Modelling weather risk preferences with multi-criteria decision analysis for an aerospace vehicle launch

Decision-making under weather uncertainty is a challenge in several fields. When the decision process involves many stakeholders, frequently with different interpretations of the meteorological information, the process is even more complex. This work provides a quantitative decision model with a new index (called the weather decision index, WDI) to support the stakeholders in making real-world choices according to their preferences regarding the uncertainty of weather information. The integrated model combines several methods such as problem structuring, multi-criteria analysis, scenario planning and probabilistic weather forecast techniques. As a demonstration, the model was applied in the sounding rocket launch mission in the Brazilian Space Programme. The WDI captured stakeholders' behaviour related to three meteorological information attributes (probability, lead-time and variables) and modelled the most important judgements of the decision maker; low probability or an extended lead-time depreciates the meteorological information, and weather variables are not considered in the decisions, even with forecasts of extreme events. Modelling with the WDI brings a new perspective in weather-related decision problems. The choice of alternatives no longer depends on a necessarily simplified optimization analysis, but rather on the decision maker's preferences about the possibly nonlinear trade-offs between forecast reliability and lead-time. The findings also increase understanding of the forecast decision maker's preferences and how to improve weather risk communication. The WDI provides a starting point for several applications, including early warning systems or climate change adaptation, for which reliable uncertainty estimates are accessible.