Search

The estimation of expected damage and losses from natural hazards requires that uncertainties in the modelling and knowledge of future events be quantified and taken into consideration. This is true not only in a fully probabilistic context but also in future scenario calculations, particularly when looking at two or more cascading hazards in which the link between them is not univocal. An offshore earthquake that triggers a tsunami would be one such case. Even if the moment magnitude and rupture size and location of the earthquake were fully defined, it is not possible to know a priori the slip distribution along the rupture and the subsequent co-seismic topographic displacements. Many feasible slip distributions can be associated with the same moment magnitude and dimensions of the rupture, and these lead to a distribution of subsequent topographic displacements and, with that, a diversity of tsunami outcomes. Exactly how much variety exists in the resulting tsunamis, in terms, for example, of maximum wave height or maximum flow velocity at points of interest, and, ultimately, damage to buildings and losses, is the question driving the present study, which is part of the “risk workflow for CAScading and COmpounding hazards in COastal urban areas” (CASCO) project. The ultimate objective is to understand the relevance of this uncertainty and whether it needs to be modelled in the whole damage/loss calculation chain.To investigate this, 500 realisations of stochastically generated rupture slip have been produced for the 1908 Mw 7.1 Messina earthquake, whose rupture source is taken from the Italian Database of Individual Seismogenic Sources (DISS). The subsequent realisations of ground surface deformation (at the bottom of the sea and on land) were used as input to run realisations of the resulting tsunami in the Strait of Messina, eastern Sicily and western Calabria with the TsunAWI software. Maximum wave heights, maximum absolute velocities and maximum flux can vary significantly for selected observation points along the coast and within the Messina Strait. While a weak correlation has been identified between these tsunami outputs and inputs such as the maximum initial co-seismic vertical displacement, a stronger correlation has been observed with respect to the distance to the centroid of rupture slip. So far, results indicate that the uncertainty in the co-seismic slip along the rupture and the subsequent vertical displacements has a relevant impact on the resulting tsunami, suggesting that this source of uncertainty should not be entirely neglected in models. Using these tsunami outputs to estimate damage to buildings in the area allows us to understand the ultimate final impact on damage and loss calculations, and to develop and test strategies to reduce the resulting computational demand.

Numerical simulations of the tsunami inundation processes require a highly nonlinear scheme. The main inundation properties, such as theflow depth and velocity depend critically on topographical imprints and bottom friction parameters. Here, we investigate the tsunami inundation in Lima and Callao resulting from the extensive 1746 (Mw 9.0) earthquake that ruptured along the Peruvian coast.Two numerical tsunami codes have been used in this analysis based on shallow water equations. We determine the relative importance of different parts in these equations with a focus on nonlinear terms. Particular focus is put on the momentum advection, bottom friction, and volume conservation in different mesh (triangular meshes and nested grids). We determine the influence on large-scale quantities like inundation extent and volume, flow velocities, and small-scale fluctuations. In that respect, also sensitivities regarding the bottom friction parameters are investigated.

The results of an extensive series of numerical experiments of the GNOM-LS model for modelling the physical and energy characteristics of tsunami waves generated by landslides are presented. Based on the published data on the tsunami on 28 September 2018 in Palu Bay, we analysed the sensitivity of the distribution of wave heights along the coastline formed by the landslide system, depending on the characteristics of these landslides and model parameters. The complexity of the work lies in the lack of a holistic picture of the initial information about landslides, their number and accurate measurement data on the height of the waves of the event. We attempted to restore these conditions by comparing numerical simulations for various initialisations of the landslide system with available observational data. It is revealed that the simulated system has a very high sensitivity to the initial conditions and characteristics of landslides. An essential task of the work is interpreting a complex picture of the nonlinear interaction of tsunami waves with minor changes in the initial characteristics of landslides. Based on the numerical simulation of single landslides and a complete system of landslides, an analysis of the complex structure of the nonlinear interaction of tsunami waves is carried out.

Special issue EPOS a Research Infrastructure in solid Earth: open science and innovation .-- 21 pages, 8 figures, Tsunamis constitute a significant hazard for European coastal populations, and the impact of tsunami events worldwide can extend well beyond the coastal regions directly affected. Understanding the complex mechanisms of tsunami generation, propagation, and inundation, as well as managing the tsunami risk, requires multidisciplinary research and infrastructures that cross national boundaries. Recent decades have seen both great advances in tsunami science and consolidation of the European tsunami research community. A recurring theme has been the need for a sustainable platform for coordinated tsunami community activities and a hub for tsunami services. Following about three years of preparation, in July 2021, the European tsunami community attained the status of Candidate Thematic Core Service (cTCS) within the European Plate Observing System (EPOS) Research Infrastructure. Within a transition period of three years, the Tsunami candidate TCS is anticipated to develop into a fully operational EPOS TCS. We here outline the path taken to reach this point, and the envisaged form of the future EPOS TCS Tsunami. Our cTCS is planned to be organised within four thematic pillars: (1) Support to Tsunami Service Providers, (2) Tsunami Data, (3) Numerical Models, and (4) Hazard and Risk Products. We outline how identified needs in tsunami science and tsunami risk mitigation will be addressed within this structure and how participation within EPOS will become an integration point for community development, With the institutional support of the ‘Severo Ochoa Centre of Excellence’ accreditation (CEX2019-000928-S)

Extreme climatic and geophysical events pose a threat to societies and have the capacity to cause significant damage and losses whenever they occur, both in their immediate aftermath and in the medium- to long-term. Their consequences can be amplified even further when more than one event affects the same geographical areas within a short time. Be it cascading hazards, in which one event triggers the next, or simply hazards that happen to occur simultaneously (“compounding” hazards), estimation of their cumulative consequences is challenging because the action of one event affects the exposure and vulnerability to the next one. While the efforts from the research community to develop multi-hazard perspectives have increased considerably in recent years, multiple remaining challenges require strongly-coordinated efforts across different disciplines and areas of expertise to tackle them with the most appropriate tools.With a multidisciplinary team of scientists from four different Helmholtz research centres in Germany, we have started working on the CASCO project (2022-2024), in which we will develop an integrated risk workflow for CAScading and COmpounding hazards in COastal urban areas by focusing on a series of events occurring around Mount Etna (Italy). The case-scenario starts with a strong earthquake that triggers a submarine collapse at the eastern flank of Mount Etna, an area already known to be unstable, and both the earthquake and the landslide trigger a tsunami that hits the coasts of Sicily and Calabria. Almost concomitantly, a heatwave or heavy rainfall happens to affect the same regions, further stressing the population that had been affected by the combined effects of the earthquake and tsunami.The project will be directed towards the modelling of the cascading earthquake, landslide and tsunami events, the compounding heatwave and rainfall, as well as their immediate impacts in terms of cumulative damage and casualties. Moreover, the medium- to long-term response in urban dynamics and the effect of these extreme events on the economic development of the affected populations will be explored. By focusing on a tangible scenario, CASCO will not only tackle the challenges associated with bringing together the whole risk chain (which will be valid beyond our case-study) but also produce outcomes that help increase awareness of such extreme events and the need for societies to develop suitable strategies to strengthen their resilience and improve their disaster response.

Submesoscale features with profound impact on ocean dynamics and climate-relevant fluxes are frequently observed in the upper ocean including Arctic region. Yet, modelling these features remains a challenge due to the difficulties in the parameterization of submesoscale processes and high resolution required, in particular, in the polar regions. The most effective way to study such phenomena is joint modelling and observational work. Several autonomous observation platforms have been deployed as part of Multidisciplinary drifting Observatory for the Study of Arctic Climate (MOSAiC) experiment within an approximately 50 km radius around the central observatory. Data from these buoys in combination with data from the central observatory provide a unique opportunity to reconstruct 3D water properties and velocity by constraining a numerical model that resolves the dynamics of the (sub-)mesoscale. It turns out that a minimum root mean square error between results of an optimal interpolation and observations indicates a characteristic length scale of about 7.5 km, corresponding approximately the first-mode barolinic Rossby radius in the area of investigation. However, results of the interpolation are questionable at the sub-mesoscale due to the distribution of the buoy observations in time and horizontal space. In order to describe the in-situ data to achieve a better characterization and understanding of (sub-)mesoscale dynamics we developed and applied a modification of the 3D regional model FESOM-C. The observed temperature and salinity were used to nudge the model to obtain an optimized solution at the resolution of the models. A series of simulations with different horizontal resolutions and model parameters make it possible to analyze the ability of models of this type to reproduce the observed dynamics, to estimate eddy kinetic energy and power spectra, and to compare findings with the observations used to nudge the model. We will show the eddy-induced fluxes and characteristics of eddies along the track of the beginning winter MOSAiC drift.

Using the depth (z) and density (ϱ) frameworks, we analyze local contributions to AMOC variability in a 900‐year simulation with the AWI climate model. Both frameworks reveal a consistent interdecadal variability; however, the correlation between their maxima deteriorates on year‐to‐year scales. We demonstrate the utility of analyzing the spatial patterns of sinking and diapycnal transformations through depth levels and isopycnals. The success of this analysis relies on the spatial binning of these maps which is especially crucial for the maps of vertical velocities which appear to be too noisy in the main regions of upwelling and downwelling because of stepwise bottom topography. Furthermore, we show that the AMOC responds to fast (annual or faster) fluctuations in atmospheric forcing associated with the NAO. This response is more obvious in the ϱ than in the z framework. In contrast, the link between AMOC and deep water production south of Greenland is found for slower fluctuations and is consistent between the frameworks.

Space geodetic estimates of ocean bottom pressure (OBP) derived by several analysis centres are evaluated. To this end, an array of 14 in situ bottom pressure recorders has been deployed between South Africa and Antarctica. The continuous measurement period of four years (2011 to 2014) and a recorder spacing of roughly 2.8 degrees latitude allows an in-depth analysis of bottom pressure variability.Our goal is to relate OBP from GRACE to in situ observations and detect which spatial and temporal features are reproduced. The recorders in the southern part of the transect generally tend to be in better agreement with GRACE and better reflect longer spatial scales of ocean bottom pressure. Over the vast expanse of the Antarctic Circumpolar Current annual and semi-annual cycles are weak (about 1cm equivalent water height (EWH)) and not reproduced well by GRACE. Variability in general amounts to a standard deviation of 2cm. This level is well captured and correlations on the order of 0.5 are found.Mean values and trends of OBP cannot be identified due to the instrumental setup. Close to the Agulhas Retroflection, signals of up to 30cm EWH are found, which cannot be resolved by GRACE. Our analysis reveals: GRACE OBP possesses longer space and time scales than in situ OBP and it misses eddy-scale signals. Filtering with DDK4 appears to be preferable to DDK6.

The dynamics of the boundary layer of the ocean significantly affect the interaction between ocean and atmosphere and, as a result, global climate. The sub-ice boundary layer of the ocean and its dynamics have not been thoroughly studied because of the extremely difficult conditions for observation, in particular during winter. Current understanding of spatial-temporal variability of (sub)mesoscales of the upper Arctic Ocean is extremely limited.At the same time, one of the most important features of the upper ocean layers are the small-scale processes that influence and possibly determine the vertical and horizontal transport of heat, salt, and biologically relevant substances. As a consequence, mathematical models, in particular climate models, experience serious difficulties in parameterization of processes not resolved by the models because of the lack sufficient knowledge to detail the spatial variability at the (sub-)mesoscale.To a better characterization and understanding of (sub)mesoscale dynamics and its role in vertical transport of energy and mass we apply a 3D regional ocean model FESOM-C. The observed vertical hydrological structure and a corresponding reconstructed horizontal temperature and salinity fields were imposed as a part of the forcing for the numerical model. These fields and information about the vertical hydrological structure were utilized by the model as initial conditions and for constraining (nudging) during the spin-up period. After the initial spin-up period, once the model had adjusted to our initial conditions, we performed several free runs.We expect that our 3D numerical studies of eddy properties will contribute to a better characterisation and understanding of (sub)mesoscale dynamics in the Arctic Ocean and its role in the vertical transport of energy and mass.

The City of Venice and the surrounding lagoonal ecosystem are highly vulnerable to variations in relative sea level. In the past ~150 years, this was characterized by a secular linear trend of about 2.5 mm/year resulting from the combined contributions of vertical land movement and sea-level rise. This literature review reassesses and synthesizes the progress achieved in understanding, estimating and predicting the individual contributions to local relative sea level, with focus on the most recent publications. The current best estimate of historical sea-level rise in Venice, based on tide-gauge data after removal of subsidence effects, is 1.23 ± 0.13 mm/year (period from 1872 to 2019). Subsidence thus contributed to about half of the observed relative sea-level rise over the same period. A higher – yet more uncertain – rate of sea-level rise is observed during recent decades, estimated from tide-gauge data to be about 2.76 ± 1.75 mm/year in the period 1993–2019 for the climatic component alone. An unresolved issue is the contrast between the observational capacity of tide gauges and satellite altimetry, with the latter tool not covering the Venice Lagoon. Water mass exchanges through the Gibraltar Strait currently constitute a source of substantial uncertainty for estimating future deviations of the Mediterranean mean sea-level trend from the global-mean value. Subsidence and regional atmospheric and oceanic circulation mechanisms can deviate Venetian relative sea-level trends from the global mean values for several decades. Regional processes will likely continue to determine significant interannual and interdecadal variability of Venetian sea level with magnitude comparable to that observed in the past, as well as non-negligible differential trends. Our estimate of the likely range of mean sea-level rise in Venice by 2100 due to climate change is presently estimated between 11 and 110 centimetres. An improbable yet possible high-end scenario linked to strong ice-sheet melting yields about 170 centimetres of mean sea-level rise in Venice by 2100. Projections of natural and human induced vertical land motions are currently not available, but historical evidence demonstrates that they can produce a significant contribution to the relative sea-level rise in Venice, further increasing the hazard posed by climatically-induced sea-level changes.

A newly developed coastal model, FESOM-C, based on three-dimensional unstructured meshes and finite volume, is applied to simulate the dynamics of the southeastern North Sea. Variable horizontal resolution enables coarse meshes in the open sea with refined meshes in shallow areas including the Wadden Sea and estuaries to resolve important small-scale processes such as wetting and drying, sub-mesoscale eddies, and the dynamics of steep coastal fronts. Model results for a simulation of the period from January 2010 to December 2014 agree reasonably well with data from numerous regional autonomous observation stations with high temporal and spatial resolutions, as well as with data from FerryBoxes and glider expeditions. Analyzing numerical solution convergence on meshes of different horizontal resolutions allows us to identify areas where high mesh resolution (wetting and drying zones and shallow areas) and low mesh resolution (open boundary, open sea, and deep regions) are optimal for numerical simulations.

The study is dedicated to the tidal dynamics in the Sylt-Rømø Bight with a focus on the non-linear processes. The FESOM-C model was used as the numerical tool, which works with triangular, rectangular or mixed grids and is equipped with a wetting/drying option. As the model’s success at resolving currents largely depends on the quality of the bathymetric data, we have created a new bathymetric map for an area based on recent studies of Lister Deep, Lister Ley, and the Højer and Rømø Deep areas. This new bathymetric product made it feasible to work with high resolution grids (up to 2 m in the wetting/drying zone). As a result, we were able to study the tidal energy transformation and the role of higher harmonics in the domain in detail. The tidal ellipses, maximum tidally-induced velocities, energy fluxes and residual circulation maps were constructed and analysed for the entire bight. Additionally, tidal asymmetry maps were introduced and constructed. The full analysis was performed on two grids with different structures and showed a convergence of the results as well as fulfillment of the energy balance. The tidal residual circulation and asymmetric tidal cycles largely define the circulation pattern, transport and accumulation of sediment and the distribution of bedforms in the bight, therefore the results are necessary and useful benchmarks for further studies in the area, including baroclinic and sediment dynamics investigations.

Sea level change is an important indicator of global warming. In order to predict future sea level changes, it becomes more and more important to understand the complex contribution of different components (steric changes, melting of ice sheets and glaciers, hydrology,…) to the total sea level change on global and regional scales.To distinguish between steric and mass-related sea level changes, we consider (1) satellite altimetry, which measures total sea level change, and (2) Gravity Recovery and Climate Experiment (GRACE) and GRACE Follow-On (GRACE-FO) data to derive mass changes within the Earth system. The information of both methods can be combined in a joint inversion approach to derive the individual components of sea level change. However, there is a gap of 11 months between GRACE and GRACE-FO and there are numerous (bi-)monthly gaps since 2011, which means that there is no dedicated gravity field mission to derive ocean mass changes during these times.In this contribution, we use time-variable gravity field data from the Swarm satellite mission and Satellite Laser Ranging (SLR) as an additional source of information on mass changes within the inversion approach. Thus, we are able to derive inversion results even in times without GRACE(-FO), albeit at the expense of an inevitably lower spatial resolution. We compare results with GRACE(-FO) data to those without GRACE(-FO) data. Furthermore, we quantify the leverage of Swarm and SLR on the low-resolution mass signal at basin scale in the combined inversion approach during the GRACE(-FO) lifetime.

Ocean bottom pressure (OBP) variability serves as a proxy of ocean mass variability, the knowledge of which is needed in geophysical applications. The question of how well it can be modeled by the present general ocean circulation models on time scales in excess of one day is addressed here by comparing the simulated OBP variability with the observed one. To this end, a new multi-year data set is used, obtained with an array of bottom pressure gauges deployed deeply along a transect across the Southern Ocean. We present a brief description of OBP data and show large scale correlations over several thousand kilometres at all time scales red using daily and monthly averaged data. Annual and semi-annual cycles are weak. Close to the Agulhas Retroflection, signals of up to 30 cm equivalent water height (EWH) are detected. Further south, signals are mostly intermittent and noisy. It is shown that the models simulate consistent patterns of bottom pressure variability on monthly and longer scales except for areas with high mesoscale eddy activity, where high resolution is needed to capture the variability due to eddies. Furthermore, despite good agreement in the amplitude of variability, the in situ and simulated OBP show almost no correlation.

A newly developed coastal model FESOM-C based on three-dimensional unstructured meshes and finite volume is applied to simulate dynamics of the southeastern part of the North Sea. Variable horizontal resolution enables using meshes that are coarse in the open sea but refined in the shallow areas (which include the Wadden Sea and the estuaries) to resolve important small-scale process (such as wetting and drying, sub-mesoscales eddies and dynamics of steep coastal fronts). Model results for the simulation for the period between January 2010 and December 2014 agree reasonably well with data from numerous autonomous observation stations with high temporal and spatial resolution, located in the region, data from ferry boxes and glider expeditions. The analysis of numerical solution convergence on meshes with different horizontal resolutions allows identifying areas where high mesh resolution (wetting and drying zones, shallow areas) and low mesh resolution (open boundary, open sea, and deep regions) are optimal for numerical simulations.

The nonhydrostatic boundary problem for an arbitrary three-dimensional domain with a seamount is considered. The problem is integrated into curvilinear boundary-fitted coordinates on a nonuniform grid. In order to identify nonhydrostatic effects the grid is condensed on the slopes of the seamount preserving a coarse resolution in the rest of the domain, where the problem is solved in the hydrostatic approximation. Calculation results for the nonhydrostatic tidal dynamics and hydrology of the Strait of Messina in the area of a seamount are presented.

Non-hydrostatic boundary–value problem is considered in the 3-D domain representing a strait with an underwater sill. The equations are integrated in the boundary–fitted coordinates on irregular grid. For detection of non-hydrostatic effects the grid has increased resolution at sill slopes. The method is applied for simulation non-hydrostatic dynamics in the Strait of Messina.

The long-wave dynamics of the Lombok Strait, which is the most important link of the West Indonesian throughflow connecting the Pacific and Indian Ocean waters, was simulated and analyzed. A feature of the strait is its extremely complex relief, on which water transport creates a field of pronounced vertical velocities, which requires consideration of the nonhydrostatic component of pressure. The work presents a 3-D nonhydrostatic model in curvilinear coordinates, which is verified on a test problem. Particular attention is paid to the method of solving the 3-D elliptical solver for a nonhydrostatic problem in boundary-matched coordinates and a vertical &sigma, level. The difference in transport through the Lombok Strait is determined by the difference in atmospheric pressure over the Pacific and Indian Oceans. Based on the results of the global simulation, the role of these factors in terms of their variability is analyzed, and the value of nonhydrostatic pressure in the dynamics of the Lombok Strait is revealed and evaluated. The vertical dynamics of the Lombok Strait are considered in detail based on hydrostatic and nonhydrostatic approaches.

Monitoring changes in eukaryotic microbial communities is critical for understanding ecosystem dynamics, trophic interactions and the impacts of climate change. Long-term time series are an important tool for monitoring changes in ecological communities, but time series from a single location may not be representative of regional dynamics. In the German Bight, the Helgoland Roads time series is such a long-term series. Here, we consider the spatial dynamics of the eukaryotic microbes as an indicator of the representativeness of the Helgoland Roads site for the coastal German Bight, which is located in the North Sea. The eukaryotic microbial community in the German Bight was analysed at Helgoland Roads and two coastal stations (Cuxhaven and Wilhelmshaven) between March and October 2016 using metabarcoding. In addition, an oceanographical model was used to check for potential hydrological connectivity between the stations during the sampling period. Our results showed that the communities were different at the three stations. Helgoland was dominated by dinoflagellates, whereas the coastal stations had more diverse communities. Furthermore, differences were observed in the dinoflagellate and diatom communities between the three stations. Lagrangian particle tracking applied to the model results, showed limited connectivity between Helgoland and the coastal stations in 2016. The differences between Helgoland and the coastal stations were correlated with the different hydrological regimes and associated nutrient contents. Our observations suggest the presence of different eukaryotic microbial communities separated by complex hydrological conditions in the coastal German Bight.

We describe FESOM-C, the coastal branch of the Finite-volumE Sea ice -- Ocean Model (FESOM2), which shares with FESOM2 many numerical aspects, in particular, its finite-volume cell-vertex discretization. Its dynamical core differs by the implementation of time stepping, the use of terrain-following vertical coordinate and formulation for hybrid meshes composed of triangles and quads. The first two distinctions were critical for coding FESOM-C as an independent branch. The hybrid mesh capability improves numerical efficiency, since quadrilateral cells have fewer edges than triangular cells. They do not suffer from spurious inertial modes of the triangular cell-vertex discretization and need less dissipation. The hybrid mesh capability allows one to use quasi-quadrilateral unstructured meshes, with triangular cells included only to join quadrilateral patches of differt resolution or instead of strongly deformed quadrilateral cells. The description of the model numerical part is complemented by test cases illustrating the model performance.

In this article, the tsunami model TsunAWI (Alfred Wegener Institute) and its application for hindcasts, inundation studies, and the operation of the tsunami scenario repository for the Indonesian tsunami early warning system are presented. TsunAWI was developed in the framework of the German-Indonesian Tsunami Early Warning System (GITEWS) and simulates all stages of a tsunami from the origin and the propagation in the ocean to the arrival at the coast and the inundation on land. It solves the non-linear shallow water equations on an unstructured finite element grid that allows to change the resolution seamlessly between a coarse grid in the deep ocean and a fine representation of coastal structures. During the GITEWS project and the following maintenance phase, TsunAWI and a framework of pre- and postprocessing routines was developed step by step to provide fast computation of enhanced model physics and to deliver high quality results.

Characterization of interior ocean variability is necessary for understanding climate. Water mass evolution shapes ocean-atmosphere interactions and contributes to determine timescales for global and regional climate variability. However, a robust assessment of past state and variability of the ocean interior is prevented by sparseness/shortness of historical subsurface observations and uncertainties affecting proxy-based reconstructions. Here, we propose a novel approach to infer past large-scale interior ocean variability with unprecedented accuracy and temporal resolution. It exploits links between stratification determined by “large-scale” water mass distributions and local dynamics. We characterize interannual interior ocean variability in the Mediterranean Sea in the early 20th century contained in tidal measurements in the Strait of Messina, and demonstrate the general applicability of our method, paving the way to a new approach to analyze historical oceanographic records: Regions where different water masses are known to collide can thus act as magnifying glasses for basin-scale interior ocean variability, hence providing “liquid archives” for climatology.

The Lena plume dynamics in the Lena Delta region of the Laptev Sea are explored in simulations performed with the Finite Volume Coastal Ocean Model (FVCOM) on a mesh with the horizontal resolution 0.4–5 km. The impact of wind and tides on the Lena plume propagation is analysed based on simulations for the summer season of 2008 and also on idealised experiments. All main Lena River freshwater channels (Trofimovskaya, Bykovskaya, Tumatskaya and Olenekskaya) produce buoyant outflows in the summer season. The surface plume buoyancy signature proves to be highly variable in time, especially in case of upwelling favourable wind events. Winds stronger than 6 m s−1 can already turn the dynamics of flows from all main freshwater channels to the wind-driven state. During the summer season, the bulk of freshwater from the Lena River stays in the eastern Laptev Sea because of location of the main Lena River freshwater channels, their large Kelvin numbers and light summer winds. Westward and northward plume excursions are wind-driven, and the model skill in simulating them depends on the available wind forcing. The main mechanism of tidal influence in the freshwater plume zone is through tidally induced mixing, except for the northern vicinity of the delta, where residual circulation may contribute to the plume eastward transport significantly.

General ocean circulation models are not perfect. Forced with observed atmospheric fluxes they gradually drift away from measured distributions of temperature and salinity. We suggest data assimilation of absolute dynamical ocean topography (DOT) observed from space geodetic missions as an option to reduce these differences. Sea surface information of DOT is transferred into the deep ocean by defining the analysed ocean state as a weighted average of an ensemble of fully consistent model solutions using an error-subspace ensemble Kalman filter technique. Success of the technique is demonstrated by assimilation into a global configuration of the ocean circulation model FESOM over 1 year. The dynamic ocean topography data are obtained from a combination of multi-satellite altimetry and geoid measurements. The assimilation result is assessed using independent temperature and salinity analysis derived from profiling buoys of the AGRO float data set. The largest impact of the assimilation occurs at the first few analysis steps where both the model ocean topography and the steric height (i.e. temperature and salinity) are improved. The continued data assimilation over 1 year further improves the model state gradually. Deep ocean fields quickly adjust in a sustained manner: A model forecast initialized from the model state estimated by the data assimilation after only 1 month shows that improvements induced by the data assimilation remain in the model state for a long time. Even after 11 months, the modelled ocean topography and temperature fields show smaller errors than the model forecast without any data assimilation.

Finite-volume discretizations can be formulated on unstructured meshes composed of different polygons. A staggered cell-vertex finite-volume discretization of shallow water equations is analyzed on mixed meshes composed of triangles and quads. Although triangular meshes are most flexible geometrically, quads are more efficient numerically and do not support spurious inertial modes of triangular cell- vertex discretization. Mixed meshes composed of triangles and quads combine benefits of both. In particular, triangular transitional zones can be used to join quadrilateral meshes of differing resolution. Based on a set of examples involving shallow water equations, it is shown that mixed meshes offer a viable approach provided some background biharmonic viscosity (or the biharmonic filter) is added to stabilize the triangular part of the mesh.