EAR-RASS and Radiosonde Data

We have adopted machine learning (ML) to retrieve vertical profiles of specific humidity (q) in the troposphere by integrating measurements from the Equatorial AtmosphereRadar (EAR) and its Radio Acoustic Sounding System (RASS). It addresses the relation between q with refractive index gradient (M ) and turbulence-related parameters, which are determined by temperature (T ), echo power (Pr ), and Doppler spectral width (\u03c3) obtained from EAR-RASS measurements. These three parameters are trained against q from radiosonde using the regression models. Among various ML methods, Ridge regression emerges as the best performer, highlighting the model\u2019s simplicity, efficiency, and robustness. Inclusion of T from RASS significantly improves the model\u2019s performance, making it more effective than models that rely solely on moment data of Wind Profiling Radar (WPR)