Strat-trop interaction and Met Office seasonal forecasting Adam Scaife, Jeff Knight, Andrew Marshall, Sarah Ineson and Alberto Arribas Hadley Centre, Met Office, Exeter, UK. European winter climate and the NAO Stratospheric influence: Mean and Extremes Trop-strat vs trop models: NAO and ENSO GloSea4: tests and plans Conclusions Adam Scaife

North Atlantic Oscillation Dominant pattern of Atlantic-European weather variability Related to changes in the position and strength of the Atlantic Storm track NAO r = 0.62 NET* r = 0.67 r = 0.84 © Crown copyright 2004 * NET: 15W-45E,30-65N CET Adam Scaife

Stratospheric influences Change in NAO index Observations have a large increase in NAO Control run has very little increase in NAO (includes GHG, aerosols, observed SST etc, c.f. Rodwell et al. 1999) Impose a body force in the model’s stratosphere (c.f. Norton 2003) => Increase in stratospheric wind from 1960s to 1990s => Increase in NAO similar to observed value NAO Change in surface pressure U50hPa © Crown copyright 2004 Adam Scaife

Winter climate and stratospheric change Model Temperature Observed Temperature European T trends 1960s-1990s HadAM3 ctl 0.15K/decade HadAM3 expt 0.59K/decade Observations 0.53K/decade Model Precipitation Observed Precipitation © Crown copyright 2004 Scaife et al. GRL, 2005 Adam Scaife

Observed vs forced changes in extremes Observations Model (forcings) Model (forcings + NAO) Observed changes larger than modelled changes with all anthropogenic forcings Signs of dipole across Europe in observed data Scaife et al., J.Clim., 2007 © Crown copyright 2004 Adam Scaife

Winter 2005/6: cold Europe case study December 2005 January 2006 February 2006 Colder than 1970-2000 over much of Europe 2nd coldest in 10 years using area mean T Record snowfall in parts of central Europe Late winter colder than early winter Extreme stratospheric sudden warming in January Zonal wind through the winter (c.f. Baldwin and Dunkerton 2001, Charlton et al. 2004) © Crown copyright 2004 Adam Scaife

Winter 2005/6 Tropospheric models (HadAM/HadGAM) 50/25 members HadISST as a boundary condition Tropospheric Model + stratospheric perturbation 25 members HadISST as lower boundary condition Perturbed stratosphere from 1st Jan Troposphere-stratosphere model Zonal wind at 50hPa © Crown copyright 2004 Adam Scaife

Winter 2005/6 Old model New model New trop-strat model Old model + strat warming Observed Anomalies New model New trop-strat model Cold European signal from Atlantic SSTs Cold European signal from stratospheric warming Stratospheric warming unpredictable from SST alone Scaife and Knight, QJ, in prep. © Crown copyright 2004 Adam Scaife

ENSO effects on Europe L60 HadGAM L38 HadGAM 50hPa gph PMSL ENSO events produce a –ve NAO response (e.g. Bronniman et al. 2004) Only the weakest 2/3 of observed events and didn’t occur in trop model (Toniazzo and Scaife 2006) Upper level component appears in trop-strat models (Hamilton, 1993, Manzini et al., 2006) L60 HadGAM L38 HadGAM 50hPa gph PMSL © Crown copyright 2004 Adam Scaife

Possible mechanism Atlantic SST Atlantic SST NORMAL Stratospheric Polar Vortex WEAK Stratospheric Polar Vortex WEAK Stratospheric Polar Vortex Rossby waves NAO - NAO - NAO --- Wind stress + heat fluxes Heat fluxes Atlantic SST Atlantic SST and vice versa in +ve NAO winters © Crown copyright 2004 Adam Scaife

Decadal variability Stratosphere-troposphere Model (HadAM3) Stratosphere-troposphere Model (HadGAM1) NAO TRENDS 1960/61-1990/91: Observations: 11.9 hPa Strat-trop HadAM3: 10.3 ± 1.2 hPa Strat-trop HadGAM: 7.8 ± 1.5 hPa Trop HadAM3: 5.6 ± 1.4 hPa Possibility of improved signal to noise ratio in ensemble decadal predictions Troposphere Model (HadAM3) © Crown copyright 2004 Adam Scaife

A simple metric to judge models: HadAM3 Observations HadGAM Zonal Wind (m/s) NAO (hPa) NAO (hPa) NAO (hPa) r(U50,NAO) = 0.0 r(U50,NAO) = 0.5 r(U50,NAO) = 0.8 NAO (hPa) U50 (m/s) Ratio HadAM3 10.3 0.5 21.0 HadGAM 7.8 3.2 2.4 OBS 11.9 5.6 2.1 © Crown copyright 2004 Adam Scaife

GloSea4 – next Met Office seasonal forecast system We are currently testing an extended stratospheric model as a likely candidate for use in the next Met Office seasonal forecast system GloSea4, due in 2009. Skill measures and mechanisms using case studies: L38 vs L60 – overall improvement? Volcanic years, cold globe but warm Europe? Weak/strong stratospheric jet years, predictable? Cold/warm Europe? El Nino years, weak polar vortex and cold Europe? Development of an operational volcanic “switch” © Crown copyright 2004 Adam Scaife

Conclusions The NAO dominated mean and extreme European winter climate shift from the 1960s to 1990s and the stratosphere played an important role. A picture is emerging which links the stratosphere to the NAO (sometimes because of ENSO) and to cold, blocked European winters. Extended model experiments produce a –ve NAO response to El Nino and potentially larger decadal responses to SST but the winter 2005/6 stratospheric anomaly is not predictable from SST information alone. Tests are underway to include initial atmospheric conditions to evaluate a trop-strat model as the next Met Office seasonal forecast model. © Crown copyright 2004 Adam Scaife