Preferred Modes of Variability and Their Relationship with
Preferred Modes of Variability and Their Relationship with Climate Change Seok-Woo Son and Sukyoung Lee The Pennsylvania State University Department of Meteorology
Annular Mode - Dominant internal variability of the atmosphere v Leading EOF of SLP SH NH [u] v Zonally symmetric v Quasi-barotropic v Useful for understanding internal variability v Useful for understanding climate change (? ) SLP Thompson et al. 2000
Thompson et al. 2000 Kushner et al. 2001 SH [u] response to global warming NH Annular Mode SH Annular Mode pressure (h. Pa) NH [u] trend 1968 -1997 latitude
“Spatial pattern” of annular mode ≈ recent trend in the observed and simulated zonal-mean circulation To what extent annular mode is capable of predicting zonal-mean climate change?
Purpose and Approaches Evaluate the predictability of zonal-mean climate change by annular mode in terms of their spatial structures. Total 49 simulations by differing radiative heating in a simple GCM Structure of [u] in the statistically steady state ( [u] ) Internal variability of [u] with a help of EOF 1 and EOF 2 Annular mode vs. Climate change Annular mode – EOF 1 of [u] (regressed against PC 1 time series) Climate change – difference of [u] between any two adjacent runs
Numerical Model Ø A dynamic core of GFDL GCM (symmetric boundary cond. ) Ø R 30 L 10 but zonal wave number 15 Ø Driven by relaxing T toward Te with timescale of 30 days Ø Dissipated by linear friction and 8 th order hyperdiffusion Te(C, H) = Tbase + ΔTe(C, H) C : high-latitude cooling (K/day) H : tropical heating (K/day)
Numerical Model (Cont. ) Ø Total 49 realizations C (0. 00, 0. 17, 0. 33, 0. 50, 0. 67, 0. 88, 1. 00) K/day H (0. 00, 0. 33, 0. 67, 1. 00, 1. 33, 1. 67, 2. 00) K/day Ø Statistics are derived from the last 4500 days of each 5000 -day integration. Data of both hemispheres are used. (C, H)=(0. 17, 1. 67) (C, H)=(0. 17, 0. 33) (C, H)=(0. 83, 0. 33) [u] Single Jet Intermediate Jet Double Jet
[u] : Structure of Westerly Jets Ø Strong C & weak H → Double Jet SJ Ø H ≥ 1. 00 K/day → Single Jet WJ DJ
Internal variability of the jets One-point correlation of 250 -h. Pa [u]' [u] & EOFs SJ Zonal-index (Jet Meander) WJ Transition DJ Poleward Propagation
Time series of PC 1 and PC 2 Correlation PC 1 vs. PC 2 SJ Zonal-index (Jet Meander) WJ Transition DJ Poleward Propagation: i. Correlation between PC 1 & PC 2 is very high ii. Var(EOF 2) is comparable to Var(EOF 1)
Collocates with intermediate- and double-jet Shading γ ≥ 0. 5 Shading χ ≥ 0. 5
Annular mode & Climate change in the mode. I Ø Annular mode : EOF 1 of [u] • [u] is regressed against PC 1 time series, unit of m/s. Ø Climate change : Difference of [u] between two adjacent runs • δ[u]H (0. 50, 1. 00) = [u] (0. 50, 1. 33) - [u] (0. 50, 1. 00) • δ[u]C (0. 50, 1. 00) = [u] (0. 67, 1. 00) - [u] (0. 50, 1. 00) δ[u]H (0. 50, 1. 00) δ[u]C (0. 50, 1. 00)
Predictability of Climate change by Annular mode I. Global measure : pattern correlation between EOF 1 and δ[u] from 150 -950 h. Pa and 10 -80˚ EOF 1 & δ[u]C EOF 1 & δ[u]H Shading correlation ≥ 0. 8 Predictability is always poor in a poleward propagation regime.
Poor predictability of δ[u]H in a zonal-index regime v Annular mode in the model is associated with eddy fluxes. v δ[u]C is associated with eddy fluxes. • Increase of C → enhances extratropical baroclinicity v δ[u]H is associated with both eddy fluxes and mean-meridional circulation. • Increase of H → enhances subtropical baroclinicity and intensifies Hadley circulation v Predictability of δ[u]C would be better than that of δ[u]H.
Summary Structure of Westerly Jet • Strong C & weak H → Double Jet • H ≥ 1. 00 K/day → Single Jet Internal Variability • Strong C & weak H → Poleward propagation (Comparable effect of EOF 2) • Weak C & strong H → Zonal index (Dictated by EOF 1) • Broad transition zone Predictability of Climate change by Annular mode • Dependent on the dominant internal variability • Relative good in a transition regime
Application to the Southern Hemisphere Ø Applied to the SH climate change at equinoctial condition Global warming at SH → ENSO-like tropical heating & enhanced extratropical baroclinicity (Son and Lee 2005 a) → increase of H and C. Ø Structure of the jet Wide range of interannual variability from single- to double-jet states Ø Internal variability Both poleward propagation and zonal index (e. g. , Feldstein 1998; Hartmann and Lo 1998) with γ ≈ 0. 5 and χ ≈ 0. 3 (Son and Lee 2005 b). [u]: structure of the jet SH EOF 1 & δ[u]C SH EOF 1 & δ[u]H SH
Application to the Southern Hemisphere (Cont. ) v Predictability is marginally good in the SH-like parameter regime. v Annular mode may not be useful for understanding paleoclimate change. Slight climate drift to the poleward propagation regime → poor predictability. EOF 1 & δ[u]C SH EOF 1 & δ[u]H SH
Any comment and suggestion are welcome. Thank you! Contact information Seok-Woo Son: sus 141@psu. edu
Dependency of internal variability to the mean flow Ø The meridional radiation of the waves is prohibited if the PV gradient of the ambient flow is sufficiently sharp (e. g. , Hoskins and Ambrizzi 1993) Ø Poleward propagation of westerly anomalies may occur only when the PV gradient is relatively weak and broad. The latitudinal distance over which the value of 250 -h. Pa quasi-geostrophic PV gradient ([q]y) is greater than 60% of its maximum value. Shading for ≥ 35˚.
Prediction of Climate-change ‘Direction’ by Annular mode? Ø Climate change direction (positive or negative phase of annular mode) is determined not by the annular mode but by the nature of external forcing. Ø Climate change associated with H increase (warming at tropics) → negative phase of annular mode (out of phase). Ø Climate change associated with C increase (broadening of extratropical baroclinic zone) → positive phase of annular mode (in phase). [u] (0. 50, 100) - + δ[u]H (0. 50, 100) δ[u]C (0. 50, 100)
Prediction of Climate-change ‘Direction’ ? (Cont. ) Climate change in SH: tropical warming & enhanced extratropical baroclinicity (Son and Lee 2005 a) → increase of H and C. Climate change in SH is in phase with SH annular mode. SH [u] response to global warming SH Annular Mode By the overwhelming effect of enhanced baroclinicity (C) over tropical warming (H) ? Kushner et al. 2001
Predictability of Climate change by Annular mode II. Local measure : latitudinal distance between extrema of EOF 1 and δ[u] at 250 h. Pa • δφC : between EOF 1 and δ[u]C • δφH : between EOF 1 and δ[u]H EOF 1 & δ[u]C (line A) • measured at both subtropics and extratropics δφC A
Ø Weak latitudinal dependency of δ[u]C prediction by annular mode. δφC (low-latitude) Shading δφ ≤ 2˚ δφC (mid-latitude) δφH (low-latitude) δφH (mid-latitude) Ø Poor predictability of δ[u]H in a zonal-index regime is due to the mid-latitudes. Ø Predictability is generally good when γ ≤ 0. 5 or Var(EOF 1) ≥ 2 • Var(EOF 2) Shading γ ≥ 0. 5
Prediction of Climate-change ‘Amplitude’ by Annular mode? II. Local measure : Compare amplitude of 250 -h. Pa |EOF 1| and |δ[u]| at 250 h. Pa EOF 1 & δ[u]C (line A) A δφC
Prediction of Climate-change ‘Amplitude’ by Annular mode? shading: δφC ≤ 2˚ shading: δφH ≤ 2˚ ratio |δ[u]|/|EOF 1| difference (|δ[u]| - |EOF 1|) • Ratios of |δ[u]C| to |EOF 1| are 0. 3 to 0. 8. • Ratios of |δ[u]H| to |EOF 1| are 1. 0 to 2. 5 Ratios vary only by a factor of two! Predictable? No theories yet!
- Slides: 25