Blocking in the Spectral Element Atmospheric Model: Effect of Topography and Resolution

BLOCKING IN THE SPECTRAL ELEMENT ATMOSPHERIC MODEL: EFFECT OF TOPOGRAPHY AND RESOLUTION Aimé Fournier, Joseph Tribbia National Center for Atmospheri Resear h Boulder CO 80307-3000 USA fournier u ar.edu Mark A. Taylor Albuquerque NM 87108 USA www.s d.u ar.edu/ ss/staff/taylor ABSTRACT Atmospheri blo king (AB) is an important aspe t of intraseasonal weather/ limate predi tion whi h is inadequately aptured by state-of-the-art GCMs. We study the e e t of horizontal resolution (HR) on AB-like events in a primitive-equation, dynami alore-only framework, the Spe tral Element Atmospheri Model. After explaining SEAM's relative advantages w.r.t. exibility, (spe tral) a ura y and e ient parallel implementation, we extend the results of others' GCM-based studies, to the higher HR easily attainable by SEAM, in luding only Heldand-Suarez for ing and topography, not physi s. We use Z elds from 4 3600-day runs, with and without topography, and at 2 HRs up to an equivalent of T181. A novel set of quasi-obje tive, quasistationary-ridge (QSR) dynami al-statisti al diagnosti s are introdu ed: 1) the = 0:25 level, from around 35 to 65ÆN, is determined to be the zone of greatest var(Z); 2) averaging Z over this zone yields a longitude-time distribution whose maxima move with a distribution of ridge speeds; 3) the latter distribution's full width at half-max is a riterion for extra ting QSR events, su h as shown around 40ÆE in the 8-day sequen e of Z( ; ) snapshots shown in Fig. 1. The frequen y distribution of events vs duration de reases exponentially, as found by Dole and Gordon, D'Andrea et al. and others, but with a 2to-8 times smaller e-folding, 18h. We also present results of statisti ally stratifying this frequen y distribution by longitude and by amplitude. INTRODUCTION Atmospheri blo king (AB) is an important aspe t of intraseasonal weather/ limate predi tion, but state-of-the-art GCMs do not adequately model AB w.r.t. its o urren e, frequen y (F ), strength or duration. This is the rst part of a two-part study, in whi h our strategy was to fo us on e e t of model horizontal resolution, N , on model phenomena similar to AB. We also looked for e e ts of −60 −50 −40 Z(σ=0.25,t=79d) (km) −60 −50 −40 Z(σ=0.25,t=80d) (km) −60 −50 −40 Z(σ=0.25,t=81d) (km) −60 −50 −40 Z(σ=0.25,t=82d) (km) −60 −50 −40 Z(σ=0.25,t=83d) (km) −60 −50 −40 Z(σ=0.25,t=84d) (km) −60 −50 −40 Z(σ=0.25,t=85d) (km)