ABSTRACT

Marangoz, H.O., 0000. Time history–based forecasting model on bathymetric depth predictions.

Seabed observations are critical for representing wave dynamics and understanding nearshore morphological changes. In this study, high-resolution bathymetric maps were generated based on field measurements conducted over a 2.5-year period along a coastal region in Rize, Turkey. A time history–based forecasting approach was applied to predict future seabed depth changes over multiple time horizons (2, 5, 10, 20, and 50 years) using statistical models, including linear regression, cubic polynomial regression, and the autoregressive integrated moving average method. Unlike many studies that integrate hydrodynamic parameters such as wave action, sediment transport, or storm surges, this research focused on direct point–based prediction models using only limited-time observed depth values without incorporating external variables. Even though this led to increased uncertainty, some applied models failed to deliver reliable results in long-term forecasts. However, the findings show that linear regression performed more consistently than the other time-dependent models within the observed data set. The outcomes highlight the promising potential of minimalist statistical approaches for bathymetric forecasting and offer support for preliminary decision-making in coastal planning, erosion risk assessment, and sediment evolution monitoring, particularly in data-limited environments.