Abstract Stratified oceanic turbulence is strongly intermittent in time and space, and therefore generally underresolved by currently available in situ observational approaches. A promising tool to at least partly overcome this constraint are broadband acoustic observations of turbulent microstructure that have the potential to provide mixing parameters at orders of magnitude higher resolution compared to conventional approaches. Here, we discuss the applicability, limitations, and measurement uncertainties of this approach for some prototypical turbulent flows (stratified shear layers, turbulent flow across a sill), based on a comparison of broadband acoustic observations and data from a free-falling turbulence microstructure profiler. We find that broadband acoustics are able to provide a quantitative description of turbulence energy dissipation in stratified shear layers (correlation coefficient r = 0.84) if the stratification parameters required by the method are carefully preprocessed. Essential components of our suggested preprocessing algorithm are 1) a vertical low-pass filtering of temperature and salinity profiles at a scale slightly larger than the Ozmidov length scale of turbulence and 2) an automated elimination of weakly stratified layers according to a gradient threshold criterion. We also show that in weakly stratified conditions, the acoustic approach may yield acceptable results if representative averaged vertical temperature and salinity gradients rather than local gradients are used. Our findings provide a step toward routine turbulence measurements in the upper ocean from moving vessels by combining broadband acoustics with in situ CTD profiles.
Abstract The circulation pathways and subsurface cooling and freshening of warm deep water on the central Amundsen Sea shelf are deduced from hydrographic transects and four subsurface moorings. The Amundsen Sea continental shelf is intersected by the Dotson trough (DT), leading from the outer shelf to the deep basins on the inner shelf. During the measurement period, warm deep water was observed to flow southward on the eastern side of DT in approximate geostrophic balance. A northward outflow from the shelf was also observed along the bottom in the western side of DT. Estimates of the flow rate suggest that up to one-third of the inflowing warm deep water leaves the shelf area below the thermocline in this deep outflow. The deep current was 1.2°C colder and 0.3 psu fresher than the inflow, but still warm, salty, and dense compared to the overlying water mass. The temperature and salinity properties suggest that the cooling and freshening process is induced by subsurface melting of glacial ice, possibly from basal melting of Dotson and Getz ice shelves. New heat budgets are presented, with a southward oceanic heat transport of 3.3 TW on the eastern side of the DT, a northward oceanic heat transport of 0.5–1.6 TW on the western side, and an ocean-to-glacier heat flux of 0.9–2.53 TW, equivalent to melting glacial ice at the rate of 83–237 km3 yr−1. Recent satellite-based estimates of basal melt rates for the glaciers suggest comparable values for the Getz and Dotson ice shelves.
Ship-related energy pollution has received increasing attention but almost exclusively regarding radiated underwater noise, while the effect of ship-induced turbulence is lacking in the literature. Here we present novel results regarding turbulent wake development, the interaction between ship-induced turbulence and stratification, and discuss the impact of turbulent ship wakes in the surface ocean, in areas with intense ship traffic. The turbulent wake development was studied in situ , using Acoustic Doppler Current Profilers (ADCP) and conductivity, temperature, depth (CTD) observations of stratification, and through computational fluid dynamics (CFD) modelling. Our results show that the turbulent wake interacts with natural hydrography by entraining water from below the pycnocline, and that stratification influences the turbulent wake development by dampening the vertical extent, resulting in the wake water spreading out along the pycnocline rather than at the surface. The depth and intensity of the turbulent wake represent an unnatural occurrence of turbulence in the surface ocean. The ship-induced turbulence can impact local hydrography, nutrient dynamics and increase plankton mortality due to physical disturbance, especially in areas with intense traffic. Therefore, sampling and modelling of e.g., contaminants in shipping lanes need to consider hydrographic conditions, as stratification may alter the depth and spread of the wake, which in turn governs dilution. Finally, the frequent ship traffic in estuarine and coastal areas, calls for consideration of ship-induced turbulence when studying hydrographic processes.
Hordoir et al. [2015, hereinafter HALDK] have used a three-dimensional hydrodynamic model to investigate the influence of sea level rise on the inflow of saline water to the Baltic Sea. They find that the inflow of water with salinity larger than 17 g kg−1 increases with increasing sea level. Based on their analyses, they conclude that this increase is caused both by an increased cross-sectional area of inflowing water and by decreased mixing in the Danish straits. This is an interesting result with links to the discussions about decreased Baltic Sea ventilation caused by the bridges over the Great Belt and the Sound [e.g., Stigebrandt, 1992; Jakobsen et al., 2010]. The question also has relevance for the sensitivity to sea level rise of other semienclosed seas, estuaries, and fjords with complicated entrances. In this comment, I will argue that increased saline inflows with increased sea level, in models and reality, are not caused by decreased mixing. A more reasonable explanation for relatively large increases in saline water inflows to the Baltic Sea with increasing sea level is that the barotropic volume fluxes increase more in the shallowest and shortest connection (the Sound) than in the other connections (the Belt Sea), and that this increase causes decreased blocking of saline water by the Drogden sill in the Sound. The Baltic Sea is a semienclosed, brackish, sea connected to the North Sea through narrow straits with shallow sills (8–20 m) (Figure 1), and its deep basins suffer from oxygen deficiency, causing large environmental problems such as, e.g., internal sources of phosphorous in the bottoms [e.g., Conley et al., 2002] feeding large cyanobacterial blooms and maintaining hypoxic conditions. The basin water quality is highly sensitive to the saline inflows, which is the reason why the question about their strengths and the understanding of factors controlling them are important for the scientific community. Today, about 25% of the volume flux to and from the Baltic Sea goes through the Sound, but the Sound is much more important when it comes to salt inflows. During periods without major inflows, the Sound may be the major contributor of salt to the Baltic [Lintrup and Jakobsen, 1999]. (a) Bathymetry of the Danish straits in meters, and (b) relative depth increase in % per meter sea level rise. In the following, I will propose an alternative theory for increasing inflow of saline water to the Baltic with rising sea level. The theory contains two parts: (i) the volume flux through the Sound is more sensitive to sea level rise than that through the Belt Sea. (ii) The amount of dense water passing the Drogden sill in the Sound is determined by a baroclinic control in the narrow northern end of the Sound as suggested by Nielsen [2001]. When the sea level rises, the frictional resistances in the Belts and Sound decrease, mainly due to the increasing water depth. This relative depth increase will be much larger in the Sound than in the Belt Sea (Figure 1b) because of the shallow depth of the Drogden sill (<10 m) whereas a large part of the Belt Sea is deeper than 20 m. According to the model results of HALDK, the barotropic volume fluxes increase with the same relative amount as the relative increase in cross-sectional area. This means that the spatially averaged barotropic velocities remain similar as today. I will argue that this must mean that the velocities in the Sound increase whereas those in the Belt Sea decrease. The average of (2) over the whole entrance region must be zero in order to obtain unchanged average velocities. The value of the second term within the brackets must therefore be close to −1 in average, meaning that the instantaneous sea level differences between the Kattegat and the southern Baltic decrease due to decreasing resistance. The relative change of the sea level gradient (I−1∂I/∂η) must be similar for the Sound and the Belt Sea, since these two straits connect the same basins. It therefore follows logically that the Sound with a smaller depth than the mean entrance depth will experience an increasing velocity, whereas the Belt Sea with a larger depth than the mean entrance depth will experience a decreasing velocity. In order to get a rough estimate of the velocity change, we assume that the value of I−1∂I/∂η is about −(15 m)−1 and that the depth of the Drogden sill is about 8 m, yielding a relative velocity increase over the Drogden sill of about 3% per meter sea level rise. More importantly, the cross-sectional area over the Drogden sill increases with about 13%, giving a total increase in the magnitude of barotropic volume fluxes of about 16% per meter sea level rise. Nielsen [2001] showed observational evidence of baroclinically controlled flow in the narrow strait north of the Drogden sill in the Sound, and suggested that the amount of saline water passing over the Drogden sill is governed by this baroclinic control. The process is illustrated in Figure 2. At small barotropic velocities, the sill will block the lower layer velocities, by radiating internal waves upstream adjusting the lower layer pressure gradient to a value that balances the pressure gradient in the surface layer. At some point when the velocities increase, a baroclinic control will be obtained in the narrowest transect north of the sill. When the velocities increase above this point, information about the sill can no longer be radiated upstream, and the baroclinic volume fluxes in the two layers are now governed by the control, whereas the barotropic flow is still mainly controlled by bottom friction over the shallow sill. As we will see below, the upper layer volume flux does not increase much above this point and the increased flow through the strait will have to go via the lower layer. The lower layer fluid entering the Sound causes an uplift of dense water above the sill which eventually gets mixed with the upper layer fluid and passes over the sill, as sketched by Nielsen [2001], see Figure 2. The lower layer volume fluxes are thus varying in a highly nonlinear fashion in response to increased barotropic volume fluxes and changed sea level. Sketch of a controlled flow situation at the narrowest cross section (HH), and mixing over the sill (D). Reproduced from Nielsen [2001]. With an upstream upper layer depth of h10 = 15 m (h1c,init = 10 m), a surface (bottom) layer salinity of 17 (33) g/kg, and a surface layer (total) area of about 37,000 (84,000) m2 [Baltic Sea Hydrographic Commission, 2013], the control sets in at a barotropic velocity of U0c = 0.49 m s−1 or a barotropic volume flux of Q0c = 41,000 m3 s−1. This is a rather rough estimate that does not take into account the cross-channel variations in velocity and interface level caused by channel curvature and earth rotation that Nielsen [2001] observed, but it is probably as good as, or better than, what can be estimated with a coarse resolution numerical model for a similar stratification. When the barotropic velocities increase above the critical value, internal waves cannot propagate upstream past the control, and the downstream flow becomes supercritical until the flow adjusts to the downstream situation via an internal hydraulic jump. With increasing barotropic velocities, the composite Froude number in the control section remains equal to 1 (equation 3), which mainly limits the upper layer water. This means that the increasing barotropic volume fluxes mainly cause increased lower layer volume fluxes. As the difference between upper and lower layer velocities decrease, the upper layer depth increases toward the upstream upper layer depth, h10. The upper and lower layer velocities can easily be determined as functions of h1 varying between 2/3 h10 and h10 by using (5) and (4). With knowledge about how the upper and lower layer areas depend on h1 [Baltic Sea Hydrographic Commission, 2013], one can then calculate the upper and lower layer volume fluxes, and the barotropic volume flux. Figure 3 shows the upper and lower layer volume fluxes as functions of the barotropic volume flux for the stratification mentioned above. These are calculated, using a total depth, H = 30 m, in (5). Volume fluxes in the upper (black) and lower (blue) layers, and inflow salinity (red) as function of present-day barotropic volume fluxes. The dashed lines correspond to a 1 m sea level rise and a 16% increase in barotropic volume fluxes. After some time with nonzero bottom layer fluxes, the halocline in the Sound has risen above sill level, and the volume fluxes of upper and lower layer water onto the sill eventually become equal to those through the control section. Under the assumption that the water becomes totally mixed over the sill, one can calculate the salinity of inflowing water, which is also shown in Figure 3. The observed inflow salinities of 20–22 g kg−1 at volume fluxes between 45 and 60,000 m3 s−1 observed by Sellschopp et al. [2006] are seen to correspond reasonably well with this simple model. A more direct way of getting deep water over the sill is by local aspiration at the sill [e.g., Mattsson, 1996a], which can happen due to a variety of processes, two of which are discussed here for completeness—frictionless and frictional flow. Equation 8, with c− given by (9), (10), is integrated numerically until the lower layer is as thick as the sill is high, or until the upstream wave velocity becomes zero, which is identical to condition (3). With a stratification as in section 3.1 and a sill depth of h1s = 8 m, the interface reaches sill level when the barotropic velocity is U0 = 0.24 m s−1 upstream of the sill (U0s = 0.90 m s−1 over the sill). This corresponds to a volume flux of about 75,000 m3 s−1, i.e., well above the volume flux that causes control in the narrow transect to the north. Note that the aspiration calculated this way happens prior to that calculated with the momentum method used by Mattsson [1996a]. This prediction of the interface elevation caused by internal wave radiation may be exaggerated because the waves will be partly reflected from the northern end of the Sound and force the interface back toward the upstream condition. This mechanism will therefore not be important for slowly varying flows, but it may be important for intensive inflows where the bottom layer north of the sill takes some time to fill up, whereas this effect happens directly. In the following, we limit the discussion to overflows caused by the control. With an increasing sea level, the changes to the layer volume fluxes as function of barotropic volume fluxes are relatively small, i.e., the relatively deep, narrow opening continues to control the baroclinic flow in almost the same manner. If, however, the predicted fluxes and salinities are plotted as function of present-day barotropic fluxes (16% smaller than in the future), the changes are significant, Figure 3. The inflows of deep Kattegat water will occur at inflow situations where there are no dense inflows today. At situations with weak deep inflows today, the deep inflows increase dramatically with a 1 m sea level rise. As an example, a meteorological situation causing a present-day barotropic inflow of 50,000 m3 s−1 will cause an 80% larger lower layer volume flux and a 1.5 g/kg increase in salinity at 1 m sea level rise. I therefore see this as one possible mechanism causing relatively large increases in saline inflows with rising sea level. Similar effects may happen in the other straits, but since the barotropic volume flux increases are smaller there, the saline inflow increases will probably also be less dramatic there. For a more thorough analysis, one should analyze the recorded current and salinity data at the Drogden sill, to validate the analysis above, and to combine the predicted changes with probability functions for various current strengths. This would also require knowledge about the stratification in southern Kattegat. This is beyond the scope of this comment. I would, however, suggest that future model studies of changes in salt water inflows to the Baltic look into the relative changes in volume fluxes in the Belt Sea and the Sound, and that they check the model description of the dynamics and mixing processes in the Sound to make sure that these are realistically described. More observations of controlled and uncontrolled situations in the Sound and aspiration and mixing of deep water over the Drogden sill are probably needed in order to validate these parts of the models. The bathymetric data used to prepare Figure 1 are taken from the Baltic Sea Hydrograpic Commission (2013) and are available at URL: http://data.bshc.pro. There are no other new data in this work.
Abstract. Turbulent diapycnal mixing is important for the estuarine circulation between basins of the Baltic Sea as well as for its local ecosystems, in particular with regard to eutrophication and anoxic conditions. While the interior of the basins is overall relatively calm, stratified flow over steep bathymetric features is known as a source for strong turbulent mixing. Yet, current in situ observations often cannot capture dynamic and intermittent turbulent mixing related to overflow over rough bathymetry. We present observational oceanographic data together with openly accessible high-resolution bathymetry from a prototypical sill and an adjacent deep channel in the sparsely-sampled Southern Quark located in the Åland Sea, connecting the Northern Baltic Proper with the Bothnian Sea. Our data include high resolution broadband acoustic observations of turbulent mixing, in situ microstructure profiler measurements, and current velocities from Acoustic Doppler Current Profilers and were acquired during two one-week cruises in February–March 2019 and 202. A temporally reversing non-tidal stratified flow over the steep bathymetric sill created a dynamic and extremely energetic environment. Saltier, warmer, and less oxygenated deep water south of the sill was partly blocked, the reversing flow was at times hydraulically controlled with hydraulic jumps occurring on both sides of the sill, and sub-mesoscale processes in the surface layer leading to high spatial variability at small scales. Mixing and vertical salt flux rates were increased by 3–4 orders of magnitude in the entire water column in the vicinity of the sill compared to reference stations not directly influenced by the overflow. We suggest based on acoustic observations and in situ measurements that underlying mechanisms for the highly increased mixing across the halocline are a combination of shear and topographic lee waves which are breaking at the halocline interface. We anticipate that the resulting deep- and surface-water modification in the Southern Quark directly impacts exchange processes between the Bothnian Sea and the Northern Baltic Proper and that the observed mixing is likely important for oxygen and nutrient conditions in the Bothnian Sea. Our results contribute to the knowledge on turbulent mixing processes in the Åland Sea and can help to improve mixing parametrizations in numerical models of the area.
Abstract. Turbulent diapycnal mixing is important for the estuarine circulation between basins of the Baltic Sea as well as for its local ecosystems, in particular with regard to eutrophication and anoxic conditions. While the interior of the basins is overall relatively calm, stratified flow over steep bathymetric features is known as a source for strong turbulent mixing. Yet, current in situ observations often cannot capture dynamic and intermittent turbulent mixing related to overflow over rough bathymetry. We present observational oceanographic data together with openly accessible high-resolution bathymetry from a prototypical sill and an adjacent deep channel in the sparsely-sampled Southern Quark located in the Åland Sea, connecting the Northern Baltic Proper with the Bothnian Sea. Our data include high resolution broadband acoustic observations of turbulent mixing, in situ microstructure profiler measurements, and current velocities from Acoustic Doppler Current Profilers and were acquired during two one-week cruises in February–March 2019 and 202. A temporally reversing non-tidal stratified flow over the steep bathymetric sill created a dynamic and extremely energetic environment. Saltier, warmer, and less oxygenated deep water south of the sill was partly blocked, the reversing flow was at times hydraulically controlled with hydraulic jumps occurring on both sides of the sill, and sub-mesoscale processes in the surface layer leading to high spatial variability at small scales. Mixing and vertical salt flux rates were increased by 3–4 orders of magnitude in the entire water column in the vicinity of the sill compared to reference stations not directly influenced by the overflow. We suggest based on acoustic observations and in situ measurements that underlying mechanisms for the highly increased mixing across the halocline are a combination of shear and topographic lee waves which are breaking at the halocline interface. We anticipate that the resulting deep- and surface-water modification in the Southern Quark directly impacts exchange processes between the Bothnian Sea and the Northern Baltic Proper and that the observed mixing is likely important for oxygen and nutrient conditions in the Bothnian Sea. Our results contribute to the knowledge on turbulent mixing processes in the Åland Sea and can help to improve mixing parametrizations in numerical models of the area.
Abstract. The North Sea and the Baltic Sea still experience eutrophication and deoxygenation despite large international efforts to mitigate such environmental problems. Due to the highly different oceanographic frameworks of the two seas, existing modelling efforts have mainly focused on only one of the respective seas, making it difficult to study interbasin exchange of mass and energy. Here, we present NEMO–SCOBI, an ocean model (NEMO-Nordic) coupled to the Swedish Coastal and Ocean Biogeochemical model (SCOBI), that covers the North Sea, the Skagerrak–Kattegat transition zone and the Baltic Sea. We address its validity to further investigate biogeochemical changes in the North Sea–Baltic Sea system. The model reproduces the long-term temporal trends, the temporal variability, the yearly averages and the general spatial distribution of all of the assessed biogeochemical parameters. It is particularly suitable for use in future multi-stressor studies, such as the evaluation of combined climate and nutrient forcing scenarios. In particular, the model performance is best for oxygen and phosphate concentrations. However, there are important differences between model results and observations with respect to chlorophyll a and nitrate in coastal areas of the southeastern North Sea, the Skagerrak–Kattegat transition zone, the Gulf of Riga, the Gulf of Finland and the Gulf of Bothnia. These are partially linked to different local processes and biogeochemical forcing that lead to a general overestimation of nitrate. Our model results are validated for individual areas that are in agreement with policy management assessment areas, thereby providing added value with respect to better contributing to international programmes aiming to reduce eutrophication in the North Sea–Baltic Sea system.
The efficiency of mixing in stably stratified systems where the turbulent mixing is confined to intermittent patches is investigated theoretically. It is possible to define two different flux Richardson numbers for mixing in such a system. One, the small-scale flux Richardson number, Rft, is based on the initial potential energy increase caused by small-scale turbulent mixing within the patches. This is the parameter that is obtained from laboratory and numerical experiments intended to determine turbulent mixing efficiencies. The other, the large-scale flux Richardson number, Rf, is based on the final potential energy increase, obtained after the mixed fluid has spread out laterally in the system. This is the relevant parameter for determining large-scale, irreversible, changes in the stratification caused by mixing. It is shown that the large-scale flux Richardson number is always smaller than the small-scale flux Richardson number, and that the difference can be almost a factor of 2. The commonly used mixing efficiencies, 0.17–0.2, obtained from laboratory and numerical experiments of small-scale homogeneous turbulence, are a measure for the small-scale flux Richardson number Rft rather than the large-scale flux Richardson number Rf. If the maximum small-scale flux Richardson number Rft = 0.2 is relevant for mixing in oceanic patches, one should use Rf = 0.11 for the large-scale flux Richardson number. The latter value is supported by results from recent microstructure experiments in the ocean.
