• Empirical correlation
• Iceberg-calving
• Iceberg-tsunamis
• Impulse wave
• Prediction method
• spatial wave propagation

Large impulse waves are generated by the calving of icebergs in oceans, bays or lakes. These waves are called iceberg-tsunamis. Iceberg-tsunamis are particularly relevant in Greenland and Antarctic during the summer season. An iceberg-tsunami reaching an amplitude of approximately 50 m was observed at the Eqip Sermia glacier in 2014 and several smaller events recently occurred in Greenland, Patagonia and New Zealand, where in some cases harbours and boats were destroyed. The forecasting of the relevant wave features, e.g. the maximum wave height, plays a key role for hazard assessment and the mitigation of iceberg-tsunamis.
The present study focuses on iceberg-tsunami generation and propagation by analysing the experimental data of Heller (2019) and Heller et al. (2019). This included the analysis of large-scale experiments conducted in a 50 m x 50 m wave basin involving five idealised calving mechanisms. The icebergs were modelled with blocks and resistance type wave probes were used to record the water surface elevations.
The experimental data were post-processed to improve their quality and exclude the physically meaningless information, such as noise and the reflected waves from the basin walls. The relevant wave features were hence extracted from the post-processed measurements, including the maximum wave height, amplitude and the corresponding period. The largest waves were observed in the experiments where the iceberg calving was dominated by the gravity force whereas the calving mechanisms dominated by the buoyancy force generated approximately an order of magnitude smaller waves.
The maximum wave height, amplitude, the corresponding wave period and their decay were expressed as a function of six governing dimensionless parameters, including the relative released energy Er, Froude number F, relative iceberg thickness S, relative iceberg width B, relative iceberg volume V and the relative density D. These equations were derived and validated relatively well for each of the five calving mechanisms individually and for combinations.
The wave propagation of iceberg-tsunamis presented in this study was compared with the wave propagation of landslide-tsunamis from technical literature. Landslide-tsunamis are impulse waves generated by the impact of a landslide in a water body. These two phenomena show similarities with regards to the wave decay with the radial distance only. By comparing the wave decay with the radial distance and the wave propagation angle combined, the equations for landslide-tsunamis do not well describe the iceberg-tsunami decay.
The empirical approach of the present study was compared with predictions from the theoretical model of Massel and Przyborska (2013). This comparison was conducted through three numerical examples and by simulating some tests with this model. This resulted in large differences except for the iceberg-tsunami generated by an iceberg overturning and impacting horizontally on the water body.
The iceberg-tsunami observed in 2014 at the Eqip Sermia glacier was predicted well by the empirical equations of the present study, underestimating the observed maximum wave amplitude by only 22%. While the prediction of the tsunami observed at the Tasman Glacier in 2011 resulted in large differences. The 2D geometry of the lake, shoaling effects and the uncertainty of the input data justify this observed difference.