Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 126
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises 5 Population Data and Crisis Response in Mali, Mozambique, and Haiti In this chapter the committee examines three countries to assess how population data and tools were used by governments and agencies to provide humanitarian assistance following different types of natural and human-induced disasters. The selection by several of the study sponsors of these three countries—Mali, Mozambique, and Haiti—was based, in large part, upon the range of assistance resources directed to those countries in accordance with key U.S. government policy initiatives (for example, Millenium Challenge Account and the President’s Emergency Plan for AIDS Relief) and other priority lists. The country examples compare and contrast data use in disaster preparedness and response between nations with relatively good, georeferenced population data disaggregated to the level of villages (Mali and Mozambique) and one nation (Haiti) with outdated national population records at the time of the disasters. The comparisons illustrate that good, georeferenced population data exist, as indicated in Chapter 2, but that the presence of good population data sets and the tools and skills to use them does not in itself guarantee effective within-country or international aid response to a subnational population affected by a disaster (see also Chapters 3 and 4). The country examples show that adequate response to assist populations at risk depends not only on access to high-quality population data, but also on general disaster preparedness, adequate infrastructure, and communications and co-ordination within the country and between government(s) and aid agencies (see also Chapter 4; Doocy, 2007, and Landau, 2007; see also Appendix E). This analysis leads to the chapter’s main recommendations to integrate
OCR for page 127
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises national statistical offices (NSOs) in disaster response coordination and to improve subnational vulnerability analyses. For each country the committee has addressed the background to and nature of the disaster, the results of the disaster for the affected population, and the response to the crisis in terms of data and geospatial information used. A focus on these issues was deemed more important, in the committee’s view, than an exhaustive historical overview of each country. The historical context, where relevant, has been developed from several general resources and the committee found United Nations reports, in particular, to be sufficient and thorough for this purpose. The committee would like to acknowledge the information sources it employed otherwise in this chapter, as we placed high value on gathering input from individuals in the national statistical or emergency management and development offices in each of these countries to accomplish the requested task. To this end, we contacted professionals from these countries with a set of questions regarding this study, together with invitations to our workshop. With direct and very valuable comment and participation received only from Mali, the committee supplemented its factual knowledge regarding population data and its use in emergency and development situations in Mozambique and Haiti by contacting other persons and organizations external to the national offices of these countries. These included humanitarian and development organizations of different sizes and purviews, as well as the U.S. Census Bureau which had established projects for several years in Mozambique. Personal interviews were also supported by information the committee acquired and determined applicable from disaster assessment reports issued during and after the countries’ crises by various international agencies like the United Nations, Save the Children, and the World Health Organization. COUNTRY EXAMPLES Mali Background and the Disasters About two-thirds of Mali, north of 15 degrees north latitude, is covered by desert or semidesert of the Sahara and bordering short grasslands of the Sahel region (Figure 5.1). Northern Mali is inhabited principally by nomadic people of the Touareg and Arab-Berber (Moor) groups. The agricultural zone of the Niger River Basin in the south and east of the country is populated by at least six major ethnic groups including the Bambara, the Soninke, the Malinke, the Songhai, the Dogon, and the Voltaic peoples. The Human Development Index (HDI) of the United Nations Development Programme (UNDP, 2005a) places Mali among the poorest countries in
OCR for page 128
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises FIGURE 5.1 Map of Mali showing some of the primary geopolitical areas. The broad opaque band across the map represents the Sahel—the border region to the Sahara Desert in the north of Africa. SOURCE: Used with permission from http://www.maps.com. the world in terms of combined factors of longevity, health, education, and living standards for the population. Approximately 64 percent of Mali’s population lives below the country’s poverty line (UNDP, 2005b). Much of Mali’s population thus falls under the definition of “vulnerable” or at-risk as defined in Chapter 1. Two crises are discussed for Mali in this chapter: a locust infestation and a civil rebellion. These crises became humanitarian disasters affecting vulnerable groups of people and demonstrate how both natural and human-induced events exerted pressure on subnational populations in terms of food security, health, and economic stability. This discussion of Mali draws heavily from Konaté (2007; see also Appendix E), in a technical paper prepared specifically for the committee. Locust Invasion of 2004. Beginning in June of 2004 during the agricultural season, the first swarms of desert locusts moved from the spring breeding grounds in Morocco and Algeria to the Sahel. Intensive control operations in Northwest Africa, mounted by the relatively resource-rich countries of Morocco, Algeria, Tunisia, and the Libyan Arab Jamahiriya,
OCR for page 129
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises contained the locusts (FAO, 2004a). The situation was different in Sahelian West Africa when the swarms moved southward into Senegal, Mauritania, Mali, and Niger (FAO, 2004b). The Sahel is a poorer region with substantial subsistence agriculture and limited resources for locust control and surveillance. Under these conditions, Mali suffered its worst locust infestation in 15 years beginning in July 2004 (Konaté, 2007; see also Appendix E). The destruction in the agricultural belt of Mali was not complete; however, harvests and pasture were severely affected in several areas, and the situation was compounded by a drought triggered by an early end to the rainy season. In all, two-thirds of the country was affected by the locust invasion. The area struck most severely by the locust hazard lies between 14 and 21 degrees north, and includes the regions of Timbuktu (Timbouctou), Mopti, Gao, and Kayes (Figure 5.1). The majority of the population in these regions of Mali was vulnerable to the locust infestation, and a food security crisis ensued in Mali and other West African countries (UNCAP, 2005). According to the coordinator of the Unit for Migratory Locust Control (ULCP) and the African Project for Emergency Control of Migratory Locusts (PALUCP), “As technicians we expected this, because in 2003 we saw a build-up (of locusts) in the region. We treated almost 40,000 hectares in Mali…. From the month of February (2004) the FAO [Food and Agriculture Organization] launched an appeal to the international community announcing that as soon as rains arrived in the Sahel there would be locusts. By the month of March, we had drawn up an action plan that we presented to the Ministry of Agriculture and all the other Ministries involved, as well as to the development partners. Unfortunately there was no reaction. We then presented this plan more than ten times before the month of July. There was no response. It was from July that we began to have the first swarms, and people, or at least the authorities, began to move. First an operational headquarters was set up … From the month of August … It was from September that we began to receive responses from partners” (Konaté, 2007; see also Appendix E). A small part of a locust swarm can consume the acreage of cropland needed to provide food for 2,500 people. The 2004 invasion affected both crops and pastureland. Approximately 88 percent of Mali’s poor live in rural areas. Approximately 1.7 million farmers (UNCAP, 2005) were affected, and a nutritional survey conducted in the affected areas by the Malian government and the UN World Food Programme (WFP) indicated acute malnutrition rates up to 16 percent in Gao and Kayes. The worst-hit provinces were Gao (Bourem) on the Niger River, Kidal in the remote Adrar des Iforras hills of the northeast, and in the regions of Kayes (Nioro) and Koulikoro (Nara) (WFP, 2005) (Figure 5.1). In addition to the food crisis due to crop destruction by locusts, the added effects of the drought led to an unseasonal population movement: nomadic herders migrated earlier and
OCR for page 130
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises to different locations than usual in an effort to find water and food for their livestock. In some cases, the herders were in direct competition with sedentary, subsistence farmers for the same scarce resources, thus exacerbating the crisis situation (UNCAP, 2005). Touareg Rebellion. Between June 1990 and March 1996, Mali experienced a civil war in the north known as the Touareg rebellion (Konaté, 2007; see also Appendix E). This rebellion, which took place directly after the 1987-1989 locust infestation, resulted from years of dissatisfaction of the people of northern Mali with what they interpreted as uneven distribution of national wealth and services and political marginalization (UNHCR, 1998). The Touareg people initiated the revolt in 1990 and were joined by Arab-Berbers. From 1991 to 1995, negotiations between the Malian government and the insurgents, with mediation by international groups and local elders from various ethnic communities, facilitated several efforts to obtain peaceful settlement to the conflict. However, the various components of the primary peace agreement established in 1991 proved difficult to implement, and insurgent activities were renewed periodically during the succeeding four years throughout much of the northern territory of Mali (the Azaouad). By mid-1995, security conditions in northern Mali had finally improved through negotiation mediated by cooperative efforts between the international community and local elders. The militias agreed to lay down all arms in April 1996 at a symbolic container, the “Flame of Peace,” constructed in the middle of Timbuktu under the aegis of the United Nations (UNHCR, 1998). In addition to the injury and death of combatants and civilians and the disruption of daily activities and services associated with the armed civil action, the rebellion caused displacement of 150,000 refugees into neighboring countries of Niger, Algeria, Burkina Faso, and Mauritania (UNHCR, 1998) (Figure 5.1). The population crisis that resulted from this conflict thus involved (1) repatriation of refugees who were returning to their homeland when peace was restored, and (2) assistance to the local populations who remained in northern Mali and faced shortages of food and water and disrupted access to basic services after six years of civil strife. Data Used in Response to the Crises. Population data have been collected in Mali during the past several decades (Konaté, 2007; see also Appendix E), and some of these data are digitally georeferenced to the village level. Existing population data in Mali at the time of the 2004 locust invasion included two National Censuses of Population and Habitat (1976, 1998); three Demographic and Health Surveys (1987, 1995-1996, 2001); budget and consumption surveys (1987, 1989); a national household activity survey (1989); and various agricultural surveys. During the 1990-1996 civil conflict, some of these same data also would have been available, with
OCR for page 131
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises the exception of the most recent national census (1998) and Demographic and Health Survey. Other international and national institutions based in Mali also have produced information on population and related samples of different sizes. Among the international institutions involved in these activities are the Famine Early Warning System Network (FEWS NET), the African Sub-Saharan Economic and Statistical Observatory (AFRISTAT), and the Sahel Institute (Konaté, 2007; see also Appendix E). Although not the “gold standard” described in Chapter 2, Mali had, at the time of these crises, some population, demographic, and geospatial data to use in prevention and relief efforts, although the census in Mali does not include questions on ethnicity, a practice continued since the period of French colonial rule (Konaté, 2006). The committee was unable to obtain specific information on the accessibility of the existing population data to outside agencies either before, during, or after the crises, or the degree to which these data were employed and disseminated by the government prior to and during the crises. The fact that the locust invasion and ensuing food crisis occurred at all, despite early warning and calls to the international community to react, indicates an ineffective response that either did not employ the population and demographic data or did not employ it early enough to alleviate the food security crisis that ensued (UNCAP, 2005). Box 5.1 gives examples of the manner in which population data in Mali can be used in the future to prepare for and respond to the danger of locust invasions. The committee did not locate published analyses of the use (or lack of use) of previously existing population data in connection with the Touareg rebellion; such data could have been employed for purposes of relief and repatriation of refugees displaced to neighboring countries during the con- BOX 5.1 Applications of Population and Demographic Data in Preparedness and Response to Threat of Locusts To coordinate local communities prevention, surveillance, and control measures, mitigating the need to respond to a food security crisis and reducing expenses of air support to conduct intensive crop protection spraying and surveillance activities To plan food stores (to mitigate effects of future locust invasions) and to distribute more and specialized crop seeds (for planting during curtailed growing seasons) To identify those areas and people who are most vulnerable to the threat of an agricultural crisis and the potential damage from a locust invasion To coordinate food relief when a crisis ensues
OCR for page 132
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises flict, although as noted above, existing national census data were 20 years old at the time of the repatriation and likely of limited use unless they had been supplemented in the intervening years by additional surveys (household, health, or other surveys as described in Chapter 2). Although not a substitute for national census data, the baseline from which aid and development responses can be centrally coordinated, systematic collection by relief organizations of the vital characteristics of displaced and affected persons through local surveys in refugee camps provided information essential for relief and repatriation efforts during and after the conflict. The UNHCR (1998) report provided specific evaluation of the collection and use of refugee population data by UNHCR and the WFP in 1995-1998 related to these repatriation and relief efforts. The experience gained in collecting and using these data is illustrative of the network of factors involved in estimating and putting to use subnational population data collected during and immediately following a disaster, and is explored briefly here. The refugee and repatriation situation was one focused upon distributing food aid, ensuring access to clean water, and facilitating cross-border registration and peaceful transfer of refugees back to Mali. These activities were part of an international effort with challenges that included regional coordination of relief efforts between Mali, the four countries hosting the refugees (Niger, Mauritania, Burkina Faso, and Algeria), and international aid agencies, primarily UNHCR and the WFP (UNHCR, 1998). Central to the relief efforts was identification of the numbers of refugees in various camps within the countries of asylum. Censuses within the camps were conducted, but only the camp census conducted in Mauritania produced reliable figures. A former government in Mauritania had apparently produced some refugee numbers early in the crisis, but a review by UNHCR officials of these numbers by observation of aerial surveys, clinic and school attendance, and number of inhabited dwellings indicated that this initial number of refugees, 85,000, appeared to be inflated. A new census of refugees conducted through coordination between the government of Mauritania, the UNHCR, the WFP, and other donor partners revised that original number by half. These data allowed accurate amounts of food, clothing and transportation to be provided to these people and facilitated border crossings and resettlement of the people back into Mali. Unreliable census figures for refugees in the other asylum countries were attributed to varying government policies in countries and lack of sufficient international (UNHCR) staff in the field. Dispersal of refugees in isolated centers would not have been a significant factor contributing to these inaccurate counts had sufficient staff with appropriate government support been engaged in the census-taking processes in these countries. The relative accuracy of the refugee census in Mauritania greatly facilitated effective refugee repatria-
OCR for page 133
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises tion from this country compared to repatriation from the other asylum countries (UNHCR, 1998). Mozambique Background and the Disaster Flooding, especially due to cyclones, is a perennial problem in Mozambique (Figure 5.2). In February-March of 2000, severe flooding affected 4.5 million people representing nearly 25 percent of the country’s population. Due to this event, 700 fatalities were recorded and more than 650,000 people were displaced. In the ensuing several years, flooding again caused displacement of 220,000 people in the same central provinces (in 2001), while in 2003 flooding affected 100,000 people in the northern provinces (Gall, 2004). In addition to immediate food crises for hundreds of thousands of people displaced from their homes, the floodwaters and lack of sanitation increased risk of waterborne diseases, and flooding raised the risk of longer-term food security issues as livestock died and cropland and seeds FIGURE 5.2 Map of Mozambique. SOURCE: Used with permission from http://www.maps.com.
OCR for page 134
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises were destroyed. Shelter and clothing, and a basic transportation infrastructure, essential to the distribution of assistance, were also lacking. As is the case with Mali, information from the HDI of the UNDP places Mozambique among the poorest nations of the world (UNDP, 2005c). The floods should be set in the context of the significant impact of the war for independence (1962-1974) and the civil war (1976-1992) in Mozambique, coupled with several years of drought that produced a devastated economy, a crippled infrastructure, and millions of refugees and internally displaced persons (IDPs) (http://www.state.gov/r/pa/ei/bgn/7035.htm). Although by 1996, the government had begun to register real annual economic growth (UNGA, 2002), Mozambique did not have the resources to respond adequately to and recover from disasters of these scales in the absence of humanitarian and development assistance. In 2000, the year of the most extensive flooding, external assistance accounted for 23 percent of the country’s gross domestic product (GDP; UNDP, 2002). The repetition of these floods—categorized as sudden-onset disasters—in consecutive years increased the vulnerability of the populations because the recovery and development period between flooding events was very short. Data Used in Response to the Crisis At the time of the first flooding in 2000, Mozambique had conducted (1997) and published a national population census (INE, 1999) and a Demographic and Health Survey (MDHS, 1998). The INE (Instituto Nacional Estatistico) data were used during the crisis, primarily by the National Institute for Disaster Management (INGC) of Mozambique, an institution established in 1999 under the Ministry of Foreign Affairs and Cooperation to manage natural disasters and to serve as the central body with which international organizations (donors and aid agencies) would interact in a disaster response situation. The INGC had worked together with UNDP, the WFP, the International Federation of the Red Cross and Red Crescent Societies (IFRC), and domestic emergency teams in a simulated disaster exercise in 1999 (UNGA, 2002). However, at the onset of flooding in early 2000, the INGC was still completing its final structural arrangements, the protocol for the government and international organizations to follow during disaster response was not yet fully established (UNICEF, 2000), and the magnitude of the flooding was beyond what the disaster simulations had attempted. The INGC, in conjunction with the Mozambique National Statistical Office, was nonetheless able to use and distribute the 1997 census data, in digital, georeferenced form and make them available to international aid organizations present in the country throughout the emergency response period. The accessibility of the georeferenced data was the end product of a development initiative for Mozambique carried out between
OCR for page 135
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises 1997 and 1999 by the U.S. Census Bureau and the Mozambique NSO, and supported financially by the U.S. Agency for International Development (USAID; G. Ferri, personal communication, January 2007). The U.S. Census Bureau had been asked by USAID to conduct a technical assistance and capacity-building program with the Mozambique NSO associated with preparation, execution, and analysis of the Mozambique 1997 national census. An officer from the U.S. Census Bureau with expertise in census data processing, analysis, and dissemination was assigned to Mozambique in 1997-1999 as a local adviser for this purpose. The end product of this work in 1999 was a compact disc (CD) containing census microdata, some of which had been georeferenced, with summary tables and thematic mapping tools. The project was being completed at the time of the 2000 flood, and USAID recalled the U.S. Census Bureau officer to Mozambique to provide technical support to the Mozambique NSO during the flood relief period. The digital census data on CD were linked to geographic information systems (GIS) and new local population surveys conducted during and directly after the flood to produce updated population and demographic data in real time to relief agencies for use in aid acquisition and distribution. After that flood, the Mozambique NSO took full responsibility for the census data and its management (G. Ferri, personal communication, January 2007). Direct uses of the 1997 census data are cited in numerous agency reports on the flooding crisis, for example: “The government has estimated that roughly two million people have been affected by flooding, including 650,000 IDPs. The 650,000 figure cited by most agencies is an estimate based upon census data conducted in 1997, combined with information on which areas were flooded. This figure is also used by WFP to calculate emergency food beneficiaries. Of those displaced, 463,000 were living in 121 accommodation centres and an unknown number were in isolated areas. According to government figures, an additional 300,000-400,000 people were seriously affected in their needs for medical and other non-food assistance. An additional 900,000 people were indirectly affected, according to the figures” (OFDA, 2000; http://cidi.org/disaster/00a/0122.html). The government in this case refers to the INGC. A contingency plan for the 2001 flood was developed by the INGC and international organizations using experience and data gained from the 2000 flood; these contingency plans allowed the WFP, for example, to have pre-positioned food stores in place prior to the onset of the floods in 2001. The 2001 floods, however, brought slightly different challenges because of topographic and demographic differences among the flooded areas: the highly populated southern part of the country was affected in 2000, while the dispersed population in the center of the country, with correspondingly difficult access conditions, was most affected in 2001 (UNGA, 2002). One
OCR for page 136
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises example of the visualization of the georeferenced population data during the 2001 flood is the thematic map shown in Figure 5.3. Other examples of these types of maps were also produced with various themes (e.g., rainfall, flood warnings, population) and show the flexibility of georeferenced population data applied to the disaster relief situation. Population figures based on estimates from the 1997 census could be revised during both of the flooding crises as refugee shelters and centers for displaced persons were established. Field personnel provided estimates of numbers of people and demographic characteristics to coordinate distribution of aid and supplies. During the 2001 flood, the WFP hosted the mapping center where data on populations, floodwaters, shelters, and other factors were collected and updated. Subsequent to the emergencies, resettlement programs were not entirely successful and the need for improvements in disaster assistance led to an assessment of shelter access and an analysis of the socioeconomic impacts of the 2000 floods. A rapid flood assessment and the development of a hazard mitigation strategy were conducted under an emergency aid agreement between Austria and Mozambique (Gall, 2004). Part of that emergency aid assistance involved a geospatial site suitability analysis of shelters based on remotely sensed data on land cover, population, and in situ measurements (global positioning system [GPS] measurements and interviews) (Gall, 2004). The study found that spatial analyses should be incorporated into the disaster management procedures of the INGC and development agencies alike. Furthermore, the research concluded that spatially referenced data sets were critical; that sharing of these among the stakeholders was essential; and that detailed, up-to-date, and complete georeferenced data were required for vulnerability assessments including the provision of health care and food aid at the subnational level (Arndt and Tarp, 2001; Gall, 2004). Further efforts to increase disaster preparedness through better analysis of population data after the floods also emerged through technical cooperation between USAID, INE, and the International Programs Center (IPC) of the U.S. Census Bureau in 2002-2003. A Census Bureau staff member with expertise in statistical data handling and population data was sent to Maputo at the request of USAID and INE to help develop estimation procedures for the 2002-2003 Inquérito aos Agregados Familiares (IAF, a national household income and expenditures survey) results from the first two quarters, to finalize the weighting procedures and standard errors for the IAF, and to review the sampling considerations for a follow-up annual poverty indicator survey (see http://www.census.gov/ipc/www/imps/activ0303a.htm). At present, the U.S. Census Bureau with support from USAID is also providing some technical assistance to the Mozambique NSO as it prepares to conduct the 2007 national census (G. Ferri, personal communication, January 2007).
OCR for page 138
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises Haiti Background and the Disaster Small island states are particularly vulnerable to natural and human-induced disasters, and research suggests that Haiti (Figure 5.4) is among the most vulnerable (Pelling and Uitto, 2001; see also Chapter 1). The states suffering the greatest number of serious disasters in the 1970s and 1980s were island states (13 out of 25), and the statistical vulnerability of island states to disasters has been known for 15 years (Briguglio, 1993). Analytically, Pelling and Uitto (2001) tied increased vulnerability to lack of integration with the world economy, local poverty, and the resilience of the local economy. Moreover, poverty always increases vulnerability to disaster—one of the principal and most robust findings of international disaster research. Haiti is the most disaster-prone island in the Greater Antilles—a group of islands that are, themselves, some of the most disaster-prone islands in the world (Pelling and Uitto, 2001). Haiti was near the bottom of the UNDP “Disaster Risk Index” (UNDP, 2004), with some of the worst health indicators and economic indicators in the western hemisphere (PAHO, 2001). FIGURE 5.4 Map of Haiti. SOURCE: Used with permission from http://www.maps.com.
OCR for page 139
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises Haiti is located in a region that is particularly subject to tropical storms and hurricanes and to significant flooding from the rainfall that ensues from these storms and hurricanes because of two factors: (1) deforestation of slopes, which increases runoff, and (2) concentration of population along coastal areas and near rivers. Thus, the secondary effects of storms are intensified. Tropical Storm Jeanne in 2004 did not meet the wind-speed criteria for designation as a hurricane, but the storm was slow moving and spawned extremely high rainfall as it began to affect the Haitian coast and then the highlands on May 23, 2004 (Regan, 2004). The winds from the tropical storm were not unusually damaging; however, the resulting flooding was destructive and killed more than 2,000 people, leaving several hundred thousand people displaced. Gonaives, the third most populous city, was most seriously affected (Farmer, 2005). Much of the flooding was due to the fact that 98 percent of the land is deforested, as a result both of an attempt to cultivate one sector of an export economy and of the need for domestic timber. Deforestation of mountain and hill slopes increases both the magnitude and the velocity of runoff, thereby increasing its destructive power. Thus, land degradation and deforestation are specific contributors to Haiti’s vulnerability and to the effects of severe storms such as Jeanne. While deaths were attributed to a variety of factors, a statistical basis for the link between deforestation, poverty, and enhanced population vulnerability after Tropical Storm Jeanne has been suggested by comparison between the 2,000 dead in Haiti, which has had extensive deforestation, and the 19 dead in the adjacent Dominican Republic, which has retained most of its forest cover (http://news.bbc.co.uk/2/hi/americas/3685534.stm). Data and Organizational Structures Employed in the Crises Until recently, population data for Haiti were absent or very difficult to access. This condition resulted from a combination of factors including lack of resources to conduct censuses or surveys (Figure 5.5) and government disarray following former president Duvalier’s fall in the late 1980s. The year 2004 was particularly difficult for Haiti: the coup d’etat in early 2004 was followed by a May 2004 flood and accompanying mudslides that were responsible for more than 2,000 deaths in both Haiti and the Dominican Republic. This background is important in understanding the response to Tropical Storm Jeanne and the use of population data in the disaster relief and recovery period. A national census was undertaken in Haiti in 2003 and was released on CD in May 2006, with a plan to make it available on the web, according to one UN official (H. Clavijo, personal correspondence, May 2006).
OCR for page 140
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises FIGURE 5.5 Estimated population distribution for Haiti and the Dominican Republic based on LandScan data. SOURCE: UNOSAT, LandScan 2002, Global Insight Plus / Europa Technologies Ltd.
OCR for page 141
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises However, at the time of Tropical Storm Jeanne the data were neither completely analyzed nor readily available for use by government or aid organizations. With the last census conducted in 1982, many relief organizations were thus dependent on population estimates available from public sources such as the United Nations (e.g., UN Population Information Network [POPIN], http://www.escwa.org.lb/popin/index.asp; the UN Population Fund [UNFPA], http://www.unfpa.org/publications/index.cfm?filterPub_Type=5; the UN Statistics Division, http://unstats.un.org/unsd/cdb/cdb_help/cdb_quick_start.asp); the World Health Organization (WHO, http://www.who.int/whosis/en/); and Save the Children (http://www.savethechildren.org.uk/foodsecurity/publications/manual.htm) or on data provided by local partners at shelters and health centers (I. Bray, personal communication, January 2007; D. Smith, personal communication, January 2007). The Haitian population was highly vulnerable prior to the onset of Tropical Storm Jeanne, and its vulnerability was compounded by the political and natural events of a particularly turbulent year. Given the needs of the country in terms of basic health care facilities and programs, infrastructure, and education, the committee finds it difficult to determine whether or not full access to georeferenced population data from a national census would have significantly improved international, U.S. government, and local agencies’ abilities to provide efficient relief aid to the affected populations. The response to Tropical Storm Jeanne was, in reality, a response to a population already in a vulnerable state from circumstances of the past year. However, the relief and development organizations involved in the response effort with which the committee spoke indicated that more and better demographic and epidemiological data, preferably georeferenced, would have enabled their activities to operate more efficiently for the affected population. Further correspondence with a UNDP official in Haiti revealed that population estimates after major disasters, such as after Tropical Storm Jeanne’s impact on the Haitian city of Gonaives, have been highly inaccurate and have affected the efficiency of subsequent interventions (E. Ergin, personal correspondence, 2006). STRENGTHS AND LIMITATIONS OF POPULATION DATA EMPLOYED IN DISASTER RELIEF AND DEVELOPMENT Digitally accessible, current, georeferenced national census data can play an important role as part of a disaster relief effort. The example from Mozambique demonstrated the strengths of access to georeferenced data linked to GIS, in a situation where the local government was empowered to make the data accessible and had the capacity in its own NSO to use and analyze the data. Although the aid response effort was not completely efficient and coordination between agencies, organizations, and the
OCR for page 142
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises Mozambique government could have been improved (Moore et al., 2003), the massive influx of international aid (more than 49 countries and 30 NGOs providing humanitarian assistance [Moore et al., 2003]) was distributed in a manner that assisted the people in need in a reasonable time period relative to the development of the disaster situation. The national data were available at a subnational level and, could be linked to thematic data including, for example, administrative boundaries, city and settlement locations, rivers and waterbodies, and flood levels and projections in GIS. Updates with new population surveys of displaced persons in various locations could thus be added in real time to these thematic maps. The national data, with additional information from the new surveys, were then incorporated into contingency plans by the INGC that were put into use when the next year’s floods arrived. No relief or preparedness planning effort is based solely on the existence of national, georeferenced census data, but the committee’s view is that the effective delivery of relief aid and preparation for subsequent flood events in Mozambique were enhanced by the existence of such data, and by the capacity of the local NSO in conjunction with the INGC and cooperating international agencies to administer them. Systematic data collection through surveys during and immediately after natural or human-induced disasters provides a reliable source of information for planning relief response quantities, types, and targets, allows the cost and time effectiveness of the response effort to be evaluated, and can be used in planning and executing recovery and development programs (Guha-Sapir et al., 2005; Van Rooyen et al., 2001; NRC, 2001). Effective use of accurate survey data can also serve to raise donors’ confidence regarding the manner in which their contributions are being used (I. Bray, personal communication, January 2007). These positive aspects of accurate survey data collection are particularly enhanced in the presence of current georeferenced census data. The repatriation program for Malian refugees in the aftermath of the Touareg rebellion illustrates the benefits of good data collection and the limitations of inaccurate survey data collection in refugee camps. Inaccuracies in the initial refugee data were identified by UNHCR and the WFP using analysis of aerial photographs, attendance at schools and clinics, and observation of inhabited dwellings. Inaccuracies, in this case overestimation of the refugee population, are not uncommon (Guha-Sapir et al., 2005; Van Rooyen et al., 2001; NRC, 2001) and can serve to delay and inhibit an effective response—smooth repatriation, in the case of the Malian refugees. Resampling and resurveying the refugee group in the case of the Malian refugees in Mauritania proved to be a worthwhile exercise that enhanced the efficiency of the repatriation of those displaced persons; as Van Rooyen et al. (2001: p. 218) indicated, “It is an unfortunate myth that statistical analysis is too time consuming to perform during a humanitarian crisis. On
OCR for page 143
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises the contrary, because of time constraints, lead agencies can’t afford NOT to perform accurate and statistically sound assessments. Without doing so, organizations will risk misplacing much-needed assistance.” The delays and obstacles to repatriation of refugees from the other asylum countries where refugee survey data remained inaccurate likely used more resources than were necessary. The UNHCR assessment of the situation indicated that placement of more trained personnel in the field in these other asylum countries could have solved this problem. A cost-benefit analysis of conducting accurate initial field demographic and epidemiological surveys in emergency situations gauged against time and resources lost when repatriation or relief efforts are executed ineffectively was beyond the committee’s scope, but could be useful for decision makers in weighing their allocation of human and material resources in response to disaster situations. Forced migration of populations as a response to disasters remains a challenge for responders, with one of the most basic difficulties being to obtain accurate numbers and characteristics of the people who have moved from their normal residences (see also NRC, 2001). Improvements in combining field surveys with developing technologies, including remote imagery and GPS, could be more actively pursued in an effort to achieve more consistent, reliable data for these important groups of people. Lack of data, or the existence of good data that are employed ineffectively, does not preclude disaster relief aid from reaching the populations in need but presents significant challenges to promoting timely, appropriate, and cost-effective relief response. While Mali and Mozambique both had georeferenced data, the response in Mali to the locust invasion was hampered by data quality (census data were old) and, importantly also, the speed with which the response to the crisis was effected. Warnings issued by the PALUCP and ULCP using the monitoring capabilities provided by early-warning networks, established through international development efforts, and appeals from the FAO for donor response issued long in advance of the locust invasion went unanswered by the international community. Early-warning systems designed specifically for this purpose had been employed correctly, but the necessary response to make the early warning systems useful did not arrive in a timely fashion. Population data quality or access in this situation was not the problem; when the data exist in a response vacuum, whatever the cause, they cannot be put to effective use. In Haiti, relief and development organizations responded immediately to the disaster produced by Tropical Storm Jeanne, but they did not have access to accurate and recent national census data. Public sources for population figures and estimates, including those from the United Nations, WHO, Save the Children, demographic and health surveys, and databases such as Gridded Population of the World (GPW) or LandScan, are viable options in these situations, and many organizations rely upon them—
OCR for page 144
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises because they do not themselves collect demographic or health data or do not have the capacity to analyze and display raw data. In the case of Haiti, the preexisting vulnerability of the population required response not just to the effects of the tropical storm but to a debilitating year and decades of economic instability. The committee suggests that resources necessary to collect national census data should not replace development programs that include education, basic health care, and food security, but that these activities should be combined with accurate information on the numbers, vital characteristics, and locations of the people that the programs intend to assist. National census data have a variety of uses outside of emergency situations, and the stabilization of Haiti’s government may lead to the implementation of more accurate and regular census coverage that ideally will inform humanitarian responses to natural disasters and development programs designed to decrease the vulnerability of the population. Accurate national population figures with demographic details are necessary for planning and executing an appropriate response for any emergency relief, recovery, or development activity. The locations of the populations affected by the disaster are also necessary to promote an effective response. National censuses are some of the basic data that may be available, but frequently subnational population counts are not part of these data sets, either because the national governments have not publicly released the data in a table as part of the statistical agency’s data releases, or because the data are not readily accessible to relief agencies and donor organizations. When such data are not available, it is difficult or impossible to estimate mortality and morbidity in the days and weeks following a disaster. CONCLUSIONS The examples from these three country analyses reinforce some of the main issues brought forward in the preceding chapters of this report: (1) national census data, available digitally, in georeferenced form, and disaggregated to a subnational level, are of basic importance to efficient aid allocation in large, multi-organizational, international disaster response situations; (2) digitally available, georeferenced national census data are also useful for effective planning, execution, and completion of disaster preparedness and long-term development projects; (3) population survey data collected before, during, and after a disaster that have been accurately collected to include basic demographic and epidemiological information are not a substitute for national census data, but are needed for effective delivery of relief aid and post-disaster recovery activities, including repatriation; (4) availability of good data alone does not ensure efficient relief
OCR for page 145
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises or development aid response because of the dependence of international, government, nongovernmental, and local organizations on efficient coordination and timely response to crises; (5) in lieu of either recent national census or local survey data, publicly available population data sets from international and private organizations, based on established population estimation methods or projections, can and ought to be employed in the distribution of aid. Lacking recent census information, these proxy data ought to be used with the understanding that they likely will not accurately represent the situation on the ground and will incorporate some inherent inefficiencies in delivery of relief or development aid. While the extent or frequency of a disaster should not alter fundamental preparedness schemes to respond to humanitarian crises, clearly the more extensive a disaster, or the higher the frequency of crises in a given region, the more people are affected, and correspondingly, the challenge to humanitarian assistance becomes greater. The presence of population and corresponding demographic data is prerequisite to providing timely, appropriate, and cost-effective response. New, geospatially referenced survey data can be acquired during a crisis, but responders would benefit from the existence of geospatially referenced disaggregated national data prior to the crisis that could serve as a baseline for coordinating the response. Data can then be more accurately revised using field surveys, supplemented by population databases and models, and remote estimates made using airborne and satellite data. Critical to employing any data is the coordination between and within local governments and international aid and donor organizations. Establishing a central government institute or body for disaster management purposes with a clear protocol for communication with foreign aid organizations is a basic element in facilitating efficient humanitarian response. The presence of a body such as the INGC in Mozambique, established specifically to manage disasters and coordinate prevention activities, including the development of a detailed hazard exposure and population information base in risk areas (see http://www.undp.org.mz/anmviewer.asp?a=22), was apparently useful, despite its relative infancy at the time of the first floods in 2000. The ability of government and agencies to enact preparedness schemes and coordinate relief efforts for subsequent emergencies was demonstrated in fairly coordinated responses during repeated flooding in successive years in Mozambique. International capacity-building endeavors, as demonstrated in the case of Mozambique prior to and following the flooding events, are important as well; in resource-poor countries, the engagement of external expertise to assist and train local people in analysis, management, and use of their own population data is critical to increasing disaster preparedness (see also Chapter 4) and decreasing vulnerability. Employing natural disaster exposure and risk assessments and increasing the capacity of local govern-
OCR for page 146
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises ments to conduct their own monitoring efforts (for any natural hazard or conflict) can increase the ability of institutions to manage humanitarian crises more effectively. RECOMMENDATIONS Based on the preceding discussion, the committee makes the following recommendations: Integrate the national statistical offices (NSOs) into the national preparedness and response teams for national emergencies. This role would involve the development of pre-disaster geospatial databases and experience in working at subnational levels relevant to hazards of all kinds. The aim is to improve the capacity of NSOs to generate and modify existing data in a timely fashion to enhance emergency response and crisis decision making. [Report Recommendation 2] Improve subnational analyses of vulnerability to natural disasters and conflict in order to delineate hazard zones or exposures where routine, periodic data collection ex ante could occur. The development of such georeferenced vulnerability analyses could help provide accountability to decision makers in preparedness and prevention and establish priorities for risk reduction investments by all stakeholders. [Report Recommendation 9] In constructing this chapter, as well as the rest of the report, the committee has been conscious of the need to balance information obtained from individual relief and development workers “on the ground” regarding their needs and desires for georeferenced population data collection and distribution with broader agency-level assessments and information from peer-reviewed articles on such data. The committee concludes that an independent, detailed study of one country, like Mali, Haiti, or Mozambique, and its use of population data in an emergency situation would be a very useful exercise. Such a study could include a statistically rigorous set of interviews with relief personnel at the field and central management levels, and in national statistical and emergency coordination offices. If conducted by a large international body with relatively good access to country-level data, results of such a study could serve to leverage the observations in this report which show that access to high-quality, geo-referenced population data can be of assistance in generating more effective disaster response.
OCR for page 147
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises REFERENCES Arndt, C., and F. Tarp, 2001. Who gets the goods? A general equilibrium perspective on food aid in Mozambique. Food Policy 26:107-109. Briguglio, L., 1993. The Economic Vulnerabilities of Small Island Developing States. Study, commissioned by CARICOM for the Regional Technical Meeting of the Global Conference on the Sustainable Development of Small Island Developing States, Port of Spain, Trinidad and Tobago, July. Doocy, S., 2007. Identify ways in which subnational demographic and geographic data and tools could be used to help decision makers provide useful information to populations at risk. In National Research Council, Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises. Washington, D.C.: The National Academies Press. Farmer, P., 2005. From Gonaives to New Orleans: Reflections of the Gulf Coast Tragedy. Partners in Health. Available online at http://www.familysystem.net/att/from%20gonaives%20to%20new%20orleans.doc [accessed October 4, 2006]. FAO (Food and Agriculture Organization of the U.N.), 2004a. Hunger in Their Wake: Inside the Battle Against the Desert Locust. Available online at http://www.fao.org/newsroom/en/focus/2004/51040/index.html [accessed October 4, 2006]. FAO, 2004b. Special Report: FAO/WFP Crop and Food Supply Assessment Mission to Mali, with Special Focus on Losses Due to the Desert Locust. Available online at http://www.fao.org/docrep/007/j3971e/j3971e00.htm [acccessed October 4, 2006]. Gall, M., 2004. Where to go? Strategic modeling of access to emergency shelters in Mozambique. Disasters 28(1):82-87. Guha-Sapir, D., W.G. van Panhuls, O. Degomme, and V. Teran, 2005. Civil conflicts in four African countries: A five-year review of trends in nutrition and mortality. Epidemiological Reviews 27:67-77. INE (Instituto Nacional de Estatística), 1999. Projecções Anuais da Populaçao, Pais Total 1997-2020: Moçambique, Estudos No. 1. Maputo: INE. Konaté, M.K., 2006. Care for Research and Training Support (CAREF); presentation to the committee at a workshop held on March 13-14 at the National Academies Keck Center, Washington, D.C. Presentation available through The National Academies Public Access Records Office. Konaté, M.K., 2007. Strengths and limitations of information and data analysis in responding to crisis in Mali. In National Research Council, Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises. Washington, D.C.: The National Academies Press. Landau, L.B., 2007. Cognitive and institutional limits on collecting and processing data on populations at risk: Preliminary reflections on Southern African responses to displacement. In National Research Council, Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises. Washington, D.C.: The National Academies Press. MDHS (Mozambique Demographic and Health Survey), 1998. Summary Report. Available online at http://www.measuredhs.com/pubs/pub_details.cfm?ID=211&PgName=country&ctry_id=61 [accessed October 4, 2006]. Moore, S., E. Eng, and M. Daniel, 2003. International NGOs and the role of network centrality in humanitarian aid operations: A case study of coordination during the 2000 Mozambique floods. Disasters 27(4):305-318. NRC (National Research Council), 2001. Forced Migration and Mortality. Washington, D.C.: National Academy Press, 145 pp. OFDA (Office of U.S. Foreign Disaster Assistance), 2000. Southern Africa—Floods Fact Sheet 19, Fiscal Year (FY) 2000, March 22, 2000. U.S. Agency for International Development,
OCR for page 148
Tools and Methods for Estimating Populations at Risk from Natural Disasters and Complex Humanitarian Crises Bureau for Humanitarian Response (BHR), Office of U.S. Foreign Disaster Assistance. Available online at http://cidi.org/disaster/00a/0122.html [accessed October 4, 2006]. PAHO (Pan American Health Organization), 2001. Country profile: Haiti. Health in the Americas: Data Updated for 2001. Available online at http://www.paho.org/english/sha/prflhai.htm [accessed October 4, 2006]. Pelling, M., and J.L. Uitto, 2001. Small island developing states: Natural disaster vulnerability and global change. Environmental Hazards 3:49-62. Regan, J., 2004. Flood-hit Haiti struggles to manage health disaster. Lancet 363(9424): 1880. UNCAP (United Nations Consolidated Appeals Process), 2005. West Africa 2005: Revision. New York, Geneva: United National Office for the Coordination of Humanitarian Affairs. Available online at http://ochaonline.un.org/cap2005/ webpage.asp?ParentID=5428&MenuID=5441&Page=1186 [accessed October 4, 2006]. UNDP (United Nations Development Programme), 2002. Human Development Report 2002: Deepening Democracy in a Fragmented World. New York: Oxford University Press. UNDP, 2004. Reducing Disaster Risk: A Challenge for Development. New York: United Nations, 146 pp. UNDP, 2005a. Country Fact Sheets: Mali. Available online at http://hdr.undp.org/statistics/data/country_fact_sheets/cty_fs_MLI.html [accessed March 14, 2007]. UNDP, 2005b. Human Development Reports Country Sheet: Mali. Available online at http://hdr.undp.org/hdr2006/statistics/countries/country_fact_sheets/cty_fs_MLI.html [accessed March 14, 2007]. UNDP, 2005c. Country Fact Sheets: Mozambique. Available online at http://hdr.undp.org/statistics/data/country_fact_sheets/cty_fs_MOZ.html [accessed October 4, 2006]. UNGA (United Nations General Assembly), 2002. Assistance to Mozambique: Report of the Secretary General (A/57/97-E/2002/76). Available online at http://www.reliefweb.int/rw/rwb.nsf/db900SID/OCHA-64C8PX?OpenDocument&rc=1&cc=moz [accessed March 14, 2007]. UNHCR (United Nations High Commission for Refugees), 1998. Review of the Mali/Niger Repatriation and Reintegration Programme. Inspection and Evaluation Service. Available online at http://www.unhcr.org/cgi-bin/texis/vtx/publ/opendoc.pdf?tbl=RESEARCH&id=3ae6bd488&page=publ [accessed October 4, 2006]. UNICEF (United Nations Children’s Fund), 2000. Mozambique Flood Relief. Available online at http://www.unicef.org/evaldatabase/index_14126.html [accessed October 4, 2006]. Van Rooyen, M.J., S. Hansch, D. Curtis, and G. Burnham, 2001. Emerging issues and future needs in humanitarian assistance. Prehospital and Disaster Medicine 16(4):216-222. WFP (United Nations World Food Programme), 2005. Mindful of Niger, WFP Warns of a Potential Food Crisis in the Sahel. WFP Newsroom, August 5. Available online at http://www.wfp.org/English/?ModuleID=137&Key=1359/ [accessed October 4, 2006].
Representative terms from entire chapter: