11:45 AM - 12:00 PM
[ACG36-11] Improvement of precipitation judgement by the GPM dual-frequency precipitation radar
Keywords:Precipitation, Spaceborne radar, Algorithm
The Dual-Frequency Precipitation Radar (DPR) onboard the Global Precipitation Measurement (GPM) mission’s core satellite measures the 3-D structure of global precipitation within latitudinal coverage (65S to 65N). The GPM DPR can detect light rainfall of about 15 dBZ. However, the sensitivity of the DPR is not enough to measure high-latitude snowfall smaller than 15 dBZ that occupies the majority of both frequent and amount. Although it is difficult to improve the DPR’s sensitivity because the sensitivity is directly constrained by the designed radar performance, a possible improvement of the precipitation detection is expected with data processing. The precipitation judgement in the DPR’s standard algorithm employs detection of successive certain signal in vertical direction (1D judgement) so that thin precipitation layer cannot be detected even if the precipitation echoes are higher than 15 dBZ. Since the top height of high-latitudes precipitation is not taller than that of tropical precipitation, the detection of thin precipitation layer is required to improve detection of high-latitude snowfall. The current study aims to improve the precipitation judgement by using 3-D information of precipitation signals.
In this study, we count up the successive signals higher than the certain level not only in the vertical direction but also the cross-track and along-track directions. The standard algorithm (1D judgement) detects as precipitation when the successive count in the vertical direction is greater than 5. Moreover, the proposed method (3D judgement) also detects precipitation when the successive count in the 3D direction is greater than 10 even if the successive vertical count is smaller than 5. This additional condition enables the improvement of precipitation detection for thin precipitation layer such as low-level storms and the edge of precipitating clouds.
The DPR (Ku-band channel) data in 2020 March were processed with the 3D judgement to quantify the improvement of precipitation detection. Analyzing the precipitation frequency using near-nadir angles data, the 3D judgment increased the precipitation detection by 20% relative to the original 1D judgement. Over high-latitude regions, about 60% increase was obtained because the detection of the low-level storms was effectively improved. Note that a change in precipitation amount due to the increase in the precipitation detection were limited.
In this study, we count up the successive signals higher than the certain level not only in the vertical direction but also the cross-track and along-track directions. The standard algorithm (1D judgement) detects as precipitation when the successive count in the vertical direction is greater than 5. Moreover, the proposed method (3D judgement) also detects precipitation when the successive count in the 3D direction is greater than 10 even if the successive vertical count is smaller than 5. This additional condition enables the improvement of precipitation detection for thin precipitation layer such as low-level storms and the edge of precipitating clouds.
The DPR (Ku-band channel) data in 2020 March were processed with the 3D judgement to quantify the improvement of precipitation detection. Analyzing the precipitation frequency using near-nadir angles data, the 3D judgment increased the precipitation detection by 20% relative to the original 1D judgement. Over high-latitude regions, about 60% increase was obtained because the detection of the low-level storms was effectively improved. Note that a change in precipitation amount due to the increase in the precipitation detection were limited.