Data Quality Control of Ka-band Cloud Radar for Detecting Cumulus Clouds

Ka-band cloud radar network is deployed around Tokyo metropolitan area to detect cumulus clouds which are developing to sever cumulonimbus. For the early detection of these clouds, the radars must have a high-sensitivity. But the high-sensitivity causes some new problems that we have never concerned. In this paper, we present a data quality control procedure for detecting cumulus clouds which are developing to cumulonimbus. The following methods are applied to the observed radar data, and show good performance for the cumulus detection.

1) Weak echo detection: The reflectivity of early stage cumulus cloud is mostly the same as the noise level (-30 dBZ - -20 dBZ) of our Ka-band radar. So we must detect the cloud signal from the noise. For this purpose, a moving average of the received power with a window size of 5 rays and 5 range bins is calculated. The range bin and beam widths of our radar are 150 m and 0.3 degrees. This calculation deletes the noise variance to retrieve the cloud echo whose horizontal scale is larger than 750 m.

2) Second-trip removal: Because of the high sensitivity, our Ka-band cloud radar detects second-trip echo much more than the other band radar. Signal Quality Index (SQI; a.k.a. normalized coherent power) is usually used to remove the second trip echo, but this also deletes the weak echo from early stage cumulus. We use the reflectivity comparison between high and low PRFs to remove the second-trip. Furthermore the second trip echo is simulated from wide area rainfall information observed by operational weather radar. The SQI method is applied only in the simulated area.

3) Range sidelobe removal: Pulse compression is used to increase the range resolution as well as the signal to noise ratio. The range sidelobe is not negligible in Ka-band cloud radar, because the reflectivity of our target is -30 dBZ - -10 dBZ. The data contaminated by the sidelobe is deleted by using the pre-determined range sidelobe pattern.

4) Gas attenuation: Gas attenuation of Ka-band is not negligible because it is 10 times larger than those of S-, C-, and X-bands. The path-integrated attenuation is calculated from the objective analysis or the numerical simulation data to include the attenuation into radar reflectivity.