Decadal climate variability: Societal impacts,
Phenomena, Problems, and Prospects
Vikram M. Mehta
The Center for Research on the Changing Earth System, Columbia, Maryland, USA
¾
What is decadal climate variability?
¾ From Athenian irrigation in 400 BC to Zimbabwe corn in 2000 AD
¾ Are these slowly-evolving physical phenomena or mere wiggles in time series of
observations?
¾ …so many problems in this field, we better find another problem!!
¾ Oh, don’t run away!! There is hope.
¾ Summary
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
What is decadal climate variability?
¾ Periodic or quasi-periodic phenomenon with an approximately 10-20 years period
¾ ”Bursts” or groups of higher frequency phenomena, e.g., El Niño-La Niña, hurricanes and
other tropical cyclones, extreme precipitation or heat events, …., occurring every 10-20 years
¾ Abrupt, in approximately 5-20 years, transitions between climate states
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Periods in decadal climate variability research
The pre-modern period: Statistical Sunspots-climate relationships; 400 BC-1980 AD
z
Irrigation engineer Meton of Athens (c. 400 BC): Connection between Sunspots and climate cause variability
in precipitation
z
Sir William Herschel (1801): Price of wheat in London indirectly controlled by decadal variability of Sunspots
z
Cottage industry in 19th and early 20th centuries to predict weather and climate from Sunspots
z
Disrepute due to correlation reversals, prediction failures, and inability to test hypothesized mechanisms with
models
z
Publications in refereed journals till the 1980s
The modern period:
z
z
z
z
z
z
z
z
z
z
z
First demonstrations of interannual-decadal climate variability with oceanatmosphere models; 1980-1994
Busalacchi and O’Brien, 1980, JPO; 1981, JGR
Lau, 1981: JAS
McCreary, 1983; McCreary and Anderson, 1984: MWR
Philander et al., 1983, 1984,1985: Nature, JAS
Cane and Zebiak, 1985: Science
Hirst, 1986, 1988: JAS
Schopf and Suarez, 1988: JAS
Battisti and Hirst, 1989: JAS
Zebiak and Cane, 1987: MWR
Mehta, 1991, 1992: Clim.Dyn., J.Clim.
Latif and Barnett, 1994: Science
The post-modern period: The topic of this talk: 1994Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Societal Impacts
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Water and civilizations
z
Fresh water the lifeblood of civilizations
z
Development of ancient civilizations near major rivers such as the Nile, the Euphrates-Tigris,
the Indus, the Ganges, and the Yangtze
z
Spurts in the evolution of civilizations on their banks related to extremes of flow in these and
other important river systems driven by climate variability and change
z
Association between availability of fresh water, the rise of civilization and societal well-being
or vulnerability (susceptibility to damage), depends heavily on population and population
density, internal cohesion, infrastructure and the resilience deliberately or inadvertently built
into the society
z
The vulnerability also depends on socio-economic-political-religious dynamics of the society
z
Strains due to these dynamics (sometimes with geopolitical consequences) can be improved or
exacerbated by climate variability and changes
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Impacts of water scarcity on civilizations
[Adapted from: Elhance, A.P.: Hydropolitics in the 3rd World.
United States Institute of Peace, 1999]
Decreased
Precipitation;
reduced
stream flow
Other
natural
phenomena
3
High
mortality/
morbidity
1
9
Migration/
displacement
4
Reduced
quality
of life
2
Water
scarcity
7
Social
unrest
Environ.
degradation
10
Domestic
upheaval
5
Population
growth
Basinwide
human
activities
Vikram Mehta
Diminished
agricultural
production
11
8
Economic
decline
National
insecurity
6
Diminished
industrial
production
The SORCE Science Team meeting; Meredith, New Hampshire
12
Potential
international
conflict
27-29 October 2004
Decadal climate variability and impacts in recent times
z
z
z
z
z
z
z
z
z
Long history of decadal variability in precipitation, surface temperature, river flow,
droughts, …; especially in regions where there are long records (Europe, Asia, North
America)
Decade-long droughts in the North American Great Plains and consequent societal
impacts, including out-migration of people, especially in the 1890s and the 1930s
Decadal variability in precipitation, stream flow, fish catch, and forest fires in the
Pacific Northwest, and in precipitation in the southwest U.S. and Mexico
Socio-economic-political instability in Brazil due to multiyear to decadal droughts in
northeast Brazil
The Sahelian drought in the 1970s-1980s-1990s and consequent societal impacts
Multiyear to decadal droughts in Pakistan and Afghanistan in the 1980s and 1990s
and consequent socio-economic-political upheavals, with geopolitical consequences
Decadal-multidecadal variability in the Indian monsoon rainfall
Decadal-multidecadal variability in numbers of hurricanes and other tropical cyclones
…
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Problems
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Problems in decadal climate variability research
Data
Relatively short time series of instrument observations, especially over and in the oceans; high and stable data
quality required
z
Instrument data quality, especially over the oceans, not well-known; quality of backward-extended instrument
data dubious
z
Observing system changes make identifying physical climate variability from all-inclusive model-assimilated
data difficult
z
Unambiguous interpretation of physical climate variables derived from proxy data difficult
Analysis and interpretation
z
Uncertainty about decadal spectral peaks
z
Research guided by only one analysis technique, e.g. empirical orthogonal functions, confusing and misleading
z
A large number of mechanisms proposed, but no consensus on a small subset and not enough instrument data for
verification of model-simulated variability
z
Pendulum has swung the other way: current mechanisms all about internally-generated decadal variability
Predictability
z
Long-term predictability research stymied by exclusive use of short-term climate prediction models
z
What would be useful to societies if it can be predicted at multiyear to multidecadal lead times with some skill?
Who will use the predicted information and how will it be used?
General
z
The importance of research on geophysical variability at these “civilization” timescales not recognized much in
the scientific community and in the U.S. funding agencies
z
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Phenomena
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
The North Atlantic Oscillation
Bader [Hawaii, 2004]
The positive NAO-Index phase
The NAO Index
SLP difference between Iceland and the
Azores/Lisbon/Gibraltar is averaged and
normalized from November through March
• A stronger subtropical high and a weaker
Icelandic low
• Stronger north-south pressure gradient results
in more frequent and stronger winter storms on a
more northern track
• This leads to warm-wet winters in northern Europe
and cold-dry winters in Canada and Greenland
Vikram Mehta
©Visbeck
©Hurrell
The negative NAO-Index phase
• A weaker subtropical high and a stronger
Icelandic low
• Weaker north-south pressure gradient results in
fewer and weaker winter storms on a more
southern track
• Cold-dry winters in northern Europe and
cold-snowy winters over the US east coast
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Impacts of the NAO: Economy
Bader [Hawaii, 2004]
©PREDICATE
•The annual heating oil consumption
in Norway varies in anticorrelation
with the NAO
•Correlation of NAO with
precipitation results in variability
in hydropower generation
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Arctic Ocean circulation regimes: 1900-2002
Proshutinsky and Dukhovskoy [Hawaii, 2004]*
* Presented in the CRCES-IPRC Workshop on Decadal Climate Variability; Kona, Hawaii; 23-26 Feb. 2004;
sponsored by NASA-Ocean., NSF-Clim. Dyn., and NOAA/OGP; available from http://www.DecVar.org/auditorium.php
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
River discharge anomalies (cubic km. per
year; Shiklomanov et al., 2000, updated)
Proshutinsky and Dukhovskoy [Hawaii, 2004]
AOO
Correlation
between river
runoff and AOO
is 93%.
NAO
River discharge – blue solid and dotted
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Arctic climate since1900
Proshutinsky and Dukhovskoy [Hawaii, 2004]
• SLP decreases over the Arctic
• Ice area, ice thickness, and ice volume decrease
• Air temperature, precipitation, sea level, river discharge,
permafrost temperature increase
• Wind driven ice and water circulation alternates between
cyclonic and anticyclonic
• Climatic indices show more or less stable conditions
before 1940 and then have a well pronounced decadal
variability and trends associated with the increase of
cyclonicity over the Arctic Ocean.
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Atlantic Multidecadal Oscillation
Enfield et al., 2001
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
The Atlantic Multidecadal Oscillation and decadal scale
climatic changes in the Greater Caribbean region
Diaz, Hoerling, and Eischeid [Hawaii, 2004]
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
The tropical Atlantic decadal variability and northeast
Brazil rainfall
Mehta, 1998
z
z
z
e
z
Dominant pattern of tropicalsubtropical Atlantic sea-surface
temperature (SST) variability
between 1881 and 1990 shows
opposite phase anomalies in North
and South Atlantic (Fig. a), with
maximum variances at 15ºN and
15ºS (Fig. b)
The 110-years long SST time series
modulating this pattern (Fig. c) and
its Fourier spectrum (Fig. d) clearly
show decadal and multidecadal
variability
Low-pass filtered rainfall in
northeast Brazil (solid line in Fig. e)
shows opposite phase variability with
respect to a low-pass filtered version
of the SST time series in Fig. c
Oscillations of this north-south SST
gradient (Fig. a) influence the
location of the Inter-Tropical
Convergence Zone
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
The Pacific Decadal Oscillation
Courtesy: Nathan Mantua, Stephen Hare
Univ. of Washington
z
Vikram Mehta
Major changes in northeast Pacific
marine ecosystems have been
correlated with phase changes in the
PDO; warm eras have seen enhanced
coastal ocean biological productivity
in Alaska and inhibited productivity
off the west coast of the contiguous
United States, while cold PDO eras
have seen the opposite north-south
pattern of marine ecosystem
productivity.
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Impacts of the Pacific Decadal Oscillation on the Pacific
Northwest forests and salmon catch
Courtesy: Climate Impacts Group, Univ. of Washington
Low frequency variability in Pacific Northwest tree growth
compared to the average winter, five-year running-mean, PDO.
For the period 1922-1995, the number of forest fires burning more than
1,000 acres was higher during warm phase of the PDO (blue bar) than cool
phase of the PDO (white bar) in nearly all of the Pacific Northwest’s
national forests.
Vikram Mehta
Selected Pacific salmon catch records with PDO
signatures. Black (gray) bars denote catches that are
greater (less) than the long-term median. The dotted
vertical lines mark the PDO polarity reversal times in
1925, 1947, and 1977.
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Impacts of the Pacific Decadal Oscillationon on the U.S.
temperatures and Columbia River stream flow
Courtesy: W. Patzert, JPL; and Climate Impacts Group, Univ. of Washington
z
High PDO index in February-April is followed
by above-average summer temperatures in the
western and southeast U.S. and below-average
summer temperatures in the northeast U.S.
The impact of ENSO and PDO on Columbia River summer streamflow at The Dalles, Oregon for 1900-1999. The
horizontal lines show average streamflow over each of the PDO epochs (cool: 1900-1925, warm: 1925-1945, cool: 19451977, warm: 1977-1995). The red dots are El Niño years, the blue dots are La Niña years, and the green dots are ENSO
neutral years. The figure shows "naturalized" streamflow (i.e., with the effects of the dams numerically removed) for
April-September of each year.
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Importance of the Indo-Pacific Warm Pool in the global climate
system
z
z
z
z
z
z
Warmest surface ocean water on the Earth
Annual-average SST≥28°C from
approximately 90°E-180°, 20°S-20°N;
pronounced annual cycle
Saturation vapor pressure non-linearly
related to SST⇒dramatic increase in
atmospheric moisture content and
convection when SST ≥ ≈28.5°C
THE major source of heat for the global
atmosphere
Numerous studies of possible mechanisms
of maintenance of time-average state of
the IPWP (Ramanathan and Collins, 1991;
Wallace, 1992; Fu et al., 1992; Hartman
and Michelsen, 1993; Waliser and
Graham, 1993; Waliser, 1996; Sud et al.,
1999)
Natural variability has not received much
attention; influences ENSO and marine
ecosystems at interannual timescales
(Delcroix et al., 2000; Picaut et al. 2000)
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
The Warm Pool Oscillation
Climatology + low-pass filtered sea-surface temperatures
from SODA: 1950-2001
Mehta and Mehta [Hawaii, 2004]
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
850-mb wind anomalies associated with the Warm Pool
Oscillation: 1949-1998
Mehta and Mehta [Hawaii, 2004]
Warm Pool less warm
Warm Pool warmer
a
c
b
d
2 m/s
z
z
z
z
Low-pass (≥8 years) wind anomalies; composited according to the phase of the low-pass filtered, western Pacific Warm
Pool SST
Note the equatorial westerlies (1-2 m/s) when Warm Pool less warm(Fig. a) and in the following year (Fig. b)
Note the equatorial easterlies (1-2 m/s) when Warm Pool warmer(Fig. c) and in the following year (Fig. d)
Also, note coherent mid- and high-latitude wind anomalies in both phases, especially over the North Pacific and the North
Atlantic, confirming previous results by Mehta et al. (2000) and Hoerling and Hurrel (2002)
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Association between global upper-ocean temperature and solar
irradiance at decadal-interdecadal timescales
White, Lean, Cayan, and Dettinger; 1997
a
z
c
z
z
Vikram Mehta
b
(a) First EOF and time series of 9-13 years filtered
bathythermograph SST anomalies and solar irradiance
anomalies from 1955 to 1994; first and second SST EOFs
“explain” 50% of filtered variance; SST lags solar
irradiance by ~1-2 years
(b) First EOF and time series of 18-25 years filtered
GISST SST anomalies from 1901 to 1990; “explains”
87% of filtered variance; SST lags solar irradiance ~0-4
years
(c) Correlation coefficients, with confidence intervals, as
a function of depth between filtered upper-ocean
temperature and solar irradiance anomalies from 1955 to
1994; 9-13 years on the left and 18-25 years on the right
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Influence of solar irradiance on the Indian monsoon rainfall and
Niño3 SST at decadal-multidecadal timescales
Mehta and Lau, 1997
z
z
z
z
z
Vikram Mehta
(a) All-India rainfall and Niño3 SST
index
(b & c) Decadal variations in
correlation coefficients between the
two
(d) Low-pass (≥ 24 years) filtered
All-India rainfall (shaded; std. dev.
units) varies coherently with lowpass filtered solar irradiance (dashed
line; 2* W/m2); this can be seen even
in unfiltered data (solid line).
(e) Low-pass filtered Niño3 SST
index (shaded; °C) shows less
coherent variations and nearlyopposite phase variation with lowpass filtered solar irradiance.
Differential heating/cooling of South
Asia and the oceans around it can
strengthen/weaken the monsoon, but
one or more positive feedback
processes may be necessary because
irradiance changes are small
eventhough persistent for one or
more decades.
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Association between the decadal-multidecadal WPO and
interannual ENSO (1950-2001): Dec-Jan-Feb sea surface
temperature and wind stress patterns
Mehta and Fayos, submitted, 2004a
9 events
La Niña 2 events
El Niño
WP
warmer
15 events
8 events
WP
less
warm
Approximately same number but
less warm El Niño events when WP
warmer at decadal-multidecadal
timescales.
Warmer El Niño events when WP
less warm at decadal-multidecadal
timescales; Indian Ocean
anomalously warm, mid-high latitude
ocean and atmosphere anomalies.
Warm
minus
less
warm
N/m2
ºC
Vikram Mehta
El Niño and La Niña events in highpass (≤7 years) filtered, DJF SST
and surface wind stress data: SODA
1950-2001; composited according to
decadal-multidecadal phase of the
WP SST
Very few La Niña events when WP
warmer; many more La Niña events
when WP less warm at decadalmultidecadal timescales.
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Association between the decadal-multidecadal WPO and
interannual ENSO (1950-2001): Dec-Jan-Feb sea-level
pressure and lower and upper troposphere zonal wind patterns
Mehta and Fayos, submitted, 2004b
SLP (mb) and u850 (m/s)
9 events
SLP (mb) and u200 (m/s)
WP
WP
warmer
warmer
15 events
8 events
Weak SLP and wind patterns in
the tropical Pacific when WP
warmer at decadal-multidecadal
timescales; weak global response
to weak El Niño (see previous
slide)
WP
less
warm
Warmer
minus
less
warm
SLP shading
mb
Vikram Mehta
El Niño events in high-pass (≤7
years) filtered, DJF SLP and wind
data from NCEP-NCAR reanalysis
1950-2001; composited according
to decadal-multidecadal phase of
the WPO
zonal winds contour
Strong SLP and wind patterns in
the tropical Pacific when WP less
warm at decadal-multidecadal
timescales; strong mid-high
latitude SLP and upper
tropospheric wind response to
strong El Niño (see previous
slide)
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Long-term variability of frequency and intensity of
Atlantic hurricanes
T.N. Krishnamurti (pers. comm.)
Atlantic cyclone index (1886-1991)
Interannual-decadal variability of Atlantic hurricane frequency simulated by the FSU global model
14
12
10
8
Observed
6
Forecast
4
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
1995
1994
1993
1992
1991
1990
1988
1987
1986
1985
1984
1983
1982
1981
1980
1979
0
1978
2
27-29 October 2004
North American drought reconstruction from long tree-ring records
Cook [Hawaii, 2004]
WESTERN USA DROUGHT TREND: 1900 - 2003
PERCENT AREA OF THE USA AFFECTED BY DROUGHT (PDSI<-1) BASED
ON RECONSTRUCTIONS FROM LONG TREE-RING RECORDS
100
y = -329.72 + 0.185x R= 0.232* p<0.05 2-tailed
80
A. ALL USA PDSI <-1 DAI
100
LESS CERTAIN
DAI<-1
DAI
80
60
60
40
40
20
20
0
0
B. WEST USA PDSI <-1 DAI
100
80
DAI
1900
MEDIEVAL "DRY" PERIOD?
40
20
z
0
C. EAST USA PDSI <-1 DAI
100
FROZEN GRID FIDELITY LIMIT
DAI
80
60
40
20
z
<50%
COVERAGE
0
800
900
1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000
YEAR
Vikram Mehta
1940
1960
1980
2000
YEAR
z
60
1920
z
Irregular decadal-multidecadal
variability in Drought Area Index
over the last 1000 years in the U.S.
Both the full grid (“changing
observing system”; blue line) and the
frozen grid (“constant observing
system”; red line) show this
variability
Increasing tendency and decadal
variability of droughts in the western
U.S. in the last 100 years
Entering a new period of increased
aridity in the western U.S.?
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
The Virtual Center for Decadal Climate Variability
http://www.DecVar.org
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Summary
z
Decadal-multidecadal climate variability impacts societies on civilization timescales
and is one of the forces guiding the course of history.
z
Many patterns of decadal climate variability; not obvious that all of them independent
and truly oscillatory
z
Many mechanisms proposed, all but one invoke internal variability in the coupled
climate system; the oceans’ role crucial
z
Success in understanding and prediction of short-term climate variability and long-term
climate change, especially at subcontinental to state/province scale, depends crucially
on success in understanding and prediction of decadal-multidecadal climate variability
z
Major problems, including insufficient instrument data, largely unknown quantitative
impacts on societies, largely unknown predictability
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004
Prospects
z
z
z
z
z
z
z
z
z
z
Decadal variability research in danger of becoming parochial, as in the premodern period
Must not limit our search for causes of decadal climate variability to internal
processes only
Variability of external forcings perhaps very important in exciting internal
variability; interplay between the two
Decadal climate predictability a function of external forcings and internal
variability
Decadal variability very important even if greenhouse gas emissions are
limited in the future
Growing research community
The importance of research on geophysical variability at these civilization
timescales not recognized much in the scientific community and in the U.S.
funding agencies
Systematic, permanent observing system required to provide high-quality
observations, including forcings
Systematic and large-scale effort required to assess impacts in various
societal sectors
Next Decadal Climate Variability Workshop in October 2005 somewhere in
the Maryland to West Virginia mountains
Vikram Mehta
The SORCE Science Team meeting; Meredith, New Hampshire
27-29 October 2004