Gap Filling of the Coastal Ocean Surface Currents from HFR Data: Application to the Mid-Atlantic Bight HFR Network

Abstract

igh-frequency radar (HFR) surface current data are an increasingly utilized tool for capturing complex dynamics of coastal ocean systems worldwide. The radar is uniquely capable of sampling relevant temporal and spatial scales of nearshore processes that impact event response activities and basic coastal ocean research. HFR is a shore-based remote sensing system and is therefore subject to data gaps, which are predominately due to environmental effects, like increased external noise or low signal due to ocean surface conditions. Many applications of these surface current data require that these gaps be filled, such as Lagrangian numerical models, to estimate material transport and dispersion. This study introduces a new penalized least squares regression method based on a three-dimensional discrete cosine transform method to reconstruct hourly HFR surface current data with a horizontal resolution of 6 km. The method explicitly uses both time and space variability to predict the missing value. Furthermore, the method is fast, robust, and requires relatively low computer memory storage. This paper evaluates the method against two scenarios of common data gaps found in HFR networks currently deployed around the world. The validation is based on observed surface current maps along the mid-Atlantic coast of the United States with specific vectors removed to replicate these common gap scenarios. The evaluation shows that the new method is robust and particularly well suited to fill a more common scenario with complete data coverage surrounding an isolated data gap. It is shown that the real-time application of the method is suitable for filling data gaps in large oceanography datasets with high accuracy.