Evaluation of Correlation Functions and Multi-model Forecasting of Geopotential Height and Temperature in the Troposphere and Lower Stratosphere

Statistical evaluation of three-dimensional auto- and cross-correlation functions for increments from the first guess was performed to interpolate the complex forecast (postprocessing) of geopotential height and temperature to regular grid points. The forecast fields from the ICON model were used as the first guess. Positive definiteness was provided in the evaluation. The verification of the complex forecast confirmed a significant reduction of the forecast error with a lead time up to 72 hours at standard levels in the troposphere and lower stratosphere in comparison with all input hydrodynamic forecast models.