Prediction of tropical cyclone genesis using a vortex merger index

Similar documents
Scale Interactions during the Formation of Typhoon Irving 边建谱 ELIZABETH A. RITCHIE GREG J. HOLLAND

On African easterly waves that impacted two tropical cyclones in 2004

ESCI 344 Tropical Meteorology Lesson 11 Tropical Cyclones: Formation, Maintenance, and Intensification

Vertical wind shear in relation to frequency of Monsoon Depressions and Tropical Cyclones of Indian Seas

Understanding the Microphysical Properties of Developing Cloud Clusters During TCS-08

Understanding the Microphysical Properties of Developing Cloud Clusters during TCS-08

The Formation of Tropical Cyclones 1

Tropical Cyclone Formation/Structure/Motion Studies

1. INTRODUCTION: 2. DATA AND METHODOLOGY:

Lectures on Tropical Cyclones

Relationship between the potential and actual intensities of tropical cyclones on interannual time scales

P Hurricane Danielle Tropical Cyclogenesis Forecasting Study Using the NCAR Advanced Research WRF Model

Department of Atmospheric Sciences, National Taiwan University, Taipei, Taiwan

1. INTRODUCTION. investigating the differences in actual cloud microphysics.

Convective scheme and resolution impacts on seasonal precipitation forecasts

An Objective Algorithm for the Identification of Convective Tropical Cloud Clusters in Geostationary Infrared Imagery

TROPICAL CYCLONE TC 03A FOR THE PERIOD 3 RD JUNE TO 10 TH JUNE, 1998

ESCI 241 Meteorology Lesson 19 Tropical Cyclones Dr. DeCaria

Improving our Understanding of Tropical Cyclone Genesis

Tropical Cyclone Genesis: What we know, and what we don t!

Tropical cyclone energy dispersion under vertical shears

10B.2 THE ROLE OF THE OCCLUSION PROCESS IN THE EXTRATROPICAL-TO-TROPICAL TRANSITION OF ATLANTIC HURRICANE KAREN

Twenty-five years of Atlantic basin seasonal hurricane forecasts ( )

Understanding the Global Distribution of Monsoon Depressions

DISTRIBUTION STATEMENT A: Distribution approved for public release; distribution is unlimited.

地球系统科学前沿讲座 台风研究现状和问题 林岩銮

William M. Frank* and George S. Young The Pennsylvania State University, University Park, PA. 2. Data

Tropical cyclones in ERA-40: A detection and tracking method

16C.6 Genesis of Atlantic tropical storms from African Easterly Waves a comparison of two contrasting years

Inner core dynamics: Eyewall Replacement and hot towers

(April 7, 2010, Wednesday) Tropical Storms & Hurricanes Part 2

Examination of Tropical Cyclogenesis using the High Temporal and Spatial Resolution JRA-25 Dataset

Satellite data analysis and numerical simulation of tropical cyclone formation

An Objective Algorithm for the Identification of Convective Tropical Cloud Clusters in Geostationary Infrared Imagery. Why?

Using satellite-based remotely-sensed data to determine tropical cyclone size and structure characteristics

Hurricanes and Tropical Weather Systems:

Why There Is Weather?

Using NOGAPS Singular Vectors to Diagnose Large-scale Influences on Tropical Cyclogenesis

Initialization of Tropical Cyclone Structure for Operational Application

THE EXTRATROPICAL TRANSITION OF TYPHOON WINNIE (1997): SELF-AMPLIFICATION AFTER LANDFALL

International Journal of Integrated Sciences & Technology 2 (2016) 55-61

Cloud-Resolving Simulations of West Pacific Tropical Cyclones

Understanding the Global Distribution of Monsoon Depressions

1. Introduction. 2. Verification of the 2010 forecasts. Research Brief 2011/ February 2011

Tropical Cyclones. Objectives

Understanding the Global Distribution of Monsoon Depressions

High initial time sensitivity of medium range forecasting observed for a stratospheric sudden warming

The Impact of air-sea interaction on the extratropical transition of tropical cyclones

Tropical Cyclone Intensity and Structure Changes in relation to Tropical Cyclone Outflow

Tropical Storm List

Tropical Cyclone Formation: Results

ABSTRACT 1 INTRODUCTION P2G.4 HURRICANE DEFLECTION BY SEA SURFACE TEMPERATURE ANOMALIES.

Chapter 24. Tropical Cyclones. Tropical Cyclone Classification 4/19/17

Tropical Cyclone Intensity and Structure Changes due to Upper-Level Outflow and Environmental Interactions

A more detailed and quantitative consideration of organized convection: Part I Cold pool dynamics and the formation of squall lines

KUALA LUMPUR MONSOON ACTIVITY CENT

2013 ATLANTIC HURRICANE SEASON OUTLOOK. June RMS Cat Response

Impact of ATOVS data in a mesoscale assimilationforecast system over the Indian region

ATMOSPHERIC MODELLING. GEOG/ENST 3331 Lecture 9 Ahrens: Chapter 13; A&B: Chapters 12 and 13

28th Conference on Hurricanes and Tropical Meteorology, 28 April 2 May 2008, Orlando, Florida.

The Extratropical Transition of Tropical Cyclones

TROPICAL CYCLONE GENESIS IN TC-LAPS: THE IMPORTANCE OF SUFFICIENT NET DEEP CONVECTION AND SYSTEM SCALE CYCLONIC ABSOLUTE VORTICITY

Evolution of Tropical Cyclone Characteristics

The 2009 Hurricane Season Overview

Interrelationship between Indian Ocean Dipole (IOD) and Australian Tropical Cyclones

Understanding the Global Distribution of Monsoon Depressions

A Preliminary Climatology of Extratropical Transitions in the Southwest Indian Ocean

Observation and Simulation of the Genesis of Typhoon Fengshen (2008) in the Tropical Western Pacific

Tropical Cyclone Intensification

The feature of atmospheric circulation in the extremely warm winter 2006/2007

Tropical Cyclone Hyperactivity in the Eastern and Central Caribbean Sea During the 2005 Atlantic Hurricane Season

THE IMPACT OF SATELLITE-DERIVED WINDS ON GFDL HURRICANE MODEL FORECASTS

Anticorrelated intensity change of the quasi-biweekly and day oscillations over the South China Sea

Synoptic Meteorology

2D.1 DETERMINATION OF A CONSISTENT TIME FOR THE EXTRATROPICAL TRANSITION OF TROPICAL CYCLONES

The Properties of Convective Clouds Over the Western Pacific and Their Relationship to the Environment of Tropical Cyclones

Observed Structure and Environment of Developing and Nondeveloping Tropical Cyclones in the Western North Pacific using Satellite Data

UPDATE OF REGIONAL WEATHER AND SMOKE HAZE (December 2017)

Determining Hurricane Formation in the eastern North Pacific Using the Global Lightning Dataset 360 and the Long-Range Lightning Detection Network

Roles of upper-level processes in tropical cyclogenesis

NHC Ensemble/Probabilistic Guidance Products

Comments by William M. Gray (Colorado State University) on the recently published paper in Science by Webster, et al

COLORADO STATE UNIVERSITY FORECAST OF ATLANTIC HURRICANE ACTIVITY FROM AUGUST 16 AUGUST 29, 2013

11/19/14. Chapter 11: Hurricanes. The Atmosphere: An Introduction to Meteorology, 12 th. Lutgens Tarbuck

Feel free to ask for help also, we will try our best to answer your question or at least direct you to where you can find the answer.

Towards a new understanding of monsoon depressions

At the Midpoint of the 2008

NOTES AND CORRESPONDENCE. What Has Changed the Proportion of Intense Hurricanes in the Last 30 Years?

Hurricanes are intense vortical (rotational) storms that develop over the tropical oceans in regions of very warm surface water.

PRMS WHITE PAPER 2014 NORTH ATLANTIC HURRICANE SEASON OUTLOOK. June RMS Event Response

Severe storms over the Mediterranean Sea: A satellite and model analysis

Subtropical and Hybrid Systems IWTC VII Topic 1.6

T-PARC and TCS08 (Submitted by Pat Harr, Russell Elsberry and Tetsuo Nakazawa)

Evaluating a Genesis Potential Index with Community Climate System Model Version 3 (CCSM3) By: Kieran Bhatia

Chapter 1. Introduction

Evidence for Weakening of Indian Summer Monsoon and SA CORDEX Results from RegCM

Christopher C. Hennon * University of North Carolina Asheville, Asheville NC. Charles N. Helms University of North Carolina Asheville, Asheville, NC

TROPICAL CYCLONE GENESIS. Todd B. Kimberlain and Richard J. Pasch WMO RA-IV Workshop on Hurricane Forecasting & Warning 7 March 2016

7 December 2016 Tokyo Climate Center, Japan Meteorological Agency

1. INTRODUCTION. designed. The primary focus of this strategy was the extratropical transition (ET) of tropical cyclones based on the poleward

Transcription:

GEOPHYSICAL RESEARCH LETTERS, VOL. 31, L04105, doi:10.1029/2003gl019005, 2004 Prediction of tropical cyclone genesis using a vortex merger index T. N. Venkatesh Flosolver Unit, National Aerospace Laboratories, Bangalore, India Joseph Mathew Department of Aerospace Engineering, Indian Institute of Science, Bangalore, India Received 6 November 2003; revised 24 December 2003; accepted 9 January 2004; published 19 February 2004. [1] We propose a new method for detecting tropical cyclone genesis at an early stage by supposing merger of mesoscale midlevel vortices to be a common precursor event. The merger event serves as a selection mechanism and is a possible explanation of why only a small fraction of cloud clusters which meet the necessary conditions actually develop into tropical cyclones. The detection procedure uses satellite IR images which are available in near real-time. After tests using data for the Bay of Bengal basin for the years 1999 2001, a real-time test was conducted for the post-monsoon 2002 and pre-monsoon 2003 seasons. We found that the method was successful in detecting the formation of tropical cyclones 04B (2002) and 01B (2003) about 48 hours before they reached storm strength, and no storm escaped detection. INDEX TERMS: 3374 Meteorology and Atmospheric Dynamics: Tropical meteorology; 3329 Meteorology and Atmospheric Dynamics: Mesoscale meteorology; 3360 Meteorology and Atmospheric Dynamics: Remote sensing. Citation: Venkatesh, T. N., and J. Mathew (2004), Prediction of tropical cyclone genesis using a vortex merger index, Geophys. Res. Lett., 31, L04105, doi:10.1029/ 2003GL019005. 1. Introduction [2] Understanding tropical cyclone (TC) genesis remains a scientific challenge to this day. The necessary conditions for genesis have been known for several decades [Gray, 1968] and previous theories have considered how an initial vortex which extends from the seasurface boundary layer to the mid-troposphere intensifies [Charney and Eliassen, 1964; Emanuel, 1986; Rotunno and Emanuel, 1987]. Recent observations [Karyampudi and Pierce, 2002] and studies [Ritchie and Holland, 1997; Simpson et al., 1997] point to the importance of interactions of mesoscale mid-level vortices (MCV) prior to TC formation. We adopt the hypothesis that a merger of MCVs is a common precursor event during the early stages of TC genesis. This model offers a prediction method using satellite IR images which are available in near real-time. After tests using data for the Bay of Bengal basin for the years 1999 2001, a real-time test was conducted for the post-monsoon 2002 and pre-monsoon 2003 seasons. We show that the method was successful in detecting the formation of tropical cyclones Copyright 2004 by the American Geophysical Union. 0094-8276/04/2003GL019005 04B (2002) and 01B (2003) about 48 hours before they reached storm strength, and that no storm escaped detection. [3] Current TC forecasting practice begins with detection of closed low-level circulation centers and persistent, organized thunderstorm activity [Sharp et al., 2002]. A recent advance in early detection has been the availability of QuikSCAT data of sea surface wind velocity. Katsaros et al. [2001] reported one to two days lead times based on identifying closed circulations, but non-developing systems were also detected. Sharp et al. [2002] constructed average vorticity fields and found an average of 26h lead time for the TCs which were detected, but several were missed. The procedure presented here is quite different as it is based on detecting a process (vortex mergers) which occurs at an earlier stage in TC genesis, well before even this process is completed. [4] Merger of MCVs observed in the Atlantic [Karyampudi and Pierce, 2002] Pacific [Simpson et al., 1997; Ritchie and Holland, 1997] and the Bay of Bengal suggest a crucial role for mesoscale vortex dynamics in TC genesis, particularly vortex mergers. Distributed regions of vorticity (termed vortex patches) will merge when the separation distance is below a threshold. For example [Melander et al., 1988], for a pair of identical two-dimensional circular patches of radii r 0, this critical distance between patch centers is 3.2r 0. Vorticity strength provides a time-scale and determines the time for merger. Such critical distances exist for arrangements of multiple patches, and can be determined by simple integrations. This critical behavior provides a selection mechanism, and is a reason to consider vortex merger to be a likely, common precursor process. [5] We consider that the process of TC genesis is as follows. During a TC season, in regions where favorable large scale conditions (Gray s conditions [Gray, 1968]) exist, clouds, cloud clusters, mesoscale convective systems (MCS) and associated MCVs form. Nonlinear vortex dynamics of these MCVs and the interaction with any large scale circulation which could be present, determines whether merger takes place. If the distances are large, the individual vortices will move in circular orbits around the centroid and eventually decay. If the distances are small and the geometric configuration is favorable, merger will result in bigger and longer lived vortices. If this merged vortex extends down to the top of the boundary layer an air-sea interaction mechanism (e.g., WISHE [Emanuel, 1986; Rotunno and Emanuel, 1987]) can amplify it into L04105 1of5

a mature cyclone. In this way vortex merger has a role in both selecting the vortices which will grow as well as providing a mechanism which creates the deep vortex which can then amplify. Although a theory for the full process must consider three-dimensional, baroclinic effects and air-sea interaction, a two-dimensional barotropic theory may be sufficient to capture the essential aspects of the initial stage. This naturally suggests that the salient result of that theory (merger events) can serve as a predictor of TC genesis. 2. Prediction Method [6] The method for prediction begins with identifying MCSs from satellite IR images and then measuring their sizes and spacings. Although the basis is the merger of MCVs a procedure based on IR images should be adequate. Ritchie and Holland [1997] have shown a reasonable correlation between MCVs and MCSs from observations, and theoretical support in terms of balanced dynamics and potential vorticity is available [Raymond and Jiang, 1990]. To detect mergers, an index was defined as m ¼ r* ðnþ R cg =L ; ð1þ where n is the number of MCSs, R cg is the average distance of the systems from the center of gravity, and L is the average radius of the systems. Here the function r* ðþ¼1:85 n þ 0:275ðn 2Þ ð2 n 5Þ ð2þ is an approximation for the critical distance below which mergers of a system of n equal patches occur. This approximation (2) was obtained by studying mergers of equal patches located initially at vertices of regular polygons [Venkatesh, 2003]. Immediately before a merger m will be high because then the spacing R cg /L < r*(n). At other times, or after a merger, m will be low. For an idealized system of identical circular vortex patches the critical value of m is unity. If m > 1, the average radius for the configuration is less than that required for merger. Asymmetric configurations, which are more likely in the atmosphere, will permit merger at distances greater than those for symmetric configurations. So the critical value of m can be expected to be less than unity. 2.1. Identifying MCVS [7] From an input image (Figure 1a) MCVs are identified by finding a connected region with the specified temperature and area thresholds (here, brightness temperatures are below 210 K and areas greater than 10000 km 2 ). Figure 1b shows such regions, circles of the same area as each such region, and lines connecting their centers to the centroid of the areas. The mean radii of the circles L and the mean distance between circle centers to the centroid R cg of this geometrically equivalent configuration are used in equation (1) to calculate m. 3. Study of Past Data [8] To assess the usefulness of the index as well as obtain an estimate of the actual critical value of m, past data for the Figure 1. Identification of MCS/MCV from an IR image and the computation of m. a: The input image. b: The MCVs identified and the equivalent geometric configuration of five patches. Bay of Bengal region (80 100E, 0 25N) in the postmonsoon period (October November) for the years 1999, 2000 and 2001 were considered. [9] The variation of m for the 1999 2001 post-monsoon seasons is shown in Figure 2. During this period, there were two severe TCs (26 31 October 1999 and 27 30 November 2000), one TC (12 13 November 2001) and two depressions which just reached storm strength (16 October 2000 and 27 Oct 2000) [Thapliyal et al., 2000, 2001, 2002, and UKMO website, http:// www.met-office.gov.uk/sec2/sec2cyclone/tcver.html]. One can see that there is a very good correlation between the formation of intense TCs and cyclonic storms and m. There is a peak of m and values above a threshold of approximately 0.5 one to two days before the system develops into a tropical cyclone. 4. Real-Time Prediction [10] Real-time tests were conducted during the postmonsoon season of 2002 and the pre-monsoon season of 2003. The region (2 22 N, 80 100 E) was chosen. An assumption made here is that the necessary conditions for TC formation are satisfied throughout the region during this period. This is reasonable, given the climatology of the 2of5

the index had exceeded a threshold (here, 0.5), and if so, forecast TC genesis. [11] The rectangular region (2 22 N, 80 100 E) had been chosen for simplification of data processing, but this includes significant land areas and had an undesirable effect on the forecasts which were apparent midway during the test: TC genesis forecasts were made even though the merger event had occurred close to or over land. Later a mask was used to exclude land areas. [12] The merger index plot for October and November (Figure 3a) shows three events, labeled M1, M2, and M3, when the criteria was met. During the period October 23 24 the values of m were clearly above the threshold (M1). The processed satellite images for this period show that the systems were very close to the eastern part of Bengal (inset in the figure). The mean wind is generally northwestward, so that any deep vortex would get advected over to land and weaken. An examination of the images would have indicated to a forecaster that the Figure 2. Merger Index in the Bay of Bengal during October to November 1999 2001. Thick vertical lines indicates the time when tropical storm intensity of 17 m/s was reached. a: 1999; b: 2000 (dashed vertical line marks formation of weaker systems); c: 2001. region. Satellite IR images from the Cooperative Institute for Meteorological Satellite Studies, University of Wisconsin-Madison [CIMSS website] were processed to estimate the number of systems, average sizes and R cg. Images were available at three hour intervals and the merger index was estimated at all the available times in a particular month. To filter out the small scale fluctuations and also compensate for gaps in the data, a running mean of 8 consecutive values (1 day average) was used. A set of programs were written to run at 00:00 UTC everyday. The programs would download IR images for the previous 24 hours, process these images to calculate the merger index for each image and find the running mean, check whether three consecutive values of Figure 3. Merger Index in the Bay of Bengal. a: Postmonsoon 2002. M1, M2 and M3 are merger events. The thick vertical line indicates when TC 04B reached storm intensity. TCP indicates when the prediction was made. Inset figures show pseudo-colored IR satellite images of cloud regions in the Bay of Bengal during events M1 and M2. The land areas are shaded lightly. b: Pre-monsoon 2003. 3of5

above 0.5 as shown in Figure 4 of maximum winds and merger index. Figure 4. Variation of m (thick line) and the maximum wind speeds (dashed line) of the tropical cyclone 01B (May 2003). Wind speed in metres per second has been divided by 33 (thus storm = = 0.5 and hurricane = =1.0). chances of intensification were minimal. Thapliyal et al. [2003] classify this system as a depression. [13] The second event (M2) was during the period November 11 12. The processed satellite images for this period show that for this period the systems extended over the eastern part of Bengal. This event corresponded to a short-lived TC. A deep depression was formed near this region by 1200 UTC of 10th which intensified into a cyclonic storm by 11th (See Thapliyal et al. [2003]). This storm further intensified into a severe cyclonic storm at 0600 UTC 12th and weakened rapidly and crossed the coast by 0900 UTC. For this case, the time when m crossed the threshold corresponds to the time when the system intensified into a cyclonic storm and around 24 hours before intensification into a severe cyclonic storm. [14] Towards the end of November 2002, tropical storm 04B formed in the Bay of Bengal. It reached a maximum wind speed of 23 m/s on the 24th. The formation was a very slow process. It took nearly 36 hours to reach storm strength. The values of m crossed the threshold during the middle of the 20th and the prediction was made at 0300 UTC on the 21st (since the processing was done once a day at this hour). In this case, the system was deep in the Bay and given the general nature of the mean wind, it was easy to expect that the system resulting from their interaction would remain over the sea for a sufficiently long time to intensify. Initially, before land areas were masked, mergers were detected on October 6 and November 2 as well. Smaller peaks at these times can be seen in Figure 3a. In both cases the mergers occurred over land and, consequently, the systems did not intensify. [15] In the pre-monsoon cyclone season (April May) of 2003 also, the real-time test was carried out. In April there was no event and this can be seen from m variation also (Figure 3b). In May, the procedure predicted a TC formation on the 9th. A system (named 01B) indeed formed and was classified as a tropical storm on the 11th. It should be noted that the prediction was more than two days in advance. TC 01B had an unusual track and intensity history (See NRL website: http://www.nrlmry.navy.mil/tc_pages/tc_home. html). It first intensified, moved north westwards, and then nearly stopped and moved eastwards finally making landfall over Burma. There was a period of re-intensification which too was preceded by a second crossing of the merger index 5. Summary and Conclusions [16] We proposed and tested the hypothesis that mergers of mid-level mesoscale vortices is a common precursor event to amplification by air-sea interaction. There is observational and theoretical evidence in support of this view. An important consequence is that detection of merger events can provide an early alert. Real-time tests for the post-monsoon season of 2002 and pre-monsoon season of 2003 resulted in two successful predictions, one false alert from a merger which took place close to land, and one alert which was followed very quickly by intensification. The successful predictions were over two days ahead of the event, and, significantly, no system escaped detection. More refined algorithms for separating clusters of interacting systems and a better critical value of the index may improve prediction quality. Even from the three seasons examined before performing the real-time tests, a value of 0.6 could have been selected for the merger index. Then, the marginal case M2 would have been omitted. But, such a finely defined merger index is perhaps unnecessary for providing an alert. Secondly, necessary conditions for TC genesis were assumed to exist everywhere during a month from climatology. Instead, the regions considered could be selected based on actually meeting these conditions. If the post-merger amplification process is included, then there is another threshold vortex strength. However, even in its present form, the merger hypothesis presents a method for the prediction of TC genesis with current technological capabilities. Since it is also at an early stage in genesis there can be a considerable impact on early detection and observations, leading to better track and intensity forecasts. [17] Acknowledgments. We thank Prof. B. N. Goswami, IISc. for helpful comments of this work, and Dr. U. N. Sinha, NAL for encouragement and initiatives taken to support this work. Image data made available freely by CIMSS from their website (http://www.ssec.wisc.edu/tropic/ tropic.html) has been an essential part of our studies. References Charney, J. G., and A. Eliassen (1964), On the growth of the hurricane depression, J. Atmos. Sci., 21, 68 75. Emanuel, K. A. (1986), An air-sea interaction theory for tropical cyclones. Part I: Steady state maintenance, J. Atmos. Sci., 43, 585 604. Gray, W. M. (1968), A global view of the origin of tropical disturbances and storms, Mon. Weather Rev., 96, 669 700. Karyampudi, V. M., and H. F. Pierce (2002), Synoptic-scale influence of the Saharan air layer of tropical cyclogenesis over the Eastern Atlantic, Mon. Weather Rev., 130, 3100 3128. Katsaros, K. B., E. B. Forde, P. Chang, and W. T. Liu (2001), QuikSCAT Facilitates Early Identification of Tropical Depressions in 1999 Hurricane Season, Geophys. Res. Lett., 28(6), 1043 1046. Melander, M. V., N. J. Zabusky, and J. C. McWilliams (1988), Symmetric vortex merger in two dimensions: Causes and conditions, J. Fluid Mech., 195, 303 340. Raymond, D. J., and H. Jiang (1990), A theory for long-lived Mesoscale Convective Systems, J. Atmos. Sci., 47, 3067 3077. Ritchie, E. A., and G. J. Holland (1997), Scale interactions during the formation of typhoon Irving, Mon. Weather Rev., 125, 1377 1396. Rotunno, R., and K. A. Emanuel (1987), An air sea interaction theory for tropical cyclones. Part II: Evolutionary study using a non-hydrostatic axisymmetric numerical model, J. Atmos. Sci., 44, 542 561. Sharp, R. J., M. A. Bourassa, and J. J. O Brien (2002), Early Detection of Tropical Cyclones Using Seawinds-Derived Vorticity, Bull. Am. Meteorol. Soc., 83(6), 879 896. 4of5

Simpson, J., E. A. Ritchie, G. J. Holland, J. Halverson, and S. Stewart (1997), Mesoscale interactions in tropical cyclone genesis, Mon. Weather Rev., 125, 2643 2661. Sundqvist (1970), Numerical Simulation of the development of tropical cyclones with a ten level model, Tellus, 22, 359 390. Thapliyal, V., D. S. Desai, and V. Krishnan (2000), Cyclones and depressions over north Indian Ocean during 1999, Mausam, 51, 215 224. Thapliyal, V., D. S. Desai, and V. Krishnan (2001), Cyclones and depressions over north Indian Ocean during 2000, Mausam, 52, 455 462. Thapliyal, V., A. B. Mazumdar, and V. Krishnan (2002), Cyclones and depressions over north Indian Ocean during 2001, Mausam, 53, 265 270. Thapliyal, V., A. B. Mazumdar, and S. Sunitha Devi (2003), Cyclones and depressions over north Indian Ocean during 2002, Mausam, 54, 579 584. Venkatesh, T. N. (2003), A vortex merger theory for tropical cyclone genesis, Ph.D. Thesis, Indian Institute of Science. T. N. Venkatesh, Flosolver Unit, National Aerospace Laboratories, PB 1779, Bangalore 560017, India. (tnv@flosolver.nal.res.in) J. Mathew, Department of Aerospace Engineering, Indian Institute of Science, Bangalore 560012, India. ( joseph@aero.iisc.ernet.in) 5of5