Toy_DataAssimilation

==================================================================================

This notebook is intended to give an easy introduction into data assimilation using the 4DVAR method, focusing on the fundamentals of forming a tangent linear model, using a forward Euler time stepping scheme. Data assimilation is how we integrate observations into a numerical model to improve initial conditions and therefore improve forecasts.

Here, it is performed first on the logisitic growth problem then on a Lorenz '63 model.

We use Jupyter notebooks to drive the python code. The instructions for a setting up a jupyter environment that will work for this project are below. Please contact me with any issues:

[email protected]

Big thanks to Aneesh Subramanian, Bruce Cornuelle, and Ian Eisenman for some fundamental underpinnings of this project.

==================================================================================

To run please follow these instructions:

Download Anaconda: https://www.anaconda.com/distribution/

Create an anaconda environment (more on environments here: https://docs.conda.io/projects/conda/en/latest/user-guide/concepts/environments.html)

==================================================================================

In Your Command line:

step 1

conda create -n AdjLorenz python=2.7 numpy matplotlib ipykernel

step 2

source activate AdjLorenz

[if this command doesn't work try: "conda activate AdjLorenz"]

step 3

python -m ipykernel install --user --name AdjLorenz --display-name "Python2.7 (AdjLorenz)"

Then start a Jupyter Notebook

command line:

step 4

jupyter notebook

==================================================================================

step 5

This will open a browser window showing your computer file architecture. Select "Adjoint Model for Data Assimilation.ipynb"

a notebook will open, and change the kernel to AdjLorenz [kernel > change kernel > Python2.7 (AdjLorenz)]

Use 'shift + enter' to run each cell in the notebook.