RADAR processing and products

RASTA

RASTA measures both power and velocity for each antenna (up to 6 antennas looking up and down). However, these measurements cannot be used directly for scientific purpose. The radar reflectivity must be calibrated and range corrected in order to convert the it into dBZ. The background noise is also removed using thresholding technique knowing the background noise characteristics.

There are several RASTA products available:

  • L2 products for each antenna (note that this product also contains raw data)
  • L2 WIND product (standard)
  • L2 WIND/ice microphysics product (optional as it depends on campaigns)
  • Spectral analysis product (available only on demand)

Note that Level 1 data are not distributed as our L2 contain everything.

Calibration

RASTA nadir reflectivity is usually calibrated using the ocean surface return technique (Li et al. 2005; Tanelli et al. 2008). The calibration of the other antennas is either directly derived using ground surface (in a similar manner as the nadir) or by comparing with nadir reflectivity as a reference.

During the integration phase of the aircraft, the zenith measurements are taken from the ground in SAFIRE’s home facility while a calibrated BASTA radar is deployed. It allows us to derive a first calibration estimate. During the field campaign, the radar is often observing the ocean surface backscatter at different altitude and incidence angle. The surface echo is identified and compared against an ocean surface echo model. ERA5 hourly reanalyses data are used to correct the radar signal from gaseous attenuation and the wind information is used for characterising the surface wind.

Single antenna level 2 product

RASTA measures both received signal power and along sight Doppler velocity for each antenna. The data are processed in the FPGA using the pulse pair technique. However, these measurements cannot be used directly for scientific purpose. The radar reflectivity must be calibrated (previous section) and range corrected in order to convert it into dBZ. The background noise is also removed using a thresholding technique knowing the background noise characteristics.

The figure represents the level 2 RASTA processing which is applied for each antenna. ERA5 reanalysis, aircraft’s information and surface type are external information that are used in the generation of this product.

RASTA L2 processing flowchart

Cloud WIND retrieval

The 3D wind is retrieved using an optimal estimation approach (inspired from Protat and Zawadzki, 1999), which consists in using an iterative process to adjust the state vector containing Vx (along track wind component), Vy (cross track) and Vz (vertical). The method uses as input the radial Doppler velocity of each antenna (from L2 data).

RASTA wind L2 product contains also some information regarding the nature of the hydrometeors. Temperature field from ERA5 is used to support the ice/liquid distinction. The melting layer is detected using the vertical velocity as the rapid conversion of ice into rain leads to a strong vertical gradient in vertical velocity. This product comes with a radar_classification_flag indicating nature of the observed targets.

Example of wind retrieval and target classification – THINICE 2022 (ATR42)

BASTA

The radar reflectivity is calibrated and range corrected in order to convert the it into dBZ (same as RASTA). The background noise is also removed using thresholding technique knowing the background noise characteristics.

BASTA is a bistatic radar, therefore we need to correct for overlap effect. Fortunately, this overlap is not an issue for the small BASTA antennas and has been checked by looking at rainy profiles.

The closest gates of the radar cannot be used, mainly due to the coupling effect – direct return into the receiver. This cleaning is done using the entire data base by looking at the normalized value (normalized by the first gate value).

As BASTA is a Doppler RADAR the line of sight velocity of the hydrometeors is measured however we need to correct it from aircraft motion and folding.

In the L1 data the reflectivity is calibrated, background noise is removed and a mask is available. Data are also geolocated. This product is available for two range resolutions (12.5m and 25m and up to 12km)

Example of reflectivity map (L1) derived from BASTA during the THINICE field campaign (Svalbard) in 2022

There is a L2 version for some campaigns including a target classification and dedicated extra information related the campaign. In this product the two range resolutions are mixed for an optimal product.

Search