§4DVar requires an adjoint
model to back-propagate information -- this can be a royal pain to develop!
§The EnKF
can get around needing an adjoint by using a filter-lag rather than a
fixed-interval Kalman smoother.
However, the need to propagate multiple time steps in the state makes it less efficient than the 4DVar
method
§Both give
a low-rank estimate of the a posteriori covariance matrix
§Both can
account for dynamic errors
§Both
calculate time-evolving correlations between the state and the measurements