-
Monsoon: Its Annual Cycle, Natural
Variability, Trend and Future Change
Tim Li
University of Hawaii
1. Classic monsoon definition
2. Monsoon onset and seasonal progression
3. Intra-seasonal (sub-seasonal) oscillation
4. Inter-annual (year-to-year) variation: relationship with
ENSO
5. Monsoon inter-decadal variation
6. Trend and future change under global warming
-
1. Monsoon basics
-
Original meaning of monsoon:
derived from the Arabic word
for season
Character: Seasonal reversal
of the wind direction
In JJA, heated land low
pressure cyclonic flow
(due to Earths rotation),
northward cross-equatorial
flow (due to land-ocean
thermal contrast)
In DJF, cold land mass
high pressure
What is Monsoon?
-
JA-JF
Seasonal differential precipitation and winds (JA-JF)
IM
EAM
WNPM
WNPM
IM
EAM
Wang, Clemens and Liu 2003
Top: domains for
three sub-monsoon
systems
IM: westerly, north-
south T gradient
EAM: southerly, east-
west T gradient
WNPM: hemispheric
asymmetric SST
gradient
Bottom: area-averaged
rainfall evolution
WNPM strongest, with a peak
phase lag to IM and EAM.
-
Why Monsoon is important: Precipitation and its social
relevance
80 85 90 95 100 105 110 115 120 125 130
25
30
35
40
45
50
80 85 90 95 100 105 110 115 120 125 130
25
30
35
40
45
50
8
7
6
5
80 85 90 95 100 105 110 115 120 125 130
25
30
35
40
45
50
80 85 90 95 100 105 110 115 120 125 130
25
30
35
40
45
50
80 85 90 95 100 105 110 115 120 125 130
25
30
35
40
45
50
80 85 90 95 100 105 110 115 120 125 130
25
30
35
40
45
50
80 85 90 95 100 105 110 115 120 125 130
25
30
35
40
45
50
500
300
150
80 85 90 95 100 105 110 115 120 125 130
25
30
35
40
45
50
80 85 90 95 100 105 110 115 120 125 130
25
30
35
40
45
50
50
The month of maximum precipitation
The precipitation during JJA
From Gadgil (2003)Monsoon
zone
-
5/21
5/11
5/01 6/15
7/20 8/10
6/21
7/01
9/15
6/01
6/11
6/21
7/11
5/11
6/01
6/01
5/21
7/01
7/11
5/21
4/21
6/11
Wang and LinHo 2002(Climatology 1979-2001)
-
7
Seasonal march of East Asian summer monsoon:
- a stepwise northward advance
During the period from early May to mid May, the South China Sea
monsoon onsets.
The monsoon rain then progresses northward to the Yangtze River
valley in early to mid June,
and finally penetrates northern China (3441N) in mid July.
The rainy season in northern China generally lasts for one month
and ends in the early or middle part of August.
From the end of August to early September, the monsoon rain belt
rapidly moves back to southern China.
-
Li, K. et al. 2013, Clim. Dyn.
Time-latitude sections of
7-day running mean OLR
(shaded, W m-2, with
values lower than 220 are
plotted) and surface wind
(vector, m s-1) fields
along 85E-95E.
-
Time-latitude section of composite OLR (shaded, W m-2), rainfall
(contour,
mm day-1), and surface wind (vector, m s-1) fields along
85-95E.
The first-branch northward-propagating ISO leads to the monsoon
onset!
-
Composite evolution of first-branch northward-propagating
ISO
-
Distribution of the background
convective instability field during
the FNISO phase (from day -10
to day +10)
winter vs. summer
-
Background convective instability distribution
(a) Time-latitude section of the background convective
instability field (shaded, K) and the OLR
perturbation associated with FNISO (contour, Wm-2) along
80-100E; (b) Time evolution of
and its partial contributions due to the temperature (red) and
moisture (green) changes. se
-
The warmest SST occurs in late April over BoB !
-
Monthly TC number
during 1981-2009 in
BoB
Composite map of the FNISO over the BoB. Purple dots denote the
time (relative to the monsoon onset time) and
latitude of intense TC (Category 4 or 5) when it reached its
maximum intensity. Green dots denote the genesis time and
latitude of these super cyclones. (The OLR and wind fields are
averaged over 85E-95E. Y-axis is latitude and x-axis is
relative time to monsoon onset in BoB.
-
Contrast of the monsoon rainfall over India and China:
Indian:
1. Steady rainfall zone along the west coast due to the
topography
effect.
2. Strong inter-annual variations over the monsoon zone, due to
the
synoptic/supersynoptic disturbance from the northern BoB
China:
1. Quasi-Steady wave propagation from the south to north,
zonal
band structure
2. Much larger amplitude of the inter-annual variations along
three
bands
-
16
Asian summer monsoon components
(Figure from Yihui Ding)
-
2. Monsoon natural variability
I: intra-seasonal oscillation
-
Monsoon Intraseasonal Oscillation In Boreal Summer
-
Time series of daily precipitation rate
estimates averaged over 1015N, 7580E
for JunSep 1987 (mm day-1).
Timelatitude section of daily precipitation rate
estimates along 7580E for JunSep 1987 and
1988. Contour interval is 5 mm/day with the
first contour at 5 mm/day.
From Lawrence and Webster (2001): Interannual Variations of the
Intraseasonal
Oscillation in the South Asian Summer Monsoon Region. J.
Climate
-
Classic Madden-Julian Oscillation (MJO)
0 day
5 day
11 day
16 day
22 day
28 day
34 day
40 day
Eastward
Madden and Julian, 1972
-
Observed
Horizontal
Structure of
MJO:
Kelvin-
Rossby wave
couplet with
BL friction
leading
convection
Hendon
and Salby
1994
C
C
A
A
-
Sperber and Slingo 2003
Observed Vertical Structure of MJO:
BL Convergence leads Convection
-
The Life Cycle of the Madden Julian Oscillation (Northern
Winter)
-
Regression maps are calculated with 25-80-day band-pass filtered
OLR and 850-mb wind fields against filtered OLR time series at the
reference box. Only anomalies that are statistically significant
are plotted.
Boreal summer ISO eastward and northward propagation
-
Real-time Multi-variable MJO (RMM) index
(Wheeler and Hendon 2004)
EOF1variance: 12.8%
EOF2variance12.1%
OLRsolidu850dashu200dot
OLRsolidu850dashu200dot
EOF1Maximum MJO convection over maritime continent,
first-baroclinic zonal wind structure
EOF2maximum convection over western Pacific
2001
-4
-3
-2
-1
0
1
2
3
4
-4 -3 -2 -1 0 1 2 3 4
RMM1
RMM2
1
2 3
4
5
67
8
W. Pacific
W. HemisphereMaritime
cont.
Indian Ocean
-
Impact of MJO on summer
circulation ad rainfall in East Asia
-
WWB
Asian Monsoon
Australian
MonsoonIOD
Global Impacts of MJO
West Africa
Monsoon
North America
Goswami et al (2003)
Maloney and Hartmann (2000)
McPhaden (1999)
dry Kelvin WavesEl Nino
-
Propagation tendency
(vector) and variance
(shaded) of summer OLR
(MJJASO)
The length of vectors
represents the magnitude
of the lagged correlation
coefficients
10-20 days
20-70 days
Quasi-biweekly oscillation (QBWO) vs. MJO
-
South China Sea (110o-120oE,10o-20oN)
Lagged correlation maps (10-20 days)
-
South China Sea (110o-120oE,10o-20oN)
Lagged correlation maps (20-70 days)
-
South China Sea
(110o-120oE,10o-20oN)
Lagged correlation maps (10-20 days)
Day -7
Day -5
Day -3
Day 0
-
South China Sea (110o-120oE,10o-20oN)
Lagged correlation maps (20-70 days)
Day -25
Day -20
Day -15
Day -10
Day -5
Day 0
-
Bay of Bengal
(85o-95oE,10o-20oN)
Lagged correlation maps (10-20 days)
Day -6
Day -4
Day -2
Day 0
-
Bay of Bengal (85o-95oE,10o-20oN)
Lagged correlation maps (20-70 days)
Day -25
Day -20
Day -15
Day -10
Day -5
Day 0
-
100%
50%
Normalized Useful Predictability (%)
Wea
ther
ISV
EN
SO
PD
O
AC
C
WCRP/WWRP Seamless Prediction
Courtesy of Duane Waliser
-
Issue on monthly forecast products
November monthly mean
rainfall anomaly is near zero.
Is the zero rainfall forecast
useful and benefit to
public/society?
Two predictability sources for
monthly forecast:
Boundary forcing (e.g., SST)
Initial value problem (MJO)
For latter, pended mean
forecast is needed at 5-30-
day lead?
-
350525mm1525mm150mm55
=2 +1.5 1 0.2 0.04 +0.01
A Spatial-Temporal Projection Model for rainfall prediction in
Fujian
-
Non-filtering method for real time use
For real-time forecast, time filtering is not applicable to
extract ISO signal.
A simple non-filtering method was developed:
(1) Remove mean and annual cycle by subtracting 90-day low-pass
filtered
climatological data.
(2) Remove the effects of interannual, decadal variability and
trend (signals
with the period larger than 60 days) by subtracting the mean of
the last 30 days.
(3) Remove the effect of synoptic-scale disturbances (signals
with the period
smaller than 10 days) by taking the mean of the last 5 days.
AC
X ' X X
30d
X '' X ' X '
5
, is the ISO variability with the period around 10-60 dayd
* *X X '' X
-
Non-filtering method to extract 10-60-day signal
Raw heavy rainfall index in Fujian
Extracted data
The non-filtering method could reasonably extract the 10-60-day
intraseasonal
component.
-
filtered (shad) vs. non-filtered (cont) DJF MJO_U850
1990-2009 Pattern corr. coef. = 0.81
RMSE = 1.6 m/s
-
Predictand Y (t)Normalized heavy rainfall index
Each predictor X (i, j, n, t)Normalized large-scale fields
Spatial-Temporal Coupled Pattern
t
COV(i, j, n)= Y(t) X(i, j, n, t)
S T Coupled Pattern Projection
P
i, j, n
X (t)= COV(i, j, n) X(i, j, n, t)
Transfer FunctionF PY (t) = X (t) +
Fitting
Procedure
Rainfall Forecast YF (tp)
ForecastProcedure at tp
X (i, j, n, tp)
Xp(tp)
Six predictors:
OLR, U850, U200,
H850, H500, H200
STPM Procedure
1. Normalize predictand Y and predictor field X.
2. Construct spatial-temporal coupled co-variance patterns for
the region where Y and X are significantly correlated.
3. The coupled pattern projection is obtained by multiplying the
co-variance field with each predictor.
4. Transfer function is constructed with a linear regression
method.
5. Heavy rainfall forecast is performed based on the coupled
pattern projection and transfer function.
i, j : spatial gridn: preceding n pentads(n=6 in this study)
Spatial-Temporal Projection Method (STPM)
-
Independent forecast skill for 2008-2012based on TCC (models
were built based on 1996-2007 training period,
total forecast:90points, 95%sig=0.21)
lead+5d lead+10d lead+15d lead+20d lead+25d lead+30d
STPM_OLR 0.26 0.3 0.25 0.22 0.17 0.19
STPM_U850 0.2 0.25 0.21 0.16 0.11 0.15
STPM_U200 0.25 0.27 0.27 0.26 0.18 0.2
STPM_H850 0.28 0.35 0.34 0.3 0.22 0.25
STPM_H500 0.24 0.29 0.3 0.3 0.18 0.21
STPM_H200 0.22 0.28 0.2 0.23 0.14 0.17
SVD_OLR 0.3 0.32 0.28 0.28 0.2 0.21
SVD_U850 0.31 0.32 0.29 0.27 0.2 0.21
SVD_U200 0.29 0.32 0.25 0.3 0.21 0.19
SVD_H850 0.26 0.33 0.27 0.26 0.15 0.18
SVD_H500 0.31 0.32 0.26 0.3 0.15 0.18
SVD_H200 0.21 0.25 0.17 0.24 0.14 0.15
SVD_index 0.44 0.47 0.32 0.27 -0.03 0.14
MVR 0.23 0.33 0.29 0.27 0.26 0.27
ens_STPM 0.29 0.34 0.31 0.29 0.19 0.23
ens_SVD 0.36 0.4 0.33 0.33 0.17 0.22
-
2013
-
3. Monsoon natural variability
II: inter-annual variation
-
Webster et al. 1998
Observed negative correlation between Indian monsoon and El
Nino
-
El Nio
Observed seasonal
evolution of interannual
SST anomaly in the
equatorial eastern
Pacific
SST anomaly pattern
-
Normal condition
El Nino condition
J. Bjerknes (1969) first
termed the equatorial
atmospheric overturning
circulation as Walker
circulation.
Reversed Walker
circulation anomaly, with
negative convective
heating anomaly over the
maritime continent.
-
Q1: How does El Nino
remotely affect Indian
monsoon?
Q2: By the summer after
peak El Nino, the eastern
Pacific SST becomes
normal. How can El Nino
has a delayed impact on
EAM?
Monsoon - ENSO RelationshipObservational fact:
1.Indian monsoon is drier during El Nino developing summer
2.East Asian monsoon (EAM) meiyu-rainfall significantly
increases 6 months after
the peak of El Nino (EAM becomes wetter during El Nino decaying
summer).
El Nino developing year El Nino decay year
-
Response of circulation and rainfall anomalies to El Nino
Suppressed convective heating in maritime continent
Atmospheric Rossby wave response drought over India
El Nino composite in JJA(0), vector: 850-hPa wind, shaded: prec.
anomaly
-
Gill modelFig. 1 Solutions for heating symmetric
about the equator in the region |x|
-
Anomaly dry AGCM experiment:
- 3D summer mean flow and anomalous
heating in MC are specified.
How does the El Nino remotely
impact the Asian monsoon?
Large-scale east-west overturning ?
Equatorial asymmetric response to a
symmetric El Nino forcing, why?
Wang, Wu and Li, 2003, J. Climate
-
Rainfall anomaly over China in summer of 1998, 6 months
after a peak El Nino in winter 1997
80 90 100 110 120 130
20
30
40
50 200
150
100
50
-50
-100
-150
-200
1998
-
Leading mode of Monsoon Interannual Variation
850 hPa winds and local SST
anomalies (19572001)
Biennial tendency associated with
ENSO turnabout
Distinct evolutions of anticyclonic
anomalies over SIO and WNP
Wang, Wu, Li 2003, J. Climate
-
Wang and Zhang
2002, JC
-
A positive thermodynamic air-sea feedback mechanismWang et al.
2000
El Nino heating atmospheric Rossby wave response
cold SSTA in WNP anomalous AC EAM
-
UV850OMEGA500 SST 200hPa velocity potential
Q3: As the local SSTA dissipates quickly in JJA(1), what
maintains the anomalous
anticyclone in WNP ? How does the IO SSTA affect the WNPM? (Wu
et al.
2009, JC; Xie et al. 2009, JC)
Q4: What is the relative role of the remote IO forcing vs. local
SSTA in affecting
circulation anomaly in WNP? (Wu et al. 2010, JC)
12 El Nino composite (1950-2006)
-
How does the basin-wide IO warming in JJA (1) impact the
WNPM?
IO equatorial heating Kelvin wave response Anticyclonic
shear
of the Kelvin wave easterly Ekman pumping induced PBL
divergence Suppressed WNP monsoon heating Anomalous
anticyclone
Wu, Zhou and Li
2009, J. Climate
-
Relative contribution to WNPAC by
IO SSTA and WNP SSTA
Vorticity anomaly over 10-35N, 115-160E
Wu, Li, Zhou, J. Climate, 2010
-
Inter-monsoon Relationships among IM, AM and WNPM
Wang, et al. 2005
Lagged correlations among WNPM, IM and AM for 1979-2005 using
CMAP and NCEP2 data
Lagged correlations among WNPM, IM and AM for 1979-2005 using
GPCP2 and JRA data
FromIndia
JJA(0)
Australia
DJF(0)
SCS/WNP
JJA(0)
Australia
DJF(0)
SCS/WNP
JJA(0)
ToAustralia
DJF(1)
India
JJA(0)
Australia
DJF(1)
SCS/WNP
JJA(0)
India
JJA(0)
Correlation
coefficient0.29 -0.28 -0.41 0.37 -0.64
FromIndia
JJA(0)
Australia
DJF(0)
SCS/WNP
JJA(0)
Australia
DJF(0)
SCS/WNP
JJA(0)
ToAustralia
DJF(1)
India
JJA(0)
Australia
DJF(1)
SCS/WNP
JJA(0)
India
JJA(0)
Correlation
coefficient0.35 -0.26 -0.32 0.32 -0.70
Red highlighted number indicates that the correlation is
statistically significant.
Gu, Li, et al. 2010, J. Climate
-
Schematic for AM-WNPM in-phase relation ENSO impact: remote vs.
local process
-
Schematic for out-of-phase relation between
WNPM and AM
El Nino developing
phase
El Nino decaying/La
Nina developing
-
Schematic for out-of-phase WNPM-IM relation
-
4. Monsoon interdecadal
variation
-
Global mean surface
temperature shows a
rising trend in recent
decades. Removing this
trend, one may find a
nature oscillation mode
on decadal timescale.
The PDO index was high
after 1976/77 (regime
shift) and stayed pretty
high till the late 1990s.
Associated with the high
PDO index (or warm
phase PDO) is a
strong Aleutian Low
anomaly.
Pacific Decadal Oscillation (PDO)
-
Decadal change of JJA rainfall
-
Temperatureshading) and meridional circulation (vector)
E. Asian regional average (105-122 E)
-
Why did the temperature have a cooling trend
in North China in past 40 years?
Li, C., T. Li, J. Liang, D. Gu, A. Lin, and B. Zheng, 2010:
Interdecadal variations of meridional
winds in the South China Sea and their relationship with summer
climate in China. Journal of
Climate, 23, 825-841.
High-latitude (NAO) impact (Yu and Zhou 2004)
Tropical forcing (Li et al. 2010)
-
The difference (Phase II minus Phase I) of wind (vector, units:
ms-1,
dark vectors denote that the difference exceeds the 95%
significance
level), geopotential height (contour, units: gpm,) and
temperature
(shaded, units: C) fields at 850 hPa.. NCEP reanalysis data
(1958-
2005) were used.
-
Vertical profiles of
temperature (a, units: C)
and geopotential height
(b, units: gpm)
anomalies averaged over
(100~110E35~45N) and
meridional wind
anomalies (c, units: m.s-1)
averaged over
(110~120E20~30N) at Phase I
(solid line) and Phase II
(dashed line)
-
Fig. 5 Meridional-vertical section of
difference (Phase II minus Phase I) fields
for a) temperature (C), b) geopotential
height (gpm), c) zonal wind (ms-1), d) p-
vertical velocity (Pa s-1) and e) meridional
wind (ms-1) averaged between 100~110E.
-
Vertical profiles of a)
anomalous specific
humidity (unit: gkg-1),
b) apparent heat source
Q1 (unit: Ks-1) and c)
apparent water vapor
sink Q2 (unit: Ks-1 )
averaged over
(100~11035~45N)at Phase I (solid line)
and Phase II (dashed
line).
-
The difference (Phase II minus Phase I) fields for a) NCEP OLR
(units: Wm-2lessthan -2 Wm-2 is shaded; symbol * denotes the OLR
difference exceeding the 95%
significance level ) and b) meridional-vertical streamfunction
averaged over 105-130E.
-
Enhanced convection over southern SCS (105-120E0-10N) in
association with the tropical ocean warming
Anomalous descending motion over
midlatitude East Asia (Hadley circulation)
Decrease in humidity and
increase in outgoing
longwave radiation into
space in the midlatitude
East Asia
Decrease in local
tropospheric
temperature and
thickness
Local negative
(positive)
geopotential height
anomaly at upper
(lower) levels
Local convergent
(divergent) flows at
upper (lower) levels
Weakening in land
ocean thermal
contrast
Weakening of the
LLMW over SCS
A tropical SST forcing hypothesis
-
Prior to 1978, YRV
and SEC are both
wet after a peak El
Nino.
After 1978, YRV is
wet while SEC is
dry after a peak El
Nino.
Chang, Zhang, Li, 2000 JC
-
79
Contour lines for 5870 gpm of 500 hPa geo-potential height for
each summer
1980-99
1958-77
NCEP/NCAR ERA40
-
5. Monsoon trends in past 30 years
and future change under global
warming
-
Trends of the global monsoon
precipitation during 1979-2008
Tim Li, Pang-chi Hsu and Bin Wang
International Pacific Research Center,
University of Hawaii, Honolulu, Hawaii
81
Hsu, P.-C., T. Li, and B. Wang, 2011: Trends in Global Monsoon
Area and Precipitation
over the Past 30 Years. Geophys. Res. Lett., 38, L08701,
doi:10.1029/2011GL046893
-
The GPCP data shows that the global monsoon shows an
increasingtrend since 1980. (Wang and Ding 2006)
BUT, the CMAP shows an decreasing trend in the global monsoon
rainfall. (Zhou et al. 2008)
Questions:
Why does the global monsoon trend show an inconsistence
between
GPCP and CMAP? Is it due to the uncertainty of data or the
analysis
methodology?
Note that the global monsoon area is defined in the past 30
years
with a fixed region for each year. Does the global monsoon
area
change annually?
Motivation
82
-
Observed monthly precipitation 1979-2008
1) Global Precipitation Climatology Project (GPCP)
2) CPC Merged Analysis of Precipitation (CMAP)
interpolated onto a 1 longitude by 1 latitude grid
Definitions of global monsoon
Global Monsoon Area (GMA)
Regions in which (1) annual range precipitation > 2mm/day
[MJJAS-NDJFM]
(2) local summer precipitation > 55 % annual rainfall
[NH: MJJAS SH: NDJFM]
Global Monsoon Precipitation (GMP)
total summer monsoon rainfall falling in the monsoon domain
[NH: JJA SH: DJF]
Data & Definitions of global monsoon index
83
-
Trends of GMP (defined based on a fixed GMA)
This result is consistent with Wang and Ding (2006) and Zhou et
al. (2008).
Linear trend (29yr)-1 GPCP (%) CMAP (%)
GMP in
climatological
GMA
glb 28.46 (0.06) -19.92 (-0.04)
glb_lnd 5.44 (0.02) 33.40** (0.16)
glb_ocn 23.02+ (0.10) -53.32 (-0.18)
Man-Kendall test
+ 80%
* 90%
** 95%
*** 99%
increasing trend decreasing
trend
84
-
Trends of GMP (defined based on a varying GMA)
increasing trend increasing trend
The GMP based on a varying GMA show increasing trends in both
the GPCP
and CMAP.
Linear trend (29yr)-1 GPCP (%) CMAP (%)
GMP in
yearly varying
GMA
glb 127.65 (0.25) 23.02 (0.04)
glb_lnd 31.58 (0.13) 70.80** (0.34)
glb_ocn 96.07+ (0.34) -47.79 (-0.14)
85
-
Trends of GMA
Trends of global monsoon area are both increased in the GPCP and
CMAP.
increasing trend increasing trend
contour:
clim_GMA
shading:
blue
extension
orange
shrink
86
-
Trends of global monsoon intensity
The GMP and GMA are both increased over time, how about the
GMI?
decreasing
trend
decreasing
trend
Since the rate of increased GMA is larger than it of GMP, the
GMI shows
decreasing trends in GPCP and CMAP.
Global Monsoon Intensity (GMI) rainfall amount per unit area
87
-
Summary
The trend of global monsoon precipitation (GMP) in a fixed
global
monsoon area (GMA), is increased in the GPCP but it is decreased
in the
CMAP. It reveals the inconsistent trends between GPCP and
CMAP.
Calculated based on a varying GMA, the trends of GMP in GPCP
and
CMAP are consistent, both of which are increasing during
1979-2008.
The GMA shows an increasing trend in both the GPCP and CMAP,
which is associated with the increased GMP.
Since the increased rate of GMA is larger than it of GMP, the
global
monsoon intensity (GMI), which is defined as rainfall amount per
unit
area, shows a decreasing trend in both datasets.
The decreased GMI in the GPCP is contributed mainly by the
land
monsoon while it is contributed by the oceanic monsoon in the
CMAP.88
-
Tim Li1, Pang-chi Hsu1, Jing-Jia Luo2,
Hiroyuki Murakami3, Akio Kitoh3 and Ming Zhao4
1International Pacific Research Center, University of Hawaii,
USA2Research Institute for Climate Change, JAMSTEC, Japan
3Meteorological Research Institute, Tsukuba, Ibaraki,
Japan4Geophysical Fluid Dynamics Laboratory, Princeton, New Jersey,
USA
Changes in global monsoon precipitation
under global warming
Hsu, P.-C., T. Li, J.-J. Luo, H. Murakami, A. Kitoh, and M.
Zhao, 2012: Increase of global
monsoon area and precipitation under global warming: A robust
signal? Geophys. Res. Lett., in
press.
-
Seeking for consistent and robust global monsoon changes
indifferent high-resolution AGCMs forced by
different SST warming patterns
Model Control Runs Global Warming Experiments
MRI
ECHAM5
T106
(~ 1.125)
T106_pd
AMIP-type
run with
observed
SST
T106_mw
Globally uniform SST
warming (2.24C) derived by
ECHAM5/MPI-OM simulated
SST anomaly between A1B
and 20C3M
T106_sw
Spatially-varying SST
warming derived by
ECHAM5/MPI-OM simulated
SST anomaly between A1B
and 20C3M
MRI
ECHAM5
T319
(~40km)
T319_pd
ECHAM5/
MPI-OM
20C3M
SST
T319_sw
Model is forced by
ECHAM5/MPI-OM simulated
SST in A1B scenario
Japan
MRI-JMA
T959
(~20km)
MRI_pd Historical
HadISST MRI_sw.e
18-model ensemble mean of
future SST warming in CMIP3
plus the present-day
interannual variations
US GFDL
HiRAM
C180
(~50km)
GFDL_pd Hsitorical
HadISST GFDL_sw.e
18-model ensemble mean of
future SST warming in CMIP3
-
Robust global monsoon changes under global warming
Global Monsoon Precipitation (GMP)Regions in which (1) annual
range
precipitation > 2mm/day (2) local
summer rainfall > 55 % annual
rainfall
Global Monsoon Precipitation (GMP) total summer monsoon rainfall
in GMA
Global Monsoon Intensity (GMI)global monsoon precipitation
amount
per unit area0
2
4
6
8
10
12
14
16
18
GMA GMP GMI(%)
Change rates in global monsoon
index_avg
composite
Change rate (%) EXP_mw EXP_sw T319_sw MRI_sw.e GFDL_sw.e
GMA 7.75 4.58 8.66 7.15 6.59
GMP 19.80 10.35 15.18 9.85 7.36
GMI 11.22 5.61 7.14 2.51 0.88
GMA, GMP and GMI increase consistently among individual
models
-
-2
0
2
4
6
8
10
GMA GMP GMI
(%)
K-1
CMIP5_index_avg CMIP5_MME_prep
CMIP3_index_avg CMIP3_MME_prep
Normalized change rates in global monsoon
CMIP3 and CMIP5 Results
-
GMA changes under global warming
High-resolution AGCMs perform well in capturing the major GMA in
the
present-day simulations (red contour)
Significant GMA expansions occur in the oceanic monsoon and the
South
American and African land monsoon regions (blue shaded
areas).
-
Rainfall changes under global warming
Contour: Climatology of
difference in precipitation
between global warming
and present-day.
Dark shading: Consistent
change in 5 models
(100% consistency)
Light shading: Consistent
change in 4 models
(80% consistency)
Rainfall increases
over the ITCZ and
SPCZ where are the
wettest regions while it
shows decrease in the
tropical western Pacific
and Indian Oceans. Not
always rich-get-richer.
-
Moisture diagnosis of enhanced GMP
GMP = < q>
-
Relative contributions of thermodynamic and dynamic effect
The thermodynamic component via increasing water vapor content
plays the
major role in strengthening GMA while global SST warming.
The dynamic effect associated with weakening monsoon circulation
and surface
wind speed contributes negatively to the GMP.
-20
-10
0
10
20
30
40
q*D q*D q*D
Decompositions of moisture convergence
T106_mw T106_sw T319_sw MRI_sw.e GFDL_sw.e
-5
0
5
10
15
20
q*V q*V q*V
Decompositions of evaporation
T106_mw T106_sw T319_sw MRI_sw.e GFDL_sw.e
pd pd
< q * D > =
< q * D > < q D > < q D >
dynamic thermodynamic nonlinear
E s a
E s a pd s a s apd
E = [ L C (q -q )]
= L C [ *(q -q ) + * (q -q )+ * (q -q )]
V
V V V
dynamic thermodynamic nonlinear
-
Summary of future GM change
By comparing the ECHAM5/MRI/GFDL model simulations driven
by present-day observed SST and future warming scenarios, we
note an increasing trend for both the global monsoon area
(GMA)
and overall global monsoon precipitation (GMP) amount. The
global monsoon intensity (GMI) also shows an increasing
trend.
The signal appears robust across different model physics and
different future SST warming patterns.
A moisture budget analysis shows that the enhanced global
monsoon precipitation is attributed to both the increased
evaporation and moisture flux divergence under global
warming.
The increase of moisture is primarily responsible for the
increase
of moisture flux convergence in the future warming scenario.
-
ThanksDiamond Head
http://www.tonyandkitty.com/gallery/album01/Diamond_Head?full=1
-
Impact of MJO on winter
circulation ad rainfall in East Asia