Assessing the utility of frequency dependent nudging for reducing biases in biogeochemical models

Bias errors, resulting from inaccurate boundary and forcing conditions, incorrect model parameterization, etc. are a common problem in environmental models including biogeochemical ocean models. While it is important to correct bias errors wherever possible, it is unlikely that any environmental model will ever be entirely free of such errors. Hence, methods for bias reduction are necessary. A widely used technique for online bias reduction is nudging, where simulated fields are continuously forced toward observations or a climatology. Nudging is robust and easy to implement, but suppresses high-frequency variability and introduces artificial phase shifts. As a solution to this problem Thompson et al. (2006) introduced frequency dependent nudging where nudging occurs only in prescribed frequency bands, typically centered on the mean and the annual cycle. They showed this method to be effective for eddy resolving ocean circulation models. Here we add a stability term to the previous form of frequency dependent nudging which makes the method more robust for non-linear biological models. Then we assess the utility of frequency dependent nudging for biological models by first applying the method to a simple predator–prey model and then to a 1D ocean biogeochemical model. In both cases we only nudge in two frequency bands centered on the mean and the annual cycle, and then assess how well the variability in higher frequency bands is recovered. We evaluate the effectiveness of frequency dependent nudging in comparison to conventional nudging and find significant improvements with the former.     

Improvement of morphodynamic modeling of tidal channel migration by nudging

State-of-the-art process-based models have shown to be applicable to the simulation and prediction of coastal morphodynamics. On annual to decadal temporal scales, these models may show limitations in reproducing complex natural morphological evolution patterns, such as the movement of bars and tidal channels, e.g. the observed decadal migration of the Medem Channel in the Elbe Estuary, German Bight. Here a morphodynamic model is shown to simulate the hydrodynamics and sediment budgets of the domain to some extent, but fails to adequately reproduce the pronounced channel migration, due to the insufficient implementation of bank erosion processes. In order to allow for long-term simulations of the domain, a nudging method has been introduced to update the model-predicted bathymetries with observations. The model-predicted bathymetry is nudged towards true states in annual time steps. Sensitivity analysis of a user-defined correlation length scale, for the definition of the background error covariance matrix during the nudging procedure, suggests that the optimal error correlation length is similar to the grid cell size, here 80–90 m. Additionally, spatially heterogeneous correlation lengths produce more realistic channel depths than do spatially homogeneous correlation lengths. Consecutive application of the nudging method compensates for the (stand-alone) model prediction errors and corrects the channel migration pattern, with a Brier skill score of 0.78. The proposed nudging method in this study serves as an analytical approach to update model predictions towards a predefined ‘true’ state for the spatiotemporal interpolation of incomplete morphological data in long-term simulations.     

Refined seaward boundary conditions for modelling shallow water waves in semi-enclosed water bodies

Based on the theory of characteristics, this research elaborates on the numerical treatment of two types of seaward boundary conditions for modelling long-wave dynamics in truncated estuarine and coastal domains. These seaward boundary conditions are devised for the solution of the fully non-linear shallow water equations in the time domain. The first type is the clamped boundary, at which the water level variation is given and the velocity is computed along the characteristic line going out of the domain. The second type is the non-reflecting boundary, where the incident wave information is introduced and the reflected waves from inside the computational domain are allowed to escape at the same time. The essence of its numerical implementation is to distinguish the inward and outward characteristics and to disconnect the incoming characteristic relation from the actual flow inside the domain. Compared with previous techniques, the present method includes extra terms in the derivation to account for the effects of the uneven bed, bottom friction and shape of the characteristic lines. A shock-capturing finite difference method is used to solve the shallow water equations in the deviatoric format, but the seaward boundary algorithms constructed herein are generic and applicable to other solvers. The necessity of these refinements is highlighted by simulating the tidal oscillation in the Persian/Arabian Gulf, periodic wave runup on the coastline and the wave resonance in a narrow harbour. It is found that neglecting the bed slope at the boundary may result in biased mean water levels in the prediction.     

Antarctic Intermediate Water variability in the northern New Zealand region

In the region between 30°S and northern New Zealand, vertical salinity profiles through the core of the Antarctic Intermediate Water (AAIW) show a high degree of spatial and temporal variability, and this variability is much larger than that found in nearby ocean areas. Characteristic features are interleaving of salinity layers and large changes in the salinity minimum between adjacent stations. Quantifying the changes through the calculation of an intrusion index highlights the degree of variability and the importance of boundary mixing along the New Zealand slope. However, the main cause of the variability is the meeting and mixing of higher salinity AAIW, arriving from the north‐west (having travelled around the subtropical gyre) with lower salinity AAIW arriving by more direct entry from the north‐east. These waters meet in the region through the action of the meso‐scale eddy field. Present data indicate that where strong salinity interleaving occurs, the length scales are of the order of 10 km and the time scales are of the order of a few days. Resolution of the processes at work will require studies on finer scales than presently available.     

A numerical model for wave propagation in curvilinear coordinates

Using the perturbation method, a time dependent parabolic equation is developed based on the elliptic mild slope equation with dissipation term. With the time dependent parabolic equation employed as the governing equation, a numerical model for wave propagation including dissipation term in water of slowly varying topography is presented in curvilinear coordinates. In the model, the self-adaptive grid generation method is employed to generate a boundary-fitted and varying spacing mesh. The numerical tests show that the effects of dissipation term should be taken into account if the distance of wave propagation is large, and that the outgoing boundary conditions can be treated more effectively by introduction of the dissipation term into the numerical model. The numerical model is able to give good results of simulating wave propagation for waters of complicatedly boundaries and effectively predict physical processes of wave propagation. Moreover, the errors of the analytical solution deduced by Kirby et al. (1994) [Kirby, J.T., Dalrymple, R.A., Kabu, H., 1994. Parabolic approximation for water waves in conformal coordinate systems. Coastal Engineering 23, 185–213.] from the small-angle parabolic approximation of the mild-slope equation for the case of waves between diverging breakwaters in a polar coordinate system are corrected.     

The morphological response of a nearshore double sandbar system to constant wave forcing

Results of 2DH morphodynamic computations are presented to quantify the temporal evolution of the crescentic patterns emerging in a double nearshore bar system in response to constant wave boundary forcing. Sixteen different conditions varying both offshore wave height and angle of wave incidence were applied. The mean length scales of the emerging irregular crescentic patterns are linearly proportional to the local longshore velocity over the inner and outer bars. For similar longshore velocities, the length scales of the outer bar are larger than of the inner bar. This is explained by accounting for the difference in water depth above the bar crest. The variable morphological response times can be explained by including additional bathymetrical parameters. The active volume of the bar, defined by the breaker index, plays an important role in this response time. With larger active volumes the bar responds more rapidly to identical boundary conditions. Also, bars with a smaller total volume respond more quickly. This faster response is due to the steeper active volume of the bars. Different initial perturbations resulted in different locations of the emerging features, showing that their location is sensitive to the initial bathymetry. However, the range in length scales and response times due to the different perturbations was significantly smaller than those obtained for the different hydrodynamic conditions. Based on the present findings we hypothesize that morphological length scales in the field are rarely in equilibrium with the concurrent offshore wave height and angle of incidence owing to the slow response of the sandbars under constant conditions relative to the stochastic nature of natural wave forcing.     

Improvement on open boundaries on a time dependent numerical model of wave propagation in the nearshore region

An improvement on the simulation of outgoing waves on a time dependent numerical model for water wave propagation in the nearshore region is presented. The governing equations consist of a system of first order partial differential equations (PDEs), the equation of continuity and the equation of motion. A comparative study of first order radiation boundary conditions (BCs) and first order radiation BCs combined with sponge layers is presented for cases where outgoing waves leave the numerical domain of calculation through the open boundary. A reduction of spurious reflections from the numerical open boundaries can be obtained with an irrelevant increase in terms of computational cost.     

Stochastic simulations of ocean waves: An uncertainty quantification study

The primary objective of this study is to introduce a stochastic framework based on generalized polynomial chaos (gPC) for uncertainty quantification in numerical ocean wave simulations. The techniques we present can be easily extended to other numerical ocean simulation applications. We perform stochastic simulations using a relatively new numerical method to simulate the HISWA (Hindcasting Shallow Water Waves) laboratory experiment for directional near-shore wave propagation and induced currents in a shallow-water wave basin. We solve the phased-averaged equation with hybrid discretization based on discontinuous Galerkin projections, spectral elements, and Fourier expansions. We first validate the deterministic solver by comparing our simulation results against the HISWA experimental data as well as against the numerical model SWAN (Simulating Waves Nearshore). We then perform sensitivity analysis to assess the effects of the parametrized source terms, current field, and boundary conditions. We employ an efficient sparse-grid stochastic collocation method that can treat many uncertain parameters simultaneously. We find that the depth-induced wave-breaking coefficient is the most important parameter compared to other tunable parameters in the source terms. The current field is modeled as random process with large variation but it does not seem to have a significant effect. Uncertainty in the source terms does not influence significantly the region before the submerged breaker whereas uncertainty in the incoming boundary conditions does. Considering simultaneously the uncertainties from the source terms and boundary conditions, we obtain numerical error bars that contain almost all experimental data, hence identifying the proper range of parameters in the action balance equation.     

Nudging of vertical profiles of meteorological parameters in one-dimensional atmospheric model: A step towards improvements in numerical simulations

In this article, we describe a simple yet effective method for insertion of observational datasets in a mesoscale atmospheric model used in one-dimensional configuration through Nudging. To demonstrate the effectiveness of this technique, vertical profiles of meteorological parameters obtained from GLASS Sonde launches from a tiny island of Kaashidhoo in the Republic of Maldives are injected in a mesoscale atmospheric model — Advanced Regional Prediction System (ARPS), and model simulated parameters are compared with the available observational datasets. Analysis of one-time nudging in the model simulations over Kaashidhoo show that incorporation of this technique reasonably improves the model simulations within a time domain of +6 to +12 Hrs, while its impact on +18 Hrs simulations and beyond becomes literally null.     

植被斜坡岸滩海啸波消减数值模拟研究

An explicit one-dimensional model based on the shallow water equations(SWEs) was established in this work to simulate tsunami wave propagation on a vegetated beach. This model adopted the finite-volume method(FVM)for maintaining the mass balance of these equations. The resistance force caused by vegetation was taken into account as a source term in the momentum equation. The Harten–Lax–van Leer(HLL) approximate Riemann solver was applied to evaluate the interface fluxes for tracing the wet/dry transition boundary. This proposed model was used to simulate solitary wave run-up and long-periodic wave propagation on a sloping beach. The calibration process suitably compared the calculated results with the measured data. The tsunami waves were also simulated to discuss the water depth, tsunami force, as well as the current speed in absence of and in presence of forest domain. The results indicated that forest growth at the beach reduced wave energy loss caused by tsunamis. A series of sensitivity analyses were conducted with respect to variable parameters(such as vegetation densities, wave heights, wave periods, bed resistance, and beach slopes) to identify important influences on mitigating tsunami damage on coastal forest beach.     

Two-dimensional scour simulations based on coupled model of shallow water equations and sediment transport on unstructured meshes

A two-dimensional scour model based on coupled system of shallow water equations (SWEs) and sediment transport on unstructured mesh is developed. The coupled system of hydrodynamic and morphodynamic equations is solved by finite volume method using Godunov scheme. Roe's approximate Riemann solver is used to calculate the inviscid fluxes. The use of unstructured mesh makes the model applicable to complex domains. However, it is difficult to evaluate the eigenvalues and eigenvectors of the Jacobian matrix in the global coordinate. The method proposed herein to deal with this difficulty is to transform the system into the local coordinate with one of the axes in the same direction as the interface outward normal vector. In the local coordinate system, the Jacobian matrix is simplified and the eigenvalues are analyzed using asymptotic method. Regular expansion breaks down when the flow is near critical. Uniformity of the expansion is achieved by changing the scales. Rotational invariance theorem is used to relate the interfacial fluxes in the global and local coordinate systems. Special treatment of the source term on unstructured grid makes the scheme stable and physically balanced (both mass and momentum). The method proposed in this paper for the eigen-system is very efficient comparing to iterative numerical methods. Results from the test cases show good agreement with the experiments.     

Large eddy simulation of stratified mixing in two-dimensional dam-break problem in a rectangular enclosed domain

Mixing in both coastal and deep ocean emerges as one of the important processes that determines the transport of pollutants, sediments and biological species, as well as the details of the global thermohaline circulation. Both the observations, due to their lack in space and time resolution, and most coastal and general circulation models due to inadequate physics, can only provide partial information about oceanic mixing processes. A new class of nonhydrostatic models supplemented with physically based subgrid-scale (SGS) closures, or so-called large eddy simulation (LES), is put forth as another tool of investigation to complement observational and large-scale modeling efforts.However, SGS models have been developed primarily for homogeneous, isotropic flows. Here, four SGS models based on Smagorinsky eddy viscosity and diffusivity are tested for stratified flows in the context of 2D dam-break problem in a rectangular enclosed domain. This idealized testbed leads to a number of simplifications about the initial conditions, boundary conditions and geometry, while exhibiting the dynamically complex characteristics of stratified flows involving the interaction of shear-induced mixing and internal waves. Direct numerical simulations (DNS) at high resolutions are taken as benchmark solutions. Under-resolved simulations without SGS terms (so-called DNS¹) are used to quantify the impact of SGS stresses. The performance of LES is assessed by using the time evolution of the volume fraction of intermediate density water masses generated by mixing. The simulations are conducted using a nonhydrostatic high-order spectral element model Nek5000 developed to exhibit minimal numerical dissipation and dispersion errors, which is advantageous to quantify accurately the impact of SGS stresses.It is found that all tested SGS models lead to improved results with respect to those from DNS¹. Also, SGS models allow for simulations with coarse resolutions that blow up in DNS¹ due to lack of adequate dissipation where needed. The SGS model in which the vertical eddy diffusion is modulated via a function that depends on the Richardson number Ri shows the most faithful reproduction of mixed water masses at all resolutions tested.The sensitivity of the results to the tunable parameter of the SGS model, to changes in the Ri-dependent function and resolution of the turbulent overturning scales is shown.     

高阶边界元方法求解规则波在三维局部渗透海床上的传播

基于线性势流理论,利用高阶边界元法研究了规则波在三维局部渗透海床上的传播。根据Darcy渗透定律推导出渗透海床的控制方程,利用渗透海床顶部和海底处法向速度和压强连续条件得到渗透海床顶部满足的边界条件。根据绕射理论,利用满足自由水面条件的格林函数建立了求解渗透海床绕射势的边界积分方程,采用高阶边界元方法求解边界积分方程进而得到自由水面的绕射势和波浪在局部渗透海床上传播过程中幅值的变化情况。通过与已发表的波浪对圆柱形暗礁的时域全绕射结果对比,证明了本文建立的频域方法计算波幅的正确性和有效性。利用这一模型研究了三维矩形渗透海床区域上波浪的传播特性,并分析了入射波波长、海床渗透特性系数等参数对波浪传播的影响。     

数值波浪水槽中完全非线性浅水波仿真特性研究

应用基于势流理论的时域高阶边界元方法,建立一个完全非线性的三维数值波浪水槽,通过实时模拟推板造波运动的方式产生波浪。通过混合欧拉-拉格朗日方法和四阶Runge-Kutta方法更新自由水面和造波板的瞬时位置。利用所建模型分别模拟了有限水深波和浅水波,与试验结果、相关文献结果和浅水理论结果吻合较好,且波浪能够稳定传播。系统地讨论造波板的运动圆频率、振幅和水深等对波浪传播和波浪特性的影响,并对波浪的非线性特性进行分析,研究发现造波板运动频率、运动振幅以及水深均将对波浪形态和波浪非线性产生显著影响。结果为真实水槽造波机的运动控制以及波浪生成试验提供了依据,便于实验室设置更合理的参数来准确模拟不同条件下的波浪。     

波浪沿斜面传播过程的数值模拟

精确模拟非线性波沿斜面传播过程非常困难,为此论文从势函数的边界积分方程出发,建立了一种时域内二维波浪模拟的数值模型,主要用来模拟完全非线性波浪的传播变形过程。论文的数值模型使用高阶二维边界元方法,采用可调节时间步长的基于二阶显式泰勒展开的混合欧拉-拉格郎日时间步进来求解带自由表面的线性或完全非线性波浪传播问题。在计算区域一端造出线性或非线性的周期性波浪,另一端采用消除反射波的人工粘性吸收边界。通过与现有理论比较证明了论文数值方法所得结果是准确可靠的。     

Nonlinear response of membranes to ocean waves using boundary and finite elements

The behavior of a highly deformable membrane to ocean waves was studied by coupling a nonlinear boundary element model of the fluid domain to a nonlinear finite element model of the membrane. The hydrodynamic loadings induced by water waves are computed assuming large body hydrodynamics and ideal fluid flow and then solving the transient diffraction/radiation problem. Either linear waves or finite amplitude waves can be assumed in the model and thus the nonlinear kinematic and dynamic free surface boundary conditions are solved iteratively. The nonlinear nature of the boundary condition requires a time domain solution. To implicitly include time in the governing field equation, Volterra's method was used. The approach is the same as the typical boundary element method for a fluid domain where the governing field equation is the starting point. The difference is that in Volterra's method the time derivative of the governing field equation becomes the starting point.The boundary element model was then coupled through an iterative process to a finite element model of membrane structures. The coupled model predicts the nonlinear interaction of nonlinear water waves with highly deformable bodies. To verify the coupled model a large scale test was conducted in the OH Hinsdale wave Research Laboratory at Oregon State University on a 3-ft-diameter fabric cylinder submerged in the wave tank. The model data verified the numerical prediction of the structure displacements and of the changes in the wave field.The boundary element model is an ideal modeling technique for modeling the fluid domain when the governing field equations is the Laplace equation. In this case the nonlinear boundary element model was coupled with a finite element model of membrane structures, but the model could have been coupled with other finite element models of more rigid structures, such as a pontoon floating breakwater.     

Time domain three-dimensional fully nonlinear computations of steady body–wave interaction problem

Three-dimensional fully nonlinear waves generated by moving disturbances with steady forward speed without motions are solved using a mixed Eulerian–Lagrangian method in terms of an indirect boundary integral method and a Runge–Kutta time marching approach which integrates the fully nonlinear free surface boundary conditions with respect to time.A moving computational window is used in the computations by truncating the fluid domain (the free surface) into a computational domain. The computational window maintains the computational domain and tracks the free surface profile by a node-shifting scheme applied within it. An implicit implement of far field condition is enforced automatically at the truncation boundary of the computational window.Numerical computations are applied to free surface waves generated by Wigley and Series 60 hulls for the steady problem. The present numerical results are presented and compared with existing linear theory, experimental measurements, and other numerical nonlinear computations. The comparisons show satisfactory agreements for these hydrodynamic problems.     

Numerical study of propagation of ship waves on a sloping coast

The aim of this paper is to investigate the propagation of ship waves on a sloping coast on the basis of results simulated by a 2D model. The governing equations used for the present model are the improved Boussinesq-type equations. The wave breaking process is parameterized by adding a dissipation term to the depth-integrated momentum equation. To give the boundary conditions at the ship location, the slender-ship approximation is used. It was verified that, although ship waves are essentially transient, the Snell's law can be applied to predict crest orientation of the wake system on a sloping coast. Based on simulated results, an applicable empirical formula to predict the maximum wave height on the slope is introduced. The maximum wave height estimated by the proposed method agrees well with numerical simulation results.