bay of bengal cyclones
DESCRIPTION
about bay of bengal cyclones..TRANSCRIPT
National Conference
Bay of Bengal Tropical Cyclone Experiments
(BOBTEX
1
Cyclone Warning Division
India Meteorological DepartmentMausam Bhawan, Lodi Road
New Delhi
Extended Abstract
National Conference
on
Bay of Bengal Tropical Cyclone Experiments
(BOBTEX-2011)
New Delhi 1-2 November 2011
Cyclone Warning Division
Meteorological DepartmentMausam Bhawan, Lodi Road
New Delhi-110003 (India)
Extended Abstracts
Ministry of Earth Sciences
Bay of Bengal Tropical Cyclone Experiments
Meteorological Department
New Delhi
1-2 November 2011
Cyclone Warning Division
India Meteorological Department
Mausam Bhawan, Lodi Road
New Delhi-110003
Extended Abstracts
National Conference
on
Bay of Bengal Tropical Cyclone Experiments
(BOBTEX-2011)
Forward
During the past few years, huge technological advancements have been achieved elsewhere in the world to observe the inner core of the cyclone through aircraft probing. Accordingly, Ministry of Earth Sciences conceived a programme in 2008 for aircraft probing of tropical cyclones over the Bay of Bengal which resulted in the commencement of Forecast Demonstration Project (FDP) in 2008 with Multi-institutional mechanism and IMD as nodal agency. FDP programme is aimed to demonstrate the ability of various NWP models to assess the genesis, intensification and movement of cyclones over the north Indian Ocean with enhanced observations over the data sparse region and to incorporate modifications into the models which could be specific to the Bay of Bengal. The lessons learnt during the pre-pilot and pilot phases of FDP campaign helped in improved monitoring and prediction of cyclonic disturbances during 2008-2010. During the Final Phase of the FDP programme (15 Oct.-30 Nov., 2012) India is planning to take up aircraft probing of cyclones over the Bay of Bengal with hiring of aircraft and drop sonde experiments. Considering all these, a two days National Conference on Bay of Bengal Tropical Cyclone Experiments (BOBTEX-2011) is organised in New Delhi during 01-02 November 2011. There are valuable research papers dealing with various aspects of cyclones over the Bay of Bengal, especially in relation to FDP from leading experts of both research and operational meteorological environments. It is intended that the conference will be a first step towards an ongoing focus on impact of surface-upper air and space based observations in operational cyclone forecasting and NWP modelling in the north Indian Ocean. I am glad to inform that a volume of Extended Abstract of the research papers of the national conference, BOBTEX-2011 is brought out which will be very helpful as guidance material for further research on cyclone and planning of future FDP campaigns. I thank Cyclone Warning Division of IMD, New Delhi, for organising BOBTEX-2011. My special thanks are due to Dr. M Mohapatra and Dr Naresh Kumar for bringing out this Extended Abstract of the Proceedings of the conference. I also thank Prof. T.N. Krishnamurti for agreeing to deliver the keynote address; Prof. S.K. Dube, Prof. U. C. Mohanty, Mr R.C. Bhatia and Mr. S. Raghavan for agreeing to deliver the lead talks; Prof. J. Shukla for agreeing to chair the Panel discussion and concluding session of the conference and Mr D.R. Sikka for reviewing the extended abstracts of the proceedings. IMD, New Delhi Ajit Tyagi 01. November 2011 Director General of Meteorology
Contents
Page Synoptic and Climatological Aspects
1. Outcome and challenges of the Forecast Demonstration Project 1
on Landfalling Cyclones over the Bay of Bengal.
Ajit Tyagi, M. Mohapatra, D.R. Sikka* and B. K. Bandyopadhyay
2. Utility of Tropical Cyclone Module for monitoring and 8
prediction of cyclonic disturbances over the North Indian Ocean.
M Mohapatra, Naresh Kumar and B. K. Bandyopadhyay
3. Climatology and intensification of Bay of Bengal Cyclonic storms 10
K.Seetharam
4. Study of wind shear, squall lines and cloud top temperatures in
association with Tropical cyclone 13
Charan Singh
5. Performance of modified CLIPER model for tropical cyclone track
prediction over the north Indian Ocean 16
R. P. Sharma, M. Mohapatra and B. K. Bandyopadhyay
6. Possible causes for absence of cyclogenesis over the Bay of
Bengal during October-November 2009 19
S. Adhikary and M. Mohapatra
7. WARD Cyclone – A Case Study 21
S.R. Ramanan, K.V. Balasubramanian and M.Veerakumar
8. Upper Ocean Observations during the passage of cyclone JAL-2010 26
Anitha Gera, M Ravichandran and A. K. Mitra
9. Salient features of JAL Cyclone of November 2010 – A case Study 27
D. C. Gupta
10. Characteristics of VLF atmospherics during tropical cyclone ‘AILA’
and several other thunderstorms over North-East India 30
Rakesh Roy, Abhijit Choudhury, Anirban Guha and Barin Kumar De
11. The Role of India Meteorological Department Telecommunication
Infrastructure on Forecast Demonstration Project (FDP) program of Tropical
Cyclones over Bay of Bengal 31
Sankar Nath
12. Evaluation of Cone of Uncertainty in Tropical Cyclone Track Forecast
over north Indian Ocean Issued by India Meteorological Department 32
D. P. Nayak and M. Mohapatra
Satellite and Radar Applications in Cyclone Monitoring
13. Observational aspects including DWR for cyclone monitoring 35
S. Raghavan
14. Observations of Cyclones from Space-Based Platforms: Current Status
and future Prospects 36
R.C. Bhatia
15. Early Detection of Global Tropical Cyclogenesis using OSCAT Data 37
C. M. Kishtawal and Neeru Jaiswal
16. Objective Detection of Center of Tropical Cyclone in Remotely Sensed 38
Infrared Images
Neeru Jaiswal, C. M. Kishtawal, P. K. Pal
17. Analysis of tropical cyclones by using microwave imageries of other 39
polar orbiting satellites over Indian region
Suman Goyal and A. K. Sharma
18. Estimation of intensity of tropical cyclone over Bay of Bengal 40
using Microwave imagery
T. N .Jha, M Mohapatra and B .K .Bandyopadhyay
19. Making a complete picture – radar composite 43
B. Arul Malar Kannan, Suresh Chand and S.K. Kundu
20. Study of Tropical Cyclone AILA using Doppler Weather Radar data 44
D. Pradhan
Heavy Rainfall, Gale Wind and Storm Surge
21. Storm surge and coastal inundation 46
S. K. Dube
22. Numerical modeling of Tide-Surge interaction in the Bay of Bengal 49
Jismy Poulose
23. Outlook of tide and storm induced current off Gopalpur coast 50
Susant Kumar Misra, P. Chandramohan, A. S. N Murty, J. K. Panigrahi,
R. Mahadevan, M. M. Mahanty and J. K. Sahu
24. Estimation of pressure drop within the tropical cyclone and height 51
of associated storm surge using Doppler velocity data
D.Pradhan, Anasuya Mitra
25. Tropical Cyclones Wind Radii prediction over North Indian ocean 53
M. Mohapatra and Monica Sharma
26. Drop size distribution Characteristics of cyclone and convective precipitation 57
observed over Semi-arid-zone in India
S.Balaji Kumar, S.B.Surendra Prasad, U.V. Murali Krishan
and K.Krishna Reddy
27. Changes in extreme daily rainfall associated with cyclonic disturbances 58
over Andaman & Nicobar Islands in a warming climate
Naresh Kumar, M. Mohapatra, A. K. Jaswal and B. P. Yadav
28. Monitoring Formation and Movement of the Depression of
16-23 June 2011 using DWR, Satellite Products and Synergy and Utility of
Implimenting a Real time Nowcasting in IMD for filling the forecasting Gap 60
Rajendra Kumar Jenamani
29. Forecasting of rainfall from landfalling cyclone using satellite derived
rain rate data: A case Study for cyclone ‘Aila’ 62
Habibur Rahaman Biswas and P.K.Kundu
30. Unprecedented flood in river Mahanadi in Orissa in September, 2008
and its impact on economic development 63
S.C.Sahu and S.K.Dastidar
31. Deep Depression without Heavy Rainfall 64
Bikram Singh, R.C. Vashisth, B.P. Yadav and Charan Singh
32. Lessons from IRENE 65
S. Raghavan
NWP Applications in Cyclone Prediction
33. NWP models applications in Tropical Cyclone Predictions
over the Bay of Bengal 66
U. C. Mohanty*, Krishna K Osuri and S. Pattanayak
34. IMD’s recent initiatives for improved Tropical Cyclone track and intensity
forecast over Indian region using Hurricane WRF Model 69
Y.V. Rama Rao, T.S.V. Vijay Kumar, Zhan Zhang, K. Naga Ratna,
A.K. Das, D.R. Pattanaik, S.K. Roy Bhowmik and Ajit Tyagi
35. Impact of cyclone bogusing and regional assimilation on tropical
cyclone track and intensity predictions 70
Manjusha Chourasia, R. G. Ashrit, John P George
36. Numerical Simulation of Tropical Cyclones in Bay of Bengal 71
R. D. Kanase and P. S. Salvekar
37. Tropical Cyclone Genesis Potential Parameter (GPP) and it’s application
over the North Indian Sea 74
S. D. Kotal and S. K. Bhattacharya
38. Track Prediction of North Indian Ocean Tropical Cyclones using ARW model 75
Krishna K. Osuri, U. C. Mohanty, A. Routray and M. Mohapatra
39. On the Implementation and the ability of the Ensemble Prediction System
for tropical cyclone track and strike probability for North Indian Ocean 76
K. Naga Ratna
40. Ocean atmospheric coupled model to estimate energy and path of
cyclone near the coast 77
Ramkrishna Datta
41. Track, intensity and few dynamical aspects of ‘AILA’ as simulated by
operational NWP model of the IAF 79
Wg Cdr TP Srivastava and Wg Cdr Anil Devrani
42. Analysis of Barotrophic Energetics of Tropical Cyclone Khai-Muk 85
S.Balachandran
43. Performance evaluation of spectrum of cyclones over North Indian Ocean
using RAMS model 86
Ancy Thomas, Basanta kumar Samala and Akshara Kaginalkar
44. An Observational and Modeling Study of the Tropical Cyclone PHET 87
Jagabandhu Panda, R. K. Giri and Harvir Singh
45. Large-Scale Characteristics of Rapidly Intensifying Tropical Cyclones
over the Bay of Bengal and a Rapid Intensification (RI) Index 89
S. D. Kotal and S. K. Roy Bhowmik
46. Development of the Lagrangian Advection model for prediction of tropical
cyclone track over the Indian Ocean 90
Sanjeev Kumar Singh, C. M. Kishtawal, Neeru Jaiswal, and P. K. Pal
47. Extended Range Forecast of Tropical Cyclone Genesis Based on Coupled
Model Outputs 92
D. R. Pattanaik, M. Mohapatra, Y. V. Rama Rao and Ajit Tyagi
48. Impact of Resolution and Data Assimilation on the prediction of the
cyclone “JAL” over Bay of Bengal using WRF (NMM) and grid
point statistical interpolation scheme 95
K. Naga Ratna
49. Study of JAL cyclone track using WRF cumulus parameter schemes 96
M. Venkatrami Reddy, S. Balaji Kumar, S. B. Surendra Prasad
and K. Krishna Reddy
50. Impact of data assimilation system for simulation of tropical cyclones
over Bay of Bengal with WRF-NMM modeling system 97
Sujata Pattanayak and U C Mohanty
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 1
Outcomes and Challenges of Forecast Demonstration Project (FDP) on Landfalling Cyclones
over the Bay of Bengal
Ajit Tyagi, M. Mohapatra, D.R. Sikka* and B. K. Bandyopadhyay
India Meteorological Department
Mausam Bhavan, Lodi Road, New Delhi-110003
*40, Mausam Vihar, New Delhi-110051
1. Introduction During the past few years huge technological advancements have been achieved elsewhere
in the world to observe the inner core of the cyclone. Accordingly a programme has been evolved
for improvement in prediction of track and intensity of tropical cyclones over the Bay of Bengal
resulting in planning of the Forecast Demonstration Project (FDP). FDP programme is aimed to
demonstrate the ability of various NWP models to assess the genesis, intensification and movement
of cyclones over the north Indian ocean with enhanced observations over the data sparse region and
to incorporate modification into the models which could be specific to the Bay of Bengal based on
the in-situ measurements and following the actual track through Satellite and Radar observations.
FDP Programme is scheduled to be implemented in three phases, viz., (i) Pre- pilot phase (15 Oct-
30 Nov. 2008, 2009, (ii) Pilot phase (15 Oct-30 Nov. 2010 and 2011) and (iii) Final phase (15 Oct-
30 Nov. 2012). India is planning to take up aircraft probing of cyclones over the Bay of Bengal
during 15 Oct.-30 Nov., 2012 with hiring of aircraft and dropsonde experiments.
To accomplish the above objective, the initiative was carried out with following priorities.
(i) Observational upgradation
(ii) Modernisation of cyclone analysis and prediction system
(iii) Cyclone analysis and forecasting procedure.
(iv) Warning products generation, presentation & dissemination,
(v) Confidence building measures and capacity building
2. Implementation of FDP during 2008-2010 Various strategies were adopted for improvement of observation, analysis and prediction of
cyclone. Several national institutions participated for joint observational, communicational & NWP
activities during the pre-pilot and pilot phases of FDP campaign during 2008-10. There were 23
days of intense observation period (IOP) in association with cyclonic disturbances (CDs) during
2008 and 2010. and no IOP during 2009, as there was no CD during FDP period over the Bay of
Bengal.
Enhanced observations during Intense IOP helped in improved monitoring and prediction of
CDs. The additional data was collected from Sagar Kanya cruise, enhanced AWS network of the
coast, high wind speed recorders (HWSRs), Doppler Weather Radars (DWRs), five activated buoy
observations from the Bay of Bengal, Oceansat-II observations and microwave imagery products.
The comparison of observational systems before and after FDP indicates a significant improvement
in terms of Radar, AWS, high wind speed recorders over the region (Table 1). It has resulted in
reduction in landfall point location error from 55 km to 25 km (Mohapatra et al, 2011)
Table 1. Observatory network by end of 2007 and 2010
Observational system Network by end of 2007 Network by end of 2010
Surface synoptic observatory network 559 559
Pilot balloon observatory network 62 62
Radiosonde/Radiowind network 35 39
Buoy network 6 12
AWS network 125 524
HWSR - 12
DWR 5 12
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 2
To ensure the availability of the data and forecast products from various national and
international sources at Cyclone Warning Division, IMD, New Delhi, an institutional mechanism
was developed in consultation with all the stake holders. A standard operation procedure (SOP) has
been prepared for monitoring and prediction of cyclonic disturbances and issue of warning. It
includes the road map and check lists for this purpose.
The tropical cyclone analysis, prediction and decision-making process was made by blending
scientifically based conceptual models, dynamical & statistical models, meteorological datasets,
technology and expertise. Conventional observational network, automatic weather stations (AWS),
buoy & ship observations, cyclone detection radars and satellites were used for this purpose. A new
weather analysis and forecasting system in a digital environment was used to plot and analyse
different weather parameters, satellite, Radar and numerical weather prediction (NWP) model
products. An integrated fully automated forecasting environment facility was thus set up for this
purpose. The manual synoptic weather forecasting was replaced by hybrid systems in which
synoptic method could be overlaid on NWP models supported by modern graphical and GIS
applications to produce
• high quality analyses
• Ensemble of forecasts from NWP models at different scales - global, regional and
mesoscale
• Prediction of intensity and track of tropical cyclone and storm surge
• Specialized warning information to various sectors including Govt. and non-Govt. agencies,
The Tropical Cyclone Module installed in this forecasting system has the facilities to serve
the above purpose. The automation of the process has increased the efficiency of system, visibility
of IMD and utility of warning products. The products before and after initiative are shown in Fig.1.
The improvement in monitoring and forecasting tools and techniques are shown in Table 2.
Fig.1. Comparison of weather analysis products before and after the initiative
3. Outcome of FDP-2008-2010 Salient features of achievements are described below.
(a). Cyclone track and intensity forecast : For comparison, the 24 hr track forecast errors and the skill scores during 2003 and 2010 are
shown in Fig.2 (RSMC, New Delhi, 2009, 2010, 2011. The figures clearly indicate the gradual
improvement in the cyclone forecast by IMD, as the error has decreased and the skill has increased.
The average landfall error was less than the long period average error for the landfalling cyclones
over the north Indian Ocean. It is also very much comparable to the forecast errors over other
Ocean basins including north Atlantic and Pacific Ocean basins. Considering, the intensity forecast,
the average 24 hrs wind forecast error has been about 10 knots (Table 3) for these cyclones.
(ii) After initiative(Isobaric analysis at mean sea
level during cyclone, Phet at 00 UTC of 03 June
2010)
(i) Before initiative
(Isobaric analysis at mean sea level)
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 3
Table 2. Comparative analysis of tools and techniques by the end of 2007 and 2010.
Parameters Tools/technique by end of 2007 Additional tools/technique by
end of 2010
Genesis Synoptic, satellite (visible & IR imagery),
NWP analysis (T254), coarser resolution
ECMWF, UKMO, NCEP, Quikscat, Ascat,
AMV
Microwave imagery,
Oceansat-II
Location
monitoring
Ship, Buoy, limited AWS, Quikscat, Ascat,
AMV
Enhanced AWS network,
GPSsonde, buoy, Oceansat-II
Intensity
monitoring
Satellite (Visible and infrared imagery), Radar,
Quikscat, Ascat, AMV
Microwave imagery,
enhanced DWR network,
buoy network, Oceansat-II
Genesis forecast Synoptic, satellite, radar Microwave imagery,
Dynamical statistical model
Track forecast Synoptic, satellite, radar, CLIPER, Limited
NWP guidance (Coarser ECMWF, UKMET,
NCMRWF (T80), LAM, MM5, QLM),
High resolution ECMWF,
IMD GFS(382),Experimental
(T574), NCEP GFS, ARPS
(Meteo-France), NCMRWF,
MME, Experimental HWRF,
WRF (ARW), WRF (NMM),
modified CLIPER, ISRO GA
technique
Strike
probability
- Strike probability based on
EPS and super EPS
Intensity
forecast
- Dynamical statistical model
Rapid
intensification
- Dynamical statistical model
Comparing the landfall forecast errors, the 24 hour mean error has been significantly less
during last three years (2008-2010). It is about 100 km against the long period average error of
about 150 Km(Fig. 3).
24 hr Track Forecast Error (km)
203
165
142
181
131110
136127
0
50
100
150
200
250
2003 2004 2005 2006 2007 2008 2009 2010
Year
Err
or
(km
)
24 hr track Forecast Error (km)
Linear (24 hr track Forecast Error (km))
3 per. Mov. Avg. (24 hr track Forecast Error (km))
Fig.2 (a). 24 hr cyclone track forecast errors of IMD during 2003-2010.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 4
24 hr Forecast Track Skill Score (%)
6.3
29.5
18.621.9
13
24.1
53
13
y = 1.8357x + 14.164
R2 = 0.0983
0
10
20
30
40
50
60
2003 2004 2005 2006 2007 2008 2009 2010
Year
Skill S
co
re (%
)
24 hr Track Forecast Skill Score
Linear (24 hr Track Forecast Skill Score)
3 per. Mov. Avg. (24 hr Track Forecast Skill Score)
Fig.2.(b): 24 hr cyclone track forecast skill scores of IMD during 2003-2010.
Fig. 3. Landfall forecasterrors of IMD during 2003-2010
Table 3. Official average intensity forecast error of 2010
Lead Period
(hrs)
Intensity Error (knots) No. of
Observation verified Average Absolute Average RMS
12 1.0 8.1 11.3 55
24 4.5 12.2 16.4 49
36 8.7 15.3 20.4 37
48 13.4 16.5 21.9 29
60 19.6 20.9 26.8 23
72 21.0 21.0 28.3 19
The performance of NWP models have increased along with the introduction of NWP
platforms like IMD GFS, WRF, HWRF and ensemble prediction system (EPS) The mean track
forecast errors of NWP models during 2010 are given in Table 4. The performance of multi-model
ensemble (MME) prediction is reasonably good. The 48 hours track forecast errors by MME
technique of IMD is about 200 km.
24 hr Landfall forecast errors (km) during 2003-2010
0
100
200
300
400
500
600
2003 2004 2005 2006 2007 2008 2009 2010
Year
Err
or
(km
)
24 hr forecast error Linear (24 hr forecast error)
12 hr Landfall forecast errors (km) during 2003-2010
0
50
100
150
200
250
300
350
2003 2004 2005 2006 2007 2008 2009 2010
Year
Err
or
(km
)
12 hr forecast error Linear (12 hr forecast error)
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 5
Table 4. Mean Track forecast errors of NWP models for cyclones during 2010
(b) Cyclone Warning Services The format and content of bulletins have been changed significantly as shown in Table 5.
These improvements have contributed to effective management of cyclone by disaster managers.
The time of issue and frequency of bulletins have been standardized. The frequency of
bulletin has also been increased along with the increase in number of users. The lead time of the
forecast has been increased upto 72 hrs. The design of the bulletin has been revised with inclusion
of prognostic and diagnostic features, observed and forecast track and intensity in Tabular form and
storm surge guidance for all member countries of WMO/ESCAP Panel. The observed and
forecast track and intensity of the cyclone were updated in cyclone page of IMD website time to
time, based on the tropical cyclone advisory bulletin issued by Cyclone Warning Division of IMD,
New Delhi. The cone of uncertainty in the forecast has been introduced with effect from the
cyclone, ‘WARD’ during December, 2009. It is helpful to the decision makers as it indicates the
standard forecast errors in the forecast for different periods like 12, 24, 36, 48, 60 and 72 hrs. The
improvement in delivery services of cyclone warning after the intiative as compared to prior to
initiative are shown in the Table 6.
Table 5. Comparison of cyclone warning products and bulletins before and after the initiative
SN Parameters Bulletin issued before
initiative
Bulletin issued after
initiative(2010)
1 Date and time of issue of bulletin Date only Both date and time
2 Current location, intensity Yes Yes
3 Past movement Yes Yes
4 Forecast validity period Upto 24 hrs Upto 72 hrs(+6, +12, +18, +124,
+36, +48, +60 and +72 hrs)
5 Quality of forecast track and
intensity
(Qualitative) Quantitative.
6 Landfall point and time Qualitative Quantitative with lati/long of
landfall and time
7 Prognostic and diagnostic features Nil Detailed features are explained
in the Technical bulletin.
08 Graphical presentation of
observed and forecast track
No Yes
9 Adverse weather (Heavy rain,
Gale wind and storm surge)
Storm surge for Indian
coast only
For coasts of all member
countries of WMO/ESCAP
Panel
10 Advice and action suggested Yes Yes, but more specific
AVERAGE 12 hours 24 hours 36 hours 48 hours 60 hours 72 hours
ECMWF 54 71 102 170 202 246
NCEP-GFS 158 178 177 236 253 334
JMA 195 96 176 203 232 268
IMD-MM5 118 141 241 350 363 356
IMD-QLM 103 144 167 181 256 311
IMD-MME 72 104 140 205 190 244
IMD-T382 94 124 164 212 246 290
IMD-WRF-VAR 155 137 236 253 234 265
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 6
Table 6. Beneficiaries feedback of cyclone warning services before and after initiative
SN Parameters Beneficiaries feedback
before initiative
Beneficiaries feedback after initiative
(e.g. 2010)
1 Number of deaths Higher Less
2 Loss due to evacuation of
people due to uncertainty in
forecast
Higher Less
3 Quality of warning
presentation
Poor Good
3 Appreciation by disaster
management agencies
Limited Appreciation by central & state Govt
agencies, and neighbouring countries
4 Number of warnees Less, e.g. six in 2003-
04 at national level
More, e.g. Fifteen in 2009-10 at
national level
5 Number of visitors to
cyclone page of IMD’s
website
Less (No counter) Significantly higher. Number of
visitor during last cyclone, PHET
(June 2010) : 40, 000 (Approx)
(c). Loss of lives due to cyclones The loss lives due to cyclone has reduced significantly due to many factors including
improvement in early warning system of cyclone. Characteristics of two similar severe cyclones
crossing Andhra Pradesh coast near Machhilipatnam in 2003 and 2010 are shown hear as example
to compare the loss of human lives.
Cyclone period : 17-21 May 2010 11-16 December 2003
Cyclone category Severe cyclone Severe cyclone
Point of landfall South of Machhilipatnam South of Machhilipatnam
Maximum wind at landfall 100 kmph 100 kmph
Landfall forecast error 24 hr lead time 55 km 257 km
48 hr lead time 115 km No forecast issued
72 hr lead time 207 No forecast issued
Loss of human lives 06 81
4. Challenges of FDP
With repeated attempt, the aircraft probing of TCs could not be possible till now. It is major
challenge for FDP-2012. The FDP on landfalling TCs over the Bay of Bengal with aircraft
probiong facility will help us in minimising the error in monitoring and hence prediction of tropical
cyclone track and intensity forecasts (Martin and Gray 1993). In addition, this project will help in
the following.
(a) Validation of Dvorak technique over the NIO
(b) Validation of pressure–wind relationship in TCs over the NIO
(c) Understanding and prediction of structure of TCs over the NIO.
(d) Development/validation of wind conversion factor for converting 3-minute average wind to 1-
minute average wind (used in Dvorak’s technique) and 10-min average wind (as required for
preparation of standardised international best tracks archives)
(e) Reanalysis of best tracks with modified pressure–wind relationship, wind adjustment and
modified Dvorak classification of intensity
(f) Improvement/validation of performance of numerical weather prediction models
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 7
The other major challenges include (i) assimilation of regional data and development of
suitable global and regional models for cyclone prediction with suitable modification of model
physics, resolution and initial and boundary conditions (ii) development of ensemble prediction
system based on IMD GFS and WRF models.
5. Conclusions The FDP on landfalling cyclones over the Bay of Bengal has helped in improvement of
monitoring, forecasting and warning of cyclones over the north Indian Ocean. The observational
network, tools and technologies, especially the NWP models have improved significantly during
2008-2010. As a result, the 24 hr forecast track error has reduced from 163km during 2003-2007 to
141 km during 2008-2010. However, the main challenge of the FDP is still to be realised with the
introduction of aircraft probing of cyclones and dropsonde experiments.
References
Martin JD, Gray WM (1993) Tropical cyclone observation and forecasting with and without
aircraft
reconnaissance. Weather Forecast 8:519–532
Mohapatra, M., B. K. Bandyopadhyay, Ajit Tyagi, 2011, Best track parameters of tropical cyclones
over the North Indian Ocean: a review, Natural Hazards, DOI 10.1007/s11069-011-9935-0.
RSMC, New Delhi (2009) Report on cyclonic disturbances over the North Indian Ocean during
2008. IMD, New Delhi
RSMC, New Delhi (2010) Report on cyclonic disturbances over the North Indian Ocean during
2009. IMD, New Delhi
RSMC, New Delhi (2011) Report on cyclonic disturbances over the North Indian Ocean during
2010. IMD, New Delhi
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 8
Utility of Tropical Cyclone Module for monitoring and prediction of cyclonic disturbances
over the North Indian Ocean
M Mohapatra, Naresh Kumar and B. K. Bandyopadhyay
India Meteorological Department
Mausam Bhavan, Lodi Road, New Delhi-110003
1. Introduction The tropical cyclone (TC) forecast & warning operations and decision-making process
should be made by blending scientifically based conceptual models, meteorological datasets,
technology and expertise (IMD, 2003). The tropical cyclone module (TCM) available in synergie
system since the end of 2009 provides a digitized platform for the above purpose As adverse
weather warning depends on the track forecast, this TCM helps in accurate prediction of adverse
weather and hence effective management of TC. This TCM is helpful in improving (i) cyclone
analysis and forecasting procedure and (ii) warning products generation, presentation &
dissemination. All these aspects are presented and analysed herewith.
2. cyclone analysis and forecasting procedure A new weather analysis and forecasting system in a digital environment has been established
at National Weather Forecasting Centre, New Delhi to plot and analyse different weather
parameters, satellite, Radar and numerical weather prediction (NWP) model products. An
integrated fully automated forecasting environment facility is thus available for this purpose. The
manual synoptic weather forecasting has been replaced by hybrid systems in which synoptic
method could be overlaid on NWP models supported by modern graphical and GIS applications to
produce
• high quality analyses
• Ensemble of forecasts from NWP models at different scales - global, regional and
mesoscale
• Prediction of intensity and track of tropical cyclone
• Specialized warning information to various sectors
Fig.1. Strategy adopted for cyclone analysis and forecasting The major highlights of the strategies followed for monitoring and prediction of cyclone are
shown in the Figure 1. The TCM installed in this forecasting system has the following facilities.
• Analysis of all synoptic, satellite and NWP model products for genesis, intensity and track
monitoring and prediction
• Preparation of past and forecast tracks upto 120 hrs
• Depiction of uncertainty in track forecast
Action Synopic
Users*
End
forecast
Initial conditions
(Observations)
Synoptic
Satellite
Forecaster
NWP
Model
Numerical
forecasts
Runs of different
Models,
Consecutive runs
from the same
model,
Ensemble runs
("choosing the
best member") *Central / State Govt/ Media/ Public
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 9
• Structure forecasting (Forecast of wind in different sectors of cyclone)
However all the data are not still available in TCM through synergie system. For better
monitoring and prediction, addition help is taken of ftp and websites to collect and analyse:
• Radar data and products from IMD’s radar network and neighbouring countries
• Satellite imageries and products from IMD and international centres
• Data, analysis and forecast products from various national and international centres
Fig.2. Utility of Modernised cyclone analysis and forecasting system using TCM Averag e time (minutes ) c ons umed by R S MC , New D elhi
to is s ue the warning bulletin
189180
152 155
y = -13x + 201.5
R2 = 0.84
0
20
40
60
80
100
120
140
160
180
200
2007 2008 2009 2010
Y ear
Tim
e (
Min
ute
s)
Fig.3. Average time consumed by RSMC, New Delhi to issue cyclone warning bulletin since
last three hourly synoptic observations To ensure the availability of the data and forecast products from various national and
international sources at Cyclone Warning Division, IMD, New Delhi, an institutional mechanism
was developed in consultation with all the stake holders. A standard operation procedure (SOP) has
been prepared for monitoring and prediction of cyclonic disturbances and issue of warning. It
includes the road map and check lists for this purpose.
3. Warning products generation, presentation & dissemination Various steps were taken by the nominee to improve product generation, presentation and
dissemination, which could enhance the users’ response for effective cyclone disaster management.
A few highlights of the initiative are discussed herewith. A few examples of products generated
using TCM are shown in Fig. 2.The time of issue and frequency of bulletins have been
standardized. The frequency of bulletin has also been increased with reduction in time required for
issue of bulletin as shown in Fig.3. The design of the bulletin has been revised with inclusion of
prognostic and diagnostic features, observed and forecast track and intensity and adverse weather in
graphical form.
4. Conclusions The TCM is a very good tool for monitoring and prediction of cyclonic disturbances and
associated adverse weather. However, it needs to be used in conjunction with other data,
information and products available from national and international centres.
References : IMD, 2003, Cyclone Manual, India Meteorological Department, Mausam Bhavan, Lodi Road, New
Delhi
Display of wind radii envelop
Display of wind radii envelop
Comparison of various
model predictions
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 10
Climatology and intensification of Bay of Bengal Cyclonic storms
K.Seetharam
Meteorological Centre, Hyderabad
It has long been hypothesized the large scale atmospheric forcings for enhanced convection,
warmer sea surface temperatures, low level voriticity and windshear and availability of mid-
tropospheric humidity are favourable for cyclogenesis and intensification over Bay of Bengal
tropical storms. The cyclonic storms and severe cyclonic storms data sets for a period of 120 years
1891-2010 formed over Bay of Bengal were collected from the e-atlas of India Meteorological
Department. It is seen from the data sets that during the period 1891-2010 506 cyclonic storms
were formed over Bay of Bengal (on an average 8) and 221 of them intensified in to the Severe
Cyclonic storm stage (44%) with an average of 4. Overall, on an average 50% of the cyclonic
storms formed over Bay of Bengal intensified in to Severe Cyclonic Storms. The trend was little
erratic prior to 60s but there is a continuous and systematic decrease in the cyclonic storms over
Bay of Bengal from 60s onward up to 2000 and started again rising in 2001-2010. Examination of
the plots of data sets showed overall decreasing trend in the total number of cyclonic storms and
overall increasing trend in the total number of severe cyclonic storms when fitted with linear trend.
When a 6th
degree polynomial trend was fitted to the data sets with forward forecast for next 5
years, the trends were alternating with different periods in case of both cyclonic storms and severe
cyclonic storms but the forecast showed increasing trend in both cyclonic storms and severe
cyclonic storms beyond 2010. Further the data sets have been divided in to two epochs. The epoch I
is the period 1891-1950 and epoch II is the period 1951-2010. Comparison of epoch I and epoch II
showed that the total number of cyclonic storms formed in the Bay of Bengal was 284 out of which
94 intensified up to the stage of Severe Cyclonic storms (33%) in the epoch I and total number of
cyclonic storms formed in the Bay of Bengal was 221 out which 127 intensified up to the stage of
Severe Cyclonic storms (57%) in epoch II. The intensification of the systems is stronger during the
epoch II than the epoch I even though there is a decrease in the total number of cyclonic storms
over Bay of Bengal from epoch I to epoch II. Moreover, further examination of the data sets on the
decadal scale showed that 35 cyclonic storms formed in the Bay of Bengal during the decade 1981-
1990 out of which 22 (63%) intensified in to Severe Cyclonic Storms and during the decade 2001-
2010 32 cyclonic storms formed in the Bay of Bengal out of which only 11 (34%) intensified in to
Severe Cyclonic Storms. In this paper the environmental conditions like SSTs and Relative
Humidity during the two contrasting decades 1981-1990 & 2001-2010 were compared. The sea
surface temperatures (SSTs) taken are the extended Kaplan SSTs taken for decades 1981-1990 &
2001-2010 from the NCEP/NCAR reanalysis data sets. The study indicated the unusual warming in
the West Central Bay and East Central Bay is leading to the intensification of cyclonic storms over
Bay of Bengal. The Relative Humidity is also taken from NCEP/NCAR reanalysis. The study of
the humidity pattern between 1000 hPa and 500 hPa levels indicated low humidity in lower levels
and higher humidity during the period 1981-1990 in comparison with the period 2001-2010 with
negative N-S gradient in both levels.
Fig. 1 Decade wise Cyclonic Storms/Severe Cyclonic Storms over Bay of Bengal (1891-2010)
with linear trends fitted
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 11
Fig.2. Year-wise Cyclonic Storms and Severe Cyclonic Storms over Bay of Bengal (1891-
2010) fitted with 6th
degree polynomial trend
Fig. 3. Year-wise Cyclonic Storms and Severe Cyclonic Storms over Bay of Bengal (1891-
2010) fitted with linear trend
References Briegel, Lisa M., William M. Frank, 1997: Large-Scale Influences on Tropical Cyclogenesis in the
Western North Pacific”, Monthly Weather Review, 125, pp 1397–1413.
Joseph P.V. and Prince K. Xavier., (1999), “Monsoon Rainfall and Frequencies of Monsoon
Depressions and Tropical Cyclones of recent 100 years and an outlook for the first decades
of the 21st century., Meteorology beyond-2000, Proceedings of National Symposium
Tropmet-99., 16-19 Feb 1999, Editors A.K. Bhatnagar et al., Indian Meteorological Society,
Chennai Chapter., 364-371.
Kalnay, E. and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project Bulletin of
American Meteorological Society, Vol. 77, No. 3, pp 437-471.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 12
Ramesh Kumar M.R & Sankar S, 2010, “Impact of global warming on cyclonic storms over north
Indian Ocean”, Indian Journal of Marine Sciences Vol. 39(4), pp. 516-520
Seetharam,K, 2004,”Statistics of cyclonic disturbances in the North Indian Ocean”, Mausam, 55
No.4, pp 698-704.
Sikka, D.R, 1977,”Some aspects of life history structure and movement of monsoon depressions”,
Pure and Applied Geophysics, 115, pp1501-1529.
Sujata Mandke K and Usha Bhide V, 2003,”A study of storm frequency over Bay of
Bengal”,Journal of Indian Geophysical Union, Vol.7, No.2, pp 53-58
Fig.. 4 Year wise Cyclonic storms and Severe Cyclonic Storms (2001-2010)
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 13
Study of wind shear, squall lines and cloud top temperatures in association with
Tropical cyclone
Charan Singh
India Meteorological Department
Mausam Bhavan Lodi Road, New Delhi-110003
1. Introduction In Tropical cyclone (TC) is a most disastrous weather phenomena, it causes huge damage to
the life and property all around the world. To understand its dynamics mainly inner core and effects
of surrounding environment mechanism is very essential. In north Indian Ocean (NIO), this
becomes very important as some TCs cause huge amount of rainfall, but in some cases very less
rainfall is observed. These are described as moist and dry air environment tropical cyclones. Orissa
super TC & Nargis are the examples of moist air environment and Ogni & Sidar are the dry air
environment TCs. Vertical Wind Shear (VWS) play an
important role in genesis, intensification and weakening of
TCs as at the time of genesis and strengthening phase, it
enhances the mixing of moist air in vertical column and on
weakening phase it enhances dry air mixing which causes
rapid weakening of the system. Squall lines form ahead of the
TCs due to increase in instability and transport of energy
from the TCs. Squall lines along with rainband clouds
associated with TCs are the main causes of rain and
thunderstorms. TCs, which made landfall over Indian coasts,
the associated rainfall is mainly confined to the right
forward sector followed by left forward sector, it also
depends upon the season and depth convection. TCs
formed just before or after of the southwest monsoon,
generally cause more rainfall than other TCs. Also, TCs
associated with Cloud Top Temperature (CCT) ≤ -600C
cause very heavy rainfall (15-25 cm in 24 hours) over the
respective area. It is observed that the rainfall also
depends on the speed of TCs as slow moving TCs cause
much more rainfall rather than fast moving TCs.
2. Role of Vertical Wind Shear, Vertical Wind Shear (VWS) of horizontal winds is
generally considered as a resultant vector wind between 200
and 850 hPa level. The main function of the wind shear is to
sustain the cloud clusters in a vertical form. When a cloud
clusters develops over the sea surface due to higher Sea
Surface Temperature (SST) or Ocean Heat Contents (OHC),
which depends on the vertical profile of the sea, the sea
surface transmits the energy to the air parcels, where
atmosphere is already unstable. Initially air parcel lifts to the
free level convection then moist air mixing starts and system
starts grow. In the presence of the sufficient OHC, it
continues to strengthen [Fig. 1] (RSMC, New Delhi report, 2002) and the height of the cloud
increases with decrease in the VWS and start mixing of moist air. As VWS starts increase (≤ 12
kts), the dry air mixing increases and upper parts of TCs also start drifting along strong upper air
winds [Fig.2] (RSMC, New Delhi report, 2007). As a result TCs start losing its intensity.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 14
3. Role of squall lines and rainbands: Squall lines are a kind of linear organization
meso-scale convective systems, which cause torrential
rainfall and thunderstorms. They may appear ahead of
landfalling TCs in general. The criteria to define squall
line is similar to Parker and Johns’s (2000). The region
exceeding 40 dBz reflectivity must extend longer than
100 km for at least 2-3 hours and convection of this
region is organized in linear or quasi linear shape with
an apparent common leading edge. The squall lines
sometimes are seen separated from the rainbands of
TCs [Fig. 3] (RSMC, New Delhi report, 2001). The
analysis of radar images show an apparent moisture
increase towards the formation position of the squall
line obviously due to the transportation of moisture
through the outer flow of the approaching TCs. In
general landfalling TCs cause squall lines in its front
quadrants.
4. Cloud top temperature: In TCs, maximum rainfall occur in the area of
maximum convection zone (Corbosiero, K. L.,
and J. Molinari, 2002). According to
(Raghavan, 1991). The maximum low level
convergence appears to occur in the right sector,
which contributes to this maximum in the right
rear sector and the formation of convective
‘streamer’ bands in the rear. To assess the
strength of convection, CTT from infra-red
satellite imagery is used as the proxy. Colder
CCT of convective clouds suggests that the
vertical extent of the cloud is more. Therefore, CTT is used as a measure of convection strength.
Isotherm analysis of CTT reveals that for most of the cases, convection generally tends to be
enhanced over the region to the right of the track. In general, about 70% convection is to the right
of track [Fig. 4 &5], which is well in agreement with the distribution of rainfall. In most of the
cases, intense to very intense convection is observed in the inner core of the TCs during the landfall
processes. The asymmetries in convection are also observed. It is due to unequal VWS in vertical
levels (Corbosiero, K. L., and J. Molinari, 2002). Very severe cyclonic storm over Bay of Bengal
during October 15-19, 1999 moved in a north-northeasterly direction before landfall near Gopalpur
(Orissa) in the early morning of 18 October, 1999. Satellite imagery received at 0230 hours IST of
18 October [Fig. 5] shows dense cloud mass with CTT -80- to -400C, spread over about 350 kms
diameter elongated along the track [Fig.4] of the TC. Study reveals that the maximum rainfall has
occurred within 150 kms radius from the landfall point and it was located on both sides along the
track. It is clear in this case that the maximum rainfall occurred over the area which lay under the
most convective cloud cover. The right first quadrant in this case also got good amount of rainfall
that ranged between 25-30 cm and it further extends beyond 300 kms of diameter. Fig. 6 shows that
the 24 hours accumulated rainfall is, to a certain extent inversely proportional to the speed of
movement of TCs with the best fit for speed range 4-10 knots. The maximum rainfall from a
landfalling TCs moving with speed in the range 4-10 knots can be estimated
by 905.64)(3429.5)( +−= speedXextremeR , with a standard deviation of 8.9 mm which is quite
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 15
large. Large standard deviation arises due to smaller data set used in the study. Thus assessing
extreme rainfall amount in ranges will give a better result rather than quantifying it with a single
number. Studies by (Singh and Bandyopadhyay, 2007) suggest that most of the TCs over NIO
basins move with a transitional speed between 4-12 knots. Therefore, the above emperical relation
could be used as a first guess by the operational forecasters to assess the extreme rainfall that could
occur in association with landfalling TCs.
5. Conclusion: Understanding of TCs inner core structure and effects of surrounding environment
mechanism is very essential to forecast the intensity and movement of TCs. In general, moist air
environment TCs cause more rainfall and bigger in size in comparison to dry air environment TCs.
Wind shear play an important role in genesis, intensification and weakening of TCs. At the time of
strengthening, the wind shear is generally ≤ 12 kts and at weakening stage it is generally ≥ 13 kts.
However, it is very difficult to fix the criteria of threshold value of VWS. Squall lines along with
rainbands clouds associated with TCs are the main causes of rain and thunderstorms. For TCs,
which made landfall over Indian coasts, the associated rainfall is mainly confined to the right
forward sector followed by left forward sector, and also depends upon the season and CCTs.
During landfall process, the CCT ≤ -600C cause very heavy rainfall (15-25 cm in 24 hours) over the
respective area. It is observed that the rainfall also depends on the speed of TCs as slow moving
TCs cause much more rainfall than fast moving TCs. Understanding the dynamics of the TCs and
the surrounding environment is still needed for prediction of location specific extremely heavy
rainfall.
References: Corbosiero, K.L., and J. Molinari, 2002: The effect of vertical wind shear on the distribution of
convection in tropical cyclones. Mon.Wea.Rev. 130, 2110-2123.
India Meteorological Department: Forecasting manual IV-23 weather radar as an aid to forecasting
1991 by S Raghavan.
Parker, M.D.and R.H.Johnson, 2000: Organizational models of mid-latitude meso-scale convective
systems. Mon.Wea.Rev.128, 3413-3436.
RSMC, New Delhi report, 2007, 2008 and 2001: A report on cyclonic disturbances over north
Indian ocean during 2007. Published by IMD, New Delhi.
Singh, C., and Bandyopadhyay, B.K. 2004: Behaviour of tropical cyclones along the east coast of
India prior to landfall. Mausam, 58, 2, pp 273-279.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 16
Performance of modified CLIPER model for tropical cyclone track prediction
over the north Indian Ocean
R. P. Sharma, M. Mohapatra and B. K. Bandyopadhyay
India Meteorological Department
Mausam Bhavan, Lodi Road, New Delhi-110003
E-Mail : [email protected]
1. Introduction India Meteorological Department (IMD) is the nodal agency for prediction of cyclonic
disturbances over the north Indian Ocean. The cyclone forecasts have improved steadily in the
recent decade (RSMC, New Delhi, 2011) due to improvement in monitoring and forecasting
technique, analysis tools and knowledge and human expertise. However, the very basic climatology
and persistence (CLIPER) models for the prediction of TC motion still have a number of
applications in a forecast office and continue to be developed. Some of these applications, not all of
which refer directly to the forecast process, are to 1) provide a convenient frame of reference upon
which the performance of more sophisticated models can be assessed, 2) enable the assessment of
‘‘forecast difficulty,’’ 3) provide a convenient way to generate bogus TC tracks, 4) provide a ‘‘first
guess’’ forecast, and 5) provide a reasonable forecast in portions of basins where deviations from
climatology and persistence are small (Pike and Neumann, 1987, Bessafi et al, 2002). Over the
north Indian Ocean, the CLIPER model was first developed by Sikka and Suryanarayana (1968) for
forecasting the movement of tropical storm for 24 hr period. Neumann and Mandal (1978)
developed a modified CLIPER model based on data of 1282 storms during 1877-1974 including
depressions to forecast the track upto 72 hrs in the interval of 12 hrs. However, there has been
significant improvement in monitoring technique leading to error in estimation of location and
intensity errors in recent decades (Mohapatra et al, 2011). Hence, it is felt that the existing CLIPER
model should be modified with inclusion of cyclone data set upto recent years for better reference
model. In this study, we present the characteristics of modified CLIPER model and its performance
with respect to cyclones during forecast demonstration project (FDP) period (2008-10).
2. Characteristics of modified CLIPER The modified CLIPER model is based on the data set of all the cyclones and depressions during
1891-2009 over the north Indian Ocean based on cyclone e-Atlas published by IMD (2008). It
includes the same parameters as predictors which was used by Neumann and Mandal (1987). It
uses the regression equation based on primary predictors to forecast the track upto 72 hrs in the
interval of 12 hrs. The predictors include the current and previous 12-hr position, the day of the
year, and the intensity of the system (depression/cyclone). The initial motion of the storm
(persistence) is the most important predictor for this model.
3. Performance of modified CLIPER model over the north Indian Ocean (NIO) The performance of the modified CLIPER has been evaluated by calculating the track
forecast errors of old and modified CLIPER models with respect to six hourly (00, 06, 12 and 18
UTC) best track data of tropical cyclones (TC) over the north Indian Ocean in post monsoon season
(October to December) during recent three years (2008-2010). This period is considered as
maximum data were collected during this period under forecast demonstration project (FDP) on
landfalling cyclones over the Bay of Bengal. For this purpose, the best track data have been
collected from the reports on cyclonic disturbances over the north Indian Ocean published by
RSMC, New Delhi (2009, 2010, 2011). There were six cyclones during this period as mentioned in
Table 1. The results are presented and discussed in the following sections.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 17
Table 1. Details of TCs considered under the study
S.N. TC Period
1 Cyclonic storm, RASHMI 26-28 October 2008
2 Cyclonic storm, KHAI MUK 11-13 November 2008
3 Cyclonic storm, NISHA 26-28 November 2008
4 Cyclonic storm, WARD 10-15 December 2009
5 Very severe cyclonic storm, GIRI 20-23 October 2010
6 Severe cyclonic storm, JAL 4-8 November 2010
3.1. Track error of modified CLIPER model The mean track errors of the modified CLIPER model based on the data of six TCs under
consideration are presented in Table 2. It is observed that the track error increases with increase in
forecast time period. The mean 12, 24, 36, 48, 60 and 72 hr track forecast errors are 97, 180, 256,
363, 461 and 540 respectively. Compared to 68, 150, 216, 292, 361, 392 km found in earlier study
of Mandal and Neumann (1978) based on data of 1877-1974and 62, 147, 240, 338, 431, 517 km by
Bessafi et al (2002) based on data of 1988-1997 for the year as a whole including storm and
depression. The results indicate that the forecast difficulty level is higher during the post monsoon
season. It may be due to the typical tracks of the systems including northeastwards and
southwestward recurvature in case of TC GIRI and WARD respectively.
Table 2. Track forecast error based on modified CLIPER Model
Lead time(hrs) Error (Km) No. of cases
12 97 49
24 180 37
36 256 25
48 363 16
60 461 10
72 540 8
The modified CLIPER model has helped to provide better guidance also in respect of larger
spatial coverage and larger lead period compared to older model. To illustrate this fact, the number
of additional cases of forecasts for which guidance is available from the modified CLIPER model is
shown in Table 3. It is due to improvement in climatological database.
Table 3. No. of additional forecasts available from modified CLIPER Model due to
improvement in climatological database.
Lead time(hrs) No. of cases
12 8
24 10
36 4
48 2
60 3
72 3
4. Conclusions: The modified CLIPER model provides better spatial and temporal prediction coverage. It
could provide prediction for more lead period and more geographical area due to increase in
climatological database. It needs to be further validated for entire north Indian Ocean and during
the entire satellite era since 1960 to analyse its efficiency over the region.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 18
References Bessafi, M., A. Lasserre-Bigorry, C. J. Neumann, F. Pignolet-Tardan, D. Payet, and M. Lee-Ching-
Ken, 2002, Statistical Prediction of Tropical Cyclone Motion: An Analog–CLIPER
Approach, Weather and Forecasting, 17, 821-831.
Neumann, C. J., and G. S. Mandal, 1978: Statistical prediction of tropical storm motion over the
Bay of Bengal and Arabian Sea. Indian J. Meteor. Hydrol. Geophys., 29, 487–500.
Pike, A. C., and C. J. Neumann, 1987: The variation of track forecast difficulty among tropical
cyclone basins. Wea. Forecasting, 2, 237–241.
Mohapatra, M., Bandyopadhyay, B.K. and Tyagi, Ajit, 2011, Best track parameters of tropical
cyclones over the North Indian Ocean: a review, Natural Hazards, DOI/10.1007/s11069-011-
9935-0
Sikka, D. R., and Suryanarayana, R., 1968, India Met. Dep., Sci. Rep., 76, 268pp
RSMC, New Delhi, 2009, Reports on Cyclonic disturbances over the north Indian Ocean during
2008, published by IMD, New Delhi.
RSMC, New Delhi, 2010, Reports on Cyclonic disturbances over the north Indian Ocean during
2009, published by IMD, New Delhi.
RSMC, New Delhi, 2011, Reports on Cyclonic disturbances over the north Indian Ocean during
2010, published by IMD, New Delhi.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 19
Possible causes for absence of cyclogenesis over the Bay of Bengal
during October-November 2009
S. ADHIKARY1 AND M. MOHAPATRA
2
India Meteorological Department,
Mausam Bhawan, Lodi Road, New Delhi – 110 003 Email: [email protected]
1 and [email protected]
2
1. Introduction: India has a coastline of about 7,516 km of which 5,400 km is along mainland. The entire
coast is affected by cyclones with varying frequency and intensity. Though only about five
cyclones develop over north Indian Ocean during a year out of 80 cyclones developing over the
globe, more than 75% of the human deaths occur over the north Indian Ocean rim countries due to
the cyclones. Thus the disaster managers and planners need the prediction of cyclogenesis well in
advance. Tourism, insurance and re-insurance companies also make use of seasonal forecasts in
their policy decision.
To predict cyclogenesis, at first detailed understanding about causes of formation and
absence of cyclone is required. There are many case studies regarding causes of cyclone but few
case studies about absence of cyclone. The Madden-Julian oscillation (MJO) can be considered as
one of the factors for occurrence or absence of cyclone. As all cyclonic disturbances (CDs) are
initially formed from these convective cloud clusters (Kalsi, 2002), the modulation of activity
during MJO passage is likely to be an important factor to cyclogenesis. However, the effect of the
MJO on the dynamical parameters may also have an important role to play.
There are very few cases in the recorded history of CD over the north Indian Ocean (NIO) when
there has been no CD over the Bay of Bengal during October to November. There have been five
such years, viz. 1895, 1911, 1920, 1957 and 2009 during 1891 - 2010. However out of the five
years, 2009 is the only year in the satellite era when data on sea areas are most reliable (Mohapatra,
et al 2011). Considering this, a case study has been taken to explain the possible causes of absence
cyclogenesis in the Bay of Bengal during October – November 2009.
2. Data and methodology: In this case study, we have tried to find out whether only MJO is responsible for the absence
of cyclone during October – November 2009 or some other dynamical parameters are also
responsible. In our study, we have taken MJO data from the Centre for Australian Weather and
Climate Research, Australia (http://www.cawcr.gov.au). The NCEP/NCAR reanalysis (Kistler et al,
2001) is utilized dynamical parameters. The reanalysis product provides global data on a 2.50
latitude and 2.50 longitude grid for a large number of dynamical and thermo dynamical parameters
including zonal and meridional wind, mean sea level pressure (MSLP), geopotential height, sea
surface temperature (SST) etc. The reanalysis products are a combination of assimilated
observations along with model derived approximations. The region of study has been taken as 500
E to 1200 E and equator to 40
0N, which is reasonably large to take into account the physical
relationship between the cyclogenesis and the large scale field parameters.
The statistics of cyclonic disturbances over the Bay of Bengal during October - November
have been taken from the best track data set published as Cyclone E-Atlas by IMD.
3. Result and discussion: Analyzing the MJO index (Wheeler and Hendon, 2004), the MJO was in the phase 4 and 5
during 1 – 10 October and it continued to be in phase 8 and 1 which are not favourable for
cyclogenesis over the NIO (Mohapatra and Adhikary, 2011). Though the phase was favourable
during 1 – 10 Oct., the amplitude was less which did not support intensification (Fig. 1). During
November, the MJO index lay over phase 3, 4 and 5 during 7th
to 23rd
November. Among those
days, during 7-12 November there was higher amplitude resulting development of cyclone, Phyan
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 20
over Arabian Sea and no development of cyclogenesis in the Bay of Bengal.
Besides MJO, all the Gray parameters (Gray, 1968) for cyclogenesis have been analysed to
find out the possible causes for the absence of CD during October to November 2009. The middle
layer relative humidity (RH) was not favourable as it was significantly below normal. Geo potential
height anomaly and OLR anomaly were positive over the BoB indicating suppressed convection.
During this period SST was near normal. Also there was dominating mid latitude westerlies
penetrating into Indian region leading to reverse pressure gradient with low over Tibetan region and
high over south BoB in October and November 2009.
4. Conclusion: The MJO is not the only determining factor for the occurrence / non-occurrence of CDs
over the BoB. However its interaction with the dynamical and thermo dynamical features plays an
important role in cyclogenesis over the BoB.
References: Kalsi, S. R., 2002, Use of satellite imagery in tropical cyclone intensity analysis and forecasting,
Meteorological Monograph, Cyclone Warning Division, No. 1/2002, IMD, New Delhi – 110
003.
Kistler, R. et al, 2001, The NCEP/NCAR 50 years reanalysis: Monthly means CD-ROM and
documentation, Bulletin, American Meteorological Society, 82, 247 – 267.
Mohapatra, M. and Adhikary, S., Modulation of cyclonic disturbances over the north Indian Ocean
by Madden - Julian oscillation, MAUSAM, 62, 3 (July 2011), 375-390.
Mohapatra, M., Bandyopadhyay, B. K., Tyagi, Ajit., 2011, Best track parameters of tropical
cyclones over the North Indian Ocean: a review", Natural Hazards, DOI 10.1007/s11069-
011-9935-0
Wheeler, Matthew C. and Hendon, Harry H., 2004, “An all-season real-time multivariate MJO
Index: Development of an index for monitoring and prediction”, Mon. Wea. Rev., 132, 1917-
1932.
Fig. 1: The amplitude and phase of MJO index (Wheeler and Hendon, 2004) during Oct. –
Nov. 2009.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 21
WARD Cyclone – A Case Study
S.R. Ramanan, K.V. Balasubramanian and M.Veerakumar
Regional Meteorological Centre, Chennai
Introduction
A cyclonic storm (CS) WARD (10-15 December, 2009) developed over southwest Bay of
Bengal (BOB) and crossed northeast Sri Lanka coast, close to south of Trincomalee as a deep
depression (DD) between 0800 and 0900 UTC of 14th
December, 2009. It weakened into a
wellmarked low pressure area (WML) over north Sri Lanka at 0300 UTC of 15th
December, 2009.
It then emerged into Gulf of Mannar and became insignificant on 16th
December. Cyclone WARD
followed a rare track, as it moved initially in a northerly direction (Fig.: 1) and moved in a
westsouthwesterly direction across Sri Lanka. It was a slow moving system, as it travelled at the
average rate of 200 km per day (8 km per hour). Before landfall it weakened into a depression. In
this study various features of WARD cyclone is discussed.
Muthuchami and Sridharan (2008) have studied the intensification and movement of CS in
BOB during post-monsoon season. Krishna Rao (1997) has described the synoptic methods of
forecasting tropical cyclones (TC). Desai and Walkar have observed that a CS recurve to the north
and to the northeast when there is a passage of middle and upper troposheric westerly trough.
Occurrence of higher preceipitable water content, higher air temperature at 300 hPa level and
higher upward vertical velocity in lower levels may be indicative of future movement of a CS
(Rameshchand and Mohapatra – 2007). Raj et. al. (2007) considered that the SCS of BOB during
post monsoon season are characterised by the presence of low OLR in the west/northwest and
front/left front sector. Research over the past four decades has established that environmental forces
at large radii have a significant impact on tropical cyclone intensification. It has also been
established that environmental vertical wind shear has a detrimental effect on TC strength. This fact
has been confirmed in recent studies. While small amounts of vertical shear have been seen as
beneficial to TC development shears above 8-12m/s have proven deleterious to TC intensity and
structure. {Levi Thatcher and Zhaoxia Pu (2011)}. Lei Yang et. al. (2011) considered that the West
Pacific Subtropical High of winter-time may be a critical modulator of TC tracks in north Indian
ocean region, specifically during post-winter-monsoon period. Moreover, such a strong weather
system is associated with Northeast Trade Wind and East Asia Winter Monsoon. How westward
this system extends is predictable when information on whether TC in BOB moves westward is
available. However this mechanism and causal relationship between the West Pacific Subtropical
High as well as possible modulated systems will have to be investigated in the future.
Earlier to WARD cyclone of 2009 there is only one CS during 28th
November to 7th
December 1996 which had unusual movement (Fig. 2). This system made first loop over central
BOB near longitude 87.00 E on the night of 30
th November, 1996 and later had a second loop near
the coast of Andhra Pradesh during the night of 4th
December, 1996. Thus, the system created a
unique history in its movement over the BOB. There is no parallel example in the past when a
cyclone executed two loops in the BOB. The very rare southward movement of a tropical cyclone
has been captured well by the altostratus warming from 0758UTC/3rd
December 1996 (Suresh,
2005). WARD cyclone is another such CS which initially moved northwards, then west-
southwestwards and thus followed a rare track.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 22
Results and discussion
Genesis Convective cloud clusters were seen over the south east BOB during first week of
December, 2009 in association with an active Inter-Tropical Convergence Zone (ITCZ). This
developed into a WML over southwest and adjoining south east Bay on 10th
morning {Fig. 3(a)}. It
concentrated into a depression at 0900 UTC of 10th
near 6.50
N/85.00
E.
Intensification and movement It further intensified into a deep depression (DD) near 7.0
0 N/84.0
0 E at 0000 UTC of 11
th
December {Fig. 3(b)}. While moving northward, it intensified into a CS WARD at 0900 UTC of
the same day near 8.50 N/84.5
0 E. It continued as a CS and moved slowly northward till 0600 UTC
of 12th
December {Fig. 3(c)}. It then moved west-south westwards and weakened into a DD over
south west BOB at 1800 UTC of 12th
near 9.50 N/83.5
0 E. Continuing to move in a west-south
westerly direction, it crossed northeast Sri Lanka coast close to the south of Trincomalee between
0800 and 0900 UTC of 14th
as a DD. It weakened further into a depression over north Sri Lanka
close to Trincomalee at 0900 UTC of 14th
and into a WML over Sri Lanka at 0300 UTC of 15th
December. It emerged into the Gulf of Mannar as a LOPAR at 1200 UTC of the same day and
became less marked at 0900 UTC of 16th
December. The track of the system is shown in Fig. 1.
The storm was tracked mostly on the basis of satellite imageries {(Fig 3 (a) to (g)}.
Environmental features The environmental features like sea surface temperature, vertical wind shear of horizontal
wind, mid-tropospheric humidity, low level convergence, upper level divergence were favourable
for cyclogenesis over the southwest and adjoining southeast Bay of Bengal during the first week of
December. The sea surface temperature was about 28-30 deg. C over this region. The vertical wind
shear of horizontal wind was low to moderate (10-15 knots) throughout the life period of the
system except 12th December evening when the wind shear became moderate to high (20-30 knots)
{Fig. 6 (a) and (b)}. It led to the weakening of the system over the sea. However, from 13th
onwards, the wind shear became low to moderate again favouring intensification of the system.
Though, the system did not intensify further due to its interaction with land surface, as it lay close
to Sri Lanka, it could maintain its intensity of deep depression due to favourable vertical wind shear
{Fig. 6 (a) and (b)} and high ocean thermal energy.
The system was close to the upper tropospheric ridge in association with the upper
tropospheric anti-cyclonic circulation throughout the life period of cyclone leading to slow
movement of the system. It was more dominant in the initial phase leading to near northerly
movement of the system till 12th December morning. However, the system lay to the south of a
well defined mid-tropospheric anti-cyclonic circulation, which guided the system to move in a
southwesterly direction from 12th
December onwards {Fig. 5 (a) and 5(b)}. As the system came
closer to Sri Lanka coast, the upper tropospheric flow also supported the system to move in a west
southwesterly direction.
Realised weather The system was away from the Tamil Nadu coast for most part of its life cycle. When the
system weakened in to a Low pressures area and emerged into a Gulf Mannar as a trough it gave
rainfall in Tamil Nadu on 15th
and 16th
December, 2009. Past studies have indicated that whenever
a trough of low lies over Gulf of Mannar with upper air wind favourable up to even one Km is
sufficient enough to generate wide spread rainfall over coastal area leading to active/vigorous
monsoon over coastal Tamil Nadu.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 23
Fig. 1: Track of WARD cyclone
(source: IMD)
Fig. 2: Track of CS – TC08B –
28th
November 1996 to 7th
December 1996 with unsual
southwestward movement.
(a) 10.12.2009 0600 UTC
- Cluster of clouds in
association with ITCZ
forming a WML
(b) 11.12.2009 0600 UTC
DD 8.0/84.5
(c) 12.12.2009 0600 UTC -
CS 10.0/84.5 – tightly
curved banding wrapping
into the center.
(d) 13.12.2009 0600 UTC - S-
ly/SWly movement DD
9.0/83.0
(e) 14.12.2009 0600 UTC
Just before crossing. DD
8.5/81.5
(f) 15.12.2009 0600UTC
Weakened into a WML
over Sri Lanka
(g) 16.12.2009 0600UTC
LOPAR over Gulf of Mannar
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 24
Fig. 3 : Sattellite imageries of WARD Cyclone (source; Dundee – www.dundee.ac.uk)
(a) 12.12.2009 0000 UTC
10.0/84.5 – CS with wind
speed 35 kts.
(b) 12.12.2009 1200 UTC -
10.0/83.5 –convection being
sheaed to east of the storm
centre
(c) 13.12.2009 – 0600 UTC –
9.0/83.0 – DD - wind speed
20 kts -
(d) 13.12.2009 – 1800 UTC –
9.0/82.5 – DD-interaction
with land area
(e) 14.12.2009 - 0000 UTC (f) 14.12.2009 – 1200 UTC
(g) 15.12.2009 – 0000 UTC Fig. 4 (a) to (g): Satellite derived winds showing low level
winds. The low level circulation is strong initially with
stronger wind speeds and reducing when WARD started
interacting with the land mass of Sri Lanka . source ;
RAMMB – rammb.cira.colostate.edu
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 25
Fig. 5 (a) and (b): Mass weighted deep layer
mean wind in two layers (a) 200 to 850 hPa and
(b) 500 to 850 hPa from the balanced 3-D wind
field derived from the AMSU temperature
retrievals. The area averaging is in an area
contained within 0 to 600 km from the centre of
the CS.
Fig. 6 (a) and (b): Area averaged vertical wind
shear in two layers (a) 200 to 850 hPa and (b)
500 to 850 hPa from the balanced 3-D wind
field derived from the AMSU temperature
retrievals. The area averaging is in an area
contained within 0 to 600 km from the centre of
the CS.
Conclusions
CS WARD initially moved in a northerly direction but later moved in a est and
southwesterly direction since it was close to upper tropospheric ridge till 12th
December 2009
morning. Due the same reason it moved very slowly too. Later it moved west and southwestwards
because the upper tropospheric flow supported the system to move in that direction. The vertical
wind shear became moderate on 12th
and due to this the system weakened into a DD.
Reference: 1. Desai. D.S., and Walkar.B.D., Recurving cyclonic storms during 1870-74, Mausam, 48, 3
(July, 1997) 421-28.
2. Krishna Rao. A.V.R., Tropical cyclones – synoptic methods of forecasting, Mausam, 48, 2
(April, 1997), 239-256.
3. Lei Yang et. al., ‘Recent Hurricane Research - Climate, Dynamics, and Societal Impacts’, p-
227-246, Edited by: Anthony Lupo, Published by InTech, April 2011.
4. Levi Thatcher and Zhaoxia Pu, ‘Recent Hurricane Research - Climate, Dynamics, and
Societal Impacts’, p-270-286, Edited by: Anthony Lupo, Published by InTech, April 2011.
5. Muthuchami.A and Sridharan.S.,Intensification and movement of cyclonic storms in Bay of
Bengal during post monsoon season, Mausam, 59, 1 (January, 2008), 51-68.
6. Rameshchand and Mohapatra, Diagnostic study of re-curving cyclone – ‘Mala’ over Bay of
Bengal, Mausam, 61, 1 (January 2010), 11-18.
7. Raj et. al., Severe cyclonic storm of North Indian Ocean, Mausam, 58, 4 (October 2007),
481-500.
8. Suresh.R, ‘Foreshadowing the tracks of tropical depressions and cyclonic storms and
understanding their thermo dynamical structure over Bay of Bengal and Arabian sea using
TOVS and ATOVS data’, ITSC XIV Proceedings, Beijing, China, 25-31 May 2005,
published by Cooperative Institute for Meteorological Satellite Studies · Space Science
and Engineering Center / University of Wisconsin-Madison,
http://cimss.ssec.wisc.edu/itwg/itsc/itsc14/proceedings/A43.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 26
Upper Ocean Observations during the passage of cyclone JAL-2010
Anitha Gera1, Ravichandran M
2 and Mitra A. K
1,
1National Centre for Medium Range Weather Forecast
2Indian National Centre for Ocean Information Services
Severe Cyclonic Storm Jal which developed from a low pressure system occurred in the
Bay of Bengal during 4-8th
Nov 2010. The estimated lowest central pressure was 988hpa. The
cyclone passed very near to a buoy located at 8 N 85.5E. The buoy recorded upper ocean and met
observations during the passage of JAL. These observations are of much value especially in the
context of the scanty observations under such severe weather events. These observations therefore
analysed to enhance our understanding of the air sea interaction processes and the oceans role
during the passage of a cyclone. In addition there are a few Argo floats which recorded the upper
ocean temperature and salinity profiles of the upper ocean near to the track of the JAL-2010.The
evolution of the mixed layer temperature, salinity and currents are examined as the cylone traversed
along its path. Relevant satellite data are also used to complement these in-situ observations.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 27
Salient features of JAL Cyclone of November 2010 – A case Study
D. C. GUPTA
Meteorological Office, Port Blair, India
Email: [email protected] Abstract:
A severe cyclonic storm “JAL” (04- 08 Nov. 2010) developed over the Bay of Bengal from
the remnant of a depression which moved from northwest Pacific Ocean to the Bay of Bengal
across southern Thailand. It moved west-northwestwards and intensified upto severe cyclonic storm
on 6th
November, 2010. However due to lower ocean thermal energy and moderate to high vertical
wind shear, the severe cyclonic storm, JAL weakened gradually into a deep depression and crossed
North Tamil Nadu – South Andhra Pradesh coast, close to North of Chennai near 13.30 N and 80.3
0
E around 1600 UTC of 07th
November 2010.
The unique features associated with system was that it weakened into a deep depression
over Bay of Bengal before landfall. The convective clouds were sheared to the west to a large
extent on the date of landfall, 7th November 2010. As result of this more rainfall occurred over the
interior parts of South India than over coastal regions.
2. Introduction:
The low pressure system over Indian region are classified on the basis of the maximum
sustained winds speed associated with the system and the pressure deficit/ number of closed isobars
associated with the system. The pressure criteria are used, when the system is over land and wind
criteria is used, when the system is over the sea. The system is called as low if there is one closed
isobar in the interval of 2 hPa. It is called depression, if there are two closed isobars, a deep
depression, if there are three closed isobars and cyclonic storm if there are four or more closed
isobars. Considering wind criteria, the system with wind speed of 17-27 knots is called as
depression, the system with wind speed 28-33 knots is called as deep depression, the system with
wind speed 34-63 knots is called cyclonic storm, a severe cyclonic storm if the wind speed is 64-
119 knots and a super cyclonic storm if the wind speed is 120 knots or more. Based on the above
criteria, a severe cyclonic storm formed over Bay of Bengal and crossed over North Tamil
Nadu/south Andhra Pradesh coast and gave copious amount of rainfall over southern peninsula of
India and caused damage of lives & property.
3. Life history (Genesis, intensification/ movement and dissipation) of JAL cyclone:
3.1 Genesis of JAL cyclone: A depression formed over the West Pacific Ocean on31st Oct.2010 in association with an
active Inter –Tropical Convergence Zone (ITCZ). It moved west north-westwards across Southern
Thailand and emerged as a low pressure area over the South Andaman sea on 2nd
November.
Animated imageries indicated merging of mesoscale convective clusters along with increase in
deep convection from 3rd
to 4th
November, 2010. As result of further improvement in convecting
band, the well marked low pressure area continued to move west- northwestwards and concentrated
into a depression at 0000 UTC of 04th
November over Southeast Bay of Bengal near Lat. 8.00
N
and 92.00 E. the track of the system is shown in fig.1.
3.2 Intensification and movement of JAL cyclone: The system intensified into a Deep Depression in the early morning of 5
th November and
into a cyclonic storm ‘JAL’ at 0600 UTC of the same day with centre near lat. 9.00
N and long.
87.50
E about 900 km east-southeast of Chennai. The cyclonic storm ‘JAL’ over southeast Bay of
Bengal continued to move west-northwestwards and intensified further into severe cyclonic storm
in the early morning of 6th
November, 2010. However as the severe cyclonic storm, JAL moved to
southwest Bay of Bengal closer to coast.
3.3 Dissipation of JAL cyclone:
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 28
It entered into a region of lower ocean thermal energy and moderate to high vertical wind
shear in association with the strong easterlies in the upper tropospheric level. The high wind shear
led to westward shearing of the convective clouds from the system centre and Lower Ocean thermal
energy led to unsustainability of convection over the region. Due to these two factors, the severe
cyclonic storm ,JAL weakened into a cyclonic storm at 0600 UTC of 7th
November2010 over
southwest Bay of Bengal with centre near lat.12.50N and long.82.5
0E, about 250 km east-southeast
of Chennai. It weakened further into a deep depression and crossed North Tamil Nadu –South
Andhra Pradesh coast, close to north of Chennai near 13.30N and 80.3
0 E around 1600 UTC of 07
th
November2010. It continued to move west –northwestwards and further weakened into a
depression at 0300 UTC and into a well marked low pressure area over Rayalaseema and adjoining
south interior Karnataka at 0600 UTC of 8th
November2010. The weakening of the system before
landfall could be attributed to lower ocean heat content, though SST was higher than threshold
value.
It emerged into the east central Arabian Sea on 9th
Nov and then moved initially
northwestwards towards Saurashtra &Kutch and adjoining Pakistan coast during 9-11 November. It
then moved northeastwards across Saurashtra & Kutch and adjoining Pakistan coast and became
less marked on 12th
November, 2010. The satellite imageries of the system are shown in fig.2.
4.0 Realized weather and associated damage:
(a) Rainfall: Rainfall occurred at most places with heavy to very heavy rainfall at a few places over North Tamil
Nadu, Pudducherry, coastal Andhra Pradesh, Rayalseema, south interior Karnataka and coastal
Karnataka.
(b) Wind: Squally winds with maximum wind speed reaching upto 60 kmph has been reported from different
observatory stations of IMD along North Tamil Nadu –south Andhra Pradesh coast. Ennore Port of
Tamil Nadu reported 33knots (61kmph) in the forenoon of 7th
Nov. 2010. The wind speed
decreased at the time of landfall, as the system weakened gradually and crossed as a deep
depression.
(c) Damage: Andhra Pradesh: Eleven people died in Andhra Pradesh , hundreds of houses were damaged and
standing crops over about 15000 hectares were destroyed. A loss of about 83 crores was estimated.
Tamil Nadu:
Five persons lost their lives about 100 pucca/kutcha houses were either fully or partially damaged.
Many boats were damaged and some were missing due to floods. Rail, road and air transport were
affected due to heavy rain. Sea water inundated in low lying areas.
5.0 Conclusions: (a) The severe cyclonic storm JAL weakened into a deep depression over southwest Bay of Bengal
before landfall.
(b) the high wind shear led to westward shearing of the convective clouds from the centre of the
system.
(c) the copious amount of rainfall occurred over interior part of southern Peninsula than its coastal
area.
Reference:
IMD, 2011, RSMC, Report on,”Cyclonic disturbances over North Indian ocean during 2010”.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 29
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 30
Characteristics of VLF atmospherics during tropical cyclone ‘AILA’ and several other
thunderstorms over North-East India
Rakesh Roy , Abhijit Choudhury, Anirban Guha and Barin Kumar De
Department of Physics, Tripura Central University, Suryamaninagar-799022, Tripura
E-mail: [email protected]
Lightning discharges radiate most of the electromagnetic energy in the very low frequency
(VLF, 3-30 kHz) and extremely low frequency (ELF, 3 Hz-3 kHz) bands. The exact electrical
processes inside thunderclouds are not yet exactly understood. It is difficult to accurately construct
an empirical model to explain the spectral character of radiated electromagnetic energy from
electrified thundercloud. To understand the physical mechanism precisely, more in situ
experimental data are required.
In the present work, we analyzed the data of VLF atmospherics at four discrete frequencies
received at the Department of Physics, Tripura University during the period from April-October,
2009. We selected temporal data from 76 active thunder-active days over North-East India for the
present investigation. Results show a total of nine different types of characteristic features in the
variation of atmospherics especially during the monsoon period. They are named as Gradual Fall
Gradual Rise (GFGR), Gradual Rise Sudden Fall (GRSF), Gradual Rise Gradual Fall (GRGF),
Gradual Fall Sudden Rise (GFSR), Sudden Rise Gradual Fall (SRGF), Sudden Rise Sudden Fall
(SRSF), Sudden Fall Sudden Rise (SFSR), Sudden Fall Gradual Rise (SFGR) and Spiky. During
the monsoon thunder active days, among all the patterns, GRGF occurred in most of the cases at all
frequencies with an average occurrence number of 37 at each frequency. During our observational
period, the severe tropical cyclonic storm “AILA” (RSMC Designation BOB02, JTWC Designation
02B) occurred over the Bay of Bengal during 23-26 May 2009. Among several characteristic
features during normal monsoon period, SRSF dominated the atmospherics on the 25th
May, 2009
with an average occurrence of 86 numbers in each frequency during the period, when the cyclone
struck the coastal areas of the Bay of Bengal. The VLF atmospherics at different frequencies for
25th
May, 2009 have also been analyzed statistically. The rise rate and fall rate of the atmospherics
for all the patterns are also analyzed for both for the monsoon days and the cyclone active day and
comparative study is performed. The possible interpretation of the observed variations in
atmospherics is explained on the basis of the dynamic electrical activity that occurs inside the
thunder-cloud during cyclonic activity. It appears from the analysis that VLF atmospherics
recorded in North-East India can be used as an efficient remote sensing tool to investigate the
electrical activity of severe thunderstorms and cyclones over Bay of Bengal.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 31
The Role of Meteorological Department Telecommunication Infrastructure on Forecast Demonstration Project (FDP) program over Bay of Bengal.
Sankar Nath
India Meteorological Department,
Mausam Bhavan, Lodi Road, New Delhi 110003, India
The high speed secured communication link to exchange data and warning information is
needed as the Forecast Demonstration Project (FDP) program is aimed to demonstrate the ability
of various NWP models to assess the genesis, intensification and movement of cyclones over the
North Indian ocean with enhanced observations over the data sparse region and to incorporate
modification into the models which could be specific to the Bay of Bengal based on the in-situ
measurements and following the actual track through Satellite and Radar observations. The Main
objective of Telecommunication in IMD is to provide Meteorological data and processed
information to forecasters and users in quickest time to meet their time bound operational
requirements. Over the time advancements in technologies have always been adopted in IMD.
Low speed (up to 300 bps) point t o point teleprinter links were replaced by medium speed (2.5
Kbps) point to point x.25 links and then point to point lease line TCP/IP high speed (64 Kbps)
links were introduced and now high speed (256 Kbps to 2 Mbps) any to any MPLS VPN links
and high speed internet connectivity have been implemented. VPN links provide secure very fast
communication links therefore requirement of transfer of large volume of ASCII and binary data
in short time is possible with it. More over data is simultaneously accessible at many locations
therefore it has overcome the problem of dependency of one center on other center for data
requirements. All AMSS centres, RMCs, MCs, Radar and RS/RW stations have been provided
with VPN links with speeds varying between 256 Kbps to 2 Mbps and speeds can further be
increased depending upon data flow requirements. A central element in a high technology
Meteorological communication environment -TRANSMET AMSS is receive, check and forward
automatically, the meteorological data and products according to the WMO standards.
TRANSMET interconnects our Meteorological sub-systems procured under modernization
project of IMD including High Performance Computer System (HPCs) to run the numerous
numerical weather prediction model installed at NWP section; share in real time our data and
product internally and from/to the meteorological world. It has the audio-visual warning system
for warning message reception and send the whole message to predefined users through e-mail
when a message with particular header is received. It also has the ability to retrieve message from
E-mail and submit that message to GTS.Warning messages can be diverted to predefined FAX
and Mobile numbers through SMS. It can exchange satellite, RADAR, model etc. data file to
predefined users as soon as those are received through FTP.
The data in various research groups under FDP program were exchange through the FTP
server installed at Regional Telecom Hub (RTH) New Delhi as well as through Email-
Group .This FTP server is accessible through high speed internet connectivity. The exchange of
warning message to IMD’s and disaster management group of different state through SMS were
demonstrated through Transmet AMSS.The information of FDP program is also displayed in
IMD website http://www.imd.gov.in.It is clear from the above discussion that the Meteorological
communication has a vital role for data and processed information to forecasters and users in
quickest time to meet their time bound operational requirements as well as conduct any research
program.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 32
Evaluation Of Cone Of Uncertainty in Tropical Cyclone Track Forecast over north Indian
Ocean Issued by India Meteorological Department
D. P. Nayak and M. Mohapatra
India Meteorological Department
Mausam Bhavan, Lodi Road, New Delhi-110003
E-Mail : [email protected]
1. Introduction The "cone of uncertainty"-also known colloquially as the "cone of death," "cone of
probability," and "cone of error"-represents the forecasted track of the center of a tropical cyclone
(TC) and the likely error in the forecast track based on predictive skill of past years as well as
numerous additional details about the TC. The India Meteorological Department (IMD) introduced
the cone of uncertainty (COU) in TC track forecast in December 2009 with effect from TC, WARD
over the Bay of Bengal. Comparing the other Ocean basins, the National Hurricane Centre (NHC),
USA introduced the COU in 2002. Prior to this, IMD issued quantitatively the 24 hr forecast track
and intensity in the interval of 06 hrs in 2003. In 2008, IMD extended this track and intensity
forecast upto 72 hrs (every 6 hrs upto 24 hrs and every 12 hrs subsequently upto 72 hrs). Though
there was no COU prior to WARD cyclone, IMD used to mention the probable area of landfall, by
indicating the likely landfall area between two coastal stations (say, the cyclone is likely to cross
Andhya Pradesh coast between Machilipatnam and Nellore). Hear an attempt is made to evaluate
the COU issued in the graphics by IMD during 2010.
2. Data and Methodology The COU represents the probable position of a TC’s circulation center, and is made by
drawing a set of circles centered at each forecast point—06, 12, 18, 24, 36, 48, and 72 hours for a
three-day forecast. The radius of each circle is equal to the average official track forecast errors of
20(35), 40(75), 60(115) and 80(150) nautical miles (km) based on data of 2003-2008 (Table-1). As
the official track forecast beyond 24 hrs period was not issued by IMD, the radius of circle is taken
as 110(200), 135(250), 165(300) and 190(350) nautical miles(km) based on average errors of quasi-
lagrangian model (QLM) of IMD used for track forecasting during 1999-2008. The cone is then
constructed by drawing a tangent line that connects the outside boundary of all the circles. Over the
Atlantic and Pacific Oceans, the COU is drawn by considering the official average track forecast
errors during past five years. The COU includes several elements, viz: the forecast track line, the
“cone” symbolising the averaged forecast error, landfall area, and background elements, such as the
legend, scale, and underlying map. An example is shown in Fig.1. The frequency of official track
forecast error for five cyclones (Laila, Phet, Bandu, Giri and Jal) during 2010 (RSMC, New Delhi,
2011) lying within and outside COU for all forecast periods has been calculated and analysed.
3. Results and discussion. The statistics of track forecast errors lying between and out side the COU are given in
Table2. It is found that the frequency of errors lying outside the COU increases with increase in
lead period of the forecast. The observed tracks of the TCs are thus expected to lie within 55-75
percent of time. It is in agreement with those over other Ocean basin. The entire track of the TC is
expected to remain within the COU roughly 60-70% of the time over the north Atlantic Ocean and
Pacific Oceans (NHC, 2008). Further, it is found that the frequency of observed track lying outside
the COU is higher in case of recurving TCs. The frequency of such cases was maximum in case of
PHET, which had a rarest of rare track (RSMC, New Delhi, 2011)
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 33
Fig.1. A typical example of observed and Forecast track of depression which later on became
the severe cyclonic storm Jal
Table 1. Radius of the circle based on standard error used to construct COU.
Forecast period (hrs) Standard error (kms)
12 75
24 150
36 200
48 250
60 300
72 350
Table 2. Statistics of official six hourly track forecast error lying within and outside the COU.
Forecast
Period
(hrs)
No. of track forecast
error within COU
No. of track forecast
error outside COU
Total No. of
Forecast
12 37(66) 19(34) 56(100)
24 29(62) 18(38) 47(100)
36 27(75) 9(25) 36(100)
48 17(61) 11(39) 28(100)
60 13(59) 9(41) 22(100)
72 11(55) 9(45) 20(100)
All periods 134(56) 75(44) 209(100)
Figures inside the parentheses are the percentage frequencies.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 34
4. Conclusions
There have been impressive strides in both forecast accuracy and lead time in recent years.
The COU represents the state of the art in forecast products. The frequency of errors lying outside
the COU increases with increase in lead period of the forecast. The observed track of the TCs are
thus expected to lie within 55-75 percent of time. The image has been widely adopted and
disseminated to the public by the media, in part because it is a graphic and thus telegenic (Lundgren
and McMakin 1998). Yet the COU is a complicated figure, containing multiple messages,
presented by multiple graphical elements. Hence there is a need for conducting a survey to evaluate
the users response to this graphical product, as it is done in other basins. (Broad etal, 2007)
References :
Kenneth Broad, Anthony Leiserowitz, Jessica Weinkle, And Marissa Steketee, Bams, 2007,
Misinterpretations of the “Cone of Uncertainty” in Florida during the 2004 Hurricane
Season, BAMS, May 2007, 1-17
Lundgren, R. E., and A. H. McMakin, 1998: Risk Communication: A Handbook for
Communicating Environmental, Safety and Health Risks. Batetelle Press, 473 pp
NHC, USA, 2008, "Definition of the NHC Track Forecast Cone". National Oceanic and
Atmospheric Administration. http://www.nhc.noaa.gov/aboutcone.shtml.
RSMC, New Delhi, 2011, Report On Cyclonic Disturbances Over The North Indian Ocean During
2010, IMD, New Delhi,
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 35
Observational Aspects including DWR for Cyclone Monitoring
S. Raghavan
G1, Prathyeka Apts., New No 12, Old no. 7, 1st Trust Link St., Mandaiveli, Chennai- 600028
Email: [email protected]
The long time dream of routine aircraft reconnaissance of Tropical Cyclones (TCs)
affecting the Indian coasts appears close to realisation. This should enable better understanding of
TCs and, more importantly, more effective forecasts and warnings. Aircraft observations over the
Atlantic and Pacific have over the years contributed the most to our knowledge of TC structure and
behaviour. In situ instrumentation, dropsondes and (Doppler and perhaps Polarimetric) Radar
(helical scan and configurations like pseudo-dual Doppler) need to be deployed on board.
Processing software is quite complex and needs to be robust.
Aircraft reconnaissance is taxing in terms of resources and needs to be fully exploited. The
possibility of mounting an instruments package when required and releasing the aircraft for other
uses at other times needs to be explored. At the surface, traditionally we were dependent on hourly
observations from manned coastal observatories and erratic reports from ships. The establishment
of Automatic Weather Stations (AWS) over land and data buoys over the ocean in the last few
years is a great step forward. AWS’s have helped in determining landfall in recent TCs.
The geostationary satellites have been our mainstay in detecting systems at sea and
estimating their intensity for nearly three decades. Polar orbiting data have been in use from earlier
times though continuous coverage is not there. ISRO’s Oceansat scatterometer data have been very
useful in the case of the recent hurricane IRENE in the USA. The recent launching of Megha-
Tropiques may be of great help. The Tropical Rainfall Measuring Mission (TRMM) satellite data
have been widely used and we may look forward to the Global Precipitation Mission (GPM) too.
In India, ground-based Radar has contributed greatly to improved forecast of TCs over the
past 40 years. We had only non-Doppler analogue radars in the last century but we could improve
TC position determination, track extrapolation and mapping of rainfall distribution significantly.
We learnt a great deal about cyclone structure and behaviour in well-developed as well as weak
systems. We also developed the concept of Radius of Maximum Reflectivity and used it to provide
inputs for storm surge forecasting. With the introduction of Doppler radars in the last 10 years we
are able to get the wind field at close range and hopefully determine maximum winds in TCs close
to the coast. IMD is expanding the Doppler radar network and is likely to induct polarimetric
radars. We are organising networking of radars but more needs to be done. There are various
processing techniques using radar data e.g. the Velocity Track Display. The assimilation of radar
reflectivity factor and Doppler velocity in numerical models has been shown to improve TC
intensity and structure analyses and forecasts significantly. A team at Florida State Unviersity has
developed a rain rate initialization for numerical models that utilize radar resolution rains. We need
to adopt all such techniques.
Besides the expanded ground-based radar network of IMD the deployment of (fully) mobile
radars during TC situations is highly desirable. Clear air ST radars/wind profilers located not far
from the coast have a great potential in the matter of understanding of structure of landfalling
cyclones. There are prospects of establishment of more profilers in the near future. None of these
observational platforms can be viewed in isolation. The real time integration of ground-based,
aircraft-based and satellite data into NWP models needs to be pursued keeping in view the
importance of human judgment. Now several organisations have observational facilities as well as
capability to run models. It is necessary that all concerned organisations collaborate effectively on
a day-to-day basis and share their facilities, data and products. A note of caution is necessary.
Despite the existence of all these facilities and more in the US there has been criticism that the
forecast of intensity change has not been good in the case of Hurricane IRENE which hit the US
east coast in August. This has implications for the continued support we can get from government
for efforts such as BOBTEX to improve our understanding and forecast of tropical cyclones and for
the creation of expensive facilities.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 36
Observations of Cyclones from Space-Based Platforms: Current Status and future Prospects
R.C. Bhatia
Retired ADGM, India Meteorological Department, New Delhi.
Capabilities of meteorological satellites to provide vital observations on Tropical Cyclones
are well known since more than last four decades, Most important are the frequent pictures of earth
cloud cover in the visible, infrared and water vapour channels obtained from Geostationary
meteorological satellites together with the capability of generating a number of quantitative
products from these data. R&D efforts of last several years at the Cooperative Institute for
Meteorological Satellite Studies, Wisconsin have culminated into development of Advance
Dvorak's technique (ADT) for automated analysis of Tropical Cyclones. This technique is currently
operational for North Atlantic and Caribbean Oceans. It is also being used on Experimental basis in
Satellite Division of IMD. Experience of last two years or so in IMD has shown that while
conventional Dvorak Technique works well for cyclones over the Indian seas, current experience of
using ADT over Indian seas does not permit its use on operational basis over our region. More
studies are needed to understand the observed variations between results of ADT and conventional
Dvorak Technique.
R&D efforts of several years at UW-CIMSS have also resulted in the improvements of
quantitative products derived from imagery data. These products have certainly improved the
analysis of Tropical Cyclones and have provided useful information on predicting the future
intensity/ movement of Tropical Cyclones. Quality of currently operational quantitative products
derived from data of Indian satellites is limited by the coarser resolution of the imaging instruments
in use at present on satellites of INSAT/ KALPANA series. With the availability of high quality of
data from the new satellite of INSAT series ( INSAT-3D ) from next year (2012 ) there is a very
good possibility of making further major improvements in the quality of derived products which
will help in better analysis of Tropical Cyclones. Data obtained from the Microwave based
instruments onboard current satellites of Polar orbiting series also complement the conventional
observations based on visible, infrared and water vapour channels. Particularly, the warm core
anomaly observed in the upper troposphere environment of the Tropical Cyclone is very useful as it
is related to the Intensity of the cyclone. Recently launched Megha-Tropique satellite will provide
this data which will improve cyclone analysis over Indian seas. Recently started 3 new METOPS
receiving stations in India will also provide useful products based on microwave data over Indian
regions.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 37
Early Detection of Global Tropical Cyclogenesis using OSCAT Data
C. M. Kishtawal and Neeru Jaiswal
Atmospheric Sciences Division, Atmospheric & Oceanic Sciences Group
Space Applications Centre (ISRO), Ahmedabad-380015, India
Email: [email protected], [email protected]
In the present work, a technique has been developed to predict the global tropical
cyclogenesis. The technique is based on the premise that there is some similarity between the low
level wind circulations of the systems that turn in to tropical cyclones at later stage. This similarity
of wind patterns has been measured quantitatively by computing the “matching index” between the
given wind pattern and the wind speed signatures of developing systems, available from past
observations. For this purpose a database has been formed that contains the low level wind patterns
of the early stages of the systems of that turn into tropical cyclones. The QuikScat derived wind
data of the period 2000-09 have been used to form a database. The India’s polar orbiting satellite
Oceansat-2 was launched by Indian Space Research Organisation (ISRO) on 23rd
September, 2009
for applications pertaining to ocean studies and meteorology. The OSCAT derived wind fields have
been used to predict the genesis of tropical cyclone (TC) formed all over the globe during the year
2011. In the present work, the tropical cyclogenesis of ten cyclones formed in the year 2011 in the
North Atlantic Ocean (viz., Arlene, Bret, Emily, Harvey, Irene, Katia, Lee, Maria, Nate and
Ophelia), ten cyclones formed in East Pacific and five cyclones formed in the West Pacific have
been discussed. The mean prediction lead time of the technique was found as 70 hours.
Keywords Cyclogenesis, OSCAT, QuikScat, tropical cyclone, scatterometer, vector block matching.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 38
Objective Detection of Center of Tropical Cyclone in Remotely Sensed Infrared Images
Neeru Jaiswal, C. M. Kishtawal, P. K. Pal
Atmospheric Sciences Division, Atmospheric & Oceanic Sciences Group
Space Applications Centre (ISRO), Ahmedabad-380015, India
Email: [email protected], [email protected], [email protected]
In the present work, an objective technique has been presented to fix the center position of
TC in the satellite generated infrared images. The basis of the technique is to determine the point
around which the fluxes of the gradient vectors of brightness temperature (BT) are converging.
First, the variance of brightness temperature at each pixel from its neighboring pixels is computed
and then the flux of the gradient of variance values is computed. Next, a line parallel to the
gradient vector at each pixel is drawn across the image, and the locations where these lines intersect
each other are stored in a density matrix. The score values accumulated in the density matrix are
averaged and location with the highest score is identified. This indicates the location where many
lines intercept that indicates a common point that the corresponding gradients are directed toward
(or away from). This position is considered to be the probable center location of the cyclone. This
location is further corrected by matching the BT distribution around a close neighborhood (11x11
pixel) to the 2D Gaussian distribution. The location where the best match is found is fixed as the
center of tropical cyclone. The technique has been tested over the Kalpana Satellite generated
(approximately 900) IR images of the cyclones that formed during the period 2009-10. The
technique has been used in fully automated mode for the five cyclones viz., Phyan, ward, Laila,
Phet, and Jal. The half hourly sequential IR images during the life period of each cyclone is
analysed and the center position is determined. The track of cyclone obtained by the automatically
determined center position is compared with the the observed track obtained from Cooperative
Institute of Meteorological Satellite Studies (CIMSS).
Key words: tropical cyclones, cyclone center, geostationary satellite, Kalpana satellite, infrared
image, gradient vectors, image variance, flux.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 39
Analysis of Tropical Cyclones by Using Microwave Imageries of other Polar Orbiting
Satellites over Indian Region
Suman Goyal and A. K. Sharma India Meteorological Department
Mausam Bhavan, Lodi Road, New Delhi-110003
Since 1982 after the launch of first Geostationary satellite by India, center and intensity of
tropical cyclones is estimated in Sat. Met. Div. by using Visible & IR imageries and applying
Dvorak’s technique operationally which have proven to be invaluable in forecast applications. But
now with the advance of satellite techniques and availability of microwave imageries operationally
from satellites NOAA, DMSP, Metops, AQUA, TERRA etc. These imageries were utilized to
analyse T.C’s, LAILA, PHET, GIRI and JAL by using the same technique. Eye appearance and
accuracy in center determination of cyclone is found to be better in microwave imageries as
compared to IR/Visible images of INSAT. Intercomparison of the intensity and center as measured
by INSAT Visible / IR imageries was also done with other agencies like SSD NOAA, JTWC and
CIMSS Wisconsin.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 40
Estimation of intensity of tropical cyclone over Bay of Bengal using Microwave imagery
T. N. Jha, M. Mohapatra and B. K. Bandyopadhyay
India Meteorological Department
Mausam Bhawan, Lodi Road, New Delhi -110003.
E-mail: [email protected]
1. Introduction: Genesis, intensification and movement of Tropical cyclone (TC) storms over north Indian
ocean are mainly monitored by Infrared (IR) and visible cloud imageries as surface data over
ocean are scanty. Dvorak’s technique (1975) is used to determine genesis and intensity of TCs
which is absolutely based on IR and visible cloud imageries taken by geostationary satellites. The
technique is imprecise as high degree of skill is required to recognise cloud features and patterns
as well as images are of low resolution. Further this technique has limitation at night due to
unavailability of visible imageries. Forecasters essentially require maximum wind speed to issue
disastrous warning likely to hit coastal areas. Microwave is powerful electromagnetic radiation for
atmospheric sounding which is unaffected by clouds as well as transparent to dense cloud mass
due to high weighting function of microwave frequencies in middle atmospheric region. With
higher resolution microwave imageries abundantly available from polar orbiting satellites and are
very useful to monitor genesis and intensity cyclones through measurement of brightness
temperature from various layer of TC. However its application over north Indian ocean has been
limited as it has been operationally applied since 2010 only . Therefore objective of this paper is to
examine Advance Microwave Sensor Suite(AMSU) imageries particularly in frequencies range at
37, 85 and 91 GHz and to establish a relationship between brightness temperature and average
wind speed over Bay of Bengal during cyclonic disturbances formed during Forecast
Demonstration Project (FDP) campaign of 2008 -10 .
2. Data and Methodology: The data of central location, central pressure , Maximum Sustained Wind (MSW) and
Brightness temperature in respect of five cyclones formed over Bay of Bengal during FDP in
2008 - 2010 , viz., “Rashmi”, “Khai muk “, “Nisha” , “Jal” and “ Giri “ have been retrieved
from website of U S Navy TC page (www.nrlmry.navy..mil/) which are observed by various polar
satellites. Best track data have also been obtained from annual report published by Regional
Specialized Meteorological Centre(RSMC) New Delhi ,2009 and 2011. Parameters based on
microwave imageries are interpolated every 6 hourly interval using central differencing scheme in
order to match the data with best track .The MSW based on best track and microwave observations
are compared and analysed . Similarly brightness temperature at centre of the disturbances has been
extracted using the imageries to establish its relation with MSW. Out of the five TCs, three could
attain the intensity of cyclonic storm ( Rashmi, Nisha, Khai muk) , one severe cyclonic storm (
JAL) and one very severe cyclonic storm ( Giri) . Details of classification of cyclones over Indian
seas are given in cyclone manual published by I.M.D (2003). Cyclonic storm “Khai muk “ and
severe cyclonic storm “ Jal ” weakened into deep depression before land fall over south Andhra
Pradesh coast near Kavali and north Tamil Nadu coast north of Chennai respectively .Very severe
cyclonic storm “ Giri “ rapidly intensified over east central Bay of Bengal and crossed Myanmar
coast near Sittewe. The mean surface wind speed in respect of the cyclonic disturbances ( category
wise) corresponding to brightness temperature 230, 240, 250, 260 and 270 Kelvin has been
analysed over Bay of Bengal .
3. Results and discussion: 37, 85 and 91 GHz frequencies based coordinates of system , wind speed and brightness
temperature for each pass of satellites have been scrutinized and found that 37GHz brightness
temperature imageries do not reflect location specific thermal structure distinctly . The location of
centre based on microwave observations 85 and 91 GHz differs from that of best track by about 20
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 41
Km (Table-1). The lowest mean track difference of 18 Km is found in respect of very severe
cyclonic storm and highest 25 Km in respect of depression.
Table-1. Mean location error of best track compared to Microwave ( in Km)
Cyclonic disturbances Mean Error
Depression 25
Cyclonic storm 22
Severe cyclonic storm 21
Very Severe Cyclonic storm 18
Table -2. Mean MSW (Knots) based on best track and Microwave products.
Cyclonic disturbances Microwave based
on MSW
Best track
based MSW
Error(Microwave -
best track)
Depression 35 27 8
Cyclonic storm 46 40 6
Severe cyclonic storm 67 56 11
Very Severe cyclonic Storm storm 116 88 28
Microwave observations generally overestimate MSW compared to best track wind in
respect of all the five cases irrespective of degree of intensity of cyclonic disturbance and
overestimation vary in the range 5 - 35 Knots . Table-2 shows MSW of depression, cyclonic
storm, Severe cyclonic storm and Very Severe cyclonic storm as per best track. The lowest error of
8 Knots is found in case of depression and the highest of 28 knots in respect of very severe
cyclonic storm.
Brightness temperatures vary in the range of 230 – 270 Kelvin over Bay of Bengal. Fig .1
shows that MSW abruptly increases when central brightness temperature rises to 260-270 Kelvin
at pressure level 250 hPa in order to enhance low level convection leading to intensification of
depression/ cyclone into severe cyclonic storms over Bay of Bengal.
Fig.1. Relation between the brightness temperature and maximum sustained wind speed of
TCs
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 42
4. Conclusions:
Following salient of feature of cyclonic disturbances may be drawn from the study.
• The microwave imageries at the frequencies 85 and 91GHz are found to be useful for
monitoring prediction of genesis and intensity of TC over Bay of Bengal.
• Mean difference in location of TC based on best track and microwave is found to be about
22 Km .It decreases with increase in intensity of system.
• Microwave overestimates MSW by about 11 – 28 knots in respect of severe and very severe
TCs and 6- 8 knot only in case of depression and marginal TCs.
• Brightness temperature in the ranges 230 -250 Kelvin is favourable for genesis of TC over
Bay of Bengal and 260- 270 Kelvin is favourable for intensification of TC into severe TC.
The microwave based brightness temperature does provide lead time to predict
intensification of TC into severe TC.
References:
Cecil, D. J and Zipser, E.J, 1999, ”Relationship between tropical cyclone intensity and satellite
based indicators of inner core convection- 85GHz ice scattering signature and lightening
”,Mon .Wea .Rev, 103 -123.
Dvorak, V. F., 1975, “ Tropical cyclone intensity analysis and forecasting from satellite imagery “.
Mon. Wea. Rev., 103, 420-430.
India Meteorological Department, 2003, ”Cyclone Manual”,
Kidder,S.Q. 1979, “ Determination of tropical cyclone surface pressure and winds from satellite
microwave data”. Technical No. 307.
India Meteorological Department, 2009, “ RSMC annual report” of 2008”,
India Meteorological Department, 2011, “ RSMC annual report” of 2011”,
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 43
Making Complete Picture – Radar Composite
B. Arul Malar Kannan#, Suresh Chand,and S.K. Kundu
India Meteorological Department
Lodi Road, New Delhi, 110003
IMD under its modernization has undertaken installation of state of art Doppler weather
radars, these 12 numbers of S-Band Single Polarization radars and the 2 numbers of Dual
polarization radars have high end signal processing receivers with a provision of modifying
inclusive of the processing technique, spurious data filtering, smoothening etc. This customization
during data collection enhances the data quality enabling for its immediate use with less further
processing and quality control.
The paper describes the unified IMD radar network conceived and used in creating specific
products such as Radar Satellite overlay and an operational National Radar Composite achieved by
India Meteorological Department in the recent years.
The main points deliberated in the paper includes
1. Networking concepts and the one in place at IMD
2. Creating a common scan strategy between IMD radars
3. With differing radar processing technologies, way to bring into a common platform.
4. Conversion of radar data from individual proprietary format (Rainbow, BEL, IRIS) to a
unified format for use in custom software in creating a composite.
5. Admissible tolerance time limit, elevation angels limit etc between referred radars.
6. Data sets at overlapping region, various techniques and the one being used
7. Choosing an appropriate map projection for the composite
8. Various useful composite products that can be used
9. Automated product generation and animation in real-time for end-users (WEB-update).
10. Integrating echoes on other regions not covered by Radar from INdian SATellite (INSAT)
HDF5 Asia region data sets.
11. Necessity of a composite for cyclone analysis.
12. Future developments and plans.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 44
Study of Tropical Cyclone AILA using Doppler Weather Radar Data
D. Pradhan
Doppler Weather Radar, Kolkata
A severe cyclonic storm (AILA) formed in the Bay of Bengal during 22-25 May 2009 and
hit the southern coastal region of Kolkata at 0630 UTC. The system further moved northerly as
severe cyclonic storm and crossed Kolkata on 25th
May 2009 during 1000 -1300 UTC. Cyclonic
storm moved slightly northwestwards and then towards north of Kolkata. Doppler weather radar
installed at Kolkata monitored the movement of the cyclonic storm during 0000-1800 UTC of 25th
May 2011 till it weakened as a depression and predicted the track at least 12 hours in advance. As
estimated from velocity images (PPI_V & VVP_2) maximum wind speed associated with the
cyclone was of the order of 75 knots (130 km/h) at an altitude of 0.9 km whereas the max wind at
0.3 km (almost surface wind) of the order of 60 knots at 0635 UTC. It was analyzed from the DWR
reflectivity & velocity images that AILA was a wide core system with no eye formation having two
spiral bands. The maximum vertical extent of the system as measured by DWR was 9 km (where
radar reflectivity reduced to 20 dBz) and average speed of movement before the landfall was 22
km/h. Heavy rain occurred at Kolkata and surroundings since 0600 UTC till 1400 UTC of 25th
May. Other features of this cyclone have been analyzed using DWR velocity data. This is also
established that VVP_2 product is a very good indicator of the arrival/crossing of the system over
the station. The study has been found very suitable for the researchers in understanding the
structure of the cyclone.
Max_Z Pictures of AILA on 25
th May 2009
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 45
PPI_V Pictures of AILA on 25
th May 2009
VVP-2 Pictures of AILA on 25
th May 2009
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 46
STORM SURGE AND COASTAL INUNDATION
S. K. Dube
Indian Institute of Technology, Hauz Khas, New Delhi 110016
Abstract: The destruction due to storm surge flooding is a serious concern along the coastal regions of the
countries around the Bay of Bengal. About 300,000 lives were lost in one of the most severe
cyclones that hit Bangladesh (then East Pakistan) in November 1970. More recently the Nargis
cyclone of May 2008 killed about 140,000 people in Myanmar as well as caused enormous
property damage. Thus, provision of precise prediction and warning of storm surges is of great
interest in the region. The main objective of the present paper is to highlight the recent developments
in storm surge prediction in the Bay of Bengal and also the future plan to enhance storm surge
forecasting capability in the region.
1. Introduction Storm surges are an extremely serious hazard along the east coast of India, Bangladesh,
Myanmar, and Sri Lanka. Although Sri Lanka is affected only occasionally by the storm surge,
however tropical cyclones of November 1964, November 1978 and cyclone of November 1992 have
caused extensive loss of life and property damage in the region. Storm surges affect Myanmar to a
much less extent in comparison with Bangladesh and India. Notable storm surges, which have affected
Myanmar, have been during May 1967, 1968, 1970, 1975, 1982, 1992, 1994, 2008 and 2010; of which
the 1982, 1994, and 2008 (Nargis) caused the heaviest loss of life and damage. Nargis generated storm
surge in excess of 4 m near Ayeyarwady deltaic region. The entire deltaic coast of Myanmar was
flooded with surges ranging from 1.5 - 4.5m.
A detailed review of the problem of storm surges in the Bay of Bengal is given by Ali (1979),
Rao (1982), Murty (1984), Murty et al. (1986), Gönnert et al. (2001), Dube et al. (1997) etc. Although
the frequency of tropical cyclones in the Bay of Bengal is not high compared to northwest Pacific, the
coastal regions of India, Bangladesh and Myanmar suffer most in terms of loss of life and property
damage. The main factors contributing to disastrous surges in the head Bay of Bengal may be
summarized as (Ali, 1979): (a) shallow coastal water, (b) convergence of the bay, (c) high
astronomical tides, (d) thickly populated low-lying islands, (e) favorable cyclone track, and (f)
innumerable number of inlets including world's largest river system (Ganga-Brahmaputra-Meghna).
The purpose of the present paper is to give a review of recent developments in predicting the storm
surges and associated coastal flooding in the Bay of Bengal.
2. Operational storm surge predictions system for the Bay of Bengal In India, the study of numerical storm surge prediction was pioneered by Das (1972).
Subsequently several workers attempted the prediction of storm surges in the Bay of Bengal
(Ghosh, 1977; Johns and Ali, 1980; Johns et al., 1981). Dube et al. (1994), developed a real-time
storm surge prediction system for the coastal regions of India, Bangladesh, Myanmar, and Sri
Lanka. IIT model can be run in a few minutes on a PC in an operational office. One of the
significant features of this storm surge predication system is its ability to investigate multiple
forecast scenarios to be made in real time. This has an added advantage because the meteorological
input needed for surge prediction can be periodically updated with the latest observations and
forecast (data assimilation) from National Weather Services.
Under the auspices of Tropical Cyclone Programme of the World Meteorological
Organization (WMO) the technology (IIT Model) has been transferred to the National
Meteorological and Hydrological Services of the region. Present IIT model predicts only residual
storm surge at the coast line (i.e., water level over and above normal astronomical tides). With the
advantage of simplicity in operation, this model has been used to produce and disseminate timely
warnings to serve public safety. From cyclone season of 2009, Regional Specialized
Meteorological Centre (RSMC) New Delhi is using IIT Model for providing ‘storm surge
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 47
Pegu●
guidance’ to the countries of the region.
3. Validation of Models The reliability of IIT models have been tested using data from several severe cyclones,
which struck the coastal regions of the countries in the Bay of Bengal during last 50 years. The
following sections describe result of numerical experiments carried out to simulate the surges
generated by 2008 Nargis cyclones. The model computed surges is in good agreement with the
available observational estimates.
Fig. 1: Simulated peak surge (m) for 2008 Nargis cyclone (Dube et al., 2009)
Figure 1 depicts the Time history of the movement of Nargis and model computed surge
contours along the coast of Myanmar. The Storm surge model is integrated with a pressure drop of 65
hPa and radius of maximum winds of 25 km obtained from India Meteorological Department. It may
be seen that a maximum surge of 4.5 m is occurred close to the landfall point. The Deltaic region of
Ayeyarwady is affected by surges between 2.5 - 4 m. The Myanmar coast from Pyapon to Yangon is
flooded with a surge of more than 2m. The computed surge values at Pegu and Moulmein are 2.5 m
and 1.5 m respectively. During this cyclone the surge of the order of 4 m was reported by the
Department of Meteorology and Hydrology, Yangon. This is in good agreement with our simulated
sea level elevations.
4. Future Plan to Enhance Storm Surge Forecasting Capabilities for North Indian Ocean While the storm surge prediction for India in particular, and for the North Indian Ocean
region in general, is generally satisfactory, improvements are needed both in storm surge model as
well as meso-scale NWP model to further enhance storm surge forecasting capability in the region.
Keeping this in view IOC-UNESCO/JCOMM organized two Advisory Workshops
(http://www.jcomm.info/SSindia) at Indian Institute of Technology Delhi, during 14-17 July 2009
and 11-15 February 2010, where the international experts on storm surges have worked with the
regional modelling experts to review the current status/performance of an operational storm surge
forecasting model (IIT-D Model) for the North Indian Ocean region and addressed requirements for
upgrading and improving model performance. Workshops also discussed initiatives of the India
Meteorological Department and other Indian national agencies to improve infrastructure required
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 48
for improved prediction of cyclone and associated surges (IOC-UNESCO, 2009 and IOC-UNESCO,
2011). This is a line of activities following the recommendations made at the 1st JCOMM
Scientific and Technical Symposium on Storm Surges (2-6 October 2007, Seoul,
Korea: http://www.surgesymposium.org). This activity is designed and conducted under the
framework of the UNESCO project on “Enhancing regional capabilities for Coastal Hazards
Forecasting and Data Portal Systems”.
REFERENCES Ali, A., 1979: Storm surges in the Bay of Bengal and some related problems, Ph.D. Thesis, University
of Reading, England, 227 pp
Das, P. K., 1972: A prediction model for storm surges in the Bay of Bengal, Nature 239: 211-213
Das, P. K., M. C. Sinha, and V. Balasubramanyam, 1974: Storm surges in the Bay of Bengal. Quart. J.
Roy. Met. Soc. 100:437-449
Dube, S. K., P. C. Sinha, A. D. Rao, and P. Chittibabu, 1994: A real time storm surge prediction
system: An Application to east coast of India, Proc. Indian Natn. Sci. Acad. 60, 157-170
Dube, S. K., A. D. Rao, P. C. Sinha, T. S. Murty, and N. Bahulayan, 1997: Storm surge in the Bay of
Bengal and Arabian Sea: The problem and its Prediction. Mausam 48:283-304
Dube, S. K., Indu Jain, A. D. Rao and T. S. Murty, 2009: Storm surge modeling for the Bay of
Bengal and Arabian Sea, Natural Hazards, 51, 3-27.
Ghosh, S. K., 1977: Prediction of storm surges on the east coast of India. Ind. J. Meteo. Geophys,
28:157-168.
Gönnert, G., S. K. Dube, T. Murty, and W. Siefert, 2001: Global storm surges: theory,
observations and applications. Die Kueste 623 pp
IOC-UNESCO, 2009: Advisory workshop on enhancing forecasting capabilities for North Indian
Ocean storm surges, 14-17 July 2009, IOC Workshop Report no. 223, Paris, UNESCO, 37
pp.
IOC-UNESCO, 2011: Advisory workshop on enhancing forecasting capabilities for North Indian
Ocean storm surges, 11-15 February 2011, IOC Workshop Report no. 239, Paris, UNESCO,
41 pp.
Johns, B., and A. Ali, 1980: The numerical modelling of storm surges in the Bay of Bengal. Quart. J.
Roy. Met. Soc. 106:1-8.
Johns, B., S. K. Dube, U. C. Mohanty, and P. C. Sinha 1981: Numerical simulation of the surge
generated by the 1977 Andhra Cyclone. Quart. J. Roy. Met. Soc. 107:915-934.
Murty, T. S., 1984: Storm Surges: Meteorological Ocean Tides, Department of Fisheries and Oceans,
Ottawa, Canada.
Murty, T. S., and R. F. Henry, 1983: Tides in the Bay of Bengal. Journal of Geophysical Research
88(c-10):6069-6076.
Murty, T. S., R. A. Flather, and R. F. Henry, 1986: The storm surge problem in the Bay of Bengal.
Prog. Oceanog. 16:195-233.
Rao, A. D., 1982: Numerical storm surge prediction in India. Ph.D. thesis, IIT Delhi, New Delhi, 211
pp.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 49
Numerical modeling of Tide-Surge interaction in the Bay of Bengal
Jismy Poulose
Centre for Atmospheric Sciences
Indian Institute of Technology, New Delhi 110016
Bay of Bengal is one of the most vulnerable area of Indian Ocean to the storm surges which
is associated with tropical cyclones. Shallow nature of Bay of Bengal and high tidal range are the
major reasons for the inundation due to storm surge. Nature of tidal phase at the time of land
crossing of cyclone is important to predict the total water level and inundation along the affected
coastal area. Surge amplitude and arrival time of peak surge can be affected by the tide (Johns et al,
1985 and Sinha et al, 2008). The objective of the paper is to study the non- linear interaction of tide
with the surge for cyclone GIRI using IIT-Delhi numerical storm surge model. Tidal solution
generated using open boundary radiation condition is the initial condition for this vertically
integrated shallow water model. Experiments are done to study the tide-surge interaction at the time
of high tide and peak surge. Figure 1 shows that the surge height at the place of landfall varies
according the tidal phase. A positive interaction occurred at the time of peak surge.
Fig. 1. Tide-surge interaction at the place of landfall during severe cyclone GIRI
References:
Johns, B., Rao, A. D., Dube, S. K. and Sinha, P. C. (1985) Numerical Modelling of tide-surge
interaction in the Bay of Bengal. Phil. Trans. R. Soc. Lond. A313, 507-535
Sinha P. C., Indu, J., Bhardwaj, N., Rao A. D., and Dube, S. K. (2008) Numerical modelling of
tide-surge interaction along Orissa coast of India. Natural Haza
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 50
Outlook of tide and Storm Induced Current Off Gopalpur Coast
Susant Kumar Misra*, P. Chandramohan*, A. S. N Murty**, J. K. Panigrahi**,
R. Mahadevan*, M. M. Mahanty*** and J. K. Sahu****
* INDOMER Coastal Hydraulics(P) Ltd, 63 Gandhi Road, Alwar Thirunagar, Chennai
** Department of Marine Sciences, Berhampur University, Berhampur – 760 007
*** NIOT, Ministry of earth sciences, NIOT Campus, Chennai
**** Costal Marine Construction & Engineering Limited, Mira Road (E), Maharashtra
E-mail: [email protected]
In the recent past, there has been increased involvement in the Exclusive Economic Zone
(EEZ) and International waters for exploration and exploitation of living and nonliving ocean
resources. The coastal regions are vulnerable to seaborne events and have significant impact on the
socio-economic condition of the coastal community. The increasing trend in investments towards
myriad coastal infrastructure causes the situation alarming. The new CRZ notification of 2011
seems to be an imperative step of government of India in this context. Coastal Regulation Zone
forms taking into consideration of the tidal line and is silent on the natural hazards and associated
water level rise. In this context, the current study on tide and storm induced current supplements the
knowledge on coastal water level at Gopalpur. This paper presents the analysis of the tide and
storm induced current.
Attempts were made to study the tide and storm induced current along the Gopalpur coastal
region by using the MIKE 21 Hydrodynamic module of DHI Software. The study area experiences
semi diurnal tide with an average spring and neap tidal range of about 1.0 m and 0.4 m respectively.
The average spring tidal range is about 2.39 m and neap tidal range is 0.85 m (Chandramohan et al
1994; Mohanty et al 2010; Mishra et al 2011). While the simulated tide induced current magnitudes
were observed to be around 0.213 m/s for all the periods of the constituents, the order of magnitude
of M2 and S2 tidal constituents (0.099 m/s and 0.034 m/s) observed to be reducing. Similarly, the
directions of tide induced currents during the flood and the ebb phases of tides were observed to
remain almost same. The variation of measured current speeds lay between 0.02 and 0.40 m/s and
the direction predominantly remained within the sector of South - West. For the Hydrodynamic
module, the authors used the parameters of 1977 Andhra cyclone which were obtained from
“Cyclone e Atlas of IMD” and Ghosh (1981). The present simulation shows that the storm induced
currents are approximately 0.80 m/s. The present estimation of the height of storm surge also agrees
well with the estimations of other authors (Subbaramaya et al 1979; Johns et al 1981; Ghosh 1981
and Murty 1984).
Keywords: Tide, Storm induced current, cyclone tracks & parameters, MIKE 21 Hydrodynamic
module &, Toolbox, Gopalpur coast.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 51
Estimation of pressure drop within the tropical cyclone and height of associated storm surge
using Doppler velocity data
D.Pradhan1, Anasuya Mitra
2
1. Doppler Weather Radar, India Meteorological Department, 13th floor New Secretariat
Building, K.S. Roy Road, Kolkata-700001, mail:- [email protected]
2. Junior Research Fellow, India meteorological Department, Mausam Bhawan, New Delhi-
110003, mail: [email protected]
Strong winds and high storm surge are the critical factors associated to Tropical cyclones in
the Bay of Bengal (India).The exact prediction of landfall location, time, wind velocity and
expected storm surge may save thousands of human lives. The large pressure drop within the eye
and the storm surge height are functions of maximum velocity in the eye wall region. These factors
are basic indicators of severity of a cyclone.
A study of five tropical cyclones during post-monsoon seasons in the Bay of Bengal (Fig 1 -
6) has been conducted using Doppler Weather radar radial velocity data to estimate the pressure
drop in the eye of the cyclone and the height of storm surge. Existing empirical relation between
maximum velocity and central pressure drop (Vmax=K√P-Pc) has been modified in terms of radial
velocity measured by the Doppler radar. At present the value of K=14.2 is being used in the above
relation by IMD but a new value of K=13.637 has been found by the authors in this study. The
storm surge height is also calculated for these cyclones using an empirical relation suggested by
SAARC Meteorological Research Centre, Dhacca (Bangladesh) and is found that the values are
very close to the actual occurrence as reported by media and measured by meteorological agencies.
Fig.1. Severe Cyclone (Nov 12, 2002- 0700 UTC) Fig. 2. OGNI (Oct 29, 2006- 0218 UTC)
The study concludes that apart from intensity of a cyclone in terms of eye diameter, radar
reflectivity (precipitation contents and estimated rainfall) and wind speed, central pressure drop and
storm surge height may also be estimated with a very high accuracy using radial velocity data from
Doppler weather radar in the range of 250 km. So far no such study has been carried out in India
for measurement of central pressure drop and storm surge height using DWR data in the Bay of
Bengal coast, present study may be useful in getting estimates for central pressure drop, intensity of
the cyclone and expected storm surge height and additionally may be used for the validation of the
parameters derived from the satellite. The study may also be applied for the validation of the model
output related to the prediction of track and landfall of a tropical cyclone in the Bay of Bengal.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 52
Fig. 3. SIDR (Nov 15, 2007- 1140 UTC) Fig.4.PCAPPI_V of SIDR at 2 km
(15Nov 2007-1140 UTC)
Fig. 5--RASHMI (Oct 26, 2008 -0644 UTC) Fig.6-- AILA (May 25, 2009-0644 UTC)
Key Words:-Tropical Cyclone, Storm Surge, Bay of Bengal, Central Pressure drop, Eye wall,
Radial velocity, Doppler weather radar.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 53
Tropical Cyclones Wind Radii prediction over North Indian Ocean
M. Mohapatra and Monica Sharma
India Meteorological Department
Mausam Bhavan, Lodi Road, New Delhi-110003
E-Mail : [email protected]
1. Introduction: India Meteorological Department (IMD) is the nodal agency for Tropical Cyclones (TC)
monitoring and prediction over the North Indian Ocean. The TC forecast issued by Cyclone
Warning Division of IMD, New Delhi contains forecasts of TC wind field for 3 days in the interval
of 12 hr period. This forecast is issued four times a day based on 00, 06, 12, 18 UTC observations.
The Cyclone Warning Division of IMD introduced the monitoring and forecasting of TC wind radii
during TC, Giri over the Bay of Bengal in October, 2010. The characteristic features of this
forecast, the methodology adopted to generate the forecast and the limitations and future scope are
presented and discussed herewith.
2. Characteristic features of TC wind radii monitoring and prediction: The TC wind radii forecasts are generated in terms of the radii of 34kts, 50kts and 64kts
(1kt = 0.52 ms-1
or 1.85 kmph) winds in four geographical quadrants around the tropical cyclone
(thereafter referred to individually as R34, R50 and R64 for 34kts, 50kts and 64kts wind thresholds
respectively or collectively as wind radii in units of nautical miles (1nm=1.85km)). These wind
radii represent the maximum radial extent of winds reaching 34kts, 50kts and 64kts in each
quadrant. A theoretical description of the concept is shown in Fig.1. The wind radii forecasts are
issued over the sea area only as per the requirement of the users. The thresholds of 34kts, 50kts and
64kts are chosen as the wind of 34kts corresponds to gale wind threshold, the threshold of 50kts
wind is the requirement of mariners and the threshold of 64kts corresponds to wind with hurricane
force.
Fig.1. Radii of surface wind thresholds used by IMD for TC forecasting
3. Methodology adopted for TC wind radii monitoring and prediction: The initial estimation and forecast of the wind radii of TC is rather subjective and strongly
dependent on the data availability, climatology and analysis methods. The subjectivity and reliance
on climatology is amplified in North Indian Ocean in the absence of aircraft observations.
28 kt
34 kt
50 kt
64 kt
Northeast Northwest
Southwest Southeast
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 54
However, recently with the advent of easily accessible remote sensing derived surface and near
surface winds (e.g. Ocean Sat., Special Sensor Microwave Imager (SSMI), low level atmospheric
motion vectors and Advanced Microwave Sounder Unit (AMSU) retrival methods) and advances in
real time data analysis capabilities, IMD introduced TC wind radii monitoring and prediction
product in Oct.,2010. The initial wind radii estimates have become less subjective due to the tools
and products mentioned above. An example of the wind radii forecast issued during TC, Giri (20-
23 October, 2010) is shown in Fig.2.
Fig.2. A typical graphical presentation of cyclone wind forecast during cyclone, GIRI
While better initial estimates of R34, R50 and R64 are becoming available, forecasting these wind
radii remains a difficult task. It is mainly because of the fact that we do not have any objective wind
radii forecast methods and current Numerical Weather Prediction (NWP) models fail to produce
forecasts that are better than climatology (Knaff et al, 2006, Knabb et al, 2006). The road map
followed for monitoring and forecasting of wind radii is given below.
3.1. Roadmap for wind radii monitoring and forecasting over north Indian ocean
(i) Date and time of initial condition
(ii) Official location and Intensity (T/ C.I. No., maximum wind and centre position
(iii) Initial TC wind radii
a. Wind radii based on Oceansat wind
b. SSMI based wind radii
c. Wind radii based on lower level atmospheric motion vectors
d. Wind radii by AMSU retrieval method
e. Wind radii based on global and regional NWP model analyses
f. Wind radii based on DWR wind retrieval
g. Official TC wind radii based on S.N. (a-e).
(iv) Forecast TC wind radii
a. Official forecast of TC intensity and track upto 72 hrs.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 55
b. Persistence forecast based on initial wind radii and past 12 hrs trend.
c. Climatology of TC corresponding to initial condition (i.e. with respect to
location and intensity of TC)
d. Climatological forecast of TC wind radii based on the climatological data base
e. NWP Model forecasts of 10 metre wind radii
• Select the model most appropriate to initial condition
• Compose the wind field distribution to the actual wind
• Calculate the wind radii in four quadrants for the threshold of 34kts, 50kts and
64kts surface wind
f. Official TC wind radii forecast in four quadrants for the threshold of 34kts, 50kts and
64kts based on S.N. (a-e)
A few remotely sensed products for monitoring of TC wind radii are shown in Fig.3.
4. Limitations and future Scope: Wind radii forecasts are somewhat dependent on track and very sensitive to the initial
vortex initialisation in NWP models and intensity forecasts. Over the past several years, there have
been large improvements in track skill ( RSMC, New Delhi, 2011) and modest improvements in the
intensity skill like other Ocean basins. However, it is still important to note that the intensity and
track errors at 24 hrs (say) are still of the order of 15kts and 130 km (RSMC, New Delhi, 2011)
respectively. These errors, particularly the intensity errors negatively affect wind radii forecasts.
The poor intensity forecast is particularly pronounced when intensity forecast fail to or falsely
forecast winds that exceed the 34kts, 50kts and 64kts thresholds. In the absence of a reliable NWP
model, a common approach is to aid and assess the forecast products with a statistical model that
employs a combination of persistence of the initial conditions and trends of initial conditions along
with climatology. Hence, an attempt is being made to develop a climatology and persistence
(CLIPER) model for TC wind radii forecast over the North Indian Ocean like those over north
Atlantic and Pacific Ocean (Knaff et al, 2007). This will provide basic guidance which will be
always available to the forecaster and serve as a reference forecast for verifying other techniques.
5. References:
Knabb R., E. Rappaport, M. Mainelli, J. Franklin, C. Lauer and A. Krautkramer, cited 2006: Progress
toward operational implementation of tropical cyclone wind probability products
(http//www.ofcm.gov/ihc06/Presentations, knabb.ppt#492,31,Slide31)
Knaff, C.P. Guard, J.P.Kossin, T.P. Marchok, T. Smith, and N.Surgi, 2006: Operational guidance
and skill in forecasting tropical cyclone structure change. Workshop Topic Reports of the
Sixth WMO International Workshop on Tropical Cyclones, Tropical Meteorology Research
Programme Rep. 72, 29 pp.(http://www.wmo.ch/web/arep/arep-home.html.)
Knaff John A., Charles R. Sampson, Mark DeMaria, Timothy P. Marchok, James M Gross and
Colin J. McAdie, 2007, Statistical tropical Cyclone wind radii prediction using Climatology
and Persistence, weather and Forecasting, 22, 781-791
RSMC, New Delhi, 2011, Reports on Cyclonic disturbances over the north Indian Ocean during
2010, published by IMD, New Delhi.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 56
Fig.3. Lower level atmospheric motion vectors and surface wind derived from
atmosphericTMI, AMSRE and SSMI around 0000 UTC of 22 October 2010 in
connection with cyclone, GIRI.
(a) (b)
(d) (c)
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 57
Drop size distribution Characteristics of cyclone and convective precipitation observed over
Semi-arid-zone in India
S.Balaji Kumar, S.B.Surendra Prasad, U.V. Murali Krishan and K.Krishna Reddy
Semi-arid-zonal Atmospheric Research Centre (SARC),
Yogi Vemana University, Kadapa
The tropical cyclones (TC) are destructive weather storms characterized by a large low
pressure system originate over oceans and move to the coastal areas-bringing large scale
destruction by violent winds and very heavy rainfall. The TCs formed over Bay of Bengal (BoB)
mostly causes severe damage to the life and economy to coastal regions of Tamil Nadu, Andhra
Pradesh and parts of Kerala. Normally, the coastline districts of AP are affected by cyclones and
floods, whereas during the passage of JAL cyclone, Rayalaseema a semi-aird-region (particularly
Kadapa district), the western and northern parts of Andhra Pradesh also experienced heavy
precipitation. For the present study the micro physical variation of the tropical cyclone, JAL
during 4-7 Nov. 2010 in the BoB is carried out. To understand the effect of cyclone, we carried
out an experiment using cyclonic and non-cyclonic precipitation influence a set of ground based
instruments like Automatic Weather Station, Mini Boundary Layer Mast, Micro Rain Radar, laser
based parsivel Disdrometer and also GPS satellite.
JAL cyclonic storm is formed on 4th
November 2010 at 12:30 IST at 90N and 88.90
0E with
a wind speed of 35 kt. The storm gets intensified with 45 kt wind speed on 5th
November at 00 IST
and located at 9.300N and 88.60
0E. The same pattern continued on 00:30 IST and 06:30 IST.
Further the storm moved towards 9.800N and 86.40
0E on 5
th November at12:30 IST with a wind
speed of 55 kt. On 6th
November, the storm wind speed increased and named as cyclone-1 at
10.500N & 85.70
0E and also more or less same magnitude of wind pattern continued from 00:30
IST to 12:30 IST. After that, the wind speed decreased. On 7th
November at 06:30 IST and 12:30
IST the JAL Cyclone wind speeds were 50 and 35 kt, respectively. On 8th
November the JAL
cyclone is declared as tropical depression with wind speed of 25 kt. Our observational analysis on
characteristic variation of raindrop size distribution (RSD) shows distinctly different DSD
variations during cyclonic and non-cyclonic precipitation. The increase of non cyclonic rain rates
arises from the increases of both drop concentration and drop diameter while the increase of the
rain rate in the cyclonic precipitation is mainly due to the increase of median and large drop
concentration. In the cyclonic precipitation the maximum rain drop diameter does not exceeds 4
mm even at higher rain rate but for non cyclonic precipitation rain drop diameter exceeds 4 mm for
higher rain rates. The higher rain rate with large diameter is due to the local convection and
surrounding environmental atmospheric conditions, where as the cyclonic precipitation with less
than 4 mm diameter observed at Kadapa is due to stratified moisture flow movement from BoB.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 58
Changes in Extreme Daily Rainfall associated with Cyclonic Disturbances over Andaman &
Nicobar Islands in a Warming Climate
Naresh Kumar, M. Mohapatra, A. K. Jaswal and B. P. Yadav
¹India Meteorological Department, Lodi Road, New Delhi-110003
²India Meteorological Department, Pune - 411 005
Email: [email protected]
The studies related to the variation in extreme weather like heavy rainfall are very important
as these events has major impacts on environment and cause considerable damage throughout the
world each year. Manton et al. (2001) studied the trends in extreme daily rainfall and temperature
in southeast Asia and the south Pacific and found decline trend in the frequency of extreme rainfall
at most of the places. Likely cause of extreme weather may be due to increased warming.
According to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change
(IPCC. 2007), the rate of increase of the global average surface temperature is 0.074°C ±.018°C
over the past 100 yr (1906–2005). Extreme weather events are very much important to coastal areas
also. According to Firing et al. (2004), Islands communities are perhaps at the most risk for climate
change events. Andaman & Nicobar Islands is situated in tropical warm pool region and is a
hazardous area influenced by tropical cyclone formation. The studies related to tropical warm pool
regions are very limited. But there are few studies related to Pacific warm pool region. Kruk (2008)
found decreasing trend in heavy rainfall events across Hawaiian Islands and Alaska. He attributed
decrease in heavy events across Hawaiian Islands may be due to possible poleward shift in the
observed Pacific storm track as found by Yin (2005).
Therefore objective of this paper is to examine the frequency of heavy rainfall events (>64.5
mm in past 24 hours) over the region and its linkage with cyclonic disturbance (CDs) over the
region. For this purpose, heavy rainfall data of all the observatories located in Andaman & Nicobar
Islands for their common period between 1961-2000 has considered. The frequency of the CDs
over Andaman & Nicobar Islands are also computed for the period 1961-2000 from IMD (2008).
The average annual frequency of CDs affects Andaman & Nicobar Islands is 2 out of 9 CDs over
the Bay of Bengal during 1961-2000. The monthly percentage of CDs for the period 1961- 2000
over Andaman & Nicobar Islands is shown in Figure 1.
Fig.1. Frequency of CDs over Andaman & Nicobar Islands
There are two peaks in CDs frequency over the region; one is in month of May and another
in month of November. These peaks over the region are mainly due to passes of Inter Tropical
Convergence Zone crosses twice in a year (May and October-November).
In heavy rainfall events, there is decreasing trend in annual as well as all seasonal basis over the
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 59
region. In yearly basis, around 9% reduction in frequency of heavy rainfall from mean yearly
rainfall based on 1961- 2000 data is found. In seasonal basis, there is around 35% reduction in
frequency of pre-monsoon heavy events and 25% reduction in frequency of post-monsoon heavy
events from their respective seasonal mean from 1961-2000.
In CDs events, a significant (confidence level >95%) decreasing trend is found for yearly
frequency between 1961-2000. In percentage wise, it is around 50% of its average annual CDs
between 1961-2000. In season wise, there is around 40% reduction in pre-monsoon season and
around 35% reduction in post-monsoon season from their respective seasonal mean from 1961-
2000. This implies, this reduction in heavy events over Andaman & Nicobar Islands may be
attributed to the strong negative trend in CDs over the area between 1961-2000. Similar trend in
heavy events are also found in west Pacific warm pool region (Kruk, 2008). This also indicate the
reduction in heavy events over the tropical warm pool region.
References: Firing, Y. and M. A. Merrifield (2004), “ Extreme sea level events at Hawaii: influence of
mesoscale eddies”, Geophys. Res. Lett., 31, L24306 doi: 10.1029/2004GRL021539.
IMD-Cyclone e-Atlas (2008) Tracks of cyclones and depressions in Bay of Bengal and the Arabian
Sea 1891-2007.
Kruk MC (2008). Evaluating the impacts of climate change on rainfall extremes for Hawaii and
coastal Alaska. 24th Conference on Severe Local Storms, American Meteorological Society.
Manton, MJ, PM Della-Marta, MR Haylock, KJ Hennessy, N Nicholls, LE Chambers, DA Collins,
G Daw, A Finet, D Gunawan, K Inape, H Isobe, TS Kestin, P Lefale, CH Leyu, T Lwin, L
Maitrepierre, N Ouprasitwong, CM Page, J Pahalad, N Plummer, MJ Salinger, R Suppiah, VL
Tran, B Trewin, I Tibig, and D, Yee (2001) Trends in extreme daily rainfall and temperature in
southeast Asia and the South Pacific: 1916-1998, Int. J. of Climatol, 21, 269-284.
Yin JH (2005) A consistent poleward shift of the storm tracks in simulations of 21st century
climate, Geophys. Res. Lett., 32, L18701, doi:10.1029/2005GL023684.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 60
Monitoring Formation and Movement of the Depression of 16-23 June 2011 using
DWR, Satellite Products and Synergy and Utility of Implimenting a Real time Nowcasting in
IMD for filling the forecasting Gap
Rajendra Kumar Jenamani
Meteorological Watch Office, IMD,
New ATS Building (Room No.-211, 2nd
Floor),
IGI Airport, New Delhi-110037
[email protected]/[email protected]
Various types of weather systems with different spatial and temporal scales are observed during the
summer monsoon. These systems play very dominant role in the behaviour of monsoon circulation and
accompanying rainfall over India. The weather systems observed during Indian summer monsoon can be
basically divided into two parts; (i) synoptic disturbances of transient characteristics and (ii) semi-permanent
systems of quasi-permanent characteristics. Synoptic weather systems during Indian summer monsoon
consist of monsoon disturbances, off-shore trough/vortex along west coast of India, mid-tropospheric
cyclones, cyclonic circulations and western disturbances. Amongest all the monsoonal weather systems,
monsoon depressions are recognized as the main rainfall-producing synoptic weather systems over India.
These are nothing but intese low-pressure areas at the surface with associated upper air cyclonic circulations.
Normally, most of these are formed over the head Bay off Orissa-West Bengal coast and move in a west-
north-westerly direction along the monsoon trough. These disturbances produce cupious rainfall while
moving through central India.
In pre-eighties(Rao, 1976), importance had been given to the structure and associated rainfall
distribution of monsoon depressions and it’s statistical studies, while most of the studies in recent years are
related to its formation, mechanism and mathematical modeling(Jenamani 2001, Jenamani and Dash 2004
and see ref 4-11). Monsoon disturbances formed during June & September are associated with the onset and
withdrawal phases of the monsoon. Four depressions formed during just ended monsoon season of 2011 as
against the normal of 4-6 monsoon depressions per season (June - September). Out of these, two Depressions
(that formed on 11th June over Arabian Sea & the other during 22nd -23rd, July over Land) had a short life
span. The Depression formed during 16th -23rd, June intensified into a DD. Its subsequent west
northwestward movement was responsible for the advance of the monsoon over the most parts of the
country. The fourth Depression formed towards the end of the season (22nd – 23rd, Sept.) weakened before
moving towards northeast. In the present study, attempts has been made to document how well DWR and
Satellite at real time has been able to detect the Formation and Movement of the Monsoon Depression
formed during 16-23 June 2011 using DWR, Satellite Products and Synergy Analysis workstations in view
of the system was having longer life period and moved along the monsoon trough zone, establishing the
through and bringing the monsoon rain/onset to over major part of eastern, central and Northern parts of
India. We have also used various real time forecast issued by NWFC for finding whether heavy/very heavy
rainfall realized along its path were captured by 24 hours and 48 hours forecast updated at 6-hours to find
their accuracies and inherent problems which limits such forecast accuracies.. Present analysis shows being
the depression forming on the onset phase of monsoon there were quick organization and re-organization of
intense convective cloud systems associated with this depression at various phases of its movement due to
which convention forecast updated at 6-hours have been found to be have limited accuracies. During when
depression was moving across West Bengal, DWR shows reflective in Max Z reaching as high as 54dbz on
morning of 17 June when the system was near Kolkota. When, the system was monitored at 24X7, using
various observing and analysis systems system such as DWR’s 10-minute products, 30-minute Satellite data
at real time, with 3-hourly analysis of various other data e.g. AWS, synop etc by Synergy, it shows such new
technology based monitoring system now have attended full potential of implementing Operational
Nowcasting in IMD which will fill the forecasting Gap arising because of inherent limitation of present
forecast system to capture such high variability at 0-6 hour time scale and district-wise heavy rainfall
warnings.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 61
References
H.R.Hatwar, Rajendra Kumar Jenamani, S.R.Kalsi and S.K.Subramanian, 2005, “Synoptic
Weather Conditions during ARMEX”, Mausam, 56, 7-18.
Jenamani Rajendra Kumar and Dash, S. K., 2005, “A Study on the Role of Synoptic and Semi-
permanent Features of Indian Summer Monsoon on it’s Rainfall Variations during Different
Phases of El-Nino” Mausam,, 56, 4, 825-840
Jenamani Rajendra Kumar, 2004, “Distinct synoptic patterns associated with pre-break onset phase
and revival of normal monsoon phase”, Mausam, 55 , 591-598
Jenamani Rajendra Kumar, 2007, “Does Break Monsoon always mean Subdued Rainfall over
India? -An analysis of Role of Off-Shore Trough in this aspect”, Mausam, 58, 572-579
Jenamani Rajendra Kumar, Dash, S. K. and V. Thapliyal, 2004. “Decadal and epochal variation of
frequency and duration of monsoon disturbances and their secular relationships with rainfall
over India”, Mausam., 55, 3, 397-408.
Jenamani Rajendra Kumar, H. R. Hatwar, S.R.Kalsi and S.K.Subramanian, 2007, “Another
Deficient monsoon 2004-A comparison with drought year 2002 and possible causes”,
Mausam, 58, 161-176
Jenamani, R. K. and Bhan, S. C., 2008, “Exceptional rainfall event of 26th
July, 2005 over
Mumbai- Radar Echoes and Rainfall”, Mausam, 59, 3, 366-376
Jenamani, Rajendra Kumar, 2001, "Mathematical modelling of weather systems during Indian
summer monsoon", Ph. D. Dissertation, Utkal University, Bhubaneswar.
Rao, Y. P., 1976 "Southwest Monsoon” Met. Monograph, India Meteorological Department,
Synoptic Meteorology, No. 1 /1976.
S. K. Dash, R. K. Jenamani and S. Sudhansu, 2004, “Decreasing frequency of monsoon
depressions over Indian region and associated parameters”, Current Science, 86, 10, 1404-
1411.
S. R. Kalsi, R. K. Jenamani, H. R. Hatwar and, 2006, “Meteorological features associated with
severe drought 2002 ” Mausam, 57, 3, 459-474
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 62
Forecasting of rainfall from landfalling cyclone using satellite derived rain rate data: A case
Study for cyclone ‘Aila’
Habibur Rahaman Biswas and P.K.Kundu*
Regional Meteorological Centre, Kolkata
*Jadavpur University, Kolkata
Email: [email protected]
An important major threat to life and property of east coast of India including West Bengal
Coast is very heavy rainfall from landfalling tropical cyclones originated over Bay of Bengal.
Forecasting magnitude of rainfall from landfalling tropical cyclone is very difficult job. With the
advent of weather satellites, no tropical cyclone anywhere over the globe goes undetected or evades
the eyes of meteorologists. Satellite derived rain rates through cloud area of tropical cyclone can be
used to forecast potential tropical cyclone rainfall accumulations. In the present study, estimation of
24 hours rainfall over Coastal stations before landfall of tropical Cyclone ‘Aila’ has been analysed
using tropical rainfall measuring mission(TRMM) satellite rain rates data and observed storm track
of Aila. Magnitude of estimated rainfalls for the case of cyclone ‘Aila’ nearly match with observed
rainfall over coastal stations. Study explores the feasibility of forecast for 24 hours rainfall from
landfalling cyclone over Bay Bengal using satellite estimate rain rate and storm forecast track.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 63
Unprecedented flood in river Mahanadi in Orissa in September, 2008 and its impact on
economic development
S.C.Sahu and S.K.Dastidar
Meteorological Centre, Bhubaneswar
State of Orissa is prone to Meteorological hazards such as heat wave, tropical cyclone and
flood. Flood in every year cause damage to kharif crops mainly paddy besides loss of lives and
properties. Northern part of Bay of Bengal is prone to cyclone genesis-that is sea surface
temperature and other meteorological conditions are favourable for formation of low pressure and
its intensification particularly during the monsoon season. The circulation pattern is such that they
move in north-west or north-north-west or west-north-west direction . Orissa coast is directly under
the track of these moving monsoon depressions or deep depressions. Out of seven low pressures
formed over Bay of Bengal during monsoon period of 2008, three systems intensified into
depression and one among them further intensified into deep depression on 16th
September,2008.
Deep depression on 16th
Sept.2008 crossed Orissa coast near Chandbali causing heavy rainfall in
catchment areas of Mahanadi. It caused massive flood surpassing earlier records of flood in
1834,1955 and 1982, 15.81 lakh cusec water were released on 20-9-2008 at Munduli near Cuttack.
19 districts are affected due to inundation of flood water by breaching of 477 embankments and
2895 road breaches. Due to this flood, 4,78,854 hectares of crop area damaged and 40,95,547
numbers of people are affected.
Floods in last 30 years in Mahanadi river has been analysed and rainfall data of rain gauge
stations in Mahanadi catchment are taken for calculation of area-depth relation for large storms.
One day average rainfall on 17th
September,2008 are recorded in districts of Jajpur (211.8 mm) ,
Bhadrak (201.4 mm) , Cuttack ( 136.7 mm) , Kendrapara (205.3 mm) , Bargarh (71.2 mm) , Puri
(52.7 mm) , Khurda (63.1 mm) , Kalahandi (135.2 mm) , Bolangir ( 131.4 mm) , Sonepur ( 131.0
mm) , Kandhamal (177.0 mm) and Boudh ( 60.0 mm) . These districts are in the catchment areas of
Mahanadi. To mitigate the suffering of people during flood , food droppings were made for more
than a week in affected areas and on precautionary measure before occurrence of flood, 3,91,907
numbers of people were evacuated to safer places.
After crossing the coast, the depression maintained its intensity over land for another 12
hours which caused heavy rainfall in interior districts of Orissa and also in Upper Mahanadi
catchment areas in the state of Chattisgarh. Attempt has also been made to study the synoptic
situation of the system for causing such copious rainfall. When deep depression moved to land,
mainly development of cumulonimbus cloud and thundery shower occurred behind the trough line.
Low level convergence behind the trough and divergence ahead of the trough are agreeing to
equation for the conservation of potential vorticity.
(f + ξ )/ ∆p = К
Where f= Coriolis parameter, ξ = Relative vorticity ( cyclonic +ve) and
∆p = Depth of air column
Air overtaking trough line is moving both upwards ( f increasing ) and towards a zone of
cyclonic curvature ( ξ increasing), so ∆p must increase to keep left hand side of equation remain
constant. This vertical expansion of the air column necessitates horizontal contraction
(convergence). Conversely, there is divergence in the air moving southward ahead of the trough
and curving anticyclonically.
Analysis of OLR and SST will also be done for occurrence of heavy rainfall for this deep
depression.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 64
Deep Depression without Heavy Rainfall
Bikram Singh, R.C. Vashisth, B.P. Yadav and Charan Singh
India Meteorological Department
Mausam Bhavan, Lodi Road, New Delhi-110003
A depression formed over Bay of Bengal at 0000 UTC of 19
th October 2011. It rapidly
intensified into a deep depression at 0300 UTC of the same day and crossed Bangladesh-Myanmar
coast south of Cox’s Bazaar (Bangladesh) as deep depression by following north-easterly track with
high speed of movement. The uniqueness of the system was that no heavy rainfall occurred and the
movement was faster than the average while it was near the centre of the anti-cyclone. The main
reason for fast movement seems to be the shearing influence due to southward shifting of Sub-
tropical Westerly Jet (STWJ) and the dry cold environment was not conducive for heavy
precipitation. The detail features, causes of fast movement and no heavy rainfall associated with
the system are analysed and discussed in this paper.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 65
Lessons from IRENE…
S. Raghavan
G1, Prathyeka Apts., New No 12, Old no. 7, 1st Trust Link St., Mandaiveli, Chennai- 600028
Email: [email protected]
The long time dream of routine aircraft reconnaissance of Tropical Cyclones (TCs)
affecting the Indian Coast appears close to realisation. This should enable better understanding of
TCs and, more importantly, more effective warnings. Aircraft reconnaissance is taxing in terms of
resources and needs to be fully exploited. The possibility of mounting an instruments package when
required and releasing the aircraft for other uses at other times needs to be explored. Airborne
Doppler Radar needs to be deployed as it is the device which has led to most of the knowledge of
cyclone structure. Besides its recognised uses, radar can contribute inputs to storm surge
forecasting.
The real time integration of ground-based, aircraft-based and satellite data into NWP
models needs to be pursued keeping in view the importance of human judgment.
The reason for the present title is the context of the Hurricane IRENE in the US. Many
comments have been made that the intensity forecast was not good and that the weakening close to
the coast produced an anticlimax and perhaps over-warning. The criticism of over-warning does
not seem justified. Intensity forecast is particularly difficult. The double eyewall feature seen on
radar which is being commented upon is an indicator of a severe TC but not a reliable predictor of
intensity change. The Oceansat scatterometer data have been very useful in the case of IRENE.
The user is interested not in the phenomenon but its IMPACT. Operationally therefore it is
important to put out warnings with graphics indicating the various possible scenarios and
explaining the uncertainties, while keeping close liaison with disaster managers. This is being
effectively done in the US where weather telecasters are qualified meteorologists. In India we often
give a “deterministic” type of forecast and keep changing it in the light of observations without
explaining the background to the public. While this may be justified scientifically it projects a poor
image of the Meteorological Service with a loss of credibility. It may be better to give a
probabilistic forecast explaining the uncertainties and taking recipients into confidence.
It is also necessary to create greater awareness of the importance of pro-active preparedness
among administrators and ensure more funding for that rather than just for relief measures after the
event.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 66
NWP models applications in Tropical Cyclone Predictions over the Bay of Bengal
U. C. Mohanty*, Krishna K Osuri and S. Pattanayak
Indian Institute of Technology Delhi, Hauz Khas – 110016, New Delhi, India
In recent days, Weather Research and Forecasting (WRF) model, the state-of-the-art
mesoscale model, is used worldwide, both in research and operational environments, for the
simulation of high impact extreme weather events such as tropical cyclones (TCs), heavy rainfall
events, and severe thunderstorms. In respect of TC prediction, a number of studies demonstrate
high degree of performance of WRF modeling system over different global cyclone basins. In the
present study, we provided a detailed evaluation of model performance of WRF model for the
simulation of TCs over the Bay of Bengal (BoB).
First, relative performance of widely well known Mesoscale Model version 5 (MM5) and
the Advanced Research WRF (ARW) model are evaluated which clearly indicate that the ARW
model could provide better track and intensity prediction of BoB cyclones compared to that of the
MM5. The improvement with ARW is not noticeable up to 48 hours and after that the improvement
is significant (Figure 1). The mean track error is reduced to almost half with ARW model
compared to that of the MM5 model. There is considerable improvement of about 8%, 5%, 5% and
73% in intensity prediction at 12, 24, 48 and 72 hour forecast respectively, over that of the MM5
model (Pattanayak and Mohanty, 2008). The performance of Hurricane WRF (HWRF) is also
studied for one severe cyclone “Mala” and reveals that the improvement in intensity is significant
compared to the improvement in track prediction. The track forecast error for the cyclone Mala
varies from 180 to 300 km from 24 to 72 hour forecast (Pattanayak et al, 2011).
The operational utility of the HWRF needs high computation requirements. However, a
detailed evaluation of real time predictions of the ARW model for North Indian Ocean (NIO)
cyclones is analyzed and can be found at Osuri et al. (2011a). Before going for real time prediction
of BoB TCs, the ARW model is customized to the same domain by simulating number of cyclones
with different physical parameterization schemes (Osuri et al., 2011d). The 12 TCs of the BoB
during 2007–2010 are predicted in real time at different times and have a total of 71 forecast cases.
The mean track forecast error from the ARW model varies from 170 km to 350 km for 24 to 72
hour forecast length. The mean landfall errors are in the range of 60 to 140 km in 24 to 72 hour
prior to observed landfall time. From the systematic errors, it may be noted that, the ARW model
has right side (eastward) bias and slower (southward) bias for all forecast intervals.
A number of experiments are conducted to assess the impact of assimilation of different
sources of data on initialization and forecast of the ARW model for the simulation of TCs over the
BoB and Arabian Sea. The inclusion of QSCAT and SSMI satellite-derived winds, through a 3-
dimensional variational (3DVAR) data assimilation system into the ARW model initial condition,
improves the initial position in 11 cases out of 13 by 34% (Osuri et al., 2011b). From Figure 3, the
24-, 48-, 72- and 96-hour mean track forecast improves by 28%, 15%, 41% and 47%, respectively,
based on 13 cases (of Narigs with 5 cases, Gonu with 4 cases, Sidr with 2 cases, and KhaiMuk with
2 cases). The landfall prediction is significantly improved in 11 cases by about 37%. Further, the
intensity prediction also improves by 10–20%. Kinematic and thermodynamic structures of TCs are
also better explained, as it could simulate heat and momentum exchange between sea surface and
upper air. Due to better simulation of structure, intensity and track, the 24-hour accumulated
rainfall intensity and distribution are also well predicted with the assimilation of satellite-derived
winds (Osuri et al. 2011b).
Further, the impact of Global Telecommunication system (GTS) data and the Doppler
Weather Radar (DWR) data on simulation of TCs are also examined. Assimilation of the DWR
data significantly improves track and intensity. The mean radial wind error is reduced from 2.59 m
s-1
(FNL analyses) to 1.67 m s-1
in the DWR analysis. The DWR experiments show better temporal
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 67
evolution of intensity with better simulation of surface latent heat flux, relative vorticity at 850 hPa,
and inner core structure of TCs. The improved upper-level divergence and steering flow helps for
better track prediction in the DWR experiments with a large gain in skill, particularly at longer
forecast intervals. Figure 4 indicates that the mean track errors (of all 16 cases) are less for the
DWR experiment tracks and varies from 50 to 250 km from the 12 hr forecast to the 72 hour
forecast. The track errors in case of GTS experiment ranges up to 400 km while in the CNTL, the
errors are even higher and up to 500 km. The gain in skill of the DWR data ranges from 33% to
74% from the 12 to 72 hour forecast. Out of 16 cases, CNTL and GTS could predict the landfall in
8 and 10 cases, while, the DWR experiment succeeds in 14 cases with minimum errors. The
landfall time errors are also reduced in the DWR experiments in most of the cases as compared to
those of others. The mean landfall error of CNTL (8 cases), GTS (10 cases), and DWR (14 cases) is
78, 64, and 66 km respectively. Considering the same 8 cases as that of the CNTL experiment, the
mean errors are 68 and 42 km for the GTS and DWR experiments, respectively. The model-
simulated reflectivity at landfall and the 24-hour accumulated rainfall are also well simulated in the
DWR experiments as compared to CNTL and GTS experiments.
Conclusions: In view of the above results, the following broad conclusions can be drawn:
The WRF modeling system performs better in real-time predictions of the Bay of Bengal
tropical cyclones compared to the MM5 system. HWRF model also provides better track and
intensity prediction of TCs over the BoB. However, the model performance is significantly
improved with data assimilation using additional high density remote sensing data such as satellite
derived winds and DWR radial wind and reflectivity products. Therefore, the high resolution
mesoscale models can provide a good guidance for the track, intensity and landfall prediction of the
Bay of Bengal TCs to the operational forecasters.
References: Krishna K. Osuri, U. C. Mohanty, A. Routray and M. Mohapatra, 2011a: Mean track errors of
Landfalling tropical cyclones of 2007-09 over Indian seas as evident from WRF-ARW
modeling system, Submitted to Atmospheric Research (Under review).
Krishna K. Osuri, U.C. Mohanty, A. Routray and M. Mohapatra, 2011b: Impact of Satellite
Derived Wind Data Assimilation on track, intensity and structure of tropical cyclones over
North Indian Ocean, International Journal of Remote Sensing, 1 – 26,
DOI:10.1080/01431161.2011.596849.
Figure 1: Mean track errors of TCs over the BoB from ARW-
WRF and MM5 model
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 68
Krishna K. Osuri, U. C. Mohanty, D. Niyogi, A. Routray, and D. V. Bhaskar Rao, 2011c, Improved
Prediction of Bay of Bengal Tropical cyclones through Assimilation of Doppler Weather Radar
Observations, Submitted to JGR (Under review).
Krishna K. Osuri, U. C. Mohanty, A. Routray, Makarand A. Kulkarni and M. Mohapatra, 2011d:
Sensitivity of physical parameterization schemes of WRF model for the simulation of Indian
seas tropical cyclones, Natural Hazards, DOI 10.1007/s11069-011-9862-0
Pattanayak S, Mohanty UC (2008) A comparative study on performance of MM5 and WRF models
in simulation of tropical cyclones over Indian seas. Current Science, 95(7):923–936.
Sujata Pattanayak, U. C. Mohanty, and S. G. Gopalakrishnan, 2011: Simulation of very severe
cyclone Mala over Bay of Bengal with HWRF modeling system, Natural Hazards, vol. 52, DOI
10.1007/s11069-011-9863-z.
Mean track forecast errors for BoB cyclones
-8
23
40
48 51 51
0
100
200
300
400
500
600
700
800
12 24 36 48 60 72
Forecast Length (hour)
Po
sitio
n E
rro
rs (km
)
-20
-10
0
10
20
30
40
50
60
Gain
in
Skil
l (%
)
ARW-WRF
Persistence
Gain in Skill (%)
Figure 2: Mean (of 71 cases) track errors of BoB cyclones from the real time
predictions of ARW-WRF model. The gain in skill (%) of the model
with respect to persistence track is also shown in line graph.
Figure 3: Mean vector displacement errors (VDEs in km) in 12-hr interval for CNTL
and 3DVAR experiments (a) Nargis (mean of 5 cases), (b) Gonu (mean of 4
cases), (c) Sidr (mean of 2 cases) and (d) KhaiMuk (mean of 2 cases).
12
0
18
1
17
2
22
1 24
4 26
7
26
2
15
5
11
5 12
7 15
1
14
5
13
1
11
2
10
3
11
3
0
50
100
150
200
250
300
12 24 36 48 60 72 84 96
(b) TC Gonu
90
21
3
25
2
30
0
16
6
11
2
22
4
14
1
78
20
8
21
4
15
1
11
0
17
7
60
14
3
0
50
100
150
200
250
300
350
6 12 18 24 30 36 42 48
(d) TC KhaiMuk
13
8 19
5 23
0
30
6 37
8
58
6
60
9
12
9
89
94
56
0
50
7
34
6
28
7
13
0
11
9
0
100
200
300
400
500
600
700
6 12 18 24 30 36 42 48
(c) TC Sidr
Me
an
VD
Es
(km
)
86
17
7
27
5 31
8 37
3
46
5
58
9
13
5
98
86
33
5
26
7
22
0
18
5
18
1
11
0
0
100
200
300
400
500
600
700
12 24 36 48 60 72 84 96
(a) TC Nargis
Me
an
VD
Es
(km
)
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 69
IMD’s recent initiatives for improved Tropical Cyclone track and intensity forecast over
Indian region using Hurricane WRF Model
Y.V. Rama Rao1, T.S.V. Vijay Kumar
2, Zhan Zhang
2, K. Naga Ratna
1, A.K. Das
1, D.R.
Pattanaik1, S.K. Roy Bhowmik
1 and Ajit Tyagi
1
1India Meteorological Department, New Delhi
2Environmental Modelling Centre (EMC), NCEP, USA
During the last 15 years, IMD running various numerical models for Tropical Cyclone track
prediction such as Limited Area Model (LAM), Quasi-Lagrangian Model (QLM) for operational
numerical guidence. With the operationalisation of High Power Computing System (HPCS) for
numerical modelling at IMD, New Delhi, IMD started high-resolution Global model (35 km
horizontal resolution) for medium range and WRF model for short range track and intensity
forecast. Recently under Indo-US joint collaborative program, IMD adapted HWRF model for
Tropical Cyclone track and intensity forecast for North Indian Ocean (NIO) region for its
operational requirements.
The Weather Research and Forecast (WRF) system for hurricane prediction (HWRFTM
) is
operational at National Centre for Environmental Prediction (NCEP), USA since 2007, providing
deterministic forecast guidance to the National Hurricane Center (NHC) for the Atlantic and North
Eastern Pacific basins. This advanced hurricane prediction system was developed at Environmental
Modelling Centre (EMC), NCEP to address the Nation's next generation hurricane forecast
problems. The HWRFTM
is a high resolution coupled air-sea-land prediction model with a movable
nested grid and advanced physics for high resolution. This model is currently coupled to the
Princeton Ocean Model (POM) in the Atlantic basin. The HWRFTM
has the capability to address
hurricane structure and rainfall forecast problems in addition to advancing wave and storm surge
forecasts. Continued advancements in track and intensity prediction are important focus areas of
this prediction system.
The basic version of the model HWRFV (3.2+) which was operational at EMC, NCEP was
ported on IMD IBM P-6/575 machine with nested domain of 27 km and 9 km horizontal resolution
and 42 vertical levels with outer domain covering the area of 800x80
0 and inner domain 6
0x6
0 with
centre of the system adjusted to the centre of the observed cyclonic storm. The outer domain covers
most of the North Indian Ocean including the Arabian Sea and Bay of Bengal and the inner domain
mainly covering the cyclonic vortex with moving along the movement of the system. The model
has special features such as vortex initialization, coupled with Ocean model to take into account the
changes in SST during the model integration, tracker and diagnostic software to provide the graphic
and text information on track and intensity prediction for real-time operational requirement.
As part of model validation, case studies were undertaken to test the ability of the model for Indian
Seas for Very Severe Cyclonic Storm ‘GIRI’ formed during 20-23 October 2010 and Severe
Cyclonic Storm ‘JAL’ formed during 4 to 7 November 2010 over the Bay of Bengal. The model
was integrated for 5-days forecast with basic input from GFS spectral fields using Gridpoint
Statistical Interpolation (GSI) assimilation system. Also the six hourly cycling of the HWRF model
was tested to run the model in cycling mode. In this run only the atmospheric model (HWRF) was
tested. The Ocean Model (POM-TC) and Ocean coupler requires the customization of Ocean
Model for Indian Seas. In this regards, IMD is expecting to work in collaboration with INCOIS,
Hyderabad which is running the Ocean Models (POM)/Hybrid co-ordinate ocean model (HYCOM)
to support in porting the Ocean Model with Indian Ocean climatology and real time data of SST
over Indian Seas.
The detailed model configuration and validation results along with the limitations and future
plans will be discussed.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 70
Impact of cyclone bogusing and regional assimilation on tropical cyclone track and intensity
predictions.
Manjusha Chourasia, R. G. Ashrit, John P George
National Centre for Medium Range Weather Forecasting (NCMRWF),
Ministry for Earth Sciences, A-50, Institutional Area, Phase-II
Sector-62, NOIDA, U.P., Pin : 201 309
Email : [email protected]
An attempt is made to assess the impact of tropical cyclone bogusing in WRF assimilation
and forecast system for cyclone track and intensity prediction in short range forecast. Three
tropical cyclones formed in the year 2010 are chosen as study cases; namely 'LAILA' (Bay of
Bengal), 'GIRI' (Bay of Bengal) and 'PHET' (Arabian Sea),.The operational NCMRWF T382L64
analysis and forecasts are used to provide initial conditions for WRF model. The WRF model is
integrated upto 72 hr for producing the cyclone track and intensity forecast. In the tropical cyclone
bogusing scheme used in this study, the existing cyclone vortex in the initial condition is removed,
and an artificial cyclone vortex is ingested at observed location by supplying observed data of
tropical cyclone centre along with intensity and radius of maximum wind. Four sets of model
experiments were carried out: (1) The control experiment (CNTL); Cold start run with interpolated
global T382L64 model analysis initial condition without performing any bogusing or assimilation.
(2) The assimilation experiment (VAR); Initial condition is prepared by regional assimilation
system (WRF 3DVAR) without cyclone bogusing.(3) The cyclone bogusing experiment (BOG);
Model is run with T382L64 global model interpolated output initialized with bogusing without
assimilation. (4) In the forth experiment,(BOGVAR); the initial condition of the model is prepared
with both cyclone bogusing followed with WRF data assimilation .
The impact is demonstrated in terms of track error, central pressure, maximum sustained
wind speed etc. Results indicate a remarkable impact of cyclone bogusing on the initial condition.
All three cyclones can be located in the initial conditions (00 Z) of bogus (BOG and BOGVAR)
experiments which were otherwise absent in no-bogus (VAR and CNTL) experiments. The track
prediction is considerably improved in terms of direction of movement as well as cyclone location
and is close to observations in the BOGVAR experiments.
LAILA
0
100
200
300
400
500
600
700
800
900
0 6 12 18 24 30 36 42 48 54 60 66 72
Forecast time, hr
Tra
ck e
rro
r, k
m
BOGVAR
VAR
Fig.1. Tropical Cyclone 'LAILA' comparison of track errors for 'BOGVAR' and 'VAR'
experiments Results with bogus followed by assimilation has given significant reductions in track errors.
Figure 1. Shows comparison between BOGVAR and VAR experiments track errors during
forecast hours for tropical cyclone 'LAILA'. Here it is evident that bogusing (BOGVAR) has
improved track predictions by reducing track errors. The maximum reduction in track error is 76.81
% in 'LAILA', 87.30 % in 'GIRI' and 51.55 % in 'PHET' respectively in BOGVAR experiment in
comparison to VAR experiment. Maximum sustained wind speed and minimum central pressure
are more close to observations in BOGVAR in comparison to VAR for tropical cyclones 'LAILA'
and 'GIRI'. Whereas in case of 'PHET' the trend in pressure drop and increase in wind speed did not
show significant improvement.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 71
Numerical Simulation of Tropical Cyclones in Bay of Bengal
R. D. Kanase and P. S. Salvekar
Indian Institute of Tropical Meteorology, Pune-411008, India
[email protected] and [email protected]
In last decade total 29 cyclones were formed in BoB, 9 in pre-monsoon and 20 in post
monsoon. Out of which 7 cyclones were VSCS and 6 cyclones were SCS. It is desirable to have
accurate prediction of the track & landfall for effective implementation of the disaster management.
For this purpose meso-scale numerical models based on well defined dynamical and physical
processes can be used.
In this study, simulation of four cases of sever cyclonic storms [Laila (17-21 May 2010),
Aila(23-26 May 2009), Jal (4-8 Nov. 2010) and SCS (11-16 Dec. 2003)]is carried out using
Mesoscale Model WRF(Skamarock et.al. 2005-WRF Version 3, NCAR Technical Note) and
NCEP FNL reanalysis data with the combination of Cumulus scheme-BMJ, Planetary Boundary
Layer scheme-YSU and Microphysics scheme such as WSM-6 class microphysics. Three two way
interactive nested domains [60km, 20km and 6.6km ] and observed low pressure as the initial state,
model integration is performed to evaluate prediction of track and intensity in terms of Central Sea
Level Pressure (CSLP) & Maximum surface wind speed (MSW) of the storm. The errors in track
prediction are calculated in terms of vector displacement errors compared to the observed (IMD)
track of the storms.
Results
1. Pre-monsoon cyclones: Laila (17-21 May 2010) - A depression formed on 17
th May in BoB, moving in the northwest
direction, intensified into a SCS at 09:19-05-2010, crossed AP Coast near Bapatla between 1100-
1200 UTC of 20th
May 2010.Track of Laila cyclone is well simulated by WRF model (fig.1a) with
minimum/maximum track error is about 40km/160km (fig.1b).
Aila (23-26 May 2009) - A depression formed over BoB on 23rd
May 2009 moved in
northward direction, intensified into a SCS at 06:25-05-2009. It crossed West Bengal coast close to
east of Sagar Island between 0800 &0900UTC of 25th
May 2009. Track & intensity of Aila are very
well simulated. The track error ranges from 12 to 84 km (fig. 2a, 2b) upto 72 hrs of
integration and thereafter continuously increases.
2. Post-monsoon cyclones: Jal (4-8Nov. 2010)- A depression formed over BOB on 00UTC of 4
th Nov. 2010, moved in
north-west direction and intensified into SCS at 21:05-11-2010. It crossed north Tamilnadu & south
AP coast close to north Chennai around 16:07-11-2010. Initially upto 42hrs of integration, the track
error is around 30km, then it is upto 200km, till 96hrs.(fig. 3a1-3b).
SCS (11-16Dec. 2003) - A depression is formed over the southeast BoB on 12:11-12-2003,
moved in north-west direction and intensified into SCS at 12:14-12-2003 and crossed the coast near
Machilipatnam around mid-night on 15-12-2003. Upto 75hrs of integration the track error is
within 100km and afterwards it increases continuously (fig.3a2-3b).
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 72
Conclusions
These results clearly demonstrate that the WRF with model configuration CPs-BMJ, PBLs-
YSU, MPs-WSM6 is suitable for track prediction of severe cyclonic storms. Intensity of Aila is
very well simulated as compared with IMD whereas for other three cyclones intensity is over
estimated.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 73
Laila Aila Jal SCS-
03
I
M
D
CSLP(hP
a)
986 968 988 990
MSW
(m/s)
29 31 31 29
W
R
F
CSLP(hP
a)
972 962 978 976
MSW
(m/s)
45 36 46 44
Table: Intensity of cyclones
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 74
Tropical Cyclone Genesis Potential Parameter (GPP) and it’s application
over the North Indian Sea
S. D. Kotal and S. K. Bhattacharya
NWP Division,
India Meteorological Department
Mausam Bhavan, Lodi Road, New Delhi-110003
E-mail: [email protected]
An analysis of tropical cyclone genesis potential parameter (GPP) for the North Indian Sea
is carried out. The genesis potential parameter developed by Kotal et al. (2009) is computed based
on the product of four variables, namely: vorticity at 850 hPa, middle tropospheric relative
humidity, middle tropospheric instability and the inverse of vertical wind shear at all grid points
over the area. The GPP at a grid point is considered under the conditions that all the variables
vorticity, middle tropospheric relative humidity, middle tropospheric instability and the vertical
wind shear are greater than zero and it is taken as zero when any one of these variables is less or
equal to zero. The variables are computed using the European Centre for Medium Range Weather
Forecasts (ECMWF) model data. Forecast of the genesis parameter up to seven days is also
generated on real time using the ECMWF model output. Higher value of the GPP over a region
indicates higher potential of genesis over the region. Region with GPP value equal or greater than
30 is found to be high potential zone for cyclogenesis. The analysis of the parameter and its
effectiveness during cyclonic disturbances in 2010 affirm its usefulness as a predictive signal (4-5
days in advance) for cyclogenesis over the north Indian Sea.
Key Words: Tropical cyclone, Genesis potential parameter, Vorticity, Moisture variable,
instability and vertical wind shear.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 75
Track Prediction of North Indian Ocean Tropical Cyclones using ARW model
Krishna K. Osuri1*
, U. C. Mohanty1, A. Routray
2 and M. Mohapatra
3
1C A S, I I T, Hauz Khas, New Delhi – 110016
2 NCMRWF, Noida-201307
3 India Meteorological Department, Lodi Road, New Delhi.
A detailed evaluation of performance and systematic bias of Advanced Research Weather
Research and Forecasting (ARW) model in predicting the movement of tropical cyclones (TCs)
over the North Indian Ocean (NIO) is undertaken in the present study. There were 16 TCs formed
over the NIO during 2007–2010. These TCs are initialized at different times and have a total of 97
cases. Based on location of genesis, the TCs have been divided into the Arabian Sea (AS) and the
Bay of Bengal (BoB) cyclones. The BoB TCs have been further, divided into recurving, northward
moving, and westward moving TCs. In addition to usual forecast errors, systematic bias in zonal
and meridional direction, cross-track (CT) and along-track (AT) error components relative to
persistence track are also calculated to analyze the gain in skill of model forecast.
The overall skill of ARW model increases with forecast length in track prediction compared
to persistence track. The mean track error varies from 130 km to 350 km from 12 to 72 hour
forecast and the mean landfall position errors are 120, 90 and 52 km for 72, 48 and 24 hour forecast
for the NIO cyclone. The mean time errors (standard deviation) are -7 (15), -3 (13) and -5 (6) hours
for 72, 48 and 24 hour forecast. The mean initial vortex position errors are 78, 71, 68 and 88 km for
northeast/east recurving, northward moving, westward moving systems of the BoB and AS systems
respectively. ARW model exhibits negative skill for short-term (12 hour) forecast for all categories
of systems. The model has a tendency to over-predict east-ward movement of the TCs over the
NIO. ARW forecasts are in general slower as compared to actual speed of the systems and hence
behind to the observed position for all the forecast lengths. As a result, model yields delayed
landfall. Further, as the mean CT errors are less compared to AT errors i.e., ARW model errors are
elliptical in nature with its major axis along the track. As the CT error components are smaller, the
landfall position error is comparatively smaller. Further analysis indicates that the model shows
better performance for post-monsoon cyclones. The track prediction of severe cyclones is better as
compared to weak cyclones over the NIO.
Key words: North Indian Ocean, Landfalling tropical cyclones, Track errors, ARW model
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 76
On the Implementation and the ability of the Ensemble Prediction System for tropical
cyclone track and strike probability for North Indian Ocean
K. Naga Ratna
India Meteorological Department, Lodi Road, New Delhi -110 003.
Email: [email protected]
Implementation of Ensemble Prediction System (EPS) over North Indian Ocean and the
ability to predict the probability that a tropical cyclone will fall within a certain area is evaluated.
The software provided by the TIGGE, has been generalised and implemented for North Indian
Ocean (NIO) to produce Ensemble and Deterministic forecasts for the tropical cyclones and also
Strike Probability along the Indian coast and neighbouring countries.
Ensemble forecasts issued by the European Centre for Medium-Range Weather Forecasts
(ECMWF), National Center for Environmental Prediction (NCEP) and the Met Office (UKMET)
were evaluated for the cyclones JAL, LAILA, PHET and GIRI that formed during 2010, over North
Indian Ocean. The ECMWF model with horizontal resolution of 45km and vertical resolution of
62 levels produces forecast data twice a day for 50 member ensembles; while the UKMO model
having horizontal resolution of 90km and 38 vertical levels has data frequency 2 times a day for 24
member ensembles; NCEP model discretized with resolution of approximately 90km in horizontal
and 28 levels in vertical sends data four times daily for 20 members are available. In the North
Indian Ocean, the ensemble mean of ECMWF, UKMO, NCEP and ECMWF+UKMO+NCEP
(ALL) tracks and intensity are comparable in skill. It is revealed that the strike probability circles
of the ECMWF ensemble could capture the best track with a skill of 70% for 24-48 hours forecasts
and were over dispersive beyond 48hours. UKMET ensemble yielded improvements in the short
range. NCEP ensemble forecasts revealed that the tracks forecasts are better in the short range, the
tracks are found to be deviated more after 48 hours. Further evaluations were done for the track
forecasts and intensity forecasts for all the three model ensembles. The ability has been evaluated
interms of track and intensity for the ensemble means of the three models and for all 10 member, 20
member and 30 member means. The evaluation the EPS thus done will be presented.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 77
Ocean Atmospheric Coupled Model to Estimate Energy and Path of Cyclone
near the Coast
Ramkrishna Datta
Regional Meteorological Centre,
Alipore, Kolkata
E-mail: [email protected]
The tropical atmospheric phenomena like Cyclone, Typhoon, Hurricane etc. cause a violent
massive disturbance . The ‘EYE’ region of such phenomena can be regarded on the basis of
abstract idea of fluid dynamics . The said EYE can be imagined as the combination of fluid
dynamical sourceof strength +m and fluid dynamical sink of strength –m at a small distance apart.
. So the EYE can be assumed to constitute a fluid dynamical two dimensional doublet of finite
strength µ. This is here the object doublet. Now the seashore can be regarded as the real line x in
the two dimensional complex plane where as y is the imaginary axis lying on the sea
perpendicular to x . It is assumed that there are no flows of fluid across the real line x (seashore ).
Then the said object doublet can be placed on the sea at a perpendicular distance a from the real
line x in the complex plain (z=x+iy) making an angle 180 degree with the real axis x. Therefore the
image doublet will be at just opposite side of the real line x .
Here the fluid can be regarded as non viscous, incompressible fluid and it is moving with certain
velocity U at infinity in the direction of x axis. The motion of the fluid is wholly two dimensional
in the complex plane z . Now the complex potential w on the whole system which consists of
object doublet, image doublet and the stream velocity U parallel to x axis is given by
w = + - Uz (1)
or � � = q = � U + � (2)
To determine the pressure at any point on the wall we use the Bernoulli’s equation
+ q² = C (constant) (3)
We get
+ m
Y-a
xis
Seashore as the axis of X
+ m
- m
- m
a
U
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 78
= using P = π, q = U when z = ∞
For any point on the sea shore z = x
= (2µ (a²-x²) + U(a²+x²)² ) (4)
=0 gives x=0 , ±a√3
We see that When x= ±a√3
> 0 if µ > 4a²U (5)
And < 0 if µ < 4a²U (6)
Equations (5)and (6) describe analytically the strength of the tropical systems.
If µ > 4a²U . Pressure minimum at x= ±a√3 on the sea shore. Such system coupled with sea
shore minima strike the shore vigorously. This explain Orissa super cyclone 1999.
If µ < 4a²U we get sea shore maximum at x= ±a√3 . Such system coupled with zonal stream
moves towards right in the Bay of Bengal and gulf of Mexico. Eventually which explain
analytically that the cyclones of the Bay of Bengal are stronger than that of at Atlantic ocean. The
same observation had been written by Sir John Eliot, the first director general of observatory of
India Meteorological Department in 1889.
Key words :- Fluid dynamical Source, Sink, Object doublet, image doublet and complex potential.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 79
Track, Intensity and few Dynamical Aspects of ‘Aila’ as Simulated by Operational NWP
Model of the IAF
Wg Cdr TP Srivastava1 and Wg Cdr Anil Devrani
2
1Met Faculty, Air Force Administrative College, Red Fields, Coimbatore.
2Air Force Centre For Numerical Weather Prediction, Subroto Park, New Delhi.
1. Introduction Operational NWP Model of the IAF uses ARW core of WRF (Version 3.1.1) in a two way
nested configuration at resolutions of 18 and 6 km as per the domains and schemes shown in Fig.1.
Initial and boundary conditions of 0000 UTC and 1200 UTC from NCEP GFS are used for model
integration of 75 hours. The products of these two operational runs are made available to the field
forecasters daily by 1600h and 0400h.
2. History of ‘AILA’ Under the influence of an upper air cyclonic circulation, a low pressure area formed over
the southeast Bay of Bengal on 22nd May morning. It subsequently concentrated into a depression
and lay centered at 1130 hours IST of 23 May 09 near 16.5º N/88.0º E about 600 km south of Sagar
Island. The depression moved northwards, intensified into a deep depression and lay centred at
0830 hours IST of 24 May 09 near 18.0ºN/ 88.5ºE. It further intensified into a cyclonic storm
‘AILA’ at 1730 hours IST of 24th May and lay centred near 18.5ºN/ 88.5ºE. It continued to move
northwards and intensified into a severe cyclonic storm at 1130 hours IST of 25 May 09 and lay
centred over northwest Bay of Bengal near 21.5ºN / 88.0ºE close to Sagar Island. The system
crossed West Bengal coast close to the east of Sagar Island between 1330 to 1430 hours IST as a
severe cyclonic storm with wind speed of 100 to 110 kmph. The lowest estimated central pressure
was about 967 hPa at the time of landfall. After the landfall, the system continued to move in a
northerly direction, gradually weakened into a cyclonic storm and lay centred at 2030 hours IST of
25 May 09 over Gangetic West Bengal, close to Kolkata. It continued its northerly movement,
weakened into a deep depression and lay centred at 0830 hours IST of 26 May 09 over Sub-
Himalayan west Bengal & Sikkim, close to Malda. It weakened into a depression and lay centred at
1130 hours IST of 26 May 09 over the same region close to Bagdogra. By 1430 hours IST of 26
May 09, it weakened further and was seen as a well marked low pressure area over Sub-Himalayan
West Bengal and became less marked by 27 May 09.
3. Prediction by IAF Model To understand the efficacy of the IAF model towards enhancing advance warning of the
impending adverse weather, track, intensity, few dynamical products, 3 hourly rainfall pattern and
composite radar reflectivity generated in the finer nested domain of 6km by using the 0000UTC
initial conditions of 23 May 09 (D-2) and 24 May 09 (D-1), valid for the period from 24 May 09 /
0600 UTC to 26 May 09 / 0600 UTC, have been discussed in the study.
(a) Predicted Track and Intensity of ‘AILA’. Formation and intensification of the system
was captured reasonably well on D-2. As shown in the Fig.2 (a&b), a general northerly track
was predicted by the initial conditions of D-2 and D-1. Deviations in prediction from the
observed track were more from the initial conditions of D-2 than that of D-1. Landfall was
predicted 187km East at three hours later in comparison to the actual location and time on D-2.
D-1 had relatively better prediction as the error reduced to nil though the landfall time was three
hours early. The isobaric patterns confirmed well with the actual pattern throughout the
predicted period but the forecast values of central pressure were 18 – 20 hPa higher, more so
when the system intensified into a severe cyclonic storm.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 80
(b) Vorticity at 850 hPa and Divergence at 200 hPa. After the onset of South West Monsoon
over Andaman Sea and adjoining south Bay of Bengal by 20th May 2009, increase in the
southerly surge resulted in increase in relative vorticity over the South East Bay of Bengal. It
led to the formation of a low pressure area over the region on 22 May 09. Due to presence of
high magnitude of the low level relative vorticity that was commensurate with the values of
upper level divergence around the centre of the system, intensification of the system continued.
By 1730 hours IST of 24th May it intensified into a cyclonic storm ‘AILA’ and lay centred at
18.50oN / 88.50
oE. Juxtaposition of higher values of low level convergence and upper level
divergence maintained the strength of the system before it started weakening after 0000 UTC of
26 May 09. The predictions of Vorticity at 850 hPa and Divergence at 200 hPa on D-2 were
relatively lower as the model did not intensify the system into a severe cyclonic storm and it
weakened the system into a deep depression much before the predicted land fall. However,
these values were more realistically predicted by the model on D-1.
(c) Vertical Velocity at 500 hPa. Low level convergence if overlaid by upper level divergence
will lead to higher positive values of upward vertical velocity at the level of no-divergence
(LND). Over the Indian region 500hPa is the representative of the LND. Higher positive values
of the vertical velocity at 500hPa predicted by the model, match well with the convective cloud
patterns as shown by the imageries of Kalpana-I, of similar times.
(d) Moisture Convergence at 850 hPa. Moisture advection is horizontal transport of moisture,
which plays a very important role in the development of precipitation. If little moisture is
available, it is unlikely that precipitation will occur. However, if any system is supplied with an
abundance of moisture, there is an increased likelihood that heavy precipitation will be realized.
The maximum moisture convergence as predicted on D-2 and D-1 match well with the areas of
precipitation. Three hourly rainfall patterns as shown by the TRMM 3B42 V6 match well with
the areas of high magnitude of moisture convergence. This rich moisture supply was enough for
showers and thunderstorms to develop as indicated by the radar echoes of Kolkata DWR of
similar times. It is to be noted that the precipitation was located in the region where the
strongest moisture convergence was predicted.
(e) Total Cloud Cover. This product is still in experimental mode. Modifications have been
done for the display of predicted clouds by suppressing or enhancing the values of low, medium
and high clouds to get the best possible realistic picture by comparing it with the Kalpana – I
image of the same time, in the hind cast mode. To make the product more meaningful 6 hourly
predicted rainfall patterns has been superimposed over the predicted total cloud cover. In the
case discussed here, this product matched well with the corresponding actual Kalpana-I IR
images of the similar times. Vertical Velocity at 500hPa, Moisture Convergence at 850hPa,
Total Cloud Cover with 6hourly precipitation predicted for 25 May 09 / 0600 UTC along with
IR imagery of Kalpana-I and 6 hourly rainfall given by TRMM 3B46 V6 valid for 25 May 09 /
0600UTC are shown in Fig. 3(a–h).
(f) Hourly Rainfall Pattern. TRMM 3B42V6 products were used for qualitative validation of
the model predicted rainfall. The three hourly pattern of rainfall predicted by the model on D-2
and D-1 matched reasonably well with the satellite derived rainfall patterns shown by TRMM
product. The model over-predicted the rainfall to a certain degree in comparison to the rainfall
shown by TRMM 3B42V6. A snapshot of comparison of 3 hourly rainfall at 0600UTC on 25
May 09 are shown in Fig. 4(a-c)
(g) Maximum Radar Reflectivity. Maximum Radar Reflectivity was also simulated
reasonably well. The predicted patterns of D-2 and D-1 matched well with the actual Maximum
Reflectivity shown by the DWR of Kolkata. Comparison of this product was done both with
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 81
and without cumulus parameterisation scheme in the nested domain of 6km. It was seen that the
patterns produced without employing cumulus parameterisation scheme were more realistic.
The same is shown for 0600 UTC on 25 May 09 in Fig. 5(a-e).
4. Conclusion Predictions pertaining to track, intensity and rainfall etc. of ’AILA’ from the finer domain
of 6 km of the Operational NWP Model of the IAF had provided sufficient warning time to the
users in the affected areas. The forecasts of D-1 were relatively better and more realistic in
comparison to the one generated on D-2, with advance warning of more than 20hours and 40hours
for its landfall, respectively. As demonstrated by the model generated patterns of maximum radar
reflectivity, cumulus parameterisation in the finer domain of 6km can be avoided to improve the
forecast and economise the computational cost.
Fig.1:IAF Model : WRF Version 3.1.1 (ARW)
Resolution : 18 Km, 6 Km (Double Nested) Physics
Options : Thompson, Thompson Cumulus Schemes :
Grell, Grell PBL : MYJ TKE, YSU
Fig.2(a): Track & Intensity : Initial Condition of
23 May 09/0000UTC
Fig2(b): Track & Intensity : Initial
Condition of 24 May 09/0000UTC
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 82
PREDICTIONS BASED ON INITIAL CONDITIONS OF 23 MAY 09 / 00Z
Fig.3(a):Vertical
Velocity at 500hPa
valid for 25 May
09 / 0600UTC:
Fig.3(b):Moisture
Convergence at 850 hPa
valid for 25 May 09 /
0600UTC:
Fig.3(c):Total Cloud
Cover & 6hrly Pptn
valid for 25 May 09 /
0600UTC:
Fig.3(d): TRMM 3B42V6
6h Rainfall for
0600Z / 25May09
PREDICTIONS BASED ON INITIAL CONDITIONS OF 24 MAY 09 / 00Z
Fig.3(e):Vertical
Velocity at 500hPa
valid for 25 May
09 / 0600UTC:
Fig.3(f):Moisture
Convergence at 850 hPa
valid for 25 May 09 /
0600UTC:
Fig.3(g):Total Cloud
Cover & 6hrly Pptn
valid for 25 May 09 /
0600UTC:
Fig.3(h):Kalpana-I, IR
Image valid at 25 May 09 /
0600UTC:
Fig.4(a): TRMM 3B42V6 3h Rainfall
for 0600Z / 25May09
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 83
Fig.4(b): Model Predicted 3h Rainfall
0600Z / 25May09
IC:23May09 / 0000UTC
Fig.4(c): Model Predicted 3h Rainfall
0600Z / 25May09
IC:24May09 / 0000UTC
Fig.5(a): Max Reflectivity by Kolkata
DWR at
0614Z / 25May09
Fig.5(b): Model Predicted
Max Reflectivity 0600Z /
25May09
IC:23May09 / 0000UTC (with
CP)
Fig.5(c): Model Predicted
Max Reflectivity 0600Z /
25May09
IC:24May09 / 0000UTC
(with CP)
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 84
Fig.5(d): Model Predicted
Max
Reflectivity 0600Z /
25May09
IC:23May09 / 0000UTC
(without CP)
Fig.5(e): Model Predicted
Max Reflectivity 0600Z /
25May09
IC:24May09 / 0000UTC
(without CP)
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 85
Analysis of Barotrophic Energetics of Tropical Cyclone Khai-muk
S.Balachandran
India Meteorological Department
Regional Meteorological Centre, Chennai
email ID: [email protected]
Tropical cyclones (TC) are intense atmospheric vortices characterized by extreme winds,
torrential rain, and destructive storm surges. When a major hurricane makes landfall one or more of
these processes can cause immense property damage and loss of life. Considerable progress has
been made in recent decades unlocking the physical and dynamical mechanisms by which
hurricanes form, and by which they change their structure and intensity. Climatologically, about
80% of all the tropical cyclones on the globe form near or within the ITCZ . This motivates
searching for mechanisms that favor tropical cyclogenesis within the context of ITCZ dynamics.
Studying dynamical mechanisms of perturbation growth in tropical cyclones is important from a
perspective of designing ensemble prediction system and adaptive observations for tropical
cyclones. It is widely accepted that sea surface temperatures (SSTs) and vertical shear are primary
factors controlling the genesis and development of tropical cyclones (TCs).
In the present study, the barotrophic energy conversion processes associated with TC Khai-
muk which formed over the North Indian Ocean and affected the eastern coastal region of southern
peninsular India during 13-16 November 2008 are studied.The special feature of this TC was that it
did not retain its TC intensity for even a day. It intensified into a Cyclonic Storm (CS) at 12 UTC
of 14th
but weakened into Deep Depression (DD) by 15th
/06 UTC over the sea itself and crossed
coast as a DD. The analysis of the instability of the background flow along with change of eddy
kinetic energy and barotropic conversion are presented during life cycle of TC Khai-muk are
presented. .
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 86
Performance evaluation of spectrum of cyclones over North Indian Ocean
using RAMS model
Ancy Thomas, Basanta kumar Samala and Akshara Kaginalkar
*Centre for Development of Advanced Computing,
Pune Unversity Campus, Pune - 411007, India
In the present study, the performance of non-hydrostatic Regional Atmospheric Model
(RAMS) model in simulating the tropical cyclones of different intensities, formed during pre and
post monsoon are analyzed. The cyclones Orissa, Sidr, Mala, h04B, 01A, and Agni occured over
North Indian Ocean are studied. This study reveals the ability of model in down-scaling the
cyclone track simulation, cyclone intensity in terms of lowest sea level pressure, winds at 850 hpa
and 200 hpa and thermodynamical features associated with the development of cyclones such as
vertical wind shear, mid tropospheric humidity and sea surface temperature. It is found that the
model could simulate the track of all cyclones reasonably well except for h04B. The track error
increases with the simulation time. The model overestimates the lowest mean sea level pressure in
comparison with observations. Model is able to represent the low level circulation and upper air
divergence of wind during all the cyclone cases . The model simulations are in agreement with the
criteria of low vertical wind shear required for the cyclogenesis and relative humidity above 60% at
700 hpa and 500 hpa satisfies the condition of high moisture required for the deep convection
during the intensification of cyclones.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 87
An observational and modeling study of the tropical cyclone Phet.
2Jagabandhu Panda*,
1R. K. Giri** and
1, 3Harvir Singh***
1Satellite Meteorology Division, India Meteorological Department, New Delhi, India
2School of Physical and Mathematical Sciences, Nanyang Technological University, Singapore 3HCL Info-systems, National Centre for Medium Range Weather Forecasting (NCMRWF),
NOIDA, U. P., India
*[email protected], ** [email protected], ***[email protected]
The accuracy of numerical weather prediction (NWP) depends on the quality of forecast
model and initial conditions. In this study, recent and advanced Weather Research and Forecasting
(WRF) mesoscale modeling system (ARW core) is used with a combination of Yonsei University
PBL scheme, Dudhia short wave scheme, RRTM long wave scheme WSM 3-class microphysics
and unified Noah land-surface model in order to study the characteristic features of the tropical
cyclone PHET that occurred over the Arabian sea (in 2010) and affected the coastal areas of several
countries. A comprehensive sensitivity analysis is carried out with respect to various cumulus
convective parameterizations including Grell-Devenyi ensemble scheme, Kain-Fritsch scheme,
Betts-Miller-Janjic scheme and Grell-3D scheme for the prediction of track and intensity of the
cyclonic storm PHET. The initial and boundary conditions for the simulations are derived from
global operational analysis and forecast products of the National Center for Environmental
Prediction-Global Forecast System (NCEP-GFS) available at 1olon/lat resolution in these model
simulations. However, the model initial conditions are further modified using KALPANA-1
atmospheric motion vectors and OCEANSAT-2 surface winds through a three dimensional
variational technique within ARW modeling system (WRF-3DVAR). The simulated results of
extreme weather parameters including the rainfall, wind field, track and intensity of the cyclone are
critically analysed comparing with those observed/predicted by India Meteorological Department
(IMD), New Delhi. Further, the model simulated results are qualitatively examined alongside the
satellite observations from METEOSAT, MODIS and KALPANA-1 in order to understand the
model performance as when compared to the observations. Several parameters derived from
satellite observations are also analysed including outgoing long-wave radiation (OLR), quantitative
precipitation estimate (QPE; rainfall), sea surface temperature (SST), relative vorticity, upper
tropospheric humidity (UTH) and the track of the cyclone (figure shown below) in order to
understand the genesis of the storm. The observational analysis reveals relatively higher values of
SST (~26.5oC), relative vorticity and UTH (90-100%) during the life span of PHET. A higher
negative correlation (~ -0.96) between the OLR and QPE corresponds to the observed maximum
value of QPE when there is minimum OLR and the cyclone reaches its maximum intensity after a
day of attending this state. The model also simulated the extreme weather parameters reasonably
well and the performance is slightly improved further through the satellite data assimilation using
surface winds and atmospheric motion vectors.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 88
Figure: Track of cyclone PHET
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 89
Large-Scale Characteristics of Rapidly Intensifying Tropical Cyclones over the Bay of Bengal
and a Rapid Intensification (RI) Index
S. D. Kotal and S. K. Roy Bhowmik
India Meteorological Department, NWP Division, New Delhi-110003
E-mail: [email protected]
A rapid intensification index (RII) is developed for tropical cyclones over the Bay of
Bengal. The RII uses large-scale characteristics of tropical cyclones to estimate the probability of
rapid intensification (RI) over the subsequent 24-h. The RI is defined as an increase of intensity 30
kt (15.4 ms-1
) during 24-h, which represents approximately the 93rd percentile of 24-h intensity
changes of tropical cyclones that developed over the Bay of Bengal during 1981-2010. It is found
that 32% of all very severe cyclonic storms (VSCS) and all super cyclonic storms (SUCS)
underwent RI phase at least once during their lifetime. No cyclonic storm (CS) and severe cyclonic
storm (SCS) underwent RI phase. Various large-scale variables associated with the RI cases are
compared to those of non-RI cases. These comparisons show that the RI cases generally occur at
higher latitude and are intensifying at a faster rate during the previous 12-h than the non-RI cases.
The statistical analysis also shows that the RI cases are embedded in regions where the upper-level
divergence, lower-level relative vorticity and relative humidity are more and vertical winds shear is
weak. Finally, the initial wind speed of RI cases is higher and tends to move with a faster
translational speed than the non-RI cases. The RII technique is developed by combining threshold
(index) values of the eight variables for which statistically significant differences are found between
the RI and non-RI cases. The probability of RI is found to be increases from 0% to 100% when the
total number of indices satisfied increases from zero to eight.
Key words: Tropical cyclone, Rapid intensification, Probability, Vorticity, Divergence, Vertical
wind shear, Bay of Bengal.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 90
Development of the Lagrangian Advection model for prediction of tropical cyclone track over
the Indian Ocean
Sanjeev Kumar Singh, C. M. Kishtawal, Neeru Jaiswal, and P. K. Pal
Atmospheric Sciences Division, Atmospheric & Oceanic Sciences Group
Space Applications Centre (ISRO), Ahmedabad-380015, India
Email: [email protected]
A new model has been developed for track prediction of Indian Ocean cyclones. The Model
utilizes environmental steering flow using the forecasts from a high resolution global model and the
effect due to earth’s rotation (the Beta-effect) to determine the future movement of cyclone. The
model is based on the dynamical frame work and the time for running is very less. A new approach
based on vertical profile of potential vorticity (PV) is used to determine the weights for different
vertical levels for computation of mean steering flow. The effect of environmental flow over the
cyclone track is also examined by removing the existing cyclonic feature from the mean wind
fields. For this, a new approach based on vortex pattern matching has been used to identify the
cyclone vortex and to remove it from mean wind fields.
The data used in the present work for the computation of cyclone trajectories are the high
resolution 0.5°×0.5° forecasted atmospheric wind fields and temperature from Global Forecast
System (GFS), which is the global NWP computer model run by NOAA. The wind fields and
temperature are taken for the North Indian Ocean domain at every 6-hour interval of 26 pressure
levels (10 mb-1000 mb) for 0 to 72 hours forecast. The Joint Typhoon Warning Center (JTWC)
best track analysis data has been used for defining the initial position of cyclonic vortex.
The present Lagrangian Advection model has been used to forecast the 6-hourly track of six
tropical cyclones (viz., Nargis, Khai-Muk, Aila, Phyan, Laila and Jal) which were formed in the
North Indian basin during the period 2008 to 2010. The maximum PV based approach has been
used for determining the optimal steering levels which is adapted from the study by Hoover and
Morgan (2006). The important step in this model is to form a synthetic cyclone (Jaiswal and
Kishtawal, 2011) which was used to remove the cyclonic features from the environmental flow.
The forecast errors for all the cyclone cases have been computed with respect to JTWC analysis
best track.
To limit the size of the presentation, the forecasted tracks of one of the above cyclones,
“Nargis”, are shown in Fig.1 (a-f) respectively. The predicted mean track errors of the Lagrangian
Advection model w.r.t. JTWC analysis best track for six cyclones for 12-72 hours are shown in
Fig.2.
References: Hoover, B.T. and Morgan, M.C. (2006) Effects of cumulus parameterization on tropical cyclone
potential vorticity structure and steering flow. Preprints of the 27th AMS Conference on
Hurricanes and Tropical Meteorology, April 23-28, 2006. Monterey, CA, paper 8B.5.
Jaiswal, Neeru and Kishtawal, C.M. (2011) Prediction of tropical cyclogenesis using scatterometer
data. TGRS, 2153862.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 91
Fig.2: Mean track error of the all six cyclones
Fig.1: Six forecasts ((a) 72 hours prediction from 28-Apr-00Z, (b) 72 hours prediction from 29-Apr-00Z, (c) 72 hours
prediction from 30-Apr-00Z, (d) 60 hours prediction from 01-May-00Z, (e) 36 hours prediction from 02-May-00Z
and (f) 12 hours prediction from 03-May-00Z) generated for NARGIS Cyclone by the Lagrangian Advection model.
a: 72-H Forecast from 28-Apr-00Z b: 72-H Forecast from 29-Apr-00Z c: 72-H Forecast from 30-Apr-00Z
e: 36-H Forecast from 02-May-00Z
d: 60-H Forecast from 01-May-00Z f: 12-H Forecast from 03-May-00Z
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 92
Extended Range Forecast of Tropical Cyclone Genesis Based on Coupled Model Outputs
D. R. Pattanaik*, M. Mohapatra, Y. V. Rama Rao and Ajit Tyagi
India Meteorological Department, New Delhi
Email-* [email protected]/[email protected]
Over the North Indian Ocean the months of October-November are known to produce
cyclones of severe intensity in the Bay of Bengal, which cause damages to life and property over
many countries surrounding the Bay of Bengal. The strong winds, heavy rains and large storm
surges associated with tropical cyclones are the factors that eventually lead to loss of life and
property. Rains (sometimes even more than 30 cm/24 hrs) associated with cyclones are another
source of damage.
There are two cyclones formed during the post monsoon season of 2010 (Fig. 1). The first
one “Giri” initially seen as a low pressure area on 19th October over the east central Bay of Bengal
and neighbourhood, intensified into a tropical cyclone at 0600 UTC of 21st and became a very
severe cyclonic storm (VSCS) at 0300 UTC of 22nd, which crossed the Myanmar coast on 22nd.
The second cyclonic storm of the season “Jal” formed in the Bay of Bengal was first observed as a
low pressure area over the south Andaman Sea and neighbourhood on 2nd November, which
intensified into severe cyclonic storm (SCS) at 2100 UTC of 5th. It crossed north Tamil Nadu-
south Andhra Pradesh coasts, close to north of Chennai between 1700 & 1800 hrs UTC of 7th
November and caused lot of damage in Tamilnadu and south Andhra Pradesh coast associated with
not only strong wind but also due to heavy rainfall associated with the cyclone.
With the improvement in numerical model and use of wide ranges of non conventional data
in the assimilation system of the model there has been considerable improvement in the forecast
skill of tropical cyclones particularly in the short range up to 72 hr. However, the forecasting of
genesis of tropical cyclone and associated rainfall in the extended range time scale (about 10 days
to 2 weeks in advance) is very useful in many respects. In the present study an attempt is made to
forecast the genesis of tropical cyclone and also the associated rainfall activity in the extended
range time scale over the north Indian Ocean for the cyclones “Giri” and “Jal” using the multi-
model ensemble techniques.
The multi-model extended range forecasts are prepared based on the coupled model outputs
from ECMWF and NECP. The outputs from these two models are used and the multi-model
ensemble forecasts are generated on every Friday with forecast anomaly for week 1 (Monday to
Sunday) and week 2 (subsequent Monday to Sunday). The low level relative vorticity, low level
convergence, wind shear and the rainfall forecasts are analysed to consider the genesis of tropical
cyclones. The operational forecast for days 05-11 of weekly mean wind from NCEP CFS and
ECMWF coupled models based on 14th
Oct, 2010 initial condition (Figs. 2a & 2b) indicates
cyclonic circulation at low level over the central Bay of Bengal during the period from 18-24
October associated with the severe cyclone “Giri”. The centre of cyclonic circulation in case of
CFS forecast is (Fig. 2a) closer to observed location of the system when compared with the
ECMWF forecast (Fig. 2b). The genesis of the cyclone “Jal” was very much captured in both the
coupled models even in the forecast valid for 12-18 days based on the initial condition of 21
October, 2010 as indicated by cyclonic circulation over the Tamil Nadu coast during 01-07
November (Fig. 3a & 3b). The MME forecast valid for 01-07 November based on 28 Oct and 21
Oct initial conditions (with forecast period of days 05-11 and days 12-18 respectively) also clearly
indicated large positive rainfall anomalies over the Tamil Nadu coast and adjoining coastal Andhra
Pradesh region (Fig. 3c & 3d) like that of observed rainfall anomalies.
Thus, the extended range forecast indicates very well the genesis and also the associated
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 93
rainfall distribution due to the tropical cyclones of post monsoon season of 2010. The other
dynamical parameters like the low-level vorticity, wind shear, humidity etc are also analysed in the
model forecast fields to understand the genesis of both the severe cyclones “Giri” and “Jal”.
.
Fig. 1 : Cyclonic disturbances of post monsoon season from October-December, 2010. The
dark black lines indicate two severe cyclones “Jal” and “Giri”.
Fig. 2 : Forecast 850 hPa weekly mean wind during the cyclone “Giri” valid for days 05-11
(18-24 Oct 2010) based on initial condition 14 Oct. (a) based on NCEP CFS and (b)
based on ECMWF coupled models.
(a) NCEP CFS 850 hPa forecast mean wind (kts)
Day 12-18 based on 21 Oct, valid for 01-07 Nov 2010
(b) ECMWF 850 hPa forecast mean wind (kts)
Day 12-18 based on 21 Oct, valid for 01-07 Nov 2010
(b) ECMWF 850 hPa forecast mean wind (kts)
Day 05-11 based on 14 Oct, valid for 18-24 Oct 2010
(a) NCEP CFS 850 hPa forecast mean wind (kts)
Day 05-11 based on 14 Oct, valid for 18-24 Oct 2010
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 94
Fig. 3 : Forecast 850 hPa weekly mean wind during the cyclone “Jal” valid for days 12-18
(01-07 Nov 2010) based on initial condition 28 Oct from (a) based on NCEP CFS and
(b) based on ECMWF coupled models. The corresponding MME forecast rainfall
anomalies valid for 01-07 Nov, based on (c) 28 Oct and (d) based on 21 Oct initial
conditions.
(c) MME forecast rainfall anomaly (mm/day)
Valid for days 05-11 (01-07 Nov, 2011), IC=28 Oct
(d) MME forecast rainfall anomaly (mm/day)
Valid for days 12-18 (01-07 Nov, 2011), IC=21 Oct
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 95
Impact of Resolution and Data Assimilation on the prediction of the cyclone “JAL” over Bay
of Bengal using WRF(NMM) and grid point statistical interpolation scheme.
K. Naga Ratna
India Meteorological Department,
Lodi Road, New Delhi – 110 003.
Email: [email protected]
In this study, performance of WRF (NMM) model and the regional Grid-point Statistical
Interpolation(GSI) data assimilation scheme used to simulate the cyclone “JAL” that formed over
Bay of Bengal during postmonsoon season 2010, has been evaluated. A severe cyclonic storm,
JAL (4-8 November 2010) developed over the Bay of Bengal from the remnants of a depression
which moved from northwest Pacific Ocean to the Bay of Bengal across southern Thailand. It
moved westnorthwestwards and intensified upto severe cyclonic storm on 6 November. However
as the severe cyclonic storm, JAL moved to the southwest Bay of Bengal closer to India coast and
weakened gradually into a deep depression and crossed north Tamilnadu – south Andhra Pradesh
coast on 7 November.
WRF(NMM) version 3.2 model was integrated with regional GSI data assimilation scheme
for 72 hours from 4 -7 November 2010 at 9km and 3km horizontal resolutions and 38 levels in
vertical. The model could capture the direction of the movement and the landfall of JAL cyclone
which was predicted 48 hours before real-time with landfall forecast error of 56km and time error
of 2 hours delay. The impact of the conventional datasets, as AWS along the east coast, Ship data
of SagarKanya data during its cruise over the Bay of Bengal for the cyclone period, SagarPurvi and
SagarPaschimi stationed near the coast have been assimilated. The impact of the datasets for the
simulations of the cyclone JAL and performance of the model has been evaluated. The cyclone
track and intensity were evaluated for both the horizontal resolutions 9km and 3km. The track
performance with 9km yielded improvement with optimized track error of 50km during 24 to 48
hour period. The intensity of the cyclone as simulated with the 3km resolution model is more
compared to the 9km resolution. Performance parameters thus computed for the validation exercise
are (a) Direct Position Errors (b) Zonal (latitudinal biases DY) (c) Meridional biases (longitudinal
biases DX) and evaluated against the CLIPER model.
In Extension to above study, Hurricane Weather Research Forecast Model (HWRF) adapted
from NCEP, USA has been implemented at India Meteorological Department, New Delhi. WRF
(NMM) being the base of the HWRF model, has the similar data assimilation, regional data
assimilation system. The data assimilated in WRF (NMM) and regional GSI system for the
cyclone JAL has been used on experimental basis for case study in data assimilation with HWRF
model. The experiments done for the cyclone JAL with data assimilation using HWRF model are
compared with above WRF (NMM) model results.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 96
Study of Jal Cyclone Track Using WRF Cumulus Parameter Schemes
M. Venkatrami Reddy, S. Balaji Kumar, S. B. Surendra Prasad and K. Krishna Reddy
Semi-arid-zonal Atmospheric Research Centre (SARC)
Department of Physics, Yogi Vemana University, Kadapa, Andhra Pradesh
Tropical cyclones that form over the Bay of Bengal and Arabian Sea during pre-monsoon
(April-May), early monsoon (June), late and post monsoon (September-November) cause
vulnerable damage to lives and property over the coastal regions of India with strong winds,
heavy rain and tidal wave. Though the general behavior of the movement of the tropical cyclones
(TC) is known, it is desirable to have timely and reasonably accurate prediction of the tracks and
their intensities. The numerical models based on fundamental dynamics and well-defined physical
processes provide a useful tool for understanding and predicting tropical Cyclones. For accurate
forecast of TC, it is essential that numerical models must incorporate realistic representation of
important physical and dynamical processes as they play crucial role in determining genesis,
intensification and movement. In the present study, numerical simulation experiments on severe
cyclone "JAL" is formed. For “JAL” track prediction, a fully compressible, non-hydrostatic
Advanced Research Weather Research and Forecasting (ARW-WRF) model with Arakawa C-grid
is used. The advanced research WRF model was run at grid spacing of 27 km, 9 km and 3 km. The
cyclone track study is done with National Center for environmental prediction (NCEP), final
analysis fields (NCEP FNL) or the reanalysis data with 1.0 x 1.0 degree grid resolution used as
initial and lateral boundary conditions for the WRF model. In the JAL cyclone track prediction,
WRF modeling was performed by changing cumulus schemes such as Kain Fritsch (KF), Betts-
Miller-Janjic (BMJ), Grell-Devenyi (GD) and New Grell (NG) without changing microphysical
properties, PBL, and Radiation Schemes. The track observed with Kain Fritsch (KF) scheme is
well compared temporally and spatially with Indian Meteorological Department(IMD) observed
track and the remaining Betts-Miller-Janjic (BMJ), Grell-Devenyi (GD) and New Grell (NG)
schemes too suitable with IMD observed track only spatially. The cyclone centre pressure,
maximum cyclone surface wind speed obtained from the model are well compared with the IMD
data. The variation of pressure, temperature and humidity parameters from the Automatic Weather
Station at Yogi Vemana University, kadapa (14.47°N; 78.82°E), a semi arid region of India, during
the cyclone landfall was analyzed and compared with the modeled parameters. The results are in
reasonable in good agreement.
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 97
Impact of data assimilation system for simulation of tropical cyclones over Bay of Bengal with
WRF-NMM modeling system
Sujata Pattanayak and U C Mohanty
Centre for Atmospheric Sciences,
Indian Institute of Technology, Delhi
Hauz Khas, New Delhi-110016
In recent years the need for higher quality and more versatile data assimilation techniques
has becomes widely accepted. Additional demands on data assimilation techniques have resulted
due to several different reasons. First, the increased density, frequency and quality of data network
with increased observing systems (land as well as remote sensing platforms) needs the
improvement in data assimilation systems. Second, the quality and quantity of numerical weather
prediction (NWP) models necessitate appropriate and more realistic representation of initial state of
the atmosphere using better data assimilation techniques.
The study represents the impact of observational datasets for simulation of tropical cyclones
over Bay of Bengal with Non-hydrostatic Mesoscale Model (NMM) core of Weather Research and
Forecasting (WRF) system. Though WRF 3-dimensional variational data assimilation system
(WRF-Var) provides improved initial conditions to Advanced Research WRF (ARW), a unified
WRF-Var utility has been developed to be used by the WRF-NMM core, as well. The upgraded
code has been successfully tested and implemented to simulate three recent very severe cyclonic
storms i.e. the pre-monsoon cyclones Nargis (27 April to 03 May, 2008) & Aila (23 to 26 May,
2009) and the post-monsoon cyclone Jal (04 to 07 November, 2010) developed over Bay of
Bengal. A total of 24 cases (8 different initial conditions for Nargis starting from 28April 2008; 7
different initial conditions for Aila starting from 22 May 2009; 9 different initial conditions for Jal
starting from 03 November 2010) at every 00 and 12 hr are selected for an enhanced assessment on
the performance of the WRF-NMM modeling system. For this purpose, two sets of numerical
experiments are carried out with the meso-scale model WRF-NMM. In the first experiment, i.e. in
the control simulation (CNTL), the model has been integrated up to 96 hours in a single domain
with the horizontal resolution of 9 km along with 51 levels up to a height of 30 km in the vertical
for all the cases. The initial and lateral boundary conditions to a limited area model are usually
provided from the large scale analysis and forecasts available at different NWP centers in the
world. The NCEP Global Forecast System (GFS) analyses and forecasts (1º x 1º horizontal
resolution) are used to provide the initial and lateral boundary conditions respectively.
Again, in order to improve the initial analysis fields for the model integration, an attempt
has been made to initialize WRF-NMM model with WRF-VAR system. Hence, in the second
experiment, i.e. in the data assimilation (DA), the impact of the observational data sets has been
investigated by incorporating the available conventional and non-conventional data sets over Indian
region. Hence, for all the above mentioned cases (Nargis, Aila and Jal), the model is integrated with
corresponding different initial conditions each, as described above with the improved initial
conditions through WRF-Var system.
The improvement in model integration is verified statistically and analytically. The vector
displacement errors (VDEs) in track forecast are calculated with respect to the observed track
provided by the India Meteorological Department (IMD). The mean improvements in VDEs of
33%, 15%, 10%, 14% and 15% are seen (Fig 1) at 00hr, 24hr, 48hr, 72hr and 96hr respectively in
track prediction with the DA experiments than the control simulations. Improvements in model
performance are also noticed in respect of the intensity prediction. The analysis on the
observational datasets such as (a) zonal wind (m/s), (b) meridional wind (m/s), (c) temperature (°K)
and (d) specific humidity (kg/kg) clearly indicate the improvement in model initial condition with
assimilation. In case of Nargis, the root mean square errors (RMSE) for different variables with and
BOBTEX-2011
Cyclone Warning Division, India Meteorological Department, New Delhi 98
without assimilation indicate the improvement of 50% with the assimilation technique. It may also
be noted that, the buoy and ship data does not show significant improvement like any other
datasets. It may be due to the scarcity in the data coverage and constrained to only surface data. In
case of Aila, RMSE reduces significantly (more than 300%) with AIREP (aircraft temperature and
winds) observations at the initial time. It may also be noticed that, the buoy and ship data
significantly improve the initial condition to the model integration. In case of Jal, significant
improvement is noticed in zonal and meridional wind components with AIREP and buoy datasets.
Further, the structure of the above mentioned cyclones has been investigated. For this purpose, few
diagnostics such as east-west cross section of horizontal wind (m/s), vertical velocity (m/s),
vorticity (x10-5
s-1
) and moisture convergence etc. representing the structure of the cyclones also
shows the improvement in DA experiments than that of control simulations.
Keywords: WRF-NMM, WRF-Var, Tropical cyclone, Track, Vector Displacement Error, Structure
0
50
100
150
200
250
300
350
400
00hr 12hr 24hr 36hr 48hr 60hr 72hr 84hr 96hr
CNTL
DA
Mean vector displacement errors (km)
Fig 1 Mean vector displacement errors from all the 24 cases with
both CNTL and DA experiments