11

Decadal modulation of the ENSO impact on the Hydroclimate over

the United States

Kingtse MoClimate Prediction Center

NCEP/NWS/NOAA

22

All drought indices indicate

Cold ENSO=> Drought overthe Great Plains, Southwest and Gulf states

Composites of drought indeces

33

QuestionsQuestions

The impact of ENSO does not always obey The impact of ENSO does not always obey the ‘composites’. WHY? the ‘composites’. WHY?

Each ENSO is differentEach ENSO is different ENSO impact on regional climate is ENSO impact on regional climate is

modulated by decadal modesmodulated by decadal modes

44

Decadal modesDecadal modes TrendsTrends Atlantic Multi decadal oscillation (AMO) Atlantic Multi decadal oscillation (AMO) Pacific Decadal Oscillation (PDO)Pacific Decadal Oscillation (PDO) PDO Barnett and Cayan (1998), Gershunov PDO Barnett and Cayan (1998), Gershunov

and Barnett (1998)and Barnett (1998)

Decadal: fluctuations with time scales longer than 6 yrs

55

Observational DataObservational Data100 yr data set is not long enough100 yr data set is not long enough

P and TsurfP and Tsurf Monthly Precipitation (P) and surface temperature

(T) over the United States from 1915-2006 (U. of Washington)

SST and SLP SST- Smith reconstructed SSTs 1915-2006 SLP – Trenberth and Paolino (1987) Anomaly- departure from monthly mean

Climatology from 1915-2006

66

Atlantic Multi decadal OscillationAtlantic Multi decadal Oscillation

AMO warm 1930-1959,1992-2006

AMO cold 1915-1925,1965-1990

AMO composite warm-cold phase

Decadal influence is small.

You can filter to get clearer signal, but the percentage of variance is small, so the net influence is small.

77

AMO influence through ENSOAMO influence through ENSOComposite of SPI6 with all seasons together for different phase of the AMO

Mo et al. 2009

88

Trends Trends Composites were formed for each season for Composites were formed for each season for

(a) 1915-1960 and (b) 1962-2006(a) 1915-1960 and (b) 1962-2006 ENSO – Nino 3.4 index over 0.8 std (less -0.8 ENSO – Nino 3.4 index over 0.8 std (less -0.8

std)std) Results are given as composites: warm –cold Results are given as composites: warm –cold

weighted by the number of events weighted by the number of events Statistical significance is determined by the Statistical significance is determined by the

Monte Carlo methodMonte Carlo method

99

Tsurf:

Strongest signal is in winter and Spring

In the recent decades

1. Cooling over the Southeast, the Gulf States has been weakening .

2. Warming over California and cooling over the Plains occur in Spring

Surface temperature ENSO composites warm-cold

1010

ENSO-TENSO-TIt used to be…..

1. Warm ENSO cold over the South and warming over the North.

2. Things are changing

• Cooling over the South has been weakening in winter.

• For California and the Southwest, spring is warmer in the recent decades.

• The great Plains is cooler in spring

1111

ENSO-P

1. Strongest impact is in winter

2. In the recent decades, influence on California increases

3. Over the Pacific Northwest, impact comes in Fall

1212

Changing in SSTAsChanging in SSTAs

Ref: Yeh et al. (2009) , Kug et al. (2009), Kao and Yu (2009), Ashok et al. (2007)

Warming centered

at the eastern Pac

Warming centered

At the Central Pac

1313

SLP anom compositesSLP anom composites

1414

ENSO teleconnections are changingENSO teleconnections are changing The impact of ENSO on P and Tsurf over the United The impact of ENSO on P and Tsurf over the United

States is not stationary.States is not stationary. More warm ENSO events in the recent decadesMore warm ENSO events in the recent decades Warming over the Northern U. S. and cooling over Warming over the Northern U. S. and cooling over

the Southeast and the Gulf states are weakeningthe Southeast and the Gulf states are weakening Impact over the Southwest and California intensified Impact over the Southwest and California intensified

in the recent decades (Warm=> wet, Cold=> drought)in the recent decades (Warm=> wet, Cold=> drought) Impact over the Pacific Northwest comes early in Impact over the Pacific Northwest comes early in

Fall. (Warm= > drought, cold=> wet)Fall. (Warm= > drought, cold=> wet)

With such small sample, are these real???

1515

Two kinds of ENSOTwo kinds of ENSO

Yeh et al.2009(nature)

X

x

Eastern PAC

Central Pac

1616

Composites of EP & CP ENSOComposites of EP & CP ENSOP Tair

C Pac

7 events

E Pac

21 events

Events were taken from Yeh et al. (2009)

•Different type of ENSO contributes to the variability of impact

•If models are able to capture two types of ENSO and impact, then they can be used for diagnosing phycical mechanisms

1717

IPCC runs (focus on winter)IPCC runs (focus on winter) 2020thth Century simulations from 1880 to 2000 runs Century simulations from 1880 to 2000 runs

(20C3m)(20C3m) Models:Models: GFDL_CM2.1 (700yrs)GFDL_CM2.1 (700yrs) GFDL_CM2.0 (412 yrs)GFDL_CM2.0 (412 yrs) MicroC3.2-medres (453 yrs)MicroC3.2-medres (453 yrs) UKMO_HadCM3 (280 yrs)UKMO_HadCM3 (280 yrs) MPI_Echam5 (814 yrs)MPI_Echam5 (814 yrs) These are simulations with realistic ENSO (Meehl et These are simulations with realistic ENSO (Meehl et

al. 2007, Oldenborgh et al 2005)al. 2007, Oldenborgh et al 2005)

Data from the Earth system grid WCRP CMIP3

1818

IPCC (JFM)IPCC (JFM)1. For each run, compute

anomalies as the departure from the model mean

2. Form SSTA Nino 34 and Nino 3 indices

3. Select events when both indices are greater than 0.6-0.8 standard deviations

4. EP events when Nino34> Nino3 + with a clear Maximum over the eastern Pacific, to 0.3

5. CP events when Nino3>Nino34 +with a clear maximum over the central Pacific

1919

Ensemble ipcc modelEnsemble ipcc model Eastern Pac Central Pac

Eastern Pac Central Pac1. Tsurf : more

north-south warming/ cooling for the E-PAC and East-west contrast for the C-PAC

2. Cooling over the South weakens for the C –PAC

3. More rainfall over California and the Southwest and less for the Southeast for the C-PAC

2020

Physical Mechanisms

Eastern Pac:

1. Positive P anomalies are located over the area extending from 160W to 90W.

2. Walker circulation puts less rainfall over northern Brazil

3. The weakening of Hadley circulation implies more rainfall over the Southeast.

Central Pac:

1. Positive P anomalies are located over the deadline

Physical mechanims were examined using the GFDL_CM2.1 model runs

Precip (shaded), Chi wind (vectors)

2121

GFDL CM2-1GFDL CM2-1

E-PAC

1. the North American jet extends to the Atlantic

2. Strm 200 –more zonal structure

Central Pac

1. jet extends to the west coast

2. Pacific-North American wavetrain

2222

Decadal changes of ENSO impactDecadal changes of ENSO impactFor forecast and monitoring:• Impact on the U. S. depends on both ENSO pattern as well as

magnitudes.

• Pay attention to the precip anomalies in the Tropics.

• Different tropical convection pattern different wind and circulation responses different impact

• These changes are not linear. The OCN magic may not work