Towards limiting uncertainties in the estimates of mass, salt and heat transport between the Baltic Sea and the North Sea

A new conceptual framework for the assessment of the physical state of the Baltic Sea was introduced recently. The approach is based on the analysis of mutual variability of well-established climate indicators such as basin-wide ocean heat content (OHC) and freshwater content (FWC). Previous studies have reported a positive trend in OHC and a negative trend in FWC. The increase in OHC has been attributed to the rising air temperature over the Baltic Sea, yet the decline in FWC remains largely unexplained. It was noted that neither salt transport to the Baltic Sea, net precipitation, nor total river runoff accounted for the FWC's downward trend. We suggest that more accurate estimates of mass, salt and heat transport between the Baltic Sea and the North Sea are needed, than currently available. This study is designed as the first step towards the goal by analysing model reanalysis data and observations. 

The Baltic Sea is a brackish marginal sea connected with the North Sea through the narrow Danish Straits (Øresund, Great Belt and Little Belt) and the Skagerrak and Kattegat transitional zones. A complex geometry and bathymetry of the area complicates the estimation of the transport values using traditional methodology like numerical modelling and solely observation based interpolation methods. We analyse volume, heat and salt transports across sets of transects in Skagerrak/Kattegart transition zone and the southern Baltic Sea using the Baltic Sea Physical Reanalysis data (NEMO, 1 nautical mile resolution, 56 vertical layers, 1993–present). The derived transports are compared to the long-term observational Baltic Saline Inflow (SBI) series.

Our results quantify the consistency and difference in transport between neighbouring model’s transects over  the transition zone between the North Sea and Baltic Sea. They show comparable temporal variability between area-mean model transports and  the SBI index and spectral analysis indicates that the reanalysis captures the dominant temporal scales of inflow variability, while differences in amplitude suggest sensitivity to area choice. 

We develop purpose specific data-driven models to link these two data sources into mass, heat and salt exchange estimators for the North Sea and Baltic Sea connective area, using attention-based Transformer architectures to learn the time-dependent relationships between reanalysis predictors and available observations, and to estimate volume, heat, and salt exchange through the North Sea–Baltic Sea gateway.