• No results found

Mesoscale data assimilation for simulation of heavy rainfall events associated with south-west monsoon

N/A
N/A
Protected

Academic year: 2023

Share "Mesoscale data assimilation for simulation of heavy rainfall events associated with south-west monsoon"

Copied!
12
0
0

Loading.... (view fulltext now)

Full text

(1)

MESOSCALE DATA ASSIMILATION FOR SIMULATION OF HEAVY RAINFALL EVENTS ASSOCIATED WITH SOUTH-WEST MONSOON

ASHISH ROUTRAY

CENTRE FOR ATMOSPHERIC SCIENCES INDIAN INSTITUTE OF TECHNOLOGY, DELHI

HAUZ KHAS, NEW DELHI-110016, INDIA

DECEMBER 2009

(2)

© Indian Institute of Technology Delhi (IITD), New Delhi-2009.

All rights reserved.

(3)

MESOSCALE DATA ASSIMILATION FOR SIMULATION OF HEAVY RAINFALL EVENTS ASSOCIATED WITH SOUTH-WEST MONSOON

by

ASHISH ROUTRAY

Centre for Atmospheric Sciences Submitted

in fulfillment of the requirements of the degree of

DOCTOR OF PHILOSOPHY

to the

INDIAN INSTITUTE OF TECHNOLOGY, DELHI

DECEMBER, 2009

(4)

Dedicated to

My parents and my wife Bini

(5)

CERTIFICATE

This is to certify that the thesis entitled "MESOSCALE DATA ASSIMILATION FOR SIMULATION OF HEAVY RAINFALL EVENTS ASSOCIATED WITH SOUTH-WEST MONSOON" being submitted by Mr. Ashish Routray for the award of the degree of DOCTOR OF PHILOSOPHY is a record of the original bonafide research work carried out by him. He has worked under our joint guidance and supervision and has fulfilled the requirements for the submission of the thesis. The results presented in this thesis have not been submitted in part or full to any other University or Institution for award of any degree/diploma.

(Dr. U. C. Mohanty) Professor,

Centre for Atmospheric Sciences Indian Institute of Technology, Delhi Hauz Khas, New Delhi-110016, INDIA.

(Dr. Someshwar Das) Scientist-G,

National Centre for Medium Range Weather Forecasting A-50, Institutional Area, Phase-II

Sector-62, Noida, UP- 201307,

INDIA.

(6)

ACKNOWLEDGEMENTS

I would like to thank all those who have contributed directly or indirectly towards completion of this research work. I will endeavor to thank everyone involved, apologies for any glaring omissions. At the outset, I express my sincere gratitude to Prof. U.C. Mohanty, my thesis supervisor, for his valuable guidance, constructive criticism, constant encouragement, affection and love extended throughout this research work.

I am extremely grateful to Dr. Someshwar Das, my thesis supervisor for providing necessary help, constant encouragement, advice and supervision which enabled me to complete this work in the present form. I am much influenced by his humbleness during my association with him.

I am grateful to Prof. O. P. Sharma, Head, CAS, for his encouragement and providing the necessary facilities. Many Thanks to all the faculty members of CAS for their suggestions and support during the research.

I express my thanks to Dr. S. R. H. Rizvi, NCAR, USA for providing me the much needed inputs and valuable suggestions at the start of this research work.

I sincerely acknowledge National Center for Atmospheric Research (NCAR), USA for providing the model and wrfhelp for the timely help extended during crisis periods. The National Center for Environmental Prediction (NCEP) and NCAR, USA are also gratefully acknowledged for providing the analysis datasets.

India Meteorological Department is highly appreciated and acknowledged for providing the observations to compare the model results.

I express my thanks to Ananda-da, Mandal-da, Nelson-Sir, and Bhatlaji for their kind help and cooperation in initial stages of my research at IIT Delhi. I wish to thank all my colleagues and friends, past and present at CAS, IIT Delhi, in particular, Sudhanshu, Subrat, Sankalp, Palash, Sujata, Senthil, Swagat, Jagabandhu, Narender, Makrand, Krishna, Dipak, Kiran, Neetha and others for sharing companionship and making the stay at IIT pleasant. My sincere thanks are also due to all the supporting staff members of CAS for their all sorts of help.

I gratefully acknowledge the financial support of Council of Scientific and Industrial Research, India and Indian National Centre for Ocean Information Services, Government of India.

I thank my family for their unwavering support and encouragement. I thank my parents for instilling a sense of wonder and exploration in me. I thank my brothers, sisters, in-laws and friends especially Kalyani and Lipu for being all along with me. Finally, I am extremely grateful to my wife, Bini and my daughter, Prachi for their patience, cooperation and understanding during the final stage of the thesis.

(Ashish Routray)

(7)

ABSTRACT

During the south-west monsoon (SWM) season heavy to very heavy rainfall events over most parts of the country are caused by organized meso-convective systems. Improving the simulation of the heavy rainfall events is important as such events routinely result in flooding and significant loss of life and property over the Indian monsoon region.

There have been considerable improvements in the field of mesoscale prediction over past few decades using high-resolution state-of-art mesoscale models and these models with advance data assimilation techniques are recently proved to be more successful for the prediction of heavy rainfall events associated with convective activities as well as synoptic systems. Performance of the mesoscale models is sensitive to the quality of initial conditions.

This thesis deals the improvement of numerical simulation of heavy rainfall events during SWM season with improved initial condition through mesoscale data assimilation system. The Weather Research and Forecasting (WRF dynamical core ARW) modeling system along with three dimensional variational (3DVAR) data assimilation technique are used to improve the simulation of these intense events. The conventional, non-conventional and Doppler Weather Radar (DWR) radial velocity and reflectivity data are used in the assimilation system.

Extensive sensitivity experiments with two planetary boundary layer [PBL; Yonsei University Scheme (Y) and Meller-Yamada-Janjic (M)] and three cumulus convection [CU;

Kain-Fritsch (KF), Betts-Miller-Janjic (BM) and Grell-Devenyi (GD)] schemes are conducted to find the possible best (optimal) combination of the physical parameterization schemes for simulation of four monsoonal heavy rainfall events (three along the west coast of India due to the presence of mid-trpospheric circulation (MTC)/off-shore trough and one along the east coast of India due to monsoon depression) during SWM season. Among the three CU schemes considered in this study, the Betts-Miller-Janjic (BM) CU scheme is found to produce more accurate simulation of the heavy rainfall events and track of the monsoon depression (MD) in

1

(8)

combination with the two PBL schemes. The Vector Displacement Errors (VDEs) are reduced in the BM CU scheme throughout the forecast period. The statistical skill scores also revealed that the BM CU scheme is suitable for simulation of heavy rainfall events. In particular, the combination of BM CU scheme with Yonsei University (YSU) PBL scheme provides optimal combination of physical parameterization schemes in simulation of the heavy rainfall events.

The different assimilation strategies (cold start and cyclic) are investigated using WRF- 3DVAR (hereafter WRF-Var) analysis scheme. A series of cold start and cyclic based assimilation experiments with WRF-Var analysis system are conducted using conventional and non-conventional observations for a 30—days period (15 July to 15 August 2005). From the statistical skill score, it is concluded that the cyclic mode assimilation enhanced the performance of the WRF-Var over Indian region. Further highlights the impact of improved model initial condition through WRF-Var using conventional and non-conventional observations on simulation of heavy rainfall event along the west coast of India. The results indicate that, compared to the global analysis, the 3DVAR data assimilation technique substantially improves the overall simulation over the Indian monsoon region.

The impact of assimilation of Doppler Weather Radar (DWR) radial velocity and reflectivity in WRF-Var data assimilation system for the simulation of MD over Bay of Bengal (BOB) is investigated. The Kolkata DWR data is used for the assimilation experiment along with the other conventional and non-conventional data. Assimilation of DWR data with a cyclic mode gradually extracts useful information from Doppler data and improved the model forecast skill. There is a reasonably good depiction of the MD in the model initial condition due to the assimilation of the DWR data. The statistical skill scores also revealed that the precipitation forecasts during day-1 and day-2 are significantly improved after assimilation of DWR data.

The average VDEs in track forecast of the MD is reduced by 51% due to the assimilation of DWR data.

ii

(9)

The evaluation of model performance is carried out using the WRF-Var based improved initial condition in simulating seven heavy rainfall events (two rain events due to MDs over BOB and five heavy rainfall events along the west coast of India) during SWM season. The DWR data obtained from Kolkata and Machlipatanam stations are used in the assimilation cyclic along with conventional and non-conventional observations for simulating the MDs cases. Similarly, the conventional and non-conventional observations are used for simulation of five heavy rainfall cases along the west coast of India.

The intensity and location of the MDs are well represented in model initial time as well as in simulations after assimilation of DWR data. The average VDEs are reduced by 50-60% in the DWR simulation as compared to the other simulations. The time series of minimum mean sea level pressure and maximum surface wind show the intensity and structure of the MDs are better simulated due to DWR data assimilation.

The unprecedented highly localized Mumbai heavy rainfall event (944 mm within 24hrs) is also studied. The amount, intensity, timing and spatial distribution of this unusual rain event is simulated with reasonably good accuracy in the assimilation experiment. The analyses of the dynamical parameters at the location of heavy precipitation revealed that the maximum convergence, maximum vertical velocity and vorticity preceding the mature stage are found in the assimilation simulation.

The orientation of off-shore trough, intensity and location of the MTC along the west coast of India during the heavy rainfall events are well represented in the assimilation simulation. The statistical skill scores are improved in the assimilation simulations, which gives better rainfall forecasts. The assimilation experiments are able to capture the location and amount of rainfall over the west coast of India reasonably well as compared to without mesoscale data assimilation simulations.

iii

(10)

CONTENTS

Page No.

Abstract i-iii

Contents iv-vi

List of Figures vii-xiii

List of Tables xiv-xv

CHAPTER — I: Introduction 1-17

1.1 Introduction 1

1.2 Rain-bearing Systems during Monsoon Season 2

i) Monsoon depression 2

ii) Mid Tropospheric Cyclones 3

iii) Off-shore trough 4

iv) Monsoon trough 4

1.3 Numerical Weather Prediction 5

1.4 Non-hydrostatic Mesoscale Models 6

1.5 Data Assimilation 9

1.6 Simulation of Heavy Rainfall Events during SWM Season 11

1.7 Objectives and Scope of the Thesis 13

CHAPTER — II: Sensitivity Study with Physical Parameterization

Schemes for Simulation of Heavy Rainfall Events 18-55

2.1 Introduction 18

2.2 Model Description 20

2.2.1 Vertical Coordinate 21

2.2.2 Flux-Form Euler Equations 22

2.2.3 Inclusion of Moisture 23

2.2.4 Governing Equations in the WRF Modeling System 24

2.2.5 Model Physics 25

a) Cumulus Convection 26

b) Planetary boundary layer (PBL) 26

2.3 Synoptic Conditions during the Monsoonal Heavy Rainfall Events 27

2.4 Experimental Design 28

2.5 Results and Discussion 31

2.5.1 Mean Sea Level Pressure (MSLP) 31

2.5.2 Wind fields at 2.1 km Height 35

2.5.3 Wind fields at 4.5 km Height 39

2.5.4 Precipitation 42

2.5.5 Additional Case Studies 48

2.6 Conclusions 54

iv

(11)

CHAPTER — III: Data Assimilation System and its Impact Studies 56-94

3.1 Introduction 56

3.2 Synoptic Situation during 25-28 June 2005 59

3.3 WRF-Var Analysis System 61

3.3.1 Background error covariances (B) 63

3.4 Numerical Experiments 64

3.4.1 Calculation of Background Error Covariances 64 3.4.2 Different Approach of Assimilation with WRF-Var

Analysis System 65

3.4.3 Impact of Observations 65

3.5 Data Pre-processing 67

3.6 Result and Discussion 69

3.6.1 Single Observation Test 70

3.6.1.1 Single temperature observation 71 3.6.1.2 Single u-wind component observation 72 3.6.2 Verifications against Observations and Analyses 73 3.6.3 High-Resolution Reanalysis and Model Initial Condition 75

3.6.4 Impact on the WRF Model Simulation 81

3.6.4.1 Wind fields 81

3.6.4.2 Vorticity 87

3.6.4.3 Precipitation 89

3.7 Conclusions 93

CHAPTER — IV: Impact of Doppler Weather Radar Radial Velocity

and Reflectivity on Mesoscale Simulations 95-125

4.1 Introduction 95

4.2 Methodology for Doppler radar data Assimilation 97

4.3 Synoptic overview of Monsoon Depression 98

4.4 Processing of Indian DWR Data 100

4.5 Numerical Experiments 102

4.6 Results and Discussion 103

4.6.1 Improved Model Initial Condition 103

4.6.2 Impact of DWR Data on Model Simulations 107

4.6.2.1 MSLP 107

4.6.2.2 Wind 109

4.6.2.3 Simulated Rainfall and Composite Reflectivity 113

4.6.2.4 Relative vorticity 119

4.6.2.5 Vertical Cross-sections 120

4.6.2.6 Track Prediction 123

4.7 Conclusions 124

V

(12)

CHAPTER—V: Evaluation of Model Performance in Simulation of

Heavy Rainfall Events using Improved Initial Condition 126-167

5.1 Introduction 126

5.2 Synoptic Overview 128

5.2.1 Monsoon Depression 128

5.2.2 Heavy Rainfall Events along West Coast of India 130

5.3 Numerical Experiment 132

5.4 Results and Discussion 133

5.4.1 Heavy rainfall due to MDs 134

a) Mean sea level pressure (MSLP) 134

b) Wind 137

c) Precipitation 140

d) Track Prediction 143

5.4.2 Heavy rainfall due to off-shore trough/MTC 145

5.5 Conclusions 165

CHAPTER — VI: Summary of Results and Future Prospects 168-176

6.1 Summary of Results 168

6.2 Future prospects 175

REFERENCES 177-191

Acronyms 192-193

BIO-DATA 194-198

vi

References

Related documents

The expansion of flood forecasting activity was done using the rainfall runoff model by developing MIKE 11 model using the daily rainfall received through India Meteorological

The identification of active and break events is based on the daily rainfall data averaged over the monsoon core zone which is coherent with respect to intraseasonal variation and

The LPSs that lead to medium and large EREs tend to occur during the positive phase of MISO when an active monsoon trough is present over central India.. On the other hand, LPS-noex

Synoptic conditions associated with Dahanu heavy rainfall event, 2016, Mumbai floods, 2005 and historical heavy rainfall events over North Konkan region are considered in

CHAPTER 3: Simulation of mesoscale features associated with intense western 81-133 disturbances over western Himalayas using a mesoscale

The soil erosion simulation model has been verified by corn- paring the simulated total runoff and soil loss amounts,resulting from a number of rainfall events,with the reported

Instigating ROMS-SST as a boundary condition to the WRF model improved the TC intensity, spatial structure, and rainfall location for the TC Phailin, while a minor improvement

(ii) Out of 122 days of the monsoon period, the impact- based warning was issued on 16 occasions and the severe weather was observed on eight numbers of days in Dehradun city with