Chapter 4 El Ni

1
Chapter 4El Niño and Year-to-Year Climate
Prediction
4.1 Recap of El Niño basics
4.2 Tropical Pacific climatology
4.3 ENSO mechanisms I Extreme phases
4.4 Pressure gradients in an idealized upper layer
4.5 Transitions into the 1997-98 El Niño
4.6 El Niño mechanisms II Dynamics of transition
phases
4.7 El Niño prediction
4.8 El Niño remote impacts teleconnections
4.9 Other interannual climate phenomena and
prospects
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
2
4.1 Recap of El Niño basics
Supplementary Fig. Reynolds SST data set From
chapter 1
Climatology 1982-2001 (C)
Sea Surface Temp. Dec. 1997
Anomaly (Dec.97 SST-Clim.)
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
3
December 1997 Anomalies of precipitationduring
the fully developed warm phase of ENSO
Recap Figure 1.8
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
4
DJF Low-level wind anomalies during the 1997-98
El Niñorelative to the 1958-98 climatology
Recap Figure 1.9
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
5
December 1997 Anomalies of sea level
heightduring the fully developed warm phase of
ENSO
Recap Figure 1.10
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
6
Sea surface temperature climatology - January
4.2 Tropical Pacific climatology
Recap Figure 2.16
Sea surface temperature climatology - July
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
7
Recap Figure 2.13
Precipitationclimatology - January
Precipitationclimatology - July
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
8
Equatorial Walker circulation
Recap Figure 2.14
Adapted from Madden and Julian, 1972, J. Atmos.
Sci., and Webster, 1983, Large-Scale Dynamical
Processes in the Atmosphere
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
9
Pacific in three-dimensions under "Normal"
conditions
Figure 4.1

• Atmosphere
• Trade winds blow across Pacific air rises
  in convergence zone over the warm SSTs in the
  west.
• Ocean
• Thermocline 100m deeper in west sea level 40cm
  higher see 4.4
• Pressure gradient in ocn. (eastward) balances
  wind stress in vert avg

Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
10
Pacific in three-dimensions under "Normal"
conditions
Figure 4.1

• Recall equatorial upwelling wind stress
  Coriolis force either side of equator give
  surface divergence
• Shallow thermocline in east upwelling
  brings up colder water
• Equatorial undercurrent above thermocline flows
  eastward

Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
11
Pacific basin under El Niño conditions
4.3 ENSO mechanisms I Extreme phases
Figure 4.2a

• Warmer SST in east rainfall tends to spread
  east
• Trade winds weaken
• Unbalanced eastward PGF in ocean
  anomalous currents (in vert
• avg through layer above thermocline)
  thermocline deepens in east
• Upwelling brings up water less cold than normal

Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
12
Pacific basin under La Niña conditions
Figure 4.2b

• Cooler SST in east rainfall concentrated in
  west
• Trade winds strengthen
• Westward wind stress exceeds eastward PGF in
  ocean
• anomalous currents along Eq.
  thermocline shallows in east
• Upwelling brings up water colder than normal

Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
13
"It is the gradient of SST along the equator
which is the cause of ... the Walker
circulation. An increase in equatorial easterly
winds is associated with an increase in
upwelling and an increase in the east-west
temperature contrast that is the cause of the
Walker circulation in the first place. ... On
the other hand, a case can also be made for a
trend of decreasing speed ... There is thus
ample reason for a never-ending succession of
alternating trends by air-sea interaction in the
equatorial belt, but just how the turnabout
between trends takes place is not yet clear. 1969
Jakob Bjerknes
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
14
The Bjerknes feedbacks (warm phase)
Figure 4.3

• Positive feedback loop reinforces initial anomaly

Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
15
The Bjerknes feedbacks (cold phase)
Figure 4.3 Supplemental

• Positive feedback loop reinforces initial anomaly

Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
16
The El Niño Pumpkin
Neelin, 2011.
17
Idealized upper ocean layer
4.4 Pressure gradients in an idealized upper layer
Figure 4.4
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
18
Two positions of the thermocline,
indicatingregion of thermocline anomalies
Figure 4.5
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
19
Buoy from the TAO array
4.5 Transitions into the 1997-98 El Niño
Figure 4.6
Courtesy of the Pacific Marine Environmental
Laboratory
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
20
Global tropical moored buoy array
(the original TAO array in the Pacific augmented
by subsequent programs)
Figure 4.7
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
21
The transition into the 1997-98 El Niño warm
phase (Jan. 1997)
Figure 4.8
After figures courtesy of David Pierce, Scripps
Institute of Oceanography.
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
22
The transition into the 1997-98 El Niño warm
phase (Apr. 1997)
Figure 4.8 cont.
After figures courtesy of David Pierce, Scripps
Institute of Oceanography.
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
23
The transition into the 1997-98 El Niño warm
phase (Sep. 1997)
Figure 4.8 cont.
After figures courtesy of David Pierce, Scripps
Institute of Oceanography.
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
24
The transition into the 1997-98 El Niño warm
phase (Jan. 1998)
Figure 4.8 cont.
After figures courtesy of David Pierce, Scripps
Institute of Oceanography.
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
25
The transition into the 1998-98 La Niña cold
phase (May 1998)
Figure 4.9
After figures courtesy of David Pierce, Scripps
Institute of Oceanography.
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
26
The transition into the 1998-98 La Niña warm
phase (Sep. 1998)
Figure 4.9 cont.
After figures courtesy of David Pierce, Scripps
Institute of Oceanography.
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
27
The transition into the 1998-98 La Niña phase
(Jan. 1999)
Figure 4.9 cont.
After figures courtesy of David Pierce, Scripps
Institute of Oceanography.
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
28
Schematic of an equatorial jet
4.6 El Niño mechanisms II Dynamics of transition
phases
Figure 4.10

• deep thermocline high pressure in upper ocean,
  H
• current can flow along Eq. (Coriolis0)
• equatorial jet balance of deep thermocline and
  current anomalies near equator (with PGF
  Coriolis note change in CF with latitude
  crucial)

Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
29
Schematic of an equatorial jet showing that it
canextend itself eastward but not westward
Figure 4.11
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
30

• Currents carry mass affects pressure
• Deep thermocline extends eastward where mass
  added (edge moving eastward Kelvin wave)
• NB something has to continually supply mass in
  the west for the jet to persist

• Shallow thermocline and westward currents also
  give equatorial jet ( just switch sign of anoms)
• Low is extended by removing water ( in the ocean
  upper layer), so also extends eastward

Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
31
Kelvin wave front at the eastern edge of an
equatorial jet
Figure 4.12
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
32
Response of the ocean to a westerly wind anomaly
Re onset and demise of El Niño warm phase
Figure 4.13

• To east of the wind anomalies, equatorial jet
  (Kelvin wave) extends east, deepening thermocline
  (H)
• (recall warms SST)
• To west, inflow of water to jet (in oc. upper
  layer) comes from off the equator (but little
  effect on SST)
• shallow thermocline in west extends westward
  (Rossby wave), as mass transferred to east by jet
• when reaches western boundary, can no longer
  supply mass by extending shallow region
• Weakening of jet extends eastward, ending warm
  phase
• As wind anomalies weaken, shallow thermocline
  extends eastward transition to cold phase

Wind anoms currents
Deep Thermocline anoms
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
33
Response of the ocean to a easterly wind anomaly
Re Onset and demise of La Niña cold phase
(supplementary Fig.)

• Same as Figure 4.13 but anomalies of opposite
  sign

Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
34
ENSO transitions
Meantime, in the West Pacific (subsurface)
Recall feedbacks that strengthen El Niño
Delay no surface effect until
And vice versa
Onset of La Niña cold phase
Figure 4.14
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
35
Forecast of SST anomalies (as three month
averages)
4.7 El Niño prediction

• Forecast of the onset of the 1997-98 El Niño
• From National Center for Environmental Prediction
  climate model
• Data through March 1997 (previous wind stress
  anomalies, ocean subsurface temperatures, SSTs,)
• Data assimilation process includes
  interpolation of sparse observations to all model
  grid points, balancing terms in model
  equations,
• climate model runs forward in prediction mode
  (from April)

Figure 4.15
Courtesy of the National Center for Environmental
Prediction
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
36
Supplementary Figure NCEP Forecast vs.
Observation
Courtesy of the National Center for Environmental
Prediction
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
37
Commonly used index regions for ENSO SST anomalies
Recall Figure 1.5
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
38
Series of forecasts of SST anomalies
averagedover the Niño-3 region of the equatorial
Pacific
Figure 4.16

• Track record of forecasts
• From forecasts made each month, collect all the
  forecast SST anomalies at 3-month lead (i.e.
  after each forecast had gone three months into
  the future), 6-month lead, 9-month lead,
• E.g. March 1997 forecast shown
• Compare each forecast to the SST anomaly that was
  later observed (solid line)
• skill decreases with longer lead still useful at
  9 months

Courtesy of the National Center for Environmental
Prediction
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
39
4.7.1 Limits to skill in ENSO forecasts

• Loss of skill in ENSO forecasts
• Imperfections in the forecast system
• -e.g., model errors, scarcity of input data (can
  be improved, if )
• Fundamental limits to predictability
• -weather unpredictable beyond two weeks (chaos
  theory) slightly different initial conditions
  lead to later weather patterns as dissimilar as
  weather maps chosen at random (except for
  aspects determined by sea surface temperature)
• - weather noise acts like a random forcing
  on slow ocean-atmosphere interaction
• e.g. in the Bjerknes hypothesis, SST gradient
  determines average strength of Tradewinds. But in
  a particular month, storms or other transient
  weather events can cause equatorial Easterlies to
  differ from this, causing a greater or lesser
  change of currents than you would expect from the
  SST anomalies

Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
40
Effects of weather noise on the ENSO cycle
Figure 4.17

• Schematically, random weather events cause cycle
  to depart from the evolution it would otherwise
  have had
• Cumulative effects cause departure from prediction

Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
41
An ensemble of forecasts duringthe onset of the
1998-99 La Niña

• Start coupled model from different ocean initial
  conditions (leading also to changes in atm. )
• Initial differences grow Þ ensemble of prediction
  runs
• Ensemble spread gives estimate of uncertainty
• Spread tends to grow with time (due to weather
  noise coupled feedbacks)
• Ensemble mean gives best estimate

Figure 4.18
Courtesy of the European Centre for Medium-range
Weather Forecasting.
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
42
Supplementary Figure
ECMWF forecast of the 09/10 El Nino from May
2009 with overlaid observations
Courtesy of the European Centre for Medium-range
Weather Forecasting.
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
43
Supplementary Figure
ECMWF forecast from March 2010 predicting
transition to La Niña of 2010
Courtesy of the European Centre for Medium-range
Weather Forecasting.
Courtesy of the European Centre for Medium-range
Weather Forecasting.
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
44
Supplementary Figure
ECMWF forecast from March 2010 with overlaid
observations for verification
Courtesy of the European Centre for Medium-range
Weather Forecasting.
Courtesy of the European Centre for Medium-range
Weather Forecasting.
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
45
Supplementary Figure
ECMWF forecast from Sept. 2010 with overlaid
observations for verification
Courtesy of the European Centre for Medium-range
Weather Forecasting.
Courtesy of the European Centre for Medium-range
Weather Forecasting.
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
46
Supplementary Figure
ECMWF forecast from March 2011
Courtesy of the European Centre for Medium-range
Weather Forecasting.
Courtesy of the European Centre for Medium-range
Weather Forecasting.
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
47
Regions with statistically reliable relation of
precipitation andsurface air temperature to El
Niño and La Niña
4.8 El Niño remote impacts teleconnections
Figure 4.19

• Impact regions change with seasonal climatology

• La Niña similar but opposite sign

Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
48
Patterns of typical response to El Niño
observedfor northern hemisphere winter
Figure 4.20
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
49
Jet stream and storm track changesassociated
with El Niño or La Niña
Figure 4.21
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
50
Schematic of shift of probability distribution of
precipitation, e.g. in Southern California,
during El Niño

• E.g., find value of precip which only 1/3 of
  winters exceed, and ask what fraction of El Niño
  winters exceed it
• Probability of rainy winter enhanced (but far
  from certain)

Figure 4.22
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
51
4.9.1 Hurricane season forecasts
Factors that affect tropical cyclone development
Figure 4.23
NOAA GOES-9 satellite photo of hurricane Linda.
NASA Goddard Space Flight Center. Initial
rendering by Marit Jentoft-Nilsen. Cross-section
follows Emanuel, K. A., 1988, Am. Sci.
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
52
Effect of ENSO on number of Atlantic named
storms (tropical storms and hurricanes) in
July-Oct. each year

• avg 8-9
• Regression
• La Niña 10
• El Niño 6
• But large scatter ( increases w earlier SST)

Figure 4.24
Tang and Neelin, Geophys. Res. Lett., 2004.
Tropical storm sustained winds gt 18 m/s
hurricane winds gt 33m/s (74 mph) Category 5 gt
69 m/s
Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
53
4.9.2 Sahel drought
Recap Figure 2.13, zoomed on Africa
Precipitationclimatology January JulyBox
shows averaging region for Fig. 4.25 (next slide)

• Sahel
• region at margin of African monsoon (seasonal
  movement of convection zones)
• On border with arid regions, just south of
  Sahara desert
• Receives all its rainfall in June-Sept. when
  convection zone moves north

Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
54
Sahel drought annual rainfall anomalies (13-20N,
15W-20E)
Figure 4.25
Data from Hulme, Global Precipitation and Climate
Change, 1994.

• Sahel region has experienced decades of drought
  from 1970s to present, compared to 1950s and
  1960s
• Also has year to year variation
• Three hypotheses
• (i) Land surface change increases albedo. More
  sunlight reflected gives less energy transferred
  from surface to atmosphere to drive convection
• (ii) (Most likely) SST anomalies in Atlantic and
  Indian oceans cause the drought by
  teleconnections
• (iii) Possible anthropogenic contribution by
  aerosols/warming

Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP
55
Supplementary Fig. Lake Chad (Africa) shrinking
due to drought
b.) 12/14/87
a.) 12/08/72
c.) 12/18/2002
LandSat images (US Geological Survey)
56
4.9.3 The North Atlantic Oscillation (NAO) and
annular modes

• Northern and Southern Annular Modes low pressure
  near pole, high at mid-latitudes (positive phase)
  or vice versa
• NAO roughly the Atlantic sector of N. Annular
  Mode
• surface pressure shown extends to stratosphere
• Winds enhance/reduce jet (in pos./neg. phase),
  shifting end of storm track north/south
  impacts European precipitation
• atmospheric origin but includes decadal vbty/trend

Neelin, 2011. Climate Change and Climate
Modeling, Cambridge UP