Sea Surface Temperature-Convection Relationship in Tropical Oceans with Special Emphasis to Intraseasonal
Variability of Indian Summer Monsoon
A thesis submitted in partial fulfillment for the degree of Doctor of Philosophy
in
Atmospheric Science
by
Sabin TP
Faculty of Marine Sciences
Cochin university of Science and Technology Cochin 682 016, India
June 2011
Declaration of Authorship
I, Sabin TP, declare that this thesis entitled,Sea Surface Temperature-Convection Re- lationship in Tropical Oceans with Special Emphasis to Intraseasonal Variability of Indian Summer Monsoonand the work presented in this thesis is a bonafide record of research work carried out by me under the supervision of Dr. C.A. Babu, in the Depart- ment of Atmospheric Scinces, Cochin University of Science and Technology, in partial fulfilment of the requirements for the award of the Ph.D. degree under the Faculty of Marine Sciences, Cochin University of Science and Technology. I further confirm that the subject matter of the thesis has not formed the basis for the award of Degree or Diploma of any University or Institution.
Sabin T.P.
Cochin 16 June 2011
i
Certificate
This is to certify that this thesis entitled,Sea Surface Temperature-Convection Rela- tionship in Tropical Oceans with Special Emphasis to Intraseasonal Variability of Indian Summer Monsoon is an authentic record of the research work carried out by Mr. Sabin T.P. under my guidance in the Department of Atmospheric Sciences, Cochin University of Science and Technology, in partial fulfilment of the requirements for the Ph.D. degree of Cochin University of Science and Technology under the Faculty of Ma- rine Sciences, and no part thereof has been presented for the award of any degree in any university.
Dr. C.A. Babu
Cochin 16 June 2011
ii
Acknowledgements
My PhD work, fulfilled through this thesis has produced challenging and interesting moments and experiences in my life. I could successfully complete the work with the help of many people, who have strongly supported me in each phase of the thesis work. It is now my turn to express my gratitude to all of them.
I would like to express my immense gratitude to my thesis supervisor Dr. C.A. Babu, Depart- ment of Atmospheric Sciences, Cochin University of Science and Technology, for his guidance, valuable suggestions, and continuous encouragement throughout the course of this thesis work.
I am very grateful to Prof. P.V. Joseph, whose enthusiasm and integral view on research and his knowledge and approach towards serious research, motivated me throughout this work. I am indebted to him for teaching me the ethics of sincere and fruitful research, which, I am sure, will guide me throughout my life.
I thank Dr. K.R. Santosh, Head, Department of Atmospheric Sciences, who is also my doc- toral committee member for providing all the facilities and valuable suggestion throughout my research period. Prof. H.S. Rammohan, Dean, Faculty of Marine Sciences and Prof K. Ma- hankumar, Dean, Faculty of Environmental Sciences, have kindly helped me at many occasions of crisis during the course of the research work. I express my sincere gratitude to both of them.
I extend my thanks to Prof. C.K. Rajan, Sri. B Chakrapani and Dr. V. Madhu of Department of Atmospheric Sciences and Dr. R. Sajeev, Dr. A.N. Balchand and Sri. P.K. Saji, of Department of Physical Oceanography. My thanks are due to Dr. Anu Simon, Dr. V Hamza, Dr. Venu G.
Nair, Dr. G. Bindu, Dr. G. Mrudula, Dr. M.G. Sreedevi and Sri. G. Yasodaran for their timely support at all occasions. I thank the non-teaching staff of Dept. of atmospheric Sciences for their help.
The sincere company of S. Abhilash, Johnson Zaccharia, P. VijayaKumar, K. Prasanth, K.K.
Baiju and Shaiju P has always helped to be happy in my hostel life. I extend my thanks to my lab-mates Asha S. Phillip, Gireesh, Aneesh and all other research scholars of Department of Atmospheric Sciences, and inmates of CUMS hostel. Many thanks to you all for your friendship and support. My special thanks is due to P. VijayaKumar for critical proof-reading of this thesis.
iii
Acknowledgements iv I thankfully remember the motivation from Dr. R Krishnan, Director, Center for Climate Change Research (CCCR), Dr. Milind Mujumdar and Dr. Karumuri Ashok, Scientists CCCR, IITM, Pune. I express my sincere thanks to Prof. BN Goswami, Director IITM for his advice and support to complete this PhD work. I extend my thanks to all my colleagues in CCCR and my friends in IITM.
The financial support from Indian Space Research Organisation (MOP-Space Application Cen- ter, Ahmadabad, India) and University Grants Commission’s Research Fellowship in Science for Meritorious Students (UGC-RFSMS), Government of India are gratefully acknowledged.
I believe, it is God’s blessing that made the completion of this work possible. Special thanks to my father, sister and brother-in-law and their kids Haritha and Hitha for their good wishes, love and support. I extend my gratitude to my wife Remya, who has always been a constant source of energy, support and encouragement. My deepest sense of gratitude goes to my dear mother whose soul will be so happy in this moment. I dedicate this thesis to her.
Preface
Asian Summer Monsoon has been a subject of intensive study for over five decades now. Since irrigation facilities are inadequate, as revealed by the fact that less than half of the cultivated land is irrigated in India, agriculture remains largely and profoundly dependent on monsoon.
A good monsoon results in a robust growth of the economy as a whole, while a poor mon- soon leads to a sluggish growth. Frequent or prolonged breaks during the monsoon season can lead to drought conditions. Long breaks in critical growth periods of agricultural crops lead to substantially reduced yield. Thus the better understanding of the Asian monsoon system is significant for the sustainable development of the country. Atmosphere is in contact with the ocean surface and sea surface temperature (SST), has an important role in producing deep con- vective clouds and weather systems in the tropics. Thus in the variability of the Asian summer monsoon, atmosphere-ocean interaction has a very important role. Using the currently available high resolution and more accurate satellite observations of SST and precipitation we conducted an enhanced and comprehensive study of the SST-convection relationship over the global tropics particularly in relation to the Active-Break cycle of the Asian summer monsoon.
This study extensively investigates the problems and prospects of SST-convection relations over monsoon and non monsoon areas. Over the summer warm pool areas of Indian and west Pacific oceans (monsoon areas) where the zone of maximum SST is about 20 degrees of latitude away from the equator: (a) Convection is related to the SST-gradient that generates low level moisture convergence and upward vertical motion in the atmosphere. Regions of SST maxima have low SST gradients and therefore feeble convection. Over the non monsoon areas (b) Convection initiated by SST gradient produces strong wind fields particularly cross-equatorial Low Level Jetstreams (LLJ) on the equator ward side of the warm pool and both the convection and LLJ are seen to grow through a positive feed back process. Thus large values of convection are associated with the cyclonic vorticity of the LLJ in the atmospheric boundary layer. In contrast with this, in the Inter Tropical Convergence Zone (ITCZ) over the east Pacific ocean and the South Pacific Convergence Zone (SPCZ) over the west Pacific ocean, low level winds converge in the zone of maximum SST which lies close to the equator producing elongated bands of deep convection which increases with SST for the full range of SST’s. The low level wind divergence has large and significant linear correlation with convection in both the warm pool and ITCZ/SPCZ areas,
v
Preface vi but the linear correlation between SST and convection is large only for the ITCZ/SPCZ. These findings have important implications for the modeling of large scale atmospheric circulations and the associated convective rainfall over the tropical oceans.
Using TMI SST we found the existence of a large tongue of cold SST in the Bay of Bengal during the monsoon season in the latitude belt 3◦N to 10◦N. It is in between two warm pools one around the equator and the other in the northern Bay of Bengal. A narrow tongue of low SST is seen in June, which was not existing in May. This tongue of low SST expands in east-west and north-south during July. There is further expansion of the area of the cold water in August. We have named this area as theCold Poolof Bay of Bengal. From June to August, the temperature of the cold pool falls. Low-level winds of the summer monsoon have alongshore component over the west coast of Kerala (India) and Sri Lanka, which generate strong coastal upwelling.
The cold pool was speculated as (a) due to the spreading of this upwelled cold water by the ocean currents and (b) the open ocean upwelling during break monsoon spells. The significant intraseasonal variability of the cold pool SST and its role in producing the SST gradient over the Bay of Bengal is also discussed.
In the Active-Break cycle of the monsoon north Indian Ocean and west Pacific Ocean are ac- tively involved. A coupled ocean-atmosphere interaction mechanism that explains all the phases of the Active-Break cycle is described through this thesis. A typical Active-Break cycle of the Asian summer monsoon is taken as beginning with maximum SST over the north Bay of Ben- gal when the oceans to its west and east from longitude 40◦E to 160◦E, and between latitudes 10◦N and 25◦N also has maximum SST. During this pentad the cold pool of the Bay of Bengal has its minimum SST. An area of convection takes genesis over the Bay of Bengal immediately in the zone of large SST gradient north of the cold pool and it pulls the monsoon Low Level Jetstream (LLJ) through peninsular India. Convection and the LLJ westerlies then spread to the western Pacific ocean in which convection and LLJ have grown in a positive feed back process.
It is hypothesised that the cyclonic vorticity to the north of the LLJ axis is acting as a flywheel maintaining the convection during the long active phase against the dissipating effect of atmo- spheric stabilization by each short spell of deep convection. When SST over north of 10◦N has cooled and the convection weakens there, when the LLJ turns clockwise over the Arabian Sea and flows close to the equator in the Indian Ocean. A fresh band of convection develops at the
Preface vii equator and latitude 10◦S over the Indian Ocean, which is nourished by the cyclonic vorticity of the LLJ through the moisture supply. Solar radiation and light winds make the SST over the north of 10◦N warm rapidly with coherent reduced wind and convection and a new Active-Break cycle begins. Thus SST, convection, LLJ and the net heat flux at the ocean surface have impor- tant roles in this new way of looking at the Active-Break cycle as a coupled ocean-atmosphere phenomenon.
Variation of net heat flux at the ocean surface is found to control the SST variation in the Active- Break cycle. The net heat flux at the ocean surface of the composite Active-Break cycle showed that the SST variations over the north Bay of Bengal and to its west and east from longitude 40◦E to 160◦E, and between latitudes 10◦N and 25◦N are mainly caused by the net heat flux on the ocean surface, assuming a mixed layer of depth about 20 m which is the typical mixed layer depth there. It is shown that these fluxes control the day to day changes in SST there.
We have found that the convection in the Bay of Bengal peaks about five days after the peak in SST gradient over the Bay of Bengal and the convection over the west Pacific ocean peaks after another five days. The importance of monitoring the daily SST gradients over the Bay of Bengal to foreshadow the development of the Active-Break cycle of Asian summer monsoon is shown through this chapter.
In the recent decade 2000 to 2009 there were three major droughts in the seasonal monsoon rainfall of India namely in the years 2002, 2004 and 2009 which were El Nino years. These monsoon seasons had 8 spells of break monsoons with duration of more than 7 days (long break monsoon spells). The convection anomalies associated with these long breaks were studied.
We found that monsoon droughts associated with the El Ninos have been caused by the large rainfall deficiency over central and northwest India during these long break monsoon spells. It is inferred that it is the long breaks that cause the monsoon droughts in India usually attributed to El Ninos. The role of the tropical Pacific sea surface temperature anomalies in forcing the precipitation variability over the north west Pacific and the Indian monsoon region is examined.
Contents
Declaration of Authorship i
Certificate ii
Acknowledgements iii
Preface v
List of Figures xi
List of Tables xiv
Abbreviations xv
Symbols xvii
1 Introduction 1
1.1 Sea Surface Temperature and deep moist convection in the atmosphere
above . . . 1
1.2 Ocean atmosphere coupling in the annual cycle . . . 6
1.3 Monsoon Intraseasonal variability in relation to the ocean . . . 9
1.4 Prediction and Predictability of monsoon system. . . 17
1.5 In this Thesis . . . 19
2 Data and Methods 21 2.1 Datasets . . . 21
2.1.1 NCEP/NCAR Reanalysis. . . 21
2.1.2 QuikSCAT wind . . . 22
2.1.3 TMI SST . . . 22
2.1.4 OLR . . . 23
2.1.5 Precipitation . . . 24 viii
Contents ix
2.1.6 IPCC-AR4 Model Data. . . 25
2.2 Methods of analysis . . . 25
3 SST-Convection relation over tropical oceans 26 3.1 Warm Pool SST and the associated Convection . . . 27
3.1.1 Pre-monsoon season in north Indian Ocean . . . 27
3.1.2 Monsoon season in north-west Pacific Ocean . . . 29
3.1.3 Active - Break cycle in Bay of Bengal . . . 32
3.2 SST and Convection over ITCZ and SPCZ areas . . . 34
3.2.1 ITCZ over the north-east Pacific Ocean . . . 34
3.2.2 SPCZ over the south-west Pacific Ocean. . . 38
3.3 Summary and discussions . . . 40
3.3.1 Observational findings . . . 40
3.3.2 Modeling support . . . 44
4 The Cold Pool of Bay of Bengal 48 4.1 The Cold Pool in observations . . . 48
4.2 The Cold Pool in IPCC models . . . 53
5 An Ocean-Atmosphere Interaction Mechanism for the Active-Break Cycle of the Asian Summer Monsoon 56 5.1 Intraseasonal Variability of the boreal summer Monsoon . . . 56
5.2 Seven year July-August mean SST, Convection and 850 hPa wind . . . 60
5.3 Composite AB cycle in SST, Convection and 850 hPa wind . . . 61
5.3.1 Composite pentad SST . . . 62
5.3.2 Composite pentad GPI rain (convection). . . 66
5.3.3 Composite pentad 850 hPa wind . . . 69
5.4 Proposed mechanism for the AB cycle . . . 72
6 Intraseasonal variation of convection in the Asian summer monsoon in re- lation to SST and Net Heat Flux 78 6.1 Net Flux and ISO . . . 78
6.2 SST and Heat Flux at the ocean surface . . . 81
6.3 SST gradient, convection (rainfall) and LLJ . . . 86
6.4 Case studies . . . 88
6.5 Summary . . . 91
7 Long Break Monsoon spells in all India drought monsoon years of the re- cent decade 2000-2009 93 7.1 Introduction . . . 94
7.2 Identification of Long Break spells . . . 97
7.3 Results and discussions . . . 98
Contents x 7.3.1 The anomalous Oceanic and Atmospheric features during 2002
2004 and 2009 . . . 98
7.3.2 SST gradient over The Bay of Bengal . . . 102
7.3.3 Prolonged Break spells of 2002, 2004 and 2009 . . . 104
7.4 Summary . . . 109
8 Summary and Conclusions 110 8.1 Scope for Future Studies . . . 116
Bibliography 118
List of Figures
1.1 Variation of OLR with SST . . . 2 1.2 SST - Convection relation in the global tropics-from observation . . . . 3 1.3 Climatology of SST and Convection for different seasons . . . 7 1.4 Climatology of wind at 850hPa and 200hPa for different seasons . . . . 8 1.5 (a) Daily variance of OLR (Jun-Sep) from 1979 to 2007 and (b) 20-90
day band pass filtered variance for the same period. . . 10 1.6 Composites of OLR and Wind associated with active/break period . . . 12 1.7 Intraseasonal variability as a self regulating coupled system . . . 13 1.8 Schematic representation of evolution and northward propagation of
ISO mode . . . 17 3.1 SST-Convection relation over north Indian Ocean during 16 April to 15
May . . . 28 3.2 Divergence-Convection and vertical velocity-Convection relation over
over north Indian Ocean during 16 April to 15 May . . . 28 3.3 SST-Convection relation over west Pacific during July and August . . . 30 3.4 Divergence-Convection and vertical velocity-Convection relation over
west Pacific during July and August . . . 30 3.5 SST-Convection relation over north Indian Ocean during July and August 33 3.6 Divergence-Convection and vertical velocity-Convection relation over
north Indian Ocean during July and August . . . 33 3.7 SST-Convection relation over ITCZ of east Pacific during July and August 35 3.8 Divergence-Convection and vertical velocity-Convection relation over
ITCZ of east Pacific during July and August . . . 35 3.9 SST-Convection relation of eastern and western side of ITCZ . . . 37 3.10 SST-Convection relation over SPCZ of west Pacific during Januar and
February . . . 39 3.11 Divergence-Convection and vertical velocity-Convection relation over
SPCZ area during January and February . . . 39 3.12 The daily a) SST gradient and GPI b) 850hPa zonal wind and wind
shear for -10 to +20 days of the AB cycle . . . 41 3.13 Schematic diagram showing the relation between SST, Convection, Wind
flow and LLJ in Warm pool, ITCZ and SPCZ areas . . . 43 xi
List of Figures xii 3.14 Relation between variations of the fraction of the region<0 with 500-
hPa vertical velocity and SST for observations and CGCM simulation over the Indian Ocean and the western Pacific . . . 45 3.15 Over the ITCZ area of east Pacific Ocean MPI ECHAM5 simulated
Monthly mean SST, precipitation, 500hPa Vertical velocity (Pascal/s), Monthly mean surface wind and its Divergence of July and August of 100 years 1900 to 1999 . . . 45 3.16 MPI ECHAM5 simulated, SST - Convection mm/day relation in monthly
data for July and August of ITCZ and SPCZ area . . . 46 3.17 Over the SPCZ area of west Pacific Ocean MPI ECHAM5 simulated
Monthly mean SST, precipitation, 500 hPa vertical velocity, Monthly mean surface wind and Divergence of July and August of 100 years 1900 to 1999 . . . 47 4.1 The Cold Pool of the Bay of Bengal . . . 49 4.2 Average zonal wind (850hPa) over peninsular India for monsoon of 1998 50 4.3 Average SST of North Bay, Cold Pool and its gradient for monsoon of
1998 . . . 51 4.4 Hovmuller of OLR during 01 June to 30 September 1998 . . . 52 4.5 Hovmuller of SST during 01 June to 30 September 1998 . . . 52 4.6 The average monthly mean SST for 1950-2000 for the 6 models that
simulate the basic features of observed Cold Pool. . . 54 4.7 The vertical profile of temperature averaged over the Cold Pool and
north Bay (ARGO and model) . . . 55 5.1 Composites of SST, GPI rain and the 850 hPa wind of the period 01
July to 31 August of the 7-years 1998 to 2005(excluding 2002) . . . 61 5.2 The SST of the north Bay of Bengal for the 11 cases of AB cycle . . . . 63 5.3 SST of the 11 case composite of pentads -2 to +5 of the AB cycle . . . 64 5.4 SST anomalies of the 11 case composite of pentads -2 to +5 of the AB
cycle . . . 64 5.5 Composite of GPI rain in mm/day of pentads -2 to +5 of the 11 cases of
AB cycle. . . 66 5.6 Composite of TMI rain rate in mm/hr of pentads -2 to +5 of the 11 cases
of AB cycle. . . 66 5.7 Hovmuller of 11 case composite daily GPI and composites of GPI rain
from -12 to +12 of a 40 day AB cycle . . . 68 5.8 Composite of 850 hPa wind in ms−1 of pentads -2 to +5 of the 11 cases
of AB cycle. . . 70 5.9 Linear correlation coefficient (LCC) between GPI rain and 850 hPa
zonal wind. . . 71 5.10 11 AB cycle composites of SST, 850hPa wind and GPI over northern
and equatorial Indian Ocean . . . 72
List of Figures xiii 5.11 Anomalies of temperature and water vapor mixing ratio during the ac-
tive and weak phases of monsoon. . . 75 6.1 SST anomalies in◦C of the 11 case composite of pentads 0 and 4 of the
AB cycle. . . 80 6.2 The heat flux components for 0 and 4 pentads of AB Cycle . . . 83 6.3 Net heat Flux and SST variation for pentad -2 to 5. . . 84 6.4 Net heat flux in Wm−2 of the composite AB cycle (mean of 11 cases)
for pentads -2 to 5. . . 85 6.5 Lag correlation between SST gradient and GPI . . . 87 6.6 Five day moving average of SST gradient and convection of the Bay of
Bengal and the 850 hPa zonal wind through peninsular India for 1999 . 88 6.7 Five day moving average of SST gradient and convection of the Bay of
Bengal and the 850 hPa zonal wind through peninsular India for 2002 . 89 6.8 Five day moving average of SST gradient and convection of the Bay of
Bengal and the 850 hPa zonal wind through peninsular India for 2005 . 90 6.9 The daily average rainfall of India of 11 May to 30 September for 2002
and 2005. . . 91 7.1 % Departure of all India seasonal monsoon (JJAS) from the normal and
daily observed rainflall variation over the Indian landmass for a) 2002 b) 2004 and c) 2009 . . . 95 7.2 Observed JJAS SST and GPCP anomalies for the monsoon season of a)
2002, b) 2004 and c) 2009. . . 99 7.3 Observed velocity potential anomalies (106m2s−1- shaded) at 200 hPa
and divergent wind flow in (ms−1) for the monsoon season (June-September) of a) 2002, b) 2004 and c) 2009. . . 100 7.4 Anomalies of OLR, gph and 850 hPa wind longitude pressure section
during long breaks . . . 101 7.5 SST gradinet and standardised GPI anomaly for 2002, 2003 and 2004 . 103 7.6 GPI anomaly 850hPa wind and relative vorticity for a break case of
2002 (case-1) . . . 106 7.7 GPI anomaly 850hPa wind and relative vorticity for a break case of
2004 (case-2) . . . 107 7.8 GPI anomaly 850hPa wind and relative vorticity for a break case of
2009 (case-3) . . . 108
List of Tables
3.1 Linear Correlation coefficient between SST and Convection, Divergence and Convection and SST and Divergence and the multiple correlation coefficient between these parameters for the different areas of the Global tropics. . . 37 7.1 List of break monsoon days identified for the 3 monsoon droughts dur-
ing 2000 to 2009. . . 98
xiv
Abbreviations
ABcycle ActiveBreak cycle
AMIP AtmosphericModelIntercomparisonProject BoB Bay ofBengal
BOBMEX BayOfBengalMonsoonExperiment CAPE ConvectiveAvailablePotentialEnergy
ECMWF EuropeanCentre forMedium-RangeWeatherForecasts ENSO ElNinoSouthernOscillation
FGGE FirstGARP GlobalExperiment
GARP GlobalAtmosphericResearchProgram GPCP GlobalPrecipitationClimatologyProject GPI GlobalPrecipitationIntex
HRC HighlyReflectiveClouds IOP IndianOceanPanel
IPCC Inter-governmentalPanel onClimateCchange ISO IntraSeasonalOscillations
ISV IntraSeasonalVariability
ITCZ InterTropicalConvergenceZone
JASMINE JointAirSeaMonsoonInteractionExperiment LLJ LowLevelJetstream
MJO MadanJulianOscillations MLD MixedLayerDepth
NASA NationalAeronautics andSpaceAdministration xv
Abbreviations xvi NCAR NationalCenters forAtmosphericResearch
NCEP NationalCenters forEnvironmentalPrediction OLR OutgoingLongwaveRadiation
QuickSCAT QuickScattermeter
SPCZ SouthPacificConvergenceZone SST SeaSurfaceTemperature
TEJ TropicalEasterlyJetstream TMI TRMMMicrowaveImager
TRMM TropicalRainfallMeassuringMission
Symbols
Qnet Net Flux W m−2
Qe Latent Heat Flux W m−2
Qsen Sensible Heat Flux W m−2
Ql Net Longwave Radiation W m−2
Qs Net Insolation W m−2
ρ Density of Air kgm−3
Cp Specific Heat Capacity of air at Constant Pressure J(kg.K)−1
Ts Sea Surface Temperature ◦C
θ Near Surface Air Potential Temperature ◦C Ce Turbulent Exchange Coefficients for Latent Heat m2s−1 Csen Turbulent Exchange Coefficients for Sensible Heat m2s−1
qs Sea Surface Specific Humidity ss
qa Near Surface Atmospheric Specific Humidity g/kg
e Water Vapor pressure Pa
k Cloud Cover Coefficient –
Q0 Outgoing Longwave Radiation W m−2
C Cloud Amount %
xvii
Dedicated to my father and mother . . .
xviii
Chapter 1
Introduction
Atmosphere is in contact with the ocean surface and Sea Surface Temperature (SST) has an important role in producing deep convective clouds and weather systems in the atmosphere, particularly in the tropics. Atmosphere and ocean interact with each other at all time scales. In the variability of the Asian summer monsoon, atmosphere-ocean interaction has a very important role as shown by recent research.
This thesis has studied the SST-Convection relationship, particularly in relation to the Active-Break cycle of the Asian summer monsoon.
1.1 Sea Surface Temperature and deep moist con- vection in the atmosphere above
The relationship between SST and large-scale convection has been studied in a variety of contexts during the 1960s and 1970s. One important aspect of the character of deep convection revealed through such studies is that it generally occurs more frequently and with more intensity as SSTs become higher. (eg. Bjerknes, 1966, 1969). According to Gadgil et al., (1984) organized convection over the tropical Indian Ocean occurs when SSTs exceed a threshold or critical value (Tc), but once the threshold is crossed, the intensity of convection is no longer dependant on the SST (Figure1.1). Graham and Barnett, (1987) while confirming this finding for the Pacific and Atlantic Oceans
1
Chapter 1. Introduction 2 finds that when SSTs are above Tc, surface wind divergence is closely associated with convection. It was also found that areas of persistent divergent surface flow coincide with regions where convection appears consistently suppressed even when SSTs are above Tc.
23 24 25 26 27 28 29 30 31
160
180
200
220
240
260 *
*
*
*
*
*
*
*
*
**
**
*
**
*
**
*
*
*
*
**
*
*
*
*
**
*
*
*
*
*
*
**
*
*
**
*
*
*
*
**
***
*
*
*
*
**
*
*
**
*
*
*
**
*
**
*
***
*
*
*
*
**
***
**
**
*
*
*
*
*
***
**
*
*
*
*
**
**
**
**
**
*
*
*
*
*
**
**
***
**
**
*
*
*
*
*
*
*
*
**
**
**
**
**
*
*
*
*
*
*
*
**
**
***
**
**
**
*
*
*
*
**
**
**
**
**
**
**
**
*
*
*
*
*
*
*
**
**
**
**
***
**
*
*
*
*
*
*
**
**
**
***
**
**
**
**
*
*
*
*
**
**
**
**
*
**
*
***
**
**
*
*
*
*
*
*
**
***
** **
*
*
**
**
*
*
*
**
**
** **
** **
***
***
**
*
*
*
*
**
**
**
** **
** **
*
***
**
*
*
*
*
**
*
**
**
** **
** **
*
*
*
*
*
*
*
**
**
*
**
** **
** *
**
**
**
*
*
*
*
*
*
*
**
** **
** **
*
***
*
*
*
*
*
*
*
***
**
***
*
***
*
*
*
*
*
*
**
**
**
**
**
*
*
*
*
*
*
*
*
**
**
*
**
*
*
*
*
*
S S T (oC )
OLR
24 26 28
180
200
220
240
260
280
S S T (oC )
a) b)
30
Figure 1.1: a) Variation of OLR (Wm−2) with SST over the Indian Ocean (15◦S-20◦N, 60◦E-100◦E) for July (1982-1998). (b) Variation of the mean and standard deviation of OLR with SST. (from Gadgil 2003, based on Gadgil et
al.,1984).
Studying data from the global tropics Waliseret al.,(1993) found that the character- istic of the relationship between SST and deep convection is such that at temperatures between 26◦C and 29◦C convection increases with increasing SST but above about 29- 30◦C, the intensity of convection observed tends to decrease with increasing SST. They found that maximum convective activity does not occur over the warmest ocean but rather the warmest SST occurs under clear or less convective skies. The SST - convec- tion relation found by them using monthly mean SST data of the entire global tropical oceans and monthly mean convection represented by Highly Reflective Clouds (HRC) and Outgoing Longwave Radiation (OLR) are given in Figure1.2.
According to Zhang, (1993) high SST is necessary for deep convection but there is no evidence of a critical SST for exciting convection. He found that the increase
Chapter 1. Introduction 3 in convection with SST is smooth and continuous and no distinct changes in the SST- convection relation occur at any value of SST. While the surface moisture convergence induced by large scale SST gradient (Lindzen and Nigam, 1987) promotes large scale lifting of the moist air in the boundary layer and thereby generation of deep convective clouds, local mean SST is also important for convection. The moisture supply from surface convergence together with fluxes of latent and sensible heat from the warm sea surface into the atmosphere make conditions favorable for deep convection. He however has no explanation for the decrease of convection above SST of 29-30◦C as observed by Waliseret al.,(1993).
Figure 1.2: SST - Convection relation in the global tropics (from Waliseret al., 1993) for monthly values in 2◦ x 2◦ latitude Longitude squares of global tropics 25◦S to 25◦N for the period 1975 to 1985. (a) SST (◦C) and HRC and (b) SST (◦C) and OLR. The right vertical axis specify the number of parameter pairs (marked by the line graph) used. It may be noted that very high SST values also have large numbers of the order of 102 to 103. The standard deviation of
the means is delineated by the shading.
In regions of large scale ascending motions in the atmosphere there is reduction of OLR (increase in deep convection) with respect to increasing SST and the rate of OLR reduction is found to be a strong function of the large scale motion field (Lau et al., 1997). OLR sensitivity to SST is approximately -4 to -5 Wm−2C−1 in the SST range 27-28◦C under conditions of weak large scale circulation. Under the influence of strong ascending motion in the atmosphere this rate can increase to -15 to -20 Wm−2C−1
Chapter 1. Introduction 4 for the same SST range. They found that there is no threshold of SST like 29-30◦C beyond which convection decrease with increase of SST. They also found that under the influence of strong large scale rising motion, convection does not decrease but increase monotonically with SST even at SST higher than 29-30◦C. The reduction in convection observed in high SST situations is likely to be caused by large scale subsidence forced by nearby or remotely generated deep convection. In central Pacific where the local correlation of SST to OLR is largest, they found that the large scale circulation has relatively stronger control than local SST on OLR variance. On monthly to inter- annual time scale they found that 26 % of OLR variance is due to SST and 44 % due to upper tropospheric divergence.
Deep convection has been shown by a modeling study to be generated by SST gradi- ent in the tropical areas of the Pacific Ocean (Lindzen and Nigam, 1987). SST gradient has been found to be an important forcing mechanism of the low level tropical flow and convergence. A large fraction of the total convergence in the eastern tropical Pacific is forced by the meridional gradients in SST whereas in the central and western tropical Pacific, it is the zonal gradients of SST that are the major contributors. They find that the zonal gradients in SST, although smaller than the meridional gradients, can force a response comparable to that forced by the latter.
The relation between SST gradients and atmospheric boundary layer convergence and the generation of deep convection over the tropical oceans was examined in a recent modelling study by Back and Bretherton (2009). A linear mixed layer model which re- produces observed surface winds and convergence over the tropical oceans was used to examine the relative roles of boundary layer and free tropospheric processes on the dis- tribution of convergence and convection under climatological conditions. They found that the distribution of convergence is primarily due to SST-gradient. The authors con- cluded that the climatological distribution of boundary layer convergence is primarily a function of the pattern of SST-gradients and is better regarded as a cause rather than a consequence of deep convection.
Chapter 1. Introduction 5 The nature of the feed backs between atmospheric convection and surface conditions of the Bay of Bengal is found important for understanding the variability of convection there. Measurements during field experiments such as BOBMEX (Bhat et al., 2001) and JASMINE (Websteret al.,2002) have given large amount of information about the interaction between SST and convection on the intraseasonal time scale. Deep con- vection is an atmospheric phenomenon, an outcome of the destabilizing tendency of the atmospheric surface boundary layer on the one hand and the stabilizing/inhibiting influence of the stably stratified atmosphere above on the other. SST is important be- cause the properties of the atmospheric boundary layer are closely linked to the SST, due to continuous interactions of the upper layer of the ocean and the atmosphere im- mediately above (Bhatet al.,1996). When the atmosphere is free of clouds, the ocean surface receives net heat and SST increases leading to increased deep convection in the atmosphere. When deep convection occurs, the clouds block solar radiation and latent heat flux increases due to increased winds. Hence the ocean cools. This can lead to reduced convection. Thus, whereas atmospheric convection depends upon the SST, the SST changes in response to atmospheric convection and the system is thus coupled.
Therefore understanding the seasonal evolution and the intraseasonal variation of SST and air sea coupling are of critical importance in the studies related to deep convection.
The author finds that SSTs are above the convection threshold Tc (Gadgilet al.,1984) from mid-March onwards over the Arabian Sea and throughout the year over the Bay of Bengal.
In summer the Bay of Bengal forms a part of the warm pool covering the tropi- cal east Indian and west Pacific Oceans (Joseph 1990a and Vinayachandran and Shetye 1991). The warm pool has high climatological SST accompanied by persistent deep convection in the atmosphere. From the data obtained from moored buoys Sengupta and Ravichandran (2001) illustrate the relation between Intraseasonal oscillation of SST and surface heat fluxes. Monsoon ISO consist of alternating episodes of active and sup- pressed atmospheric convection moving northward over this area. Negative/positive SST anomalies generated by fluctuations of net heat flux at the ocean surface move
Chapter 1. Introduction 6 northward following regions of active/suppressed convection (Sengupthaet al., 2001).
A coherent relationship among convection, surface fluxes and SST has been found on intraseasonal time scales in the Indian Ocean, Maritime Continent and west Pacific re- gions of the tropics by Woolnough et al., (2000), using lag-correlation analysis and compositing technique on 15 years of data. Prior to the maximum in convection there are positive short wave radiation and latent heat flux anomalies into the sea surface fol- lowed by warm SST anomalies about 10 days before the maximum in deep convection.
Coincident with the maximum in convection there is a minimum in the short wave flux, followed by a cooling due to increased evaporation associated with enhanced westerly wind stress leading to negative SST anomalies about 10 days after the convection has reached maximum. The spatial scale of the anomalies is about 60 degrees longitude, consistent with the scale of the intraseasonal oscillation
1.2 Ocean atmosphere coupling in the annual cy- cle
The local and remote variations of SST have a significant influence on monsoon. But how this mode of SST variability alter the monsoon circulation and associated convec- tion is not yet understood properly. Diagnostic and modeling studies suggested that there were persistent large-scale, low frequency changes in the heat transport and stor- age of the Indian Ocean. These changes were linked to wind forcing (Wyrtki,1973;
Anderson and Rowlands, 1976) in early 80’s. It was noted that there was a substan- tial wind driven meridional oceanic heat flux (Hasntenrath and Lamb, 1978; Hsuing et al., 1989; Hastenerth and Greischer, 1993). Using observational data and model estimates, Loschnigg and Webster (2000) show a strong annual cycle of ocean heat transfer with a seasonal peak with the direction of the flux out-of-phase with solar heat- ing. During the northern hemisphere summer the heat flux is directed southwards to the winter hemisphere, reversing during the boreal winter. The seasonally reversing heat flux is suggestive of a possible feedback between the ocean and the atmosphere that
Chapter 1. Introduction 7 may regulate the strength of the monsoon (Loschingg and Webster, 2000; Webster et al.,2002; Loschnigget al., 2003). The atmospheric circulation respond thermodynam- ically through variations in heating and water vapor supply in surface fluxes in to the atmosphere producing feedbacks as suggested by Ramanathan and Collinss (1991) in their natural thermostat hypothesis posed in an attempt to explain why the warm pool of the western Pacific Ocean appears to be constrained to remain in the 28-30◦C SST change.
20N
0
20S 20N
0
20S 20N
0
20S 20N
0
20S 60E 90E 120E 150E 180
2 4 6 8 10 12
22 23 24 25 26 27 28 29 30
DJ
AM
JA
ON
DJ
AM
JA
ON
60E 90E 120E 150E 180
SST GPCP
Figure 1.3: Climatology of SST(◦C) and Convection (mm/day) for different seasons: December-January, April-May, July-August and October-November.
Figure 1.3 shows the annual cycle of the mean SST and the precipitation (GPCP) for the four seasons: December to January (DJ), April to May (AM), July to August (JA) and October to November (ON). The transition of SST is clearly seen from season to season. During the premonsoon the SST over Indian Ocean north of 10◦S attains a maximum (> 29◦C) which leads to the warm pools of the summer season. During boreal summer northern part of Indian Ocean (north of 10◦N) begins to rise in its tem- perature which extend up to western Pacific. With the progress of south west monsoon surface winds become stronger over the Arabian sea leading to reduction in SST. The
Chapter 1. Introduction 8
25N
0
20S 25N
0
20S 25N
0
20S 25N
0
20S
60E 90E 120E 150E 180 60E 90E 120E 150E 180
DJ
AM
JA
ON
DJ
AM
JA
ON
850 hPa 200 hPa
Figure 1.4: Climatology of wind at 850hPa and 200hPa (ms−1) for different seasons: December-January, April-May, July-August and October-November.
winter time SST is generally cooler by 1◦C or more compared to other seasons. Most notable is that SST appears to change little during the boreal spring over much of the north Indian Ocean despite weak winds and high insolation. Similarly the winter SST in the north Indian Ocean is only a degree or so cooler than in the fall or the spring despite significant reduction in the solar heat flux.
During the summer, maxima in precipitation (GPCP) occur over the south Asian region with an extension towards west Pacific and the near equatorial southern hemi- sphere. During the boreal winter, an elongated band extends across the Indian Ocean and Australia with a maximum over Indonesia and north of Australia. The lower tropo- spheric circulation of the Asian summer monsoon is stronger than in all other seasons and possesses a concentrated cross equatorial flow (Findlater, 1969). A weak reverse flow during the boreal winter is also observed (Figure 1.4). South of 10◦S persistent southeasterly trade winds exist throughout the year. During the Asian summer mon- soon season June to September strong Tropical Easterly Jet is a persistent feature in the upper tropospher. Subtropical Jet stream (STJ) is strong during the other seasons
Chapter 1. Introduction 9 north of 20◦N. The upper tropospheric flow pattern during the summer emphasises the thermal contrast between continents and oceans (Krishnamurti 1971a,b).
1.3 Monsoon Intraseasonal variability in relation to the ocean
Intraseasonal variations are now recognized as the elementary building blocks of the monsoons; hence, monsoon simulation and prediction hinges on the ability to simulate it. Recent studies provide solid evidence that a successful simulation of Monsoon In- traseasonal Oscillation and MJO requires a coupled model with an active ocean that modulates atmospheric convection through intraseasonal SST fluctuations (Kemball- Cooket al., 2002; Fuet al.,2003; Zheng et al., 2004; Rajendranet al., 2004a). Even though the model simulation is not adequate at the present time, the skill of statistical predictions is evidence that models can be improved, in part based on a better under- standing and representation of oceanic upper-layer physics, and on a better choice of initial oceanic conditions. Fig 1.5(a) gives the daily variance in the outgoing longwave radiation (OLR) for the monsoon season June to September. High variability is notice- able over the equatorial Indian Ocean, the northern Indian Ocean including the Arabian Sea and Bay of Bengal and the northern tropical west Pacific. Of the three regions, the northern tropical west Pacific exhibits the largest variance. Fig1.5(b) Shows the vari- ance at intraseasonal time scales during the same season (10-90 day band pass filtered).
The pattern of the variance is almost similar in both (a) and (b) and so the total mon- soonal rain is assumed to be a manifestation of intraseasonal variance (Annamalai and Sperber 2005).
Monsoonal intraseasonal oscillations occurs over the region of warm pools of the Indian and west Pacific Oceans where the mixed layer is shallow and hence can easily be modulated by the winds and the heat and moisture fluxes. The upper-ocean stratification under rapidly changing wind speed and precipitation and salinity variations determine the mixed-layer depth.
Chapter 1. Introduction 10
a) Variance
a) 20-90 day filtered variance
Figure 1.5: (a) Daily variance of OLR (Jun-Sep) from 1979 to 2007 and (b) 20-90 day band pass filtered variance for the same period.
Pronounced SST anomalies in the Indian Ocean are observed during boreal sum- mer when the intraseasonal oscillation is propagating northward. During the convec- tive phase, cloud cover reduces insolation and with strong westerlies latent heat flux increases, producing surface cooling. A warm SST anomaly leads the northward prop- agating convective anomaly by 1-2 weeks (1/4 ISO cycle) and a cold SST anomaly fol- lows the convective anomaly by a similar lag. The magnitude of these SST anomalies, in the Bay of Bengal, can be much larger than the near- equatorial SST anomalies asso- ciated with the eastward propagating MJO (Sengutpa and Ravichandran 2001). There is shallower and more stably stratified mixed layer with a deeper barrier layer in the BoB due to enhanced precipitation and large river runoff than in the western equatorial Pa- cific (Bhatet al.,2001; Sengupta and Ravichandran 2001; Websteret al.,2002). Hence, the mixed layer in the Bay of Bengal is more sensitive to the intraseasonal surface heat flux variations.
Understanding how SST and the upper ocean mixed layer vary intraseasonaly is important both for coupled simulation and for validation of air-sea interaction theories.
Chapter 1. Introduction 11 Not much work has been done on the mechanisms of SST variability in the northern In- dian Ocean associated with northward propagating intraseasonal oscillations during bo- real summer. Recent field campaigns (JASMINE and BOBMEX) and new satellite SST products have revealed some similarities and differences with the behavior of the near- equatorial warm pool. Observations in the Bay of Bengal reveal a near-quadrature rela- tionship between net surface heat flux variations associated with intraseasonal oscilla- tions in convection and SST, consistent with the idea of SST variation is primarily driven by the surface heat flux variation (Sengupta and Ravichandran 2001). The variation of mixed layer depth and penetration of short wave radiation, Sengutpa and Ravichandran (2001) showed qualitative agreement between the surface heat flux variation and ob- served intraseasonal SST tendency. Strong mean winds ( 10 ms−1), strong rainfall and runoff in the region maintains a stable mixed layer upon a deep barrier layer, which prevents mixing of cold sub-thermocline water into the mixed layer (Bhatet al.,2001).
A modeling study by Loschnigg and Webster (2000) suggest the oceanic poleward heat transport integrated across the entire Indian Ocean varies in association with the pole- ward propagating intraseasonal events. Their result also implies that the intraseasonal oscillation drives changes in the wind-driven circulation of the Indian Ocean, but it is not clear how or whether these changes in heat transport relate to the associated SST changes.
Intraseasonal variability of the monsoon occurs over a vast region extending over the entire equatorial Indian Ocean and the Bay of Bengal, and extend up to west Pa- cific. Figure 1.6 shows the composite precipitation for the active and the break period of the monsoon. During the active period a coherent positive precipitation anomaly zone exists over the northern Arabian sea extending through north peninsular India to the west Pacific through Bay of Bengal. During the composite break period the pre- cipitation maxima are along the equator and along the foothills of Himalaya with a considerable minimum over peninsular India. Such large scale low frequency variabil- ities are seen in precipitation, OLR and wind fields (Webster et al., 1998, Joseph and Sijikumar, 2004) and bring periods of extended drought and excess rainfall on 30-60
Chapter 1. Introduction 12 day time scales throughout the monsoon region. Studies of Krishnamurtiet al.,(1988) to the latest studies such as Wanget al.,(2006) and Annamali and Sperber (2005) sug- gest that the intraseasonal variability of the monsoon exists over the entire Indo-West Pacific Ocean basins.
Figure 1.6: Composites for active/break monsoon days in Jun-Aug of 1979-90.
(a/c) OLR (Wm−2), (b/d) 850-hPa wind vectors. (from Joseph and Sijikumar 2004).
Studies of Stephenset al.,(2004); Wanget al.,(2005); Lau and Sui (1997); Fasullo and Webster 1999, have provided evidence of tropical Intraseasonal variability as a self regulating coupled system. Stephenset al.,(2004) find that the Intraseasonal oscillation has 3 phases. 1) Destabilization phase: The atmosphere becomes increasingly unstable through a radiative cooling of upper atmosphere, the gradual warming of the SST and the development of the shallow boundary layer cumulus. 2) Convective phase: large scale convection develops that results in heavy precipitation which leads to deepening of the mixed layer, decreasing of SST, and moistening of the upper troposphere. 3) Restoring phase: In this phase strong winds associated with high humidity stabilize the atmosphere and suppress convection. Figure 1.7 shows a schematic of the 3 phases of
Chapter 1. Introduction 13 Stephenset al.,(2004). The Wanget al.,(2005) ISO sequence involves a total of eight phases.
Figure 1.7: A schematic of the: SST variation as a function of time (top) which is used to illustrate the three phases of the feedback. The atmospheric temperature and moisture change during these phases (middle). Changes of cloud conditions and associated wind field and heating changes are given (bot-
tom). (from Stephens et al., 2004).
By examining the data during 1979 from First GARP Global Experiment (FGGE) Krishnamurtiet al.,(1988) found that ISOs involve significant fluctuations of SST and latent heat flux over the BoB and western Pacific. For the ocean to play an active role in the dynamics of the eastward propagating MJO and the northward propagating ISO variability in the Indian monsoon, the surface flux variations must induce a SST varia- tion. Krishnamurtiet al.,(1988), motivated by the need to explain the long time scale of the MJO, were the first to examine SST variability associated with the MJO. Using data from the FGGE year, they showed that intraseasonal SST variability was most promi- nent across the equatorial Indian and western Pacific Oceans and had temporal phasing
Chapter 1. Introduction 14 indicative of the atmosphere forcing of the ocean on the intraseasonal timescale. Sen- guptha and Ravichandran, (2001) also studied Intraseasonal fluctuations of SST and net heat flux over the Bay of Bengal from measurements from a few moored buoys.
Measurements during field experiments such as BOBMEX (Bhatet al.,2001) and JAS- MINE (Websteret al.,2002) also show large Intraseasonal fluctuations of net heat flux.
Using high resolution SST measurements using microwave sensors, TMI on board the TRMM satellite Senguptaet al.,(2001) describe the spatial distribution of Intraseasonal variability of SST over the Indian Ocean region. Using SST wind speed and OLR data they showed that the Intraseasonal SST oscillation with an amplitude of 0.6-0.8◦C has a large horizontal scale similar to that of atmospheric ISO and possessed northward movement in the region coherent with OLR, surface wind speed, precipitation and SST during the summer monsoon season.
Vecchi and Harrison (2002) indicate that there is a phase lag between SST and con- vection which could be exploited to predict the monsoon breaks. The warm (cold) SST band follows the dry(rainy) band with the time lag of 7-10 days. The speed of north- ward movement of SST is same as that for the atmospheric field but lags behind the Qnet by about 7 days. Waliser et al., (2003a) studied an ocean model with a mixed layer with daily wind stresses and heat fluxes associated with the atmospheric ISOs.
While the net heat flux associated with ISOs is a major forcing, they found that the mixed layer depth (MLD) variations tended to contribute positively to the SST varia- tions. Goswami (2005) suggests that the relationship between SST and Qnet indicates that the Intraseasonal SST fluctuations are driven by the atmospheric ISV through Qnet.
But the quadrature relationship between SST and Qnet is not found to hold good over the equatorial Indian Ocean.
AGCMs forced with prescribed SST exhibit intraseasonal variability with spatial patterns that resemble the observed pattern (Waliser et al., 2003b; Kemball Cook and Wang, 2001; Jianget al.,2004) and include some form of northward propagation. This also support the hypothesis that the basic oscillation and the northward propagation are of atmospheric origin and that the basic temporal scale selection and northward
Chapter 1. Introduction 15 propagation of the atmospheric summer ISOs may arise from internal feedback within the atmosphere. Air sea interaction can modify these internally triggered oscillations.
Kemball-Cook and Wang (2001) find that air sea coupling enhances northward propa- gation of the summer ISOs significantly. Waliseret al.,(2004) find that air sea coupling improves the space time characteristics of the summer ISO over the Indian Ocean. Fuet al.,(2002) also find that ocean atmosphere coupling improves simulation of the ampli- tude and northward propagation characteristics of the ISOs. However Goswami (2005) suggests that some of these modeling results could be influenced by the bias of the mean states of the individual components of the coupled model. Therefore quantitative esti- mate of modification of the summer ISO by air sea coupling is still not well established.
There are so many mechanisms which explain the poleward propagation of 30-60 day mode of ISO. Webster (1983) proposed that the northward gradient of sensible heat flux is responsible for the poleward transition of the cloud band. Goswami and Shukla (1984) explained northward propagation through ground hydrology feedback mechanism. Later Gadgil and Srinivasan (1990) and Nanjundiahet al.,(1992) showed that the northward gradient of moist static stability (poleward side being more unstable than the equator side) is responsible for the poleward propagation of the cloud band.
Keshavamurthy et al., (1986) indicates that the northward movement may arise from zonally symmetric component of an eastward propagating equatorial oscillation inter- acting with the mean flow over monsoon region. Wang and Xie (1997) and Lau and Peng (1990) suggested that the northward propagation results from unstable Rossby waves interacting with the basic monsoon mean flow.
A critical question in explaining the north-west-south-east tilt of the precipitation band and associated northward propagation of ISO is what makes the rossby waves have a northward propagation component. Obviously without mean flows, the em- anated rossby waves move only westward hence it is the presence of the basic flow that induces a northward propagation component. The works of Drbohlav and Wang (2005) and Jinaget al.,(2004) identified that the effect of the easterly vertical shear is an important internal dynamic factor. Kemball-Cook and Wang (2001) and Lawrence
Chapter 1. Introduction 16 and Webster (2001) indicated that the planetary boundary layer convergence maximum, is responsible for the northward propagation of the summer monsoon 30-60 day oscil- lation. Hsu and Weng (2001) suggested that a similar frictional convergence driven by the low level vorticity is responsible for the north-westward movement of the monsoon 30-60 day oscillation.
Jianget al.,(2004) provide a physical process for the pole ward propagation based on a zonally symmetric model. They propose that a combination of vertical wind- shear mechanism and a moisture - convection feedback mechanism is responsible for the poleward propagation of the convection band. They show that the easterly mean ver- tical shear in the region gives rise to the generation of barotropic divergence in the free atmosphere north of the convection center. This leads to boundary layer convergence north of the convection maximum. The mean flow and mean boundary layer humid- ity during summer monsoon season also allow perturbation moisture convergence to be maximum north of the convection center. Both these processes contribute to the poleward propagation of the mode. Near the equator the moisture convection feedback mechanism is the dominant mechanism producing the poleward propagation.
Based on the understanding of the genesis of the zonally symmetric component of the 30-60 day monsoon intraseasonal oscillation from modelling studies (Goswami and Sukla 1984), theoretical analysis by Jianget al.,(2004) and his diagnostic studies, Goswami (2005) has given the evolution of this intraseasonal mode in the meriodnal plane over the south Asian monsoon region schematically in figure 1.8. According to Goswami (2005) low level convergence of moisture associated with SST gradient in a conditionally unstable atmosphere produces organised convection and intensifies the tropical convergence zone over the SST maximum in the equatorial Indian Ocean with subsidence and clear sky conditions over the monsoon trough region to the north. The anomalous hadley circulation has ascending motion over the equatorial Indian Ocean and descending motion over the monsoon trough region. This situation correspond- ing to the break monsoon condition. The cyclonic vorticity at the low level and the
Chapter 1. Introduction 17
Figure 1.8: A schematic representation of evolution and northward propa- gation of the meridional circulation associated with the 30-60 day mode (from
Goswami 2005).
associated boundary layer moisture convergence is maximum north of the zone of max- imum convection and makes it move northward. After about 10 days the convection zone reaches latitude about 10◦N with both the monsoon trough region and the equato- rial Indian Ocean region being under subsidence with clear sky. After another 10 days the convection zone reaches the monsoon trough area around latitude 25◦N. Now the anomalous hadley circulation has assending motion over the monsoon trough region and descending motion over the equatorial ocean. This situation corresponds to active monsoon condition. Under the clear sky condition over the equatorial region coupled with light winds there the net heat flux over the equatorial region becomes positive and the SST there increases. There after a convective band develops over the warm SST zone and the cycle gets repeated.
1.4 Prediction and Predictability of monsoon sys- tem.
In an effort that focused on active and break conditions of the Indian summer monsoon, Goswami and Xavier (2003) noted that all active (break) conditions go over to a break
Chapter 1. Introduction 18 (active) phases after about 15-20 days. Using a rainfall-based ISO index and a defi- nition of active and break conditions, they calculated the ensemble average transition from active to break (and break to active) conditions as a function of lead time. This method exhibited some skill in predicting break-to-active (active-to-break) transitions with about 10 (20) days lead time, indicating that monsoon breaks are more predictable than active monsoon conditions.
Krishnamurtiet al.,(1990) finds that the loss of forecast skill associated with low- frequency modes comes both from the errors and evolution of high wavenumber/fre- quency variability, associated with the boreal summer monsoon. In the prediction of active and break monsoon periods, it is possible to filter the initial state in order to re- move all mean state and the low-frequency modes. They further pointed out that this will delay the contamination of the low frequency modes as a result of the energy ex- changes from the higher frequency modes. This idea was tested in forecasts that used observed SST anomalies. They showed that the model forecast exhibited considerable skill at predicting the meridional motion of the 850-hPa trough-ridge system over the Indian region and the eastward propagation of the 200-hPa divergence anomaly out to about 4 weeks. In Krishnamurtiet al.,(1992, 1995), experiments using two case studies for low-frequency, wet and dry spells over China and Australia for each of their associ- ated summer monsoons with similar results, the SST anomalies were found to be vital to retaining the forecast skill.
Many attempts to simulate the ISO by numerical models have met with difficulties in reproducing its realistic propagation and spatial structure. Comparisons of the ISO in 15 general circulation models (GCMs) as part of the Atmospheric Model Inter com- parison Project (AMIP) Slingo et al., (1996) has shown that most of the GCMs have weaker than observed intraseasonal variability and most of the models tend to simulate slightly shorter periods than observations and failed to capture the seasonality of the os- cillation. Inness and Slingo (2003) note that many GCMS produce some aspects of the ISO but all have problems with determination of the amplitude, the propagation speed,
Chapter 1. Introduction 19 and reproducing the different seasonal aspects. Lin et al.,(2006) evaluated the tropi- cal intraseasonal variability in 14 CGCMs participating in the Inter-governmental Panel on Climate Change (IPCC) Fourth Assessment Report (AR4). Their results showed that current state-of-the-art GCMs have significant problems in simulating tropical in- traseasonal variability. In most of these 14 models the total intraseasonal variance of precipitation is very weak.
Indian Ocean panel (IOP) review report found that despite the critical need for ac- curate and timely monsoon forecasts, our ability to predict seasonal conditions has not changed substantially over the last few decades. The critical need for seasonal predic- tion cannot at this time be filled by coupled, numerical modelling. Dynamical predic- tion is still in its infancy and severely handicapped by the inability of models to simulate the mean monsoon structure and its year-to-year variation (Sperber and Palmer, 1996;
Gadgil and Sajani, 1998), or the intraseasonal (20-50 day period) band which controls a very large percentage of the precipitation variability (Slingo et al., 1996; Waliser et al., 2003a,b). The limited capability for prediction is also due to the unavailabil- ity of good quality ocean data as initial condition. Better data for initialization of the numerical models will in itself improve predictions. Empirical schemes (Waliseret al., 1999; Lo and Hendon, 2000; Wheeler and Weickmann, 2001;Webster and Hoyos, 2004) have shown that regional precipitation characteristics are predictable with considerable accuracy 20-30 days in advance. Good understanding of physics and dynamics and ocean-atmosphere interaction is required to improve models.
1.5 In this Thesis
The purpose of this Thesis is to study in detail the SST-Convection relation over the tropical oceans and ocean atmoshere interaction in the intraseasonal variability of the South Asian Monsoon. More specifically the aim of the thesis is:
