Previous PageTable Of ContentsNext Page

Annex XIII



The assessment of food insecurity at subnational, national, regional and global levels is of paramount importance for development planning and for directing efforts and resources to combat malnutrition and hunger. The World Food Summit (WFS), held in 1996, recognizing the need for an assessment of the extent of hunger and malnutrition, called for the development of food insecurity and vulnerability information and mapping systems (FIVIMS), indicating areas and populations affected by or at-risk of hunger and malnutrition, and elements contributing to food insecurity. To support the overall FIVIMS effort, the FAO Global Information and Early Warning System (GIEWS) is implementing a Trust Fund Project entitled Asia FIVIMS with financial support from the Government of Japan. With the collaboration of the Centre for Research on the Epidemiology of Disasters (CRED), the project has so far geo-referenced a total of 992 disaster events, or 2422 cumulative number of provincial level units covering the period of 1990-1999. These geo-reference data can be used to produce disaster frequency or occurrence maps for various disaster types for Asia. The mapped data and information on disaster-prone and vulnerable areas will be disseminated to FIVIMS data users and policy makers through an Internet-based data management, sharing and analysis tool known as the Asia Key Indicator Data System (Asia KIDS) for further analysis and decision making.

1. Introduction

1.1 The Need for Food Insecurity and Vulnerability Assessments

1. The assessment of food insecurity at subnational, national, regional and global levels is of paramount importance for development planning and for directing efforts and resources to combat malnutrition and hunger. Similarly, a knowledge of the number, nature and exact location of vulnerable populations is required by contingency planners, relief operations, and international aid organizations, in times of food emergency and/or civil strife, for accurately targeting the beneficiaries of food aid programmes. Food importing developed and developing countries also require vulnerability information since external shocks, including natural and man-made disasters, may disturb the availability of food in the international markets.

2. The World Food Summit (WFS), hosted by FAO in Rome in November 1996 with representatives from 185 countries in attendance, also recognised the need for an assessment of the extent of hunger and malnutrition, including at local level, and pledged to reduce by half the number of undernourished people not later than 2015. Specifically, Commitments II, V and VII of the World Food Summit Plan of Action request governments to "develop and periodically update, where necessary, a national Food Insecurity and Vulnerability Information and Mapping System (FIVIMS), indicating areas and populations, including at local level, affected by or at-risk of hunger and malnutrition, and elements contributing to food insecurity, making maximum use of existing data and other information systems in order to avoid duplication of efforts" (FAO 1996 and 1998). FIVIMS was considered as a required tool for member states and international organisations to adequately plan and direct their efforts and resources to meet the WFS target.

3. In this regard, the WFS also called upon FAO to play a catalytic role, within the UN family, for the further elaboration and definition of the System and for its development in a co-ordinated manner, starting with the development of guidelines for the establishment of national FIVIMS, and for the determination of suitable indicators for the assessment of food insecurity and vulnerability at national and subnational levels. FIVIMS, when fully established, will assist policy makers from both industrialised and developing nations to identify vulnerable groups and areas within the countries for adequately directing their resources and efforts for reducing the poor and undernourished, as well as for establishing policies and strategies which will ensure stable food production and supply, and improve access to safe and nutritious food by all.

1.2 The Asia FIVIMS Project and Natural Disasters Assessments

4. In order to support the overall FIVIMS effort, the FAO Global Information and Early Warning System (GIEWS) is implementing a Trust Fund Project entitled "Development of a Vulnerability Information Base, Mapping and Dissemination System for Asia, often referred to as Asia FIVIMS, with financial support from the Government of Japan. Asia FIVIMS seeks to assemble, analyse and disseminate information about food insecure and vulnerable populations in Asia, providing information on their geographic locations and the causes behind their food insecurity and vulnerability.

5. To date, various elements and shocks contributing to food insecurity and vulnerability in Asia have been considered by Asia FIVIMS. These range from food self-sufficiency to agricultural and crop production to food emergencies and disaster risks. Disaster risks are included in Asia FIVIMS analyses since they are a major cause of acute food shortages and supply in Asia as well as increasingly leading to food emergencies in the region. In Cambodia, for instance, the worst flooding in 70 years that started in early July 2000 was responsible for the estimated 374,000 hectares of rice damaged or destroyed, around 5,000 livestock killed, and US$ 10 million of estimated economic damage to crops alone (FAO/GIEWS 2000).

6. Disasters are often associated with a precipitous decline in well-being and food security status, and therefore information on disasters are required in addition to other food insecurity and vulnerability indicators being developed and applied by Asia FIVIMS, such as dietary adequacy and anthropometric indicators which capture a dimension of long-term, or chronic, food insecurity.1

7. The disaster data included in Asia FIVIMS analyses will provide an indication of the disaster risk (relative frequency), and provide information on the magnitude and types of disasters and emergencies in a country. The disaster data also provide information on causes of disaster-induced food emergencies at national and subnational levels. These are useful indicators for assessing a country's vulnerability to major shocks and disasters that may lead to very acute food insecurity and to food emergency situations.

2. CRED Natural Disaster Database

2.1 The Asia FIVIMS Project - CRED Collaboration

8. Asia FIVIMS coordinates with other international organizations and national agencies involved in vulnerability and risk assessments in order to collect data required for determining vulnerability factors, and to include them as part of the Asia FIVIMS database for further analysis. In relation to disaster data, a major project partner is the WHO Collaborating Centre for Research on the Epidemiology of Disasters (CRED) at the Université Catholique de Louvain in Brussels, Belgium. CRED maintains a database on the occurrence and effects of over 12,000 mass natural and technological disasters in the world from 1900 to the present known as the Emergency Events Database, or EM-DAT. Since mid 1999 the Asia FIVIMS project has collaborated with CRED in geo-referencing EM-DAT data to produce a time-series, digital geo-referenced data set of natural disasters at subnational level for Asia. These data will be used to investigate causalities between food insecurity and natural disasters, as well as the vulnerability related to such events. They can also be used to evaluate the significance of events and their frequency, and to allow the development of appropriate strategies to be taken by the countries.

2.2 The Emergency Events Database (EM-DAT)

9. Since 1988 CRED has been maintaining EM-DAT with support from WHO and the Government of Belgium. It provides policy makers and programme managers with historical and consistent data on disasters and their effects, according to country and disaster type, in order to serve the purposes of humanitarian action at national and international levels. It is also aimed at informing rationalised decision making as well as to provide an objective basis for vulnerability assessment and priority setting. For example, it helps policy makers to decide whether floods in a given country are more significant in terms of their human impact than earthquakes.

10. The database is continuously updated from various sources including UN agencies, non-governmental organisations, insurance companies, research institutes and press agencies.2 Specialised CRED' staff validate the data contained in EM-DAT by researching different information sources and materials. For a disaster to be entered in EM-DAT, at least one of the following criteria must be fulfilled: (1) 10 or more people reported killed; (2) 100 people reported affected; (3) a call for internal assistance; and/or (4) a declaration of a state of emergency. The core data contained in EM-DAT include: date, type of disaster,3 continent,4 country, region, deaths, affected population, injured, homeless, an estimate of damage and additional information. Until 1995, the USAID-OFDA (Office of US Foreign Disaster Assistance) is the main source of the data. A collaboration between CRED and OFDA started in 1999 assisted in completing the EM-DAT database and validate its contents, while making the EM-DAT data widely available via the Internet (

2.3 Disaster Profile for Asia

11. Asia was most affected by natural calamities and suffered disaster-induced damages and losses during the final quarter of the 21st century. This is evidenced by looking into the global statistics of natural disasters during 1975-2000 produced from EM-DAT.

12. Chat 1 shows that both disaster occurrence and people affected for Asia outnumber those for other regions in the world. The cumulative number of disaster events for Asia between 1975 and 2000 is 2,171, which accounts for 38.2 per cent of the total disasters for the world.5 The total number of people affected during the same period accumulated to 3,548,204,253 persons, exceeding 87 per cent of that for the world reported in EM-DAT. The figures appear to be remarkably high when taking into consideration that about 20 per cent of the global land belongs to Asia, which accommodates 56.9 per cent of the total world population.6

Chart 1.
Cumulative number of natural disaster events reported in 1975-2000

Chart 1.
Total number of affected people
for the same period

13. The number of disaster events for Asia recorded in EM-DAT increased gradually from 1975 to 1997, and rose sharply from 1998 (Chart 2). Throughout the period, floods, storms and earthquakes together account for two-thirds of the total disasters for Asia (Chart 3), and are responsible for more than 70 per cent of the total people killed,7 and more than 99 per cent injured.8 It is also noteworthy that more than 25 per cent of the people killed during the same period are attributable to famines bringing 270,760 people to death. The total number of people killed during 1975-2000 is 1,056,528.

Chart 2.
Number of disasters reported in EM-DAT during 1975-2000

Chart 3.
Composition of natural disasters occurred in Asia during 1975-2000 (per cent)

3. Geo-referencing the CRED EM-DAT Natural Disaster Data

14. The CRED maintains disaster information by country in the EM-DAT database. In order to adequately gauge food security situations and allow comparisons at subnational levels, Asia FIVIMS required that EM-DAT be geo-referenced at least by 1st administrative unit, or at provincial level, for all countries in Asia.9

3.1 Geo-referencing Methodologies and Remodelling of the EM-DAT Database Structure

15. All information on natural and technological disasters compiled by CRED is stored, maintained and continuously updated in the EM-DAT database. The master database table is comprised of a number of data fields which record critical information on disasters ranging from the number of people killed to estimated economic losses to actual affected areas denoted by name. The area names that are recorded in the table reflect various geographical features and administrative units, such as provinces, cities, villages, rivers, ports, and regions, and sometimes encompass large-scale areas such as the Mekong River Basin. Subnational locations affected by disaster are recorded by name only in EM-DAT and lack geographical reference co-ordinates or locational codes that are required for mapping the affected areas. These data have since been geo-referenced through a procedure that links administrative regions, other boundaries, and physical features to the descriptive location names found in EM-DAT.

16. In addition to lack of geo-location codes, the original database structure and relational model did not allow effective geo-referencing and mapping. Spatial databases usually maintain attribute information by geographical features such as administrative units, ecological zones, towns and streets. On the other hand, when a disaster affected more than one subnational or 1st administrative areas, EM-DAT stored all of the affected places' names in one data field, which are attribute information on spatial features. In other words, the disaster information is stored by disaster event rather than spatial feature. In order to geo-code the disaster information and to make it mapable, it was required to modify the original database structure and relational data model by adding a spatial data component to the database (Verelst 1999). The following new data tables were added to the database for this purpose, and a relational database model developed (Figure 1).

"Code_ADM1" Table: Unique codes for all 1st administrative areas for Asia were developed based on the digital administrative boundary maps available at FAO. The code list was then imported into the EM-DAT database. It was used for assigning a unique subnational unit code to each descriptive area name. The table contains information on all current 1st administrative units in Asia by unit name, country name and a unique identifier for each unit.

Figure 1.
Modified relational data model
and tables of EM-DAT

"Disaster_ADM1" Table: This table is added to the database for storing geo-referenced data. Each row of the table contains a 1st administrative unit code and a corresponding disaster number linking the master disaster table to the "Code_ADM1" table. The actual geo-referencing procedure populated this data table by assigning an administrative unit code from the "Code_ADM1 table" to each disaster event referring to affected place names found in the database. The original disaster event number was repeated if the disaster affected more than one administrative area.

17. In addition to the above modifications, an updated data entry module was developed for EM-DAT in order to enable efficient geo-referencing. The module provides rapid access to all the information necessary for geo-referencing in one window, coupled with search facilities, and allows users to make online queries for obtaining digital map images for affected areas, location names and co-ordinate information. A series of online geographic names reference sites were added to the database as an integral part of EM-DAT.10 These are maintained separately from EM-DAT, and are dynamically linked, providing visual and textual information required to identify the accurate location of disaster affected areas.

18. The modifications allowed the creation of maps and charts to become more or less semi-automated. By taking advantage of query builders and SQL connections from within mapping software such as ArcView GIS, it is now possible to create thematic maps of natural disasters by directly retrieving geo-referenced information from EM-DAT.

3.2 Known Limitations

19. There were at least three constraints identified in processing the EM-DAT data and producing the geo-referenced component:

  1. The number of casualties, i.e., people killed, injured and affected, is originally stored by disaster event, regardless of whether the disaster affected more than one administrative unit, thus making it impossible to disaggregate the number by each affected subnational region. As a result, the geo-referenced component cannot represent the number of casualties according to 1st administrative unit.
  2. Information on affected places is also noted for 2nd and 3rd administrative levels in EM-DAT. While geo-referencing to the 2nd or 3rd administrative level could be undertaken at national level where more detailed information on the disasters is available and where additional local knowledge can be used to assist the geo-referencing, the disaster events that contain information on affected 2nd or 3rd administrative areas (in most cases, they are district or commune names) were assigned instead to the location codes of 1st administrative units to which they belong, and stored in the "Disaster_ADM1" table for the purpose of Asia FIVIMS.
  3. Similarly, point and line-related information were converted to area-related attribute information during the course of geo-referencing. For instance, a volcano eruption occurred at Mount Lokon in April 1991 (point-related information) was assigned to the 1st administrative code of Sulawesi Utara Province, Indonesia, in which the mountain is located.11 Line-related information (e.g., roads and mountain-ridges) were treated in the same fashion.

4. Identification of Disaster-prone Food Insecure Regions

4.1 Development of Disaster Occurrence Maps

Figure 2.
Natural disaster occurrence by 1st administrative unit
for 1990-1999(East, Southeast and South Asia)

20. As of the end of March 2001, 992 disaster events for Asia recorded in EM-DAT during the period of 1990-1999 were geo-referenced to 1st administrative, or provincial, level. As a result, 2422 administrative units were geo-referenced in relation to disasters which occurred in East, Southeast and South Asia as well as part of West Asia, i.e., Iran and Afghanistan. The scope of the geo-referencing activity is to process all natural disaster data for Asia from 1975 to 2000, representing a total of 2375 disaster incidents. The geo-referencing of the remaining 1383 disaster events is expected to be finalised by mid 2001. This section illustrates some example products using the geo-referenced EM-DAT natural disaster data for 1990-1999, with focus on food insecurity and vulnerability in Asia.

21. The modifications to the EM-DAT database described in the previous section, enable the creation of disaster occurrence tables by subnational unit that are compatible with most Geographic Information Systems. Disaster frequency and/or occurrence maps for various disaster types for Asia were then produced using the geo-referenced EM-DAT database for further analysis and interpretation. An example showing disaster occurrence during the period is presented in Figure 2.12 The disaster data when mapped will allow FIVIMS stakeholders, and policy and decision makers to understand the extent and magnitude of various types of natural disasters which have occurred in Asia, as well as to investigate the effects on disaster-prone food insecure populations.

4.2 Data Validity

22. Although it is still premature to conclude the validity of maps derived from the disaster incidents that occurred during the limited time period of 10 years from 1990 to 1999, the disaster occurrence maps could be used to portray past disaster events in combination with other variables. The map below illustrates the frequency of earthquakes in 1990-1999 for Iran, Afghanistan and Pakistan (Figure 3). The earthquake occurrence map was overlaid with major tectonic contacts and faults in the region and the hypocenters of significant earthquakes recorded from 1950 to 1994.13 The map depicts that the earthquakes recorded during the 10-year period, correspond with the mapped hypocenters except for the western part of Iran.

Figure 3.
Earthquake occurrence map for 1990-1999 overlaid with tectonic ontacts/faults
and major historical hypocenters (Iran, Afghanistan and Pakistan)

23. A flood frequency map developed from the geo-referenced EM-DAT provides a similar observation on the validity of the data set. Figure 4 shows the number of flood events recorded for Bangladesh and Northeastern India during 1990-1999 compared with inundated lands (delimited in red) which were detected and mapped using satellite images form the NOAA-AVHRR, Landsat-7 and Radarsat sensors in 1985-2000.14 Most flooded regions detected by satellite imagery can be found within the subnational areas of high to middle flood frequency, despite a gap in the data collection period between the two.

Figure 4.
Flood occurrence at 1st administrative level during 1990-1999
with inundated areas delimited by satellite images for 1985-2000 (red)

24. Important to reiterate when interpreting the disaster maps produced from the geo-referenced EM-DAT, is that not all disaster events which have occurred in the world are included in EM-DAT. Disasters are entered in the database if 10 or more people were reported killed; (2) 100 people reported affected; (3) a call for internal assistance was made; and/or (4) a state of emergency was declared. Any maps produced from the geo-referenced data set, as well as the disaster frequency depicted in the maps, reflect these criteria.

4.3 Types of Natural Disasters

25. The types and nature of natural disasters vary from country to country in Asia depending on their geographic location, topographical and geological characteristics, and climate and environmental conditions. To illustrate the difference, the cumulative number of affected subnational units were classified by disaster type and expressed in percentage for Indonesia and Viet Nam (Chart 4). The left chart represents 122 cumulative number of provinces that were geo-referenced based on 134 natural disaster events recorded between 1990 and 1999 for Indonesia. During the period, seven different natural disasters affected the countries ranging from floods to earthquakes to droughts.15 When compared to other types of disasters, floods, wild fires and earthquakes are three major threats to Indonesia that affected the country according to the geo-referenced data.

Chart 4.
Cumulative number of provinces affected by disasters during 1990-1999 expressed in percentage

26. On the other hand, floods and tropical storms together account for more than 90 per cent of the total cumulative number of affected provinces in Viet Nam during the same 1990-1999 period. The number was generated based on 169 cumulative number of provinces geo-referenced or 68 disaster events. On average, Viet Nam is hit by 4 to 6 typhoons each year, and over 70 per cent of the population in Viet Nam is at risk of floods and tropical storms as reported by the Disaster Management Unit in Ha Noi.16 Floods and tropical storms alone are responsible for the deaths of, and injuries to, hundreds of people and result in tens of thousands of people being rendered homeless each year according to EM-DAT.

4.4 Impacts of Natural Disasters on Food Supply and Acute Food Insecurity

27. Disasters have been increasingly a major cause of crop losses and damages and often result in food supply shortages in Asia as noted by the FAO member states at the 25th FAO Regional Conference for Asia and the Pacific and the 16th Session of the Committee on Agriculture (FAO 2000 and 2001). Two actual cases are provided below which demonstrate impacts of natural disasters on crop production in relation to disaster frequency.

  1. Indonesia: Provinces in Indonesia which have experienced a higher frequency of disasters have had more crop areas destroyed and/or damaged. During 1998-1999, most provinces were hit by disasters from one to three times. In order to depict the phenomenon empirically, the ratio of disaster-destroyed crop lands to planted areas in 1998-1999 was compared to disaster occurrence for each province.17 On average, provinces with one disaster event occurred during the period, lost 3.83 per cent of the total planted areas per year. For provinces with two events, the ratio increased to 5.82 per cernt. Only one province was severely affected by three different disasters over the two years in Indonesia: Kalimantan Timur Province had 59.4 per cent of total planted lands damaged or destroyed. The disasters that paralysed the Kalimantan Timur's agricultural lands included two forest fires in February 1998 and in June 1999, respectively, as well as a flood in August 1998 that caused 7,000 hectares of lands to be inundated and more than 100,000 people affected.
  2. Viet Nam: The larger the scale of the disaster, the more severe the damages it can cause. This is demonstrated by most seriously affected areas in terms of crop loss during 1997-1998 in Viet Nam (Chart 5).18 The areas which substantially lost crops and crop lands are identical to the provinces most severely inundated by floods and storms, i.e., South Central Coast and Mekong River Delta Provinces as illustrated in Figure 5. It is interesting to note, however, that the provinces in North Central Coast which did not experience crop loss as severely as the South Central Coast and Mekong River Delta regions, were among most frequently hit by natural disasters between 1990 and 1999.

Chart 5.
Paddy areas inundated by floods/storms vs. crop lands
actually damaged/destroyed in 1997-1998

Figure 5.
(Left) Total paddy fields inundated by tropical storms and floods during 1997-1998;
(Middle) Total actual crop areas damaged and/or destroyed during 1997-1998;
(Right) Disaster frequency for 1990-1999

4.5 Impacts on Poor and Undernourished

28. The effect of natural disasters on transitory, or short-term, food insecurity is evidenced by the past experiences of many disaster-prone countries in Asia as described in the previous sections. Yet, the effects on long-term, or chronic, food insecurity, which is often manifested as undernourishment, is not so clearly understood. The seemingly complex nature of the causalities between the two, as well as a broad range of factors involved in such assessments, require careful examination of a number of variables associated with natural disasters and food insecurity. In addition, a lot of efforts for detailed information collection on the variables must be made.

29. Despite the lack of enough empirical studies, in most disaster-prone food insecure countries, it is the poor and undernourished who would be affected most by natural calamities due to lack of coping capacities to protect themselves against the acute external shocks, and also because they are often residing in marginalized areas where few measures have been undertaken by the responsible authorities for disaster preparedness and mitigation and emergency response.

30. An attempt was made to identify disaster-prone food insecure areas in Indonesia by using the information on poor families and on PEM (Protein-Energy-Malnutrition) prevalence among children under five years old, which is an indication of chronic food insecurity. Both data were produced by the National Food and Nutrition Surveillance System. Figures 6 and 7 show the PEM prevalence rate and the percentage of poor families in each province, respectively, while disaster frequency is depicted in Figure 8. Based on the three factors, provinces that fulfil the following criteria were selected: (1) PEM equal to or above 25 per cent; (2) poor families rate equal to or above 25 per cent; and (3) disaster occurrence equal to or more than five. They include Irian Jaya, Kalimantan Barat, Kalimantan Selatan, Kalimantan Tengah, and Sulawesi Tengah Provinces. When natural emergencies such as floods, tropical storms and earthquakes occur, people in these provinces, especially poor and food insecure, could be most severely affected by outcomes of the acute external hazards.

Figure 6.
PEM (Protein-Energy-Malnutrition) under-5 children prevalence (%) for 1997-1998

Figure 7.
Poor Families (%) by Province for 1997-1998

Figure 8.
Disaster Occurrence by Province during 1990-1999

(a) Irian Jaya; (b) Kalimantan Barat; (c) Kalimantan Selatan; (d) Kalimantan Tengah;
(e) Sulawesi Tengah; (f) Aceh

31. It is, however, premature to clearly define the causalities between disaster frequency and its impacts on poor and undernourished by using the geo-referenced data set that covers the limited time period of 10 years. In this relation, it can be noted that not all provinces of high PEM and/or poverty rates belong to high disaster frequent regions. One of such regions is Aceh Province where the PEM ratio was 37.60 per cent - second highest in the country, and where 41.8 per cent of families were suffering destitution, or below the poverty line. Yet, the province was hit by disasters only twice during 1990-1999.

5. Data Sharing and Dissemination through the Asia Key Indicators Data System (Asia KIDS)

5.1 Development of the Asia KIDS - an Internet-based, data sharing, analysis and dissemination system

Figure 9.
Asia FIVIMS homepage: gateway to National
FIVIMS and dynamic systems

32. In order to allow effective data sharing and dissemination between FIVIMS data users and policy makers, the Asia FIVIMS project is developing the Asia Key Indicators Data System (Asia KIDS). The Asia KIDS is a web-based integrated information management, analysis, and dissemination system with online mapping and data analysis capabilities (Figure 9). It makes use of the database technologies used by GIEWS and the Key Indicators Mapping System, or KIMS, being developed by FAO WAICENT. The Asia KIDS will enable system users to have rapid access via the Internet to FIVIMS related information, including data on natural disasters, in map, image, tabular and text formats maintained in a single database commonly administered by Asia FIVIMS stakeholders.

33. The geo-referenced EM-DAT natural disaster data for Asia will be shared and disseminated through the Asia KIDS to help improve disaster preparedness and mitigation, elaborate strategies and programmes to deal with various problems as well as understand the causes of food insecurity and vulnerability in disaster-prone countries in Asia. A database of key indicators is a core component of the Asia KIDS. The information base and the new technologies developed by the project will contribute directly to Global FIVIMS and the Global Key Indicators Data System to be established within the framework of the Inter-agency Working Group on FIVIMS (IAWG-FIVIMS).

5.2 Data Profile Analysis Modules

34. The Asia KIDS allows users to browse selected FIVIMS indicators, which are critical for understanding food insecurity and vulnerability situations in Asia, with various types of analytical functions developed by the Asia FIVIMS project. There are currently three major analytical modules available for data analysis (Figure 10): (a) Indicator Browsing Module; (b) Data Profile Analysis Module; and (c) Policy and Programme Intervention.

Figure 10.
Asia KIDS Indicator Browsing and
Data Profile Analysis Modules

Indicator Browser: comprises of "Map," "Table," "Chart," "Metadata," "Links," and "Find" functions. Users can first select an indicator of their interest and build a map combining other geographical information such as lower-level administrative boundaries, rivers, roads, and cities. With the "Table" function the users will obtain time-series figures for the selected indicator when available or they can create a chart in various forms for a given year. Other functions include "Metadata" showing information on the indicator, and "Links" web sites related to the indicator. "Find" will assist the users to find an indicator they are looking for from the indicator list in the database.

Data Profile Analysis for Indicators: allows more in-depth investigation of food insecurity and vulnerability situations by looking into the temporal and spatial trend of the indicator selected or a set of indicators related each other causing food insecurity and vulnerability in Asia. "Overview," "Temporal Analysis," "Spatial Analysis," and "Causal Relation Analysis" functions consist this module.

Policy and Programmes Interventions: provides information on various policies and programmes being implemented currently or undertaken in the past in the countries, as well as assists you with possible interventions and/or new policy options to alleviate food insecurity. Types of assessments include "International Interventions," "Domestic Interventions," and "Impact of Policy on Food Security."

6. Conclusions and Perspectives

35. CRED EM-DAT remains the only source in the world which provides essential information on the magnitude and effect of natural and technological disasters which have occurred world-wide since 1900 to present collated in a single database with standard criteria. The current on-going addition of a geo-referenced component to EM-DAT will allow the data to be used for assessing country and subnational level vulnerability to natural hazards which can lead to acute, transitory food insecurity, as well as for investigating the possible impact of disasters on undernourished and poor residing in disaster-prone vulnerable regions.

36. When combined with other information such as seismology data and satellite imagery, the geo-referenced EM-DAT data, though still an interim product, support the usefulness for using the data while conducting vulnerability and food insecurity assessments for Asia. Yet, there is a need to enhance data accuracy as well as develop more reliable maps of disaster-prone areas by geo-referencing the rest of the data that covers the period of 1975-1989 and 2000.

37. The causalities between the frequency and scale of disasters and the extent of crop damages and losses, i.e., short-term food shortages and insecurity, are well documented from the experiences of many disaster-prone Asian countries, some examples of which were described in this paper. More effort is, however, required to understand linkages between frequency of natural disasters and chronic food insecurity.

38. The methodologies used for geo-referencing and mapping EM-DAT can be applied for geo-referencing regional and national level databases. Using local knowledge and expertise available in-country, geo-referencing can be undertaken at more detailed levels, e.g., 2nd and 3rd administrative units, or district and commune levels, resulting in more accurate and reliable assessments of vulnerability to agricultural disasters and food insecurity, while helping contribute to the development of national capacity.


FAO. 1996. WFS96/3 Rome Declaration on World Food Security and World Food Summit Plan of Action. 13-17 November 1996. FAO. Rome.

FAO. 1998. CFS98/5 Guidelines for National Food Insecurity and Vulnerability Information and Mapping Systems (FIVIMS): Background and Principles. 24th Session of the Committee on World Food Security. 2-5 June 1998. FAO. Rome.

FAO. 2000. APRC/00/3 Food Insecurity and Vulnerability in Asia and the Pacific: World Food Summit Follow-Up. The 25th Regional Conference for Asia and the Pacific. 28 August - 1 September 2000, Yokohama, Japan.

FAO. 2001. COAG/01/6 Reducing Agricultural Vulnerability to Storm-Related Disasters. The 16th Session of the Committee on Agriculture. 26-30 March 2001, Rome, Italy.

FAO/GIEWS. Foodcrops and Shortages. No. 4 September/October and No. 5 November 2000. Rome, Italy.

National Food and Nutrition Surveillance System Team. 1999. Food and Nutrition Situation in Indonesia 1998-1999. Jakarta.

Verelst, Luc. 1999. Recommendations for Geo-referencing EM-DAT Data Base. Consultancy Report prepared for the Asia FIVIMS Project and the CRED. July 1999. Rome.

* Prepared by Naoki Minamiguchi, Vulnerability Analysis Coordinator, Global Information and Early Warning Service, ESCG.

1 In addition to the disaster variables, Asia FIVIMS is testing other risk factors, which are out of the scope of this paper. They include environmental (climate and biophysical) and socio-political (e.g., wars and conflicts) risks and demographic facators that are increasingly causing, and often, worsening food insecurity in Asia.

2 As of the end of March 2001, there are 12456 records in the database, with 8512 recorded as natural disasters and 3944 as technological (or man-made) disasters.

3 Natural disasters include avalanches, cyclones, epidemics, volcanic eruptions, famine, landslides, floods, insect infestations, hurricanes, droughts, storms, earthquakes, tsunami, typhoons and cold waves.

4 Africa, America, Asia, Europe and Oceania are the five continents included. Asia includes East, South, Southeast and West Asian countries.

5 Records for epidemics and insect infestation are not included in the statistics.

6 (Data Source) FAOSTAT ( Land: Asia = 2,678,388; World = 13,387,019 (1,000 ha). Population: Asia = 71,326,781; World = 125,382,963 (1,000). The population ratio was derived from the cumulative population figures from 1975 to 1999 for Asia and World, respectively. 1999 is the newest year for population data available from FAOSTAT as of the end of March 2001.

7 Floods, storms and earthquakes account for 11.0 per cent, 23.2 per cent and 36.6 per cent, respectively.

8 Floods (47.8 per cent), storms (23.0 per cent) and earthquakes (29.6 per cent).

9 The data for the Philippines were geo-referenced to 2nd administrative units, or at district level.

10 The web-based information bases include the Getty Thesaurus of Geographic Names Browser (, the GEONet Names Serve (, Principal Cities and Agglomerations of the World and of Selected Countries (

11 For point-related information, latitude and longitude coordinates were also recorded for future study and applications.

12 Although epidemics and insect infestation are treated as natual disasters in EM-DAT and were geo-referenced, they were excluded from the studies and map creation in this section.

13 (Data Source) USGS National Earthquake Information Center (

14 (Data Source) Dartmouth Flood Observatory (

15 18 cases of epidemics recorded and geo-referenced in the EM-DAT for Indonesia were excluded from this study.

16 (Source) Disaster Management Unit (DMU), the Standing Office of the Central Committee for Flood and Storm Control in Vietnam.

17 (Data Source) Information on damaged/destroyed crop areas were obtained from the publication "Food and Nutrition Situation in Indonesia 1998-1999" by National Food and Nutrition Surveilance System Team, Jakarta 1999.

18 (Source) Disaster Management Unit (DMU).

Previous PageTop Of PageNext Page