Implementation of Scale-Dependent Background-Error Covariance Localization in the Canadian Global Deterministic Prediction System

Abstract The approach of applying different amounts horizontal localization to ranges background-error covariance scales as proposed by Buehner and Shlyaeva was recently implemented in the four-dimensional ensemble–variational (4DEnVar) data assimilation scheme global deterministic prediction system (GDPS) at Environment Climate Change Canada operations. To maximize benefits from this reduce sampling noise ensemble-derived covariances, it necessary adopt a new weighting between climatological flow-dependent covariances that increases significantly role latter. Thus, December 2021 GDPS became first operational medium-range weather forecasting rely completely on troposphere lower stratosphere. experiments led adoption these two related changes their impacts forecasts up 7 days for various regions globe during boreal summer 2019 winter 2020 are presented here. It is also illustrated relying more amplifies positive when background ensemble generation strategy improved.