Thus, for the purpose of planning ahead (rather than responding to disasters after they happen), one needs to focus on the actual physical process that causes destruction and death. For this reason, we abandoned "hurricane" as a disaster type, introducing instead with wind velocity together with inundation (the latter being subdivided in 4 types according to velocity of the water and the rate of water-level rise, plus a type for coastal flooding).

However, deaths caused by building failures as a result of earthquakes, flooding, or high winds, are attributed to the natural hazard in question (rather than to "poor building standards"). The motivation is that the building is part of society, not nature: the disastrous effect on society happens by definition from the moment when the roof caves in, not the split second later when the people underneath dies. Still, it is important to recognize the role the building standards have in preventing a natural disaster from happening. As a general rule, all casualties that happen due to building failures are avoidable.

Definition of natural disaster, vulnerability, and risk

What is a natural disaster? The meaning of the word obviously changes with person, culture, and time. The following definition of natural disaster is from the "International Agreed Glossary of Basic Terms Related to Disaster Management (1992), UN-DHA, IDNDR, Geneva: "Situation or event, which overwhelms local capacity, necessitating a request to national or international level for external assistance."

In a project for the government of Nicaragua, we decided to modify this definition on the ground that it does not capture the intuitive notion of what a natural disaster is. Instead we adopted the following definition of natural disaster:

A physical natural event that kills people or overwhelms local capacity for damage control or recovery

First of all, this limits the definition to physical events, excluding biological events such as diseases or grasshopp invasions. Secondly, it includes any natural physical event that somebody dies from, since no loss of life is an "acceptable loss". Thirdly, it skippes the phrase about requesting external assistance, since that phrase has caused otherwise rather harmless events to be classified as disasters in certain databases.

Still, the definition means that a small earthquake in a poorly prepared country easily classifies as a disaster, whereas a much stronger earthquake in a well-prepared country does not. Thus, disasters are relative. Disasters happen because we are not adequately prepared. Disasters can to a large extent be avoided.

The threat factors in nature can be termed natural hazards, which we chose to define as follows:

Natural hazard = the threat of a dangerous magnitude of a natural process

The factors that decide whether a natural hazard event turns into a disaster or not can be summarized in the term vulnerability, which we defined as:

Vulnerability to an event = the loss from a natural hazard event

The risk, finally, is defined as the probability of occurence times the vulnerability. This is equivalent to vulnerability divided by return period. The unit is loss per year (wheather in monetary terms, lives, per cent, or some other unit).

Since the vulnerability depends on the magnitude, it has a unique value for each magnitude of each process, as shown in the graph below. Also the risk has a value for each magnitude, with a maximum somehwere.

The X-axis is the magnitude of the process (e.g., water level). The left Y-axis is the return period in years and corresponds to the blue line. The right Y-axis is for the red and green lines; the thin lines represent vulnerability (cost*1000), and the fat lines risk (cost/year).

The graph shows two examples, that could be two villages next to the same river. Village A (red) is poorer, and therefore also located on lower terrain where land is cheaper. It gets flooded before Village B (green), but the vulnerability in B increases faster once the water reaches it, since each house represents a much higher value.

The most important values for a risk mapping to find are at what probability the vulnerability starts (10% per year in Village A, 0.5% per year in Village B), at which return period the risk is highest (the 100-year event in A, the 1000-year event in B), and the corresponding values for the magnitude. This requires good quantitative data, something that is not available for Nicaragua yet.

The above graph illustrates the case when the process frecuency is a function of magnitude, which is valid for many hazards. For some, though, there is essentially only one magnitude, and thus only one probability, and one vulnerability value. This is the norm for man-made hazards, such as bridge failures and dam breaks (they either fail or don't fail).

Vulnerability of the country

Only some comments of general interest will be given on a couple of aspects of the vulnerability.

Infrastructure

After Mitch, several countries went in to rebuild Nicaragua (and Honduras), with the ambition to construct something that would prevent future disasters. Much of the destruction was bridges. In the reconstruction (in which Hydroconsult participated in the design phase in Honduras), the design return period had to be decided. A typical value is 100 years, which sounds quite reassuring for many people, including many decision-makers. Also, many of them, being economists, know that 30 years is a reasonable design period, and consider 100 years to be well on the safe side. There seems to be a bit of confusion regarding these concepts, including among people writing books about Natural Hazard Mitigation.

Return period is the inverse of probability. Thus, a 100 year return period signifies that there is a 1% risk that the bridge falls each year. During 30 years, the risk is 30% (the economists will tell you it is only 26%, which is true for the present bridge; but if you intend to rebuild the bridge—and who wouldn't?—you can't count like that).

As we all know just too well, bridge failures frequently, though not always, cause the loss of human lives. If in average one car with 2 persons aboard fall off each destroyed bridge in a lethal accident, then we have a risk level of 0.02 persons/year for each of those bridges (2 persons * 1%).

Bridge failures always bring extra costs, besides the cost of reconstruction. In fact, all it takes for an event to escalate to disaster level is a number of bridge failures. If there are 200 bridges in a country, each with 100 year return period, every 5 years 6 or more of them will fall (assuming the events are not correlated). Since there is almost always some spatial correlation, many of the bridge failures will occur in the same region during the same storm, why the event will escalate to a "natural disaster". Thus, building bridges with a return period of 100 years may be all it takes to generate natural disasters well over once per decade (depending on the number and distribution of the bridges).

There is obviously a discrepancy between the methods used for determining the design criteria for the bridges, and our notion of disaster. The discrepancy is caused by the inability to take the value of human lives and indirect costs of the bridge failures into account in the choice of design return period, i.e., safety level.

Society

Nicaragua has got a large concentration of society functions to Managua, both in the public and in the private sector. Managua is at risk from 15 faultlines through the city. Twice in the 20th century the city was destroyed by earthquakes, with movements of faultlines within the city. There are also a number of volcanic centers within the city limits, that could reactivate in a future earthquake. Being located in a spreading zone with faultlines and volcanic centers that possibly can give rise to widespread lava eruption, Managua is very exposed to natural hazards. The city is also subjected to flash floods, flooding from Lake Xolotlán, and at some risk from mass movements and from nearby active volcanoes; however, the great risk is from the local tectonic movements.

Managua being the capital and the economic centre of Nicaragua, the whole country is highly vulnerable due to the vulnerability of Managua. The most important effort to decrease the vulnerability of Nicaragua no doubt would be to decentralize the administration, and create secure systems that are not jeopardized by a destruction of Managua. Government institutions may be relocated to other cities, and incentives be used to stimulate economic growth in other parts of the country than Managua. It may even be advisable to relocate the capital itself to a safer place, based on modern hazard mapping and GIS analysis. There are safe alternatives, away from the volcanic chain (in connection to which all previous capitals have been located).