With the recent developments in the solution methods for large-dimensional nonlinear algebraic systems, fullyimplicit ocean circulation models are now becoming feasible. In this paper, the formulation of such a three-dimensional global ocean model is presented. With this implicit model, the sensitivity of steady states to parameters can be investigated efficiently using continuation methods. In addition, the implicit formulation allows for much larger time steps than can be used with explicit models. To demonstrate current capabilities of the implicit global ocean model, we use a relatively low-resolution (4°horizontally and 12 levels vertically) version. For this configuration, we present: (i) an explicit calculation of the bifurcation diagram associated with hysteresis behavior of the ocean circulation and (ii) the scaling behavior of the Atlantic meridional overturning versus the magnitude of the vertical mixing coefficient of heat and salt.

SUMMARYIn this paper two preconditioners for the saddle point problem are analysed: one based on the augmented Lagrangian approach and another involving artificial compressibility. Eigenvalue analysis shows that with these preconditioners small condition numbers can be achieved for the preconditioned saddle point matrix. The preconditioners are compared with commonly used preconditioners from literature for the Stokes and Oseen equation and an ocean flow problem. The numerical results confirm the analysis: the preconditioners are a good alternative to existing ones in fluid flow problems.

We present a new algorithm that constructs a fill-reducing ordering for a special class of saddle point matrices: the F -matrices. This class contains the matrix occurring after discretization of the Stokes equation on a C-grid. The commonly used approach is to construct a fill-reducing ordering for the whole matrix followed by an adaptation of the ordering such that it becomes feasible. We propose to compute first a fill-reducing ordering for an extension of the definite submatrix. This ordering can be easily extended to an ordering for the whole matrix. In this manner, the construction of the ordering is straightforward and it can be computed efficiently. We show that much of the structure of the matrix is preserved during Gaussian elimination. For an F -matrix, the preserved structure allows us to prove that any feasible ordering obtained in this way is numerically stable. The growth factor of this factorization is much smaller than the one for general indefinite matrices and is bounded by a number that depends linearly on the number of indefinite nodes. The algorithm allows for generalization to saddle point problems that are not of F -type and are nonsymmetric, e.g. the incompressible Navier-Stokes equations (with Coriolis force) on a C-grid. Numerical results for F -matrices show that the algorithm is able to produce a factorization with low fill.

The optimal temporal resolution for rainfall applications in urban hydrological models depends on different factors. Accumulations are often used to reduce uncertainty, while a sufficiently fine resolution is needed to capture the variability of the urban hydrological processes. Merging radar and rain gauge rainfall is recognized to improve the estimation accuracy. This work explores the possibility to merge radar and rain gauge rainfall at coarser temporal resolutions to reduce uncertainty, and to downscale the results. A case study in the UK is used to cross-validate the methodology. Rainfall estimates merged and downscaled at different resolutions are compared. As expected, coarser resolutions tend to reduce uncertainty in terms of rainfall estimation. Additionally, an example of urban application in Twenterand, the Netherlands, is presented. The rainfall data from four rain gauge networks are merged with radar composites and used in an InfoWorks model reproducing the urban drainage system of Vroomshoop, a village in Twenterand. Fourteen combinations of accumulation and downscaling resolutions are tested in the InfoWorks model and the optimal is selected comparing the results to water level observations. The uncertainty is propagated in the InfoWorks model with ensembles. The results show that the uncertainty estimated by the ensemble spread is proportional to the rainfall intensity and dependent on the relative position between rainfall cells and measurement points.

Fully-implicit primitive equation ocean models are useful to study the sensitivity of steady ocean flows to parameters, to determine bifurcations of these flows associated with instabilities and to use relatively large time steps in transient flow computations. This paper addresses a problem related to the origin of wiggles occurring in fully-implicit C-grid models. The situation considered is the computation of three-dimensional thermally-driven steady flows in a midlatitude spherical sector. We determine the reason why in a coarse resolution C-grid implicit model, the values of the lateral friction coefficients are restricted to far higher values than for the same B-grid model. The analysis also reveals why the B-grid discretization is superior for the computation of this type of flows.