Search Results
You are looking at 1 - 7 of 7 items for
- Author or Editor: Brian Emery x
- Refine by Access: All Content x
Uncertainty Estimates for SeaSonde HF Radar Ocean Current ObservationsAbstract
HF radars typically produce maps of surface current velocities without estimates of the measurement uncertainties. Many users of HF radar data, including spill response and search and rescue operations, incorporate these observations into models and would thus benefit from quantified uncertainties. Using both simulations and coincident observations from the baseline between two operational SeaSonde HF radars, we demonstrate the utility of expressions for estimating the uncertainty in the direction obtained with the Multiple Signal Classification (MUSIC) algorithm. Simulations of radar backscatter using surface currents from the Regional Ocean Modeling System show a close correspondence between direction of arrival (DOA) errors and estimated uncertainties, with mean values of 15° at 10 dB, falling to less than 3° at 30 dB. Observations from two operational SeaSondes have average DOA uncertainties of 2.7° and 3.8°, with a fraction of the observations (10.5% and 7.1%, respectively) having uncertainties of >10°. Using DOA uncertainties for data quality control improves time series comparison statistics between the two radars, with
Abstract
HF radars typically produce maps of surface current velocities without estimates of the measurement uncertainties. Many users of HF radar data, including spill response and search and rescue operations, incorporate these observations into models and would thus benefit from quantified uncertainties. Using both simulations and coincident observations from the baseline between two operational SeaSonde HF radars, we demonstrate the utility of expressions for estimating the uncertainty in the direction obtained with the Multiple Signal Classification (MUSIC) algorithm. Simulations of radar backscatter using surface currents from the Regional Ocean Modeling System show a close correspondence between direction of arrival (DOA) errors and estimated uncertainties, with mean values of 15° at 10 dB, falling to less than 3° at 30 dB. Observations from two operational SeaSondes have average DOA uncertainties of 2.7° and 3.8°, with a fraction of the observations (10.5% and 7.1%, respectively) having uncertainties of >10°. Using DOA uncertainties for data quality control improves time series comparison statistics between the two radars, with
Direction Finding and Likelihood Ratio Detection for Oceanographic HF RadarsAbstract
Previous work with simulations of oceanographic high-frequency (HF) radars has identified possible improvements when using maximum likelihood estimation (MLE) for direction of arrival; however, methods for determining the number of emitters (here defined as spatially distinct patches of the ocean surface) have not realized these improvements. Here we describe and evaluate the use of the likelihood ratio (LR) for emitter detection, demonstrating its application to oceanographic HF radar data. The combined detection–estimation methods MLE-LR are compared with multiple signal classification method (MUSIC) and MUSIC parameters for SeaSonde HF radars, along with a method developed for 8-channel systems known as MUSIC-Highest. Results show that the use of MLE-LR produces similar accuracy, in terms of the RMS difference and correlation coefficients squared, as previous methods. We demonstrate that improved accuracy can be obtained for both methods, at the cost of fewer velocity observations and decreased spatial coverage. For SeaSondes, accuracy improvements are obtained with less commonly used parameter sets. The MLE-LR is shown to be able to resolve simultaneous closely spaced emitters, which has the potential to improve observations obtained by HF radars operating in complex current environments.
Significance Statement
We identify and test a method based on the likelihood ratio (LR) for determining the number of signal sources in observations subject to direction finding with maximum likelihood estimation (MLE). Direction-finding methods are used in broad-ranging applications that include radar, sonar, and wireless communication. Previous work suggests accuracy improvements when using MLE, but suitable methods for determining the number of simultaneous signal sources are not well known. Our work shows that the LR, when combined with MLE, performs at least as well as alternative methods when applied to oceanographic high-frequency (HF) radars. In some situations, MLE and LR obtain superior resolution, where resolution is defined as the ability to distinguish closely spaced signal sources.
Abstract
Previous work with simulations of oceanographic high-frequency (HF) radars has identified possible improvements when using maximum likelihood estimation (MLE) for direction of arrival; however, methods for determining the number of emitters (here defined as spatially distinct patches of the ocean surface) have not realized these improvements. Here we describe and evaluate the use of the likelihood ratio (LR) for emitter detection, demonstrating its application to oceanographic HF radar data. The combined detection–estimation methods MLE-LR are compared with multiple signal classification method (MUSIC) and MUSIC parameters for SeaSonde HF radars, along with a method developed for 8-channel systems known as MUSIC-Highest. Results show that the use of MLE-LR produces similar accuracy, in terms of the RMS difference and correlation coefficients squared, as previous methods. We demonstrate that improved accuracy can be obtained for both methods, at the cost of fewer velocity observations and decreased spatial coverage. For SeaSondes, accuracy improvements are obtained with less commonly used parameter sets. The MLE-LR is shown to be able to resolve simultaneous closely spaced emitters, which has the potential to improve observations obtained by HF radars operating in complex current environments.
Significance Statement
We identify and test a method based on the likelihood ratio (LR) for determining the number of signal sources in observations subject to direction finding with maximum likelihood estimation (MLE). Direction-finding methods are used in broad-ranging applications that include radar, sonar, and wireless communication. Previous work suggests accuracy improvements when using MLE, but suitable methods for determining the number of simultaneous signal sources are not well known. Our work shows that the LR, when combined with MLE, performs at least as well as alternative methods when applied to oceanographic high-frequency (HF) radars. In some situations, MLE and LR obtain superior resolution, where resolution is defined as the ability to distinguish closely spaced signal sources.
Improving Surface Current Resolution Using Direction Finding Algorithms for Multiantenna High-Frequency RadarsAbstract
While land-based high-frequency (HF) radars are the only instruments capable of resolving both the temporal and spatial variability of surface currents in the coastal ocean, recent high-resolution views suggest that the coastal ocean is more complex than presently deployed radar systems are able to reveal. This work uses a hybrid system, having elements of both phased arrays and direction finding radars, to improve the azimuthal resolution of HF radars. Data from two radars deployed along the U.S. East Coast and configured as 8-antenna grid arrays were used to evaluate potential direction finding and signal, or emitter, detection methods. Direction finding methods such as maximum likelihood estimation generally performed better than the well-known multiple signal classification (MUSIC) method given identical emitter detection methods. However, accurately estimating the number of emitters present in HF radar observations is a challenge. As MUSIC’s direction-of-arrival (DOA) function permits simple empirical tests that dramatically aid the detection process, MUSIC was found to be the superior method in this study. The 8-antenna arrays were able to provide more accurate estimates of MUSIC’s noise subspace than typical 3-antenna systems, eliminating the need for a series of empirical parameters to control MUSIC’s performance. Code developed for this research has been made available in an online repository.
Abstract
While land-based high-frequency (HF) radars are the only instruments capable of resolving both the temporal and spatial variability of surface currents in the coastal ocean, recent high-resolution views suggest that the coastal ocean is more complex than presently deployed radar systems are able to reveal. This work uses a hybrid system, having elements of both phased arrays and direction finding radars, to improve the azimuthal resolution of HF radars. Data from two radars deployed along the U.S. East Coast and configured as 8-antenna grid arrays were used to evaluate potential direction finding and signal, or emitter, detection methods. Direction finding methods such as maximum likelihood estimation generally performed better than the well-known multiple signal classification (MUSIC) method given identical emitter detection methods. However, accurately estimating the number of emitters present in HF radar observations is a challenge. As MUSIC’s direction-of-arrival (DOA) function permits simple empirical tests that dramatically aid the detection process, MUSIC was found to be the superior method in this study. The 8-antenna arrays were able to provide more accurate estimates of MUSIC’s noise subspace than typical 3-antenna systems, eliminating the need for a series of empirical parameters to control MUSIC’s performance. Code developed for this research has been made available in an online repository.
Measurement of Antenna Patterns for Oceanographic Radars Using Aerial DronesAbstract
A new method is described employing small drone aircraft for antenna pattern measurements (APMs) of high-frequency (HF) oceanographic radars used for observing ocean surface currents. Previous studies have shown that accurate surface current measurements using HF radar require APMs. The APMs provide directional calibration of the receive antennas for direction-finding radars. In the absence of APMs, so-called ideal antenna patterns are assumed and these can differ substantially from measured patterns. Typically, APMs are obtained using small research vessels carrying radio signal sources or transponders in circular arcs around individual radar sites. This procedure is expensive because it requires seagoing technicians, a vessel, and other equipment necessary to support small-boat operations. Furthermore, adverse sea conditions and obstacles in the water can limit the ability of small vessels to conduct APMs. In contrast, it is shown that drone aircraft can successfully conduct APMs at much lower cost and in a broader range of sea states with comparable accuracy. Drone-based patterns can extend farther shoreward, since they are not affected by the surfzone, and thereby expand the range of bearings over which APMs are determined. This simplified process for obtaining APMs can lead to more frequent calibrations and improved surface current measurements.
Abstract
A new method is described employing small drone aircraft for antenna pattern measurements (APMs) of high-frequency (HF) oceanographic radars used for observing ocean surface currents. Previous studies have shown that accurate surface current measurements using HF radar require APMs. The APMs provide directional calibration of the receive antennas for direction-finding radars. In the absence of APMs, so-called ideal antenna patterns are assumed and these can differ substantially from measured patterns. Typically, APMs are obtained using small research vessels carrying radio signal sources or transponders in circular arcs around individual radar sites. This procedure is expensive because it requires seagoing technicians, a vessel, and other equipment necessary to support small-boat operations. Furthermore, adverse sea conditions and obstacles in the water can limit the ability of small vessels to conduct APMs. In contrast, it is shown that drone aircraft can successfully conduct APMs at much lower cost and in a broader range of sea states with comparable accuracy. Drone-based patterns can extend farther shoreward, since they are not affected by the surfzone, and thereby expand the range of bearings over which APMs are determined. This simplified process for obtaining APMs can lead to more frequent calibrations and improved surface current measurements.
Abstract
The performance of a network of five CODAR (Coastal Ocean Dynamics Application Radar) SeaSonde high-frequency (HF) radars, broadcasting near 13 MHz and using the Multiple Signal Classification (MUSIC) algorithm for direction finding, is described based on comparisons with an array of nine moorings in the Santa Barbara Channel and Santa Maria basin deployed between June 1997 and November 1999. Eight of the moorings carried vector-measuring current meters (VMCMs), the ninth had an upward-looking ADCP. Coverage areas of the HF radars and moorings included diverse flow and sea-state regimes. Measurement depths were ∼1 m for the HF radars, 5 m for the VMCMs, and 3.2 m for the ADCP bin nearest to the surface. Comparison of radial current components from 18 HF radar–mooring pairs yielded rms speed differences of 7–19 cm s−1 and correlation coefficients squared (r 2) in the range of 0.39–0.77. Spectral analysis showed significant coherence for frequencies below 0.1 cph (periods longer than 10 h). At higher frequencies no significant coherence was found, and noise levels corresponding to 6 cm s−1 rms were evident in the radar data. Errors in the radar bearing determination were found in 10 out of 18 comparisons, with a typical magnitude of 5°–10°, and a maximum of 19°. The effects of bearing errors on total vector currents were evaluated using a simple flow field and measured bearing errors, showing up to 15% errors in computed flow speeds, and up to ∼9° errors in flow directions.
Abstract
The performance of a network of five CODAR (Coastal Ocean Dynamics Application Radar) SeaSonde high-frequency (HF) radars, broadcasting near 13 MHz and using the Multiple Signal Classification (MUSIC) algorithm for direction finding, is described based on comparisons with an array of nine moorings in the Santa Barbara Channel and Santa Maria basin deployed between June 1997 and November 1999. Eight of the moorings carried vector-measuring current meters (VMCMs), the ninth had an upward-looking ADCP. Coverage areas of the HF radars and moorings included diverse flow and sea-state regimes. Measurement depths were ∼1 m for the HF radars, 5 m for the VMCMs, and 3.2 m for the ADCP bin nearest to the surface. Comparison of radial current components from 18 HF radar–mooring pairs yielded rms speed differences of 7–19 cm s−1 and correlation coefficients squared (r 2) in the range of 0.39–0.77. Spectral analysis showed significant coherence for frequencies below 0.1 cph (periods longer than 10 h). At higher frequencies no significant coherence was found, and noise levels corresponding to 6 cm s−1 rms were evident in the radar data. Errors in the radar bearing determination were found in 10 out of 18 comparisons, with a typical magnitude of 5°–10°, and a maximum of 19°. The effects of bearing errors on total vector currents were evaluated using a simple flow field and measured bearing errors, showing up to 15% errors in computed flow speeds, and up to ∼9° errors in flow directions.
Abstract
HF radars measure ocean surface currents near coastlines with a spatial and temporal resolution that remains unmatched by other approaches. Most HF radars employ direction-finding techniques, which obtain the most accurate ocean surface current data when using measured, rather than idealized, antenna patterns. Simplifying and automating the antenna pattern measurement (APM) process would improve the utility of HF radar data, since idealized patterns are widely used. A method is presented for obtaining antenna pattern measurements for direction-finding HF radars from ships of opportunity. Positions obtained from the Automatic Identification System (AIS) are used to identify signals backscattered from ships in ocean current radar data. These signals and ship position data are then combined to determine the HF radar APM. Data screening methods are developed and shown to produce APMs with low error when compared with APMs obtained with shipboard transponder-based approaches. The analysis indicates that APMs can be reproduced when the signal-to-noise ratio (SNR) of the backscattered signal is greater than 11 dB. Large angular sectors of the APM can be obtained on time scales of days, with as few as 50 ships.
Abstract
HF radars measure ocean surface currents near coastlines with a spatial and temporal resolution that remains unmatched by other approaches. Most HF radars employ direction-finding techniques, which obtain the most accurate ocean surface current data when using measured, rather than idealized, antenna patterns. Simplifying and automating the antenna pattern measurement (APM) process would improve the utility of HF radar data, since idealized patterns are widely used. A method is presented for obtaining antenna pattern measurements for direction-finding HF radars from ships of opportunity. Positions obtained from the Automatic Identification System (AIS) are used to identify signals backscattered from ships in ocean current radar data. These signals and ship position data are then combined to determine the HF radar APM. Data screening methods are developed and shown to produce APMs with low error when compared with APMs obtained with shipboard transponder-based approaches. The analysis indicates that APMs can be reproduced when the signal-to-noise ratio (SNR) of the backscattered signal is greater than 11 dB. Large angular sectors of the APM can be obtained on time scales of days, with as few as 50 ships.
Abstract
Dense arrays of surface drifters are used to quantify the flow field on time and space scales over which high-frequency (HF) radar observations are measured. Up to 13 drifters were repetitively deployed off the Santa Barbara and San Diego coasts on 7 days during 18 months. Each day a regularly spaced grid overlaid on a 1-km2 (San Diego) or 4-km2 (Santa Barbara) square, located where HF radar radial data are nearly orthogonal, was seeded with drifters. As drifters moved from the square, they were retrieved and replaced to maintain a spatially uniform distribution of observations within the sampling area during the day. This sampling scheme resulted in up to 56 velocity observations distributed over the time (1 h) and space (1 and 4 km2) scales implicit in typical surface current maps from HF radar. Root-mean-square (RMS) differences between HF radar radial velocities obtained using measured antenna patterns, and average drifter velocities, are mostly 3–5 cm s−1. Smaller RMS differences compared with past validation studies that employ current meters are due to drifter resolution of subgrid-scale velocity variance included in time and space average HF radar fields. Roughly 5 cm s−1 can be attributed to sampling on disparate time and space scales. Despite generally good agreement, differences can change dramatically with time. In one instance, the difference increases from near zero to more than 20 cm s−1 within 2 h. The RMS difference and bias (mean absolute difference) for that day exceed 7 and 12 cm s−1, respectively.
Abstract
Dense arrays of surface drifters are used to quantify the flow field on time and space scales over which high-frequency (HF) radar observations are measured. Up to 13 drifters were repetitively deployed off the Santa Barbara and San Diego coasts on 7 days during 18 months. Each day a regularly spaced grid overlaid on a 1-km2 (San Diego) or 4-km2 (Santa Barbara) square, located where HF radar radial data are nearly orthogonal, was seeded with drifters. As drifters moved from the square, they were retrieved and replaced to maintain a spatially uniform distribution of observations within the sampling area during the day. This sampling scheme resulted in up to 56 velocity observations distributed over the time (1 h) and space (1 and 4 km2) scales implicit in typical surface current maps from HF radar. Root-mean-square (RMS) differences between HF radar radial velocities obtained using measured antenna patterns, and average drifter velocities, are mostly 3–5 cm s−1. Smaller RMS differences compared with past validation studies that employ current meters are due to drifter resolution of subgrid-scale velocity variance included in time and space average HF radar fields. Roughly 5 cm s−1 can be attributed to sampling on disparate time and space scales. Despite generally good agreement, differences can change dramatically with time. In one instance, the difference increases from near zero to more than 20 cm s−1 within 2 h. The RMS difference and bias (mean absolute difference) for that day exceed 7 and 12 cm s−1, respectively.
