Estimating Vertical Air Motion within Precipitating Clouds Using a Novel Dual-Frequency Radar Method: From Single-Site Validation to Multi-Site Application for an MJO Case Study

Abstract

Very High Frequency (VHF) atmospheric Doppler radars measure vertical air motion via a Bragg scattering peak, but precipitation introduces a Rayleigh scattering peak that severely biases the vertical air motion estimate downward. The conventional two-peak fitting method is computationally unstable and frequently fails to converge, resulting in substantial data loss. To automate this process and overcome these limitations, we developed a novel, robust dual-frequency method for the VHF Equatorial Atmosphere Radar (EAR) and a co-located L-band Boundary Layer Radar (BLR) at Kototabang, Indonesia. To resolve the downward biases caused by the Rayleigh scattering peak, our approach utilizes a highly reliable peak identification routine for spectral separation and implements a weighted-filtering technique only when peaks overlap. This robust combination successfully resolves the persistent downward bias. Validation confirmed the proposed algorithm's superiority: it achieved accuracy nearly identical to manual estimates, recovered data that the conventional fitting method lost at a 20%–40% failure rate, and proved effective using the EAR's standard observation mode. As a showcase of its practical utility, the proposed algorithm was applied to VHF/UHF radar data from Darwin, Australia, for a simultaneous case study of the Madden-Julian Oscillation (MJO) at Kototabang and Darwin, revealing that the vertical motion profiles largely differed: stratiform precipitation dominated at the MJO core, while convective precipitation characterized the edge. This work provides an essential, validated tool for generating highly reliable, long-term datasets critical for enabling future rigorous statistical investigations of regional differences in tropical atmospheric dynamics.