ON THE POSSIBILITY OF CORRECTION OF THE FORECASTING FOR INEXACT OBSERVATIONS ON THE BASIS OF DATA ASSIMILATION

In this paper the application of one of the Data Assimilation methods based on the Kalman filter usage is described. The application of the Ensemble Kalman Filter in forecasting is shown on the Lorenz attractor model. The normal noise, which dispersion changes for various experiment variants, is added on the model and on the obtained measurements. When carrying out an experiment both for model, and for forecasting, the mistakes have been considered. The assessment of influence of model and filter parameters on the quality of forecasting implementation is carried out.