Date of Award

Spring 2021

Document Type

Open Access Dissertation

First Advisor

George Voulgaris

Abstract

Although High Frequency (HF) radars are used routinely for measuring ocean surface currents at high spatial and temporal resolution, their utilization for estimating ocean wave spectra is still limited, mainly because of the lack of extensive evaluation of the accuracy of wave inversion models, and lack of well-established methods, especially if swell is present in the area of study. Estimation of surface currents is based on analyzing the signal of the first-order Bragg peak, while extraction of wave information requires analysis of the signal contained in the second-order continuum of the Doppler spectra; its quality depends on a number of environmental (i.e., noise levels, ocean wave energy) and system-based (i.e., frequency of operation, range, azimuthal angle, etc.) parameters. A number of theoretical and empirical inversion methods have been developed to estimate wave parameters from the HF radar data, with the latter one being more attractive for routine operations due to easier implementation and reduced computational cost. Further, most research on HF radar wave inversion has been limited todissertation, a hybrid radar wave inversion method that treats swell and wind waves separately is introduced and evaluated using a single Very High Frequency (VHF) 48 MHz radar site, two High Frequency (HF) 12 MHz radar sites, and in situ wave measurements. Using a single VHF (48 MHz) covering the nearshore and in situ directional wave data from ranges between 0.7 and 4.2 km and beam angles between 22.3 and 55.8 deg, it was concluded that wind wave inversion of the 2nd order spectra requires normalization by using Barrick’s (1977b) weighting function. This removes no wind-wave energies from the second harmonic and corner reflection peaks and leads to better wave estimations. However, at lower operating frequencies the normalization removes some of the wind wave energy something that needs to be accounted for. Application of the weighting function in the wind wave inversion model results in empirical wind-wave regression coefficient that is not wave frequency-dependent and of similar in magnitude to those found in studies that used different radar operating frequencies but included the weighting function in the inversion. This is further confirmed using data from 12 MHz system sampling ocean conditions with significant swell energy being present at times. The applicability of the empirical wave inversion method to increase the accuracy of the estimation of ocean wave spectra and wave bulk parameters by accounting for the presence of swell waves is examined and presented. The ability of the method to estimate wave directional spectra and bulk wave parameters from inverting Doppler spectra are investigated. Doppler spectra from single beam/site and two beams/sites WERA HF radar system operated with frequency 12 MHz are used over a one-month (March 30th-April 27th, 2012) data collection. Within the radar footprint, in situ wave spectra were collected using a buoy deployed offshore of the north coast of Cornwall in the UK, and used for comparisons.

To examine the influence of swell, three different swell inversion models developed by Lipa et al. (1981), Wang et al. (2016), and an empirical method, denoted as LPM, WFG, and EMP respectively are presented and evaluated. The methods were evaluated using (1) a single beam from a single radar site, (2) two beams from a single radar site, and (3) two beams from two radar sites intersecting each other at the buoy location. The LPM swell method for two beams from two sites scenario was found to be the most accurate in estimating swell parameters (RMS Error of 0.24m), the inverted swell height correlated well with the partitioned in situ swell measurements. The swell spectrum can be reconstructed from the inverted swell wave heights and combined with the wind wave inversion results to create the total directional wave spectrum. The method presented in this dissertation is fully dependent on information from HF radar data and does not no need calibration against in situ data for implementation; it can be applied to any beam forming system and operating frequency.

Share

COinS