Data Assimilation Demo

This web-page aims to demonstrate the Kalman Filter with some simple linear toy models. First choose model and data assimilation parameters and then click on "Run assimilation".

Model parameters

Model
Equation(s) Model code	The function model should return the state vector at the next time step given the state vector x at the time step n. function model(n,x) { // swap elements return [x[1],x[0]]; }
State vector size
Number of time steps
True initial condition xi

Data assimilation parameters

Method	(Extended) Kalman Filter Ensemble Kalman Filter Optimal Interpolation Nudging 4DVar
Covariance matrix of initial condition error Pi
Covariance matrix of model error Q
Observed every x grid points
Model time steps between observations
Observation error variance
Seed for random numbers
Relaxation time-scale for nudging Number of ensemble members Max. number of inner loops
Variance inflation Max. number of outer loops

Contact: Alexander Barth, a.barth at uliege.be. SANGOMA