Monitoring Ocean Wave height in the Northeast Atlantic Using Terrestrially based microseism data

Current methods employed to track the spatiotemporal evolution of ocean wave mainly include insitu buoys, numerical ocean wave modeling, and satellite altimetry. Each method has its own strengths and weaknesses in terms of spatial and temporal resolution. For example, buoys provide high temporal resolution, but lower spatial resolution compared to numerical wave forecast modeling and satellite altimetry.

This study explores an alternative method to investigate the feasibility of constructing an ocean wave monitoring system utilizing land-based seismic amplitudes. The proposed method relies on the correlation between secondary microseism amplitudes detected on land and their causative ocean wave heights. .

In this method, we implemented a supervised Artificial Neural Network (ANN) to quantify the nonlinear relationship between secondary microseism amplitudes recorded on land and the associated ocean wave heights.. The ANN was trained using seismic amplitudes data from seismic stations distributed across Ireland and Buoy data or numerical simulated ocean wave height data in the Northeast Atlantic. Subsequently, the trained ANN was utilized to estimate significant Wave Height (SWH) at specific location(s). The estimated wave heights exhibit a similar statistical distribution to in-situ wave height observations, with normally distributed differences. Since the approach is purely data-driven, its implementation is straightforward and holds potential as a reliable, low-cost operational tool.

The comparison between our results and the measured wave height data demonstrates a strong correlation, particularly for smaller wave heights, where the estimates show excellent accuracy. For larger wave heights, while the estimates are not as accurate, they still provide reasonably reliable approximations, highlighting the robustness of this  approach, across a range of ocean wave conditions.