Extreme storm surge modelling in the North Sea

This study shows that storm surge model performance in the North Sea is mostly unaffected by the application of temporal variations of surface drag due to changes in sea state provided the choice of a suitable constant Charnock parameter in the sea-state-independent case. Including essential meteorological features on smaller scales and minimising interpolation errors by increasing forcing data resolution are shown to be more important for the improvement of model performance particularly at the high tail of the probability distribution. This is found in a modelling study using WAQUA/DCSMv5 by evaluating the influence of a realistic air-sea momentum transfer parameterization and comparing it to the influence of changes in the spatial and temporal resolution of the applied forcing fields in an effort to support the improvement of impact and climate analysis studies. Particular attention is given to the representation of extreme water levels over the past decades based on the example of the Netherlands. For this, WAQUA/DCSMv5 is forced with ERA-Interim reanalysis data. Model results are obtained from a set of different forcing fields, which either (i) include a wave-state-dependent Charnock parameter or (ii) apply a constant Charnock parameter (α _{C h} =?0.032) tuned for young sea states in the North Sea, but differ in their spatial and/or temporal resolution. Increasing forcing field resolution from roughly 79 to 12 km through dynamically downscaling can reduce the modelled low bias, depending on coastal station, by up to 0.25 m for the modelled extreme water levels with a 1-year return period and between 0.1 m and 0.5 m for extreme surge heights.     

Adjoint free four-dimensional variational data assimilation for a storm surge model of the German North Sea

In this paper, a data assimilation scheme based on the adjoint free Four-Dimensional Variational(4DVar) method is applied to an existing storm surge model of the German North Sea. To avoid the need of an adjoint model, an ensemble-like method to explicitly represent the linear tangent equation is adopted. Results of twin experiments have shown that the method is able to recover the contaminated low dimension model parameters to their true values. The data assimilation scheme was applied to a severe storm surge event which occurred in the North Sea in December 5, 2013. By adjusting wind drag coefficient, the predictive ability of the model increased significantly. Preliminary experiments have shown that an increase in the predictive ability is attained by narrowing the data assimilation time window.     

Climate change and North Sea storm surge extremes: an ensemble study of storm surge extremes expected in a changed climate projected by four different regional climate models

The coastal zones are facing the prospect of changing storm surge statistics due to anthropogenic climate change. In the present study, we examine these prospects for the North Sea based on numerical modelling. The main tool is the barotropic tide-surge model TRIMGEO (Tidal Residual and Intertidal Mudflat Model) to derive storm surge climate and extremes from atmospheric conditions. The analysis is carried out by using an ensemble of four 30-year atmospheric regional simulations under present-day and possible future-enhanced greenhouse gas conditions. The atmospheric regional simulations were prepared within the EU project PRUDENCE (Prediction of Regional scenarios and Uncertainties for Defining EuropeaN Climate change risks and Effects). The research strategy of PRUDENCE is to compare simulations of different regional models driven by the same global control and climate change simulations. These global conditions, representative for 1961–1990 and 2071–2100 were prepared by the Hadley Center based on the IPCC A2 SRES scenario. The results suggest that under future climatic conditions, storm surge extremes may increase along the North Sea coast towards the end of this century. Based on a comparison between the results of the different ensemble members as well as on the variability estimated from a high-resolution storm surge reconstruction of the recent decades it is found that this increase is significantly different from zero at the 95% confidence level for most of the North Sea coast. An exception represents the East coast of the UK which is not affected by this increase of storm surge extremes.     

Results of a 36-hour storm surge prediction of the North Sea for 3 January 1976 on the basis of numerical models

Summary To explore the feasibility of forescasting North Sea storm surges by integrating numerically a combined atmospheric-oceanographic physical model, the severe storm and the resulting water levels occurring on 3 January 1976 were simulated as a first step into this direction. For this purpose, the atmospheric model was run with a resolution of 8 levels in the vertical and a horizontal grid spacing of 1.4° in latitude and 2.8° in longitude on the hemisphere. The initial conditions are based upon observations of 2 January 1976, 12 GMT, i.e. about 24 hours before the storm reached its greatest intensity in the southern parts of the North Sea.The surface wind predicted by the atmospheric model was converted into stress values through a bulk formula which then entered the vertically integrated North Sea model, to yield the desired water elevations in a 22 km-grid. Moreover, also the observed wind, stemming from a careful re-analysis of the storm situation, was handled in the same way. The numerically obtained results were compared with gauge measurements at a number of coastal stations.

---

Ergebnisse einer 36stündigen Vorhersage der Nordsee-Sturmflut am 3. Januar 1976 aufgrund von numerischen Modellen

Zusammenfassung Zur Untersuchung der Möglichkeiten, Sturmfluten mit physikalischen Modellen der Atmosphäre und der Nordsee vorherzusagen, wurde als erster Schritt in dieser Richtung die schwere Sturmflut an der deutschen Nordseeküste vom 3. Januar 1976 nachgerechnet. Hierzu wurde das im Sonderforschungsbereich 94 entwickelte Atmosphärenmodell auf der Grundlage der hydro- und thermodynamischen Gleichungen mit einer Auflösung von 8 Flächen in der Vertikalen sowie 1,4° in der geographischen Breite und 2,8° in der geographischen Länge auf der gesamten Nordhalbkugel über 36 Stunden numerisch integriert. Als Eingangsdaten dienten die Beobachtungen vom 2. Januar 1976, 12 Uhr. dieser Termin liegt ca. 24 Stunden vor Eintritt des stärksten Sturmes im Bereich der Deutschen Bucht.Die auf diese Weise prognostizierte Windverteilung wurde über der Nordsee in Schubspannungswerte überführt, woraus das Nordseemodell mit den hydrodynamischen Gleichungen die Wasserstandswerte in einem 22-km-Netz berechnete. In entsprechender Weise wurde mit einem aus Beobachtungen während dieses Zeitraums gewonnenen Windfeld verfahren; diesem lag eine besonders sorgfältige, nachträglich angefertigte Analyse des Seewetteramts Hamburg zugrunde. Die so errechneten Wasserstände sind dann mit Pegelmessungen verglichen worden.

---

Résultats d'une prédiction d'onde de tempête de 36 heures de la mer du Nord pour le 3 janvier 1976 sur la base de modèles numériques

Résumé Afin d'étudier la possibilité de prévior les ondes de tempête en mer du Nord en intégrant numériquement un modèle physique combiné atmosphère-océan, la forte tempête du 3 janvier 1976 et des niveaux d'eau qui en ont résulté ont été simulés comme premier pas dans cette voie. Dans ce but, on a fait fonctionner le modèle atmosphérique avec une résolution de 8 niveaux dans le plan vertical et un réseau horizontal aux pas de 1,4° en latitude et de 2,8° en longitude sur l'hémisphère. Les conditions initiales sont basées sur les observations du 2 janvier 1976 à 12h GMT, c'est-à-dire environ 24 heures avant que la tempête n'atteignît sa plus forte intensité dans les parties méridionales de la mer du Nord.Le vent en surface prédit par le modèle atmosphérique était converti par l'intermédiaire d'une formule globale en valeurs de tensions qui étaient alors introduites dans le modèle de la mer du Nord intégré verticalement, pour obtenir les hauteurs d'eau désirées, dans un réseau de 22 km de côté. En outre, le vent observé, issu d'une nouvelle analyse soignée de la situation de tempête, était aussi traité de la même manière. Les résultats obtenus numériquement furent comparés aux mesures fournies par les marégraphes en un certain nombre de stations côtières.

     

Source term balance in a severe storm in the Southern North Sea

This paper presents the results of a wave hindcast of a severe storm in the Southern North Sea to verify recently developed deep and shallow water source terms. The work was carried out in the framework of the ONR funded NOPP project (Tolman et al. 2013) in which deep and shallow water source terms were developed for use in third-generation wave prediction models. These deep water source terms for whitecapping, wind input and nonlinear interactions were developed, implemented and tested primarily in the WAVEWATCH III model, whereas shallow water source terms for depth-limited wave breaking and triad interactions were developed, implemented and tested primarily in the SWAN wave model. So far, the new deep-water source terms for whitecapping were not fully tested in shallow environments. Similarly, the shallow water source terms were not yet tested in large inter-mediate depth areas like the North Sea. As a first step in assessing the performance of these newly developed source terms, the source term balance and the effect of different physical settings on the prediction of wave heights and wave periods in the relatively shallow North Sea was analysed. The December 2013 storm was hindcast with a SWAN model implementation for the North Sea. Spectral wave boundary conditions were obtained from an Atlantic Ocean WAVEWATCH III model implementation and the model was driven by hourly CFSR wind fields. In the southern part of the North Sea, current and water level effects were included. The hindcast was performed with five different settings for whitecapping, viz. three Komen type whitecapping formulations, the saturation-based whitecapping by Van der Westhuysen et al. (2007) and the recently developed ST6 whitecapping as described by Zieger et al. (2015). Results of the wave hindcast were compared with buoy measurements at location K13 collected by the Dutch Ministry of Transport and Public Works. An analysis was made of the source term balance at three locations, the deep water location North Cormorant, the inter-mediate depth location K13 and at location Wielingen, a shallow water location close to the Dutch coast. The results indicate that at deep water the source terms for wind input, whitecapping and nonlinear four-wave interactions are of the same magnitude. At the inter-mediate depth location K13, bottom friction plays a significant role, whereas at the shallow water location Wielingen also depth-limited wave breaking becomes important.     

Application of an unstructured mesh model to storm surge propagation in the Mersey estuary region of the Irish Sea

The storm surge period of 13–16 November 1977 when there was a major positive surge followed by a negative surge in the Irish Sea is investigated using a two-dimensional unstructured mesh model of the west coast of Britain. The model accounts for tidal and external surge forcing across its open boundaries which are situated in the Celtic Sea and off the west coast of Scotland. Although this period has been examined previously using a uniform finite-difference model, and a finite element model, neither of these could resolve the Mersey estuary which is the focus of the present study. By using a finite element model with very high mesh resolution within the Mersey, the spatial variability of surge elevations and currents within the Mersey to rapidly changing surge dynamics can be examined. The mesh in the model varies from about 7 km in deep water, to the order of 100 m in the Mersey, with the largest mesh length reaching 17 km in deep offshore regions, and smallest of order 26 m occurring in shallow coastal regions of the Mersey estuary. The model accounts for wetting/drying which occurs in shallow water coastal areas. Calculations showed that during the positive surge period, the amplitude and speed of propagation of the surge was largest in the deep water channels. This gave rise to significant spatial variability of surge elevations and currents within the estuary. As wind stresses decreased over the Irish Sea, a negative surge occurred over Liverpool Bay and at the entrance to the Mersey. However, within the Mersey there was a local positive surge which continued to propagate down the estuary. This clearly showed that although the large scale response of the Irish Sea to changing wind fields occurred rapidly, the response in the Mersey was much slower. These calculations with a west coast variable mesh model that included a high-resolution representation of the Mersey revealed for the first time how elevations and currents within the Mersey responded to Irish Sea surges that rapidly changed from positive to negative.     

Identification of extreme storm surges with high-impact potential along the German North Sea coastline

Planning and design of coastal protection rely on information about the probabilities of very severe storm tides and the possible changes that may occur in the course of climate change. So far, this information is mostly provided in the form of high percentiles obtained from frequency distributions or return values. More detailed information and assessments of events that may cause extreme damages or have extreme consequences at the coast are so far still unavailable. We describe and compare two different approaches that may be used to identify highly unlikely but still physically possible and plausible events from model simulations. Firstly, in the case when consistent wind and tide-surge data are available, different metrics such as the height of the storm surge can be derived directly from the simulated water levels. Secondly, in cases where only atmospheric data are available, the so called effective wind may be used. The latter is the projection of the horizontal wind vector on that direction which is most effective in producing surges at the coast. Comparison of events identified by both methods show that they can identify extreme events but that knowledge of the effective wind alone does not provide sufficient information to identify the highest storm surges. Tracks of the low-pressure systems over the North Sea need to be investigated to find those cases, where the duration of the high wind is too short to induce extreme storm tides. On the other hand, factors such as external surges or variability in mean sea level may enhance surge heights and are not accounted for in estimates based on effective winds only. Results from the analysis of an extended data set suggest that unprecedented storm surges at the German North Sea coast are possible even without taking effects from rising mean sea level into account. The work presented is part of the ongoing project “Extreme North Sea Storm Surges and Their Consequences” (EXTREMENESS) and represents the first step towards an impact assessment for very severe storm surges which will serve as a basis for development of adaptation options and evaluation criteria.     

Some environmental impacts of a storm surge barrier

In 1976 the Dutch government decided to close the Eastern Scheldt estuary with a storm surge barrier instead of a dam as was previously envisaged. In this way safety is bought and the excellent Eastern Scheldt ecological system is preserved. The cost of this compromise is estimated at 8.7 billion Dutch guilders. However what does this compromise mean for the Eastern Scheldt ecosystem?     

The statistics of wave height and crest elevation during the December 2012 storm in the North Sea

This paper concerns new field measurements of wave height and crest elevation probability distributions as measured in the North Sea during a storm in December 2012. The water surface elevation was recorded by Saab WaveRadar REX instruments mounted on eight fixed-jacket platforms in addition to a Datawell Directional Waverider buoy. The storm generated an easterly sea state which peaked well in excess of the 100-year wave height for that direction in the region. Furthermore, 19 freak waves occurred during the storm according to the definition as reported by Haver (2000). The present study demonstrates that the significant steepness and spectral bandwidth during the storm remain almost constant. Consequently, there is little change in the commonly applied design wave height and crest elevation probability distributions throughout the storm. Whilst the bulk of the recorded data was in good agreement with the theoretical distributions, it was demonstrated that when the wind speed was larger than 25 m/s, the measured crest elevation lies above the second-order Forristall distribution.     

Spectral analysis of storm surge in Hong Kong Victoria Harbour

Based on a linear model the dynamic characteristics of Victoria Harbour (Hong Kong) is obtained by means of spectral analysis of the storm surge hydrographs. The results show that the harbour is an ideal one which has a small gain factor and a flat response in the frequency range from 0 to 6 × 10⁻⁵ Hz. The results also show that the power spectra possess the narrow band features which indicates that the periodic components associated with tidal motions are predominant over the random components. The power spectrum corresponding to a frequency of 2.3 × 10⁻⁵ Hz is likely to be associated with the astronomical tides. The peaks in the power spectra at zero frequency suggest that the pumping mode of oscillations is dominant in a storm surge. This mode of oscillations represents the temporal variations in mean sea level. To demonstrate the full potential of the present model, more case studies should be conducted when surge as well as non-surge data are available.     

Discontinuous Galerkin methods for modeling Hurricane storm surge

Storm surge due to hurricanes and tropical storms can result in significant loss of life, property damage, and long-term damage to coastal ecosystems and landscapes. Computer modeling of storm surge can be used for two primary purposes: forecasting of surge as storms approach land for emergency planning and evacuation of coastal populations, and hindcasting of storms for determining risk, development of mitigation strategies, coastal restoration and sustainability.Storm surge is modeled using the shallow water equations, coupled with wind forcing and in some events, models of wave energy. In this paper, we will describe a depth-averaged (2D) model of circulation in spherical coordinates. Tides, riverine forcing, atmospheric pressure, bottom friction, the Coriolis effect and wind stress are all important for characterizing the inundation due to surge. The problem is inherently multi-scale, both in space and time. To model these problems accurately requires significant investments in acquiring high-fidelity input (bathymetry, bottom friction characteristics, land cover data, river flow rates, levees, raised roads and railways, etc.), accurate discretization of the computational domain using unstructured finite element meshes, and numerical methods capable of capturing highly advective flows, wetting and drying, and multi-scale features of the solution.The discontinuous Galerkin (DG) method appears to allow for many of the features necessary to accurately capture storm surge physics. The DG method was developed for modeling shocks and advection-dominated flows on unstructured finite element meshes. It easily allows for adaptivity in both mesh (h) and polynomial order (p) for capturing multi-scale spatial events. Mass conservative wetting and drying algorithms can be formulated within the DG method.In this paper, we will describe the application of the DG method to hurricane storm surge. We discuss the general formulation, and new features which have been added to the model to better capture surge in complex coastal environments. These features include modifications to the method to handle spherical coordinates and maintain still flows, improvements in the stability post-processing (i.e. slope-limiting), and the modeling of internal barriers for capturing overtopping of levees and other structures. We will focus on applications of the model to recent events in the Gulf of Mexico, including Hurricane Ike.     

Evaluation of weather forecast systems for storm surge modeling in the Chesapeake Bay

Accurate forecast of sea-level heights in coastal areas depends, among other factors, upon a reliable coupling of a meteorological forecast system to a hydrodynamic and wave system. This study evaluates the predictive skills of the coupled circulation and wind-wave model system (ADCIRC+SWAN) for simulating storm tides in the Chesapeake Bay, forced by six different products: (1) Global Forecast System (GFS), (2) Climate Forecast System (CFS) version 2, (3) North American Mesoscale Forecast System (NAM), (4) Rapid Refresh (RAP), (5) European Center for Medium-Range Weather Forecasts (ECMWF), and (6) the Atlantic hurricane database (HURDAT2). This evaluation is based on the hindcasting of four events: Irene (2011), Sandy (2012), Joaquin (2015), and Jonas (2016). By comparing the simulated water levels to observations at 13 monitoring stations, we have found that the ADCIR+SWAN System forced by the following: (1) the HURDAT2-based system exhibited the weakest statistical skills owing to a noteworthy overprediction of the simulated wind speed; (2) the ECMWF, RAP, and NAM products captured the moment of the peak and moderately its magnitude during all storms, with a correlation coefficient ranging between 0.98 and 0.77; (3) the CFS system exhibited the worst averaged root-mean-square difference (excepting HURDAT2); (4) the GFS system (the lowest horizontal resolution product tested) resulted in a clear underprediction of the maximum water elevation. Overall, the simulations forced by NAM and ECMWF systems induced the most accurate results best accuracy to support water level forecasting in the Chesapeake Bay during both tropical and extra-tropical storms.     

Observation impact analysis methods for storm surge forecasting systems

This paper presents a simple method for estimating the impact of assimilating individual or group of observations on forecast accuracy improvement. This method is derived from the nsemble-based observation impact analysis method of Liu and Kalnay (Q J R Meteorol Soc 134:1327–1335, 2008). The method described here is different in two ways from their method. Firstly, it uses a quadratic function of model-minus-observation residuals as a measure of forecast accuracy, instead of model-minus-analysis. Secondly, it simply makes use of time series of observations and the corresponding model output generated without data assimilation. These time series are usually available in an operational database. Hence, it is simple to implement. It can be used before any data assimilation is implemented. Therefore, it is useful as a design tool of a data assimilation system, namely for selecting which observations to assimilate. The method can also be used as a diagnostic tool, for example, to assess if all observation contributes positively to the accuracy improvement. The method is applicable for systems with stationary error process and fixed observing network. Using twin experiments with a simple one-dimensional advection model, the method is shown to work perfectly in an idealized situation. The method is used to evaluate the observation impact in the operational storm surge forecasting system based on the Dutch Continental Shelf Model version 5 (DCSMv5).     

Development and evaluation of a storm surge warning system in Taiwan

An operational storm surge forecasting system aimed at providing warning information for storm surges has been developed and evaluated using four typhoon events. The warning system triggered by typhoon forecasts from Taiwan Cooperative Precipitation Ensemble Forecast Experiment (TAPEX) has been executed with two storm surge forecasting scenarios with and without tides. Three numerical experiments applying different meteorological inputs have been designed to assess the impact of typhoon forcing on storm surges. One uses synthetic wind fields, and the others use realistic wind fields with and without adjustments to the initial wind fields for the background circulation. Local observations from Central Weather Bureau (CWB) weather stations and tide gauge stations are used to evaluate the wind fields and storm surges from our numerical experiments. The comparison results show that the accuracy of the storm surge forecast is dominated by the track, the intensity, and the driving flow of a typhoon. When the structure of a typhoon is disturbed by Taiwan’s topography, using meteorological inputs from real wind fields can result in a better typhoon simulation than using inputs from synthetic wind fields. The driving flow also determines the impact of topography on typhoon movement. For quickly moving typhoons, storm forcing from TAPEX is reliable when a typhoon is strong enough to be relatively unaffected by environmental flows; otherwise, storm forcing from a sophisticated typhoon initialization scheme that better simulates the typhoon and environmental flows results in a more accurate prediction of storm surges. Therefore, when a typhoon moves slowly and interacts more with the topography and environmental flows, incorporating realistic wind fields with adjustments to the initial wind fields for the background circulation in the warning system will obtain better predictions for a typhoon and its resultant storm surges.     

Investigating the effects of a summer storm on the North Sea stratification using a regional coupled ocean-atmosphere model

The influence of a summer storm event in 2007 on the North Sea and its effects on the ocean stratification are investigated using a regional coupled ocean (Regional Ocean Modeling System, ROMS)-atmosphere (Weather Research & Forecasting model, WRF) modeling system. An analysis of potential energy anomaly (PEA, Φ) and its temporal development reveals that the loss of stratification due to the storm event is dominated by vertical mixing in almost the entire North Sea. For specific regions, however, a considerable contribution of depth-mean straining is observed. Vertical mixing is highly correlated with wind induced surface stresses. However, peak mixing values are observed in combination with incoming flood currents. Depending on the phase between winds and tides, the loss of stratification differs strongly over the North Sea. To study the effects of interactive ocean-atmosphere exchange, a fully coupled simulation is compared with two uncoupled ones for the same vertical mixing parameters to identify the impact of spatial resolution as well as of SST feedback. While the resulting new mixed layer depth after the storm event in the uncoupled simulation with lower spatial and temporal resolution of the surface forcing data can still be located in the euphotic zone, the coupled simulation is capable to mix the entire water column and the vertical mixing in the uncoupled simulation with higher resolution of the surface forcing data is strongly amplified. These differences might have notable implications for ecosystem modeling since it could determine the development of new phytoplankton blooms after the storm and for sediment modeling in terms of sediment mobilization. An investigation of restratification after the extreme event illustrates the persistent effect of this summer storm.