Annual%20cycle%20of%20the%20West%20African%20rainfall

1
A Multiscale Analysis of the West African
Monsoon Chris Thorncroft Department of
Atmospheric and Environmental Sciences University
at Albany
2
A Multiscale analysis of the West African Monsoon
(1) Annual Cycle of Rainfall and associated
Water Vapour Transport (2) Interannual
Variability of the coastal rainfall in Spring
(3) The African Easterly Wave Life-Cycle (I)
Genesis (II) Baroclinic Developments (III) West
Coast Developments (4) Final Comments and
Future Work
3
Annual cycle of Water Vapor Transport in the West
African Monsoon region

• Chris Thorncroft, Hanh Nguyen (University at
  Albany)
• Chidong Zhang (RSMAS, Miami)
• Philippe Peyrille (MeteoFrance, Paris)

4
The Coupled Monsoon System
Key features of the WAM Climate System during
Boreal summer
AEJ
5
Meridional Circulations
North-South Section along the Greenwich Meridian
AEJ
50oC
90oC
?
?e
?
?e
20oC
60oC
6
Meridional Circulations
Shallow Meridional Circulation (SMC) over ocean,
especially in Spring
AEJ
50oC
90oC
?
?e
?
?e
20oC
60oC
7
Annual Cycle of Mean Rainband
Data GPCP (Global Precipitation Climatological
Project). Resolution pentad on a 2.5o grid.
Averaged from 10oW to 10oE over 23 years
(1979-2001). c.f. Gu and Adler (2004)
8
Aims

• To document the climatological mean annual
  evolution of the water vapour transport and
  associated three-dimensional pattern of moisture
  convergence in the WAM and tropical Atlantic
  regions. (revisiting Cadet and Nnoli (1987))
• To relate this to the regional circulations and
  low-level thermodynamic conditions, especially
  those linked to the Atlantic cold tongue and
  Saharan heat low.
• To improve understanding of the various phases
  of the annual cycle of WAM rainfall

9
Data

• Observations, reanalysis and operational analysis
  data including
• pentad 2.5o GPCP
• Reynolds SST 1o, weekly and daily
• Reanalysis from the ECMWF daily 2.5o ERA40
• The period of study is 1979-2001

10
Relationship between SKT and surface meridional
wind
Relationship between rainfall and surface
conditions
SKT and rainfall
Warming over the continent due to the surface
solar heating. Rapid cooling of the ocean
surface south of the equator between April and
June ? rapid rise in MSLP Acceleration of
southerly winds across the equator. c.f.
Okumara and Xie (2004)
MSLP and VWND
11
Relationship between rainfall and surface
conditions
Equivalent potential temperature

• Peak rainfall always lies south of thetae peak
• Gradient in thetae still important
• Location of heat low important for poleward
  extent

12
Total Column Moisture Flux Convergence
13
Total Column Moisture Flux Convergence
Peak in moisture flux convergence linked to heat
low shallow meridional circulation acts to
moisten the column and extend the rainfall
polewards (c.f. Sultan and Janicot (2000,2003),
Hagos and Cook (2008))
14
Total Column Moisture Flux Convergence
Peak in moisture flux convergence over ocean
15
Total Column Moisture Flux Convergence
Rapid shift and increase in moisture flux
convergence towards coast between April and May
16
Total Column Moisture Flux Convergence
Rapid reduction in moisture flux convergence
during June linked to end of coastal rains
17
Total Column Moisture Flux Convergence
Rapid increase in moisture flux convergence
beginning of July linked to Sahelian rainfall
onset
18
Meridional Moisture Fluxes
Mid-levels (850-500mb)
Impact of Heat Low SMC
Low-levels (sfc-850mb)
19
Meridional Moisture Fluxes
Mid-levels (850-500mb)
Equatorward moisture flux at mid-levels enhances
moisture flux convergence in rainy zone
enhances rainfall there? Polewards of this there
is dry advection inhibits rainfall there?
Low-levels (sfc-850mb)
20
Meridional Moisture Fluxes
Mid-levels (850-500mb)
Low-levels (sfc-850mb)
Marked increase in cross-equatorial moisture
fluxes during April-May Linked to cold tongue
development and coastal onset
21
Schematic evolution
ITCZ
1. Ocean phase (Feb-April) -Main rainband is
broad with peak values just poleward of the
Equator (1oN ). The rainfall is located mostly
over the warmest water (gt28oC) with little over
the land. -At the end of this period the cold
tongue starts to develop, resulting in a broad
region of SSTs close to the equator falling below
28oC. - Does the heat low SMC impact the
surface winds?
SMC
HL
SST
22
Schematic evolution

• 2. Coastal phase (May-mid-June)
• -Cold tongue development associated with a rise
  in equatorial surface pressure, and an
  acceleration of southerlies and associated
  moisture flux towards the coast.
• -Marked moisture flux convergence, just
  equatorward of the land (4oN) is associated with
  the highest rainfall of the annual cycle, and the
  first rainy season for coastal regions of West
  Africa.
• c.f. Zheng, Eltahir and Emanuel (1999)
• Okumara and Xie (2004)
• Gu and Adler (2004)
• Caniaux et al (2009)
• - Peak rainfall is located over warmest water

ITCZ
SMC
HL
SST
23
Schematic evolution

• 3. Transitional Phase (End of June)
• June represents a period where the environment
  becomes less favorable for convection in the
  coastal region. This is consistent with coastal
  upwelling and a reduction of SSTs there.
• Intense coastal rainfall can only be transient?
• Why doesnt it rain more in June?
• Does this weakening promote the perception of a
  jump often discussed in the literature?

ITCZ
SMC
HL
SST
24
Schematic evolution
4. Sahelian Phase (July-August) - Between June
and July the peak in moisture flux convergence
reaches 10oN and increases rapidly consistent
with the observed Sahelian rainfall onset. - In
July and August moisture flux divergence is
present over the coastal region consistent with
continued suppression of rainfall there. c.f.
Sultan and Janicot (2000.2003) Sijikumar et
al (2006) Ramel et al (2006) Hagos
and Cook (2007)
ITCZ
SMC
HL
SST
25
ERA40 vs NCEP1 rainfall
Wet bias in Spring? Dry bias in Sahel in
Summer Dry bias in Spring?
26
Total Column Moisture Flux Convergence
ERA40 NCEP1 can a strong heat low SMC
suppress convection south of it?
27
Concluding remarks on Annual Cycle

• At some level the coastal onset seems easier to
  understand than the Sahelian onset with peak
  rainfall following the peak in SSTs
• What processes determine the nature and
  variability of the cold tongue (role of heat low,
  sub-surface ocean structure, Atlantic ocean
  variability, radiation)?
• Why is cold tongue development more rapid in the
  Atlantic than in the Pacific?
• Can climate models represent these coupled
  processes?
• Need more in situ observations in the tropical
  East Atlantic!
• Need more work on nature and causes of
  variability of coastal rains (next)

28
Interannual Variability of Coastal Rainfall
29
Resolution pentad on a 2.5o grid. Averaged from
10oW to 10oE over 10 years (1998-2007).
TRMM

• Similar patterns
• broad and weak in winter.
• a rapid shift of the southern limit of the
  rainband in May.
• most intense in spring.
• a marked decline in June-July.
• rapid shift to the Sahel in summer.
• steady retreat following surface solar heating.
• Different intensities
• strong in winter for CMAP.
• strong in spring for TRMM and CMAP.
• weak in summer for CMAP.

CMAP
GPCP
30
Resolution pentad on a 2.5o grid. Averaged from
10oW to 10oE over 10 years (1998-2007).
TRMM
Definition Coastal Onset defined in terms of the
rapid reduction in rainfall over the equatorial
region and the associated reduction in the width
of the ocean rainfall.
CMAP
GPCP
31
Composite diagrams
TRMM
Period 1998-2007
tOC 9 May
CMAP
tOC 10 May
GPCP
tOC 10 May
32
Interannual variability
Coastal onset 11 May range 13
pentads
Coastal length 7 pentads range 11
pentads
Sahel onset 6 July range 7 pentads
Hypotheses Delayed cold tongue development
delays coastal onset Strong heat low delays
coastal onset
33
Relationship between the coastal onset and the
SST cooling
34
Relationship between the coastal onset and the
Saharan Heat Low
35
A comparison of 3 years 2005-2007
36
A comparison of 3 years 2005-2007
07
07
06
06
05
05
37
2005
Large variation in the coastal onset. Earliest
cold tongue development in Spring 2005 earliest
coastal onset. Strongest HL in Spring 2007
during the oceanic regime ? possible role in
delaying the coastal onset via subsidence
11
20
2006
21
3
2007
23
29
38
Interannual variability
Coastal rainfall intensity 8.2 mm/d range
9.1 mm/d
39
Relationship between the coastal rainfall and the
SST
MAY
MARCH
APRIL
FEBRUARY
40
Relationship between the coastal onset and the
Sahel onset
41
Concluding remarks on Interannual Variability

• The West African coast is characterized by
  marked interannual variability in rainfall both
  in terms of the onset (of the coastal phase) and
  amounts.
• Onset date is influenced strongly by the timing
  of the cold tongue development as well as the
  intensity of the heat low.
• Rainfall amounts are correlated with SSTs in the
  Pacific and SE Atlantic suggesting predictability
  with several months lead-time.
• Onset of the coastal phase is correlated with the
  Sahelian onset. Sahelian onset tends to occur
  roughly 2 months after the coastal onset.

42
Variability of Synoptic Weather Systems

TD-filtered OLR (AEW-activity) Peaks in
summer We know little about the nature and
causes of AEW-variability
Kelvin-filtered OLR Peaks in Spring Key synoptic
system for pre-coastal phase and possibly the
coastal phase
43
3. African Easterly Waves
OLR and 850 hPa Flow Regressed against
TD-filtered OLR (scaled -20 W m2) at 10?N, 10?W
for June-September 1979-1993
Day 0
Streamfunction (contours 1 X 105 m2 s-1) Wind
(vectors, largest around 2 m s-1) OLR (shading
starts at /- 6 W s-2), negative blue
Kiladis, Thorncroft, Hall (2006)
44
3. African Easterly Waves
Objectively diagnosed troughs (solid lines),
African Easterly Jet (dashed), with PV (315K) and
IR from METEOSAT (courtesy Gareth Berry)
45
AEW life-cycle phases

• Phase I Genesis (e.g. Thorncroft, Hall and
  Kiladis, 2008)
• Phase II Baroclinic growth (e.g. Berry and
  Thorncroft, 2005)
• Phase III Tropical Cyclogenesis (e.g. Hopsch,
  Thorncroft, Tyle, 2010)

46
Phase I Genesis Two Theories for the Genesis of
AEWs
I AEWs are generated via a linear mixed
barotropic-baroclinic instability mechanism
AEJ satisfies the necessary conditions for
barotropic and baroclinic instability Burpee
(1972), Albignat and Reed, 1980). Therefore we
expect AEWs to arise from small random
perturbations consistent with a survival of
the fittest view. Continues to be the consensus
view.
315K PV
925hPa q
47
Two Theories for the Genesis of AEWs
I AEWs are generated via a linear mixed
barotropic-baroclinic instability mechanism
(evidence against!)

• The AEJ is too short!
• The jet is typically 40-50o long.
• It can only support two waves at one time.
• It is therefore not possible for AEWs to develop
  via a linear instability mechanism.
• The AEJ is only marginally unstable!
• Hall et al (2006) showed that in the presence of
  realistic boundary- layer damping the AEW growth
  rates are very small or zero.
• It is therefore not possible for AEWs to develop
  sufficiently fast to be important.

48
Two Theories for the Genesis of AEWs
I AEWs are generated via a linear mixed
barotropic-baroclinic instability mechanism
(evidence against!)

• The AEJ is too short!
• The jet is typically 40-50o long.
• It can only support two waves at one time.
• It is therefore not possible for AEWs to develop
  via a linear instability mechanism.
• The AEJ is only marginally unstable!
• Hall et al (2006) showed that in the presence of
  realistic boundary-layer damping the AEW growth
  rates are very small.
• It is therefore not possible for AEWs to develop
  sufficiently fast to be important.
• So what can account for the existence of AEWs,
  their genesis and intermittancy?

49
Two Theories for the Genesis of AEWs
II AEWs are generated by finite amplitude
forcing upstream of the region of observed AEW
growth. Carlson (1969) suggested the importance
of convection and upstream topography for the
initiation of AEWs. Others pushed the linear
instability hypothesis. More recent
observational evidence has been provided
by Berry and Thorncroft (2005) case study of
an intense AEW Kiladis et al (2006) composite
analysis Mekonnen et al (2006) climatological
view c.f. Farrel, B. (1987)
50
Satellite imagery

• METEOSAT-7 Water Vapour channel.
• Shown every 6 hours from 30th July 2000 00z to
  4th August 2000 18z.
• Convective outbursts in the first day of the
  sequence preceded the dynamical signal.

51
(No Transcript)
52
OLR and 850 hPa Flow Regressed against
TD-filtered OLR (scaled -20 W m2) at 10?N, 10?W
for June-September 1979-1993
Day 0
Streamfunction (contours 1 X 105 m2 s-1) Wind
(vectors, largest around 2 m s-1) OLR (shading
starts at /- 6 W s-2), negative blue
53
OLR and 850 hPa Flow Regressed against
TD-filtered OLR (scaled -20 W m2) at 10?N, 10?W
for June-September 1979-1993
Day-4
Streamfunction (contours 1 X 105 m2 s-1) Wind
(vectors, largest around 2 m s-1) OLR (shading
starts at /- 6 W s-2), negative blue
54
OLR and 850 hPa Flow Regressed against
TD-filtered OLR (scaled -20 W m2) at 10?N, 10?W
for June-September 1979-1993
Day-3
Streamfunction (contours 1 X 105 m2 s-1) Wind
(vectors, largest around 2 m s-1) OLR (shading
starts at /- 6 W s-2), negative blue
55
OLR and 850 hPa Flow Regressed against
TD-filtered OLR (scaled -20 W m2) at 10?N, 10?W
for June-September 1979-1993
Day-2
Streamfunction (contours 1 X 105 m2 s-1) Wind
(vectors, largest around 2 m s-1) OLR (shading
starts at /- 6 W s-2), negative blue
56
OLR and 850 hPa Flow Regressed against
TD-filtered OLR (scaled -20 W m2) at 10?N, 10?W
for June-September 1979-1993
Day-1
Streamfunction (contours 1 X 105 m2 s-1) Wind
(vectors, largest around 2 m s-1) OLR (shading
starts at /- 6 W s-2), negative blue
57
OLR and 850 hPa Flow Regressed against
TD-filtered OLR (scaled -20 W m2) at 10?N, 10?W
for June-September 1979-1993
Day 0
Streamfunction (contours 1 X 105 m2 s-1) Wind
(vectors, largest around 2 m s-1) OLR (shading
starts at /- 6 W s-2), negative blue
58
Idealised Modeling Study Thorncroft, Hall and
Kiladis (2008)
59
Consequences

• Significance for weather prediction
• A significant convective outbreak in the Darfur
  region will favor the formation of a train of
  AEWs to the west over sub- Saharan Africa within
  a few days.
• For daily-to-medium range forecasts of AEWs, it
  is important to monitor, and ultimately predict,
  the nature of the upstream convection.

60
Consequences

• Significance for longer timescales
• In addition to considering the nature of mean
  AEJ, we should consider the nature and
  variability of finite amplitude convective
  heating precursors.
• It is likely that BOTH are important

61
Phase II Baroclinic Development - Scale
Interactions
62
Conceptual framework (ii) Baroclinic growth.
q Max
700hPa Trough
63
Conceptual framework (ii) Baroclinic growth.
q Max
700hPa Trough
64
Conceptual framework (ii) Baroclinic growth.
q Max
700hPa Trough
65
PV-theta analysis of AEWs Scale Interactions

Synoptic-Mesoscale Interactions
66
PV-theta analysis of AEWs Scale Interactions

Synoptic-Mesoscale Interactions
67
PV-theta analysis of AEWs Scale Interactions

Synoptic-Mesoscale Interactions From a PV-theta
perspective, the heating rate profiles are
crucial to know and understand.
68
PV-theta analysis of AEWs Scale Interactions

Synoptic-Mesoscale Interactions From a PV-theta
perspective, the heating rate profiles are
crucial to know and understand.
Mesoscale-Microscale Interactions Ultimately
these profiles are influenced by the nature of
the microphysics!
69
Phase III West Coast Developments Role of
Guinea Highlands!
AEWs often get a boost before they leave
Africa associated with mergers of PV from
upstream and in situ generation. The Guinea
Highlands region is one of the wettest regions of
tropical North Africa.

GPCP rainfall (mm/day) for Aug-Sep, 1997-2007)
70
Importance of Guinea Highlands Region
Coherent cyclonic centers are tracked within the
ITCZ at 700hPa and in the low-level baroclinic
zone at 850hPa

Average vorticity tracking statistics for
June-July-August at 700hPa and 850hPa based on
ERA40 using methodology of Thorncroft and Hodges
(2001).
71
Importance of Guinea Highlands Region
Composites of East Atlantic Developing and
Non-Developing AEWs (1979-2001)



Developing Non-Developing
Hopsch , Thorncroft and Tyle 2010
72
4. Final Comments on Phase I Genesis



AEWs are forced by upstream finite amplitude
precursors including, most importantly,
upstream convection, that is most commonly
triggered by topography over Darfur, but
sometimes the Ethiopian Highlands. Other forcing
is possible including that associated with
midlatitude troughs. What are the causes of
intra-to-interannual variability of AEW-activity?
What are the relative roles of variability in
upstream precursors and variability in the AEJ?
73
4. Final Comments on Phase II Baroclinic
Developments



AEWs interact with and develop in association
with MCSs The PV-theta framework is ideal for
studying scale interactions - but it remains a
challenge for us to assess the diabatic heating
profiles and associated PV structures at the
mesoscale. Ongoing research is utilizing radar
data from AMMA to shed light on thee mesoscale
structures. High resolution modeling offers a
useful tool to study these structures (c.f. YOTC)
74
4. Final Comments on Phase III West Coast
Developments



AEWs tend to intensify at the West African coast.
The nature of the resulting structures can impact
the probability of tropical cyclogenesis
downstream. What are the relative roles of
upstream PV and in situ generated PV? Why do
some intense AEWs weaken? What are the relative
roles of the SSTs, Saharan air layer, dry air in
the upper-troposphere, midlatitude troughs.?
75
4. Final Comments Intraseasonal Variability



There exists significant intraseasonal
variability in AEW activity that is yet to be
fully described and understood. Ideal timescale
for studying interactions between weather and
climate