Solar Influence on Stratosphere-Troposphere Dynamical Coupling Mike Blackburn (1), Joanna D. Haigh (2), Isla Simpson (2), Sarah Sparrow (1,2) (1) Department of Meteorology, University of Reading, UK (2) Space and Atmospheric Physics, Imperial College London, UK Earth Simulator Center, 12 November 2007

Outline Introduction - influence of the 11-year solar cycle on climate Observed atmospheric variability - regressions Model experiments to investigate how the tropospheric response over the 11-year cycle could be produced by a dynamical response to stratospheric heating equilibrium response to heating spin-up ensembles - mechanisms Comparison of two different stratospheric heating perturbation cases Relationship to internal annular variability

Observations of total solar irradiance >2 solar cycles Absolute values uncertain ~0.08% (1.1Wm -2 ) variation C. Frölich, PWDOC

Reconstruction using solar indices Extrapolate an index which correlates with TSI over the observed period Several indices! IPCC: change in radiative forcing since 1750: Wm -2 Conversion TSI to RF: 4 disc-area 0.7 albedo Sunspot number (grey) Amplitude of sunspot cycle (red) Length of sunspot cycle (black) aa geomagnetic index (green) IPCC TAR

Proposed Amplification Mechanisms Solar UV and impact on stratospheric O3 (Haigh 1994) - solar cycle variation ~7% at 200nm (cf 0.08% in TSI) absorption by O 3 stratospheric heating downward IR flux into troposphere dynamical impacts on troposphere changes in O 3 Modulation of low-level cloud cover (Svensmark & Friis- Christensen 1997) - assumed mechanism involving galactic cosmic rays

Solar index regressions using reanalysis data Crooks & Gray (2005) ECMWF reanalyses (ERA-40) Observed stratospheric temperature signal solar max - solar min

Temperature changes over the 11-year cycle Non-uniform. Increase of ~1K in equatorial stratosphere, decreasing towards the poles. Banded increase in temperature in mid-latitudes. Figure: Haigh (2003) Multiple regression analysis of NCEP/NCAR reanalysis,

Circulation changes over the 11-year cycle Weakening and poleward shift of the mid-latitude jets. Weakening and expansion of the Hadley cells. Poleward shift of the Ferrell cells. Haigh and Blackburn (2006) Multiple regression analysis of NCEP/NCAR reanalysis, DJF,

GCM simulation (UGAMP GCM) (a) Control run (b) Difference between solar maximum and solar minimum: Total solar irradiance Stratospheric ozone Similar response found using the Met Office model GCM response to solar irradiance & ozone (DJF) Haigh, Science (1996); QJRMS (1999)

The Hypothesis Are the tropospheric changes observed over the 11-year solar cycle a response to perturbations in the tropical (lower) stratosphere, which are a response to enhanced UV absorption at solar maximum? Investigate using idealised stratospheric heating experiments in a simplified atmospheric GCM: Can we reproduce the tropospheric response? What (dynamical) mechanisms are involved?

Simplified GCM - dynamical core model Based on University of Reading primitive equation model: (1) Spectral dynamics: T42 L15 No orography Newtonian cooling – idealised equinoctial radiative- convective equilibrium temperatures T e (lat,height) (2) Boundary layer Rayleigh friction Experiments: 1.Equilibrium response to perturbations to stratospheric T e (Haigh, Blackburn & Day, J.Clim., 2005) 2.Spin-up ensembles 200 x 50-day run (1)Hoskins & Simmons (1975) (2) Held & Suarez (1994)

Comparing Uniform and Equatorial Heating 5K 0K 5K Equatorial heating (5K) (E5) Uniform heating (5K) (U5) Weakening and poleward jet shift? How does the tropospheric response depend on the heating distribution?

(b) (c) (a) Simplified GCM equilibrium response zonal wind (ms -1 ) Run C U5 - C E5 - C

Vertically-integrated budget of zonal momentum

Haigh et al (2005) - Equatorial heating gave a similar tropospheric response to that seen over the Solar cycle Coherent displacement of the jet and storm-track How does this arise? Spin-up ensemble for the equatorial heating case: –200, 50-day runs Ensemble spin-up Experiments: 5K 0K 4.5K 0.5K

Change in temperature over the spin-up Control Equilibrium (Equatorial heating (5K) – Control)

Change in zonal wind over the spin- up Equilibrium (Equatorial heating (5K) – Control) Control

increases decreases Changes in eddy momentum fluxes are in the right sense to drive meridional circulation changes. Mean meridional circulation Horizontal Eddy Momentum Flux [uv]

Anomalous meridional circulations are accompanied by zonal wind accelerations in the troposphere: increases decreases Mean meridional circulation Zonal mean zonal wind [u]

Comparison with zonally symmetric model. Eddy forcing remains fixed at its value of the control run. Heating perturbation applied and the model run as before. Not much response in the troposphere, particularly at mid/high latitudes. it is altered eddy momentum fluxes that are important in driving the tropospheric circulation changes. Full 3D modelNo change in Eddy fluxes [mmc] [u]

Whats causing the change in eddy momentum fluxes? E-P Flux Refractive Index C=8ms -1

Days 0 to 9 of the spin-up: Change in E-P Flux and Change in Change in : a) Only changing b) Only changing

Days 40 to 49 of the spin-up: Change in E-P Flux and Change in Change in : a) Only changing b) Only changing

Contributions to the change in PV gradient (days 0 9): Meridional Curvature Third term (only changing ) Total change in PV gradient

Outline of mechanism: Altered vertical temperature gradients Zonal wind accelerations stratosphere/tropopause Change in horizontal eddy momentum flux Changes in mean meridional circulation Zonal wind accelerations in the troposphere. Altered horizontal temperature gradients

Comparing Uniform and Equatorial Heating: 5K 0K 5K Equatorial heating (5K) (E5) Uniform heating (5K) (U5) Weakening and poleward jet shift. Weakening and equatorward jet shift.

E-P Flux E5 (days 0 9 )U5 (days 0 9 )

E5 (days )U5 (days ) E-P flux and n 2

Conclusions (1) The tropospheric response to increased Solar activity could be produced by a dynamical response to increased heating in the stratosphere. Changes in eddy momentum flux are important in driving circulation changes in the troposphere. Feedback with changing zonal wind in the troposphere influencing eddy propagation. Change in vertical temperature gradient around the tropopause and its localisation in latitude is important in determining the direction of the jet shift.

Relationship with internal annular variability Internal Variability –Empirical Orthogonal Functions (EOFs) –Phase space trajectories –Vertically integrated zonal momentum budget –EP Flux and zonal wind anomalies Dynamical mechanisms

Equilibrium Response U5: Jets weakened and shifted equatorwards. E5: Jets weakened and shifted polewards. Control Run U5 - ControlE5 - Control Latitude (equator to pole) Height Haigh et al (2005)

Leading Modes of Variability EOF 1 (51.25%) EOF 2 (18.62%) Latitude (equator to pole) Height Mean state differences from idealised forcing experiments project strongly onto the leading modes of variability in the control run.

Projections of Mean State Differences The signal of the experiments can be viewed as displacements in principal component (PC) phase space. Mean state differences project most strongly onto EOF1 and EOF2. PC2 Amplitude EOF Number Amplitude (ms -1 ) U5-Control E5-Control PC1 Amplitude Phase Space Poleward Equatorward Broader, Weaker Narrower, Stronger

Internal Variability: Phase Space Trajectories At low frequencies circulation is anticlockwise with a timescale of ~46 days. At high frequencies circulation is clockwise with a timescale of ~ 7 days. Arrow ×2 Arrow ×½ Unfiltered Periods Longer than 30 Days Low Pass Filter Periods Shorter than 21 Days High Pass Filter PC1 PC2

Zonal Wind Evolution: Low Frequency Jet strengthens and moves polewards. New subtropical jet grows forming double jet structure. Poleward jet collapses and merges with the new subtropical jet.

Zonally-averaged zonal momentum equation: Integrated through depth of atmosphere: or: Vertically Integrated Zonal Momentum Budget

Vertically Integrated Momentum Budget: Low Frequency

Vertically Integrated Momentum Budget: High Frequency

Phase Space View of Momentum Budget Surface stress points slightly in advance of the origin in phase space. Eddies change behaviour at high and low frequencies. PC1 PC2 PC1 PC2 Low Pass High Pass

EP Flux Anomalies: Low Frequency Low PC1 Composite High PC1 Composite EP Flux anomalies reinforce current state. Subtle differences between the wind anomalies and the EP Flux cause phase space circulation

EP Flux Anomalies: High Frequency Low PC1 CompositeHigh PC1 Composite Less LC1 More LC1 More LC2 Less LC2 longitudelatitude Thorncroft et al (1993)

Conclusions (2) Tropospheric response to stratospheric temperature changes project strongly onto dominant modes of annular variability. Distinct difference in behaviour at high and low frequencies: –Low frequency: poleward migration (quasi-equilibrium) –High frequency: equatorward migration (strongly evolving) Eddies drive the phase space trajectory at high and low frequencies: –Eddies are balanced more strongly by surface stress at low frequencies leading to a slower circulation –High frequency eddy anomalies reflect past baroclinicity; feedback understood in terms of LC1/LC2 behaviour

- Thank you -

Multiple regression of zonal mean T (200hPa) NCEP-NCAR reanalysis - solar variability (red) - volcanic aerosol (green) - QBO (cyan) - NAO (blue) - ENSO (black) - trend (straight black line) - amplitude/phase of annual & semi-annual cycles 35°S 35°N 35°S T at 35°S T (200hPa) regressions Haigh (2003)