1. Types of natural hazards and their distribution

Figure 14 shows data relating to cyclones, floods, earthquakes and volcanic eruptions likely to cause damage. Besides the principal cyclone tracks, lines of equal frequency built out of data standardised over a hundred years have been represented (after NEUMANN, 1993 in ROUX & VILTARD, 1997).

Statistically two types of cyclone tracks dominate: the ESE-WNW ones that mainly affect the Philippines, Vietnam: the cyclones of the ESE sector strengthening as they go over the positive thermal anomaly in the South China Sea; the others correspond to a displacement of cyclones in the SSW sector and affect preferentially Bangladesh or the northern Myanmar coast. The cyclones also grow stronger as a result of intense evaporation arising from the positive thermal anomaly in the Gulf of Bengal waters. Figure 15 illustrates the high frequency of the cyclones which affect preferentially the low deltaic coast of Bangladesh. The cyclones also affect the Arakan plains and mountains causing strong rains because of the orographic effect. The effective tracks of certain cyclones like the Forrest typhoon (November 1992) and to a certain extent the Colleen typhoon (October 1992) recall that major events can be left out by statistical plottings.

In addition to wind and rain effects, cyclones are characterised by destructive storm surges which affect the more highly exposed coastal fringes in Vietnam, Myanmar and particularly Bangladesh and the Philippines. Figure 15 shows that cyclones cause storm surges which can go upstream for more than 100 kilometres. Certain parts of the coast are more frequently subject to storm surge effects than others; this depending on the topography and the possibilities of shelter. Figures 16 and 17 illustrate this phenomenon in the Philippines and Vietnam.

The easily inundated areas and, moreover, those prone to mass movements or droughts cannot easily be represented at this scale, hence the drawing up of national maps. Figures 17 and 18 give some of the examples in Vietnam and Bangladesh. Floods and landslides are favoured in sectors statistically most affected by cyclones though they depend notably on the relief, lithology and the development of vegetation.

The areas prone to earthquakes with a magnitude greater than 5 are given by the Swiss Reinsurance Company (1992) and four zones of decreasing exposure have been identified. All the coasts of the region under study are exposed to tsunamis and particularly those nearest to earthquake epicentres. Here again, as shown in figure 19, the topography and shelter conditions may cause locally differences as concerns exposure. The data related to volcanism have also been taken from the Swiss Reinsurance Company and RANTUCCI (1994). The affected zones, are directly related to the regional geodynamic framework. South East Asia lies in the area of the planet where the Indian, Eurasian and Pacific plates come into contact. These contacts correspond to the destruction of the lithosphere by subduction, hence a high deep focus earthquake frequency. Their epicentres surround this southern extension of the Eurasian plate constituted by the continental part of South East Asia.

The map of potential threats (figure 14) and figure 20 showing the possibilities in terms of variety and hazard intensity , highlight three groups of countries.

Fig. 14 - Hazards map

Key to figure 14 1. Lines of equal frequency of cyclones using data normalized over 100 years. Reference periods: - 1945-1988 on the North West Pacific - 1891-1989 on the Bay of Bengal Source: Neumann (1993), in Roux et Viltard (1997) 2. Principal tracks of the major cyclones . Source: Neumann (1993), in Roux et Viltard (1997) 3. Observed and/or partly estimated paths of three cyclones: Eli, Colleen and Forrest (1992). After data from the "Water Resources Journal", June 1993, No. 177. 4. Areas vulnerable to river floods. Miscellaneous sources: Asian Development Bank (1991), DHA (1994), etc. Exposure to earthquakes 5. Zones prone to very high exposure 6. Zones prone to high exposure 7. Zones prone to average exposure 8. Zones prone to low exposure Source: Swiss Reinsurance Company (1992) 9. Areas of recent or active volcanism Sources: Swiss Reinsurance Company (1992), Rantucci (1994)

* The Philippines and Bangladesh

These two countries both have a very high degree of exposure to several threats but show notable differences. Bangladesh is affected almost only by hydro-meteorological phenomena (cyclones, floods, and droughts), while all hazard types are potentially present in the Philippines.

* Myanmar, Vietnam and Laos

Myanmar, Vietnam and to a certain extent Laos are affected by most of hydro-meteorological and induced (mass movements) phenomena and have a potentially high degree of exposure. The earthquake threat is low except in Myanmar. Volcanic activity does not exist.

* Thailand and Cambodia

The principal threat in these countries is flooding.

Historical data on natural disasters in the areas affected by hazards will supplement the information given above.

Fig. 15 - Cyclone and storm surge hazardareas in Bangladesh (Source: Asian Development Bank, "Disaster mitigation in Asia and the Pacific", Manila, 1991, p. 142.)

Fig. 16 - Historical storm surges in the Philippines (Source: Asian Development Bank, "Disaster mitigation in Asia and the Pacific", Manila, 1991, p.250.)

Fig. 17 - Areas subject to flooding in Vietnam (Source: DHA, "Strategy and action plan for mitigating water disasters in Viet Nam", 1994, p. 1.0.)

Fig. 18 - Flood types in Bangladesh (Source: Asian Development Bank, "Disaster mitigation in Asia and the Pacific", Manila, 1991, p. 144.)

Fig. 19 - Tsunami prone areas in the Philippines (Source: PHTVOLCS, "Geologic hazards and disaster preparedness systems", 1987, p.53.)

Fig. 20 - Types and relative intensity of hazards faced by the seven countries (According to our hazard analysis and "Economic and Social Commission for Asia and the Pacific", 1995, p. 4)

2. Disaster frequency and distribution

According to the CRED-UCL database, the 7 countries studied considered as a whole were above all affected by cyclones (up to 60% of the 700 events recorded between 1900 and 1996) and more than 25% of the events are floods (¹¹). The other events show much lower frequencies, always less than 5% (Figure 21). These percentages are globally equivalent to those for the period 1972-1996 (Figures 21, 22 and Figures 25 to 29). In comparison with the other phenomena, there are many more deaths and affected people registered for floods and cyclones; approximately 95% of the deaths and affected people result from these two destructive phenomena. It has however been noted that the cyclones have caused many more deaths than the floods. The floods have been responsible for many more affected people (almost three times) and are most likely a plague to the concerned economies.

¹¹ It is sometimes difficult to establish the difference between effects registered under "cyclones" item and those under "floods'. The effects of storm surges are generally put under "cyclones" whereas the effects of some floods though related with cyclones can be found sometimes under the second item. This has a tendancy of increasing the significance of cyclones.

The analysis of the frequency of events by country and by phenomenon (figures 22 to 29; appendices 1, 2 and 3) allows, to a certain extent, to confirm the diagnosis established in terms of potentialities and the following conclusions can be drawn:

* in terms of the raw frequency data, both Bangladesh and the Philippines have by far experienced the greatest number of catastrophes during the last decades.

- The frequency of the major catastrophes is higher in the Philippines; nearly 47% of events were registered between 1972 and 1996 (figure 22). At the same time the diversity of the phenomena that occurred confirms the established diagnosis in terms of potentiality. The phenomena that occur emphasise the cyclones very high recurrence (more than 6 cyclones per year). The earthquakes and volcanic eruptions are relatively frequent. There are likely to be many more floods (about 2 floods per year) and other phenomena like mass movements but, as mentioned above, the minor events are not present in the database because of the statistical limits used.

- 27% of the catastrophic events, essentially of hydro-meteorological origin (3 cyclones and 2 floods on average per year), have occurred in Bangladesh. However two of the events are of earthquake origin.

- There are a significant number of catastrophes that have occurred in Vietnam and Thailand (67 and 37 respectively between 1972 and 1996), however these numbers are lower than those of the previous countries. They are mainly of hydro-meteorological origin. Earthquakes also occur (except in Thailand) as can be assumed from the map of potential threats.

- Myanmar, Laos and Cambodia have been less frequently hit by disasters, though there is a doubt about the reliability of the information especially in the case of Cambodia. With the exception of Myanmar which was affected by some earthquakes, all the countries were affected solely by events of hydro-meteorological origin.

* in terms of the phenomena that occurred, it is clear that cyclones, floods and to a certain extent droughts are, by far, the main destructive phenomena known and potential for all the seven countries. They are therefore a common concern. Earthquakes are a threat to the Philippines, Bangladesh and Myanmar, and, to a certain extent, Vietnam and Thailand. The registered number of major earthquakes in the last decades seems to be relatively low compared to the potentialities.

Volcanism is only present in the Philippines. In the last 25 years, five of its principal active volcanoes erupted (Canlaon, Bulusan, Taal, Mayon, and Pinatubo).

* There is generally a relatively strong relationship between the events that have occurred and the potential threats (¹²). This excludes the earthquakes, floods, droughts likely to occur more frequently and certain countries that may be affected to a greater extent (Laos and Cambodia). On the other hand, some more significant distortions appear in comparing the frequency maps, number of deaths and affected people (figures 10 to 12). The most obvious distortions have been observed in the Philippines and Bangladesh. There are a more significant number of disasters in the Philippines, but proportionally many more deaths and affected people in Bangladesh. Besides the physical component, the above facts highlight the importance of the human factor. It is therefore essential to analyse the vulnerability criteria in order to have a better estimation of the risks.

¹² It is difficult to determine precisely the reasons for the distortions observed. They can be related both to the real frequency of the events and the insufficiency of available data. The period 1972-1996 has been preferentially analysed because it is during this period that the data are more reliable as compared to the whole period 1900-1996.

Fig. 21 - Events, deaths and affected people from 1900 to 1996, according to the nature of disaster phenomena.

1900-1996

	EVENTS	%	DEATHS	%	AFFECTED	%
Cyclones / storms	416	59.4	673,279	88.3	131,223,459	27.6
Floods	180	25.7	64,276	8.4	327,364,163	68.9
Earthquakes	30	4.3	10,298	1.4	2,064,262	0.4
Droughts	20	2.9	18	0.0	13,314,118	2.8
Volcanic eruptions	19	2.7	6,331	0.8	1,012,489	0.2
Landslides	17	2.4	1,256	0.2	39,122	0.0
Tsunamis	9	1.3	5,940	0.8	60,918	0.0
Cold waves	9	1.3	886	0.1	0	0.0
Total	700	100	762,284	100	475,078,531	100

1900-1971

	EVENTS	%	DEATHS	%	AFFECTED	%
Cyclones / storms	117	74.1	480,693	96.1	22,237,737	45.0
Floods	18	11.4	11,470	2.3	27,119,433	54.8
Earthquakes	10	6.3	1,626	0.3	334	0.0
Droughts	1	0.6	0	0.0	0	0.0
Volcanic eruptions	9	5.7	5,612	1.1	96,899	0.2
Landslides	1	0.6	40	0.0	0	0.0
Tsunamis	2	1.3	532	0.1	0	0.0
Cold waves	0	0.0	0	0.0	0	0.0
Total	158	100	499,973	100	49,454,403	100

1972-1996

	EVENTS	%	DEATHS	%	AFFECTED	%
Cyclones / storms	299	55.2	192,586	73.4	108,985,722	25.6
Floods	162	29.9	52,806	20.1	300,244,730	70.5
Earthquakes	20	3.7	8,672	3.3	2,063,928	0.5
Droughts	19	3.5	18	0.0	13,314,118	3.1
Volcanic eruptions	10	1.8	719	0.3	915,590	0.2
Landslides	16	3.0	1,216	0.5	39,122	0.0
Tsunamis	7	1.3	5,408	2.1	60,918	0.0
Cold waves	9	1.7	886	0.3	0	0.0
Total	542	100	262,311	100	425,624,128	100

Source: CRED database.

Fig. 22 - Types of events by country (1972-1996)

	Cyclones	Floods	Droughts	Earthquakes	V. eruptions	Others	Total
PHILIPPINES	155	51	5	15	10	18	254
BANGLADESH	78	51	4	2	0	12	147
VIETNAM	43	20	1	1	0	2	67
THAILAND	15	20	2	0	0	0	37
MYANMAR	5	10	0	2	0	0	17
LAOS	3	7	6	0	0	0	16
CAMBODIA	0	3	1	0	0	0	4

Source: CRED database

Fig. 23 - Types of events (1972-1996): comparison between countries in absolute value

Fig. 24 - Types of events (1972-1996): comparison between countries in relative value

Fig. 25 - Number of disasters caused by cyclones or storms (1972-1996) (Source: CRED database)

Fig. 26 - Number of disasters caused by floods (1972-1996) (Source: CRED database)

Fig. 27 - Number of disasters caused by earthquakes (1972-1996) (Source: CRED database)

Fig. 28 - Number of disasters caused by volcanic eruptions (1972-1996) (Source: CRED database)

Fig. 29 - Number of disasters caused by droughts (1972-1996)