Four-dimensional data assimilation

Four-dimensional data assimilation (FDDA) A technique by which high-frequency observations are used to update the model state during the time integration step, improving the fidelity of meteorological model predictions. 

Four-dimensional data assimilation Method of analyzing the state of the atmosphere or ocean by using a prediction model in such a way to best fit observed data during a period of 6-12 hours. 

Meanwhile, the idea was formulated about resolving the full set of primitive hydro- and thermodynamic equations with all the boundary conditions specified successively correcting the current model fields of the temperature, salinity, and SLE by their observed values with the use of this or that kind of assimilation algorithm [35,36]. This approach is sometimes referred to as a four-dimensional analysis. Strictly speaking, it has little in common with the initial diagnostic methods. They are joined only by the common goal - the hydrodynamic calculations of the fields of currents from the data of observations of the temperature, salinity, and sea level. Therefore, in this section, we consider the results of application of all the above-mentioned approaches. 

FIGURE 2 Schematic diagram for atmospheric forecast-analysis system. The two processes in the square enclosed by dashed lines can now be combined in the four-dimensional variational data assimilation using a continuous stream of observed data. See text for details. 

The method of analyzing the state of the atmosphere by utihzing a continuous stream of different kinds of observations to produce the evolution of flow that best fits all the observations during a period of assimilation window of 6 to 12 (or more) hours is referred to as four-dimensional variational (4DVAR) data assimilation. 

---