Concepts for a pattern‐oriented analysis ensemble based on observational uncertainties

摘要: The idea is proposed of an analysis ensemble deterministic, model-independent analyses. based on random perturbations irregularly distributed observations. purpose implementing to define uncertainties in fields due their observational background and errors. As one possible application, the uncertainty information could, future, be used confidence intervals for verification measures depending reference data. system VERA a high-resolution Central European observation network are as testbed development methodology. Several approaches defining weights perturbation investigated compared. Basic determined by sophisticated data quality control scheme producing error estimates These can combined with additional attempting include more explicitly spatial structure observed ensemble. provided either principal component time series or 2D-discrete wavelet transform. Strengths weaknesses different adjustments discussed. It shown that wavelet-based approach lead useful results several meteorological parameters tested. Copyright © 2011 Royal Meteorological Society