Estimating Liquid Water Content Using Dual-Frequency Radar and Bayesian Neural Network

Liquid Water Content (LWC) is a pivotal parameter that describes the mass of the water in a cloud in a specified amount of dry air, crucial for research in cloud physics and meteorology. This study explored a novel approach to estimating the LWC of cloud layers from radar observational data by utilizing dual-frequency radar (35 GHz and 95 GHz) in tandem with Bayesian Neural Network (BNN). The dual-frequency radar utilizes differential attenuation between the two distinct frequencies to directly assess the LWC in clouds, with the variance proportionate to the overall LWC in the surveyed volume. However, due to atmospheric perturbations and other factors, standalone radar observations might not suffice for high-precision LWC evaluations. To address this, we incorporated BNN, capable of handling the inherent uncertainties in radar data and offering a more accurate estimation for LWC. Preliminary results demonstrate that the combination of dual-frequency radar and BNN can effectively assess the LWC of cloud layers, showcasing superior accuracy and resolution compared to conventional methods. This methodology provides a potent tool for a deeper understanding of cloud physical processes and further refinement of climate models.