Physics and Chemistry of the Earth, Parts A/B/C
Volume 34, Issues 17–18,
2009
, Pages 894-903
Author links open overlay panel, , ,
Abstract
This paper describes an extraordinary tsunami-like event that occurred on 21 June 1978 that encompassed the middle and south Adriatic Sea. The flood had its culmination in Vela Luka, where a maximum wave height of 6m was reported. This paper contains a detailed description of the event as seen by eyewitnesses, its outreach along both the eastern and western coasts, and the aftermath and recovery activities in Vela Luka. All available records have been collected and analysed to detect the source and the generating mechanism of the long ocean waves. Seismic generation is fully excluded from the consideration, while a submarine landslide seems rather unrealistic as it does not explain the characteristics of the measured ocean waves. Therefore, the source of the event was presumably in the atmosphere, where a travelling disturbance was detected that had the capability to resonantly transfer energy to the ocean via the Proudman resonance mechanism. Although these data prove the proposed mechanism, the final confirmation for such a scenario should come from a process-oriented numerical modelling study.
Introduction
The inhabitants of the city of Vela Luka (Fig. 1), situated in a hidden bay on Korčula Island, experienced large seiches (local name – šćiga) a number of times, reaching and even overtopping the piers and the city promenade. However, residents were not prepared for a chain of events that attacked this picturesque city in the early hours of 21 June 1978, especially as no weather storms and low wind had been forecast for the region. The sea suddenly began to rise, overtopping the piers and breaking into the city. A number of sea strokes occurred until the end of the morning hours, when the sea retreated to its bed, leaving widespread chaos and damage all over the city.
The resemblance of the event with a tsunami was too obvious, but the problem was in its generation and source. Namely, there was no recorded earthquake at that time in the Adriatic Sea. Two other sources of a tsunami came to the researchers’ minds in the following days: a submarine landslide and an atmospheric process. Also, a proper assessment of the event presumes that all of the collected data, reports, and material should be treated equally and examined in a collaborative work, but a problem arises in the outreach of the event. Namely, anomalous sea level oscillations were recorded on 21 June 1978 on both sides of the Adriatic Sea, along the Yugoslav (at that time) eastern Adriatic coast and along the western Italian shore. However, the researchers did not jointly evaluate all of the information collected on both sides of the Adriatic Sea and the reports were trying to assess the origin of this event based on a limited portion of the available material. This study intends to bridge this gap in the investigations and to present all of the materials, eyewitness reports, data, and theories in one place.
The Italian Tsunami Catalogue (Tinti et al., 2004) classifies this event as a tsunami with unknown origin; however, several hypotheses have been suggested by researchers: (i) Bedosti (1980) suggested that a submarine landslide occurred along the 200m isobath between Termoli and Vasto; (ii) Zore-Armanda (1979) hypothesised that the tsunami waves were generated by an earthquake that occurred in the Aegean Sea; (iii) Hodžić, 1979/1980, Hodžić, 1986 assumed that cyclonically generated open ocean waves freely propagated towards Vela Luka Bay and excited local seiches; (iv) Orlić (1980) offered the Proudman resonance theory as an explanation, where the approaching ocean waves are constantly forced and amplified by atmospheric gravity waves. These theories will be carefully assessed in Section 6. In fact, it seems that this event was the strongest tsunami-like event in the Adriatic Sea in the 20th century, classified as a four on the Sieberg–Ambraseys tsunami intensity scale. Moreover, the 1978 event was the second largest tsunami in intensity in the history of catalogued Adriatic Sea events (Tinti et al., 2004), just after the 1627 Gargano tsunami whose intensity has been estimated at a five on the Sieberg–Ambraseys scale and that, together with the earthquake, claimed more than 5000 victims (Tinti and Piatanesi, 1996). It should also be mentioned that all of the other catalogued tsunami events were of seismic origin (Maramai et al., 2007). Therefore, a proper estimate of the source of the 1978 event is a high priority in the list of actions to be accomplished in assessing Adriatic tsunamis.
Apart from tsunamis listed in the Italian Tsunami Catalogue, some events were sporadically reported along the eastern Adriatic Sea shoreline, such as the 25cm high tsunami recorded (at Bar) by the 1979 Montenegrin earthquake (Orlić, 1983/1984), or the 10cm tsunami generated by the 2003 Makarska earthquake (Herak et al., 2001). It should be noted that the annals of the Dubrovnik Republic indicate a strong tsunami during the 1667 destructive earthquake as, “ships severely hit the ground three times,” in the Dubrovnik harbour. Paulatto et al. (2007) assessed various scenarios for seismically generated tsunamis in the Adriatic and numerically modelled a maximum tsunami height of 5m in Dubrovnik for the worst case scenario based on a magnitude 7.5 earthquake with an epicentre just outside of Dubrovnik. Aside from the 1978 event which, according to the preliminary explanations given by Orlić (1980) and Hodžić (1986), may be classified as a so-called meteorological tsunami. Further tsunami-like events were recently attributed to an atmospheric source, such as the 2003 middle Adriatic event (Vilibić et al., 2004), the 2007 Ist event (Šepić et al., 2009), and the 2008 Mali Lošinj event. The 2003 event caused substantial damage to coastal infrastructure, destroying a large portion of the shellfish farms in Mali Ston Bay due to severe currents, and resulting in several million Euros in damage.
This paper will attempt to attribute the event of 21 June 1978 that occurred in the middle and south Adriatic Sea to its source by examining all of the available eyewitness reports, sea level and other data, and by assessing the observed properties versus common characteristics of different types of tsunamis in the literature. In addition, an overview of aftermath activities will be given to show the capacities of the local and national authorities and civil protection agencies present at the time to mitigate the consequences and the impact to the population and coastal infrastructure. All of the hypotheses collated in the literature and the media will be carefully assessed and a full explanation of the event will be presented and discussed, including the pathways for further research activities.
Section snippets
Vela Luka disaster of 21 June 1978
In the early morning hours of 21 June 1978 the sea began to rise in the city of Vela Luka (Vučetić and Barčot, 2008). The first wave and overtopping of the quays and piers occurred at 04:15 UTC, flooding the basements of sea front houses, breaking house walls, and snapping mooring lines of boats and pushing them ashore. Several minutes later, the sea retreated, emptying the top of the bay and leaving aground the moored boats and ships (Fig. 2a). The culmination of the flooding was reached
Outreach of the event
The city of Vela Luka was not the only location on the Adriatic Sea that was affected by the severe waves on 21 June 1978. A number of eyewitness reports from other places along the eastern and western Adriatic Sea coastline can be found in the newspaper archives and technical reports. A succinct summary of these reports, containing the maximum heights of the all witnessed and measured waves, is illustrated in Fig. 3.
Severe (2–4m high waves) were reported by eyewitnesses in the western sector
Instrumental records
As mentioned previously, no earthquake occurred in the Adriatic Sea on 21 June 1978, thus no seismographs are available for inspection. When regarding submarine landslides, a recent geomorphological study of the seabed at the western edge of the South Adriatic Pit by Minisini et al. (2006) documents the occurrence of only marginal submarine landslides in the present geological timeframe with no capacity to induce the recorded waves. In this section all available sea level charts will be
Aftermath activities, damage assessment
Immediately after the destructive waves receded, local authorities proclaimed a state of natural disaster emergency for the city of Vela Luka (Vučetić and Barčot, 2008). Simultaneously, the civil protection service instructed the local population about the use of drinkable water, electrical installations, medicines, and other emergency issues to prevent the occurrence of contagious diseases and to avoid any casualties caused by defectiveness in electrical installations upon the return of the
Assessment of existing theories
Four different hypotheses about the source and the generation of the tsunami-like waves were introduced after the 1978 event. Zore-Armanda (1979) tried to connect the event with the earthquake in the Aegean Sea. Bedosti (1980) and a team of Croatian scientists introduced a possibility that a submarine landslide was responsible for the generation of the observed waves. Finally, Hodžić, 1979/1980, Hodžić, 1986, Orlić, 1980 related the event to an atmospheric mechanism, namely freely propagating
Conclusions
This paper attempts to present all available material from both Adriatic Sea coastlines, which were used in an assessment of the great Adriatic flood that occurred on 21 June 1978. Careful inspection of the possible generation mechanisms for the observed tsunami-like waves favours travelling atmospheric waves, which created long ocean waves through a long-distance resonant energy transfer from the atmosphere. Thus, these meteotsunami waves hit coastal areas and had the largest amplitudes in the
Acknowledgements
We thank the colleagues and organisers of the International Symposium on Meteotsunamis, Vela Luka, Croatia, 19–21 June 2008 (see details at http://www.izor.hr/~vilibic/vela_luka, with special thanks to Tonko Barčot and Vanjo Žuvela). We would also like to acknowledge the enthusiastic individuals who took photos during the Vela Luka flood of 21 June 1978, and appreciable credit should go to all of them. The Hydrographic Institute and Meteorological and Hydrological Service of the Republic of
References (20)
- M. Herak et al.
Did the Makarska earthquake of 1962 generate a tsunami in the central Adriatic archipelago?
Journal of Geodynamics
(2001)
- S. Tinti et al.
Numerical simulations of the tsunami induced by the 1627 earthquake affecting Gargano, southern Italy
Journal of Geodynamics
(1996)
- I. Vilibić
Numerical simulations of the Proudman resonance
Continental Shelf Research
(2008)
- I. Vilibić et al.
Destructive meteotsunamis along the eastern Adriatic coast (overview)
Physics and Chemistry of the Earth
(2009)
- Bedosti, B., 1980. Considerazioni sul maremoto adriatico (tsunami) del 21.6.1978. Supplemento Bollettini Sismici...
- M. Hodžić
Occurrences of exceptional sea-level oscillations in the Vela Luka Bay (in Croatian)
Priroda
(1979/1980)
- Hodžić, 1986. Long Gravity Waves Generated by Cyclones and Free Bay Oscillations (seiches) in the Adriatic. PhD Thesis,...
- ICG-NEAMTWS, 2008. Tsunami Early Warning and Mitigation System in the North Eastern Atlantic, the Mediterranean and...
- R.S. Lindzen et al.
Banded convective activity and ducted gravity waves
Monthly Weather Review
(1976)
- A. Maramai et al.
Investigation on tsunami effects in the central Adriatic Sea during the last century – a contribution
(Video) 16 Cold Cases That Were Finally Solved Many Years Later - Cold Cases Solved CompilationNatural Hazards and Earth System Sciences
(2007)
There are more references available in the full text version of this article.
Cited by (38)
High-frequency sea level oscillations in the Mediterranean and their connection to synoptic patterns
2015, Progress in Oceanography
Citation Excerpt :
A strong resemblance to meteotsunami coastal impacts may be perceived – as the coastal infrastructure in the Mediterranean is not adapted to several metres high meteotsunami waves. If such waves impacted certain bays and harbours, the flooding in coastal cities and the damage to marine infrastructure (shellfish farms, nautical marina, yachts and boats) could be quite substantial, on the order of tens of MEuros (Vilibić et al., 2004; Monserrat et al., 2006; Vučetić et al., 2009). Coastal flooding is not the only potential consequence of such events; associated currents in harbour constrictions can be quite strong and affect the safety of navigation and ship mooring lines in harbours (Lopez and Iglesias, 2014).
This paper contains a basin-wide assessment of high-frequency sea level oscillations in the Mediterranean Sea for the period from 2010 to 2014. Sea level series with temporal resolutions of 1min were taken from 29 tide gauges that had been operational for at least 4years with high-quality time series. The contribution of high-frequency sea level variance (2min–6h) to the total and residual sea level variance was estimated to vary between 0.4% and 9.5% and between 0.6% and 12.1%, respectively, but to become dominant at some stations during extreme high-frequency events. A total of 36 high-frequency sea level events were extracted from the series, with some occurring locally in one of the four selected regions of Spain, Sardinia, Sicily and the Greece–Ionian Sea (6 events), some occurring over two or three regions (19 events), and some occurring in all of the selected regions within a period of 1–3days (11 events). The basin-wide events normally propagate from the western to the eastern parts of the basin, with an average velocity of approximately 30km/h. Synoptic patterns associated with high-frequency events were analysed, and a high resemblance to patterns associated with meteotsunamis was observed: (i) a cyclone with a centre W-NW from the affected area, (ii) a strong thermal front in the lower troposphere, and (iii) a forefront of an unstable strong mid-tropospheric jet placed over the affected area. The strong coherence between high-frequency sea level events and synoptic patterns introduces the possibility of a timely forecast of these events.
Mediterranean sea-level variability and trends
2012, The Climate of the Mediterranean Region
Natural disasters in Croatia
2022, Lijecnicki Vjesnik
Types of natural disasters and their occurrence
2022, Lijecnicki Vjesnik
(Video) Venice | Wikipedia audio articleMeteotsunamis in Orography-Free, Flat Bathymetry and Warming Climate Conditions
2022, Journal of Geophysical Research: Oceans
Recommended articles (6)
Research article
Development of an unstructured-grid wave-current coupled model and its application
Ocean Modelling, Volume 104, 2016, pp. 213-225
An unstructured grid wave-current coupled model was developed by coupling the SWAN (Simulating Waves Nearshore) wave model and ADCIRC (Advanced Circulation model) ocean model through the Model Coupling Toolkit (MCT). The developed coupled model has high spatial resolution in the coastal area and is efficient for computation. The efficiency of the newly developed SWAN+ADCIRC model was compared with that of the widely-used SWAN+ADCIRC coupled model, in which SWAN and ADCIRC are coupled directly rather than through the MCT. Results show that the directly-coupled model is more efficient when the total number of computational cores is small, but the MCT-coupled model begin to run faster than the directly-coupled model when more computational cores are used. The MCT-coupled model maintains the scalability longer and can increase the simulation efficiency more than 35% by comparing the minimum wall clock time of one day simulation in the test runs.
The MCT-coupled SWAN+ADCIRC model was used to simulate the storm surge and waves during the typhoon Usagi which formed in the western Pacific on September 17, 2013 and landed at Shanwei, China. Three numerical experiments were performed to investigate the effect of wave-current interaction on the storm surge and waves. The results show that the coupled model can better simulate the storm surge and waves when considering the wave-induced radiation stress, the wave effect on the wind stress drag coefficient and the modulation of current and water level on waves. During the typhoon Usagi, the effect of wave radiation stress could result in a maximum of 0.75m increase in the extreme storm surge, and the wave induced wind stress could cause a −0.82∼0.48m change of the extreme storm surge near the coastal area. Besides, the radiation stress forced currents cannot be ignored either in the study of mass transport at coastal zones. Results of this study are useful for understanding the wave-current interaction processes and the development of the operational prediction technique for storm surge and waves.
Research article
Examining tidal impacts on seasonal circulation and hydrography variability over the eastern Canadian shelf using a coupled circulation-ice regional model
Progress in Oceanography, Volume 189, 2020, Article 102448
A coupled circulation and sea ice model is used to examine the tidal impacts on the seasonal variability of circulation, hydrography and sea ice over the eastern Canadian shelf (ECS). The model performance is assessed using in-situ and satellite remote sensing observations. The tidal impacts have significant spatial variability, which are relatively small over the Labrador and Newfoundland Shelves, moderate over the Scotian Shelf (ScS), and significant over the St. Lawrence River Estuary (SLRE), northwestern Gulf of St. Lawrence (GSL), southwestern ScS, Gulf of Maine (GoM), Bay of Fundy (BoF), and the northern flank of Georges Bank (GeB). The tidal impacts on the seasonal mean circulation are greater in winter than in summer in the SLRE, while greater in summer than in winter over several other areas in the GSL, the GoM-BoF and the southwest ScS. The tidal impacts on temperature and salinity are the most significant near fronts, where both tidal mixing and frontal circulation play important roles. The tidal residual circulation, especially that due to tidal rectification in the GSL and GoM-BoF, spreads the large tidal impacts generated near fronts into broader areas. The changes in circulation and stratification also account for the reduced sea ice concentrations in the GSL.
Research article
Spatio-temporal variability of tidal asymmetry due to multiple coastal constructions along the west coast of Korea
Estuarine, Coastal and Shelf Science, Volume 151, 2014, pp. 336-346
(Video) Second Workshop on Coyotes in Urban AreasAt least 19 remarkable dikes and land reclamations have been constructed since 1970 along the west coast of Korea, which resulted in a reduction in tidal flat area of almost 50%. Both the reduction in tidal flats and the artificially simplified coastal line have distorted the spatio-temporal tidal hydrodynamics; to quantify this, we analyzed and evaluated tidal asymmetry by phase differences of the principal semi-diurnal lunar constituent M2 and its first over-tide M4. Moreover, we applied the ADCIRC model to quantitatively investigate near- and far-field impacts on tidal variations using the gamma parameter, tidal energy flux, and dissipation rate. Through this study, we found that the tidal regime around the Incheon harbor area in Gyeonggi Bay has changed from ebb- to flood-dominant due to multiple nearby reclamations, in particular to the construction of the Siwha dike. The Saemangeum dike caused near-field de-amplification of M2 but far-field amplification in the Shandong area of China. In addition to allowing a traditional asymmetry approach using phase difference, analyses based on the gamma parameter and tidal energy variations could distinctly improve spatial understanding of anthropogenic impacts on coastal tidal hydrodynamics.
Research article
Evaluation of WAVEWATCH III performance with wind input and dissipation source terms using wave buoy measurements for October 2006 along the east Korean coast in the East Sea
Ocean Engineering, Volume 100, 2015, pp. 67-82
In the winter, in the East Sea (ES), storm waves due to moving developed lows are frequently reported and cause extensive coastal disasters. During October 2006, there were extensive damages along the east Korean coast due to high storm waves induced by winter storms passing over the ES. This paper investigates the performance of a wave model, WAVEWATCH III, for the rough sea conditions in October 2006 with respect to the wind input and dissipation terms because the wave-breaking dissipation is the least known source term, which acts as a tuning knob for the closure of the action balance equation. Three package-like wind input-dissipation parameterizations, the WAM3 type (WAM-equiv), Tolman and Chalikov terms (TC96), and WAM4 type and its variant (WAM4+), are used experimentally for their performances under the same wind forcing obtained from atmospheric modelling with WRF. Overall, all experiments illustrate good accordance with observed wave characteristics. Among them, the WAM4+ results exhibit the best performance based on the Taylor diagram and index of agreement. In terms of dissipation behaviour, the TC96 results depict high energy losses at high frequency over 0.25Hz, whereas the WAM-equiv runs display less dissipation at the same frequency. The WAM4+ results lie between those of the WAM-equiv and TC96.
Research article
The role of seaward morphology on wave transformation onto and across a microtidal shore platform
Continental Shelf Research, Volume 224, 2021, Article 104472
Waves are a primary erosive agent on intertidal shore platforms. In microtidal environments, as waves cross a shore platform the energy transforms from gravity to infragravity frequencies. The morphology of the intertidal zone is a key boundary condition for this energy transformation. Waves approaching the shore however often interact with the seabed at some distance from the intertidal platform. In this study we explore this wave interaction over a gently sloping subtidal ramp (1km wide) down to 11m water depth fronting a semi-horizontal intertidal platform (180m wide). Five non-directional pressure sensors were installed on the sea floor from 5 to 11m depth for one week with a further four placed on the intertidal platform for two tidal cycles during the offshore deployment. It is found that a shift in wave frequency starts to occur at the base of the seaward ramp at 7m water depth, with the primary development of infragravity waves occurring at less than 5m depth. The offshore energy signal was also tidally modulated with infragravity frequency waves hypothesised to be able to escape seaward across the platform edge at high tide. This new exploration of nearshore wave energy shows that wave dynamics over the entire platform surface should be considered rather than just focussing on the intertidal zone.
Research article
Thermohaline conditions and circulation in the Gulf of Thailand during the northeast monsoon
Continental Shelf Research, Volume 225, 2021, Article 104487
Hydrographic data from the Gulf of Thailand (GoT) reveal two bottom saline water areas (BSWA1 and BSWA2) with salinity (S)≥33.0 psu and where the water columns are highly stratified during December–January. Observational results from a seafloor acoustic doppler current profiler verify that the current is mainly dominated by a barotropic current. The monthly average sea-level-anomaly (SLA) and barotropic current vector maps confirm a cyclonic circulation in the northern GoT during November–January. The BSWA1 is located near the center of the cyclonic circulation. The BSWA1 and cyclonic circulation simultaneously enter their weakening phases, which suggests that the cyclonic circulation convergence plays a key role in maintaining the BSWA1. The BSWA2 occurs as a narrow strip concentrated along the eastern slope of the basin in the southern GoT during December–January. The southern GoT also has higher SLA and forms a congregation area of diluted water with S<31.0 psu that enters from the north and south of this region during December–January.
Copyright © 2009 Elsevier Ltd. All rights reserved.