1. 11-year Solar Signal in Transient Climate Simulations Lesley Gray NCAS University of Oxford Oxford: Dann Mitchell, Scott Osprey Met Office: Neal Butchart, Steve Hardiman, Sarah Ineson, Adam Scaife Reading: Manoj Joshi Imperial: Indrani Roy

2. Questions How well do we model solar influence on climate? - focus on Atlantic / European response Can we use models to improve understanding of mechanisms, given that we have limited observational time records? 1. Brief description of observations of Solar Influence on Climate. 2. Summary of 3 prime mechanisms for solar irradiance influence (‘top-down / bottom-up’ influences). 3. Model analysis Analysis of very long time-varying climate runs (CMIP5)

3. Regression analysis of ERA data Annual average 1979-2008 Frame and Gray 2010 Gray, Rumbold and Shine 2009 Regression analysis of SAGE satellite dataset Annual average 1985-2003 Soukharev and Hood 2006 Observations Smax minus Smin Solar Maximum: More UV radiation => higher temps More ozone => higher temps (early work of Labitzke; Haigh) TemperatureOzone

4. Observations: Zonal winds NCEP DJF zonally-averaged zonal winds 1979-2002 Haigh,Blackburn,Simpson,Sparrow Climatology (m/s) Smax minus Smin (m/s) NCEP zonal winds/temps 1979-1999 winds temps Kuroda and Kodera 2002 Stratosphere Troposphere Smax minus Smin +ve NAO Pattern at surface

5. Roy and Haigh 2010 Observations: Mean Sea Level Pressure (max-min) HadSLP2 1956-2004 Woollings et al 2010 See also Ineson et al 2011 1856-1905 ERA-40 1958-2001 Response is REGIONAL Solar max minus min can reach ~5-8 hPa in Atlantic but amplitude / sign varies with time => scepticism... (min-max)

6. NOAA Extended Reconstructed SST Hadley Centre HadISST 1871-present White and Liu 2008Meehl et al. 2009 Observations: Surface temperatures Response is REGIONAL and very small ~tenths K globally No agreement on spatial pattern; depends on how solar max/min are defined (Roy and Haigh 2010) Zhou and Tung 2010)

7. Smax minus Smin temperature 2 ‘top-down’ routes: Polar route: planetary waves / SSWs (only during winter) Equatorial route: synoptic-scale waves (all year round) Mechanisms: top-down altered planetary wave propagation = > fewer sudden stratospheric warmings (SSWs) +ve temp anomaly at stratopause +ve NAO at surface in Smax +ve temp anomaly lower strat => increased horizontal temp grad. => altered synoptic wave propagation westerly subtropical wind anomaly POLAR ROUTE EQUATORIAL ROUTE Solar Maximum: More UV radiation => higher temps More ozone => higher temps

8. Mechanisms: bottom-up The ‘bottom-up’ mechanism through total solar irradiance (TSI): Increased solar absorption during Smax in cloud-free subtropical oceans, increases evaporation; increased moisture converges into precipitation zones, intensifies precipitation and upward vertical motions, which strengthens Hadley and Walker circulations; stronger subsidence in subtropics gives positive feedback that reduces clouds and allows increased solar forcing. Cubasch, van Loon, Meehl, White

9. The MODEL Coupled ocean-troposphere-stratosphere Unified model (HadGEM2-CC) Stratosphere resolving coupled ocean-atmosphere Atmosphere: N96 1.875° x 1.25° 60 levels 0-84km (high-top) Ocean: 1.0° x 0.83° 40 levels includes non-orographic GWD scheme; interactive carbon cycle but no tropospheric or stratospheric interactive chemistry Historical+future ‘all-forcings’ CMIP5 simulations: 1x ensemble 1860-2100 2x ensembles 1960-2100 Increasing greenhouse gases (RCP8.5); monthly ozone variations are imposed, including ozone hole development + recovery; aerosols, solar cycle. 520 years

10. Lean et al. up to 2005; idealised in future (average of last 4 cycles) Spectrally partitioned based on traditional view i.e. NOT using recent SIM UV observations Solar Irradiance Variations Corresponding variations also imposed in ozone fields 520 years = 47x 11-yr cycles

11. Multiple Linear Regression Analysis with autoregressive noise model (AR1) 8 regression indices: CO2, ozone, 11-year solar cycle, long-term solar trend, volcanic aerosol, ENSO 3.4 index, 2 x QBO indices

12. 95% significance 90%+95% significance 95%+99% significance TemperatureStratospheric winds Tropospheric winds Model Results: Smax minus Smin

13. 520 years 47 solar cycles 95+99% confidence Model Results: Mean sea level pressure Smax minus Smin DJFMAM JJASON

14. Model Observations 1956-20041856-1905 520-year period 45-year period ‘X’ Model – Obs Comparison: DJF Smax minus Smin in MSLP

15. 45-year period ‘X’ 80%+95% confidence All years DJF What’s going on during ‘rogue’ 45-periods? Possible non-linear interactions with QBO? ENSO? SON Model Zonal winds DJF

16. 95%+99% confidence Model Results:Surface Temperatures Smax minus Smin 520 years 47 solar cycles SONDJF

17. Summary Model appears to capture both top-down mechanisms and bottom-up mechanism Model has periods where solar signal in troposphere appears to reverse – similar to obs BUT overall there IS an 11-year signal in mslp (primarily NAO region) and in SSTs (Europe and tropics) Still much to be understood in terms of mechanisms especially the relative timing / lagged responses and how solar interacts with QBO and ENSO.

18. Additional model runs – Amanda Maycock 3 x ensembles 2005-2070