India, home to 1.2 billion people, is one of the fastest growing economies in the world. It is also vulnerable to several types of natural disasters, including cyclones, earthquakes, droughts, and floods. According to the Ministry of Home Affairs in India, annual economic losses from natural hazards have more than tripled during the last three decades, with losses during this period estimated at over USD 48 billion1.
Roughly 60% of India's landmass and more than 5,700 kilometers of the coastline is exposed to tropical cyclones. And while cyclones in India are well known for their powerful and destructive winds, they can also produce copious amounts of rainfall, which can lead to inland flooding—a significant driver of loss in the region.
As India's economy grows and global markets become increasingly interdependent, the insurance industry is turning its attention to the potential for large losses in the region. In a positive development, sophisticated modeling tools are being developed to help insurers understand the scale of risk they face. In this article, we examine India's tropical cyclone climatology and how the soon-to-be-released AIR Tropical Cyclone Model for India can help companies assess and manage the potential losses caused by these high impact events.
Monsoons and India Tropical Cyclone Climatology
The North Indian Ocean basin generates an average of five to six storms at tropical storm strength or higher per year. While fewer storms form here than in other basins around the world, the North Indian Ocean is a breeding ground for some of the most intense cyclones in the world. Since 1990 there have been six storms that have impacted India or nearby countries with wind speeds greater than 190 km/h (118 mph and a Category 3 on the Saffir-Simpson Scale), making up 7.5% of all named storms in the basin. This is the highest percentage of super cyclonic storms in all basins worldwide.
While India cyclones can follow a number of tracks, their origins can usually be traced to the same body of water—the Bay of Bengal to India's east. Indeed, cyclones are four times more likely to form here than over the Arabian Sea (Figure 1). One such storm was the 1999 Super Cyclone Orissa, which brought high winds and record-breaking flooding in low-lying areas, damaged more than 10,000 houses, destroyed some 2.1 million hectares of farmland, and caused more than USD 100 million1 in insured losses.
Figure 1. Regional distribution of cyclones along India's coast (by state, 1891-2006) shows the significant proportion of strong storms. (Source: India Meteorology Department) The seasonality of cyclones in the region is related to a large-scale factor of the climatology—the monsoon. The North Indian Ocean tropical cyclone season extends from April to December, with peaks in activity during May to June and October to November (Figure 2). Few strong storms form in July, August, or September, which are considered prime months for tropical cyclogenesis in other northern hemisphere oceanic basins. The suppression of activity during these months is directly related to the strength and position of the North Indian Ocean Monsoon, which tends to shear storms apart and prevents them from intensifying.
Figure 2. Although tropical depressions form in abundance in the Bay of Bengal and Arabian Sea during July, August and September, monsoons tend to prevent their organization into cyclonic storms. (Source: India Meteorology Department, 1950-2009) However, once a system forms, even cyclones with weak winds can cause catastrophic damage. In 2008, Cyclone Nisha came ashore with wind speeds of just 65 km/h (40 mph and tropical storm strength on the Saffir-Simpson Scale), but it generated more than 1,000 mm (39 inches) of precipitation in some mountainous locations and damaged more than 4,000 properties.
Modeling Precipitation-Induced Flooding
As Cyclone Nisha so aptly illustrates, any model that purports to assess India's cyclone risk must capture the risk from precipitation-induced flooding. Indeed, storms needn't make landfall to produce flooding. Precipitation shields can extend hundreds of kilometers—far beyond a storm's damaging windfield. Thus even bypassing storms can produce substantial amounts of rain onshore—and if the storm is also slow moving, the potential for significant flooding is exacerbated. In this section, we take a closer look at AIR's approach to modeling the flood risk associated with cyclone activity in this region.
To determine the probability and severity of a flooding event, the AIR flood module begins by generating total event precipitation, which is determined by accumulating the hourly precipitation at each location over the entire duration of the storm. Since rainfall is accumulated over time, the forward speed of the storm is an important factor; slow-moving storms will subject any given location to higher rainfall totals. Unlike tropical cyclone winds, which generally decrease as a storm moves inland, precipitation can actually increase as the system moves inland.
After the total precipitation is calculated, it is then redistributed based on the porosity of the soil, land use/land cover and slope—all of which determine what fraction of the precipitation is absorbed. If the soil is sandy, for example, a higher fraction will be absorbed than if the soil is clay. The water that cannot be absorbed by underlying soils becomes surface runoff, or flood.