Work

A Novel Approach for Improving Cloud Liquid Water Content Profiling with Machine Learning

Year: 2025

Type: article

Abstract: Accurate prediction of Cloud Liquid Water Content (CLWC) is important in atmospheric science for understanding and forecasting weather phenomena, particularly in locations with complex microclimates. We cross-validated the Medium-Range Weather Forecasts (ECMWF) ERA5 data with high-precision radiosonde observations from the Hong Kong area. Machine learning (ML) techniques are used for interpolating radiosonde data to enhance data coverage and resolution and estimate CLWC. This is complemented by applying a metaheuristic algorithm for data cleansing, which enhances the correlation between the input features - temperature, pressure levels, relative humidity, and specific humidity - and the CLWC. Our approach, tested against various ML algorithms for robustness and performance, demonstrates marked improvements in the accuracy and reliability of CLWC profiles.The methodology developed is designed to be applicable beyond Hong Kong's unique atmospheric conditions, demonstrating how similar machine learning approaches can be adapted for various climates globally. This enhances weather prediction models worldwide, highlighting the global significance of this work in atmospheric science.

Source: Remote Sensing

DOI: 10.3390/rs17111836

Authors: Anas Amaireh, Yan (Rockee) Zhang, Pak Wai Chan, Dusan Zrnic

Institutions: Al Hussein Technical University, University of Oklahoma, Hong Kong Observatory, National Oceanic and Atmospheric Administration, NOAA National Severe Storms Laboratory

Language: English

Topic: Meteorological Phenomena and Simulations

Subfield: Atmospheric Science

Field: Earth and Planetary Sciences

Domain: Physical Sciences

SDG: Climate action

Open Access status: gold

Estimated APC paid: $3,258