Minimize Wind and wave forecasts for offshore operations and ship routing
The global demand for marine information

With global reserves of oil and gas forever diminishing, and demand increasing, exploration managers are looking to frontier areas offshore to provide new viable sources of hydrocarbons - areas which have unpredictable weather due to the sparsity and unreliability of metocean data. Firstly, in order to pinpoint sites for exploratory drilling, regional seismic surveys are undertaken for which a day of unanticipated down-time due to bad marine conditions represents an avoidable several thousand ECU loss. If a rig is eventually deployed to exploit this newly discovered reserve, both one-off and routine operations must be undertaken which can not only be disrupted by extreme events but also conditions typical of the area, such as swell waves with particular periods which can initiate resonant motion.

Many of these operations such as vessel operations, subsea operations (diving and ROVs), crane lifts, tanker loading and drilling, involve coupling of structures or unevenly distributed weighting. These constructions are particularly at risk from unacceptable motions initiated by waves which can lead to impacts and capsize events. For example, pipeline laying operations involving crane lifts, carried out from unmoored vessels using dynamic positioning, are fully at the mercy of the elements, requiring wave heights less than 3 metres for a window of at least 6 hours.
A shipping trawler in high seas  
There is significant growth in both size and number of ships navigating key routes around the Pacific rim and the Persian Gulf, and across the North Atlantic. Refit costs in dry dock for wave-induced slamming fatigue can be considerable, and currently on average three ships with a displacement of over 500 tons are sunk every week. Additionally, as coastal fish stocks become further depleted due to pollution and over-fishing, the fishing industry is having to explore more inaccessible places in which to deploy their nets.

Limitations to traditional data sources

In situ measurements of wind from ships over a 3 day period in 1990  

Forecasts of marine wind conditions can be provided by national meteorological agencies, although atmospheric models driven by the few in situ wind measurements available (gathered by ships using the major shipping routes, or in a few cases by dedicated buoy networks) may inaccurately represent atmospheric cyclonic features characteristic of storms not directly measured. Without accurate wind information, it is then impossible to accurately generate swell waves resulting from these storms in wave forecasts.


Wave models exist with which to predict marine conditions, but information on wave height, period and direction is needed to correct the output in order to accurately represent wave fields. In areas with a high level of marine activity such as the North West Shelf in Europe, organisations (including oil companies) in countries such as Norway, Denmark, The Netherlands and the UK are deploying buoy networks to measure both winds and wave heights with the spatial and temporal sampling necessary to drive models of the Shelf. These dedicated buoy networks have running costs of over a million pounds per annum for each 10 buoy network. In many cases incompatibility problems prevent neighbouring countries from sharing the data.

Even with the advent of such dedicated buoy networks, large parts of the oceans remain unobserved, including frontier areas of development where levels of marine activity are increasing, such as South East Asia. Each type of operation will have different requirements. In certain cases a ‘point-to-point’ forecast along a particular ship route may be needed or information on a specific location, for a specific time period or with a particular frequency of update.

The contribution of satellite data

ERS Scatterometer wind measurements over one day - 29th January 1991  

The active microwave sensors onboard ERS (altimeter, scatterometer, Synthetic Aperture Radar (SAR)) offer clear advantages for the study of marine winds and sea state. Firstly, they allow homogeneous, global and continuous coverage, at an improved resolution over conventional observations from ships and buoys. Due to frequent revisits, global wind fields from the scatterometer are likely to detect 3-4 of approximately 10 major cyclonic depressions which are over the Earth’s surface at any one time, hence improving atmospheric forecasts and resultant wave fields. Wave height observations, and measures of the period of swells generated by a far-off storm (sometimes thousands of kilometres away), can be used to improve wave forecasts - as the future sea state is dependent on accurate knowledge of the current situation.


Tropical Cyclone Oliver - forcing the atmospheric model using the scatterometer. Courtesy: Jet Propulsion Laboratory, USA  

Secondly, these data are of known quality and reliability, and are thoroughly validated before use. All these factors allow more accurate metocean and sea state forecasts to be produced, with a true representation of wind and wave fields. Figure 4 shows the improvement in the volume of wind measurements from in situ and satellite sources.

Improving meteorological forecasts

In various studies by the UK Meteorological Office (UKMO) and the European Centre for Medium-Range Weather Forecasts (ECMWF), the impact of the assimilation of scatterometer data was observed to be an improvement on the meteorological forecasts derived without scatterometer data of between 6-10% for surface pressure fields and 2-4% in the temperature and wind fields in the four day forecast from the global models. Figure 5 demonstrates the obvious improvements in description of cyclonic features afforded by the assimilation of scatterometer data into the operational weather prediction model used by the ECMWF. These figures show tropical cyclone Oliver during the Austral summer of 1992, located in the southern tropical Pacific where in situ observations are sparse.

Producing wave forecasts The wave model

Every wave forecast is produced using an initial state obtained by assimilating (blending) wind and wave data into a numerical model. Usually data are gathered into six hour windows centred around the assimilation times. At each time increment sea spectra are calculated, using an action balance equation where the evolution of the energy spectrum of the waves is forced by three contributory effects: energy input from wind forcing (produced by an atmospheric model), nonlinear wave-wave interaction which transfers energy between long and shorter wavelength waves, and energy dissipation from breaking waves. The wind field is used to force the wave model output which is then corrected locally with measurements of wave height, period and direction at each time step by drawing the model sea state towards the observed values using an optimal interpolation scheme.

Driving the model with ERS data

The availability of global high quality data on wind and wave fields in near real time through ERS requires the development of a suitable scheme for the rapid efficient assimilation of a high volume of data, within the lifecycle of the operational forecast.

In contrast to conventional observing systems which provide wind and wave data at fixed times and locations, satellite data are generated continuously at variable locations, requiring a four dimensional (both space and time dependent) assimilation method, achieved by assigning weights to the various data.

Assimilation efficiency also affects the resolution at which the gridded output from the model can be provided, although this is also dependent on the complexity of the model, the size of the geographical area covered and the available computing power.

Forcing the wave model with scatterometer winds

The impact of scatterometer data on the output from the VAG wave model. Courtesy Meteo-France  

Wave outputs also benefit significantly from quality wind forcing. Usually, scatterometer data are used to correct wind fields input into the the atmospheric models of meteorological offices. However, Meteo-France demonstrated the direct impact on wave forecast accuracy of scatterometer wind fields. The output from the model North Atlantic wave model VAG both when using scatterometer wind fields (SCAT) and wind vectors derived from the Meteo-France operational weather model ARPEGE (NODATA) was compared to wind vectors from drifting buoys and to altimeter wave heights, see figure 6. The error curves obtained denote a clear improvement in the presence of the scatterometer as the impact becomes relevant.

Correcting the wave fields with ERS wave heights

The magnitude of the difference between wave model output using simulated wind measurements only and ERS measured wave heights. Courtesy: European Centre for Medium Range Weather Forecasts  

Significant wave height is obtained from the radar altimeter, derived using the wave height-dependency of the rate at which the energy from the radar pulse is echoed back to the satellite. ECMWF assimilates altimeter data into the WAM wave model used to produce wave forecasts. Additonally, scatterometer winds are now used in the analysis of surface winds which are used to force the wave model. Figure 7, below, shows the differences between the wave heights output from the WAM model derived using simulated wind fields (no scatterometer data used), and the measured significant wave height from the ERS-1 altimeter for July 1993, illustrating the magnitude of the corrections needed. The new wave field produced with assimilation of altimeter data was validated with wave buoy data, and there was a reduction in standard deviation of 3% in the Western Atlantic, increasing to 15% in the Eastern Pacific where there are fewer in situ observations.

The difference the wave spectrum makes

Heights of waves alone are often insufficient to describe (and thus correct) the entire wave spectrum in the model. In contrast, the ERS SAR Wave Mode yields two-dimensional wave spectra (period and direction) at a resolution comparable to that of the wave model, every 200km along the satellite track. This information is particularly useful for local models used during dedicated support campaigns, due to the better characterisation of swell period.

Ensuring the quality of the wave data

Calibration of ERS-2 wave heights with in situ measurements. Courtesy: European Centre for Medium Range Weather Forecasts  

Satellite data must be carefully cross-calibrated against in situ wave-buoy and weather station measurements. For example, a comparison of ERS-2 altimeter measurements with buoy data for the period June 1995 to August 1996 is shown below. ERS-2 underpredicts wave heights as observed by buoys by about 8% (as compared with 15% for ERS-1, it's predecessor), which is then corrected for before assimilation.

The UK Meteorological Office

The UK Meteorological Office (UKMO) has been providing tailored products incorporating ERS data to a number of customers, predominantly shipping operators and offshore oil production companies, since 1992. Scatterometer, radar altimeter and SAR wave mode data are incorporated into the general forecast products such as the five-day ocean weather forecast. These are assimilated into numerical models which provide initial analysis fields of wind and wave data, upon which forecasts are based.

In addition to the general marine forecast products, specific support is available from the UKMO during particular projects. On average, around 1000 ship routing contracts are carried out annually with support provided to between thirty and forty customers at any one time. The model predictions are analysed by experienced master mariners and optimal routes are recommended to the ships’ masters. For the oil industry, the model outputs are extrapolated to the shallow waters around the UK to provide operational support in planning and execution of tasks such as drilling and lifting operations. A third customer area is the coastal protection and engineering industry, where users are supplied with data extrapolated from the nearest grid point to the location of interest via a transform incorporating the local coastal bathymetry. Companies such as Delft Hydraulics and HR Wallingford buy such products from the UKMO in order to provide sea-defence design services to local authorities.

The Neptune service from Meteo-France

image   image  
Wind forecast and Swell wave forecast for the North Eastern Atlantic. Courtesy Meteo-France  

Meteo-France, the French National Weather service, has introduced the Neptune service, which is based upon the numerical wave forecasts from both Meteo-France and from the ECMWF, both of which incorporate ERS altimeter and scatterometer data. The Meteo-France forecast is generated using their model of the atmosphere, ARPEGE, which ingests data from ships, buoys and infrared satellite data. ERS scatterometer data is used to provide corrections to the wind field (see previous page for more details).

The data are transmitted via the Inmarsat C satellite to an onboard desktop computer equipped with the appropriate interpretation software. The information can be viewed either in the form of a chart or in a user-defined format if the extraction of particular information, such as waves exceeding a particular height, is required. Information is included on both the current situation, and the forecast for the following days. Depending on the particular requirements of the customer, Neptune provides the following data:

  • Pressure, wind (speed and direction - figure 9) and swell (height, direction (figure 10) and period). Forecasts are provided with a spatial resolution of 0.5° longitude for a five-day period.
  • A chart of fronts and isobars containing both an analysis of the current situation and a forecast for the following three-day period.
  • Sea surface temperature analysis.

This service allows mariners to receive forecasts on a specific sea area, for example for delicate offshore operations (test drilling, pipeline deployment), or for fishing vessels. Another service allows shipping operators or yacht races to receive forecasts on a ‘point-to-point’ basis in near real-time over the course of a voyage. This service is costed according to the number of forecasts required.

ARGOSS-forecasts for offshore operations

Jacket installation. Courtesy Heeremac Engineering Services and ARGOSS  

Offshore heavy lifting operations, such as the positioning of rig structures, can be carried out only if the wave heights remain at a reasonably low amplitude for the duration of the operation and the period of the swell is sufficiently different from the natural periods of oscillation of the vessels involved. A project was carried out by Delft Hydraulics, supported by the Netherlands Remote Sensing Board, to provide sea state forecasts to Heeremac Engineering Services, an offshore engineering company, during a number of operations including heavy lifting and exploratory drilling. The figure shows one such operation - the installation of a jacket (support structure for an oil rig).

During such operations, decisions need to be made shortly before or during the task about suspension of work based on expected risk of damage or loss of life. In particular, swell (waves coming from a distant storm field) can cause strong motions of vessels if the local wind speed is low. These decisions must be based upon accurate short- and medium-term wave forecasts which include directional wave spectra at the site of the operation, to be used to forecast vessel motion, for example crane tip movement during lifting operations. The temporal resolution for the short-term (up to every three hours) and the quality of the forecast must be sufficient to allow this application.

ERS-1 radar altimeter and scatterometer products were used to correct ‘first-guess’ wave and wind fields (from the UKMO) of the numerical sea-state forecasting model (PHIDIAS) covering the North Sea and the NE Atlantic Ocean. An improved forecast was generated, which was validated using Waverider buoy and directional wave buoy measurements. It was found that assimilation of ERS significant wave height data was particularly useful in the presence of swell, as in the case of the trial carried out in August, see figure 12 above.

The forecast contained information on significant wave height, period, wave length and direction, as well as wind direction and strength. In addition, data on approaching swells were provided. These data were then used to compute parameters such as crane tip motions for lifting operations, to prevent high tension variations within the lifting equipment and dangerous impact forces between the lifted objects and other vessels. In one particular case, the improved forecast predicted a peak in the crane tip motions which was entirely missed by the ‘first guess’ fields.

In August 1995, a new private organisation, ARGOSS (Advisory and Research Group on Geo-Observation Systems and Services) was launched linking data suppliers, the space industry, research organisations and end-users. Among others, the activity of ARGOSS covers the further development of the wave forecasting system to provide fully operational systems and services. ARGOSS is providing contract research services on the use of satellite observations to improve wave forecasts. Within a partnership with around twenty other organizations ARGOSS is working on the development of an onboard decision support system to reduce risks of damage to vessels and tows. ARGOSS is contributing to a module for onboard updating of sea state analysis and forecast using both onboard measurements and satellite observations.

Current and future developments in marine forecasting

The efficiency of the assimilation of these data is improving - the European Centre for Medium-Range Weather Forecasting (ECMWF), who produce wave forecasts based on the assimilation of ERS data using the WAM model, have improved the output of their global model from 1.5° to 0.5° , due to technical advances within the system. This new resolution extends the wave forecasts from the open ocean to the coastal regions without the limitations of regional models.

Currently, ERS-2 satellite data is the only method of getting wide area wave information in near real time. In addition to the scatterometer on ERS-2, near real time satellite wind information will soon be available from the scatterometer (NSCAT) on the Japanese ADEOS platform. Following on from the ERS series, the next generation of Earth Observing satellite from ESA, ENVISAT, will be launched in 1999, and will have the capability to measure wave height, period and direction using advanced versions of the ERS SAR and radar altimeter sensors. The scatterometer measurements currently made by ERS-2 will be continued by the ASCAT instrument which will be flown on the EUMETSAT platform METOP, due for launch in 2001.

Cost versus benefit for the marine industries

The scale and immediacy of the risks facing industries which depend on the sea creates an enormous potential for application of both meteorological and sea-state forecasts. However, as with all private-sector industry, these customers are extremely cost and service sensitive, and can only invest in the necessary infrastructure if the information provided fulfills the user’s requirement in terms of accuracy, reliability, timeliness and cost.

The areas where wave forecasts are most useful are in the avoidance of extreme waveheights and certain dangerous wave periods in marine operations, such as the one shown in figure 12. This information is essential in order to plan operations as safely and efficiently as possible and to ensure objective decisions are made during weather sensitive operations, though benefits such as these are long-term and hard to quantify, as dangerous conditions do not occur during every operation. These benefits are therefore often not fully accounted for in cost-benefit analyses of these services, necessary in order to invest in the information. The further adoption of the use of wave forecasting services in general will therefore be induced by continuing illustration of the tangible benefits of such information in conjunction with initiatives in cost minimisation to enable the economical justification of the investment.

Where satellite data really comes into its own as a source of information is in frontier regions for offshore operations such as the North Western Approaches or South East Asia, or for fishing operations in the Southern Ocean. In areas where there are insufficient or sub-optimal in situ measurements being made, ERS-2 data is essential to produce optimum forecasts of winds and waves.


For further information on the services and products covered here, contact:

Jack S. Hopkins
The Meteorological Office
London Road
RG12 2SZ (UK)
Tel.: +44 1344 856684
Fax.: +44 1344 854906

J. Poitevin
42, avenue Gustave Coriolis
31057 Toulouse Cedex (France)
Tel: +33 5 61 07 80 80
Fax: +33 5 61 07 80 09

H. Wensink
P.O. Box 61
8325 ZH Vollenhove (The Netherlands)
Tel: +31 527 242299
Fax: +31 527 242016

P. Janssen
European Centre for Medium-Range Weather Forecasts
Shinfield Park
Berkshire RG2 9AX (UK)
Tel.: +44 118 9499116
Fax.: +44 118 9868450


The material illustrated here was provided by the following organisations: Shell UK Exploration & Production, European Centre for Medium-Range Weather Forecasts (ECMWF), Jet Propulsion Laboratory, Meteo-France, ARGOSS, Heeremac Engineering Services, UK Offshore Operators Association
ESA gratefully acknowledge all contributions

Read more