FLOOD LOSS ESTIMATE MODEL - Southern Illinois University

Copy and paste this link to your website, so they can see this document directly without any plugins.



Keywords

flood, that, from, damage, were, damages, Gonaives, Figure, with, Flood, Table, which, loss, land, area, water, such, Detail, Quinte, have, Haiti, This, population, model, economic, estimate, into, Damage, depth, rural

Transcript

FLOOD LOSS ESTIMATE MODEL: RECASTING FLOOD DISASTER ASSESSMENT AND
MITIGATION FOR HAITI, THE CASE OF GONAIVES
By Guetchine Gaspard
B.S., Faculté d’Agronomie et de Médecine Vétérinaire,
University of Haiti State
Haiti, 2006
A Thesis
Submitted in Partial Fulfillment of the Requirements for the
Master of Science Degree
Department of Geography and Environmental Resources
In the Graduate School
Southern Illinois University Carbondale
August 2013
THESIS APPROVAL
FLOOD LOSS ESTIMATE MODEL: RECASTING FLOOD DISASTER ASSESSMENT AND
MITIGATION FOR HAITI, THE CASE OF GONAIVES
By Guetchine Gaspard
A Thesis Submitted in Partial
Fulfillment of the Requirements
for the Degree of
Master’s of Science
in the field of Geography and Environmental Resources
Approved by:
Dr. Tonny J. Oyana, Chair
Dr. Guangxing Wang
Dr. Samuel Adu-Prah
Graduate School
Southern Illinois University Carbondale
July 1, 2013
i AN ABSTRACT OF THE THESIS OF
Guetchine Gaspard, for the Master of Science degree in Geography and Environmental
Resources, presented on May 17
th , 2013, at Southern Illinois University Carbondale.
TITLE: FLOOD LOSS ESTIMATE MODEL: RECASTING FLOOD DISASTER
ASSESSMENT AND MITIGATION FOR HAITI, THE CASE OF GONAIVES
MAJOR PROFESSOR: Dr. Tonny J. Oyana
This study aims at developing a model to estimate flood damage cost caused in Gonaives,
Haiti by Hurricane Jeanne in 2004. In order to reach this goal, the influence of income,
inundation duration and inundation depth, slope, population density and distance to major roads
on the loss costs was investigated. Surveyed data were analyzed using Excel and ArcGIS 10
software. The ordinary least square and the geographically weighted regression analyses were
used to predict flood damage costs. Then, the estimates were delineated using voronoi
geostatistical map tool.
As a result, the factors account for the costs as high as 83%. The flood damage cost in a
household varies between 24,315 through 37,693 Haitian Gourdes (approximately 607.875
through 942.325 U.S. Dollars). Severe damages were spotted in the urban area and in the rural
section of Bassin whereas very low and low losses are essentially found in Labranle. The urban
area was more severely affected by comparison with the rural area. Damages in the urban area
are estimated at 41,206,869.57USD against 698,222,174.10 17,455,554.35USD in the rural area.
In the urban part, damages were more severe in Raboteau-Jubilée and in Downtown but BigotParc Vincent had the highest overall damage cost estimated at 9,729,368.95 USD. The lowest
cost 7,602,040.42USD was recorded in Raboteau. Approximately, 39.38% of the rural area
underwent very low to moderate damages. Bassin was the most severely struck by the 2004
floods, but Bayonnais turned out to have the highest loss cost: 4,988,487.66 USD. Bassin along
with Labranle had the least damage cost, 2,956,131.11 and 2,268,321.41 USD respectively.
ii Based on the findings, we recommended the implementation and diversification of
income-generating activities, the maintenance and improvement of drains, sewers and gullies
cleaning and the establishment of conservation practices upstream of the watersheds. In addition,
the model should be applied and validated using actual official records as reference data. Finally,
the use of a calculation-based approach is suggested to determine flood damage costs in order to
reduce subjectivity during surveys.
iii
DEDICATIONS
This work is dedicated to the people of Haiti who struggle everyday against socioeconomic, political, and environmental problems. I am grateful to God as he guides me through
every step that I take. To my relatives, my friends and my family, I offer this work. May it
contribute to make them feel happier.
iv AKNOWLEDGEMENTS
I thank the professors at the Faculté d’Agronomie et de Médecine Vétérinaire of the Haiti
State University for having provided me with strong science. I owe them a lot of my
methodological and statistical skills among others. The same way, I thank the professors at the
Geography and Environmental Resources Department for making me a spatial analyst. I am
especially thankful to Dr. Tonny Oyana who has provided the bulk of what I will have gained
from the Southern Illinois University Carbondale.
v TABLE OF CONTENTS
CHAPTER PAGE
ABSTRACT ..................................................................................................................................... i
DEDICATIONS ............................................................................................................................. iii
AKNOWLEDGEMENTS.............................................................................................................. iv
LIST OF TABLES ........................................................................................................................ vii
LIST OF FIGURES ..................................................................................................................... viii
ABREVIATIONS AND ACRONYMS ......................................................................................... ix
CHAPTER 1 ................................................................................................................................... 1
INTRODUCTION .......................................................................................................................... 1
1.1 Background ........................................................................................................................... 1
1.2 Justification ........................................................................................................................... 2
1.3 Problem Statement ................................................................................................................ 2
1.4 Research Questions ............................................................................................................... 3
1.5 Definition of Terms ............................................................................................................... 3
CHAPTER 2 ................................................................................................................................... 6
LITERATURE REVIEW ............................................................................................................... 6
2.1 Damages in Gonaives ............................................................................................................ 6
2.2 Flood Hazard Assessment ................................................................................................... 10
2.3 Flood Loss Modeling .......................................................................................................... 11
2.3.1- Procedure in Flood Loss Modeling ........................................................................................... 12
CHAPTER 3 ................................................................................................................................. 19
METHODS ................................................................................................................................... 19
3.1 Study Area ........................................................................................................................... 19
3.1.1 Geography .................................................................................................................................. 19
3.1.2.- Soil Properties .......................................................................................................................... 20
3.1.3.- Water Resources ...................................................................................................................... 23
3.1.4.- Economy .................................................................................................................................. 24
3.1.5.- Land Tenure ............................................................................................................................. 26
vi 3.2 Data and Methods................................................................................................................ 26
3.2.1 Sampling Technique .................................................................................................................. 31
3.2.2 Profile of the Places ................................................................................................................... 31
3.2.3 Survey ........................................................................................................................................ 33
3.2.4 Model ......................................................................................................................................... 33
CHAPTER 4 ................................................................................................................................. 37
RESULTS AND DISCUSSIONS ................................................................................................. 37
4.1 Preliminary Results ............................................................................................................. 37
4.1.2. Model Analysis ......................................................................................................................... 37
4.1.2 Spatial Statistics Results ............................................................................................................ 40
4.2 Model Reevaluation ............................................................................................................ 40
CHAPTER 5 ................................................................................................................................. 55
CONCLUSION AND IMPLICATIONS ...................................................................................... 55
REFERENCES ............................................................................................................................. 58
VITA ............................................................................................................................................. 64
vii
LIST OF TABLES
TABLE PAGE
Table 1- Different Types of Flood Damages………........………………………..……………...11
Table 2-Direct Damage Categories, Measurement Units, Maximum Damage Costs and Data....14
Table 3. Summary of Existing Flood loss Estimation Methodology………….……………...….16
Table 4: Comparison of Different Flood Loss Models ………………………………………….17
Table 5. Cropping Calendar in the Quinte River Watershed (without irrigation)….…………....25
Table 6 : Cropping Calendar in the Quinte River Watershed (with irrigation)…………....….....25
Table 7. Definition of Variables Used in the Regression Models…………………………….....30
Table 8. Distribution of Buildings and Households for Gonaives in 200…………………….….32
Table 9. Age Distribution of the Population of Gonaives by Sex Percentage in 2009 (IHSI,
2009)……………………………………………………………………………………………..32
Table 10. Correlation Matrix of the Variables……………………………………………...……37
Table 11. Statistical Results for the Initial OLS Model………………………..………………...40
Table 12. Statistical and Spatial Autocorrelation Results for the OLS and GWR Models……...45
Table 13. Flood Damage Cost Percentages in the Urban and Rural Areas……………………...52
Table 14. Distribution of Flood Damage Costs in the Urban Area……………………………...53
Table 15. Distribution of Flood Damage Costs in the Rural Area……………………….………53
viii
LIST OF FIGURES
FIGURE PAGE
Figure 1. Flooding in Gonaives Following the passage of Hurricane Jeanne in 2004…..……..…8
Figure 2. Flooding in Gonaives Haiti before and after Hurricane Jeanne 2004………….....….....9
Figure 3. Map Showing the Location of the Municipality of Gonaives.…………..…...………..20
Figure 4. Land Use and Land Cover for the Municipality of Gonaives…………………...…….22
Figure 5.- Soil Erosion Risk Map for Gonaives ……………………………...…………………23
Figure 6. Map showing the watershed in Gonaives ……………………………………………..24
Figure 7. Processing and Preparation of the Variables……...……………………………...……29
Figure 8. Flood Damage Cost Modeling……...………………………………………………….30
Figure 9. The Variation of the Flood Loss Cost………...……………………………………….38
Figure 10. The Variation among the Independent Variables……………………….……..….….39
Figure 11. Moran’s I test result for the OLS Model……...……………………………………...42
Figure 12. Variation of Income, Slope and Density at Global and Local levels…………..…….44
Figure 13. This Map Shows the Local R
2
in the GWR Model……….……………………….…47
Figure 14. T-statistics of the GWR Model……………………………………………………….48
Figure 15. The Intercept, Income, Slope and Density Coefficients in the GWR Model….……..49
Figure 16. Interpolation of Flood Cost across Gonaives…………………………….….……….50
Figure 17. Local Damage Cost Variation Depending on Income, Slope and Density….…….....54
ix ABREVIATIONS AND ACRONYMS
B.S. : Bachelor of Science
CDMP : Caribbean Disaster Mitigation Project
CIMH : Caribbean Institute for Meteorology and Hydrology
FAO : Food and Agriculture Organization
FEMA : Federal Emergency Management Agency
GIS : Geographic Information System
GWR : Geographically Weighted Regression
H T G : Haitian Gourdes
IDW : Inverse Distance Weighted
IHSI : Institut Haïtien de Statistiques et d'Informatiques (Haitian Statistics and
Informatics Institute)
LfUG : Landesamt für Umwelt und Geologie
NDVI : Normalized Difference Vegetation Index
OAS : Organization of American States
OLS : Ordinary Least Square
OXFAM : Oxford Committee for Famine Relief
UK : United Kingdom
UNISDR : United Nations International Strategy for Disaster Risk
U.S. : United States
USAID : United States Agency for International Development
USDE : Unit of Sustainable Development
UTSIG : Unité de Télédétection et de Systèmes d’Information Géographique (Remote
Sensing and GIS unit).
VIF : Variance Inflation Factor
WFP : World Food Program (PAM)
1
CHAPTER 1
INTRODUCTION
1.1 Background
Direct damages associated with Hurricane Katrina hitting New Orleans in 2005 was
estimated at ninety billion dollars. Sixteen billion of this total cost was caused by flooding to
residential property (Jonkman, 2008). Prior to the passage of Katrina, Hurricane Jeanne floods
killed more than 2800 people as recently as 2004 in Gonaives, Haiti (Colindres, 2009). In fact, it
makes no doubt that flooding is an expensive and deadly global issue that needs to be accounted
for in risk management.
Due to Haiti’s geographical location and topographical characteristics, flooding from
hurricanes and storms is quite frequent. From 1968 to 2001, the island has been affected by 30
hurricanes and 90 floods due to heavy precipitation. The disasters are not totally of natural
origin. In fact, they are aggravated by anthropogenic activities. With oil too expensive for the
bulk of the population, charcoal from burnt trees has provided at least 85% of energy in Haiti for
decades. As a result, many citizens have relentlessly hunted and chopped down huge amount of
forest. In addition to deforestation and disasters, hoed crops are being practiced for more than 20
to 30% in areas that are not fit for that type of agriculture. Not to mention that 2.5 million cubic
meters of stones are extracted from the ground annually for construction, leaving denuded
mountain slopes that rainwater washes down unimpeded. Twenty-five out of thirty watersheds in
the country are degraded. That situation has a huge impact on agriculture, and commerce and has
severely affected the communities and the poor, in particular. Catastrophic floods are considered
a barrier to sustainable development especially because the less developed a country is, the more
prone it is to economic damages and loss of life. Haiti is identified as a country where there is a
strong link between the level of development and the impact of natural hazards. Effective flood
risk management plans consisting of forecasting and warning systems and plans for evacuation
and relief and post-flood recovery can substantially reduce losses.
2
The present study focuses on the area of Gonaives, which has been greatly damaged by
the 2004 and 2008 flood events. The goal of the study is to quantify the flood damages to the
city and the surrounding areas. This will help in planning and designing flood protection
measures, managing emergency in real-time and recovering from flood. Countries such as
Germany, Australia, United Kingdom, United States and others have conducted similar studies.
Software tools such as ArcGIS 10 and Microsoft Excel were used. This study also provides some
guidelines for flood risk mapping in developing countries.
1.2 Justification
Previous efforts to assess disaster risk in Haiti have not been extensive. According to
Felhome (2007), there have been three major efforts in hazard mapping in Haiti: an island wide
seismic map done by OAS/USDE/CDMP, an island wide atlas of probable storms effects
prepared by OAS/USDE/CDMP/CIMH, and a national multihazard map produced by OXFAM.
Some vulnerability assessment projects have been undertaken, and few mapping efforts have
been made. This includes, for instance, UTSIG (2004), FewsNet (2005) and USAID (2007). The
scale at which information is gathered is important to decision makers. Broad scale studies often
result in generalization of the spatial dimensions of risk and vulnerability, with minimization of
their complexity and variability.
To date, few studies have quantified flood risk at a useful scale in Haiti. Recent floods in
the city of Gonaives and other parts of the country such as Mapou and Fond Verrettes have
increased community, governmental and international awareness of the need to do more research
at a local level. The development of a flood loss estimation model will serve several major
applications. Examples include but are not limited to: planning and designing flood protection
measures, managing emergency in real-time and recovering from floods.
1.3 Problem Statement
Certain regions of Haiti are under a permanent threat of flooding. Gonaives, Fonds
Verrettes, Port-au-Prince and Léogane are among the areas that have particularly been hit several
times, particularly 2004 and 2008 by massive floods. The location of the Island on the hurricane
path in the Atlantic Ocean, the artificial degradation of watersheds in the upper land combined
3
with urban sprawling and land use in the bottom land amplify the vulnerability of the weakened
population to respond to the impact of a flood hazard, not to mention steep slopes in the upper
and heavy rain falls that characterize the tropical climate of Haiti. The purpose of this study is to
develop a model to estimate residential damages resulted from the floods caused by Hurricane
Jeanne in 2004. Also, flood damage maps are created showing the distribution of damages for
residential areas in the Quinte River watershed.
1.4 Research Questions
The study aims at answering two specific questions:
1. What are the damages in a residential household with an inundation depth ranging from
10 to 200 centimeters?
2. What are the flood losses for a farm household surrounding the city of Gonaives?
1.5 Definition of Terms
Disaster
According to UNISDR 2004:3, disaster can be defined as: “a serious disruption of the
functioning of a community or a society causing widespread human, material, economic or
environmental losses which exceed the ability of the affected community or society to cope using
its own resources.”
Vulnerability
“The conditions determined by physical, social, economic and environmental factors or
processes, which increase the susceptibility of a community to the impact of hazards”. Not to
mention that “these hazards might originate from the natural environment, such as droughts,
floods or sinkholes or may be anthropogenic in nature, for example nuclear meltdowns, pollution
or terrorism” (Riet, 2009).
Capacity
Capacity is defined as “A combination of all strengths and resources available within a
community, society or organization that can reduce the level of risk or the effects of a disaster”
(Riet 2009).
4
Disaster Risk
“The probability of harmful consequences, or expected losses (death, injuries, property,
livelihoods, economic activity disrupted or environment damaged) resulting from interactions
between natural or human-induced hazards and vulnerable conditions”, (UNISDR, 2004:3).
Disaster Risk Reduction
The conceptual framework of elements considered with the possibilities to minimize
vulnerabilities disaster risks throughout a society, to avoid (prevention) or to limit mitigation and
preparedness) the adverse impacts of hazards, within the broad context of sustainable
development.
Risk and Flood Mitigation
According to the Federal Emergency Management Agency (FEMA), mitigation is the “effort to
reduce loss of life and property by lessening the impact of disasters.” The agency continues to
explain that “this is achieved through risk analysis, which results in information that provides a
foundation for mitigation activities that reduce risk, and flood insurance that protects financial
investment.”
Tangible Damages
Damages caused by any natural disasters are broadly classified into two categories; they are
tangible damage and intangible damage. Tangible damages are those which can be evaluated
quantitatively in economic terms such as, damage to lifelines, buildings, etc.
Intangible Damages
Intangible damages are ones that are difficult to express in economic value, for example anxiety,
mental tremor to victims, inconvenience and disruption of social activities, etc. In case of flood
damage, both tangible and intangible damages can be of two types, direct and indirect damages.
5
Direct and Indirect Damages
Direct damages are caused by physical contact of floodwater. Indirect flood damages are those
caused through interruption and disruption of economic and social activities as a consequence of
direct flood damages. Direct and indirect damages can be subdivided into primary and secondary
categories. Commonly, primary and secondary direct as well as primary indirect damages are
evaluated in monetary currencies using both survey procedures which consist in interviewing
affected populations and stages-damaged functions where parameters like inundation and
duration are taken into account. Those functions are normally from analysis of past flood data or
description of flood damage ratio at a given depth and duration.
6
CHAPTER 2
LITERATURE REVIEW
The literature for the study is organized into four sections which are damages in
Gonaives, flood hazard assessment, flood modeling and flood loss estimate components. In the
first section, information about flood damages in Gonaives is reported whereas the process of
flood assessment is detailed in the second section. Section three emphasizes the main focus of
the study which is flood damage modeling as well as its different steps. Finally, the components
of the current flood estimation model are presented in the fourth section.
2.1 Damages in Gonaives
In September 2004, the Caribbean region as well as the east cost of the United States was
struck by Hurricane Jeanne. Few losses of life were recorded in the Dominican Republic, in
Puerto Rico and in Barbados. Essentially, Haiti was the most damaged country by far with
important damages and considerable losses of life. According to the Global Security website, as
many as 34 people died for all the counties mentioned combined excluding Haiti (2004).
Oppositely, approximately 300,000 people were affected by flooding and heavy rains. In fact, it
was officially estimated and reported that the country lost over 3006 souls of which 2826 died in
Gonaives. Other sources provided slightly different numbers for Gonaives where most damages
were recorded. The casualties in Gonaives amounted to 2800 (Colindres, 2007). The same
website reported that 80% the people living in Gonaives was affected. Furthermore, 35,000
houses were affected of which nearly 5000 were destroyed or damaged. According to the same
source, almost all 397 elementary and secondary schools were damaged and closed. The main
public hospital serving the area was damaged and became definitively non-operational. In
parallel, about 70 percent of the region’s agricultural areas were damaged.
According to a report of the Haitian Institute of Informatics and Statistics (IHSI, 2009),
the urban population of Gonaives was particularly affected. The Centreville, literally
“Downtown,” and Kasoley were the most affected among the five zones that make up the urban
area. Their population plummeted by 31% from 2003 to 2009. On the contrary, the population of
the Biénac-Gathereau area increased by 26%. At the same time, the number of the urban
households decreased by 46% against 8.7% in the rural area. That decrease is higher for the
Downtown and Kasoley areas with 55.8% and 60.6% respectively.
7
The Institute categorized the damages to houses on a severity scale of four: very severe,
severe, minor and none. In the urban area, the Raboteau-Jubilée has the highest percentage
(3.3%) in the very severe class whereas Biénac-Gathereau and Kasoley have the lowest (2.1%
and 1.1% respectively). For the severe category, Bigot-Parc Vincent has the highest percentage
(10.7%) whereas Kasoley and Biénac-Gathereau have the lowest (5.3% and 5.7%) respectively.
Biénac-Gathereau was the least damaged among the urban areas since it has the lowest
percentage (3.7%) for the minor damage category and the highest (16.0%) for the none damage
category. Also, it was noticed that 58% of houses in Biénac-Gathereau and 26.4% of those in
Raboteau underwent no damage.
The rural area was less affected by comparison with the urban area. In fact, 43% of the
houses underwent no damage against 30% in the urban area (IHSI, 2009). Among the five rural
subdivisions, Pont Tamarin was the most damaged with the highest percentages for both the very
severe and the severe categories, 3.5% and 8.7% respectively. Labranle had the lowest
percentages for the same categories, 1.4% and 2.4% respectively. The pictures in Figure 1 testify
of the damages that Gonaives underwent due to the passage of Hurricane Jeanne in September
2004. The two pictures at the top were taken for the rural area. They show the immersion of the
National Route 1, leading to the city, and an airstrip terminal. The pictures at the bottom show
damages in the urban area. The ground level was washed and people moved to the tops of houses
to live. Figure 2 presents two pictures of the Downtown of Gonaives city taken before and after
the September 2004 floods. The pictures show that the roads and the surrounding infrastructures
were severely damaged.
8
Figure 1. Flooding in Gonaives Following the Passage of Hurricane Jeanne in 2004
(Globalsecrity.org , 9/21/2004). These pictures were taken on Tuesday September 21
st , two days
after the rain had started, Saturday night and Sunday. The first picture shows the road that leads
to town at the top and a terminal at the lower part covered by water. The third and fourth pictures
are town pictures. The latter shows that people moved to the tops of houses to live.
9
Figure 2. Flooding in Gonaives Haiti before and after Hurricane Jeanne 2004 (NASA Earth
Observatory, posted on 03/25/2011:4:30pm)
10
2.2 Flood Hazard Assessment
Inform contingency planning, reduce vulnerability and identify high risk area are three
reasons to make a risk assessment (Riet, 2009). Flood damage assessment deals with measuring
quantitative, economic, and qualitative damages. In general, five conditions are required in a
flood damage assessment methodology. First, an accurate and efficient prediction of flood
inundation is performed. This defines the spatial scope of flood damage. This engineering factor
unrelated to the economic factor is required to guarantee results’ reliability. Second, an accuracy
and precision assessment in surveys of land use and assets in the damaged area is performed.
This represents the severity of potential damages in the target area and defines whether the
assessed damages can describe the characteristics of the area. Third, reasonable information on
the susceptibility of assets (or depth–percent damage relationships) is analyzed. This defines the
percent of total value of assets damaged for a range of flood inundations with respect to
structures, personal properties, and other items. This element is crucial in relating condition 2
with condition 1. Finally, generality and convenience of analysis are considered. As flood
damage assessment is utilized in the economic analysis of various flood damage reduction
projects, the methodology used must be universal and convenient to use.
Gathering hydrologic data directly from rivers and streams is a valuable but timeconsuming effort. If such dynamic data have been collected for many years through stream
gauging, models can be used to determine the statistical frequency of given flood events, thus
determining their probability. However, without a record of at least twenty years, such
assessments are difficult. In many countries, stream-gauging records are insufficient or absent.
As a result, flood hazard assessments based on direct measurements may not be possible because
there is no basis to determine the specific flood levels and recurrence intervals for given events.
Hazard assessments based on remote sensing data, damage reports, and field observations can
substitute when quantitative data are scarce. They present mapped information defining floodprone areas which will probably be inundated by a flood of a specified interval.
11
2.3 Flood Loss Modeling
Nowadays, flood loss modeling has gained more attention in risk analysis and risk
management. It is a very challenging and complex task requiring the understanding of dynamics
and the working of each and every sector concerned with flood events. Often time, researchers
develop damage estimate models for specific categories such as commerce (Kreibich, 2010),
agriculture (Tapia-Silva, 2011), loss of life (Jonkman, 2008) and so forth. The strategy of
modeling one sector is logical given the difficulty of one single model to account for all the
aspects impacted by flooding. Such a reality leads to a situation where models encountered in
literature are flawed and unpractical assuming that any different aspect from that of a model is
stable. In fact, literature categorizes flood damages into direct and indirect and also into tangible
and intangible damages (Dutta, 2001). That classification is commonplace. However, the
interpretation of what is a direct damage and what is not sometimes differs from authors
(Jonkman, 2008). That classification is based on two distinctions in a sense that direct damages
are located in the flooded area while indirect are outside. Also, tangible damages are those that
can be priced as opposed to intangible for which there are no market prices (Table 1).
Table 1- Different Types of Flood Damages
Tangible and priced Intangible and unpriced
Direct  Residences  Fatalities
 Capital assets and inventory  Injuries
 Business interruption (inside the flooded area)  Inconvenience and moral
damages
 Vehicles  Utilities and communication
 Agricultural land and cattle  Historical and cultural losses
 Roads, utility and communication infrastructure  Environmental losses
 Evacuation and rescue operations
 Reconstruction of flood defences
 Clean up costs
Indirect  Damage for companies outside the flooded area  Societal disruption
 Adjustments in production and consumption
patterns outside the flooded area
 Psychological traumas
 Temporary housing of evacuees  Undermined trust in public
authorities
Source: (Jonkman, 2008)
12
2.3.1- Procedure in Flood Loss Modeling
In literature, the vast majority of flood loss estimate models have been developed for
physical direct damages. Most of the time, indirect intangible damages are neglected for several
reasons such as lack of consistent data. Jonkman is one of the rare researchers having attempted
to develop an integrative flood loss estimate model including an indirect economic component as
well as a life loss component. Generally, direct damages related to physical impacts are
estimated by stage-damage functions or curves. Those functions are based on the relationship
between the flood features, usually depth, and the economic damages. The first step of the
process is the estimation of the structural damages to objects while the second step is the pricing
of those damages. Stage-damage functions were primarily developed in the 1960’s in the United
States by Gilbert F. White and Robert Kates and later spread across the world. Other researchers
tried, later on, to include other parameters such as flood duration and depth, contamination, and
preparedness (Kreibich et al., 2010). In simulation, flood characteristics like depth, flow
velocity, and duration can be derived from a hydrodynamic model. Those attributes coupled with
land use data as well as the application of stage-damage functions help estimate direct physical
damages. The next section provides more details about the process to derive flood loss estimates
for direct physical damages.
2.3.1.1 Direct Physical Damage
The process of estimating physical damages includes three aspects. These are
determining flood characteristics, gathering information on land use and land cover data, and on
maximum damage amounts, and utilizing the so-called stage-damage functions or curves.
Flooding patterns are based on the simulated model of SOBEK 1D-2D. The hydrodynamic
model shows how a flow of water issued from a breach impacts land use. Mathematically, the bidimensional model is based on the Saint Vincent equation which requires input data on the area,
location of breaches, and height and duration characterizing the hydraulic load. One must
account for roughness and geometry of the surface in simulation. A lack of that may not only
result in compartment of the flooded area but also in a changed flow area induced by obstacles.
The overtime output of such models based on one or multiple breaches scenarios is the water
depth, the velocity as well as the rise rate of the water. These output parameters can all be
13
described on a map and consequently be linked to economic damages. In a direct physical
damage assessment, five major types of assets at risk are identified. These are land use and land
cover, infrastructure, households, companies, public utilities and facilities. The determination of
direct physical damages is done by means of a maximum damage amount for one object. The
assumption is that the value of one object is the same countrywide for that object as presented in
Table 2. In such a procedure, site realities or regional specificities are not taken into account.
According to Jonkman 2008, the following equation is used to estimate direct physical damages
in a flooded area:
  nDh ririri
n r m i D
,,max (Equation 1)
Where Dmax,i is the maximum damage amount for an object or the land use category i, i is the
damage or the land use category, r is the location in the flooded area, m is the number of damage
categories, n is the number of locations in flooded area, hr is the hydraulic characteristics of the
flood at a particular location, αi(hr) is the stage-damage function that expresses the fraction of
the maximum damage for category i as a function of flood characteristics at a particular location
r (0≤αi(hr)≤1) and ni,r is the number of objects of damage category i at location r.
14
Table 2-Direct Damage Categories, Measurement Units, Maximum Damage Costs and
Data
Damage
category
Damage
sub-category
Measurement unit Maximum
direct
damage
amount (€)
Data source
Land use Agriculture m
2
2 CBS land use
Greenhouses 40 CBS land use
Urban area 49 CBS land use
Intensive recreation 11 CBS land use
Extensive recreation 9 CBS land use
Airports 1230 CBS land use
Infrastruct
ure
Motorways M 2100 National Road Database
Major roads 980 National Road Database
Other roads 270 National Road Database
Railways 25,000 Rail-NS
Households Low-rise housing Object 172,000 Bridgis dwelling types
Middle-rise housing 172,000 Bridgis dwelling types
High-rise housing 172,000 Bridgis dwelling types
Single-family houses 241,000 Bridgis dwelling types
Farms 402,000 Bridgis dwelling types
Vehicles 1050 Manual input combined
with bridgis persons file
Companies Mineral extraction Value added per
working place
1,820,000 D&B employment
database
Industry 279,000 D&B employment
database
Electric companies 620,000 D&B employment
database
Construction 10,000 D&B employment
database
Trading and catering 20,000 D&B employment
database
Banking and insurance 90,000 D&B employment
database
Transport
and communications
75,000 D&B employment
database
Other Pumping stations Object 750,000 Water information
system
Water
purification installations
11,000,000 Water information
system
Source: (Jonkman, 2008)
Explanatory notes: CBS: statistics Netherlands; D&B: Dun and Bradstreet.
15
2.3.1.2 Flood Loss to Agricultural Crops Using Remote Sensing
Flood loss estimate methodology for agriculture is slightly different from that of the
commercial or the residential sector. The purpose is to identify crops that are affected at the time
a flood occurs and to be able to price the related damages. In this type of modeling, the crop
type, inundation duration and the month of event have to be accounted for, besides the expected
crop value. Tapia-Silva (2011) sets up the following equation to estimate flood losses to major
crops grown in polders in Germany:
D = cv*df (Equation 2)
Where cv is the crop value, with crop value being the product of the yield and the price (cv =
yield*price), df is a damage factor which is a function of crop type, flood duration and the month
the event hits. The identification implies the understanding of the phenological phases in the life
cycle of a crop.
Identification of Affected Crops
In the modeling process of flood damages to agricultural lands, the identification of
affected crops is considered the first stage. That step can be performed using various methods in
remote sensing such as NDVI (Normalized difference vegetation index), disaggregation of
statistics and analysis of crop rotation with data mining Net Bayesian Classifier (Tapia-Silva,
2011). Using NDVI requires analysis of Landsat TM images for a period of time based on
standard spectral curves. The method can be validated, and its accuracy can be derived. That may
not be an ideal method since one may not spot small areas on a low resolution image.
Furthermore, the method has low accuracy due to misclassification which is explained by the
fact that it does not use training area.
Unlike NDVI, the disaggregation of statistical crop data method uses statistics for crops.
It takes into account the probability of a crop to be cultivated in a specific type of soil. Then, the
method is validated using reference data. Two main limitations of that method are the
misclassification of crops and the division of parcels according to soil types. The limitations
imply that other possible factors such as crop price, rotation, distance as well as parcels
accessibility drive the farmers’ decisions. Additionally, the Net Bayesian Classifiers method
16
aims at inducing a network that best portrays the probability distribution over the training data. It
consists in classifying a class variable given a set of attribute variables.
Hydrodynamic Model and Loss Estimation
Several specialized software interfaces are used to simulate floods. Flood characteristics
such as water level, flow velocity, discharge, water depth and duration can be derived and
mapped. Also, recorded water levels and flows at gauges can lead to the same modeling process,
knowing the slope of the flooded area and land use information of that area. Flood duration maps
are produced based on simulated water levels and digital elevation models (Tapia-Silva, 2011).
Several efforts have been made mainly in rich countries to develop flood loss estimate models.
Some models are presented in Table 3 and Table 4
Table 3. Summary of Existing Flood Loss Estimation Methodology
Damage categories
Japan Australia United
Kingdom
United
States
Urban Residential Detail Detail Detail Detail
Non-residential Detail Detail Detail Detail
Rural Crop damage Rough Detail Detail Detail
Farmland
damage
Detail None Detail Detail
Fishery None Detail Detail None
Infrastructure Rural Rough Rough Detail None
Damage Rough Rough Detail None
Business losses Rough Detail Detail Detail
Environmental damages None None Detail None
Source: (Dutta, 2001)
17
Table 4: Comparison of Different Flood Loss Models
Authors\References Model
development
Model
scale
Loss
functions
Impact
parameters
Resistance
parameters
Differentiation
of results
RAM (Australia)
( NRE, 2000)
Empirical –
synthetic
Micro Absolute - Object
size/value
and lead time
and flood
experience
One figure
building
structure and
contents
(Anuflood,2002)
(Australia)
Empirical Micro Absolute Water depth Object size
and object
susceptibility
One figure:
HAZUS (FEMA,
2003; Scawthorn et
al.)
Empirical –
synthetic
MicroMeso
Relative Water depth Object type Three figures:
Building,
Equipment
Inventory
Multicoloured
manual (United
Kingdom)
synthetic
MicroMeso
Absolute Water depth
and duration
Object type
and
lead time
Five figures
+ Immobile,
stock
MURL (2000)
(Germany)
Empirical Meso Relative Water depth Business
sector/
ATKIS landuse classes
Three figures
(ICPR, 200)
(Germany)
Empirical –
synthetic
Meso Relative Water depth Business
sector/
CORINE
land use
classes
Three figures
(Hydrotec, 2004)
(Germany)
Empirical Meso Relative Water depth Business
sector/
ATKIS landuse classes
One figure
(LfUG, 2005)
(Germany)
Empirical –
synthetic
Meso Relative Water depth
and specific
discharge
(m 2
/s )
Business
sector/
ATKIS landuse
classes
Three figures:
Building,
Equipment,
Inventory
Source: (Kreibich, 2010)
18
Flood Loss Estimate Components
The flood loss estimation methodology consists of two components that carry out basic
analytical processes. These are flood hazard and flood loss estimation analysis. The flood hazard
analysis module uses characteristics, such as frequency, discharge, and ground elevation to
estimate flood depth, flood elevation, and flow velocity. The flood loss estimation module
calculates potential loss estimates from the results of the hazard analysis. The potential loss
estimates analyzed through this process include physical damage to residential, commercial,
industrial and other buildings, debris generation which includes the distinction between different
types of materials, economic loss which includes lost jobs, business interruptions, and repair and
reconstruction costs, and social impacts which includes estimates of shelter requirements,
displaced households, and population exposed to scenario floods.
19
CHAPTER 3
METHODS
3.1 Study Area
3.1.1 Geography
The city of Gonaives is located at 19°27΄ north and 72°41΄ west in the northern part of
Haiti. Gonaives stands as the third most populated cities of the country with a population of
263716. The population in its entirety has decreased by 10.7% in 2009 (i.e. 235340 inhabitants)
(IHSI, 2009). That reduction is due to the exodus of people that flew, died and disappeared from
five major disasters (Jeanne, Fay, Gustav, Hanna and Ike) since 2004. Haiti has a warm and
humid climate with rainfalls ranging annually from 400 to 4000 millimeters. The average annual
rainfall for the whole country is 1400 millimeters. The watershed of Gonaives has an average
annual rainfall of 1307.96 millimeters (Prophète, 2006). Gonaives is considered a semiarid area.
The aridity is due to the effect of Foehn winds that come from the northern Atlantic Ocean. The
Municipality of Gonaives is administratively divided into five Sections. These are Labranle,
Poteau, Bassin, Pont Tamarin and Bayonnais. The urban area is composed of five sectors:
Raboteau-Jubilée, Downtown, Bigot-Parc Vincent, Kasoley and Biénac-Gathereau (IHSI, 2009).
Gonaives is crossed by the National Route 1 which connects the region to Port-au-Prince, the
capital located at 110 km south, and to the North part of the country. Figure 3 displays the
administrative subdivisions of Gonaives.
20
Figure 3. Map Showing the Location the Municipality of Gonaives with the Major Divisions.
The Quinte Watershed overhangs the highest and the most mountainous region of the
Artibonite Department and covers an area of 700 km
2
. The basin belongs essentially to the
Municipality of Gonaives and less importantly to the surrounding municipalities such as Ennery
and Marmelade. The watershed is crossed by several tributaries of the Quinte River. The most
important are Branle River from north, Ennery River from northeast, Bassin River, and
Bayonnais River draining the south side. The elevation of the terrain varies from sea level in the
southwest to 1,000 meters (Prophète, 2006).
3.1.2.- Soil Properties
The soils are Clay-loam in the Quinte Watershed (Prophète, 2006). In certain
downstream areas, the soils are suited to practice agriculture. Contrarily, other soils located in
steep slopes at about 40% susceptible to erosion are convenient to practice agroforestry.
Generally, the most part of the Quinte Watershed is made of mountains where farming is not
practiced according to soil conservation practices and techniques.
21
Annually, soils are exploited in the Quinte Watershed to produce cash and subsistence
crops. However, few farmers use conservation practices in order to reduce soil losses and to
simultaneously reduce the degradation of the ecological systems in the watershed. The
uncontrolled deforestation and the absence of soil conservation practices led to water erosion and
to the desertification process. Thus, the area is considered one of the zones exposed to
unprofitable production (Prophète, 2006).
The intensification of traditional agriculture in the country with the increase of needs in
food often has unwanted environmental and ecological consequences such as soil erosion,
salinization, and contamination of rivers with the use of chemicals. The situation is not different
from that of Haiti and the Quinte Watershed. Non sustainable agricultural practices contribute to
abuse the soil and to contaminate rivers and water springs. The entire tributaries of the Quinte
Watershed are in a state of deterioration with a relatively low vegetation cover. The lack of plant
cover exposes the arable layer of the soils to the continual washing caused by pouring rains. That
entails in every shower a significant part of land into the sea.
The type of soil hydric erosion contributes to subtract tons of sediments from the watershed, to
increase the gullying and the impoverishment of soils which considerably reduces the
agricultural space of the region (Bernardin, 1993). The soils in the Quinte Watershed are eroded
during the dry season by winds the same way they are eroded by rains during the rainy season.
However, it is worth mentioning that the hydric erosion is more severe. The risks of soil erosion
in the Municipality of Gonaives vary on a high to very high scale. Gullies formed on both sides
of the Quinte Watershed measure about 5 to 7 meters large. The small ones measure between
0.50 and 0.70 meter deep against 1.5 to 5 meters for the biggest. A land cover land use map for
the Municipality of Gonaives is presented in Figure 4. Also, an erosion map is shown in Figure
5. The geographic elements (scale bar, north arrow, legend) of the two maps (Figure 4 and
Figure 5) were modified. Their titles were taken away for visualization purposes and to comply
with the academic requirements.
22
Figure 4. Land Cover and Land Use Map of Gonaives (UTSIG/MPCE/Haïti, 1998).
23
Figure 5.- Soil Erosion Risk Map for Gonaives (UTSIG/MPCE/Haïti, 1998)
3.1.3.- Water Resources
In Haiti, the rural population suffers from a shortage of water. Access to drinking water is
very difficult. According to Magny (1991) as cited in Prophète (2006), less than 14 liters of
water are available for a person on a daily basis. The water scarcity is more severe in certain
rural zones of Gonaives such as Labranle which is one of the most disadvantaged Sections. The
supply in drinking water of the Gonaives population relies in part on the purchase of bottled
water prepared by companies located in other cities or in the neighboring Dominican Republic.
Many people consume water pulled out of wells but the water is not good because of its salinity.
In the Quinte Watershed springs are harnessed in certain area in order to provide the population
with drinking water. On the other hand, the shortage of drinking water is a major problem in the
Labranle region. It is a mountainous location where the infiltration capacity of the rain water is
low and the installation of irrigation infrastructures as well as the systems of drinkable water
conveyance is very expensive. The hydrographic network of Gonaives is presented in Figure 6.
24
Figure 6. Map Showing the Watersheds in Gonaives. The Quinte River has the largest watershed.
The major damages are generally caused when it is swollen during the cyclonic season from May
to November.
3.1.4.- Economy
The Haitian economy has known considerable decrease over the last few years. Political
instability, weakness of the public institutions, illiteracy, lack of investment, and deterioration of
the infrastructures make the economic recovery difficult. The important amounts of international
assistance granted in Haiti decreased in recent years. The 822 resolution of the OAS in 2002
supported a resumption of dialogue between international financial institutions and the Haitian
Government to find a mutual agreement on the preliminary technical conditions in the
resumption of the activities, but the results were not satisfactory (Prophète, 2006). The living
cost appreciably increased during the last five years at the rate of 15.5 % a year. Housing and
food supply prices increase while people’s income, in particular those living in rural areas,
remains very low. The annual average income of the countrymen in Haiti is estimated at
approximately 115 U.S. Dollars (Alleyn, 2006 as cited in Prophète, 2006). On the other hand, a
25
reduction in the agricultural productivity is observed in rural areas, particularly in the production
of rice. The valley of Gonaives is considered the rizicole attic of the country. The total volume of
goods exported decreases considerably whereas the imported volume increases. In the Quinte
Watershed, people have severe economic problems. The yield on the agricultural plots of land
decreases because of the deterioration of farmlands and the rarity of irrigation water during the
dry season. Table 5 and Table 6 present the cropping calendar in Gonaives. There have been no
projects of rural financing over the last few years that facilitate women to undertake non
agricultural activities such as the commercialization of staple commodities which would help
them to hold out (Prophète, 2006).
Table 5. Cropping Calendar in the Quinte River Watershed (without irrigation)
Crops Sowing Harvest
Bean April-May (1) July
August-September (2) November
Corn May September-October
Sorgum June January
Peanut May-June November-January
Pigeon Pea (Cajanus Cajan) May January
Manioc May-June After 18 Months
Sweet Potato May-June November-December
Tomato June October-November
Source : (Prophète, 2006)
Table 6 : Cropping Calendar in the Quinte River Watershed (with irrigation)
Crops Sowing Harvest
Bean November-December February-March
Sorgum February May-June
Rice May-June October-November
Carrot, beet, tomato September-November January-March
Shallot August (1) November
January (2) March-April
Onion December-January May-June
Source : (Prophète, 2006)
In the Quinte Watershed as everywhere else in the Haitian rural environment subsistence
farming is practiced. However, cash crops are produced at a small scale. Truck farming crops
(hot pepper, tomato, shallot, carrot, beet among others) are mainly produced to be marketed.
During the harvest period, small shopkeepers called “Madan Sara” buy agricultural commodities
26
from the farmers and resell them in the Downtown area of Gonaives or in Port-au-Prince. Small
vans provide transportation for products from the region of production to the points of sale.
Generally, women take care of the sale of farm products, but it is necessary to underline that the
existence of a formal and organized market to sell agroforester products is lacking.
3.1.5.- Land Tenure
In the High-Artibonite region, most of the land owners are rich. They possess large
surfaces in rural areas while living in town. Generally, farmers are owners of small surfaces.
They either use lands belongings to the State or to the big owners on the basis of a lease.
However, in the zone of Quinte Watershed, farmers are generally owners of their plots of land.
In certain cases, they occupy lands belonged to the State. In such cases they enjoy the right to
exploit trees grown on the land.
3.2 Data and Methods
The flood estimation model developed in this study combines six factors to explain the
variation of flood damage costs in Gonaives. These are income, inundation duration, inundation
depth, slope, population density and distance to the major roads. Each of those variables was
collected and processed differently. In fact, information such as receipts and expenses that allow
calculating income were collected by surveying flood victims in Gonaives. Then, the values were
inserted into an income field and joined to a spatial table containing the administrative
subdivisions of Gonaives for further analyses. Similarly, the damage costs were collected during
the same survey and spatially joined to the table. The only difference is that no calculation was
made to derive the costs. The interviewee was simply asked to estimate the amount of money he
or she would pay to repair the damages which will later have been used as the flooding cost for
that particular household. On the other hand, depth was obtained through field measurements.
This time, the interviewee was only asked to identify the spot the water reached. Once the
position was identified, the height was measured and recorded into the same questionnaire of the
survey. Then, the depth and duration variables were added to the common spatial table to be later
analyzed with the others. As for duration, the respondent was asked to estimate the time that the
water had lasted on the ground. Unlike the previous factors which were collected or derived from
survey, slope was obtained from a digital elevation model of Haiti downloaded from the divagis.org website (2009). From the DEM, the region of Gonaives was extracted, cleaned, filled,
27
projected and converted into slope percentages. Then, the percentages were reclassified into 30
categories and joined to the rest of the variables. The population density was determined and
compiled from tabulated numbers and percentages found in a report written on Gonaives in 2009
by the Haitian Informatics and Statistics Institute. Thereafter, the data were normalized joined to
the spatial table. Finally, a road layer was extracted for Gonaives from a national road network
downloaded from www.diva-gis.org (2009). A 30 rings buffer of 140 meters each was run on the
road layer. The buffer layer was overlaid with the point damage cost layer in order to measure
the distance of each point to the road. Like the six other variables, the values were inserted under
a distance field in the same spatial table. Figure 7 and Figure 8 present respectively the
preparation of the variables and the flood damage cost modeling process.
After the processing and preparation of the variables, the analysis step was ready to
taken. All the data were exported into Excel to undergo a preliminary analysis. A correlation
matrix and scatter plots of the variables were generated. That basic analysis signaled that
duration and depth factors were highly correlated. Despite the fact that redundancy was detected
between duration and depth, we decided to move forward since we could not tell which one of
the two was to be discarded. At that point, income, duration, depth, slope, density and distance
were combined as explanatory variables against flood loss cost as the dependent variable to run
the ordinary least square (OLS) regression analysis. The interpretation of the variance inflation
factor (VIF) indicator generated revealed that the value of depth was significantly high; therefore
it was excluded from the model. The same analysis was later conducted with the remaining
variables to create the primary OLS I model. To test the trustworthiness of the model, a Moran’s
I test was performed to check whether there was spatial autocorrelation in the distribution of the
costs’ residuals. It turned out that the errors were randomly distributed. Thus, there was no need
to run the geographically weighted regression (GWR) analysis since OLS I was reliable to
predict flood damage costs at a global level. However, all the variables were permuted in groups
of four to check which combination was spatially auto-correlated. Surprisingly, none of the
combinations of four variables showed any clustering in the distribution of the residuals. The
same process was done for combinations of three factors. Finally, it has been found that the
residuals issued from OLS II consisted of the combination of income, slope and population
density, deviated from the normal distribution at a global scale; therefore GWR proved to be
necessary to check the tendency at a local scale.
28
Finally, it came the time to interpolate and map the flood damage costs estimated by the
OLS I and GWR models. Several interpolation tools were unsuccessfully tested to see which one
would best explain the point features. None of them showed enough spatial variability in the
interpolation of the loss cost estimates. Among these techniques were Simple and Ordinary
Kriging, Inverse Distance Weighted (IDW), Spline interpolator, Point Density and Kernel
interpolator. Lastly, the Voronoi Map geostatistical interpolator was successfully used to create
Thiessen polygons from the point features.
In the development of this flood loss estimate model for the residential sector, it was
assumed that there was neither flood insurance, preparedness plans nor disaster experience;
therefore these variables were not considered in the model. Indirect intangible damages were not
accounted for in the model because of scarceness of data especially economic and life loss data.
Here, the inundation depth is equal to the difference between the water level and the ground level
at the entrance as defined by (Zhai, 2005) in the case of a building. Flood duration categories
were surveyed in default of maps created from water level and digital elevation models.
Conventionally, inundation duration and depth are two primary factors of flood modeling. In
this study, income, slope, population density and distance to the major roads were added as four
additional variables to build the final conceptual model.
We used a flood loss estimate model developed in Excel and ArcGIS 10 to predict
potential flood damages in Gonaives associated with income, duration, depth, slope, density and
distance. Each of these factors was collected from a different source or derived differently.
Receipts, expenses and money transfer collected through survey enabled us to determine the
income variable. The income was calculated for each household using the following formula:
 


n j Jj CPBR
1
(Equation3)
Where R is income, PB is gross product of the activity j, and C the costs induced by the activity
j. 29
DEM
Road Network
Field
Measurement
Tabular Data
Fill
Spatial Table
Slope
Calculate
Distance
DEM Fill
Depth
Survey IncomeCalculate
Extract
Slope %
Density
DistanceRoad Layer
Record
Join
Duration
Cost
Record
Base Data Extract Study Area
Figure 7. Processing and Preparation of the Variables
30
Table 7. Definition of Variables Used in the Regression Models.
Independent
Variable
Dependent
Variable
Definition
1-Income
This is the general income constituted of farming,
breeding, fishing, off-farm activities, and money
transfer from the diaspora.
2-Duration This is the time that the inundation lasted. It is
estimated in minutes.
3-Depth This is the height that the water reached above the
ground level. It is measured in centimeters.
4-Slope This the slope percentages derived from the digital
elevation model of Gonaives.
5-Density This is the population density expressed in population
per square kilometer of the different administrative
subdivisions of Gonaives.
6-Distance This is the distance in meters of the surveyed
households to the major roads crossing Gonaives.
Flood loss cost
The cost corresponds to the amount that an owner
spent or would spend to repair the damages. It is given
in local currency, the Gourdes (40 HTG = 1USD).
Income
(HTG)
Distance to
Roads
(M)
Population
density
(P/km2)
Slope
(%)
Join
Flood Cost
Estimate Maps
Delineate
Moran’s I
OLSSpatial Table Cost Estimates
Duration
(Mn)
Flood Cost
(HTG)
Figure 8. Flood Damage Cost Modeling.
31
3.2.1 Sampling Technique
Municipal Section is the least administrative unit in Haiti. The study area is divided into
five Municipal Sections which are Bayonnais, Pont Tamarin, Bassin, Poteau and Labranle. The
Quinte River crosses all the five Sections almost equitably. Also, the relief acr

PDF Document reader online

This website is focused on providing document in readable format, online without need to install any type of software on your computer. If you are using thin client, or are not allowed to install document reader of particular type, this application may come in hand for you. Simply upload your document, and Docureader.top will transform it into readable format in a few seconds. Why choose Docureader.top?

  1. Unlimited sharing - you can upload document of any size. If we are able to convert it into readable format, you have it here - saved for later or immediate reading
  2. Cross-platform - no compromised when reading your document. We support most of modern browers without the need of installing any of external plugins. If your device can oper a browser - then you can read any document on it
  3. Simple uploading - no need to register. Just enter your email, title of document and select the file, we do the rest. Once the document is ready for you, you will receive automatic email from us.

Previous 10

Next 10