Improve operational decisions with longer range forecasts

...but many companies are not making the most of the opportunity to plan ahead and avoid costly downtime.

Dr Edward Steele, Science Manager (Marine Applications), describes the benefits of his recent research that facilitates the interpretation of longer-range forecasts – and how adopting this approach could save the industry valuable time, resources and money.

Storms or calm weather, strong winds or varying wave heights – all have an impact on the viability of operations at sea. Currently, most offshore industry companies rely on deterministic forecasts that give them a single prediction. The appeal of these forecasts lies in the fact that they are easy to interpret, providing a simple yes/no outcome on which decisions can be made. However, because these are only typically reliable out to, at most, three to five days ahead, they don’t give industry decision-makers the chance to confidently plan on longer timescales.

The chaotic nature of the atmosphere and oceans means there is an inherent uncertainty associated with deterministic forecasts. This imposes a limit on how far ahead a particular detail can be predicted, but importantly the uncertainty (or ‘predictability’) varies from day to day depending on the weather situation. In contrast, probabilistic forecasts aim to quantify this uncertainty dynamically by using multiple projections, that account for both the incomplete analysis of the initial state of the system, and the approximations within the numerical schemes used to determine its evolution. The complete set of these constituent projections (or members) are referred to as an ensemble forecast and, since these are constructed such that results from each member will be equally likely, the proportion of ensemble members that forecast an event will be representative of the likelihood (or ‘probability’) of it occurring.

Finding the solution

From collaborative work with offshore industry decision-makers over the last few years, we know that if an offshore oil and gas company can factor in the likelihood of upcoming calm weather – instead of relying on a single prediction of a few days’ lead time – they will be able to make better operational decisions, earlier. However, exploiting the complex ensemble information has traditionally been difficult. Our research has focused on addressing this; making the science more accessible, useful and usable through the development of several tools that seek to combine the benefits of the simplicity of a single prediction with the skill of ensemble forecasts. For example, by applying certain criteria to the data or looking at the synoptic weather patterns over larger areas, we are able to come up with a much more reliable longer-term forecast. By presenting the information in a way that is both easy and intuitive to interpret, the full potential of the ensemble forecast can be realised.

^{Typical ensemble forecast (consisting of multiple predictions), with potential calm weather windows (green colour) indicated by the background shading to help to guide interpretation, based on known model skill (performance) and maximum operational Significant Wave Height limit of 3.5m.}

Better, earlier decisions

Unlocking the potential of an ensemble forecast increases confidence that a certain type of weather is more or less likely to occur in the coming weeks ahead. Let’s say, for instance, a company wants to plan a routine operation or maintenance activity. With just a single prediction, they could be fairly confident the calm conditions they need to successfully carry out the work are coming up in three days’ time. However, by properly using the full extent of the available uncertainty information, they would also not only have an early awareness of suitable conditions at a longer lead team, but also an indication as to if this is likely to be the only workable weather window over the coming weeks ahead. This additional context is vital for planning. It means the company’s asset manager can book equipment and coordinate staff with much greater confidence, as well as prioritising any activities accordingly.

Still keeping safe – but reducing operational costs

As the safety of employees is paramount to the offshore industry and the final approval to initiate a planned operation is made immediately prior to its start, it is essential for asset managers to only book equipment with the greatest possible confidence of being able to use it. However, from a planning perspective, as the event draws nearer in time, the greater the demand for resource grows. This potentially results in huge increases in the cost of securing offshore equipment, sometimes by over 1,000%! Thus, provided with a way to improve their decision-making, companies would be able to secure resources much earlier, ensuring much-reduced costs and avoiding losing thousands of pounds per day in unplanned downtime.

Extending the benefits beyond the marine sector

The benefits of longer-range forecasts may similarly be applied to other sectors. For example, if the marine renewable energy sector was able to integrate longer-range forecasts into their way of working, then they could streamline their operations, make them even more cost-efficient – and potentially help the UK deliver on its Industrial Strategy. Increased usability could also be exploited by other industries who require improved longer lead time decision-making and the management of specific risk, such as demand modelling for the water sector.

Translating science into services for the future

Regardless of the specific industry or sector, the decision-makers who drive businesses forward all want to reduce operational costs through gains in efficiency. A core part of our job at the Met Office is to consult and collaborate with customers to enable them to do so. This research not only helps to support this – it addresses a pressing industry need by developing the very latest in scientific and technological thinking.

For more information about this research, read Dr Edward Steele's blog.