Shallow sills and steady river flows could turn Norway’s Bolstadfjord into a stagnant lake

A land-locked fjord on Norway’s west coast survives on rare, dramatic breaths. New numerical simulations of Bolstadfjord show that its deep basin waters are refreshed only during extended periods of low river flow, when dense seawater slips over a threshold sill barely a couple of metres below the sea surface. The study, published in Ocean Dynamics, uses the Bergen Ocean Model to test how dredging, hydropower regulation and climate change could each tip the fjord toward stagnation, anoxia and hydrogen sulphide production at depth.

Bolstadfjord is connected to the outer fjord system through a narrow sill channel with a minimum depth estimated at between roughly 1 and 3 metres in historical surveys. Below that shallow gate lies a deep basin reaching about 160 metres, fed by the Vosso river, which delivers an average freshwater flow of around 113 cubic metres per second. Because the fjord’s average width of about one kilometre is far smaller than the internal Rossby radius of deformation, estimated at roughly 8 kilometres, the researchers could justify a two-dimensional depth-varying model, running 16-year simulations from 1974 to 1990 driven by measured river flows, tidal forcing with an M2 amplitude of 0.452 metres, and bi-monthly temperature and salinity observations from Sognesjøen.

The benchmark simulation reproduced the episodic character of deep water renewal. In early 1977, after river flows stayed below 50 cubic metres per second for a prolonged period while surface density outside the sill climbed to a peak of about 26.5 kilograms per cubic metre, dense water poured continuously across the sill and flushed the full depth of the basin. Dissolved oxygen in the bottom waters jumped from anoxic levels to more than 5 millilitres per litre, displacing hydrogen sulphide-rich water. Smaller renewal events in 1978, 1979, 1982 and 1987 temporarily re-oxygenated intermediate depths but never reached the bottom, and between events the lower basin steadily slid back toward anoxia, with hydrogen sulphide reappearing by the mid-1980s.

The model runs identified sill depth as a primary control on this delicate balance. When the researchers deepened the threshold from 2.3 to 3.3 metres, simulating dredging, renewal events became larger and more frequent, and the mean density excess of basin water below 50 metres stabilised near 21 kilograms per cubic metre. Shallowing the sill to 1.3 metres had the opposite effect: intrusions dwindled, and the mean density excess below 50 metres decayed monotonically to about 5 kilograms per cubic metre by 1990. An extended simulation with a sill depth of just 0.8 metres showed the basin turning fully fresh by 1999, effectively converting the fjord into an inland lake.

The team framed these outcomes with a densimetric Froude number, a dimensionless ratio comparing the momentum of the outflowing freshwater layer against the buoyancy of the dense inflow, which laboratory experiments have shown must stay below about 0.125 for saline water to intrude across a sill. For conditions during the 1977 event, this number ranged from 0.053 at a 3.3-metre sill to 0.214 at a 1.3-metre sill, correctly predicting intrusion at the two deeper thresholds and blockage at the shallowest. The same calculation for a constant mean river flow of 113 cubic metres per second yielded a value of 0.725, far above the blockage threshold.

That constant-flow scenario speaks directly to hydropower regulation. When the model replaced the naturally variable 1974–1990 river record with a steady discharge equal to the long-term mean, no significant dense water intrusion reached the basin at all. Density excess in the overlying waters declined steadily while anoxic volumes grew. The researchers emphasise that the extended low-flow windows in the natural record, not the average flow itself, are what permit renewal; any regulation regime that removes or shortens these quiet periods would suppress deep water renewal and degrade water quality at depth, even when dense seawater is available just outside the sill.

Tidal forcing, by contrast, proved comparatively marginal. Simulations with no tide at all and with a 1-metre amplitude bracketed the benchmark case, shifting equilibrium density at the basin bottom between roughly 17.7 and 19.0 kilograms per cubic metre and modestly altering anoxic water volumes through changes in tidal mixing. Crucially, the 1977 renewal still penetrated to the bottom even with tides removed entirely, confirming that the freshwater outflow, rather than tidal pumping, is the decisive gatekeeper at this site.

Climate change scenarios painted a subtler but still consequential picture. Raising the temperature of the intruding coastal water by 1 degree Celsius, the order of warming observed in North Atlantic Water at Sognesjøen between 1980 and 2010, and reducing its salinity by 1 practical salinity unit lowered basin density excess by about 0.7 to 0.8 kilograms per cubic metre and slightly increased anoxic water volumes. Adding a 25 per cent increase in river flows, representing higher rainfall runoff, pushed the reduction beyond 1 kilogram per cubic metre and, notably, prevented the simulated 1977-style event from reaching the basin bottom, implying reduced replenishment of the deepest waters and prolonged anoxia and hydrogen sulphide production.

Field measurements lend some support to these modelled trends. Comparing profiles collected in June 1932 with new measurements from May 2024 shows deeper basin waters slightly warmer and less saline in 2024, consistent with reduced density of intruding waters, although the differences were not large enough to change the dynamical regime: anoxic conditions with zero dissolved oxygen persisted at depths of 80 metres and below in both eras. The researchers caution that the time elapsed since the last major intrusion is unknown in either case, so the comparison hints at, but does not prove, a long-term decline in full deep water renewal driven by warming and dilution.

The study’s broader message is that Bolstadfjord’s connection to the sea hangs on a few metres of channel depth and the rhythm of its river. Natural sedimentation could gradually seal the sill and push the fjord toward freshwater lake status, while dredging would strengthen and renew the saline lifeline. The authors note that artificial oxygenation by injecting freshwater at depth, as trialled in Swedish and Norwegian fjords, would require more than 20 cubic metres per second of continuous flow to force renewal within a year, which they judged impractical here. Management of the sill bathymetry and of hydropower-driven flow regimes, they conclude, will determine whether the basin’s stagnant, sulphidic bottom waters can ever be flushed again.

Subject of Research: Modelled deep water renewal and anoxia in the land-locked Bolstadfjord, Norway

Article Title: Basin water renewal in land-locked fjords with shallow threshold sills: impacts from natural and anthropogenic change

Article References: Cuthbertson, A., Hanson, E., Mundal, E., Normann, E., Berntsen, J., & Aksnes, D. (2026). Basin water renewal in land-locked fjords with shallow threshold sills: impacts from natural and anthropogenic change. Ocean Dynamics, 76(11), Article 118. https://doi.org/10.1007/s10236-026-01870-z

Image Credits: AI Generated

DOI: 10.1007/s10236-026-01870-z

Keywords: Bolstadfjord, fjord oceanography, deep water renewal, threshold sill, anoxia, hydrogen sulphide, Bergen Ocean Model, hydropower regulation, climate change, densimetric Froude number, dissolved oxygen, Vosso river