Evaluation of northern hemisphere blocking climatology in the global environment multiscale (GEM) model and in the present and future climate as simulated by the CMIP5 models

The performance of the Global Environmental Multiscale (GEM) model, Canadian operational numerical model, in reproducing atmospheric low-frequency variability is first evaluated in the context of Northern Hemisphere blocking climatology. The validation is conducted by applying a comprehensive but re...

Full description

Bibliographic Details
Main Author: Dunn-Sigouin, Etienne
Other Authors: Seok-Woo Son (Internal/Supervisor)
Format: Thesis
Language:English
Published: McGill University 2012
Subjects:
Online Access:http://digitool.Library.McGill.CA:80/R/?func=dbin-jump-full&object_id=110589
Description
Summary:The performance of the Global Environmental Multiscale (GEM) model, Canadian operational numerical model, in reproducing atmospheric low-frequency variability is first evaluated in the context of Northern Hemisphere blocking climatology. The validation is conducted by applying a comprehensive but relatively simple blocking detection algorithm. Comparison to reanalysis reveals that the maximum blocking frequency over the north Atlantic and western Europe is generally underestimated and its peak season is delayed from late winter to spring. This contrasts with the blocking frequency over the north Pacific which is generally overestimated during all seasons. The biases in blocking frequency are found to be largely associated with the biases in climatological background flow. Specifically, modelled stationary wave shows a seasonal delay in zonal wavenumber 1 and an eastward shift in zonal wavenumber 2 components. Next, we extend our methodology to preliminary analyses of Northern Hemisphere blocking climatology from a subset of climate models participating in the Coupled Model Inter-comparison Project phase 5 (CMIP5). Historical integrations reveal that the maximum Euro-Atlantic blocking frequency is generally underestimated during the cold season and that significant overestimation of maximum Pacific blocking frequency occurs throughout the year in some models, as compared to reanalysis. In contrast, RCP8.5 integrations show a weak hint of reduced blocking frequency over the Pacific sector in comparison to historical integrations. However, no significant trend in terms of block duration within the RCP8.5 integrations is found. Les performances du modele Global Environment Multiscale (GEM), qui est le modele numerique operationel Canadien, a reproduire les variabilites atmospheriques de basse frequence sont evaluees en premier lieu dans le contexte de la climatologie de bloquage atmospherique dans l'hemisphere Nord. Afin de valider le modele, un algorithme de detection de bloquage qui est a la fois comprehensif et relativement simple est applique aux donnees atmospheriques. Les resultats montrent que la frequence maximum de bloquage au dessus de l'Atlantique Nord et l'Europe de l'Ouest est generalement sous-estimee et il y un delai dans la saison d'amplitude maximale puisqu'elle se produit au printemps au lieu de tard en hiver. De plus, la frequence de bloquage est generalement sur-estimee au dessus du Pacifique Nord. Il a ete trouve que les erreurs dans la frequence de bloquage sont grandement associees aux erreurs dans la circulation climatologique de l'atmosphere. En fait, les ondes stationnaires modelisees montrent un delai saisonnier dans le nombre d'onde zonal 1 et un deplacement vers l'Est des composantes du nombre d'onde zonal 2. Ayant confiance en la capacite de notre index pour identifier des bloquages atmospheriques, nous appliquons notre methodologie sur des analyses preliminaires de bloquage climatologique dans l'hemisphere Nord a partir d'un sous-ensemble de modeles climatologiques faisant partie du Coupled Model Inter-Comparison Project Phase 5 (CMIP5). Les integrations historiques revelent que la frequence maximale de bloquage sur l'Euro-Atlantique est generalement sous-estimee durant la saison froide et que la sur-estimation de la frequence maximale de bloguage sur le Pacifique se produit tout au long de l'annee dans certains modeles. En comparaison, les integrations de type RCP8.5 montrent un leger indice d'une reduction de la frequence de bloquage sur le Pacifique meme si aucune tendance significative en terme de duree de bloquage n'a ete trouvee.