DAPack(Py-DA*)– The Data Assimilation Testing-Suite!

Data Assimilation(DA) is a rich literature with tons of open questions. Once a researcher decides to test a new idea, several choices have to be made such as the model and the experimental settings. DAPack is intended to be a work-in-progress package for DA. Our vision is to provide an open-source high-level language that enables DA researchers to collaborate and to help them avoid reinventing the wheel. We want it to be as simple as possible to add new models and DA schemes to the package. This package is intended to ease the process of comparing the performance of different schemes efficiently and easly.

  • Modeled implemented so far:
    • Two-variables pendulum (0D),
    • Lorenz-3 (three-variables Lorenz) model (0D),
    • Lorenz-96 (fourty-variables Lorenz) model (0D),
    • All HyPar models (1D,2D,3D).
  • Filters implemented so far:

Installing DAPack

  • Navigate to the directory where you want to install DAPack.

  • Download DAPack from git:

    >> git clone https://bitbucket.org/emconsta/dapack.git

  • Install HyPar:

    Follow instructions here There you can find links to download and install PETSc which is optional but recommended.

  • Set DAPack variables:

    Navigate to <DAPack root path>/DAPack/ Set the variables in the file setup_dapack.py

    • Hypar_bin : Actual path of the binary file of HyPar. <HyPar root path>/bin/HyPar
    • DA_Pack_RootPATH : Actual path of the DAPacke. This is the root directory you get by cloning the repository.
    • CCompiler : Chose the number of C compiler on your machine from the list given in the variable CCompilers.

Indices and tables