Variability and extremes of the Caspian Sea's modelled wave climate

Abstract

The wave climate of the Caspian Sea has been notably underrepresented in the existing literature, leaving an important gap in the understanding of wave extremes. We address this deficiency by generating a high-resolution, idealized ice-free wave hindcast dataset for the Caspian Sea, covering 42 years (1982–2023) using the third-generation SWAN model driven by ERA5 winds. The simulated wave properties are calibrated and validated against previous lower-resolution simulations and satellite altimetry data from Copernicus Marine Service's multi-mission products. The model is run using 3-monthly average water level. The comparison by means of dynamic time warping and wavelet coherence analysis confirmed overall consistency of our simulations with earlier simulations but revealed different timing of some modelled events. The wave climate of the Caspian Sea is generally mild, with extensive spatial and seasonal variation. The mean significant wave height ($H_S$) is around 1 m in most of the sea, but certain storms can produce $H_S$ up to 6–7 m. Extreme wave heights tend to occur in areas where average wave heights are moderate, particularly near the eastern shores of the central and northern part of the basin, and near the western shores in the southern part. The probability distribution of wave heights is best represented by a Wald (inverse Gaussian) model, whereas its upper tail is more suitably fitted by a Weibull distribution. The enhanced understanding of the Caspian Sea's wave dynamics, including analysis of the properties and location of severe waves in a selection of extreme storms, provides a robust foundation for informing coastal management strategies and mitigating the impact of extreme wave events.