Coastal morphological influence for tropical cyclone track deviation along Andhra coast: GIS and remote sensing-based approach

K. Vinod Kumar, A. Bhattacharya and
C. Subramanyam

National Remote Sensing Agency, Balanagar, Hyderabad 500 0037, India

World-wide attention has been focussed on the need for better disaster mitigation programmes towards all the natural hazards. Tropical cyclone is one such natural hazard that needs better disaster management and prediction. GIS and remote sensing are two powerful tools for monitoring such disasters. Presented here is a case study where IRS-1C WiFS data coupled with historical database from Indian Meteorological Department is effectively used to demonstrate for better monitoring of tropical cyclone crossing the Andhra coast.

BAY of Bengal is one of the six regions in the world where severe tropical cyclones usually originate in the months of May, November, and December. They are well known for their extreme destructive potential and impact on human activities. Associated with the severe cyclones are strong winds, storm surges along the coast, and there is heavy rainfall which results in destruction to life and property. Proper prediction of this natural disaster requires understanding of its genesis, movement, and landfall. Historical database of this tropical cyclone, available with Indian Meteorological Department (IMD), was used to understand its genesis and movement in a geographical information system (GIS) environment. This knowledge can be effectively used for proper prediction in real time mode. Remote sensing and GIS are two important tools for such type of disaster monitoring. IRS-1C satellite data from WiFS sensor covering ground swath of 810 by 810 km was effectively used to understand the regional geomorphology of the terrain.

Historically all the tropical cyclones crossing the east coast had their landfall near the major deltas like Ganges, Mahanadi, Godavari, Krishna, Pennar, Palar, and Cauvery delta1. Authors were interested to understand the behaviour of tropical cyclone's landfall near the major deltas along the east coast. Limited studies have been undertaken world-wide regarding the influence of coastal morphology on the cyclone track movement. Most of the severe cyclones crossing Andhra coast (the present study area) had their landfall in Godavari, Krishna and Pennar deltas. Very few severe cyclones had crossed Ongole coast. Vizag coast generally experiences storms and depression of lesser magnitude and intensity.

The objectives of the study are: (1) to understand the influence of coastal geomorphology for deviation of cyclone tracks crossing the Andhra coast, and (2) use of GIS in creating a historical database for better prediction of landfall of cyclone.

IRS-1C WiFS data dated 4 April 1996 was georeferenced to Survey of India toposheet covering the area. The total study area is bounded by 80  to 99 E long. and 7   to 24 N lat. The tracks of severe storms crossing the Andhra coast were extracted from the IMD Atlas1. These were coregistered with the georeferenced satellite data for accurately overlaying the cyclone tracks on to the satellite image. For better management of the database, the severe cyclone storm tracks (wind speed = 47 knots) crossing the Andhra coast were classified as those which crossed the east Godavari, west Godavari, east Krishna, west Krishna, Ongole, and Nellore coasts. These were generated as separate vector files using Arc-Info GIS software. The tracks that had deviated without crossing the Andhra coast were vectorized as a separate layer (Figure 1). The total track width that covered the extremes of track between the north and south, crossing each of the above coasts was vectorized, and a separate file was generated for proper integration in the GIS. The georeferenced WiFS data was used to understand the regional coastal morphology of the study area. All the vector files generated in Arc-Info format were overlaid on the corrected satellite data (Figure 1) to interpret the relationship between cyclone movement and the local geomorphology.

Tropical cyclone crossing the Andhra coast generally originates between 5  and 7 N lat. belt in the intertropical convergence zones (ITCZ). Statistics of historical records have shown that the maximum number of severe storms crossing this coast originate in November, while second maxima is reached in May2. The movement of the cyclone from its origin point to the landfall is dependent on the steering current. This in turn depends on climatological and environmental factors3. One of the environmental factors is the physiography along the coast. The present study highlights the environmental influence on the movement of the cyclone track as it approaches the coast. Severe cyclone when it approaches the coast is influenced by the local pressure gradients related to the physiography of the coast. The isobars along the coast tend to be parallel to the shape of the coast4-7.

The origin of severe cyclonic storms crossing east Godavari lies between 12  and 13 N lat. and 96  to 97 E long. (Figure 1). The tracks crossing west Godavari originate between 7  and 12 N lat. and 92  and 96 E long. (Figure 1). These two tracks after their origin show overlapping regions between 12  and 16 N lat. and 84   and 96 E long. From 84 E long. onwards towards the coast, the two tracks deviate independently and cross the coast. This deviation point is approximately 350 km away from the coast. Similar deviation of tracks are noticed for other severe cyclones crossing the Krishna, Nellore and Ongole coast (Figure 1). Most of the deviation of the tracks are seen as it approaches the coast within 100 km indicating probably the influence of local geomorphology for such preferential deviation (Figure 1). Most of the severe cyclones avoid the hilly Vizag coast and the interdeltaic areas along this coastal plain.

From the satellite data, Godavari delta shows arcuate shape with an areal extent around 4163 sq km, and its orientation is east-west with convex outline towards Bay of Bengal. Krishna delta on the other hand shows lobate shape with 4600 sq km area and is oriented north-south with convex outline towards the coast. The Pennar delta near Nellore is cuspate shaped with total surface area around 1470 sq km. The delta is oriented in the east-west direction with convex projection towards the coast. The Ongole coast shows concave outline. Most of the tracks approaching these zones have more or less maintained this trend and orientation of the delta. Notice the prominent east-west trend of tracks of Godavari, and north-south trend of tracks of Krishna (Figure 1). All the landfalls are near the major deltas and coastlines with a convex projection. The isobars show a tendency to be parallel to the shape of the coast, and the mesoscale meteorological lows and highs (pressure) depend on whether the coast is concave or convex shaped4.

Probably in the present study area also the deviation of track might be related to the shape of the coast as well as the orientation of these major deltas. The local pressure gradient generated too might be related to this shape. The absolute land temperature computation along this deltaic stretch during the landfall of cyclone may establish the reasons for such preferential pressure gradient along this stretch of the coasts10, which is beyond the scope of present study.

957.jpg (35788 bytes)


The other similarities observed in these areas are the extensive land use practices because of highly fertile soil, and the presence of huge water bodies like Godavari, Krishna and Pennar rivers, and Pulicat lake. The evapotranspiration from these water bodies and fertile soil perhaps provides the necessary conducive environment for a cyclone to approach this area (a cyclone requires unlimited source of water to maintain its strength and vigour)11. Only two of the cyclone tracks were seen to cross this deviation zone, approximately within 100 km from the coast, and deviated further without crossing this coast during the last 100 years (Figure 1).

The entire database of historical tropical storm tracks and satellite data gridded to one degree interval (Figure 1) is useful for monitoring the cyclone tracks in real time mode. This database can be updated regularly when ever a new cyclone crosses the coast. All the meteorological parameters during pre- and post-cyclonic conditions can also be linked in the database for more precise prediction.

Predicting accurately the landfall of cyclone is a difficult task for any meteorologist. However, effective use of historical database is helpful for such type of prediction. In this study, GIS has been used as one of the powerful tools for such type of monitoring using the historical database. Also to understand the influence of local coastal geomorphology on the deviation of cyclone track just before crossing the coast is important for accurately predicting the landfall. Another powerful tool used is remote sensing to map the regional geomorphology of the terrain. The database generated can be updated regularly whenever a new cyclone develops and crosses the coast.

The study has revealed that accurate prediction can be made on the landfall of cyclone when the eye of the cyclone is within the range of 100 km from the coast. This is one step forward towards better disaster monitoring and management, even though the time required for taking disaster mitigation measures is less.

1.IMD Atlas, Tracks of storms and depression crossing Bay of Bengal - 1877-1995.

2.Jayanti, N. and Sen Sarma, A. K., Mausam, 1988, 38, 193-196.

3.Coastal Meteorology, National Research Council, National Academy Press, Washington DC, 1992.

4.Abbs, J. D., Mon. Weather Rev., 1986, 114, 881.

5.Abe, S. and Yashida, T., J. Meteorol. Soc. Japan, 1982, 60, 1074-1084.

6.Atlas, D., Schu Hseien Chow and William, D., Mon. Weather Rev., 1983, 111, 2, 24.

7.Mahru, Y. and Seegal, M. C., Mon. Weather Rev., 1985, 113, 170.

8.Chang, S. W., Mon. Weather Rev., 1982, 110, 1255-1270.

9.Moms, A., Tuley, E. R. and Yoshro Kurihara, Mon. Weather Rev., 1987, 115, 130.

10.Jones Rober, W. A., Mon. Weather Rev., 1987, 115, 2279.

11.Chang Tytai and Wetzel, P. J., Mon. Weather Rev., 1991, 112, 1369.


ACKNOWLEDGEMENTS. We acknowledge the technical advice and critical review by the Director, NRSA, Hyderabad. The technical suggestions of Dr S. K. Sasmal, Scientist, NRSA are duly acknowledged. The encouragement and support from colleagues of Geosciences Group are acknowledged.


Received 24 June 1998; revised accepted 24 August 1998