Print this Page



AGRIF (Adaptive Grid Refinement In Fortran) is a Fortran 90 package for the integration of full adaptive mesh refinement (AMR) features within a multidimensional finite difference model written in Fortran. Its main objective is to simplify the integration of AMR potentialities within an existing model with minimal changes.

Capabilities of this package include the management of an arbitrary number of grids, horizontal and/or vertical refinements, dynamic regridding, parallelization of the grids interactions on distributed memory computers.

AGRIF requires the model to be discretized on a structured grid, like is typically done in ocean or atmosphere modelling. As an example, AGRIF is currently used in the following ocean models: MARS (a coastal model developed at IFREMER-France), ROMS (a regional model developed jointly at Rutgers and UCLA universities) and OPA/NEMO (a general circulation model used by the French and European scientific community). AGRIF is licensed under a GNU (GPL) license and can be downloaded at its web site. More than two hundred downloads of the software have been done during the last year.

Website: Contact: Laurent Debreu (MOISE).


Tangent and adjoint models for the NEMO platform of the oceanic modelling that have been delopped by the MOISE team have been published now under Cecill license and distributed by the NEMO consortium.

Website: Contact: Arthur Vidard (MOISE).


DatIce is a data assimilation tool designed to estimate consistent chronologies for a set of deep ice cores (i.e., depth-age relationships for the ice and trapped gas). A cost function is built from bayesian inference. Its minimization provides a trade-off between background chronologies calculated with glaciological models (accumulation rate, densification, and ice flow models), and stratigraphic or chronological constraints derived from measurements on ice cores (i.e., volcanic ash layers, greenhouse gases, etc.). The cost function includes covariance error matrices, upon which confidence intervals of the solution can be assessed.

Website: Contact: Bénédicte Lemieux (MOISE).


The software Stochastic Downscaling Method (SDM) is aimed to be coupled to large-scale numerical weather forecasting models. Indeed, based on large-scale informations, SDM refines the forecast thanks to a Langevin model. This forecast is computed thanks to Monte-Carlo average on stochastic partices, for which we built a specific system of SDEs. The model inputs are thus large-scale data and the desired “zoom” (i.e. the scale to which the users to refine the computations). The outputs are fine-scale data. In addition, the code provides confidence intervals.
The software is not distributed for now, but it is used by collaborators at Ecole Polytechnique (P. Drobinski and T. Salameh, Laboratoire de Météorologie Dynamique) for validation in the Rhône Valley, in the framework of a partnership with the French Environment and Energy Management Agency ( Contact: Antoine Rousseau (MOISE).


The CompModSA package (Complex Computer Models Sensitivity Analysis) is an R package aimed at sensitivity analysis of computer models (that is, the determination of the most influential input parameters on a quantity of interest, which is an output of the model). The technique involved is a Monte-Carlo method, which is a flexible and rigorous way of estimating Sobol indices (the indices quantifying the relative influences of each input parameter). However, this method often requires a large number of model outputs, for different input parameters’ values; this makes Monte-Carlo method often computationally intensive. To overcome this problem, the CompModSA package use the so-called response surface method, which uses a restricted sample of model outputs to build a fastly-evaluated approximation to the original code. This package also comes with an indicator of the error made on the estimated Sobol indices.
Alexandre Janon (MOISE, LJK) contributed to the development of this software, which has been initiated by Curtis Storlie (Los Alamos National Lab).

Website: Contact: Curtis Storlie (Los Alamos National Lab), Alexandre Janon (MOISE).


RHEOLEF is a finite element computer environment (C++ classes and unix commands). It is designed as a convenient laboratory for computations in applied mathematics, involving finite element-like methods. It provides a set of unix commands and C++ algorithms and containers. Containers cover first the classic graph data structure for sparse matrix formats and finite element meshes. A higher level of abstraction is provided by containers related to approximate finite element spaces, discrete fields and bilinear forms. Current applications cover : Poisson problems in 1D 2D and 3D with P1 or P2 elements; Stokes problems in 2D and 3D, with P2-P1 or P1 bubble-P1 elements; linear elasticity in 2D and 3D, with P1 and P2 elements, including the incompressible and nearly incompressible elasticity; characteristic method for convection-diffusion, time-dependent problems and Navier-Stokes equations; auto-adaptive mesh based for 2D problems; axisymetric problems; multi-regions and non-constant coefficients.

Input and output are possible in various file formats for meshes generators and numerical data visualization systems (vtk, plotmtv, gnuplot). Both reference manual and users guide are available. The license is GPL. It is well suited for numerical simulation of complex systems described by partial differential equations and has been developed and improved in IDOPT for both simulation of yield stress flows (mud flows and dense snow avalanches) and for powder-snow avalanches at high Reynolds numbers.

Website: Contact: Pierre Saramito (EDP/MOISE).


Dassflow is a river hydraulics simulation software designed for variational data assimilation. The model is based on the bidimensional shallow-water equations, solved by the finite volume method using the HLLC approximate Riemann solver. The adjoint code is available, as well as all the optimization framework that is necessary to easily build up new data assimilation experiments. The software is written in Fortran 90. It is currently evolving to take into account Lagrangian data for assimilation experiments.

Website: Contact: Jérôme Monnier (IMT Toulouse).

Permanent link to this article: