wind profiler (the other two wind profilers are were standard DBS or Doppler Beam Swinging systems and are in the data catalogue under the ISS2 and MISS entries). MAPR uses spaced antenna techniques to made very rapid wind measurements. The data catalogue includes 5-minute and 30-minute winds from MAPR, whereas the DBS system just have 30-minute winds. MAPR was making measurements as fast as 30 seconds, however these are only available on a case-by-case basis.

The MAPR ISS was located about a mile southeast of Independence in a gravel pit. The site was about 20 feet below the surrounding terrain, thus shielding the system from unwanted clutter echoes and interference. An NCAR GAUS radiosonde system, a meteorological tower, a ceilometer, and a sodar-RASS (from Arizona State) were also operated at this site which was at Lat: 36d 47.279m Lon: 118d 10.705m

Raw IQ time series data were archived for the entire project. The data were processed in near real-time to find correlation functions (averaged over 30-seconds), and FCA (Full Correlation Analysis, Briggs 1984), was applied to determine winds. Post-processing also applied the Slope-at-zero-lag winds technique (Holloway et al 1997). A number of techniques were applied to combine, average, and filter the raw wind estimates, including consensus averaging and around a dozen statistical tests; with a fuzzy logic scoring scheme used to sort the results.

MAPR does pick up clutter and other unwanted signals (at T-REX bird echoes particularly on clear calm nights were a problem), however the filtering and test routines did a reasonable job of recovering the atmospheric signals. Comparisons with radiosondes gave a median deviation of 1.6 m/s or standard deviation of 2.4 m/s - a little larger than our usual s.d. of 1.5 m/s, however most of the radiosondes during T-REX were launched in IOP conditions of strong winds and so there was a larger than usual separation between the radiosondes and profiler sampling volume at altitude.

The accompanying data files are of four types, wind measurements (files with extension "windsnc_05" and "windsnc_30"), RASS virtual temperature measurements (extension ""), and spectral moments data (extension "momnc"). For example files mapr060325.windsnc_05, mapr060325.windsnc_30, and mapr060325.momnc give the 5-minute, 30-minute, and moments for March 25 2006 respectively. The winds and moments files contain samples for the whole 24-hours of each day. The moments data are the zeroth, first, and second moments of signals from each 30-second dwell. These moments are the signal strength (here SNR or Signal to Noise Ratio which can be used to estimate reflectivity), Doppler shift (from which the vertical winds are derived), and spectral width (can be used to estimate turbulence with lots of caveats). We generally recommend that only experienced profiler users analyze the moment data.

RASS was run for 5-minutes either every hour or half hour depending on the operations mode of the day. The RASS files have names like for March 5th, 2006, at 14:00 UTC. Unlike the winds and moments files, there is a separate RASS file for each averaging period (typically 5-minutes). The RASS files are listed separately to the winds and moments files on the T-REX data catalogue.

The data files are in netCDF format. This is an advanced binary format with data self-descriptors widely used in the atmospheric research community. It was devised by UCAR/Unidata and is described further on their web site

The data is arranged in time, height coordinates (the height coordinate is labeled "x" for compatibility with some legacy software). For winds (in the *.windsnc files) use variables wspd and wdir (wdir follows the meteorological convention of the direction the wind comes from, measured clockwise from north). For RASS (* files), use variable tv (virtual temperature) for calm conditions or variable tvc when significant vertical velocity is expected (such as when waves or convection are occurring).

The ISS data are plotted on our project web page where you can also find documentation such as descriptions of the instruments and a map of their location during T-REX as well as operator logs.

Please contact Bill Brown (wbrown at or phone 303-497-8774) for more information.


B.H. Briggs, 1984: "The Analysis of Spaced Sensor Records by Correlation Techniques", Middle Atmosphere Program Handbook, Vol. 13, pg. 166 - 186.

S.A. Cohn, W.O.J. Brown, C.L. Martin, M.S. Susedik, G. Maclean, and D.B. Parsons, 2001: "Clear Air Boundary Layer Spaced Antenna Wind Measurement with the Multiple Antenna Profiler (MAPR)", Annales Geophysicae, 19(8), 845 - 854.

C.L. Holloway, R.J. Doviak, S.A. Cohn, R.J. Lataitis, and J.S. Van Baelen, 1997: Cross correlations and cross spectra for spaced antenna wind profilers, Part 2: Algorithms to estimate wind and turbulence, Radio Sci., 32, 967-982.