A pragmatic approach for the numerical prediction of meteotsunamis in Ciutadella harbour (Balearic Islands)

Abstract

Abstract Ciutadella (Menorca, Balearic Islands, Spain) is well known for the large amplitude seiches of about 10.5 min period registered in its long and narrow inlet, especially during the warm season. This phenomenon, locally referred to as “rissaga”, might reach extreme wave heights (in the range 1.5 – 4 m) with a recurrence of only a few years, leading to damaging consequences in the port activities and moored vessels. The provision of as accurate as possible predictions of the rissaga risk, hours or days ahead, appears as a crucial element in helping to mitigate these consequences. We devise a chain of atmospheric and oceanic numerical simulation components aimed at capturing with low computational cost the key physical processes conducive to the vast majority of rissaga events: (i) the genesis upstream from the Balearic Islands of high amplitude atmospheric gravity waves travelling in the SW–NE direction; these mesoscale waves are synthetically triggered using a 2D nonhydrostatic fully compressible model within a vertical environment provided by a representative sounding; (ii) the oceanic response to the concomitant pressure fluctuations along the Menorca channel, in the form of long oceanic waves subject to Proudman resonance; these processes are simulated with a shallow-water model applied over a 80-m depth channel; (iii) shelf amplification, which according to theory (Green’s law) accounts for a doubling of the wave amplitude for a depth jump from 80 to 5 m; and (iv) harbour resonance within Ciutadella inlet, a crucial mechanism solved again with the shallow-water equations over the idealized 5-m deep channel. The prognostic system is successfully tested for the available set of 126 rissaga events and for a complementary set of 549 ordinary situations. Our approach discriminates fairly well non-rissaga events from high-amplitude oscillations and tends to correctly categorize the meteotsunamis among weak, moderate or strong events. We are now pursuing the real time application of the method in a probabilistic context.