Dr David Hooper

Apply Now
Instrument Support Level

Manufacturer & Model

Rutherford Appleton Laboratory in-house design

Daily Facility Charge



Calendar 2:
MST Radar

MST Radar

The 46.5 MHz National Centre for Atmospheric Science (NCAS) Mesosphere-Stratosphere-Troposphere (MST) Radar is located at the NCAS Capel Dewi Atmospheric Observatory (CDAO) near Aberystwyth in west Wales. It was previously known as the Natural Environment Research Council (NERC) MST Radar. It is operated in two modes, which cover different altitude ranges.

ST-mode (1.6 – 21.0 km)
  • Provides altitude profiles of the 3 dimensional wind vector.
  • The quality of the 30 minute averaged horizontal wind data is comparable to that of radiosonde measurements.
  • Horizontal wind data are operationally assimilated by a number of European weather forecasting agencies through EUMETNET’s E-PROFILE programme.
  • Provides a measure of turbulence intensity.
  • Provides a measure of the vertical gradient of potential temperature at upper-tropospheric altitudes and above. This allows the altitude of the tropopause to be determined.
  • The data can be used to study atmospheric phenomena over a wide range of scales. This extends down to the convective scale in the case of the horizontal wind data and down to the turbulence scale in the case of vertical velocity data.
  • The data have proved to be particularly useful for studying (amongst other things) turbulence, moist convection, mountain waves, sting jets, tropopause folds, and inertia-gravity waves.

M mode (57 – 93 km)
  • During most of the year, echoes are only observed infrequently. They tend to be weak, short-lived, and of narrow vertical extent.
  • During the mid-summer months, echoes become more frequent, strong, longer-lived, and of wider vertical extent. They are analogous to the Polar Mesosphere Summer Echoes (PMSEs) observed at higher latitudes. Their existence is known to be associated with the presence of ice crystals in the vicinity of the mesopause. The ice crystals can be seen in the form of Noctilucent Clouds (NLCs) if they grow sufficiently large.
  • The radar has been operated on a quasi-continuous basis since late 1997 and contributes to NCAS’s long term measurement programme. These observations focus mainly on the ST-mode, but an M-mode observation has been included in the standard observation cycle since 2005.

    The radar can be booked for campaign support. You will need to submit an application even if your campaign relies on the standard observation cycle used for long term measurements. The radar can alternatively be operated using a bespoke observation cycle.

    Instrument Details
    • Radar frequency: 46.5 MHz
    • Typical range resolution: 300 m
    • Typical duration of observation cycle: 5 minutes
    • Latitude: 52.424500°N
    • Longitude: -4.0054665°E
    • British National Grid Reference: SN6379882627
    • Altitude above mean sea level: 50 m

    Summary of Dataset

    Lower Very High Frequency (VHF) radar signals are primarily scattered from ‘clear-air’ targets, i.e. structures in atmospheric refractive index that have scale sizes of half the radar’s wavelength (approximately 3 m) along the beam pointing direction. Scatter from hydrometeors, aircraft and ground-based targets can also occasionally be detected. The refractive index depends on humidity (within the lowest 10 km of the atmosphere), on air density (within the lowest few 10s of km), and on free electron density (above 50 km).

    The radar return signals are parameterised by their power, Doppler shift, and spectral width. The signal power depends on the vertical gradient of (potential) refractive index and so relates to atmospheric structure. The Doppler shift depends on the radial component of the wind along the beam pointing direction. The spectral width depends on the intensity of any turbulence within the radar observation volume. However, the values must be corrected for the effects of so-called beam-broadening, which depends on the horizontal wind speed and on the width of the radar beam.

    The basic unit of MST radar observation is referred to as a dwell. This involves accumulating range-sampled receiver signals over a large number of successive transmitter pulses (typically covering a period of 21 s in the case of the NCAS radar). The time series of samples for each range gate is converted into Doppler power spectrum through the application of a discrete Fourier transform. These spectra are processed in order to identify signals and to parameterise them.

    The Doppler Beam Swinging (DBS) technique involves cycling through a sequence of dwells with different beam pointing directions. The minimum requirement for deriving the three dimensional wind vector is for observations to be made in the vertical direction and at an off-vertical angle in two orthogonal azimuths. This is referred to as the 3 beam technique. The five beam technique requires the off-vertical beam observations to be made in four orthogonal azimuths. This provides redundancy and allows for self-consistency checks to be made. The NCAS MST radar typically makes use of the 5 beam technique.

    The radar began making observations in 1989 and has been operated on a quasi-continuous basis since late 1997. Its observations contribute to a long term measurement dataset irrespective of whether or not the radar is supporting a field campaign. The standard observation cycle, which takes just over 5 minutes, makes use of both the ST and M modes at 300 m range resolution. The radar can be operated using a non-standard observation cycle on request. The vertical velocity data are typically used without averaging since fluctuations can occur over time scales of considerably less than 30 minutes.

    An application for access must be made if you would like the radar to support a field campaign. This is irrespective of whether or not you require the radar to be operated using a non-standard observation cycle. Non-standard observation cycles must typically be designed so that it is possible to derive horizontal winds at 300 m range resolution. Consequently it is important to discuss your requirements with the instrument scientist before submitting an application. You do not need to submit an application if you simply want to make use of historical data.

    Most of the radar’s hardware is over 30 years old. It is subject to a continual programme of maintenance rather than to servicing at fixed intervals. The representativeness of the horizontal wind data has been established through a succession of studies. Although the radar is not calibrated in a formal way, the EUMETNET’s E-PROFILE programme provides monthly statistics of how the measured winds compare to those generated by models. This is how it is known that the accuracy of the data is comparable to that from radiosondes.

    Refer to the Access to the data section of the Capel Dewi Atmospheric Observatory page for general details about:

    • Licensing and acknowledgements
    • Access to data files/plots
    • Availability of data files/plots
    • Centre for Environmental Data Analysis (CEDA) file naming convention and archive structure

    There are 4 levels of MST Radar products. The highest-level Cardinal data products are the most appropriate for virtually all ST-mode studies. However, Radial data products should be used for M-mode studies. They might also be appropriate for specialist turbulence-scale ST-mode studies

    This section contains details about

    Cardinal data products
    • These include:
      • Eastward wind component
      • Northward wind component
      • Upward wind component
      • (Vertical beam) radar return signal power
      • Corrected spectral width (used for turbulence studies)
      • Uncorrected spectral width (only included for reference)
      • Aspect sensitivity (used for turbulence studies)
      • Tropopause altitude
      • Tropopause sharpness
    • The data files contain multiple quality control flags. Refer to the ancillary_variables attribute of each variable to find out which one to use.
    • Cardinal data products are closely related to Cartesian data products. The main difference is that the horizontal wind components have been averaged over a nominal period of 33 minutes in order to improve their representativeness (these averaged winds are used to apply beam-broadening correction to the Vertical beam spectral widths).
    • Since Cardinal data files are currently derived from Cartesian data files, their availability typically lags behind that of Cartesian data files by a day.
    • ST-mode data plots are initially derived from Cartesian data products, but are overwritten with plots derived from Cardinal data products when they become available.
    • Cardinal data products are not available for M-mode observations.
    • Cardinal data products are only available from the version 4 dataset.

    Cartesian data products
    • These are essentially the same as Cardinal data products. However, the horizontal wind components are derived from single cycles of observation. These should not be used without averaging. A nominal period of 33 minutes (which covers 6 cycles of observation) has been found to be optimal.
    • Although Cartesian data products are available for M-mode observations, they do not add any value to the Radial data products, which should be used instead.
    • ST-mode data plots are initially derived from Cartesian data products, but are overwritten with plots derived from Cardinal data products when they become available.
    • Cartesian data products should be used from the version 3 dataset.

    Radial data products
    • These are typically only used to derive Cartesian (and hence Cardinal) data products. However, they have advantages for two specialist areas of study:
      • M-mode observations. The Cartesian data products were designed for ST-mode studies. Although they are available for m-mode observations, they offer no added value.
      • Turbulence scale ST-mode studies. Radial data files typically contain vertical velocity data at approximately 50 s intervals (as opposed to approximately 5 minute intervals for Cardinal/Cartesian data files).
    • The following products are available for each dwell:
      • Noise power
      • Signal power
      • Radial velocity
      • (Uncorrected) spectral width
      • spectral limits of signal components
    • For ST-mode observations, two potential signal components are identified within each spectrum. Only the first of these should be considered. The second one is identified in case the strongest signal component within a spectrum does not relate to a desired radar return, e.g. if it is the result of echoes from aircraft, ground clutter, or interference, and echoes from hydrometeors. A radial continuity algorithm is used to identify which of the signal components is most consistent with an atmospheric profile.
    • A single quality control flag, signal_component_is_reliable applies to all data products (except noise power, which has no quality control flag).
    • An additional variable, signal_component_reliability_details, provides details of the various quality control checks that have been carried out in order to derive the value of signal_component_is_reliable.
    • Radial data products should be used from the version 3 dataset.

    Spectral data products
    • These are typically only used to derive Radial data products. However, they may be of interest for specialist turbulence studies.
    • They contain only a single parameter: power spectral density.

    Overview of file naming convention

    Refer to the Access to the data section of the Capel Dewi Atmospheric Observatory page for general details about the file name format.

    MST Radar data files contain a string with the format AARRR, where

    • AA represents the altitude mode, i.e. st or m
    • RRR represents the range resolution (in metres), which typically has a value of i.e. 300, but can sometimes be 150 for ST-mode observations or 1200 for M-mode observations.

    Files that contain data of suspect quality within limited time-altitude regions are available through the archive. They contain the string suspect within the file name. Details of the affected region are contained within the comment global attribute of the file. Do not use these files unless you block the data within the affected region.

    Version 4 dataset
  • CEDA catalogue page
  • File name format: nerc-mstrf-radar-mst_capel-dewi_YYYYMMDD_AARRR_cardinal_TT-min-smoothing_v4-0.nc, where TT represents the nominal period (in minutes) over which horizontal winds have been averaged.
  • ST-mode CEDA archive location: /badc/mst/data/nerc-mstrf-radar-mst/v4-0/st-mode/cardinal/
  • ST-mode web access to CEDA archive (requires a CEDA account)
  • There are no version 4 data products for M-mode observations.
  • Version 3 dataset
  • CEDA catalogue page
  • Cartesian data products
  • Radial data products