Department of Civil Engineering and Surveying University of Puerto Rico Mayagüez Campus Formulation of econometric solutions of immediate resources for a disaster area in Mayagüez, Puerto Rico Coral A. Torres Cruz September 14, 2006 Dwight David Eisenhower Fellowship 1
Table of Contents 1. Introduction 4 2. Location. 7 3. Objectives.. 8 4. Methodology.. 8 5. Project Task.. 9 6. References. 10 2
Abstract Puerto Rico is an island that is very susceptible to different types of natural disaster do to its location. The ones that most affect the island are hurricanes, floods, mudslides and earthquakes. Since the 1950 s up to 2005, there has been 20 major disaster declarations made by FEMA. What would happen if a natural disaster of enormous proportions happens? Would the pertinent authorities be prepared for assisting the population? It is important that these authorities have the appropriate logistics to response correctly in case there is a disaster. They should know with accuracy what are the immediate resources needed in different areas. This was the case of for example Hurricane Katrina in witch approximately 1,800 people died because FEMA, Red Cross and the state and local government did not have the logistics capacity to fully support the massive number of Gulf coast victims, according to a Congressional Investigation. The purpose of this study is to provide robust estimates of the dynamic needs in a disaster area using econometric tools. This way new methodologies and tools to foster an accelerated convergence between the dynamic needs and supplies of critical resources in the aftermath of an extreme event can be provided to the agencies involved in the response of the after math of a natural disaster. In order to complete this study data must be acquired threw interviews to directors of agencies that distribute the resources necessary on the affected areas. Also interviews to people on the area of study (Mayagüez, Puerto Rico) will be conducted to obtain information about the immediate needs they found more relevant during a disaster. All this information will be used to make models that will provide the demand of different resources on different periods of time and different types of disasters. 3
Introduction Puerto Rico is an island that is very susceptible to different types of natural disaster do to its location. Around 60 hurricanes and 3 mayor earthquakes affected the island the past century. This island is also susceptible to tsunamis on the coastal areas and mudslides on its central region. The Mayagüez area, specifically, is very vulnerable to different types of natural disasters because of its location. It is a costal area which is at risk of tsunamis (Fig. 2) and marine tornados. Around 80% of its urban area is located on flooded zones and approximately 70 inches of rain fall every year. It has been affected by some of the biggest hurricanes that have hit Puerto Rico (Fig. 1) and in October 1918 one of the biggest earthquakes registered in Puerto Rico (7.3 Richter scale), caused a tsunami that destroyed half of the city. Because of these events Mayagüez was the first city in Puerto Rico to have a tsunami warning system. But sometimes, natural disasters are so unpredictable or are of such enormous proportion that catastrophe is impossible to prevent. The best example is Hurricane Katrina, witch caused around 1,800 deaths on the United States only and left more than $105 billion in damages. The government though that they had all the precautions necessary for an event like this, however catastrophe stroke anyway. What would happen if an event like this strikes Puerto Rico? Would the pertinent authorities respond adequately? Will the needs of people be attended in the right ways? Fig. 1, Hurricane George 4
The purpose of this study is to provide robust estimates of the dynamic needs in a disaster area using econometric tools. This way new methodologies and tools to foster an accelerated convergence between the dynamic needs and supplies of critical resources in the aftermath of an extreme event can be provided to the agencies involved in the response of the after math of a natural disaster. Fig. 2, Map of tsunami vulnerable areas This is a very important task because many times the resources available are distributed in the wrong ways. In some cases, the available resources are distributed to areas that are the least affected an in other cases the resources can t be distributed because the areas can t be reached because of the destruction. 5
To be able to provide these estimates an econometric function must be develop. The immediate resource requirements at time t (e.g., blood, water), Z(t), are likely to be a function of the characteristics of the extreme event itself, X(t), (e.g., type of event, event size, duration), and the vector of characteristics of the impacted area, Y(t), e.g., population density. = Ω( X ( t), Y ( t) ) + ε ( ) (1) Z( t) t The research problem focuses on the specification, development, and estimation of the function Ω in equation (1). In order to obtain these coefficients, data of before, during and after a disaster must be collected from the agencies responsible of functioning during a disaster as well as official records from the past events. Some of the agencies are FEMA, Agencia Estatal para el Manejo de Emergencias, Red Cross, etc. This is the topic that will be studied specifically. At any given time t, the dynamic supply of critical resources to the impacted site, S(t), is comprised of three basis components: (1) the vector of available resources on site, W(t); (2) the vector of non-solicited donations, N(t); and (3) vector of flows of critical resources provided by emergency response agencies, R(t). At any given time t: S(t) = W(t) + N(t) + R(t) (2) The proposed approach is to combine both models, optimization and feedback to improve the various stages of emergency response and recovery. Models are needed to aid in understanding the effect of disturbances and for the decisions on purchase and transportation of critical supplies; Optimization techniques are needed to make decisions that result in minimizing the objective function(s); and Feedback is needed to compensate for errors and uncertainty, update models and revise decisions about resource allocation. 6
Location Fig. 3, Mayagüez Area and its suburbs Mayagüez is the fifth-largest city and is considered one of the most important cities in the island. Mayagüez is located in the west coast, north of Cabo Rojo, Hormigueros and San Germán; south of Añasco; and west of Las Marías and Maricao. It has an area of about 76 square miles and a population of 105,000. Also, Mayagüez has become a major college town it host the third biggest campus of the University of Puerto Rico, Catolica University and Colegio de Abogados. It is also home of the only Zoo in the island, Juan A. Rivero Zoo (commonly known as Mayagüez Zoo), home of numerous birds, mammals and reptiles. The city can be access by car threw road PR-2 from South or North, PR-65 from Las Marias and PR-64 from Añasco. Also Mayagüez has its own airport, Eugenio María de Hostos - Carr. /Rd. 2 Barrio Maní. It is 3 miles North from Mayagüez. Its total area is 2 acres. It offers services for freight (cargo) transportation as well as passenger transportation. Mayagüez also has its Port of Entry in which accept entries of merchandise and commodities. It s also a point of connection between Puerto Rico and Dominican Republic through its ferry service. 7
Objectives The purpose of this study is to provide robust estimates of the dynamic needs in a disaster area using econometric tools. The fundamental purpose of the research proposed is to determine the most efficient method to close the gap between immediate resources and dynamic supplies considering supply priorities and donations needed at the disaster site. In the absence of formal decision support tools, with limited or no feedback control provided, a fairly inefficient flow of critical supplies may result and, as a result due to the lack of optimal feedback coupled with dynamic estimation of the available resources, the convergence between cumulative needs and cumulative supplies of critical resources is likely to take longer than the optimal. Methodology In order to create models that would optimize the distribution of resources during a disaster, data must be acquired to be implemented on the models. The first step is to make interviews to the directors of the different agencies involved to obtain data of how each agency response on different types of events, specifically in terms of dynamic supplies. To obtain data for the immediate resources needed a survey to a sample of the population of Mayagüez must be done. Also data from disasters on the area will help for this purpose. All the data obtained from the previous steps will be analyzed and characterized for further use. Then econometric models will be done using the analyzed data. Because a great number of models will be obtained, optimization analysis will be used to obtain the best result. 8
Project Task Literature revision Agencies Interviews Survey Data analysis Task Sept. Oct. Nov. Dec. Jan Feb. March April May Descriptive Analysis Development of multiple linear regressions 9
References 1. An econometric model for international tourism flows to Spain; Teresa Garín Muñoz; Teodosio Perez Amaral; Applied Economics Letters, 2000, 7, 525± 529. 2. Discrete Choice Models in Transportation: Proof of Concept; Joan L. Walker; January 2005. 3. Incorporating Transportation Network Structure in SpatialEconometric Models of Commodity Flow; James P. LeSage; Wolfgang Polasek; May 2006. 4. Mayagüez; http://www.topuertorico.org/city/mayag.shtml 5. Junta de Planificación; http://www.jp.gobierno.pr 6. Hurricane Katrina; http://en.wikipedia.org/wiki/hurricane_katrina 10
11