On May 2, 2008 a “very severe tropical cyclone” (category 4) made landfall in Mynamar and moved eastward along the southward facing delta of the Irrawaddy River. The path of the storm could not have been worse. Storm surges inundating the low lying coastal region were largely responsible for the loss of 138,000 people and the destruction of infrastructure and agricultural lands that may take years to repair. Formal warnings from by the Indian Meteorological Department (IMD) in the days before landfall, albeit without storm surge alerts, appear to have been unheeded and the people of the delta were completely unprepared for the arrival of a major storm. We provide a detailed description of the evolution of Nargis.

Whereas the impact of Nargis ranks highly in the list of recent disasters in the region, there have been many North Indian tropical cyclones that have incurred a similar loss of life especially in the littoral regions of the Bay of Bengal that are particularly highly populated. There are three steps by which the adverse impacts of severe tropical cyclones in the North Indian Ocean can be reduced. Noting that modes of evacuation in developing countries are essentially on foot, the first step is the provision of extended range forecasts of tropical cyclone genesis and track beyond the three days now provided by the IMD. For reasons that are unknown, global numerical weather prediction models are able to provide substantially more reliable predictions of tropical cyclones in the North Indian Ocean (and Western Pacific) than in the Atlantic. The high resolution ECMWF EPS system successfully predicted that Nargis would form and move over southern Mynamar days before either the joint Typhoon Center or the Indian Meteorological Department designated Nargis as a tropical cyclone. Similar skill was exhibited with the same model for both Sidr and Gonu, major North Indian Ocean tropical cyclones that occurred in the previous year. The second step is to provide forecasts of storm surges in addition to track and intensity. Given the large number of river deltas with shallow offshore waters in the region, such forecasts should be a matter of course. These first two steps require only the application of existing technology. The last step is more difficult and requires a commitment by individual countries to the development of a warning system that downscales the synoptic scale forecasts to regional meteorological and hydrological risks to the people in peril. The very inexpensive, simple and effective system developed over the past few years in Bangladesh will be compared to what occurred in early May 2008 in Myanmar.