Doppler Radar Data : Processing

1. Summary of the processing

 "The MIRAI Doppler radar data set" is processed with the following methods for the raw data acquired using IRIS (Vaisala Inc. Sigmet Product Line, USA).
· Filter processing
· Conversion from polar coordinate system into rectangular coordinate system
· Addition of the latitude and longitude information
Doppler radar / signal processor
Radar reflectivity
Doppler velocity*1
Position data
Filter processing

Removal of the data noise level
Post Processing
Filter processing

Removal of low coherence data
Removal of shadow area data
Removal of the speckle noise
Bias correction*2
Correction of false velocity unfolding*3
Reduction processing of the second-trip echo
Conversion from polar coordinate system to rectangular coordinate system
horizontal / vertical grid interval: 1km
Addition of the latitude and longitude information
Output data set
*1:Surveillance scan mode does not contain the Doppler velocity data. In dual-PRF(Pulse Repetition Frequency) observation, correction of false velocity unfolding is carried out in signal processor automatically.
*2:Bias correction is applied only to the cruise during which the receiver trouble occurred.
*3:Correction of false velocity unfolding is not applied to single-PRF observation.

2. Details of filter processing

2.1 Removal of the noise level data (IRIS Programmer's Manual, 2010; Katsumata et al., 2008)

The following equation is used to define the threshold between the valid signal and the noise of the raw data:
Zmin is the threshold reflectivity to be obtained (in dB);
r is the range distance in km;
Zmin(1km) is the threshold reflectivity at the range distance of 1km (in dB);
Cgas is the gaseous attenuation coefficient (dB/km).
The Zmin(1km) and Cgas are taken from the dataset in IRIS/Open RAW format.

2.2 Removal of Low coherence data

 SQI(Signal Quality Index) is the value calculated in the signal processor system "RVP7" and is used as a threshold for determining whether the Doppler velocity is available or not. SQI with the value from 0 to 1 indicates that the signal is non correlative noise (white noise) for "0" and is a pure tone for "1", i.e., all signals have the same strength. The larger SQI is, the harder the noisy bin is adopted.
 We adopt 0.3 as the SQI and remove the data with the value less than 0.3.
 Before the MR01-K05 cruise, RVP7 was not installed. Therefore, the correction using SQI is not carried out.
 See “IRIS Programmer's Manual, 2010” for detailed correction method.

Removal of Shadow area data (Katsumata et al., 2008)

 "R/V MIRAI" has some structures that become the obstacles for the radar observations.
 The data in these shadows are removed by referring to look-up tables. The look-up tables are constructed statistically from the observational data. To simplify, we assume the shadow depends only on the azimuth and elevation, not on range distance, because all obstacles existed only in the vicinity of the radars.
 In case of the MIRAI, the data in the azimuths with an anomaly of the averaged reflectivity of less than ‒3 dB (i.e., half power) are removed as the shadow direction.

2.4 Removal of the speckle noise

 In case the data do not exist in 2 consecutive bins of the beam direction, it is determined to be a speckle noise.

2.5 Correction of false velocity unfolding (Katsumata et al., 2005)

 The automatic Doppler velocity unfolding with the signal processor may mistake processing in a domain with a few data. Therefore, we judge the need of the correction of the Doppler velocity data from the following conditions:
· The discontinuity of the Doppler velocity in the range of +/-2 degree in the azimuth direction from bin of the aim.
· The discontinuity of the Doppler velocity in the range of +/-3km in the beam direction from bin of the aim.
 The unfolding is conducted up to five times at the maximum.
* The unfolding is not carried out for the single-PRF observation.

2.6 Reduction processing of the second-trip echo (Katsumata et al., 2005)

 In case that the second-trip echo exists, the large discontinuity of the reflectivity occurs between a beam next to each other. The present technique to remove second-trip echo uses this characteristic.
· For bin where horizontally changes more than 2dB/km are seen, it is flagged as the "invalid data".
· If this "invalid" data flag exists more than 70% in a range of 5km of the beam direction, the bin is removed as the second-trip echo.

2.7 Conversion from polar coordinate system to rectangular coordinate system

 On the virtual globe centered the interpolated grid, the weighted interpolation method with a Gaussian distribution is used to convert from polar coordinate system to rectangular coordinate system.
 The weighted value is expressed in the following expressions.
d : Distance from the interpolation point
W(d) : Weighted coefficient of the data which separated only distance (d) from the interpolation point
H : Half bandwidth
 The half bandwidth is set to horizontally direction in 500m and vertically direction in 250m.

2.8 Addition of the latitude and longitude information

 The central location of volume scan is set from a position of "MIRAI" at the start time of each volume scan observation.
 The conversion from the grid position (X, Y) in the rectangular coordinate system to the latitude and longitude is made with the Lambert azimuthal equal-area projection method (Nakatsuka, 2006).
 In this case, the earth is supposed to be GRS80 ellipse.

3. References

Katsumata, M., K. Yoneyama, Y. Yuuki, S. Sueyoshi, N. Nagahama, and K. Yoshida, 2005 : Noise filtering for dual-PRF observed data from R/V Mirai shipborne Doppler radar. JAMSTEC Rep. Res. Dev., 2, 29-34
Katsumata, M., T. Ushiyama, K. Yoneyama, and Y. Fujiyoshi 2008 : Correction of Radar Reflectivity Using TRMM and Distrometer, SOLA, 4, 101-104
Nakatsuka, T., 2006 : Library Software for Geophysical Data Processing and Representation (2), GSJ Open-File Report, no.442

IRIS Programmer’s Manual, 2010 :
RVP7 User's Manual, 2003 :