Directional wave measurements using an autonomous vessel
- 896 Downloads
An autonomous vessel, the Offshore Sensing Sailbuoy, was used for wave measurements near the Ekofisk oil platform complex in the North Sea (56.5º N, 3.2º E, operated by ConocoPhillips) from 6 to 20 November 2015. Being 100 % wind propelled, the Sailbuoy has two-way communication via the Iridium network and has the capability for missions of 6 months or more. It has previously been deployed in the Arctic, Norwegian Sea and the Gulf of Mexico, but the present study was the first test for wave measurements. During the campaign the Sailbuoy held position about 20 km northeast of Ekofisk (on the lee side) during rough conditions. Mean wind speed measured at Ekofisk during the campaign was 9.8 m/s, with a maximum of 20.4 m/s, with wind mostly from south and southwest. A Datawell MOSE G1000 GPS-based 2 Hz wave sensor was mounted on the Sailbuoy. Mean significant wave height (Hs 1 min) measured was 3 m, whereas maximum Hs was 6 m. Mean wave period was 7.7 s, while maximum wave height, Hmax, was 12.6 m. These measurements have been compared with non-directional Waverider observations at the Ekofisk complex. The agreement between the two data sets was very good, with a mean percent absolute error of 7 % and a linear correlation coefficient of 0.97. The wave frequency spectra measured by the two instruments compared very well, except for low Hs (∼1 m), where the motion of the vessel seemed to influence the measurements. Nevertheless, the Sailbuoy performed well during this campaign, and results suggest that it is a suitable platform for wave measurements in a broad range of sea conditions.
KeywordsSurface Waves Directional Autonomous Vessel Measurements North Sea
Accurate ocean wave observations and forecasts are in increasing demand, and of interest to users involved in the offshore industry, shipping, offshore wind energy industry and prospecting of bridges, roads and other near shore constructions. For example, along the coast of Norway, several fjord-crossing road construction projects involving long bridges are under way, and detailed mapping of the local wave climate is required.
The North Sea is an area with rough ocean conditions year round (Grabemann and Weisse 2008). A significant wave height, Hs, exceeding 19 m in the Northern North Sea is suggested to be realistic in the worst case scenarios described by Reistad et al. (2005). Aarnes et al. (2012) suggest a 100 year return value of Hs for the North Sea of 16–20 m, depending on method. However, there is a negative trend in Hs and wind speed in the region in recent decades (Young et al. 2011). There is high shipping density and substantial offshore oil activity in the area. Multiple offshore wind farms are under development and detailed information about the wave climate has practical and economical value, since construction dimensioning off shore is dependent on wave climate information.
Accurate ocean wave measurements are important for verification of ocean wave forecasts and wave climate mapping (Steward 2008). The observation network at sea is coarse, and there is a consistent lack of wave measurements to verify model predictions (Reistad et al. 2005). Long-term in situ wave monitoring programmes tend to be interrupted as a result of environmental stresses such as bio-fouling and severe weather (Herbers et al. 2012; Manov et al. 2004).
Wave-forecasting models in use by meteorological agencies are based on integrations of the directional wave spectrum discretized in direction and frequency (or wavenumber), (See e.g. Komen et al. 1994 or Steward 2008). The forecasts follow individual components of the wave spectrum in space and time, allowing each component to grow or decay depending on advection, energy input by local winds, energy sinks by dissipation, such as wave breaking and bottom friction, and repartition of energy by non-linear interactions. Detailed and accurate wave measurements to validate these models are consequently of interest.
Ocean waves are measured remotely by observing the sea state, by satellite altimeters (Queffeulou 2004) or by Synthetic Aperture Radars (SAR) (Li et al. 2008). In situ measurements are normally carried out by accelerometers or Global Positioning System (GPS) sensors mounted on floating buoys (e.g. Jeans et al. 2003) or, in the case of shallow water, Acoustic Doppler Current Profilers (ADCPs) or other wave gauges mounted on the sea floor (Herbers and Lentz 2010). Alternative methods, such as mounting of ADCPs on autonomous underwater vehicles (Haven and Terray 2015), the Surpact Waverider (Reverdin et al. 2013), and ship-mounted acoustic sensors (Christensen et al. 2013) have also been demonstrated.
Here we describe a novel methodology, namely a GPS-based 2 Hz motion sensor mounted on a small autonomous surface vehicle. This configuration can have advantages such as low costs (no ship time needed), independence of water depth, flexibility, and mobility. Other advantages using GPS based sensors are size, robustness, and no need of calibration (Herbers et al. 2012).The measurement platform is tested over a period of about 2 weeks during rough weather conditions in the Central North Sea and measurements are compared to observations carried out by a permanently installed traditional Waverider. The main aim of this study is to test the Sailbuoy for its operational capability for measuring wave parameters in rough ocean conditions.
The Sailbuoy has proven its endurance and navigation capability through various missions including a transect from Bergen, Norway to Iceland, Bergen to Scotland, a mission north of Svalbard close to the marginal ice zone, and surveys in the northern Gulf of Mexico and off Gran Canaria. For a report on the navigation capability and efficacy we refer to Fer and Peddie (2012), for a report on near-surface temperature, salinity and oxygen concentration measurements see Fer and Peddie (2013) for an application in the northern Gulf of Mexico see Ghani et al. (2014).
For the experiment described here, the Offshore Sensing Sailbuoy Wave (SB Wave hereafter) was equipped with the Datawell MOSE-G1000 wave sensor (www.datawell.nl). This is a three-dimensional motion sensor based on single GPS and measures the translational motion of the GPS antenna in three frequency or period regimes each with its own precision: high-frequency motion (1–100-s periods, 1 cm precision), low frequency motion (10–1000 s periods, several cm precision), and GPS position (infinitely long periods, 10 m precision). An indoor version of the sensor was installed in the payload section of the SB Wave, and the external GPS antenna was integrated at the rear part of the buoy (Fig. 2). The antenna is free from obstructions and is elevated above the deck to mitigate potential signal loss due to wave wash over. For details on the MOSE-G1000 we refer to the manufacturer’s reference manual.
The reference measurements at the Ekofisk platform are from a Datawell non-directional Waverider.
The SB Wave was deployed during a cruise of the research vessel Håkon Mosby (cruise number HM2015 623), on 30 October 2015, 18:00 UTC, at 56° 32′ N along the track of the vessel towards the FINO1 platform in the North Sea. Ekofisk is in block 2/4 of the Norwegian sector of the North Sea about 320 km southwest of Stavanger (Fig. 3a).
Measurement positions and duration
Position (latitude; longitude)
Measurement period (2015, UTC)
56° 45.0′ N; 3° 9.0′ E
7 November 0000–14 November 1000
56° 38.1′ N; 3° 11.9′ E
14 November 1200–20 November 0930
56° 32.9′ N; 3° 6.2′ E
The Sailbuoy was successful in station keeping and maintained a position within ±2 km of the way points (Fig. 3b).
The MOSE-G1000 sensor was set to sample at 2 Hz. The internal logging included the high-frequency (HF) string containing horizontal and vertical displacements and a data quality flag. Furthermore, every 10 s the position, together with horizontal dilution of precision (HDOP) and vertical dilution of precision (VDOP), were logged. The sensor was initialized by power (82 mA), and a new data file was written every 30 min, logging data for 25 min (3000 data points). The sensor was left on continuously, thanks to the short duration of the experiment. While the entire raw data field was logged internally, the data were also transferred to the Sailbuoy for data reduction and relay of wave parameters via satellite. The first 125 data points (of 3000; approximately 1 min) were excluded to avoid a possible contamination by filter effects or file book-keeping.
There are, therefore, two data sets as follows: (1) full resolution, raw sensor data for post processing using various methods to infer wave parameters, and (2) twice-hourly real-time relayed data including time, position and key wave parameters inferred onboard from a zero-crossing analysis of typically 2875 data points. For further details on the data processing and reduction, see Fer and Peddie (2016).
detect individual waves using zero-crossing of the vertical displacement record (retain every second zero-crossing to define a complete wave, this is loosely referred to as the number of zero-crossings),
calculate wave height and wave period for each individual wave,
sort the wave heights (book-keeping the corresponding periods),
calculate the significant wave height (Hs, average of the largest 1/3rd sorted wave heights) and the maximum wave height (Hmax), and
calculate mean zero-crossing period (T0, mean period of all the waves in the record), and the significant period (Ts, average period of the waves used to define Hs).
4 Results and discussion
The period of deployment was dominated by a series of passing lows to the west of Ekofisk and the predominant wind direction was southwest, as common for the season. This means that SB Wave was mostly downwind of Ekofisk. Mean wind speed over the 14 days of deployment was 9.8 m/s (U10m—10 min average), with a maximum wind speed of 21.4 m/s on 13 November 2130 UTC (Fig. 4). Given the vigorous wind conditions, we assume that the relative contribution of swell was minor during the field campaign.
Statistics of SB Wave to Waverider comparison for significant wave height, Hs
Raw data downloaded from the instrument are further analysed for more accurate estimates of the wave parameters, using spectral analysis. For spectral analysis of both SB Wave and Waverider data we used the DIWASP Matlab toolbox (www.metocean.co.nz) and the direct Fourier transform method (DFTM) originally developed by Barber (1961). DFTM is used with a frequency resolution of 0.01 Hz in the range 0.05 to 1 Hz, and a direction resolution of 2°. Note, however, that the Waverider is non-directional, and the directional spectra are calculated for SB Wave only. From the analysis, the significant wave height, Hs, peak period, Tp, and for SB Wave, direction of spectral peak, DTp and the dominant direction, Dp are extracted.
Finally, examples of directional spectra from SB Wave (calculated by DIWASP and the DFTM method) are shown in Fig. 15. The six examples are the same segments as in Fig. 11. Two examples for low (∼1 m), moderate (∼3.2 m) and high Hs (∼6 m) are shown to demonstrate consistency in the measurements of energy distribution in frequency and direction. It appears that SB Wave is able to observe rather well-defined wave directions, even in vigorous seas. In the upper two panels, a weak swell propagating towards the southeast can be observed together with wind sea propagating towards the northeast.
5 Summary and conclusion
Recent developments in global positioning system (GPS) technology have enabled in situ ocean wave measurements at a relatively low cost using surface-following buoys (Herbers et al. 2012). In the experiment presented here, a Datawell MOSE G1000 sensor was placed in an autonomous vessel, the Offshore Sensing Sailbuoy. A data set was collected between 7 and 20 November 2015, near the Ekofisk oil field in the North Sea. For sensor inter comparison and data validation, the measurement position was co-located (10 to 20 km) with a bottom-anchored Waverider buoy. The measurement period covers from quiescent periods with Hs on the order 1 m to energetic periods with Hs reaching 6 m with maximum wave heights in excess of 10 m. The peak period of the wave spectrum was approximately 5 s for the Hs on the order 1 m, and 10–12 s for Hs exceeding 5 m.
The Sailbuoy delivered two data sets as follows: (i) twice-hourly real-time relayed wave parameters processed on board using a zero-crossing analysis; and (ii) full resolution, raw sensor data for post processing using spectral methods to infer wave parameters. First, the real-time data set is compared with the Waverider measurements to showcase the operational merit of the Sailbuoy. The agreement is very good with a fractional error less than 10 %, and without a significant trend when binned in Hs.
Next, the wave measurements from the Sailbuoy are compared with the Waverider measurements using spectral analysis of 30 min time series of 2 Hz sampling rate. The agreement between the two data sets is good with a linear correlation coefficient of 0.97, a bias of 1 cm, a root-mean-square error of 27 cm, and a mean percent absolute error of 7 %. The observations presented here suggest that the Sailbuoy is a suitable platform for wave measurements delivering reliable real time data as well as accurate post-processed data. This is particularly true for spectral measurements for medium (∼3 m) and high (∼6 m) wave heights and for integrated values (Hs, and Tp).
The advantages of using GPS based wave sensors have already been discussed by Herbers et al. (2012). A GPS based sensor placed in the Sailbuoy could be a particularly attractive and cost-efficient alternative for short term measurement programmes (weeks to months) carried out as part of coastal and offshore construction project planning.
The emerging possibility to use robust autonomous platforms for wave measurements at relatively low costs provides an opportunity for more dense wave observations in the future, of benefit for forecasters and commercial users. The main aim of this study is to test the Sailbuoy for its operational capability for measuring wave parameters in rough ocean conditions. Future studies should involve focus on hull effects on measurements of low and high-frequency waves, comparison with other directional sensors as well as the use of multiple platforms for mapping of horizontal coherence.
This study is partly funded by the Norwegian Centre of Offshore Wind Energy (NORCOWE). The Waverider raw data and the wind data have kindly been provided by the operator of Ekofisk, ConocoPhillips. The deployment of the Sailbuoy was conducted from a University of Bergen research cruise. We thank ConocoPhillips and Skandi Marøy for retrieving the Sailbuoy at Ekofisk, and two reviewers for their comments on the manuscript.
- Barber, NF (1961) The directional resolving power of an array of wave detectors, Ocean Wave Spectra. Prentice Hall. Inc. pp.137-150Google Scholar
- Fer I, Peddie D. (2012) Navigation performance of the Sailbuoy. Bergen - Scotland mission, 12 pp. www.sailbuoy.noGoogle Scholar
- Fer I, Peddie D (2013) Near surface oceanographic measurements using the Sailbuoy. OCEANS 2013 MTS/IEEE, IEEE, 15Google Scholar
- Fer I, Peddie D (2016) Report on wave measurements using the Sailbuoy wave. www.sailbuoy.no/files/Report_Ekofisk_Nov2015.pdf
- Ghani MH, Hole LR, Fer I Kourafalou VH, Wienders N, Kang H, Drushka K. and Peddie D (2014) The Sailbuoy remotely-controlled unmanned vessel: measurements of near surface temperature, salinity and oxygen concentration in the Northern Gulf of Mexico. Methods in Oceanography, 10, 104-121, doi: http://dx.doi.org/10.1016/j.mio.2014.08.00
- Haven S, & Terray EA. (2015) Surface wave measurements from an autonomous underwater vehicle. In Current, Waves and Turbulence Measurement (CWTM), 2015 IEEE/OES Eleventh (pp. 1-7). IEEEGoogle Scholar
- Jeans G, Bellamy I, de Vries JJ, & van Weert P (2003) Sea trial of the new Datawell GPS directional Waverider. In Current Measurement Technology, 2003. Proceedings of the IEEE/OES Seventh Working Conference on (pp. 145-147). IEEEGoogle Scholar
- Komen G, Cavaleri L, Donelan M, Hasselmann K, Hasselmann S, & Janssen PAEM (1994) Dynamics and Modelling of Ocean WavesGoogle Scholar
- Steward HR (2008) Introduction to physical oceanography. Department of Oceanography. A & M University, Texas. http://oceanworld.tamu.edu/resources/ocng_textbook/PDF_files/book.pdfGoogle Scholar
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.