Robust Lidar-Radar Composite Cloud Boundary Detection Method With Rainfall Pixels Removal

Cloud vertical structure detection is essential for understanding atmospheric dynamics. Currently, cloud boundaries can be effectively identified based on lidar and millimeter-wave radar. However, how to integrate the two observation methods and remove the interference of rainfall on cloud identification are crucial for precise detection of cloud boundaries. This study develops a robust cloud boundary detection method combining radar and lidar observations with ability to identify rainfall effectively. Consistency analysis at Sheyang meteorological station using radiosonde data showed that lidar detected 41.1% of clouds and radar detected 93.3% of clouds compared to composite detection. The composite method overestimated the cloud base by 855.1 m and underestimated the cloud top by 551.2 m compared to radiosonde, as radiosonde measurements are affected not only by drift but also by rainfall, which mainly affects cloud base detection. Utilizing Doppler velocity and the lidar-radar cloud base difference improved rainfall detection by 41.4% over Doppler velocity alone. Observations are consistent with ground-based rain gauge, with Doppler velocities providing good identification of significant rainfall. Also, different cloud bases detected by lidar and radar providing additional identification of drizzle. With the rainfall removed, the error of rainwater path offered by microwave radiometry during rainfall is reduced by 32.4%. Overall, this study proposes a threshold-insensitive lidar-radar composite cloud identification method. It has good robustness and more precise detection of cloud boundaries for its ability to identify vertical rainfall regions.