Nonlinearity of atmospheric response to ENSO and MJO Hai Lin Recherche en Prévision Numérique, Environment and Climate Change Canada BIRS Workshop 17w5061, November 19-24, 2017 Banff, Alberta

Outlines Extratropical response to El Nino and La Nina Observations Numerical experiments Extratropical response to MJO

ENSO Leading source of seasonal forecasting skill in the extratropics → El Niño & La Niña Is the extratropical response to El Niño a mirror image of that to La Niña? Why?

Composite of T2m anomaly for 5 El Nino winters and 5 La Nina winters Hoerling et al. 1997, JCLIM

Composite of Z500 anomaly for 5 El Nino winters and 5 La Nina winters Hoerling et al. 1997, JCLIM

Composite of SST anomaly for 5 El Nino winters and 5 La Nina winters Hoerling et al. 1997, JCLIM

Composite of precip anomaly for 5 El Nino winters and 5 La Nina winters Hoerling et al. 1997, JCLIM

Explanation for extratropical nonlinearity One possible explanation is the nonlinear relationship between SST and precipitation in the tropics: Positive SST anomalies can initiate deep convection within the core of the east Pacific cold tongue, whereas negative SST anomalies will have their largest effect in the west Pacific warm pool region. Any other dynamical reason?

Numerical experiments Primitive equation model (Hall 2000) Specified tropical diabatic heating anomaly similar to ENSO, Fixed location 61 experiments with different heating amplitudes from negative to positive 30 members of 120 day integrations for each amplitude

Ensemble mean Z550 anomaly response

Dominant response patterns

Transient response Three sets of experiments: Each set has 50 integrations from different initial conditions Each integration lasts 30 days Daily ensemble difference → response Control run ─ climatological forcing; El Niño run ─ climatological forcing + tropical heating anomaly La Niña run ─ climatological forcing + tropical cooling anomaly The same 50 initial conditions for all three sets of experiments, so that the only difference is the thermal perturbation.

Forcing anomaly Only in temperature equation Maximum at 350 mb

Day 1

Day 2

Day 3

Day 4

Day 5

Day 6

Day 7

Day 8

Day 9

Day 10

Day 11

Day 12

Day 13

Day 14

Days 16-30

W vector Days 16-30

ω’ = - h’ / [α – cp(∂T/ ∂p + ∂T’/ ∂p)] Tropical circulation Why La Niña has a stronger response? →Stronger vertical motion due to changes in static stability _ ω’ = - h’ / [α – cp(∂T/ ∂p + ∂T’/ ∂p)] In the center of response, adiabatic process tends to balance the diabatic heating anomaly El Nino: ∂T’/ ∂p < 0 La Nina: ∂T’/ ∂p > 0

Asymmetric development of tropical waves To the west of thermal forcing: El Nino, upper troposphere easterly  suppress Rossby wave development La Nina, upper troposphere westerly  favors Rossby wave development In the center of response, adiabatic process tends to balance the diabatic heating anomaly

Feedback from transients Modified storm track Vorticity flux convergence by transients

Days 16-30 Anomaly of Z550 rms for transients El Niño La Niña

Days 16-30 Z550 tendency by transients El Niño La Niña

MJO Leading source of skill for subseasonal predictions Is there difference between extratropical response pattern to +MJO and that to -MJO?

Composites of tropical Precipitation rate for 8 MJO phases, according to Wheeler and Hendon index. Xie and Arkin pentad data, 1979-2003

Lagged probability of the NAO index Positive: upper tercile; Negative: low tercile Phase 1 2 3 4 5 6 7 8 Lag 0 +45% −42% Lag +1 +47% −46% Lag +2 +50% +42% −41% Lag +3 +48% −48% Lag +4 −39% Lag +5 (Lin et al. JCLIM, 2009)

Z500 anomaly composites following MJO phase 3 and 7

Wintertime North Atlantic weather regimes From Cassou (2008)

Observed nonlinearity

Dominant tropical subseasonal OLR patterns

MJO forcing structure

Experiments with MJO forcing Three sets of experiments: Each set has 360 integrations from different initial conditions Each integration lasts 20 days Daily ensemble difference → response Control run ─ climatological forcing; +MJO run ─ climatological forcing + MJO -MJO run ─ climatological forcing - MJO The same 50 initial conditions for all three sets of experiments, so that the only difference is the thermal perturbation.

Z500 ensemble mean response

Spread Z500 response

Z500 response to +MJO

Z500 response to -MJO

STRM200 response to +MJO STRM200 response to -MJO

Summary Asymmetric response to El Nino and La Nina In addition to the nonlinear relationship between SST and precipitation, there are dynamical mechanisms. Mechanisms include: 1) changes in static stability; 2) aymmetric development of tropical waves; and 3) feedback from synoptic-scale transients. The extratropical response to MJO is also nonlinear A better representation of such nonlinearity in numerical models would be helpful for subseasonal to seasonal predictions.