Atmospherically induced high-frequency sea-level extremes: insights from the Adriatic Sea meteotsunami monitoring and warning network

High-frequency sea-level oscillations (with periods shorter than three hours) occasionally reach significant amplitudes along the eastern Adriatic coast, posing risks to coastal infrastructure and communities. These oscillations are often driven by rapid atmospheric pressure disturbances that travel towards the shore, where local topography can amplify the resulting sea-level response - especially in bays and harbors along the Croatian coast. The most extreme events of this type are meteotsunamis. Understanding the relationship between atmospheric forcing and sea-level response is crucial for developing meteotsunami forecasting and early warning systems in vulnerable Adriatic locations.

In this study, we explore the connection between high-frequency components of 1-minute atmospheric pressure and sea-level measurements from a network of microbarograph and tide-gauge stations located in the central Adriatic along both its eastern and western coast. The measurements that cover the period from May 2017 to October 2024 underwent rigorous quality control, addressing issues such as outliers, duplicates, and data gaps. To explore atmospheric pressure and sea-level relationships, we applied multiple methods for identifying extreme events, including variance-based and wave-height-based approaches. We then examined correlations across station pairs using various temporal offsets (0 to 480 minutes) between atmospheric and ocean series, all in order to determine what kind of atmospheric pressure oscillations tend to precede extreme sea-level oscillations.

Preliminary results suggest varying degrees of correlation between atmospheric pressure and sea-level oscillations, with stronger connections observed for certain station pairs, possibly influenced by geographic proximity. In some cases, the strongest sea-level events were preceded by distinct pressure jumps, suggesting a delayed sea-level response, while in others, their co-occurrence appeared more synchronized. Both the variance-based and wave-height-based approaches provided broadly similar patterns, though the variance-based method showed limitations in capturing very short-lived atmospheric pressure disturbances. Ongoing analyses aim to refine these findings by testing alternative methods, involving event-based extraction of extremes, and estimating the propagation velocity of atmospheric disturbances - steps intended to enhance our understanding and improve the predictive capabilities of future high-frequency sea-level oscillation monitoring and warning systems.