4 The 3DVAR can also be used for a **time analysis of the observations of a given
5 dynamic model**. In this case, the analysis is performed iteratively, at the
6 arrival of each observation. For this example, we use the same simple dynamic
7 system [Welch06]_ that is analyzed in the Kalman Filter
8 :ref:`section_ref_algorithm_KalmanFilter_examples`. For a good understanding of
9 time management, please refer to the :ref:`schema_d_AD_temporel` and the
10 explanations in the section :ref:`section_theory_dynamic`.
12 At each step, the classical 3DVAR analysis updates only the state of the
13 system. By modifying the *a priori* covariance values with respect to the
14 initial assumptions of the filtering, this 3DVAR reanalysis allows to converge
15 towards the true trajectory, as illustrated in the associated figure, in a
16 slightly slower speed than with a Kalman Filter.
20 Note about *a posteriori* covariances: classically, the 3DVAR iterative
21 analysis updates only the state and not its covariance. As the assumptions
22 of operators and *a priori* covariance remain unchanged here during the
23 evolution, the *a posteriori* covariance is constant. The following plot of
24 this *a posteriori* covariance allows us to insist on this property, which
25 is entirely expected from the 3DVAR analysis. A more advanced hypothesis is
26 proposed in the forthcoming example.