The next generation meteogram: enhancing decision making for aviation stakeholders

Aviation operations are impossible without reliable weather information. Safe air traffic management requires precise meteorological predictions across various timeframes, ranging from immediate nowcasting (0-3h) to mid-range forecasting (4-30h). Given the inherent uncertainties in predictions, it is crucial to ensure that users can interpret uncertain weather data confidently. This enables effective planning and decision-making within aircraft operations, including setting approach rates and runway configurations.Currently, the aviation meteogram product provided by MeteoSwiss offers deterministic predictions and threshold exceedance probabilities at hourly intervals up to 24 hours in advance. This data is derived both from direct numerical weather model output and statistically post-processed forecasts. However, this approach underutilizes the forecast potential due to (1) possible biases in deterministic predictions, (2) limited granularity in lead time within the nowcasting range, and (3) a lack of uncertainty information passed on to decision makers.To overcome these challenges, we are transitioning from delivering deterministic to probabilistic forecasts by employing machine-learned local predictions instead of relying solely on physical model output. Additionally, we are increasing output granularity in the first three hours. This shift unlocks the full potential of information-based decision-making, leading to smoother and more economically and ecologically sustainable aviation operations. Given that probabilistic data and its implications are largely unfamiliar to our product users, a robust program of frequent training and education is essential.In this contribution, we introduce new machine learning based ensemble predictions for thunderstorms 1, visibility conditions 2, and wind at Swiss airports, focusing on nowcasting and mid-range forecasting. Our emphasis lies in their comprehensive visualization in a web-based meteogram tailored to effectively convey probabilistic information.1: abstract Attinger et al., 2024, submitted to EMS 2024.2: abstract Wehrli et al., 2024, submitted to EMS 2024.