DesInventar as a Disaster Information Management System

DesInventar Methodology

The DesInventar methodology proposes the use of historical data about the impact of disasters, collected in a systematic and homogeneous manner in the process of identifying hazards and vulnerabilities and thus risks on specific regions. Data must be collected following a set of standards and is time-stamped and geo-referenced and disaggregated to a relatively small geographic unit, usually a low level administrative unit.

The basic criteria guiding DesInventar are:

• all inventories must use the same variables to measure the effects and the same homogeneous and basic classification of events.
• The information compiled and processed must be entered in a scale of time and at a geo-referenced spatial level;
• The information comprising DesInventar inventories MUST be spatially disaggregated in order to show (and later analyse) the effects of disasters at local level. For country level disaster inventories it is recommended a minimum disaggregation level equivalent to Municipality, usually one or two levels below the first administrative/political division (Province/State/Department, depending on each country).
• The inventories can then be analyzed following a number of existing and emerging methodologies, starting with the Preliminary Analysis Methodology, which give users an immediate understanding of the impact of disasters in a country or region, the possibilities of comparative research and support to decision-making processes related to risk reduction actions (including Risk Assessments) and risk management as a whole.

As part of the project deliveries there is also an instrument that allows visualizing, in space and time, the phenomena that have been registered using charts and maps.

DesInventar Methodology Summary

A brief summary of DesInventar Methodology is presented here. Please refer to the full document, available in the home page of this site. The methodology is strongly based on a set of definitions and classifications, and the concept of a space subdivided in multiple levels of zonings, but above all it proposes:

• Disaggregating and geo-referencing of data that will later permit the analysis of the data at the minimum level of geographic resolution.
• The collection and use of data about small and medium disasters.

The methodology also discusses a series of suitable sources of information that can be used to gather the inventory information.

Basic Definitions

The core of the Methodology is contained in the definitions of “Event” and “Disaster”. These are not established to contradict or redefine much widespread definitions but to serve as the basis for the systematic work of collecting and storing the information about disasters in an orderly fashion.

“Event” is defined as any social-natural phenomena that can be considered as a threat to life, properties and infrastructure.

“Disaster” is defined as the set of adverse effects caused by social-natural and natural phenomena on human life, properties and infrastructure (an “Event”) within a specific geographic unit during a given period of time.

The different types of hazards associated with these events are thoroughly defined in the Methodology, and the definitions are available in the set of links at the left of this text.

The team of researchers that formulated and refined the Methodology did also invest a significant amount of time discussing the effects of disasters. The goal of these discussions was to reach an agreement on which effects should be tracked by the system, its definition, measuring units and other problems associated with quantitative and qualitative measures and representation of these effects.

The set of chosen effects address the definition of Disaster in the core of the methodology: effects of human life (dead, missing, injured, evacuated, relocated and directly or indirectly affected persons), properties (destroyed and affected houses as well as effects on crops and domestic animals) and on the infrastructure (transportation, communications, utilities, etc.).

Please refer rto to the Methodology and User Manual documents for detailed definitions of all the effects.

The concept of minimal resolution geographic unit

One of the cornerstones of the methodology are the concepts of geo-referencing and disaggregation of the information in the inventory. When a project is started in a given region of study one of the first steps is to select a division (zoning) that will allow researches to perform analysis and reach to the desired conclusions at a convenient level of detail.

It is a well known fact that relief organizations, emergency management agencies and media tend to aggregate the information of disaster losses with several different purposes that go from allocating budget and resources to support the relief operation to just obtain sensational data for news headlines.

However, the analytical objectives of the information gathered for disaster inventories require fully disaggregated data for each of the geographical units in the selected zoning system. Unfortunately large and medium disaster information is very often not available in its disaggregated form. Small disasters, due to its nature are easily geo-referenced, as they normally do not transcend the borders of the geographical units where they occur in.

Disaggregating data is a difficult task, which has tremendous implications on the work and the later usability of the information. On one hand, utilizing data about disasters gathered at highest geographic resolution (that is, with the smallest possible sizes of geographic units) will lead to a finer detail in analysis and results, but in the other hand it will raise dramatically the level of effort and difficulty when disaggregating data.

One of major lessons learned from the work done so far in the project is that the challenge of disaggregating the data is definitely the major difficulty that an inventory research faces.

To ease some of the burden of disaggregating data, and to allow the users to get results at different levels of detail, DesInventar considers the space as hierarchically divided into several levels of divisions. At each level of the hierarchy each geographic unit is conformed by several units at the next level of detail.

For national level databases the project normally considers a first level of geography to be equivalent to province, and a second level typically assimilated to municipalities. Adherence to the political-administrative division is recommended, as other statistical data needed in the risk analysis process is likely to be produced for that specific division. Disaster management agencies also are organized in most cases around the standard administrative division.

Researchers face the problem of disaggregating data very frequently and there are many instances where the problem has simply no solution, especially when going back in time many years and in cases where the original files are not available anymore. The methodology suggests several workarounds, some of them controversial or with severe implications in the usability of the data during the analysis phase.

Recommendations and How to's

The Methodology contains also a series of suggestions and recommendations that will help those in charge of creating the historical inventories in facing frequently happening challenges.

These, in summary, are:

• Possible alternatives to including non-disaggregated data in the system
• Classification and prioritization of Sources of Information
• Management of Discrepancies between multiple data sources
• Treatment of Long Duration events
• Treatment of "Chained" events
• Procedures and suggestions when heographic units are split or consolidated
• and several other frequent challenges...