Major Baltic Inflows (MBIs) are critical ventilation events that transport oxygenated, saline water from the North Sea into the deeper basins of the Baltic Sea (Mohrholz, 2018). Their formation depends on multiple factors, with wind direction and saltwater transport playing a particularly significant role. To evaluate the influence of MBIs on the Baltic Sea ecosystem, two numerical model scenarios were implemented.

• Closed boundary scenario: From May 2014 to January 2015, saline inflows from the North Sea were restricted by closing the model’s western boundary conditions. This modification prevented high-salinity water intrusion, allowing an isolated analysis of the effects of reduced saltwater exchange on Baltic Sea hydrography and ecosystem dynamics.

Model outputs were evaluated for changes in chl concentrations and TTI across Baltic Sea HELCOM basins to assess impacts on eutrophication status. The mean relative effect (Equation 1) was calculated as the percentage difference between the scenario and the reference simulation in 2014–2021.

To investigate the role of large-scale circulation patterns in eutrophication, a modified hydrodynamic scenario was implemented in the Baltic Sea model. In this scenario, a barrier structure was introduced to disrupt natural current into the Gotland basin. The artificial structure was positioned between longitude 19.85°–20.88° and latitude 55.92°, approximately 8 km offshore, extended 70 km in length and 4 km in width (Figure 1), thereby preventing the direct inflow of salt intrusions into the Gotland basin. This experimental setup enables for the assessment of circulation alterations on nutrient transport, retention, and eutrophication dynamics within affected regions.

To assess the effect of warming on TTI and Chl in a scenario, the air temperature in the atmospheric forcing data was increased by 1.7°C for the year 2014. All other initial and boundary conditions remained consistent with the reference simulation.

This scenario assesses the impact of a 50% reduction in nutrient loads from all rivers shown in Figure 2 that discharge into the Baltic Sea. Riverine nutrient loads were estimated using the GREEN model (Grizzetti et al., 2012; 2019) combined with the LISFLOOD hydrological model (De Roo et al., 2020), focusing on the 2010–2021 period. Total Nitrogen (TN) and Total Phosphorus (TP) fluxes, derived from an updated hydrological model (Vigiak et al., 2023), were integrated into a marine biogeochemical model to simulate eutrophication responses. Nutrient loads for the reference scenario, aggregated by major basins (Gulf of Finland, Baltic Proper, Gulf of Riga, Kattegat), are provided in Supplementary Material S6.

Hindcast simulations were conducted under a 50% reduction scenario, maintaining constant riverine discharge. The relative impact was quantified in 2014–2021: I m p a c t = S c e n a r i o − R e f e r e n c e S c e n a r i o R e f e r e n c e S c e n a r i o 100 (1)

This approach enables a comparative evaluation of nutrient reduction effectiveness.

HELCOM’s core indicators assess the average chlorophyll-a (chl) concentration in surface waters (0–10 m) during summer (June–September). The evaluation of good environmental status (GES) is based on scientifically established, sub-basin-specific threshold values that define acceptable concentration limits. These threshold values, detailed in the HOLAS II assessment, serve as the benchmark for chl evaluation.

In our study, model results were analyzed for each basin using these threshold values. If a model grid point exhibited a value below the threshold, that area was classified as meeting GES for the corresponding indicator. The percentages in the figures represent the proportion of the total possible area that achieves good status based on the indicator’s value. To ensure comparability between TTI and chl scales, a 3-year rolling average was applied to the spatial chl data.

The Trophic Transfer Index (TTI) is based on the assumption that eutrophication impacts a marine area when an increase in primary production (PP) is not accompanied by a corresponding increase in zooplankton grazing activity (Polimene et al., 2023; Tubay et al., 2013). This concept is supported by evidence showing that eutrophication is triggered by extended periods during which primary production remains ungrazed (Chislock et al., 2013; EEA, 2019; Eddy et al., 2021).

The central assumption of the TTI is that, in a healthy marine environment and over appropriate temporal scales, grazing activity should correlate with primary production (Kemp et al., 2001; Schmoker et al., 2013), regardless of the system’s trophic status. The strength of this correlation (Equation 2) is assessed by combining both the linear (Pearson) and rank (Spearman) correlation coefficients as follows: TTI = max RL , RR (2) where RL is the linear correlation coefficient between monthly depth-integrated primary production ( P P ) and grazing ( G r a z i n g both expressed in mmol N m⁻² month⁻¹: R L = c o r r c o e f P P t , G r a z i n g t , t + 1 (3) and R R is the rank correlation coefficient between PP and the Grazing to PP ratio: R R = S p e a r m a n P P t , G r a z i n g t , t + 1 : P P t (4)

The rank correlation coefficient allows the TTI to capture monotonic relationships that may be non-linear or steeper than linear. In both Equations 3, 4, t represents the months over which the fluxes are averaged.

Following the approach of Polimene et al. (2023), the TTI was calculated at each model grid point over a 3-year period to identify “problem areas” (TTI-Eutrophic Zones). A problem area is defined as a region where primary production is not sufficiently balanced by grazing activity, leading to potential ecological dysfunctions such as organic matter accumulation and anoxia. The threshold value of 0.7 was chosen based on the findings of Polimene et al. (2023), where areas with TTI values below this limit exhibited significant signs of eutrophication.

Conversely, regions where the TTI is equal to or greater than 0.7 are classified as unaffected areas by eutrophication, referred to in this study as areas in Good Environmental Status (GES_TTI), indicating a healthy balance between primary production and grazing.

Since the TTI calculation spans 3 years, the intermediate year is represented in the graphical outputs. For example, the year 2011 includes data from 2010 to 2012, and so on.

A decrease in chl concentration is considered a positive impact, while an increase indicates a negative impact. All percentage changes reported here refer to average values over the respective sub-basins, not to spatial GES areas. The closed boundary scenario had the strongest effect on chl concentrations, leading to an average decrease of 80% between 2014 and 2021 across the affected sub-basins (Figure 5). A negative impact exceeding 20% was observed in sub-basins from Kattegat to Kiel, whereas in the Bay of Mecklenburg (BoM), average concentrations decreased by 14%. The most pronounced improvement occurred under the barrier scenario in the Gulf of Finland (GoF), where average chl concentrations declined by 40%. The impact of the reversed wind scenario exceeded 7% in sub-basins from Kattegat to Bornholm. In contrast, the nutrient reduction scenario had a stronger influence than hydrographic factors only in the Gdansk Basin and Gulf of Riga (GoR). In all other sub-basins, physical drivers had a greater effect on chl concentrations than nutrient reductions.

TTI was most affected by the closed boundary scenario during 2014–2021 (Figure 6), with reduced salt inflow decreasing TTI values by more than 15% in the Kattegat to BoM region. Both the reversed wind and closed boundary scenarios had a greater impact on TTI than nutrient reduction in most basins. However, nutrient reductions were more effective in basins Gdansk, GoR, and GoF. Notably, the barrier scenario had the strongest effect in the Gulf of Finland, counteracting eutrophication mitigation efforts.

The average DIN:DIP ratio in the closed boundary scenario was more than 15% lower than in the reference scenario in basins Kattekat to BoM during 2014–2021. This decrease was reflected in both TTI and chl indicators. The DIN:DIP ratio increased in all basins under the barrier scenario except for the Gulf of Riga, though the change remained below 4%. In contrast, under the reversed wind scenario, the DIN:DIP ratio increased by 11% in basin 6. However, while TTI remained unchanged in this basin, chl concentrations showed a 6% decline.

In the closed boundary scenario, cyanobacteria concentrations increased by more than 50% in basins Kattekat to Kiel Bay and by 34% in Bay of Mecklenburg between 2014 and 2021. This increase was linked to an ∼8% reduction in bottom-layer oxygen concentrations. In other scenarios, bottom oxygen concentrations showed little change, while under the reversed wind scenario, oxygen levels improved by 5%–8% in basins 4–10. A minor oxygen increase (1%–2.5%) was also observed in the barrier scenario.

Several limitations should be acknowledged. First, annual aggregation of TTI may obscure seasonal trophic linkages and overrepresent winter-period correlations. Finer temporal resolution could help better resolve cause-effect relationships in energy transfer and nutrient cycling.

Second, the use of broad plankton functional groups may mask species-specific interactions that are critical for capturing shifts in food web structure and nutrient processing.

Third, our 11-year simulation window limits the ability to detect long-term processes such as ecological adaptation, regime shifts, and cumulative feedbacks. Consequently, while our findings highlight the short-term dominance of physical factors, the long-term effectiveness of nutrient load reductions—as underscored by Saraiva et al. (2019)—is likely underestimated.

Natural variability introduces more uncertainty into observed ecosystem indicators than previously recognized (Schimanke et al., 2012). Nevertheless, both Schimanke et al. (2012) and Saraiva et al. (2019) concluded that substantial improvements in Baltic Sea oxygen conditions can be achieved through continued nutrient load reductions—even under future climate change scenarios. Our results support this view but emphasize that ignoring the influence of physical factors may lead to misinterpretation of ecosystem responses.

For example, a lack of improvement in chlorophyll-a after nutrient reduction does not necessarily indicate management failure if improvements are observable in functional indicators such as TTI. Current eutrophication assessments under the Marine Strategy Framework Directive (MSFD, Descriptor 5) may be too narrow to fully capture the Baltic Sea’s dynamic complexity.

Accurate assessment of eutrophication trends therefore requires accounting for physical variability, especially ventilation, stratification, and temperature shifts. We advocate for an integrated assessment framework that combines state indicators (e.g., chlorophyll-a) with process-oriented metrics (e.g., TTI, stoichiometry).

Functional indicators enhance early-warning capacity, improve interpretation of ecosystem trajectories, and reduce the risk of misjudging recovery progress. Future modeling efforts should aim to resolve plankton at the species level and extend simulations to better capture the long-term effects of climate-driven hydrographic change.

Ultimately, effective eutrophication management must address not only nutrient inputs but also the stoichiometric and functional shifts shaped by physical factors.