Access Official Technical PDF

Share Summary:LinkedInEmail

A mesosphere-stratosphere-troposphere (MST) radar must detect weak atmospheric echoes across a wide altitude range while controlling interference, coordinating many antenna elements and producing usable wind profiles. The Wuhan MST radar offers a detailed example of how those requirements can be addressed through a distributed hardware architecture, configurable observation modes and a processing chain that includes clutter suppression. Comparisons with independent measurements and atmospheric models provide a further check on the resulting wind data.

MST radar architecture: antenna array and distributed transmitters

Located in Chongyang, Hubei Province, the Wuhan system is a monostatic Doppler radar operating at 53.8 MHz. Its antenna field contains 576 Yagi antennas arranged as a 24-by-24 square array over an area with a side length of 96 metres. Each antenna is driven by an individual small transmitter-receiver (TR) module, with a peak output power of 300 W. The system’s maximum peak power is 192 kW.

Rather than relying on a single high-power transmitter, the design distributes amplification across the array. Twenty-four large TR modules in the observation house connect through row and column feeding networks to the small modules in the antenna field. Switching between row and column drive states supports beam steering in the two horizontal directions. The system generally observes vertically and at 15-degree off-zenith angles to the north, south, east and west.

This arrangement also shapes maintenance and monitoring. Built-in test functions monitor the large modules, while the small modules use a modular design intended to support repair. Equal-length feeder cables provide stable channel characteristics without requiring compensation for differences between cable paths, according to the system description.

Timing, waveform control and digital processing

A radar controller coordinates transmission, reception and beam changes. Its timing generator produces the control signals used to select an observation mode, switch between row and column operation, and manage test signals and beamforming weights. A direct digital synthesizer generates coded continuous-wave signals for transmission, as well as calibration signals for the receiver channels.

The receiver chain protects and conditions incoming signals before digitization. After amplification and bandpass filtering, an analog-to-digital converter samples the signal. Digital down conversion then produces in-phase and quadrature components for subsequent processing. The system’s digital signal processing includes pulse compression, coherent averaging, fast Fourier transforms and spectrum averaging. Digital beamforming applies amplitude and phase weights to support beam formation.

Observation parameters can be selected for different altitude regions. The low mode covers approximately 3.5 to 12 km, the middle mode approximately 10 to 25 km, and the high mode approximately 60 to 85 km. The reported height resolutions are 150 m, 600 m and 1,200 m, respectively. Under normal operation, the radar cycles through modes and produces wind data at a 30-minute temporal resolution.

Why clutter suppression matters

Clear-air echoes can be weak relative to returns from nearby terrain, trees or buildings. A strong ground return can obscure atmospheric signals and bias estimates of radar parameters. The Wuhan site uses a fence to help limit ground clutter, but the processing chain also needs to address residual interference.

The reported processing approach targets ground clutter and high-frequency interference. Ground clutter typically appears as a narrow peak near zero Doppler frequency and changes relatively little over time. The processor suppresses this near-zero-frequency component, then applies a median filter at each range gate to reduce high-frequency interference. In the illustrated example, these steps reveal atmospheric signals that were difficult to distinguish in the original spectrum.

For system operators, the practical point is that usable wind retrieval depends on more than transmitter power or antenna count. Signal quality, interference rejection and the criteria used to accept or reject measurements all affect the final data product.

Validating wind measurements across altitude

The radar estimates horizontal winds using Doppler beam swinging (DBS): radial velocities from four inclined beams are combined to retrieve the horizontal components. A sample is considered invalid if interference or low signal-to-noise in any inclined beam prevents the wind solution. This makes data acquisition rate a useful measure of how consistently the system can produce valid horizontal winds at different heights.

During the 2016–2017 assessment period, acquisition rates were generally above 90% from 3.5 to 10 km in low mode and from 10 to 20 km in middle mode. Rates decreased at the upper limits of those ranges. High-mode observations were concentrated mainly between 66 and 86 km, with availability affected by the conditions required for mesospheric returns. The reported analysis found substantially better high-mode acquisition when considering daytime observations alone.

Radiosonde and reanalysis comparisons

For the troposphere and lower stratosphere, radar profiles were compared with radiosonde observations and ERA-Interim reanalysis. In a short-term comparison, zonal winds agreed well across much of the 3.5-to-23-km range. Meridional winds also agreed at most heights, although the radar measured weaker values around 14 km. The comparison involved instruments about 120 km apart, so differences in location, timing and measurement method can contribute to discrepancies.

Across monthly averages, the radar captured major features of the zonal wind pattern in ERA-Interim. Meridional wind comparisons showed larger differences, in part because those winds varied more over time and the averaging periods differed. These results support evaluating both wind components and considering the time and spatial scales represented by each dataset.

Meteor radar and mesospheric model comparisons

Mesospheric winds were compared with observations from a meteor radar and with the HWM-07 atmospheric wind model. The meteor radar comparison showed overall agreement in zonal and meridional wind trends between 76 and 86 km, with better correspondence around 81 km. Differences can arise from atmospheric variability between measurement locations and from the MST radar’s lower data availability at some mesospheric heights.

Monthly mean mesospheric trends were also broadly consistent with HWM-07, while wind magnitudes differed in some seasons. The comparisons do not imply identical results: they show how independent instruments and models can help assess whether the radar captures the main patterns while highlighting conditions that merit further investigation.

Practical implications

The Wuhan MST radar demonstrates how a large antenna array, distributed TR modules, configurable timing and digital processing can be integrated for wind observations across multiple atmospheric layers. Its validation results show the value of checking measurements against independent instruments and models, while also accounting for differences in coverage, averaging and data availability. For engineers and system designers, the broader lesson is to assess the complete measurement chain—from array and timing through interference handling and retrieval criteria—when judging the reliability and operational usefulness of radar data.

Access Official Technical PDF

About The Author