We propose an improvement of the algorithm of joint assimilation of the data on climatic temperature, salinity, and altimetric sea level in a model of circulation. Unlike the previous works, the variances of the forecast errors of temperature and salinity and the cross-covariance functions of of the forecast errors of salinity-level and temperature-level depend on the dynamics of waters. It is shown that the structure of the fields of cross-covariance functions in the upper mixed layer is formed by the vertical turbulent diffusion of the variances of forecast errors of temperature and salinity. At greater depths, these statistical characteristics are mainly determined by the vertical advection. We compared the results of calculations with and without taking into account the dynamics of the statistical characteristics. The analysis of the influence of the dynamics of these characteristics makes it possible to reconstruct the mutually adapted climatic fields of temperature, salinity, and horizontal and vertical current velocities in the Black Sea with the assimilation of data in the numerical model in each time step.