Optimizing High-Frequency Radar Data: A Study on Effective Imputation Methods for Ocean Current Analysis

This study addresses the issue of missing data in High-Frequency Radar (HFR) measurements, which are crucial for monitoring ocean surface currents and supporting maritime operations and environmental studies. However, the integrity of HFR data can be compromised by factors such as sensor malfunctions and environmental disturbances, creating a need for effective data imputation methods. The research focuses on evaluating and comparing the effectiveness of different statistical and machine learning techniques, including Seasonal-Trend Decomposition using Loess (STL) and Cubic Spline Interpolation, to accurately impute missing wind speed data. The findings of the study indicate that STL and Cubic Spline methods are superior in handling missing data, demonstrating high accuracy and strong correlation with actual data across scenarios with 5%, 10%, and 15% missing data. These methods outperform others like the Last Observation Carried Forward and Moving Average, especially when the percentage of missing data increases. The study highlights the potential of advanced imputation techniques to enhance the reliability of oceanographic data analysis, thereby improving the accuracy of ocean current models and forecasts. These advancements are crucial for enhancing maritime safety, resource management, and environmental monitoring, and they pave the way for further research into optimizing data imputation methodologies in related fields.