Economics 101 states that if you want to know if it is worth investing in a project, the basic approach is to define the costs and benefits for your base case (or what would happen if you 'do nothing') and compare them with the outcomes under different investment scenarios. We take account of risk and uncertainty, but typically focus on the sensitivity to changes in demand, costs, phasing, charging and the other assumptions related to 'normal' operations.
With weather the uncertainty is greater: to quote the UK Met Office "Everyone knows that weather forecasts go wrong sometimes!" End-users such as airports and airlines want to know if a specific 'abnormal' weather event will disrupt their normal activities, but the uncertainty in forecasting means that it is difficult to be precise. Investment in forecasting capability allows greater accuracy e.g. the "10% chance that Heathrow will get...." could be refined to "a 5% chance...", but there is still the risk of false alarms and being caught unprepared.
An airport incurs costs in preparing for an abnormal weather event; if the event occurs then disruption is minimised, but if it does not, flights may have been cancelled unnecessarily. If the event is unexpected, or the user did not have enough confidence in the forecast to take action, then the costs are much greater. Decision-makers will typically balance these costs and set thresholds of probability at which they will start to react. The concept of planned and unplanned user responses related to the provision of probabilistic weather forecasts is central to this type of investment appraisal.
On the surface, a refinement in the probability provides a "better" forecast, but is it good enough? A recent study by Helios for the UK Met Office found that although a 5% probability of bad weather is low, if the event occurs, it can have a large impact on aviation because it was unexpected. Moving from a 10% probability to 5% therefore still carried a high unplanned cost when the event occurred and this cost is greater at congested airports such as Heathrow. For aviation users a weather forecast therefore needs to satisfy the following criteria:
- High confidence (ideally less than 1% or greater than 99% probability)
- Geographical precision (airport-specific)
- Timely (to allow users to respond)
How much would you need to invest in your weather forecasting services to meet those criteria? It may be time to rethink the aviation investment case.
Contact the Author
Tel: +44 1252 451 651