Abstract
The phenomenon of ducting, influenced by variations in air moisture and temperature above the sea surface, introduces non‐standard propagation characteristics that can significantly affect marine communication and radar systems. Therefore, it is necessary to understand these duct formations in order to accurately predict system performance. Atmospheric ducts are characterized by variations in the vertical refractivity profile. The Refractivity‐From‐Clutter (RFC) technique, known for real‐time low tropospheric refractivity profiling, is the focus of this study. This paper explores the concept of employing Multiple Launch Angles (MLA) to enhance RFC inversion precision. The optimal number and distribution of launch angles, multiple‐angle cost functions, and corresponding MLA‐RFC performance computations are demonstrated via simulations. An MLA‐RFC‐capable radar is built and deployed during the Coastal Land‐Air‐Sea Interaction (CLASI) Study in Moss Landing, CA in 2021 and 2022. The Lower ATmospheric PROPagation (LATPROP) radar system is a modified, low‐cost marine radar. The capabilities of MLA‐RFC are demonstrated using multiple launch angle LATPROP radar measurements collected during the CLASI campaign. The results show that the MLA‐RFC method outperforms the single‐angle RFC approach by generating reliable launch angle weights and applying them to the inversion algorithm, which leads to reduced bias and RMS error in evaporation duct height estimates. These weights are derived from the averaged simulated energy trapped in a duct for a given evaporation duct height. Both simulation and experimental results indicate that MLA improves estimation performance by up to 77% and 67%, respectively, compared to the classical single‐launch angle RFC. Key Points MLA‐RFC improves duct estimation by using multiple radar angles instead of one Field test using LATPROP and simulation show MLA‐RFC boosts accuracy by up to 67% 77% respectively versus single‐angle radar Weighted MLA stays robust under uncertain conditions and without needing prior knowledge while non‐weighted works best with few angles