Report No: AUS0001219
. Central America Strengthen DRM+Resilience of CA cities Guidelines for Flood Risk Modelling in the Sula Valley, Honduras. . November, 2018 . URS
.
. Document of the World Bank . © 2017 The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org
Some rights reserved
This work is a product of the staff of The World Bank. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries.
Rights and Permissions
The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given.
Attribution—Please cite the work as follows: “World Bank. 2018. Guidelines for Flood Risk Modelling in the Sula Valley, Honduras. © World Bank.”
All queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: [email protected].
THE WORLD BANK GUIDELINES FOR FLOOD RISK MODELLING IN THE SULA VALLEY - HONDURAS Draft Report
Maria Carolina Rogelis 11/5/2018
Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
CONTENTS
1 Introduction ...... 3 2 Definitions ...... 4 3 Study Area ...... 7 4 Planning for flood risk assessment ...... 9 5 Methodological approaches ...... 11 5.1 Data ...... 11 5.1.1 Background site data ...... 11 5.1.2 Terrain Data ...... 13 5.1.3 Infrastructure data ...... 18 5.1.4 Rainfall data ...... 19 5.1.5 Stream flows ...... 21 5.1.6 Historical data ...... 22 5.1.7 Land Use and land cover ...... 25 5.1.8 Assets data ...... 25 5.1.9 Demographic and socio-economic data ...... 26 5.2 Flood Hazard assessment and flood mapping ...... 27 5.2.1 Hydrologic analyses ...... 29 5.2.2 Hydraulic modeling ...... 32 5.3 Exposure ...... 43 5.4 Vulnerability ...... 43 5.5 Risk ...... 50 6 Floodplain Zoning ...... 56 7 Maps ...... 62 8 Visualization and dissemination of hazard, vulnerability and risk information ...... 65 9 Community consultation ...... 67 10 Debris flows risk assessment guidelines ...... 67 10.1 Debris flow hazard modeling...... 69 10.2 Vulnerability curves ...... 72 10.3 Risk analysis ...... 74 11 Required team for flood risk assessment ...... 76 12 Guidelines on reporting ...... 76 References ...... 78
2 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
1 Introduction
Honduras is highly exposed to climate-related hazards (hurricanes, tropical storms, floods, droughts and landslides). The most damaging event so far is the Hurricane Mitch (1998) that destroyed and estimated 70% of the country´s crops and infrastructure, causing more than 10000 deaths and $ 3 billion in damage (USAID 2017). Floods are the one of the disasters that affects the most people out of all hazards in Honduras (Thieme and Jacobs 2012). One of the areas heavily affected by floods is the Sula Valley that concentrates 24.5% of the population of the country and where 5 of the main cities of the country are located. This is the area with the most rapid economic growth of the country with 80% of the manufacturing and textile industry (Caballero et al. 2008).
Flood risk assessment is an essential component in the disaster risk management and climate change adaptation planning process, and constitutes a crucial element in the sustainability of Honduras and in particular of the Sula Valley. The purpose of a risk assessment is to define the nature of the risk, answer questions about characteristics of potential hazards (such as frequency, severity), and identify vulnerabilities of communities and potential exposure to given hazard events. Risk evaluation helps in the prioritization of risk management measures, giving due consideration to the probability and impact of potential events, cost effectiveness of preventative measures, and resource availability (The World Bank 2011). Development and implementation of flood risk studies and flood mapping impacts a wide-range of key users in areas as diverse as land use planning, emergency management and community awareness. Communication of this information helps to build flood resilience within the community and informs and agencies to manage flood risk into the future (BMT WBM Pty Ltd 2017).
This document compiles good practices and guidelines for flood risk assessment in the Sula Valley, that are intended to promote future studies meeting the regional and local needs, for an effective approach to Integrated Flood Management.
Despite the variety of flooding sources in the Sula Valley, this document if mainly focused on fluvial floods. Coastal and debris flows are also addressed although in a more general manner.
3 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
2 Definitions
The following are the definitions of the main concepts used in this document.
Floods
Floods can be defined as a temporary covering of land by water outside its normal confines (Schanze et al. 2006). Figure 1 shows a classification of floods according to the origin.
Figure 1. Basic classification for the type of floods according to their origin. Source: (Díez- Herrero et al. 2009) Some types of floods are shown in Figure 4. Their detailed definition is as follows:
• Tidal flooding: Both sea and river defences may be overtopped or breached by a combination of low pressure weather systems and peak high tides. Storms with high wind speeds cause tall and powerful waves and low pressure fronts cause sea levels to rise above normal levels (RIBA 2009). • Fluvial Flooding: Flooding occurs in the floodplains of rivers when the capacity of water courses is exceeded as a result of rainfall or snow and ice melts within catchment areas further upstream. Blockages (RIBA 2009). • Pluvial Flooding: Surface water flooding is caused by rainwater run-off from urban and rural land with low absorbency. Increased intensity of development in urban areas has given rise to land with a larger proportion of non-permeable surfaces, a problem often exacerbated by overloaded and out-dated drainage infrastructure. These circumstances, combined with intense rainfall, can give rise to localised flooding (RIBA 2009).
4 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 2. Types of floods. Source: (RIBA 2009)
Hazard
A potentially damaging physical event, phenomenon and/or human activity, which may cause the loss of life or injury, property damage, social and economic disruption or environmental degradation (UN/ISDR, 2002).
Flood hazard map
Flood hazard maps provide flood hazard information in terms of the relationship between the flood probability and flood intensity. Such maps should be produced for areas where flooding could cause considerable damage. Statistical flooding probability is expressed in terms of flooding recurrence, whereas intensity is generally measured in relation to water depth. However, other characteristics such as water velocity are also used (Meon 2006).
Flood hazard zone map
Flood hazard zone maps are the result of the interpretation of flood hazard maps from the perspective of specific disciplines. They are used to apply water management data to regional planning processes and should be developed insofar as necessary. These maps are to show the probability and intensity of flooding in specific danger zones, which are generally assigned one of the following classifications: high, medium, low, and very low (Meon 2006).
Vulnerability
Even if a universal definition of vulnerability does not exist a common used definition is: The conditions determined by physical, social, economic and environmental factors or processes, which increase the susceptibility of a community to the impact of hazards (International Strategy for Disaster Reduction - UN/ISDR, 2004).
5 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Exposure
The exposure is defined as the human activities affected by the hazardous event. The flooding area shown in the flood hazard map doesn’t reveal the houses, factories or firms covered underneath. The exposure analysis is to find out the economic activities covered by the flood (Su and Kang 2005).
Damage
Damage is the amount of money to restore the area back to its original condition before the disaster (Su and Kang 2005).
Flood risk
According to the European commission "flood risk" means the combination of the probability of a flood event and of the potential adverse consequences for human health, the environment, cultural heritage and economic activity associated with a flood event.” The term risk can be defined as the probability and the amount of harmful consequences or expected losses resulting from interactions between natural or human induced hazards and vulnerable conditions (UN/ISDR, 2002) (Birkmann 2006).
Actual flood risk is the risk posed to an area, whether it is behind defenses or undefended, at the time of the study. This should be expressed in terms of the probability of flooding occurring, taking into account the limiting factors, both natural and manmade, and preventing water from reaching the development. Residual risks are the risks remaining after all risk avoidance; substitution and mitigation measures have been taken. Examples of residual flood risk include: the failure of flood management infrastructure such as a breach of a raised flood defense, blockage of a surface water channel, failure of a flap valve, overtopping of an upstream storage area, or failure of a pumped drainage system; and A severe flood event that exceeds a flood design standard such as, a flood that overtops a raised flood defense (Fhoilsiú and Oifig 2009).
Flood risk management is a ‘holistic and continuous societal analysis, assessment and reduction of flood risk’. ‘Holistic’ refers to the flood risk system which should be considered as comprehensive as possible. The term ‘continuous’ expresses the need for an ongoing assessment of flood risks, their dynamic change and effects of reduction activities. Analysis, assessment and interventions for risk reduction as common elements of a development model are dedicated to the ‘managing entity’. As far as the management of the flood risk system is concerned, society, represented by politicians, experts and individuals, could be seen as the ‘management entity’. These representatives perceive flood risks and assess whether a certain flood risk is tolerable or not and decide on risk reduction interventions (Schanze et al. 2006).
6 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
3 Study Area
The Sula Valley comprises the basins of the Ulúa, Chamelecón, Motagua, Mezapa and Lean rivers and corresponds to the Development Region 01 conformed by 20 municipalities of the departments of Atlántida, Cortés, Santa Bárbara and Yoro (SEPLAN 2014) (see Figure 3). The months with the highest precipitation are June and September and the driest are March and April. The precipitation ranges from 2001 mm in the south to 3800 mm in the coast (Omoa and Tela municipalities) (SEPLAN 2014).
The Merendon Mountains rise from the Sula Valley to a maximum height of 1700 m above mean sea level and develop a very steep topography. The landform is composed of steep mountains slopes, steep streams, alluvial fans and cones at the valley mouths and alluvial plains (JICA 1994).
The Sula Valley and its tributary basins, comprise about 25,800 km2 or 23% of the area of Honduras; the Valley proper is approximately 2,000 km2 (Schultz 1981). Severe flooding has caused substantial loss of life and property damage in the valley and the threat of floods has inhibited the realization of its full economic potential (Schultz 1981). This is an area of rapid demographic growth, with a population of 1985141 estimated for 2011. The economic activities include the agroindustry (palm tree, sugar cane, banana and coffe); maquilas (mainly textile), commerce and tourism. The main port of the country, -Puerto Cortés-, is located in this region. The route CA-5 connects Tegucigalpa and Puerto Cortés through San Pedro Sula and Choloma. This route is of high importance for transporting business and tourism passengers, export and import goods at Puerto Cortés and commodities for domestic use (JICA 1994).
Basin Total area km2 Rio Chameleco�n 4091.47 Rio Lean 1156.96 Rio Mezapa 809.12 Rio Motagua 2869.9 Rio Ulu�a 21330.55
Table 1. Total area of the basins in the Sula Valley. Source: (SEPLAN 2014) The El Cajón reservoir has been constructed in the the Ulúa river basin and the Jicatuyo and los Llanitos dams are planned in the short term. In the Chamelecón river basin El Tablón dam is planned to be constructed (SEPLAN 2014). El Tablón is a multipurpose reservoir (467 Mm3) and is expected to contribute to flood control in the Sula Valley (Bnamericas 2018).
Several flood mitigation works have been undertaken in the Sula Valley by the Comisión para el Control de Inundaciones del Valle de Sula – CCIVS, created in 2010. However, municipalities such as Potrerillos, Progreso and Tela are frequently flooded. The failure of levees is a recurrent problem. Due to the lack of economic resources for levee maintenance, the construction of dams (El Tablón, Jicatuyo and Los Llanitos) is seen as an alternative to mitigate flooding in the Sula Valley (SEPLAN 2014).
7 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 3. Sula Valley The rivers Chamelecón and Ulua have been intensively modified. The banana companies constructed canals in for the drainage of crop areas and for irrigation. These structures promoted the use of the floodplains for human settlement and cultivation (banana and palm tree). However, currently this infrastructure is deteriorated due to lack of maintenance and inadequate use (Municipalidad de Puerto Cortés 2011b).
Flooding problems in the valley were graphically demonstrated by the disastrous effects of hurricane Fifí in September 1974 (Schultz 1981) and by the Mitch Hurricane in 1998. Also, frequent localized flooding occurs in the Valley.
In 1974 the hurricane Fifí caused severe flood damages to the Sula Valley. It caused thousands of hill slope collapses in the Merendón mountains. The debris flows and floods resulted in a heavy loss of human lives and severe flood and debris related damages. Approximately 10,000 casualties were reported (JICA 1994). During the hurricane Fifi from September 18 to 19, La Mesa and Puerto Cortés recorded the maximum daily rainfalls of 340 mm and 280 mm respectively. The simulations carried out by (JICA 1994) showed that the discharges in the Sula Valley during the hurricane Fifí correspond to those caused by the daily rainfall of once in 50 years return period in the whole
8 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
basin.
In late October 1998, Hurricane Mitch, a category 5 hurricane, struck Honduras and other countries in Central America. Several days of intense rain from this tropical cyclone caused devastating floods and landslides throughout the affected area. In Honduras, 7,000 people died, 33,000 homes and 95 bridges were destroyed, and 70 percent of the road network was damaged (Mastin 2001).
4 Planning for flood risk assessment
As initial stage in any flood risk assessment project, planning is crucial. The case of the Sula Valley is particularly challenging due to the severity of the flooding impacts and size and complexity of the area.
Before starting a flood mapping project, it is important to clearly identify (World Meteorological Organization 2013):
(i) Why flood mapping is being implemented; (ii) The specific food management decisions it is intended to support; (iii) The intended audience (who is using the maps?), including the necessary dissemination process and mechanisms; (iv) The status of flood mapping in the country and gap identification; (v) The data requirements and availability; (vi) The availability of technical expertise; (vii) The future action plan, including the long-term update process, along with the financial resources required.
It is widely recognized that flood mapping is an urgent need in the Sula Valley. Land planning adequately informed by risk analysis is a priority risk management measure due to the traditionally informal occupation of the territory that has progressively increased flood risk. Other uses, such as emergency management, community awareness and risk mitigation works also require different levels of flood risk assessment. Although several efforts have been made to advance risk analysis in the area, currently the Sula Valle does not have risk information based in an integral approach. Therefore, a highly important stage in the process is to plan at short, medium and long term the development of flood risk knowledge considering the available resources. Flood assessments can be implemented in different stages that depend on a policy decision and that can ensure a proportionate methodology to assessing flood risks for a given land area for a particular decision. A staged approach should be adopted comprising appropriate screening, scoping and detailed assessments. This will ensure all the issues are identified and re-use of data and model information provides fit-for-purpose risk assessments (RIBA 2009). The stages differ mainly with regard to the technical depth of the assessment and with the approaches used (World Meteorological Organization 2013), Figure 4 shows the progressive stages for flood hazard assessment, starting from a macro-scale to a local scale of analysis. Flood risk assessments are required at different scales by different organisations for many different purposes. A hierarchy of assessments is necessary to ensure a proportionate response to the needs by avoiding the need for detailed and costly assessments prior to making strategic decisions (RIBA 2009).
The more detailed a method of damage evaluation is in terms of its spatial resolution and its differentiation of elements at risk, the more effort per unit of area is required to carry it out.
9 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Therefore, the most detailed methods are often restricted to areas of local size, while studies for areas of regional or even national size have to rely on approaches, which require less effort per unit of area and, consequently, do not provide such a high level of precision (Messner et al. 2007). Due to the difference in level of effort and therefore cost, the strategy in many countries is to carry out the detailed studies in prioritized areas, e.g densely populated areas that concentrate exposure and in relatively unpopulated areas, conduct approximate studies (NFIP 2015) (BMT WBM Pty Ltd 2017). An important principle to take into account is the fit-for-purpose approach. Fit-for- purpose maps are advocated as a result of recognising that not all communities require the same level of mapping to ascertain their flood risk (State of Queensland 2014).
Preliminary Detailed First perspective Flood prediction for Can be combined General purpose with macro-scale specific locations vulnerability Cover large areas Scale 1:25,000- Scale 1: 100,000 - Scale 1:50,000 - 1: 1:2,000 1:500,000 250,000 Extensive planning, Mainly historical and data gathering and geomorphological expertise methods
Figure 4. Stages of flood hazard assessment. Adapted from: (World Meteorological Organization 2013) These considerations imply that in the case of the Sula Valley a prioritization would be very useful to identify areas where detailed flood risk assessments are needed and where a coarser scale would produce fit-for-purpose results. This approach would promote an efficient use of resources and provide a strategy to advance risk knowledge.
The choice of mapped parameters and type will depend upon the objectives of the project, the resources available and the potential benefit achievable. The following maps are relevant: Event map, Hazard map, Vulnerability map, Risk map (World Meteorological Organization 2013). The updating process has to be defined before the mapping starts. Basic maps need to be updated according to the need, recommended updating frequency is as follows (World Meteorological Organization 2013):
• Event map: Whenever a flood occurs the information should be integrated in the available data base and shown on the map • Hazard map: Updating should occur regularly (e.g. every 10 to 15 years) or after new information is available; this might be a major event or new knowledge, know-how and technologies to assess the respective hazards • Vulnerability and risk map: The built environment is rapidly changing. New developments might considerably change the vulnerability; it might even change the hazards. Therefore, the map updating should occur at the same pace as for the hazard map or any major change in the flood plain.
10 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
To address this challenge, a committee could be created, grouping the technical institutions involved that could be in charge of setting priorities and criteria for flood risk assessment.
5 Methodological approaches
Flood risk assessments focus on (NFIP 2015):
1. Flood hazard—the probability and magnitude (e.g., depth, velocity, discharge) of flooding. Flood hazard assessment estimates the probability of different magnitudes of damaging flood conditions, such as the depth of inundation, duration of inundation, velocity of moving water, quality of water, debris content of water, or the wave height in addition to still water level (NFIP 2015). 1. Exposure—the economic value of assets subjected to flood hazard 2. Vulnerability—the relationship of flood hazard properties to economic loss
A fourth component is frequently included addressing the performance, effectiveness and behavior of flood protection and damage mitigation measures that modify the flood hazard, the exposure, or the vulnerability. This fourth component, if particularly important in the Sula Valley since the drainage system has been severely modified and a significant amount of flood mitigation works have been constructed. In addition, there is recognition that mitigation works lack maintenance (Municipalidad de Puerto Cortés 2011b).
The following subsections address the data needed for flood risk assessment and the methodological approaches for each component of flood risk.
5.1 Data
In any model application, the accuracy of the results depends on the accuracy and resolution of the key input data or suitable proxies (Czajkowski et al. 2016). Flood studies rely on a range of data. A crucial step is to collate and analyze data to check for completeness, accuracy, relevancy and spatial extent (BMT WBM Pty Ltd 2017).
The following subsections summarize the main data required for flood risk assessment.
5.1.1 Background site data
Background information of the area to be studied includes (SEPA 2014):
• Geographical features, street names and all watercourses or other bodies of water in the vicinity which may have an influence on the study area. This should include drainage outfalls and overflows. Currently, the SINIT1 (a screenshot is shown in Figure 5) provides information at national level. However, most layers are outdated. In the case of the water courses the available layers have a scale 1:50.000 and there is no detailed information on the names of the rivers and creeks. Therefore, a priority task is to update the geographical
1 Sistema Nacional de Información Territorial Honduras, http://www.sinit.hn/
11 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
database.
Figure 5. Screenshot of the SINIT website • Photographs of the site showing areas of importance such as river channel, floodplain, culverts, other structures, areas of erosion, trash lines and areas of woody debris accumulation etc. A georreferenced photo database is very useful and can be created on the basis of fieldwork carried out by the water and risk management agencies. • Agricultural pollution sources (herbicides, pesticides, fertilizers, manure, poisonous substances and nutrients), wastewater treatment plants, waste storage, septic tanks (EXCIMAP 2007). • Protected areas, nature conservation thematic databases and maps (EXCIMAP 2007). • Cultural heritage – thematic databases and maps (EXCIMAP 2007). • A detailed indication of flood alleviation measures already in place and planned, of their state of maintenance and their performance. An inventory of mitigation works is crucial, not only for flood risk analysis but also for emergency response and maintenance planning. The Comisión para el Control de Inundaciones del Valle de Sula - CCIVS, continuously plans, prioritizes and watches the condition of flood mitigation works. An example of planned projects is shown in Figure 6. The incorporation of this information into a risk information system that promotes updated information and easy access for all agencies involved in risk management would constitute a significant step forward in knowledge sharing and efficient information use.
12 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 6. Projects planned by the CCIVS. Source: email communication CCIVS
5.1.2 Terrain Data
Terrain data for flood hazard purposes differs in accuracy depending of the type of analysis to be carried out. High-accuracy digital elevation data are not required for hydrologic analyses. If no better form of digital elevation data is available, SRTM or ASTER DEMs can be used for hydrologic analyses and modeling of entire watersheds (FEMA 2003). In contrast, hydraulic analyses require high-accuracy contours, mass points and breaklines, TINs, or DEMs, but of floodplain areas only, and not of entire watersheds (FEMA 2003).
Generally terrain data for hydraulic analysis is provided in the form of a digital terrain model, usually built from data collection that can include airborne light detection-and-ranging (LIDAR) data as well as geographical survey data (Prinos 2008). Bare Earth is a classification that is free from vegetation, buildings, and other man-made structures. Bare Earth contains only the ground topography. The Bare Earth point classification is used for creating digital topographic surfaces that are the basis for hydraulic analysis and floodplain mapping (FEMA 2016).
To increase accuracy of the details with respect to buildings and terrain structures near to water bodies, the laser-derived digital terrain model is combined with geographical survey data. All buildings in and around a water body should be surveyed. Furthermore, where LIDAR do not provide a complete survey of the river cross section (e.g. where cross sections are occupied by water during the LIDAR survey), these should be surveyed at appropriate intervals for hydraulic requirements (Prinos 2008) (INDECI 2011). LIDAR can be combined with geographical survey data as shown in Figure 7.
13 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
In the framework of the PGRD project, the COPECO carried out a LIDAR survey in the Sula Valley covering mainly the floodplains of the rivers Ulúa and Chamelecón. A previous LIDAR survey exists, carried out by the USGS in 2002. This was undertaken in order to prepare flood hazard maps for several municipalities of Honduras after Hurrican Mitch, including El Progreso, Choloma and La Lima (USGS 2002a). The LIDAR is archived by the USGS and would require a request to be provided. In addition to the LIDAR, ground survey would be required to obtain cross sections in reaches where water levels were significant at the time of the survey. The definition of those reaches should be carried out through an analysis that justifies the use of the LIDAR without merging cross sections, taking into account the accuracy needs for the planned hazard mapping. An example of such analysis was carried out by the USGS (USGS 2002b).
A digital elevation model with pixel size of 5m was acquired by COPECO to cover areas that are not covered by LIDAR in the Sula Valley. It is anticipated that the accuracy of this model is lower than the LIDAR and ground survey for cross sections would be needed.
Figure 7. Combination of geographical survey data and laser scan data for extracting cross section data along the water bodies (Ministry of the Environment Baden- Wü rttemberg, 2005). Source: (Prinos 2008) For surveying cross sections beneath the water level, the method to be used depends on the depth and velocity of the stream. When water is shallow and can be waded, traditional differential leveling is used with conventional elevation rod (including rod extensions), using an attached bottom plate as necessary to prevent the rod from penetrating silt and mud. When water is deeper or too swift for wading, various hydrographic survey methods may be used. These may include small workboats with sounding poles or lead lines, or specialized hydrographic survey vessels with echo sounders, either single (vertical-beam) or multi-beam transducers (FEMA 2003). The requirements established by (FEMA 2003) for cross section surveying are shown in Figure 8.
Good practices in cross section surveying are summarized as follows:
• Cross-sections are required at all locations where changes occur in longitudinal slope, cross- sectional area, channel roughness, bridges and other channel constriction. Several cross- sections maybe required to describe abrupt changes (Nottawasaga Valley Conservation
14 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Authority 2013) (FEMA 2003). • All cross-sections should be coded from left to right looking downstream. • Overbank and channel distances between cross-sections shall be reflective of actual watercourse bends (Nottawasaga Valley Conservation Authority 2013). • Left and right bank channel stations shall be representative of actual channel low flow banks. • Maximum spacing between successive cross-sections shall be dictated by the analytical requirements of the model (Nottawasaga Valley Conservation Authority 2013). • The length between cross-sections should be based on river geometry and the assumption that gradually varied flow within a reach is valid (Nottawasaga Valley Conservation Authority 2013).
Figure 8. Requirements for cross section surveying. Source: (FEMA 2003)
• At high flows, cross-sections are expected to follow the valley feature and not the low flow channel (Nottawasaga Valley Conservation Authority 2013). • Cross-sections shall be extended across the entire floodplain, should be perpendicular to the
15 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
anticipated flow lines (approximately perpendicular to contour lines) and only positive chainages are to be utilized (Nottawasaga Valley Conservation Authority 2013) (FEMA 2003). • Cross sections must be spaced so that the geometry and hydraulic roughness of the reach between adjacent cross sections varies gradually and that variation can be estimated as linear (FEMA 2003). • Where possible, cross-sections of the channel above and below the waterline must be taken by field survey at all representative locations throughout the channel reach. Cross-sections must include the entire floodplain of the main channel and any tributaries. Cross-sections must be tied in vertically to established geodetic benchmarks and horizontally to permanent structures or GIS coordinates (Nottawasaga Valley Conservation Authority 2013). • The general slope of the flow path between adjacent cross sections should be approximately constant (FEMA 2003). • The same numbering system for the cross-sections must be used in the hydrodinamic model, as on the floodline maps and field survey notes and plots (Nottawasaga Valley Conservation Authority 2013). • For each reach studied, the sources of cross-section data and methods of measurement must be fully documented. Where more than one technique is used to acquire cross-section data, the documentation must include an explanation of how the data were merged (FEMA 2009).
In areas for which nor LIDAR nor 5m elevation model are available, an alternative terrain data is the SRTM data. However, the difficulty of obtaining a bare-earth DEM from SRTM, (which is the default surface required for flood mapping (FEMA 2003)), in vegetated areas should be considered carefully. The inability of the synthetic aperture radar (SAR) to fully penetrate the tree canopy results in significant positive biases in the SRTM ground elevation over densely forested regions, and a similar problem occurs in urban areas where radar returns from the roofs of buildings are more common than returns from ground level. Recently released global estimates of forest canopy heights enable systematic corrections to be made to SRTM data, yielding significant improvements to hydraulic models run in tropical forested catchments (Sampson et al. 2015).
For coastal areas, as is the case of the north of the study area, to properly build a coastal flood hazard map, different types of data to characterize the coastal domain are required. These are necessary to build a (Prinos 2008):
i. Digital terrain model, DTM, of the subaerial domain to be flooded (floodplain), to define the morphology of the coastal fringe where waves and surge will impact during the flood event and; ii. The bathymetry where waves and surge will propagate during their approach towards the coast (Prinos 2008). In addition to the topography of the floodplain, in coastal flooding analysis is necessary to characterize the morphology of the coastal fringe.
Regarding accuracy, several countries have issue standards to ensure a minimum level. For example, in the national guidelines of Canada the highest specification level is for a consolidated vertical accuracy of 36.3 cm, which corresponds to an equivalent contour accuracy of 0.6 m. In rural, sparsely populated areas, a lower degree of accuracy is acceptable, while in dense urban areas, higher accuracies are recommended, such as those obtained through LiDAR surveys. Minimum requirements for digital elevation models (DEMs) are considered to be 10 m by 10 m horizontal resolution (5 m by 5 m preferred) and 0.5 m vertical resolution (0.3 m preferred) (Profesional Engineers and Geoscientists of BC 2017). The datum, coordinate system and projection should match the requirements specified by the data spatial structure of the country.
16 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
The Federal Emergency Management Agency of the USA (FEMA) has reduced the complex requirements to two standard choices for digital elevation data, expressed as equivalent contour intervals (FEMA 2003):
• Two-foot equivalent contour interval for flat terrain (Accuracyz= 1.2 foot at the 95-percent confidence level). This means that 95 percent of the elevations in the dataset will have an error with respect to true ground elevation that is equal to or smaller than 1.2 feet.
• Four-foot equivalent contour interval for rolling to hilly terrain (Accuracyz = 2.4 ft at the 95- percent confidence level.) This means that 95 percent of the elevations in the dataset will have an error with respect to true ground elevation that is equal to or smaller than 2.4 feet.
Table 2 summarizes the vertical accuracy requirements established by (FEMA 2016).
Table 2. Vertical Accuracy Requirements based on Flood Risk and Terrain Slope within the Floodplain being Mapped. Source: (FEMA 2016) The European exchange circle on flood mapping (EXIMAP) in their handbook of good practices for flood mapping in Europe, indicate that minimum requirements for digital elevation models are 10 m*10 m (possibly 5 m*5 m) horizontal and minimum 0.5 m vertical resolution. Possible tools/methods to generate DEMs of the required accuracy (EXCIMAP 2007) are LiDAR, SAR and variations (IFSAR, GeoSAR, AIRSAR), orto-maps, DTM derived from digital satellite images (SPOT 5; multispectral res 10 m, panchromatic res 3, DTM/DEM derived from aerial digital ortophotos (terrain pixel size: 0,5-2,0 m; vert. res. 0,3-0,5 m, DEMs derived from the vectorised contour lines of 1:10 000 scaled digital map segments (terrain pixel size: 0,85-2,0 m; contour lines available in 1.0 m resolution, on flat territories also the median 0.5 m contour lines are given) interim terrain surfaces to be determined by non-linear interpolation.
According to FEMA the topographic mapping, cross section, and hydraulic survey requirements for each analysis option are summarized below (FEMA 2003):
• For a detailed flood hazard analysis, which will generally include a detailed hydraulic analysis, digital topographic data, cross sections (to include underwater elevations), and surveys of hydraulic structures are required. Ground surveys and either photogrammetric mapping or LIDAR-generated mapping are normally required unless suitable topographic information is already available from other sources.
17 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
• For an approximate flood hazard analysis, cross sections may be interpolated from contours on topographic maps, and underwater elevations may be interpolated from upstream/downstream data, assuming the channel bottom information has not changed significantly. Hydraulic structure surveys are not required. • For a redelineation of floodplain boundaries using more up-to-date or more accurate topographic data, the topographic data needed to update the floodplain boundaries are required, but no new cross sections or hydraulic structure surveys are required.
5.1.3 Infrastructure data
Infrastructure data includes information on specific man made features within a catchment such as roads/rail embankments, culverts, weirs, bridges, levees etc. For riverine studies, it is the infrastructure that is contained within the waterway or on the floodplain that is important. For overland flow studies, stormwater drainage infrastructure is of importance. Infrastructure can have a significant influence on flood behavior, and the availability of such information is often critical to the outcomes of the flood study. The effects on flood behavior may be intentional (such as a weir) or unintentional (such as blockage by a rail embankment) (BMT WBM Pty Ltd 2017).
For detailed flood hazard analyses, the required dimensions and elevations of all infrastructure and hydraulic structures and underwater sections adjacent to the structures must be obtained from available sources or by field survey where necessary (INDECI 2011). The determination of dimensions and elevations of hydraulic structures by aerial survey methods is not normally accepted (i.e., photogrammetry or LIDAR) (FEMA 2003).
For bridges, sufficient data for input to modeling software usually includes the following (FEMA 2003):
• Size and shape of the opening; • Upstream and downstream channel invert elevations; • Entrance conditions (e.g., wingwalls, vertical abutments) • Bridge deck thickness, low-steel elevation, and bridge parapet type (i.e., solid railing, open railing); • Roadway embankment side-slope rate; • Type and width of roadway pavement; and • Top-of-road section for sufficient length for weir-flow calculations.
For culverts, the required data includes the following (FEMA 2003):
• Size and shape; • Upstream and downstream channel invert elevations; • Entrance conditions (i.e., headwall, wingwalls, mitered to slope, projecting); • Roadway embankment side-slope rate; • Type and width of roadway pavement; and • Top-of-road section for sufficient length for weir-flow calculations.
For dams and weirs, sufficient data for input to modeling software usually includes the following (FEMA 2003):
• Top-of-dam elevation; • Normal pool elevation;
18 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
• Principal spillway type, inlet and outlet elevations, and dimensions; and • Emergency spillway (if applicable) type, elevation, and dimensions.
5.1.4 Rainfall data
Rainfall data is required to estimate the inflows into a catchment or study area. Rainfall data is used to estimate the total rainfall as well as spatial and time varying rainfall intensities across study areas and/or catchments during historical flood events. The estimated rainfall distributions are traditionally input into the hydrologic model to simulate the conversion of rainfall to runoff flows due to the historical flood events during model calibration and verification (BMT WBM Pty Ltd 2017) (Victoria State Government 2016).
The climatological stations provided by the SINIT are shown in Figure 9. The SINIT provides a layer called “Estaciones Climatológicas (1981-1999)”, therefore it is interpreted that the layer contains information until 1999. Furthermore, the layer does not provide information about the type of station. From Figure 9 it is noticeable that there are large areas that do not have coverage of stations.
From information provided by COPECO during meetings, the quality and continuity of information from rainfall stations require a close assessment to determine their usability. Due to maintenance issues and lack of standard validation procedures quality may be affected.
19 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 9. Climatological stations in the Sula Valley (1981-1999). Source: http://www.sinit.hn/
In 2012 a meteorological radar started operation by COPECO, acquired in the framework of the project “Modernización del Equipamiento de Observación Medioambiental y Protección Civil” coordinated by COPECO. Besides Honduras, the radar covers 100% of El Salvador, 90% of Guatemala, 100% of Belize and 90% of Nicaragua (ReliefWeb 2012). Figure 10 shows a screenshot of a radar image showing precipitation intensity. This information is highly valuable for hydrologic studies providing high resolution data at spatial and temporal level. Radar data despite the short time series (6 years) should be analyzed for use in flood hazard modeling. A quality assessment for hydrologic modeling purposes should be carried out.
20 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 10. Imagen del radar meteorológico actualmente en operación por COPECO. Fuente: http://copeco.gob.hn/Radar-SRI
5.1.5 Stream flows
Streamflows are available from stream gauging stations (if they have a rating table). Alternatively, hydrographers can undertake measurements during floods. The information is useful to determine flow magnitude and flood frequency (Victoria State Government 2016).
Streamflow can be of two types (BMT WBM Pty Ltd 2017):
• Event Flow Gauging: measurement of flow using current meters during an event can yield direct flow estimates during an event. These flow estimates can be compared to model predictions for calibration purposes. • Water Level and Flow Hydrograph: Water level hydrographs recorded by stream gauges can be used to assist in understanding the inflows into a catchment or study area. They are used to calibrate and verify the flood modeling and, where sufficient data exists, inform a Flood Frequency Analysis (FFA). The water levels recorded at a stream gauge site can be
21 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
used to derive stream flows at the same location and the agencies that operate the gauge will typically provide both the recorded water level, derived stream flow estimates and a guide as to expected accuracy/quality of the flow estimate.
Although water level stations exist in the study area, the quality of the data should be carefully assessed. The flood hazard studies carried out by PGRD-COPECO used stream flow time series until 1998, -the year of occurrence of Hurricane Mitch-, due to lack of continuity of the time series after that year in many of the stations. In addition, rating curves require periodic updating and time series require validation procedures2.
Since the quality and quantity of available data will affect every step in the modeling process, including modeling approach, schematization, calibration and analysis, the available data needs have to be reviewed to ensure it is adequate for the type of study proposed (BMT WBM Pty Ltd 2017). The lack of streamflow data will affect particularly the calibration of models (INDECI 2011), therefore, suitable methods and criteria to improve reliability should be taken into account into the modeling process.
5.1.6 Historical data
Historical data are very important for public awareness rising as well as for the calibration of flood modeling (as long as past and modeling conditions can be compared). Historical data interesting to be collected are (EXCIMAP 2007):
• Flood maps: A plan of the site showing any existing information on extent and depth of flood events. Information may be anecdotal, photographic or from survey results. The events should be identified with date/time, source of the data and supporting information provided on rainfall and/or return period, or probability of occurrence. Recorded data are particularly valuable and, if available, should be highlighted along with evidence of any observed trends in flood occurrence (SEPA 2014). • Water level records in river • Velocity records (gauge) • Flood marks: high water marks (for specific floods, indicated by silt deposits or debris), traditional knowledge, anecdotal information and paleoflood analysis (Profesional Engineers and Geoscientists of BC 2017) are very useful to understand the flooding mechanisms and to record the extent of past floods.
Flood level surveys are best undertaken soon after a flood, when multiple recent flood marks are available and people’s memories are fresh. Apart from the basic requirements of a licensed surveyor (preferably with experience in surveying flood marks) the following is required (Victoria State Government 2016): (i) Interviews with landholders and flood emergency response personnel about previous flood heights and extents; (ii) Flood marks should be captured on both sides of the floodplain and, if possible, spread out across the floodplain; (iii) Levels should be obtained at the locations of previous recorded flood levels if possible; (iv) A robust and standard method of documentation (consider use of laptops, spreadsheets, scanned imagery for locating and recording flood mark locations and GIS capability); (v) Field sketches and photographs; (vi) Use of suitable reference maps (hard copies and/or digital); (vii) Locations of flood marks in the appropriate coordinate system;
2 Meetings with PGRD-COPECO team
22 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
• Pictures, painting or drawing • Newspapers relating flood events • Historical reports or books on floods, focusing on damages and on protection upgrade: It is recommended that when a formal flood report is produced to document and review the flood, a reasonable attempt is made to collect flood damage data and to assess flood damages, including damage to waterways. Relevant information will include which properties and buildings were inundated and to what depth, and which roads, and other infrastructure were damaged or suffered reduced services (Victoria State Government 2016). • Aerial and satellite photos: Aerial photography provides a historic record of the extent of flooding at a given date and time. It provides useful data to confirm areas flooded and to verify flood behavior. The quality and the usefulness of the information will depend on the available budget, the coverage and the prevailing weather conditions. If flying conditions are restricted because of bad weather or there are cost issues, a hand-held camera (or video recorder) out of a plane or helicopter might be the only realistic option (Victoria State Government 2016). If weather conditions and budget permit, fully orthorectified photography (i.e. photos corrected for camera tilts and terrain distortions) is recommended. For slow-moving floods across large floodplains, it may be necessary to arrange flood photography over a number of days, to ensure flood extents close to the flood peak are captured (Victoria State Government 2016).
Flood extents from international agencies obtained from satellite imagery are available for the Sula Valley. The Centro de Agricultura Tropical (CIAT 2001) produced the “Atlas of Honduras” compiling information on the Hurricane Mitch that includes the processing of satellite SAR/RADARSAT images taken in October and November 1998. The analysis produced three layers: areas fully flooded, areas partially flooded and flooded vegetation areas. These are shown in Figure 11.
In addition to the CIAT data, other sources exist. The Dartmouth Flood observatory 3 manages a database of flooded areas obtained from MODIS imagery (see Figure 12-a). UNOSAT4 has also generated information using RADARSAT imagery (see Figure 12-b). This information is highly valuable for regional analysis and constitutes a starting point for more detailed hazard analysis. It is best used as a complement to other available hydrologic and climatic data. It is useful in preliminary assessments during the early stages of a development planning study because of the small-to-intermediate scale of the information produced and the ability to meet cost and time constraints. The data may also be applicable to other aspects of the study (OEA 1991). Potential applications for satellite photography include (Victoria State Government 2016):(i) use as a supplement to flood photography to provide an overview of flood behavior at a coarser scale; (ii) in lieu of flood photography where a high resolution is not required (e.g. large-scale rural flooding with slow moving floodwaters in sparsely inhabited areas); and (iii) As a planning tool for flood response operations.
3 http://www.dartmouth.edu/ 4 https://unitar.org/unosat/
23 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 11. Flooded areas during Hurricane Mitch. Source: (CIAT 2001)
24 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 12. a) Flood map September 6-9 de 2007. Source: Darthmouth flood observatory http://www.dartmouth.edu/~floods/2007177.html; b) Flood map November 1 2008: Source: UNOSAT
5.1.7 Land Use and land cover
Land cover can be determined by analyzing satellite and aerial imagery and constitutes important information for hydrological modeling. A full coverage currently exists for the Sula Valley. However updating is important.
Regarding land use, this is important since macro and mesoscale flood risk analysis can apply aggregated depth-damage functions attributable to land use types (EXCIMAP 2007), micro-scale studies should use at least for buildings an object-oriented assessment (Messner et al. 2007).
5.1.8 Assets data
25 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
In order to measure damage in monetary terms, information on the value of assets at risk needs to be quantified (Messner et al. 2007). The economic value of an asset can be quantified using at least two main approaches: (1) construction value (e.g. cost of material, work, and other related expenses needed to build the asset); (2) market value (e.g. cost of the asset on the local real estate market) (Sterlacchini et al. 2014). Market values were preferred in the analysis as they describe better the actual distribution of economic activities and prices (Sterlacchini et al. 2014).
The characteristics of the assets are also required for vulnerability analysis. In the case of buildings and residences, the data to be collected commonly includes:
— Information on the location, number and type as well as the elevation. This can be either primary data from field surveys or secondary data, i.e. data from already existing sources. Furthermore, the spatial resolution and level of categorization can be highly diverse. It ranges from broadly aggregated land use information (e.g. industrial area, residential area) to detailed information on the type of every single building or property (Messner et al. 2007). — Floor levels of buildings (Victoria State Government 2016). — Presence of a basement, their characteristics and usage (NFIP 2015). — The number of floors and the type of supporting foundation (NFIP 2015). — The type of structure or architecture, the type of interior and exterior finishes (e.g., brick vs. siding; wood vs. vinyl floors), and the quality of construction (NFIP 2015). — Contents of the residence or building, commonly obtained through interviews (Villanón 2003). — Coverage of public utilities coverage. This information is important in order to assess secondary impacts due to interruption or pollution (Villanón 2003). — Location and characteristics of critical infrastructure (schools, hospitals, transportation, energy supply, communication, water supply, sewerage)
A cadastral database that provides basic information on buildings and residences does not exist in the Sula Valle. The generation of this information is of great importance for flood risk assessment. This can be carried out from satellite imagery and LiDAR data (already available for the study area) (ESRI 2018) (Villanón 2003).
5.1.9 Demographic and socio-economic data
Demographic and socio-economic data are used for the assessment of exposure and vulnerability and finally for the assessment of risks in a given location (World Meteorological Organization 2013). The Instituto Nacional de Estadística – INE is the main source of demographic and socio-economic data. Limitations arise from the resolution and need for updated data. Global sources such as WorldPop5 can also provide useful data.
5 http://www.worldpop.org.uk/
26 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
5.2 Flood Hazard assessment and flood mapping
Three main different approaches can be used for flood hazard assessment and to produce flood maps (World Meteorological Organization 2013):
• Historic approach is based on past flood events: Written reports, old maps or photographs or any other document may provide relevant information to delineate flood zones. • Geomorphologic approach: Floods and flows leave distinct marks of past events in the landscape. • Modelling approach: Hydrological and hydraulic models are applied to simulate floods of a particular magnitude occurring in a channel/channel system.
The historic and geomorphologic approaches are mainly used for preliminary and general purpose flood assessments, while the modelling approach is used for detailed flood assessment (World Meteorological Organization 2013). For the Sula Valley a flood susceptibility map was developed in 2016 using a geomorphologic approach (see Figure 13). Therefore, only the modelling approach will be taken into account from this point forward.
As explained in the previous section, the selection of a modelling approach (hydrologic and hydraulic model types) depends on the level of complexity required according to the use of the information. This fit-for-purpose approach leads to the minimum required actions to meet the local needs however if funds allow, a more detailed and/or complex approach could be used. Similarly, the flood study and flood modelling process can be iterative, with opportunities to add detail or complexity to the model at a later date (BMT WBM Pty Ltd 2017). The possibility to use/operate models for several purposes and the need for an updating process and progressive increase in complexity as needed, requires that models are well documented and that all data and information used for the setting of the model are available. It is recommended that models are appropriately archived and supported by information that allows any professional with the required experience and training, to fully understand and reproduce the construction of the model, use it and update it.
27 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 13. Flood susceptibility in the Sula Valley. Source layer: (PGRD-COPECO 2016) The flood hazard studies involve the following steps, which are shown in Figure 14 (Council 2015):
1. Hydrologic analyses are conducted to estimate the river discharge rates. Depending on data availability, the discharge rate is estimated using (a) statistical analyses of historical annual maximum discharges measured at stream gages; (b) regression equations derived from observations at similar locations in the region to estimate discharge as a function of drainage area and other river basin characteristics; or (c) precipitation-runoff models, which convert rainfall to stream discharge rates. 2. Hydraulic modeling is carried out to determine the depths and velocities that correspond to the river discharge rates estimated in the hydrologic analyses. 3. Comparisons of estimated water surface elevations at river cross sections (or cells or polygons) to the ground elevations along the river are made to define the extent and properties of the inundated floodplain. If the computed water surface elevation for a point
28 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
or cell is greater than the ground elevation, then the point or cell will be inundated.
Figure 14. Steps for flood hazard assessment and mapping. Source: (Council 2015) The integrated simulation of hydrologic analyses with rainfall–runoff and flood inundation is also possible and has a tremendous advantage, especially when flood inundations occur at multiple locations in a basin, or when it is difficult to separate rainfall–runoff and inundation processes within a basin.
Details on the methods used for hydrologic and hydraulic modeling are presented in the following subsections.
5.2.1 Hydrologic analyses
There are three main approaches to estimating extreme flood generating river discharge, namely from hydrological models (rainfall-runoff models), regional/local regression and statistical analysis/flood frequency analysis (see Figure 14).
Flood frequency analysis is carried out on the maximun discharges to estimate discharges associated to different return periods (INDECI 2011).
Regional/local regression analyses work on the assumption that data from data-rich regions can be transferred to data poor ones. More specifically, these methods assume that catchments with similar characteristics will also exhibit similar flood frequency statistics. The application of this type of analyses is limited by uncertainties in observed discharge records which are known to be significant for extreme flows. Moreover, the estimation of extreme flow behavior using catchment characteristics is further hindered by human modification of river systems that change natural extreme flow behavior (Sampson et al. 2015).
29 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Rainfall-runoff models convert a spatial and temporal description of a given frequency storm over a watershed into a flood flow hydrograph at the outlet or concentration point of the watershed. A hydrograph represents the passage of a flood wave at a point usually expressed in terms of discharge as a function of time (FEMA 2018a). The main advantage of a hydrologic simulation approach is that it can be applied in sparse stream gauge settings. Furthermore, it takes into consideration the river network structure’s role in shaping the spatial pattern of flooding (Czajkowski et al. 2016).
Hydrologic models range in complexity from simple spatially lumped models to complex spatially distributed models. There are essentially three types of rainfall-runoff model (Prinos 2008):
Unit Hydrograph models: are black box models which links the input data to the output data. The model has to be calibrated with some parameters that can be derived from historical time series.
Physically based models: the physical model is based on the exact underlying equations and solves them in a precise way. This process is very data and time intensive.
Conceptual models: A conceptual model seeks to describe the hydrology of a drainage basin from rainfall to stream discharge as a sequence of interlinked processes and storages. The processes and storages a conceptual model contains are bases on underlying physical equations without solving them explicit.
Regardless of complexity their performance depends greatly on precipitation data quality (Sampson et al. 2015). Another crucial aspect that determines performance relates to calibration. The use of hydrological models where calibration data do not exist, as it is well known that the prediction of extreme flows in ungauged catchments remains a key challenge to the field of hydrology (Sampson et al. 2015).
The following three decision trees are to be used to guide the selection of the most appropriate hydrologic approach based on the existing inflow data available for the flood study. Generally, the more data sets and the longer the length, the better confidence can be applied to the flood study’s hydrology (BMT WBM Pty Ltd 2017).
If the study area contains multiple inflow data sets If the study area contains limited inflow data sets
30 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
If the study area contains poor inflow data sets:
Figure 15. Choice of most appropriate hydrologic approach based on the existing inflow data available for the flood study. Source: (BMT WBM Pty Ltd 2017) Hydrometric authorities (COPECO and CENAOS) hold flood data for many gauging stations. However, the lack of continuous and reliable data implies that a rainfall-runoff modelling approach may be more appropriate for the Sula Valley with limitations on the calibration procedure.
There are a number of steps involved in determining the magnitude of flood discharge and hydrographs. These are shown in Figure 16.
Design storms are commonly obtained from depth-duration-frequency relations. The depth values for a given frequency and duration are used to draw isohyets, or lines of constant depth, creating a map from which the rainfall depth for that particular frequency and duration can be found (FEMA 2018a). The spatially averaged depths of rainfalls with large areal extents are, in general, less than those with relatively smaller areal extents. In practice, an areal adjustment factor is applied to depth values derived from those relations (FEMA 2018a).
Calibration of runoff, sub-basin response, and routing parameters are performed through modeling major historic storms over the watershed where rainfall and outflow data are available. By comparing the measured outflow from a storm to the modeled outflow, the modeler can judge the reliability of the model and adjust input parameters accordingly (FEMA 2018a).
31 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
•Definition of watershed and subwatersed boundaries Delineation of study area
•Determined directly from data or modelling approach Selection of modelling approach
•Collect data for model setup, calibration and verification Data collection and collation for model setup
•Discharges, rainfall and soil moisture Model validation
•Simulation for design storm events Flood discharge simulation
Figure 16. Steps involved in determining the magnitude of flood discharges and hydrographs. Adapted from (World Meteorological Organization 2013)
5.2.2 Hydraulic modeling
Hydraulic modeling simulates how the flow determined by the hydrologic model moves through the river system. Inflow hydrographs, estimated using the hydrologic modeling, are applied at the upstream ends of waterways and floodplains. The essential aim of hydraulic modeling is to represent realistic flow behavior to obtain flood levels, velocities, discharges and depths (BMT WBM Pty Ltd 2017).
32 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
The choice of model depend on the work which has been undertaken in the past, what data and information are available for the catchment, type of catchment, flood risk, end-user needs and the objectives for the flood study (BMT WBM Pty Ltd 2017). Basically, the following hydrodynamic models may be distinguished (BMT WBM Pty Ltd 2017) (Merwade et al., 2008) (Tymkow et al. 2016) (see Figure 17):
— One-dimensional approach (1D): based on the assumption that the speed component compatible with the direction of the axis of the river constitutes the dominant one, use the description of the geometry of the flow area in the form of the cross-sections which are perpendicular to the direction of the flow (Merwade et al., 2008). In the case of such an approach, the impact of land-use objects, such as buildings or vegetation, on the wave distribution is usually expressed through alternative resistance factors. The use of 1D models is restricted to modelling single waterway branches, or simply connected channel systems, where flow in the floodplain is well connected to the main channel (BMT WBM Pty Ltd 2017). — Two-dimensional approach (2D): 2D systems are used in the simulation of the flows in the riverbeds, where the vertical component of the speed and acceleration vectors are much smaller than the components in the horizontal plane. The characteristics of motion parameters are therefore determined by the longitudinal coordinate. The geometry of the valleys and the riverbed in these models is represented with the use of the DTM (2.5D geometry) whereas the roughness coefficients, which amongst others include resistance associated with vegetation, are presented in a spatial manner by referring them to the individual meshes of DTM grid (Merwade et al., 2008). 2D models are capable of providing a detailed description of the flow in urban or rural floodplains and overbank areas, and can be integrated with 1D model elements (BMT WBM Pty Ltd 2017). Unsteady flow simulations of a two-dimensional model have the capability to more accurately account for the movement of water and storage within a wide area of the floodplain. The two-dimensional solution has the ability to accurately model unsteady, unconfined flows; however, rating curves are necessary to reflect control structures within the floodplain. Two-dimensional hydraulic models are used to determine flood elevations for wide floodplains caused by flat topography; for these floodplains, the basic assumption of unidirectional flow is violated, and one- dimensional models may not provide reliable results (FEMA 2009). — Combination of 1D and 2D: One important limitation of standard 2-D approaches over large domains is the inability to represent rivers whose width is considerably smaller than the grid size. Where grid scales may be limited by terrain data resolution, subgrid methods and hybrid 1-D/2-D models have emerged as potential methods for representing such channels (Sampson et al. 2015). They have the capability to provide one-dimensional unsteady flow solutions within the stream channel and two-dimensional flow solutions for the wide overbank floodplains. They have the capacity to model channel flow as one-dimensional and overbank floodplain flow as two-dimensional. Such models should be used for floodplains with clearly defined channel system (FEMA 2009). — Three dimensional approach (3D): 3D models allow for a complete description of the flow, taking into account all the components of the velocity vector. This models are rarely applied to flood mapping and are confied to the study of small areas (Prinos 2008).
33 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 17. Types of numerical models used for inundation modeling. Source: (Asselman and Team 2009) The applicability of the models according to area characteristics and data is shown in Table 3 and Table 4 presents the description of some of the most used software codes for flood modeling.
Table 3. Overview of hydraulic model types and their application. Source: (Asselman and Team 2009) Model Description Models such as Solve the one-dimensional St Venant equations are used to describe flow processes in compact Infoworks RS, ISIS, as channels. They are applied in the case of design scale modeling which can be of the order of 10s well as Mike 11, HEC- to 100s of km depending on catchment size. Surveyed cross sections of channel and floodplain, RAS (1D application) upstream discharge hydrographs and downstream stage hydrographs are provided as inputs to
34 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
and SOBEK-CF the models aforementioned. Water depth and average velocity at each cross section, inundation extent by intersecting predicted water depths with Digital Elevation Models and downstream outflow hydrograph are produced as outputs of the models (Prinos 2008). they are 2D excluding the law of conservation of momentum for the floodplain flow. They are HYDROF and used in the case of broad scale modeling or urban inundation depending on dimension cells. HYDROF simulates the process of flood volume propagation. It is an available model of medium LISFLOOD-FP cost, accuracy and medium run-time. It uses Digital Terrain Maps (DTM), time series and overflow discharges as input and produces results of flood depth, flood extent and flood duration. LISFLOOD-FP is a 1D/2DH hybrid model of the same characteristics as HYDROF. When used to simulate flood inundation in compound channels it uses DEMs, upstream discharge hydrographs and downstream stage hydrographs as input and produces outputs of inundation extent, water depths and downstream outflow hydrographs. The computation time needed is measured in hours (Prinos 2008). Are 2D models, which solve the two-dimensional shallow wave equations. Mike 21 is similar to Mike 21 and the previously analysed models, while TELEMAC 2D is a finite element 2DH available software of high cost, medium accuracy and medium run-time (hours to days). Except from flood volume TELEMAC 2D propagation, TELEMAC 2D simulates the processes of percolation and seepage (Prinos 2008).
Is a 2DH finite element model of medium cost and run-time and high accuracy. Except from flood FINEL 2D volume propagation, FINEL 2D simulates the processes of percolation and seepage. DTMs, time series and overflow discharges are used as input to the model, while flood extent, duration, flood depths and flood plain flow velocities are extracted from the model (Prinos 2008). TELEMAC 3D is an available finite element 3D model included in the category of Second TELEMAC 3D and Generation models. It is a model of high coast, accuracy and high run- time. Except from flood volume propagation, TELEMAC 3D simulates the same physical processes and is characterized of Delft-3D the same input and output as TELEMAC 2D. TELEMAC 3D and Delft-3D are usually characterized as 2D+ models, 2D plus a solution for vertical velocities using continuity only, which are predominantly applied to coastal modeling where 3D velocity profiles are important. They have also been applied to reach scale river modeling problems in research problems. Inputs needed by the models are DEMs, upstream discharge hydrographs, inlet velocity distribution and downstream stage hydrographs. Outputs produced are inundation extent, water depths, u, v, w velocities for each computational cell and downstream outflow hydrographs. The time demanded for computation is in the range of days (Prinos 2008). is a 3D finite element model of high cost, accuracy and run-time. Simulated processes, inputs FINEL 3D and outputs are similar to those of FINEL 2D (Prinos 2008).
are usually applied for local predictions of three-dimensional velocity fields in main channels and CFX, FLUENT and floodplains. They provide a 3D solution of the three-dimensional Reynolds averaged Navier Stokes equations. Inputs necessary for these models are DEMs, upstream discharge hydrographs, inlet PHOENIX velocity and turbulent kinetic energy distribution and downstream stage hydrographs. Outputs produced are inundation extent, water depths, u, v, w velocities and turbulent kinetic energy for each computational cell and downstream outflow hydrographs. Computation time is estimated in days (Prinos 2008).
Table 4. Description of some commonly used software for flood modeling The German LAWA (La�nderarbeitsgemeinschaft Wasser, 2006) gives recommendations on which model should be used as follows: 1D models can be a good choice in mountain and low mountain range areas when the stream flow takes mainly one direction. 2D models should be used in plain areas or estuary areas or areas that are affected by tide. In these areas the assumption of one dimensional flow is false. There can also be made a distinction between steady and unsteady models. When the change over time is important for the distribution of the water volume in the area, an unsteady calculating model should be used (Prinos 2008).
The recommendations of applicability according to (BMT WBM Pty Ltd 2017) are shown in Table 5.
Hydraulic Model Common applications Reference Type number
35 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Dynamic 1D/2D Suitable for small urban catchments (overland flow) where a detailed I overland understanding of the overland flood behaviour is required.
Dynamic 1D/2D Suitable for large urban (riverine) catchments where a detailed understanding of II riverine riverine flood behaviour is required including storage, timing and spatial distribution of velocity and hazard Dynamic 2D rural Suitable for large rural (riverine) catchments with a township where a moderate understanding of flood behaviour is required including storage, timing and spatial III distribution of velocity and hazard. Dynamic 1D or coarse Suitable for rural (riverine) catchment with scattered development where a broad IV 2D scale understanding of flood behaviour is required.
This type of model is only appropriate when storage or timing are not relevant and flow is largely within the watercourse and the immediate overbank area. Hence 1D steady state this type of model is generally not suitable for most flood studies and is more likely V used for individual infrastructure design (i.e. waterway crossings). Historical Information This includes a combination of historical flood extents/flood frequency VI analysis/digital elevation model etc. to define the historical flood extent, levels and depths in the area.
Table 5. Reccommended modeling approach according to application. Source: (BMT WBM Pty Ltd 2017) According to the characteristics of the Sula Valley it is expected that mainly models type II is required, although type III and type I may be suitable for some areas.
In the case of coastal areas, Figure 18 presents the physical system with the four physical zones: offshore, nearshore, shoreline response and flood inundation zone. Coastal flood studies are similar to riverine flood studies, but they also assess the effects of storm surge (water piled up against the shore during a storm) and tidal- and wind-driven wave action. The studies use data on fetch (the distance over water that the wind blows in a single direction), near-shore terrain and water depths, and wind speed to predict storm surge properties. Data on past storms from gages and historic high water marks are used with statistical and conceptual models to determine the storm surge elevations. Next, transects perpendicular to the shoreline are surveyed to determine onshore and offshore ground elevations. The elevations are then used to compute the height of wave crests and wave run-up (the rush of waves up a slope or structure). For coastal flooding, the base flood elevation is the stillwater elevation plus wave run-up, or the wave crest elevation, whichever is greater (FEMA, 2011) (Council 2015).
Coastal flood hazards can be grouped as follows (Profesional Engineers and Geoscientists of BC 2017):
1. Storm surges in combination with high tides, waves and/or river flows: Site-specific hydraulic modelling may be required to provide refined estimates of deep water storm surge to account for regional coastline type, local characteristics such as shoaling and shallow water, and nearshore features such as estuaries, spits and seawalls. 2. Tsunamis: Tsunamis can be caused by nearby or distant earthquakes and large landslides (above or below water). As tsunami wave heights are very dependent on site-specific conditions, detailed modelling is required to determine potential run-up at a given location. 3. Ongoing sea level rise due to climate change.
Table 6 presents some widely used models for coastal areas (Defra, 2003).
36 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 18. Characterization of the physical system for coastal areas. DEFRA/ Environment Agency Flood and Coastal Defence R&D Programme (2003) - Best practice in coastal flooding, R&D Technical Report FD2206/TR1, Workshop Version issued 24/09/03, HR Wallingford Report TR 132. Source: (Prinos 2008)
Table 6. DEFRA/ Environment Agency Flood and Coastal Defence R&D Programme (2003) - Best practice in coastal flooding, R&D Technical Report FD2206/TR1, Numeric Models for prevision in coastal areas Workshop Version issued 24/09/03, HR Wallingford Report TR 132. Source: (Prinos 2008) The sequence of steps for hydraulic modeling is shown in Figure 19. The recommendations on average errors are intended for detailed flood hazard analysis.
37 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
•Define upstream and downstream end of the model area Delineation of •Define computational domain study area
•Dimensions 1D, 1D2D or 2D •Coverage of the flood dynamics (steady or unsteady Selection fo •Appropriate flow resistance apporach modelling •Handling of hydraulic jumps and structures approach
•Geographic data: topography •Land use •Bathymetry Data collection •Stage discharge relationships, historic flow data, hydrographs and roughness and collation parameters
•Water depths of the observed flood events reproduced with an average error of less Model than +/- 5 cm calibration and validation
•Water levels and flow velocities Flow simulation
Determination •Delineation of inundation area of the inundated area
Figure 19. Steps involved in determining flood prone areas by hydraulic models. Adapted from (World Meteorological Organization 2013) The calibration and validation step is of high importance. This allows establishing a degree of confidence that the models are suitably representing actual site conditions (BMT WBM Pty Ltd 2017). Standard practice for hydraulic modeling includes calibration of a model to a known data set, if
38 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
available, such as a river profile, recorded flood levels, flows at gauging stations, peak flood levels from field survey, photographs and videos, anecdotal evidence of flood behavior or observed high water marks, by adjustment of model parameters, such as friction factors and other loss coefficients (Profesional Engineers and Geoscientists of BC 2017)(BMT WBM Pty Ltd 2017). Once the model is calibrated, validation is carried out, simulating other event observed in the watershed, different from the events used for calibration. If the simulation results are acceptable the model is accepted, otherwise calibration must be repeated (INDECI 2011). FEMA recommends that models should match known high-water marks within 0.5 foot (FEMA 2009). In the absence of suitable calibration data, engineering judgment must be applied to estimate the required parameters (Profesional Engineers and Geoscientists of BC 2017).
Calibration can be undertaken on the hydrology and hydraulic models separately, or on both models concurrently, as part of a joint calibration. The calibration exercise can be lengthy and iterative and the definition of calibration ‘success’ should reflect the quality of the underlying data and the nature of the catchment and flood event (BMT WBM Pty Ltd 2017).
Model sensitivity analysis is an important aspect of model establishment. It is particularly important if model calibration or validation is unlikely. Sensitivity analyses assess the degree of influence different model parameter values have on the results of the calibration and validation. Sensitivity analyses can provide an indication of the relative uncertainty associated with design model results. Minimum parameters which must be tested in hydraulic models include roughness/friction, energy losses, bridge coefficients and boundary conditions (e.g.tailwater levels from tides/sea level rise for coastal studies) (BMT WBM Pty Ltd 2017). Sensitivity assessments are also important to understand how changes in parameters affect the resulting flood behavior. These analyze impacts of climate change, (sea level rise, rainfall intensity increases and/or other climate change variations).
With respect to model application, the following are aspects to take into account:
Grid and time step resolution (Asselman and Team 2009)
— The grid must cover the entire project area and the derivation or development of the grid must be clearly documented. The Mapping Partner should carefully select cell size, not only considering the accuracy of the topographic data and computational efficiency of the model, but also mapping and floodplain management needs. Too small a cell size not only slows computations, but also creates too many elevation grids, which may not practically be presented on the floodplain (FEMA 2009). — Grid and time resolution depend on the resolution of available terrain data, the length scales of terrain features in the domain and the length and time scales of relevant flow processes. — Grid resolutions in rural areas with gentle/regular/ topography can be coarser than grids developed for areas with a more complex topography or urban areas. The minimum grid resolution that is required for flood simulations in urban areas depends on the characteristics of the city (e.g. size of the streets) and on the required information. In general, less resolution is required if only water level is to be predicted, finer resolution if the velocity field is also required for flood characterization. — In urban areas, if the user is interested in water depths only, the minimum grid cell size should equal half to once the width of the streets (e.g. 5 to 10 m in most European cities). In rural areas where villages make up a small part of the total model area, larger cells can be used, e.g. grid resolutions of 50 to 100 m are quite common in the Netherlands. — Water depths in urban areas can also be computed quite accurately by increasing the hydraulic roughness or by adopting a porosity approach. When information on flow
39 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
velocities is required, the grid cell size should always be smaller than the width of the streets. — Increasing the hydraulic roughness or application of the porosity approach does not result in accurate velocity estimates. — Too large a cell size either creates flat water-surface elevations over a large area or does not accurately define the flood boundaries. Gradual changes in elevation must be maintained from one cell to adjacent cells to avoid numerical instability; they should not use too many cells along a cross section to avoid unnecessary difficulty in maintaining surcharge in the floodway calculation (FEMA 2009). — Too big a time step, although resulting in a lesser computational time, may lead to model instability and increased mass balance errors. Thus, choosing an appropriate time step is vital for ensuring increased accuracy while reducing the computational requirements. Some software packages have a facility for specifying an adaptive time step which may be used to avoid modeler specific uncertainties (SEPA 2014).
Roughness
Relevant input for the hydraulic models is the roughness distribution and the parameters of the flood plain and in the riverbed to determine flow resistance. A reliable evaluation of the roughness parameters needs good experience. Aerial images, land use maps and biotope distribution maps can be used for a first classification of roughness zones but results require thorough field check. Roughness values selected should be well documented. The selected values should be well justified and explained (Nottawasaga Valley Conservation Authority 2013). The basis to evaluate the roughness should include:
Taking bed material samples at various locations of the stream bed and banks and determining the grain size distribution curve in the laboratory (World Meteorological Organization 2013)(FEMA 2009). Often only visual control of the bed material is possible (World Meteorological Organization 2013).
The slope of the channel (FEMA 2009).
The type and density of vegetation in the floodplain (FEMA 2009).
The degree of meandering (FEMA 2009).
The expected depth of flooding. The variation of roughness coefficient values with flood stage should be considered, depending upon factors such as the width-to-depth ratio of streams, vegetation in the channel and overbanks, and materials of the river bed (FEMA 2009).
Roughness coefficients in overbank areas should be carefully selected to represent the effective flow in those areas (FEMA 2009).
Boundary conditions
Where a control starting elevation (such as a weir) is not possible, the starting section shall be located sufficiently downstream that the reach under consideration is not significantly affected by the starting elevations, typically at least three cross sections downstream of any significant change in channel form (Nottawasaga Valley Conservation Authority 2013).
Even though detailed hydraulic models have improved in recent years, they still have significant limitations for operational use over large areas. Limitations include (Czajkowski et al. 2016): (i)
40 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
high implementation cost; (ii) excessive computational time; and (ii) large data requirements.
River morphology
Floodplain changes that may affect the water-surface elevations, flood extent and/or velocity must be incorporated in the hydraulic model. These changes should have been identified during validation of the engineering data and include (but not limited to) (FEMA 2009): (i) Development within floodplains; and (ii) Changes in the alignment of the stream, the carrying capacity of the channel, and other geo-morphological changes; (FEMA 2009).
Structures
The modeling of each hydraulic structure must be documented to include a list of each grid or element associated with the structure, and a description of the rating table including the derivation and sources of data (FEMA 2009).
For hydraulic structures that are designated to divert flood flow from its natural path, such as flood gates and diversion channels, the documentation must include the owners and operators of the structure; the date it became operational; operation, inspection, and maintenance plans; and as- built plans describing the dimensions and identifying any moving parts (FEMA 2009).
Flood risk mitigation structures such as levees, require consideration of their performace to decide the way in which they are modeled. For example, when they are considered to meet standards, are modeled as blocking flow onto the floodplain, and levees that do not meet the standards are modeled as if they fail to protect (Council 2015). In the simulation of levee failures the timing of breach initiation and the breach growth rate determine to a large extent the volume of water that flows into an area. This implies that breach initiation and growth have a very large effect on computed water depths and flood extent. Unfortunately, breach initiation and growth are difficult to forecast and simulate, due to the complex mechanisms involved and to the stochastic character of breaching initiation. The advice therefore is to always carry out a sensitivity analysis to determine the uncertainties in the model results related to uncertainties in breach initiation (timing) and growth rate (Asselman and Team 2009). Various breach locations should be assessed to establish the worst case scenario, and combinations of breaches should be included in the analysis. Two-dimensional hydraulic modelling is now the standard for representing the propagation of the flood wave from a dike breach across the floodplain. Such modelling should take into account all structures influencing the flow, such as roads, bridges and culverts, existing and future development, remaining dikes and other embankments. Output from this modelling can be used to indicate areas subject to flooding in the event of a dike breach (Profesional Engineers and Geoscientists of BC 2017).
Figure 20 shows the likelihood that a levee will fail to function as designed (breach), conditioned on channel water surface elevation. Similar functions can be developed to describe the likely performance of other local mitigation measures (NFIP 2015).
41 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 20. Levee fragility function. Source: Courtesy of David Ford, David Ford Consulting Engineers, Inc. Source: (NFIP 2015) Regarding dams, one of the concerns about dams is its safety and the possibility of serious accidents including the dam failure. This concern is particularly important for people living along the valley downstream the dam (Betâmio De Almeida 2001). Dams are also a potential danger for downstream valleys: whatever the cause and the probability of occurrence, a failure in the water retaining capability is always possible. In fact, valley safety need to be considered as an integrated concept closely involving both the dams and reservoirs as well as the downstream valley system comprising the people, the land and the economic occupancy (Betâmio De Almeida 2001).
A dam-break flood intensity (peak discharge, volume and flood hydrograph) will depend on several factors, as, among others, dam and reservoir general characteristics and the dam breach characteristics. The valley damages will depend on the valley vulnerability to dam-induced floods. This vulnerability will be a function of several other factors: flood intensity along the valley, warning system and flood time of arrival, land socio-economic occupancy and characteristics on flood prone areas and people survival capability. In fact, the valley risk will strongly depend on dam safety or response to hazards and on valley capability to cope and to survive to those induced floods. An integrated dam-valley risk management system can be conceptually composed by two parts: the risk assessment process, in which an approximate quantitative risk estimation and evaluation is made for dangerous situations; it includes the hazard identification and characterisation as well as the dam risk analysis; and the risk mitigation process, in which actions to reduce the risk will be identified and implemented (Betâmio De Almeida 2001).
Storm drainage systems
In urban areas the drainage system may be an important aspect of the management of flooding. Three scenarios can be considered: (i) the drainage system is highly important then the hydraulic model should include all drainage pipes; (ii) the drainage system is highly important then the hydraulic model should include all drainage pipes greater than x m; (iii) the pipe drainage network has a relatively minor capacity for the management of flooding, the hydraulic model does not need to include drainage pipes (BMT WBM Pty Ltd 2017). Most urban areas in Honduras correspond to case iii where the hydraulic model does not need to include drainage pipes. However, each urban area requires an individual assessment to ensure that any significant structure that may influence the model results is appropriately included.
42 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
5.3 Exposure
Identification and mapping of elements at risk are essential tasks for vulnerability assessment studies, providing one of the main spatial data layer required for a total risk calculation (van Westen et al. 2008). In general terms, elements at risk comprise the population, properties, economic activities, private and public services potentially threatened by a harmful event in a territory, either directly or indirectly (Alexander 2005; van Westen and Montoya 2009). Elements at risk are defined as objects which possess the potential to be adversely affected (Hufschmidt et al. 2005). Depending on the aim of the project and the working scale adopted, different levels of detail and accuracy have to be pursued during the data collection and storage phases (Sterlacchini et al. 2014).
Exposure includes (World Meteorological Organization 2013):
— Population density, indicating also to particular vulnerable groups (e.g. distribution of senior citizens homes)
— Residential area with building stock;
— Lifeline infrastructure like school, hospital, re guard station, reservoir, telecom station, power transformer and switch yard;
— Traffic infrastructure
— Industrial estate with indications for environmental relevance (e.g. chemical plants)
— Primary sector territories (agriculture, horticulture, pisciculture etc.)
5.4 Vulnerability
The term vulnerability is conceptualised in hazard and disaster management in various ways. As a consequence, the notion of vulnerability is as divergent as the methods and theories of disciplines involved in vulnerability research. Social scientists and natural scientists often address different issues when they are using the term vulnerability (Fuchs et al. 2011). Today, ‘vulnerability’ is defined, interpreted and applied in various ways, partly due to the need to work within a specific social and environmental context, partly due to a range of disciplines entering this research field, equipped with their own ontologies, definitions and methods (Hufschmidt 2011). However, what is common from the definitions is that vulnerability is: Multi-dimensional (e.g. physical, social, economic, environmental, institutional, and human factors define vulnerability); Dynamic (vulnerability changes over time); Scale-dependent (vulnerability can be expressed at different scales from human to household to community to country resolution; - Site-specific (each location might need its own approach) (Birkmann 2006).
In summary, definitions of vulnerability tend to fall into two broad categories that deal with vulnerability in terms of damage caused to a system by a particular hazard or climate-related event (hazards and hazard impacts approach) or in terms of ‘inherent vulnerability’ (‘social vulnerability’ for people, Adger et al. 2004) that is an intrinsic property of a system (community) before
43 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
encountering an external hazardous event (Sterlacchini et al. 2014).
Thematic dimensions of vulnerability have to be addressed within a holistic assessment process. Key thematic components are (Birkmann et al. 2013):
— Social dimension: propensity for human well-being to be damaged by disruption to individual (mental and physical health) and collective (health, education services, etc.) social systems and their characteristics (e.g. gender, marginalization of social groups). — Economic dimension: propensity for loss of economic value from damage to physical assets and/or disruption of productive capacity. — Physical dimension: potential for damage to physical assets including built-up areas, infrastructure and open spaces. — Cultural dimension: potential for damage to intangible values including meanings placed on artefacts, customs, habitual practices and natural or urban landscapes. — Environmental dimension: potential for damage to all ecological and bio-physical systems and their different functions. This includes particular ecosystem functions and environmental services (see, e.g., Renaud 2006) but excludes cultural values that might be attributed. — Institutional vulnerability: potential for damage to governance systems, organizational form and function as well as guiding formal/legal and informal/customary rules—any of which may be forced to change the following weaknesses exposed by disaster and response.
At last, vulnerability is a function of the objective of the study (which establish the number of dimensions to be included) and the temporal and spatial scale of analysis. Systematic vulnerability assessments have to meet the needs of the potential end-users, including public administrators (responsible for urban planning and development), economists, managers (dealing with services, buildings or other vulnerable facilities), insurance companies, lawmakers and policy makers (drafting building regulations or codes of practise for construction), people responsible for civil protection, relief and emergency services (whose job is to prepare contingency plans) (Sterlacchini et al. 2014).
The ultimate goal of vulnerability assessment should be to measure/quantify it as quantitatively as possible, so that subsequent evaluations can be carried out to determine if it is being reduced or not (Sterlacchini et al. 2014). Holistic approaches use indicators to assess vulnerability. Table 7 shows a summary of commonly used indicators for socio-economic vulnerability according to (Lundgren and Jonsson 2012).
Three approaches, which differ in scale and resolution, are available to determine and map tangible vulnerability and risk (World Meteorological Organization 2013):
— Micro-scale approach (community-level) is an object related method based on empirical damage data. This scale is recommendable for small catchments (size has to be defined: several 100 hectares in rural environment to less than 1 hectare in urban areas) or selected areas which require very detailed considerations. This method is based on interviews with people who are/were affected by flood or on real damage data. The (potential) damage can be assessed for each element separately. The micro-scale approach is in general too detailed for flood risk maps. It is primarily used in flood mitigation plans to determine the necessary measures of adaptation at each property as this scale allows precise recommendations for appropriate protection measures.
— Meso-scale approach is an area related method. This scale is normally used for larger catchment areas. For the process of risk assessment single land use types (residential area, industrial estate etc.) based on digital geographic data are aggregated to larger, more general ones. The
44 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
corresponding amount of losses is based on statistical economic values and can be described with the unit $/m2.
— Macro-scale approach is similar to meso-scale, but the data is even less detailed. The area of interest is larger compared to the other approaches (e.g. international river catchments). Sometimes more than one hazardous process is considered.
Table 7. Commonly used indicators for socio-economic vulnerability to natural hazards. Indicators are sorted by the number of authors that reference to the indicator, starting with the indicator with the most references. Source: (Lundgren and Jonsson 2012) The vulnerability indicators, defining the physical, economic, social and environmental vulnerability can be aggregated and combined into an overall vulnerability value (see Figure 21). One very suitable tool for combining and weighing the different vulnerability factors is Spatial Multi
45 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Criteria Evaluation. SMCE can also be used for hazard assessment, using an expert based approach (Westen and Kingma). The theoretical background for the multi-criteria evaluation is based on the Analytical Hierarchical Process (AHP) developed by Saaty (1980). The AHP has been extensively applied on decision-making problems (Saaty and Vargas 2001), and extensive research has been carried out to apply AHP to risk assessment. For implementing the semi-quantitative model, the SMCE module of ILWIS-GIS can be used (see Figure 22). The SMCE application assists and guides users when performing multi-criteria evaluation in a spatial manner (ITC 2001).The input is a set of maps that are the spatial representation of the criteria, which are grouped, standardised and weighted in a ‘criteria tree.’ The output is one or more ‘composite index map(s),’ which indicates the realisation of the model implemented (Westen and Kingma).
Figure 21. A model to integrate the vulnerability components into an overall vulnerability. Source: (Westen and Kingma)
Figure 22. Schematic procedure for spatial multi-criteria evaluation based on the analytical hierarchical process. Source: (Westen and Kingma) When assessing physical vulnerability, three major approaches are commonly used: vulnerability
46 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
curves, damage matrices and vulnerability indicators (Kappes et al. 2012). Vulnerability curves are usually building type-specific and link the intensity of a hazard to the expected damages or the cost of these damages related to the total value at risk (Kappes et al. 2012). In contrast to curves, damage matrices are a simpler and more widely applicable method (i.e. applicable for more hazards types). They are composed by classified intensities and stepwise damage levels (Menoni, 2006) (Kappes et al. 2012). An alternative, mostly rather qualitative, approach is the use of vulnerability indicators. In the socio-economic field, indicators are already widely used to consider the multiple characteristics of humans (age, wealth, health, education level etc.), institutions and/ or societies that contribute to their overall vulnerability (Kappes et al. 2012).
Regarding vulnerability curves, also called damage functions, or stage-damage curves, different types of elements at risk will show different levels of damage given the same intensity of hazard. This is illustrated in Figure 23, where the red line indicates an element at risk with a lower vulnerability than the green line (Westen and Kingma).
Vulnerability curves can be subdivided into two types (Westen and Kingma):
• Relative curves: they show the percentage of property value as the damaged share of the total value to hazard intensity. • Absolute curves: show the absolute amount of damage depending on the hazard intensity; i.e. the value of the asset is already integrated in the damage function.
Fragility curves provide the probability for a particular group of element at risk to be in or exceeding a certain damage state under a given hazard intensity. In figure Figure 23 there are four damage states defined (complete destruction, extensive damage, moderate damage, and slight damage)(Westen and Kingma).
47 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 23. Fragility curve (top figure) and vulnerability curve (down figure). Source: (Westen and Kingma). A typical approach is to determine the total value of the structure and its content; to categorize the structure according to its construction type, use, or other characteristics; and then to predict the damage corresponding to specific water depths using the vulnerability curves (NFIP 2015).
In the current state-of-the-art the main inundation parameter considered in these damage functions is inundation depth (depth-damage functions). Others, like velocity, duration and time of occurrence are rarely taken into account. As the susceptibility of elements at risk depends on their type and attributes (e.g. mode of construction), properties of similar type are grouped and expressed by one approximate damage function. The extent of this aggregation and categorization varies among the different approaches (Messner et al. 2007). Local engineers and architects should be consulted during the process for the definition of building typologies and the development of appropriate vulnerability functions (The World Bank 2011). A major shortcoming in disaster risk assessments for many developing countries, however, is the unavailability of vulnerability for different housing typologies and hazard profiles. This problem is largely due to significant proportions of the building stock being informal and to damages not being recorded in a systematic manner. Field visits that record building stock and relative damages in disaster-affected areas can therefore be used to provide an indication of vulnerability of a city’s building stock (The World Bank 2011).
Methods to construct vulnerability curves are summarized in Table 8.
48 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Table 8. Methods to construct vulnerability curves. Source: (Westen and Kingma). Among the methods presented in Table 8, the use of analytical models for the quantitative evaluation of vulnerability of buildings presents the following advantages: (a) independence of the necessity of past event inventories; (b) possibility of development of fragility functions for a range of harmful event magnitudes without any interpolation or extrapolation assumptions; (c) possibility to integrate the peculiarities of the threaten buildings for the study site; and (d) objectivity of the results. Its application is suggested for site-specific or local scale (Sterlacchini et al. 2014).
Vulnerability curves can be derived at the regional scale. Normally vulnerability studies are focused on the construction of the stage-damage curves related to a specific class of different human activities or land use categories. These curves can be used to estimate the damage of an object (a family house, a manufacturing factory or a land parcel) under a specific flood event. Since the hazard management studies are usually on a regional basis instead of a single object, an up-scaling from a point estimate to a regional one is necessary (Su and Kang 2005).
In the case of population loss estimation some models have been proposed (Jonkman et al. 2008) which use evacuation models and causalities functions to estimate the percentage of casualties depending of the flow velocity and water depth.
49 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
As example, the damage curves developed by RiskScape (Reese and Ramsay 2010) are shown in Figure 24.
Figure 24. Flood fragility curves for various building types, with inundation depth above the floor level (m) along the horizontal axis and average damage ratio along the vertical axis. Source: (Reese and Ramsay 2010) There are also depth-damage criteria for the estimate of economic costs from damages on vehicles (Scawthorn et al, 2006). Figure 25 shows an example of depth-damage functions for three categories of vehicles: cars, light trucks and heavy trucks (Universitat Politécnica de Valencia 2011).
Figure 25. Depth-damage curves for vehicles. Scawthorn et al, 2006. Source: (Universitat Politécnica de Valencia 2011). 5.5 Risk
Flood risk is the combination of the probability of a flood event and of the potential adverse consequences to human health, the environment and economic activity associated with a flood event (EXCIMAP 2007a). Flood risk maps extend the information shown on flood hazard maps by
50 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
quantifying the risk from a range of possible flood events and the consequences of each event (Profesional Engineers and Geoscientists of BC 2017).
Risk is defined as (Messner et al. 2007):
R=PxS
Where:
R: Risk
P: Probability of hazard occurrence, or the probability of a flood event
S: expected consequence, corresponding to the flood event
S=αSmax
Where:
α: damage factor (or vulnerability, degree of loss, damage ratio), which takes values from 0 (no damage) to 1 (total damage)
Smax: damage potential (total potential damage)
The annual average flood damage is the area under the graph of flood damages plotted against exceedance probability (the reciprocal of the return period in years). This is the area under the curve in Figure 26 (bottom right-hand diagram), and this curve should be derived from an analysis of several future floods with a range of severities, or the results will be unreliable (at least 3 and preferably 6 flood events should be used). It is important, in particular, to determine the annual probability of that flood at which damage begins, assuming that some floods cause no measurable damage. It will also be important in the coastal situation to gauge accurately the timing and annual probability of the failure of existing defenses, if that is the mechanism of flooding to be alleviated. Indeed, assessing the probability of failure of existing defenses remains one of the most difficult fields in the area of project appraisal for flood defense expenditure, since there is a very sparse research record on which to base the assessment of such probabilities (Messner et al. 2007).
51 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 26. Damage-probability Curve. Source: (Messner et al. 2007)
The expected annual damage, is given by ($/year):
Which is calculated on the basis of a number of flood events considered in pairs (Si, Pi) defined, where:
Figure 27 provides the classic four-part diagram summarising the inter-relation of hydrology, hydraulics and economics as the basis of calculating the damge probability curve. This represents the flood damage that can occur for each flood over the range of possible floods. This function can be developed for an individual structure category or for an entire portfolio of structures. The function is integrated to compute the expected annual damage for the full range of floods, also known as the average annual loss (NFIP 2015).
52 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 27. Procedure to obtain the damage-probability curve. Source: (Analysis et al. 2000) Flood risk assessment tools can be classified depending on whether they provide or not a numerical value for the risk (quantitative or qualitative) (Universitat Politécnica de Valencia 2011).
PARTIAL AND QUALITATIVE: Qualitative approaches are generally based on the experience and knowledge of the reality, without estimating a numerical value for probability or consequences. There are some limitations as they do not provide a whole knowledge of the existing risk. Therefore, in some cases, they can produce wrong conclusions about the requirement of applying measures for risk reduction (Universitat Politécnica de Valencia 2011). However, qualitative tools for measuring flood consequences could be the only option to estimate environmental and cultural losses, being widely used to describe social trauma and indirect economical effects of floods. In general, these tools are usually represented as a description of past events, lists of consequences or maps in a large scale, but they cannot provide inundation maps in a detailed scale. More detailed results should be performed from a quantitative analysis. However, they may assist a preliminary flood risk assessment, since historical flood events and environment knowledge are the basis for detailed flood risk assessments. (Universitat Politécnica de Valencia 2011).
COMPLETE AND QUALITATIVE: These methods estimate both risk components by means of a combination of qualitative methods to obtain flood components separately. One of the most common complete and qualitative tools are risk maps, which are obtained by combination of a quantitative estimation of flood occurrence, using hydrologic and hydraulic models, and a qualitative consequence estimation (Figure 1.4.2). Therefore, risk levels are obtained directly quantifying only one component. These methods identify areas where measures for flood risk reduction may be applied in first place, being a useful tool for planning and managing. However, their lack of accuracy, especially due to the estimation of consequences, should be considered as a limitation with regard to other analyses (Universitat Politécnica de Valencia 2011).
53 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
PARTIAL AND QUANTITATIVE: These tools compute numerically one of the risk components: probability of occurrence or consequences. They perform a numerical approximation of risk, but they have the same limitations related to partial analyses. Flood hazard maps are the most common method (Figure 1.4.3) and define the inundation area for different flood events, with an annual probability of exceedance. Therefore, they provide the probability risk component, without considering flood consequences. There are tools that compute only flood consequences and they can be useful to make a first approximation to the consequences of a severe flood. Methodologies will depend on the estimated consequences: economical losses or loss of life (Universitat Politécnica de Valencia 2011).
COMPLETE AND QUANTITATIVE: These tools obtain a numerical value of both risk components with the final purpose of obtaining a numerical value for flood risk, multiplying both values: probability and consequences. These tools rely on combining the computation of hazard maps and the estimation of flood consequences. Flood risk must be defined for an area which depends on the level of detail of the analysis to apply measures for risk reduction. In each defined area, the probability of inundation with a given depth is obtained and consequences are estimated. The sum of the resulting products of the probability of occurrence and consequences of each flood event will give the total flood risk in the area. In general, risk units are the one used for measuring the consequences divided per time, for instance a monetary unit or number of victims per year, as the hazard probability usually has units of time (Universitat Politécnica de Valencia 2011).
The conceptual framework for flood risk calculation (complete and quantitative) developed by (de Moel et al. 2009) is shown in Figure 28. From the analysis the following information can be mapped with regard to flood risks:
• Individual vulnerability parameter “Value” o Population: number of people, special groups, etc. – o Economic assets and activity: private property, lifelines, infrastructure, etc.; o type of production, number of jobs, etc. o Environmental issues: installations potentially damaging the environment • Potential adverse consequence (flood damage; loss per unit area) • Risk (loss per unit area in a given period of time)
Risk maps can be used for (EXCIMAP 2007):
• Flood risk management, decision making: Where is the greatest risk? Priority setting for measures • Flood risk management, planning: Select the best options and range of measures to reduce flood risk. • Emergency and crisis management at national / local level: number of people involved, evacuation route, safe havens/temporary refuge centers, hospital response plans, transport disruption.
54 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 28. Conceptual framework for flood hazard and risk calculations. The displayed matrix and curves are purely illustrative and based on a hypothetical case. In the matrix the yellow colour signifies low danger, the orange colour moderate danger and the red colour high danger. Source: (de Moel et al. 2009)
Importance of the assessment of indirect economic costs has been proofed several times in recent years by showing that indirect economic losses can easily exceed direct ones (Giacomelli 2005; Sterlacchini et al. 2007). For that reason, estimation of economic flows and trends should be an integral part of any risk analysis. Moreover, assessment of the indirect economic consequences has to be made not only within the hazardous area but also on a broader-regional basis (Sterlacchini et al. 2014). Indirect losses represent mostly disruptions of economic activities and can affect much larger areas; they include losses from interruption of transportation routes, tourist revenues, reduced real estate values, or costs of mitigation measures. In particular, the interruption of transportation can have serious outcomes in costs arising from (1) blockage of the services (transport of people and goods); (2) increased traffic and fuel consumption in case of availability of alternative routes, and (3) secondary costs for the economy arising from blocked or longer alternative routes (non-availability of goods, longer travel time, etc.). For example, Hazus flood analysis for census blocks includes in the total loss estimate values associated with inventory loss, relocation cost, income loss, rental income loss, wage loss, and direct output loss (FEMA 2018b)
Risk assessment methods for flood risk need to include the damage of all probable flood events. However, the occurrence of a flood is uncertain and consequently the total flood damage is not a deterministic but a statistical value. It needs to include the probability in the damage assessment. All probable flood events which can occur within a certain period of time need to be considered (World Meteorological Organization 2013).
55 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
6 Floodplain Zoning
Floodplain zoning is widely used to divide the floodplain into areas where the flood hazard is different, and define the types of development and land use that are suitable in each zone. The purpose of flood zoning is to prevent inappropriate development by only allowing certain types of development and land use in areas where the flood hazard is highest (Sayers et al. 2013).
The choice of events to be modeled, mapped and use for floodplain zoning might be influenced by potential flood risk, characteristics of the catchement, legacy issues, design levels of critical infrastructure and requirements for flood mitigation (BMT WBM Pty Ltd 2017). Therefore, floodplain zoning criteria significantly vary worldwide. The following are some examples of criteria that can be used as a base for discussion on criteria for the Sula Valley.
Queensland (Australia)
Table 9 illustrates the stardard design events proposed by (BMT WBM Pty Ltd 2017) for Queensland (Australia), where the defined flood event for land use planning is a 1% AEP (annual exceedance probability) flood.
Table 9. Standard design events in Queensland. Source: (BMT WBM Pty Ltd 2017) In Australia, the PMF (Probable maximum flood) is also modeled to provide an understanding of the full extent of the potential floodable area in a particular location (BMT WBM Pty Ltd 2017). The PMF is an extreme flood deemed to be the largest flood that could conceivably occur at a specific location (BMT WBM Pty Ltd 2017).
Canada
In Canada return periods of recommended design rainfall and snowmelt-generated floods and ice jam floods range from 20 to 2500 years. The lower return periods (20 and 200 years) are suggested for lower flood risk situations, while the higher ones (500, 1000 and 2500 years) are recommended where there is moderate, high or very high loss potential (Profesional Engineers and Geoscientists of BC 2017).
UK
The UK has adopted a hazard rating formula (HR Wallingford 2006) to characterize hazard intensity as a function of inundation depth, water velocity, and the potential for floating debris, primarily based on consideration of the direct risks to people exposed to floodwaters. The UK formula is (Profesional Engineers and Geoscientists of BC 2017):
56 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
HR = d x (v + 0.5) + DF, where
HR = (flood) hazard rating;
d = depth of flooding (m);
v = velocity of floodwaters (m/s);
and DF = debris factor (= 0, 0.5, 1 depending on probability that debris will lead to a significantly greater hazard).
It is useful to use a hazard rating classification framework as a proxy for physical hazard to persons directly exposed to inundation. For example, a UK hazard rating classification framework from Surendran et al. (2008) is summarized as follows:
Hazard Rating (HR) Hazard to People Classification < 0.75 Very Low Hazard (Caution) 0.75 – 1.25 Danger for Some (includes children, the elderly, and the infirm) 1.25 – 2.00 Danger for Most (includes the general public) > 2.00 Danger for All (includes emergency services)
Maps showing hazard ratings can be used to develop mitigation measures as part of land use planning at the local government level. These can be applied through official community plans, zoning bylaws and floodplain bylaws. Maps of hazard ratings can also be used to develop consequence assessments as a precursor to flood risk mapping (Profesional Engineers and Geoscientists of BC 2017).
Germany
Flood hazard zone maps show area-specific flood hazard categories based on the interplay between probability and intensity. Figure 29 shows the relationship between flood intensity and the probability of flooding, based on the following flood hazard categories: Extreme hazard (red), Moderate hazard (blue), Low hazard (yellow), Residual hazard (hatched). The fact that fields 2, 4 and 6 comprise two overlapping hazard categories is meant to indicate that allocation of a flood hazard category is a subject-specific weighting process that under some circumstances cannot be based on standard zoning approaches (Meon 2006).
Figure 29. Hazard categories based on the interplay between probability (horizontal) and intensity (vertical), categories: high, medium low, very low. Source: (Meon 2006)
57 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
The following flood intensity levels for water depth, water velocity and the like in specific areas are meant to be guidelines that users should adapt to their specific conditions (Meon 2006):
• High intensity refers (for example) to settings where humans and animals indoors are at risk and/or where buildings may sustain damage or be destroyed.
• Medium intensity refers (for example) to settings where (a) humans and animals are at risk outdoors, but incur only residual risk if they are indoors; and/or (b) buildings may sustain damage.
• Low intensity refers (for example) to settings where (a) human and animals located both outdoors and indoors incur only residual risk; and/or (b) some parts of buildings such as basements may sustain damage.
This qualitative classification system is subject to the classifications of physical parameters shown in table 3. Users of flood hazard zone maps should assess the technical aspects of these classification thresholds and should modify them insofar as necessary (Meon 2006).
Table 10. Examples of intensity classifications. Source: (Meon 2006) Classifications have also been devised for probability, in addition to intensity. According to published results, the following classification thresholds appear to be most useful (Meon 2006):
• Flood events with a return period of 20 years: high probability.
• Flood events with a return period of 20 to 100 years: medium probability.
• Flood events with a return period of 100 years: low probability.
• Flood events with a return period of 200 years: residual risk.
Nicaragua MET-ALARN project
Figure 30, Figure 31 and Figure 32 show the criteria for intensity, frequency and hazard levels used in the MET-ALARN project in Nicaragua.
58 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 30. Criteria for intensity levels in the MET-ALARN project. Source: (MET-ALARN 2005)
Figure 31. Criteria for frequency levels in the MET-ALARN project. Source: (MET-ALARN 2005)
Figure 32. Hazard levels in the MET-ALARN project. Source: (MET-ALARN 2005)
59 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Ireland
In Ireland (OPW 2009) there are three types or levels of flood zones defined:
Flood highest (greater than 1% or 1 in 100 for river flooding or 0.5% or 1 in 200 for Zone A – where the probability of flooding from rivers and the sea is coastal flooding); Flood moderate (between 0.1% or 1 in 1000 and 1% or 1 in 100 for river flooding and Zone B – where the probability of flooding from rivers and the sea is between 0.1% or 1 in 1000 year and 0.5% or 1 in 200 for coastal flooding); and Flood C – where the probability of flooding from rivers and the sea is low (less than 0.1% or 1 in 1000 for both river and coastal flooding). Flood Zone Zone C covers all areas of the plan which are not in zones A or B (OPW 2009).
Figure 33. Flood levels in Ireland. Source: (OPW 2009)
Saxon (Germany)
The general Saxon flood maps vary from HQ(5/10/20/25), HQ(50), HQ(100) to HQ(200/300). Flood hazard maps are produced for events with return periods of 20 (or 25), 50, 100 and 200 (or 300) years and for so- called “extreme” events (sometimes corresponding to a HQ(200/300) event). The general hazard maps displays three intensity levels (high, medium, low), depending on the water depth. In mountainous areas, these levels depend on the discharge/flow velocity. For non-public use there are five intensity levels. The hazard indication map displays water levels/depth only for the HQ(extreme) and in four levels: 0-0.5 , 0.5- 2.0, 2.0-4.0 and >4.0 m (Meyer et al. 2004).
Bavaria (Germany)
The Bavarian flood hazard maps contain information about the spatial extent of the flood, the water depth and the flow velocity (where appropriate). Following flood scenarios are covered: o Low probability (HQextreme) o Medium probability (HQ100) o High probability (HQ5, HQ10) The water levels are displayed in 5 levels: 0,00 to 0,50, 0,50 to 1,00, 1,00 to 2,00, 2,00 to 4,00 and > 4,00m. The water depths are based on a 100 year event (Meyer et al. 2004).
England and Wales
60 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
The outline flood maps (OFM) of England and Wales produced by the Environment Agency indicate return periods for following categories: o HQ:100 (rivers) o HQ:200(coastal) o HQ1000(extreme) for both rivers and coastal flood risk The maps of the NAtional Flood Risk Assessment (NaFRA) contain three categories significant, where flood risk of greater than a return period of HQ(75), medium risk which is between HQ(75) and HQ(200) and low risk, where the risk is less than HQ(200). No water level information or intensity criteria are presented on the OFM or NaFRA layer (Meyer et al. 2004).
Table 11. Mapped zones on the outline indicative flood map. Source: (Meyer et al. 2004) Defense information is not taken into account when calculating the flood risk, however flood defenses are presented on the map and the areas which are protected by the defenses are also illustrated by black hatching. Damage date has also been linked to the flood map data most commonly used for CBA purposes. However this information is not currently available to public users (Meyer et al. 2004).
Austria
The Austrian flood maps vary from HQ(1/5/10/30/50 and 100). As the maps are separated for water bodies, torrents and water ways, the maps concerning torrents include HQ(150). Water levels are only given for torrents. 0.7 m is the distinction between red and yellow hazard zones (Meyer et al. 2004).
France
The French maps contain information about the spatial extent and the water depth but in a separate way. The maps of PPRI (Flood Risk Prevention Plan or Plan de Prevention des Risques d’Inondation) show 100-year events and higher. On the AZI-maps (Atlas of flood areas or Atlas des zones inondables) no water levels but levels of danger shown. It is a combination between speed (low, medium, high velocity) and height of the flood (<0.5m; 0.5m to 1m; 1m to 2m; >2m) (Meyer et al. 2004)
Cape town (South Africa)
Figure 34 provides an example based on the flood zoning policy in Cape Town, South Africa (Sayers et al. 2013).
61 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 34. This policy is based on the approach adopted in the city of Cape Town in South Africa. The key elements of the policy are shown below. Source: (Sayers et al. 2013)
7 Maps
Flood maps are the principal communication tool used to convey various aspects of flood behavior, including depth, level, velocity, flow rate and hazard. The maps generally display a representation of peak flood values (e.g. maximum flood depth), presented as an ‘envelope’ of results, such that although the peak values do not occur everywhere at the same time, the values presented are based on taking the maximum which occurred at each computational point in the model during the entire flood. Hence, a presentation of peak levels does not represent an instantaneous point in time, but rather an envelope of the maximum values, which occurred over the duration of the flood event. In addition to peak values, flood maps can also be created from any time in the flood simulation (e.g. three hours from start of flood), or any time, which correlates to a critical value (such as the time a levee overtops in a particular design event). Flood maps are generally modified in their presentation depending on the intended end-users (BMT WBM Pty Ltd 2017).
Maps can be classified under three broad headings:
• Emergency management: Information on inundation risk areas and evacuation sites/routes, which are basic elements of flood hazard maps, are indicated on the map (see Figure 35). It is because flood hazard maps are assumed to be used at the time of evacuation (River Bureau and Ministry of Land 2005).
62 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 35. Construction of emergency management flood maps. Source: (Shigenobu 2008) • Land use planning: this includes the themes of land use planning, building controls, structural works, infrastructure, landscape and environment, insurance and coastal management. • Community education and awareness: this includes the theme of education and information.
Good practices in the preparation of maps are:
• Public maps should be simple and self-explanatory and include a legend, such that as little supporting or explanatory information as possible is required for correct interpretation (Prinos 2008). • Flood maps should include at least the following information: (a) Title (what kind of information on which location is presented), (b) Location of the map as part of the catchment or country, (c) North and scale (preferably scale pole as this allows for changes in page sizes), (d) Responsible authority or institute with address, website and or telephone number, (e) Date of publication and (f) Disclaimer, including remarks on the quality of information (Prinos 2008). • Flood maps have limitations that should be clearly noted on each map (Profesional Engineers and Geoscientists of BC 2017). • Flood maps should be reviewed about every 10 years and updated if any of the following have occurred (Profesional Engineers and Geoscientists of BC 2017): o There is a change in the design flood because of changes to the criteria, change in climate, or a significant hydrologic change in the upstream watershed. o There have been significant changes in the channel geometry as a result of a flood or other event. o Significant local subsidence has occurred that changes the land elevation in relation to SLR. o New flood hazards are identified o Significant diking works are constructed in the floodplain, particularly if the diking alignments are new. o There are changes to the official community plan within a floodplain that would nullify the assumptions made in the hydraulic modelling (e.g., a development blocking a preferential overland flow route that was included in the model). o There are significant changes in the floodplain, such as community growth and urbanization. • Visually distinguishable and standard choice of colors. See (FEMA 2018c) for an example. • Appropriate meta-data should be provided where maps are issued or downloadable in GIS format (e.g. shape, ascii). Such data should include standard meta-data (dates, responsible organisation, etc.) as well as information necessary for use of the GIS data, including the
63 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
map projection and any datum levels used. Consideration should also be given to any relevant meta- data protocols or requirements (World Meteorological Organization 2013). • Nominally, towns or locations of higher population and infrastructure density should have a larger scale to allow for easy reference of flood impacts. Different scales may also be appropriate for land use planning and emergency response maps. Readability is clearly of paramount importance for hard copy maps (Victoria State Government 2016). • Updating flood maps can involve a substantial update of the existing modelling and analysis, so the costs need to be considered. To minimise the cost of such updates, it is recommended that agencies capture and retain electronic copies of the input data sets, models, maps and calculation records that the flood study practitioner used to create the maps. This will ensure that future practitioners can more easily update these data sets (Victoria State Government 2016).
According to the information that is presented, the following types of maps can be distinguished:
• Flood extent maps: The flood extent maps are the most used flood maps. They display the areas, which are inundated for specific scenario. Historical flood maps, which are based on real events, occurred in the past, can also be produced, not only for hypothetical events. Every single scenario is showed on particular map. There are several main uses of these specific maps. They serve as a base for production of risk maps and danger maps. Also they are used in flood risk management and spatial planning (Mavrova-guirginova et al. 2010). Figure 36 shows an example of this type of maps.
Figure 36. Example of a flood extension map from Baden-Wü rttemberg (Neckar) for return period of 10, 50 and 100 years. Source: (Mavrova-guirginova et al. 2010) • Flood depth maps: The flood depth maps show the water depth in the flooded area. They are based on the flood extent maps – after the determination of the water extent, in this boundaries is calculated the water depth. They are produced for the same return periods as the flood extent maps. The depth is given in centimeters or meters, as it is appropriate. In the different countries are used different colours – from shades of blue to red, yellow and green. These maps are very useful for risk management an urban planning (Mavrova- guirginova et al. 2010). An example is shown in Figure 37.
64 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 37. Example of flood depth map (Neckar river). Source: (Mavrova-guirginova et al. 2010)
8 Visualization and dissemination of hazard, vulnerability and risk information
Properly presented, flood maps allow residents to see exactly where and how they may be impacted by floodwaters. In addition, maps help to understand the process of flooding and are needed to provide information to guide people’s behavior in case of the event. People, residing in an area where flooding is a frequent event, have devised (out of experience), many ways of coping with them (World Meteorological Organization 2013). The communication of flood hazard and risk is a complex cross-disciplinary field where different audiences need to be reached. In most cases flood maps are used to communicate the relevant information to these audiences. Generally, means of communication are tools or devices to transit information to a target audience conceived as a single unit or a number of individuals (World Meteorological Organization 2013).
The dissemination of maps and the information to be provided depends on the following two groups of users (a) Public (including general public that ought to be made aware of the risk) and (b) Professional users (public sector authorities, involved in planning, execution and maintenance of measures, and/or decision making in policy and/or crisis management, private companies)(Prinos 2008). The primary purpose of public dissemination of flood maps is to increase public awareness. This is to empower individuals for taking appropriate preparatory and response measures, and informing them about decisions such as the purchase or otherwise of a property, or the assignment of use, layout and design of an area of land (Prinos 2008).
The public dissemination via the Internet is essential (single-point update, low-cost dissemination, reducing risks of use of superseded (aged) data, etc.), although hard copies of maps should also be available in public offices (e. g. in libraries, municipalities) to make the information available to those who do not have Internet access (Prinos 2008).
65 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
An example of advanced flood hazard mapping dissemination through web information systems is the approach of Fema. This agency has developed a program to produce digital flood hazard maps where the population can access very detailed flood hazard maps for the area of interest via web (Prinos 2008). Other examples are shown in Figure 38 and Figure 39 that correspond to public flood information systems in the USA and Australia.
Figure 38. Hudson River Flooding Decision Support System. Source: http://www.ciesin.columbia.edu/hudson-river-flood-map/
66 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 39. Australia flood information system. Source: http://dnrm- floodcheck.esriaustraliaonline.com.au/floodcheck/
9 Community consultation
The local community both flood prone and otherwise has a key role to play in the development, implementation and success of a flood study. At the scoping phase or commencement of the study to assist with data collection, particularly information about past floods such as flood marks, photos or video, newspaper clippings and recollections about flood behavior (timing, depth, extent). At the calibration phase, to provide feedback on the preliminary calibration results and provide any additional information, which may assist the calibration process. Peer review of the hydraulic model prior to consultation is strongly encouraged to promote community confidence in the adopted modeling approach and preliminary results. Separate to the development of a flood study, the community should be engaged on an ongoing basis to build community flood resilience and ensure the community is prepared to respond and recover to future flood events. It is important to note that careful consideration should be given to the communication of difficult modelling issues, such as uncertainty, limitations and assumptions (Victoria State Government 2016).
A good practice is to prepare a detailed community engagement plan as part of the study in order to complement historic data, include local knowledge and diffuse advances and results. Details about participatory processes can be consulted in (Meyer et al. 2004).
10 Debris flows risk assessment guidelines
In mountain torrents the hazard is controlled by (EXCIMAP 2007):
• Water flood (discharge, velocity) • Debris flow (discharge, velocity) • Debris accumulation (height) • Erosion (depth)
According to the interpretation of aerial photographs taken soon after the hurricane Fifi and field investigation by (JICA 1994) approximately 10% of the mountain slopes in the Choloma river basin suffered slope failures. Figure 40 shows the identification of debris flow prone areas in the area of Puerto Cortés, showing that most of the creeks in the Merendon mountains can be subjected to debris flows. This preliminary analysis indicates the need for a more profound assessment that allows verification of debris flow prone areas and a prioritization for risk analysis.
67 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 40. Identification of debris flow prone areas in the Sula Valley, area Puerto Cortés. Source: (JICA 1994) Modeling of torrent activity is only partly possible it needs special knowledge and there are big differences between modeling of torrent flows and river floods. As one example of practices the Austrian approach to torrent assessment and mapping is described as follows (EXCIMAP 2007):
• Hazard zone mapping is based on a design event with a return period of approx. 150 years. Hazard zones represent the sum-line of all possible hazard scenarios within the framework of this design event. • Hazard zone mapping in Austria is normally based on an intensive survey in the field supported by the results of remote sensing methods and geo-data (e.g. geological maps, hydrological data, soil maps, vegetation maps, terrain models). Also of major importance is information on disaster events in the past, derived from chronicles, testimonials in the field (so called “silent witnesses”), disaster documentation (torrent cadastre) as well as observations by local residents. • Recently some computer based models describing hydrological processes (floods, debris transport, debris flows) have been developed; however, the application in torrent catchments is still very limited. • The outline of the hazard zones is an “expert decision” based on all available data and models and is normally done in the field. The fixed hazard zones are then digitally processed and integrated into GIS. Hazard zones are shown on plans based on the land register/cadastre in scale 1:2,000.
Debris flows can arise from water flows through entrainment of channel debris, but can also be generated by landslide, dam or glacial lake outbreak floods, other dam failures, hillslope and channel erosion and similar processes. Debris floods are highly erosive but can cause aggradation where channel slopes decrease, leading to avulsions and erosion (Profesional Engineers and Geoscientists of BC 2017). The most pragmatic approach to hazard assessment is through detailed
68 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
fieldwork to identify likely avulsion sites and routes down the fan. Erosion can be accounted for through allocation of setbacks for varying degrees of hazard. Hazard zones for flood mapping can be determined through identification of a combination of potential inundation areas and those subject to erosion (Profesional Engineers and Geoscientists of BC 2017). It is recommended that a suitably experienced geoscientist be involved with assessment of the types of hazard that could occur. If this assessment suggests that geomorphic events (debris floods and debris flows) are predominant with the return periods of concern, the hazard and risk assessments should be completed by the geoscientist with input from a hydrotechnical engineer (Profesional Engineers and Geoscientists of BC 2017).
10.1 Debris flow hazard modeling
In general, the mapping of debris and mud flow hazards in steep mountainous channels follows the same procedures as for the floods; however, the assessment is less accurate than that of flood hazards. Uncertainties with regard to probability of occurrence are higher, the flow characteristics are less determined and flow behavior less predictable. Nevertheless, a systematic procedure provides reliable results (World Meteorological Organization 2013). Figure 42 shows the procedure.
A variety of models exist for simulating mass-flows and for characterizing the hazard. Defining the spatial extent of the endangered area is very important for Quantitative Risk Assessment (QRA). It requires accurate forecast of the run-out behavior (e.g. how far and how fast will the mass travel?) expressed through quantitative spatially distributed parameters that include the run-out distance, the run-out width in the spreading zone, the displacement rate (e.g. velocity), the thickness of the moving mass and the impact pressure (Hungret al.2005). However, run-out modeling is rather complicated because of the various physical processes that happen during an event. These processes depend on the characteristics of the source area, on the type of triggering process, on the characteristics of the path and on the type and volume of material incorporated or deposited during the travel. Some of these complex processes are erosion, entrainment and deposition, changes in the rheology during one event, layering and formation of pulses and damming and breaching inside the channels. The methods and tools used in run-out analyses are very different in scale of applications: empirical or statistical techniques are generally applied to predict susceptibility at the regional scale with the purpose of hazard mapping, while process-based approaches are applied at the local scale with the purpose of designing engineered mitigation works and early-warning systems. A common problem with the empirical methods is that the scatter of the data is too large for anything but very preliminary predictions of the travel distance. The flexibility of the empirical methods allows them to be applied in local to medium-scale landslide susceptibility and hazard maps but as they do not provide kinematic parameters (velocity, kinetic energy) of the landslides these approaches can be hardly applied to site-specific analyses and in QRA (Luna et al. 2014). Table 12 summarizes current methods for debris flow hazard assesment.
Empirical Run-Out Models Based on extensive amounts of field observations and on the analysis of the relationships between the run-out distance and different landslide mechanisms, their morphometric parameters, the volume of the landslide mass, and the characteristics of the terrain. Empirical approaches are based on simplified assumptions, and although they lead to generalized results they are relatively easy to apply over larger areas.
Heuristic methods Involve the identification and mapping of landslide deposits that provides a direct measurement of the distance travelled in the past. The extent of both ancient and recent landslide deposits is the basis for defining future travel distances. Field work and photo interpretation are classical procedures used to define the spatial distribution
69 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
and extent of past landslides. The margin of the landslide deposits give an indication of the maximum reach that a landslide is able to reach in the present landscape (Hungr et al.2005).
The mass-change method Is based on the phenomenon that, as the landslide debris moves down slope, the initial volume/mass of the landslide is being modified through loss or deposition of materials, and that the landslide debris halts when the volume of the actively moving debris becomes negligible (Cannonand Savage1988). The average mass/volume-change rate of landslide debris was established by dividing the volume of mobilized material from the landslide by the length of the debris trail.
The angle of reach method The angle of reach method is based on the angle of the line connecting the crest of the landslide source to the distal margin of the displaced mass also called the angle of reach (Heim1932). This angle is also used as an index of efficiency for the dissipation of energy. Once the release source, volume and direction of the flow are known, these methods can estimate the length of the run-out.
Physically-Based Dynamic Run-Out The analytical models are based on hydro-mechanics and solve the equations of Models mass and momentum in a close-form or numerically way.
Models Based on the Solution One dimensional models analyze the movement considering the topography as a Dimension (1D or 2D) cross-section of a single pre-defined width while two dimensional models makes the analysis considering the topography in plan and cross section.
Models Based on the Solution The equations of motion can be formulated in two different frames of reference: Reference Frame (Eulerianor Eulerian or Lagrangian. Lagrangian)
Models Based on the Basal Rheology The rheology of the flow is expressed as the resistance force (Sf) that interact inside the flow and at the interface between the flow and the bed path. The most common rheologies used in the dynamic models are: the “Frictional” (or “Coulomb”) resistance (Hungr and McDougall2009), the frictional-turbulent “Voellmy” resistance (Voellmy1955), the visco-plastic “Bingham” (or “Herschel-Bulkey”) resistance(Coussot1997; Malet et al.2004), the “Quadratic” resistance (O’Brien et al.1993)(Table5.1) A thorough description of rheologies can be found in Naef et al. (2006).
Table 12. Methods for debris flow hazard assessment. Source:(Luna et al. 2014) Some commonly used run-out models are: MADFLOW, TOCHNOG, RAMMS, DAN3, FLATMODEL, SCIDDICA, 3dDMM, PASTOR, MassMov2D, RASH3D, FLO-2D, dTITAN2D, PFC and VolcFlow (Luna et al. 2014).
Regional-scale analysis, which includes scales in the range from 1:10.000 to 1:50.000, can provide an initial overview of the hazard in a specific area. The goal of a medium scale analysis is to identify all the potentially unstable areas as accurate as possible and the down-slope regions probably affected by the flow (Luna et al. 2014).
This simple approach is used in the recently developed Flow-R model that allows rockfalls and debris flow analyses at regional scales (Horton et al. 2008). The method consists in (1) the identification of the source areas, and (2) the run- out modeling. The source identification is based on the topographical assumption that from a concave slope located below a sufficiently large catchment area and exhibiting a certain slope angle, debris flows will potentially initiate (Horton et al. 2008). Although information on the spatial and temporal probability is available, the result is not, strictly speaking, a full hazard assessment but only a susceptibility assessment since no indication is given on the hazard intensity (Luna et al. 2014).
70 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 41. Qualitative debris flow susceptibility scenarios for the Barcelonnette basin (South East France). (a) Debris flow susceptibility assessment for the mid-course of the Ubaye River catchment; (b) Excerpt of the debris flow susceptibility assessment for the Riou-Bourdoux and La Valette sub-catchments; (c) Excerpt of the debris flow susceptibility assessment for the Faucon, Bourget and Sanieres sub-catchments. Source: (Luna et al. 2014)
71 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
• Based on avaliability of erodible sediment sources, Determination of availability of water and steepness of the channel. mud-flow prone watersheds
• Combination of historical an geomorphological approaches Definition of design events
• For preliminary assessments geomorphological approaches can be used. Determination of • For detailed assessment geomophologic, historical and potetially affected modelling can be used. Models: e.g. FLO-2D or RAMMS. mud flow areas
Figure 42. Assessment and mapping of debris and mud flow hazards. Adapted from: (World Meteorological Organization 2013) 10.2 Vulnerability curves Fuchs et al. (2007) obtained an empirical vulnerability function analyzing data from a well- documented debris flow event in 1997 in the Austrian Alps, linking process intensities to object vulnerability values. The elements at risk were represented by brick masonry and concrete buildings located on the fan of the torrent. Vulnerability was calculated in terms of damage ratio (that describes the amount of loss related to the overall potential damage of the structure) and the debris flow intensity Figure 43. This vulnerability function can be used as a proxy for structural resistance of buildings with respect to dynamic debris flow impacts. Akbas et al. (2009) developed an empirical vulnerability function based on observations of a debris flow that occurred on July 13, 2008, in the village of Selvetta, in the Valtellina Valley (Northern Italy). In this study Figure 44, vulnerability was calculated using an economic approach and defined as the ratio between the loss and the individual reconstruction value for each of the 13 buildings that were affected by the debris flow event. Damage-related data were obtained from official documents and an approximate reconstruction value for each building was extracted from the Housing Price Book, prepared by the Engineers and Architects of Lombardy Region (DEI 2006), according to the building type and size. The obtained ratios were coupled with the corresponding deposition height to compare the results
72 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
and to perform a critical assessment of vulnerability functions developed for debris flows by different authors. Differences in the estimated vulnerabilities suggest that there is a need for further studies with additional data to construct empirical curves that can be used with a higher level of confidence. Barbolini et al. (2004) derived vulnerability curves relating damage state with the avalanche dynamic parameters, such as velocity and flow depth. The elements at risk were alpine buildings, as well as the people inside them and people directly exposed to avalanches. The vulnerability of buildings was defined as the ratio between the cost of repair and the building value (referred as specific loss, SL) (Sterlacchini et al. 2014).
Figure 43. Relationship between debris flow intensity and vulnerability expressed by a
second order polynomial function for an event intensity (e.g. debris height) <2.5 m (Fuchs et al. 2007). Source: (Sterlacchini et al. 2014)
Figure 44. Example of fragility curve and function (dashed line) between debris flow height and vulnerability (Akbas et al. 2009). Source: (Sterlacchini et al. 2014)
Quan Luna et al. (2011, submitted) derived synthetic physical vulnerability curves Figure 45, that related the outputs of numerical dynamic run-out models (flow depth and impact pressures) with economic values of physical damage to the elements at risk (buildings). The elements at risk were
73 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
represented by single to three-storey brick masonry and concrete buildings (Quan Luna et al. 2011) (Sterlacchini et al. 2014).
Figure 45. Proposed synthetic fragility function for debris flow deposit heights (left) and impact pressures (right). Source: (Sterlacchini et al. 2014) 10.3 Risk analysis
Figure 46 shows the methodology used for debris flow risk in the area of Tresenda (Valtellina Valley, Italy), where three quantitative debris flow hazard scenarios for different return periods were prepared using available rainfall and geotechnical data. Figure 47 shows the results that were obtained (Luna et al. 2014).
74 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Figure 46. Methodological flowchart applied in the back-calculation of the Selvetta debris flow. Source: (Luna et al. 2014).
Figure 47. Results of the hazard modelling for the 100-year return period showing the calculated degree of damage to the buildings. On the left height of accumulation, on the right impact pressures. Source: (Luna et al. 2014).
75 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
11 Required team for flood risk assessment
Typical qualifications or a team of professionals may include education and experience in:
• hydrodynamic modelling • watershed hydrology • groundwater geology • extreme value statistics and trend analysis • air photograph and satellite imagery interpretation • bathymetric and land based surveying • hydrological studies, including flood frequency analysis • climate change and its effects on hydrological processes • geomatics • knowledge of fluvial and coastal geomorphology principles and applications
12 Guidelines on reporting
Good practice on flood risk analysis reporting includes:
• The technical reports should be prepared in such a manner that the entire work can be recreated by a qualified person without the need to refer to any other material. • Qualified persons are to be able to recognize and understand all the methods, approaches, basic data, assumptions and rationale used for these methods. • The modeling report should contain justification of the modeling method, modelling software used, limitations of the modeling method and of the software used as well as any assumption made for model simplification. It should also include information on uncertainties as well as any other issues which may affect the accuracy of the output (SEPA 2014). • The report should include plans and description of any structures (culverts, screens, embankments or walls, overgrown or collapsing channels etc) which may influence local hydraulics, and a summary of the findings of any hydraulic modelling including how structures impact on water levels at the site. Where culverts provide a significant flow restriction, levels and discharge rates at which flow would overtop the structure should be identified. Likely impacts of blocked culverts also need to be identified (SEPA 2014). • Discussion on the data used in calibration work including the reasons for the choice of data used in the work shall be included in the Hydraulic Report (Nottawasaga Valley Conservation Authority 2013). • In case a comprehensive calibration and validation of the hydrologic and hydraulic model is unlikely to be possible, due to a lack of calibration data for the catchment, it should be outlined how selection of model parameters in the hydraulic model will achieve reasonable accuracy in model results (BMT WBM Pty Ltd 2017). • Calibration should be fully documented, including dates, measurements, and locations of measurements of historic floods; parameters revised and rationale for revising; and the
76 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
calibration model input and output data (FEMA 2009). • Documentation of on-site observations must include photographs as well as the computations used to estimate roughness coefficients (FEMA 2009). If “n” values were adjusted based on calibration, the documentation must include a summary of the values before and after the adjustments (FEMA 2009). • Georeferenced spatial files must be submitted, which include but are not limited to: o Sub-basins. o Locations of estimated flood discharges. o Flood control structures, such as reservoirs and diversions within the reach system that affect flood flow. • The choice of temporal storm distribution must be fully documented. If the source of the distribution is not a rederal, state, or regional agency, the documentation must include a detailed description of the derivation of the distribution, including sources of data and the means of fitting those data to a particular distribution (FEMA 2018a). • The choice of areal reduction factors (or lack thereof) should be documented. (FEMA 2018a)
The following summarizes what is expected from any hydraulic modelling study (SEPA 2014):
• Statement of objective - to explain clearly the situation being modeled and the objectives of the modeling study, including details of the output required from the model. To demonstrate the modeling approach is fit for purpose. • Justification of the model - to demonstrate that the model used is suitable for this study. This should include evidence of previous applications in similar circumstances and a demonstration of experience in the application of the model. • Data collection - all relevant data collection and measurement techniques should be quoted, including expected errors and relevant quality assurance. It is expected that appropriate input data is collected to support the objectives of the study. Surveyed cross-sections of the main channel and floodplain should be comprehensive, i.e. not include ‘glass walls’ in the model. • Model calibration / boundaries - The model calibration coefficients and the procedures used to optimize the calibration must be stated clearly. The choice of upstream and downstream model boundaries must be justified. • Model validation – efforts should be made to validate the model against historical flood events/ high flow events, or the use of gauge data where available. • Sensitivity analysis - this analysis must be presented to demonstrate the effect on the key output parameters resulting from variation of input data and controlling assumptions. This is particularly important where limited data is available to validate the model or where there are some uncertainties involved. Parameters that should most commonly be tested for sensitivity include design flow, roughness and upstream/downstream boundary conditions. Culverts and small hydraulic structures that may be prone to blockage during floods deserve particular analysis (i.e. the model should be run with full and/or partial blockage to better understand the impact of such processes). • Quality assurance - to demonstrate that the model has been subject to an evaluation procedure establishing its suitability for the relevant tasks and that numerical and method checks have been carried out by an appropriately qualified person. ? Auditable - to ensure that there is a clear account of the modeling exercise for inspection by any appropriate auditors. ? Reporting - clear description of the model including the underlying principles and implicit or explicit assumptions. Also, a clear summary of the numerical output should be supplied (preferably in tabular format) in addition to likely errors, bias, sensitivity,
77 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
implications for the objectives of the study and conclusions. Diagrams showing the geographical extent of the model, plan view location of cross- sections, the resultant flood profile in long-section and also at each cross-section, should be provided. • Auditable - to ensure that there is a clear account of the modeling exercise for inspection by any appropriate auditors. • Reporting - clear description of the model including the underlying principles and implicit or explicit assumptions. Also, a clear summary of the numerical output should be supplied (preferably in tabular format) in addition to likely errors, bias, sensitivity, implications for the objectives of the study and conclusions. Diagrams showing the geographical extent of the model, plan view location of cross- sections, the resultant flood profile in long-section and also at each cross-section, should be provided.
References
Analysis, R. and Coauthors, 2000: Risk analysis and uncertainty in flood damage reduction studies. http://books.google.com/books?hl=en&lr=&id=KaOpl_cPpoUC&oi=fnd&pg=PA1&dq=Risk+Ana lysis+and+Uncertainty+in+Flood+Damage+Reduction+Studies&ots=W5el4sIhRQ&sig=kfJE1Xi- K0Ul_rQmt17rkXZ9QJc (Accessed October 25, 2012). Asselman, N., and Team, 2009: Flood Inundation Modelling Executive Summary. 24. www.floodsite.net. Betâmio De Almeida, A., 2001: Flood risk management in valleys with dams. Prog. Water Resour., 3–13. http://www.scopus.com/inward/record.url?eid=2-s2.0- 2942657724&partnerID=tZOtx3y1. Birkmann, J., 2006: Measuring vulnerability to natural hazards: Towards disaster resilient societies. U.N.U. Press, Ed. bepress,. Birkmann, J. and Coauthors, 2013: Framing vulnerability, risk and societal responses: the MOVE framework. Nat. Hazards, 67, 193–211, doi:10.1007/s11069-013-0558-5. http://link.springer.com/10.1007/s11069-013-0558-5. BMT WBM Pty Ltd, 2017: Guide for Flood Studies and Mapping in Queensland. State of Queensland, https://www.dnrm.qld.gov.au/__data/assets/pdf_file/0010/332695/guide-flood-studies- mapping-qld.pdf. Bnamericas, 2018: Presa de Usos Múltiples El Tablón. https://www.bnamericas.com/project- profile/es/represa-el-tablon-nuevo-represa-el-tablon (Accessed April 27, 2018). Caballero, Menjivar, Maradiaga, and Chaín, 2008: Estudio de Diagnóstico y Propuesta sobre Riesgo Urbano Corredor Industrial: Villanueva – San Pedro Sula – La Lima – Choloma – Puerto Cortés Zona Metropolitana del Valle de Sula, Honduras. Tegucigalpa (Honduras),. CIAT, 2001: Atlas de Honduras. http://gisweb.ciat.cgiar.org/Mitch/. Council, N. R., 2015: Tying Flood Insurance to Flood Risk for Low-Lying Structures in the Floodplain. The National Academies Press, Washington, DC, https://www.nap.edu/catalog/21720/tying-flood-insurance-to-flood-risk-for-low-lying- structures-in-the-floodplain. Czajkowski, J., L. K. Cunha, E. Michel-Kerjan, and J. A. Smith, 2016: Toward economic flood loss characterization via hazard simulation. Environ. Res. Lett., 11, doi:10.1088/1748- 9326/11/8/084006. http://iopscience.iop.org/article/10.1088/1748-9326/11/8/084006/pdf.
78 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Díez-Herrero, A., L. Huerta, and M. Isidro, 2009: A handbook on flood hazard mapping methodologies. http://books.google.com/books?hl=en&lr=&id=4nXerdbeEogC&oi=fnd&pg=PA3&dq=A+Handb ook+on+Flood+Hazard+Mapping+Methodologies&ots=pycv3HeQ2p&sig=5U1h9jveo_nYn9- eDxAlV9ntdjk (Accessed April 20, 2013). ESRI, 2018: Obtaining elevation information for building footprints. http://desktop.arcgis.com/en/arcmap/latest/extensions/3d-analyst/3d-buildings- obtaining-elevation-information-for-building-footprints.htm (Accessed May 11, 2018). EXCIMAP, 2007: Handbook on good practices for flood mapping in Europe European exchange circle on flood mapping Handbook on good practices for flood mapping in Europe. FEMA, 2003: Guidelines and specifications for flood hazard mapping partners Appendix A: Guidance for aerial mapping and surveyin. https://www.fema.gov/media-library- data/1387814416677- caa613eeca53246cb7a7dcbf342a7197/Guidelines+and+Specifications+for+Flood+Hazard+Map ping+Partners+Appendix+A-Guidance+for+Aerial+Mapping+and+Surveying+(Apr+2003).pdf. FEMA, 2009: Guidelines and Specifications for Flood Hazard and Mapping Partners Appendix C: Guidance for Riverine Flooding Analyses and Mapping. https://www.fema.gov/media-library- data/1387814760538- 48140a90bd0462c31a686a957ad7ad7c/Guidelines_and_Specifications_for_Flood_Hazard_Map ping_Partners_Appendix_C- Guidance_for_Riverine_Flooding_Analyses_and_Mapping_(Nov_2009).pdf. FEMA, 2016: Guidance for Flood Risk Analysis and Mapping Elevation Guidance. https://www.fema.gov/media-library-data/1469794589266- f404b39e73fa7a1c5ffe4447636634d4/Elevation_Guidance_May_2016.pdf. FEMA, 2018a: Guidance for Flood Risk Analysis and Mapping Hydrology: Rainfall-Runoff Analysis. https://www.fema.gov/media-library-data/1520964160255- 7c49e1753d0b2634e0c5fb4999459374/Hydrologic_Rainfall_Runoff_Analysis_Feb_2018.pdf. FEMA, 2018b: Guidance for Flood Risk Analysis and Mapping Flood Risk Assessments. https://www.fema.gov/media-library-data/1521055701532- 414dbf3beb1b381426301863a3b93912/Flood_Risk_Assessment_Guidance_Feb_2018.pdf. FEMA, 2018c: Guidance for Flood Risk Analysis and Mapping Flood Risk Map. https://www.fema.gov/media-library-data/1520966727438- 8c05ecc18e141449aa79665641bd9fe3/Flood_Risk_Map_Guidance_Feb_2018.pdf. Fhoilsiú, A., and A. G. Oifig, 2009: Flood Risk Management. Fuchs, S., C. Kuhlicke, and V. Meyer, 2011: Editorial for the special issue: Vulnerability to natural hazards-the challenge of integration. Nat. Hazards, 58, 609–619, doi:10.1007/s11069-011- 9825-5. http://download.springer.com/static/pdf/856/art%3A10.1007%2Fs11069-011-9825- 5.pdf?originUrl=http://link.springer.com/article/10.1007/s11069-011-9825- 5&token2=exp=1448575665~acl=/static/pdf/856/art%253A10.1007%252Fs11069-011-9825- 5.pdf?originUrl=http%3A%2F. Hufschmidt, G., 2011: A comparative analysis of several vulnerability concepts. Nat. Hazards, 58, 621–643, doi:10.1007/s11069-011-9823-7. http://link.springer.com/10.1007/s11069-011- 9823-7. INDECI, 2011: Manual de estimación del riesgo ante inundaciones fluviales - Cuaderno técnico No 2. Lima, Perú, 81 pp. http://bvpad.indeci.gob.pe/doc/pdf/esp/doc1743/doc1743-
79 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
contenido.pdf. JICA, 1994: El estudio de plan maestro sobre el control de erosion y sedimentacion en la cuenca piloto, Choloma, San Pedro Sula, Cortes en la Republica de Honduras. http://open_jicareport.jica.go.jp/617/617_613.html. Jonkman, S. N., M. Bočkarjova, M. Kok, and P. Bernardini, 2008: Integrated hydrodynamic and economic modelling of flood damage in the Netherlands. Ecol. Econ., 66, 77–90, doi:10.1016/j.ecolecon.2007.12.022. http://linkinghub.elsevier.com/retrieve/pii/S0921800907006155 (Accessed March 2, 2013). Kappes, M. S., M. Papathoma-Köhle, and M. Keiler, 2012: Assessing physical vulnerability for multi- hazards using an indicator-based methodology. Appl. Geogr., 32, 577–590, doi:10.1016/j.apgeog.2011.07.002. http://linkinghub.elsevier.com/retrieve/pii/S0143622811001378. Luna, B., J. Blahut, M. Kappes, S. O. Akbas, J. P. Malet, A. Remaître, and M. Jaboyedoff, 2014: Methods for Debris Flow Hazard and Risk Assessment. Mountain Risks: From Prediction to Management and Governance, T. Van Asch, J. Corominas, S. Greiving, J.-P. Malet, and S. Sterlacchini, Eds., Springer Netherlands, Dordrecht, The Netherlands, 133–177 http://dx.doi.org/10.1007/978-94-007-6769-0_5. Lundgren, L., and A. Jonsson, 2012: Assessment of social vulnerability: a literature review of vulnerability related to climate change and natural hazards. CSPR Brief., 9, 20. http://www.cspr.se/briefings/1.358869/CSPRBriefing9.pdf. Mastin, M. C., 2001: Flood-hazard mapping in Honduras in response to Hurricane Mitch Water- Resources Investigations Report 2001-4277. https://pubs.er.usgs.gov/publication/wri014277. Mavrova-guirginova, M., D. Pencheva, and M. Belyashka, 2010: MAPS IN THE CONTEXT OF THE EUROPEAN FLOOD RISK DIRECTIVE. Meon, G., 2006: Flood hazard map guidelines of the German Working Group of the Federal States on Water Issues (Bund / Länder-Arbeitsgemeinschaft Wasser (LAWA)). http://www.lawa.de/documents/Flood-Hazard-Map-Guidelines_ea8.pdf. Merz, B., A. H. Thieken, and M. Gocht, 2007: CHAPTER 13 FLOOD RISK MAPPING AT THE LOCAL SCALE : CONCEPTS AND CHALLENGES. 231–251. http://crs.itb.ac.id/media/jurnal/Refs/Critical_Review/Download/12Jun/fulltext-100.pdf. Messner, F. and Coauthors, 2007: Evaluating flood damages : guidance and recommendations on principles and methods principles and methods. MET-ALARN, 2005: Inundaciones fluviales. Managua, Nicaragua, 71 pp. Meyer, V. and Coauthors, 2004: CRUE Final Report. de Moel, H., J. van Alphen, and J. C. J. H. Aerts, 2009: Flood maps in Europe – methods, availability and use. Nat. Hazards Earth Syst. Sci., 9, 289–301, doi:10.5194/nhess-9-289-2009. http://www.nat-hazards-earth-syst-sci.net/9/289/2009/. NFIP, 2015: Tying Flood Insurance to Flood Risk for Low-Lying Structures in the Floodplain. C. on R.-B.M. for I.P. Of and N.E.S. in the N.F.I. Program, Eds. http://www.nap.edu/catalog/21720. Nottawasaga Valley Conservation Authority, 2013: NVCA Natural Hazards Technical Guide. Utopia, Ontario - Canada, https://www.nvca.on.ca/Shared Documents/NVCA Natural Hazards Technical Guide.pdf. OEA, 1991: Primer on Natural Hazard Management in Integrated Regional Development Planning.
80 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
Washington D.C., USA, http://pdf.usaid.gov/pdf_docs/Pnabj801.pdf. OPW, 2009: The Planning System and Flood Risk Management. 73 pp. http://www.opw.ie/media/Planning System and Flood Risk Management Guidelines.pdf. PGRD-COPECO, 2016: Informe Geomorfológico Municipio de Puerto Cortés Departamento de Cortés Actualización de Planificación Territorial con enfoque en Gestión de Riesgos para (7) siete Municipios: Choloma, La Lima, Omoa, Pimienta, Potrerillos, Puerto Cortés en el Departament. Tegucigalpa (Honduras), 27 pp. Prinos, P., 2008: Review of Flood Hazard Mapping. 70 pp. Profesional Engineers and Geoscientists of BC, 2017: FLOOD MAPPING IN BC APEGBC PROFESSIONAL PRACTICE GUIDELINES V1.0. https://www.egbc.ca/getmedia/8748e1cf-3a80-458d-8f73- 94d6460f310f/APEGBC-Guidelines-for-Flood-Mapping-in-BC.pdf.aspx. Reese, S., and D. Ramsay, 2010: RiskScape : Flood fragility methodology. ReliefWeb, 2012: Honduras pone en operacion primer radar meteorologico de centroamerica. https://reliefweb.int/report/honduras/honduras-pone-en-operacion-primer-radar- meteorologico-de-centroamerica (Accessed May 11, 2012). RIBA, 2009: Climate Change Toolkit 07 Designing for Flood Risk. River Bureau, and Ministry of Land, 2005: Flood Hazard Mapping Manual in Japan. http://www.icharm.pwri.go.jp/publication/pdf/2005/flood_hazard_mapping_manual.pdf. Sampson, C. C., A. M. Smith, P. D. Bates, J. C. Neal, L. Alfieri, and F. J. E., 2015: A high-resolution global flood hazard model. Water Resour. Res., 7358–7381, doi:10.1002/2015WR016954.Received. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4989447/pdf/WRCR-51-7358.pdf. Sayers, P., L. Yuanyuan, G. Galloway, E. Penning-Rowsell, W. K. Shen Fuxin, C. Y. And, and T. Le Quesne, 2013: Flood Risk Management A Strategic Approach. Schanze, J., E. Zeman, and J. Marsalek, 2006: Flood risk management: hazards, vulnerability and mitigation measures. NATO, Ed. Springer Netherlands, Dordrecht, The Netherlands, 319 pp. http://books.google.com/books?hl=en&lr=&id=g39AohmaOjgC&oi=fnd&pg=PR8&dq=Flood+Ri sk+Management:+Hazards,+Vulnerability+and+Mitigation+Measures&ots=ehhbKxaL3A&sig=hL q0wRYSDY9Nl7E3YCj6smPIr10 (Accessed March 12, 2013). Schultz, A. H., 1981: Basin Planning for the Sula Valley in Honduras. Water Int., 6, 75–82, doi:10.1080/02508068108685905. https://doi.org/10.1080/02508068108685905. SEPA, 2014: TECHNICAL FLOOD RISK GUIDANCE. 8, 1–30. SEPLAN, 2014: PLAN DE DESARROLLO REGIONAL CON ENFOQUE DE ORDENAMIENTO TERRITORIAL REGION 01 VALLE DE SULA, Resumen Ejecutivo. https://drive.google.com/file/d/0B1rHWW25PZmmekJmUjdjZmVmWjA/view. Shigenobu, T., 2008: LOCAL DISASTER MANAGEMENT International Centre for Water Hazard and Risk Management. State of Queensland, 2014: Flood mapping implementation kit. Queensland Australia, https://www.dnrm.qld.gov.au/__data/assets/pdf_file/0009/230778/flood-mapping-kit.pdf. Sterlacchini, S., S. O. Akbas, J. Blahut, O.-C. Mavrouli, C. Garcia, B. Q. Luna, and J. Corominas, 2014: Methods for the characterization of the vulnerability of elements at risk. Mountain Risks: From Prediction to Management and Governance, T. Van Asch, J. Corominas, S. Greiving, J.-P. Malet, and S. Sterlacchini, Eds., Vol. 34 of, Springer Netherlands, Dordrecht,
81 Guidelines for flood risk modelling in the Sula Valley – Honduras Draft Report
The Netherlands, 233–273 http://hdl.handle.net/2117/20764. Su, M., and J. Kang, 2005: A grid-based GIS approach to regional flood damage assessment. J. Mar. Sci. …, 13, 184–192. http://jmst.ntou.edu.tw/marine/13-3/184-192.pdf (Accessed September 12, 2012). The World Bank, 2011: ASSESSMENTS AN APPROACH FOR UNDERSTANDING. Thieme, E., and C. Jacobs, 2012: Risk Mapping Honduras Sector Disaster Risk Reduction & Emergency Aid. https://www.cordaid.org/en/wp- content/uploads/sites/3/2013/08/Honduras_risk_mapping_20120130_ETG.pdf. Tymkow, P., M. Karpina, and A. Borkowski, 2016: 3D gis for flood modelling in river valleys. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. - ISPRS Arch., 41, 175–178, doi:10.5194/isprsarchives-XLI-B8-175-2016. https://www.researchgate.net/profile/Andrzej_Borkowski2/publication/307530193_3D_GIS _FOR_FLOOD_MODELLING_IN_RIVER_VALLEYS/links/57cda82408ae83b37460e08b/3D-GIS- FOR-FLOOD-MODELLING-IN-RIVER-VALLEYS.pdf. Universitat Politécnica de Valencia, 2011: SUFRI Methodology for pluvial and river flooding risk assessment in urban areas to inform decision-making Report. USAID, 2017: CLIMATE CHANGE RISK PROFILE HONDURAS. https://www.climatelinks.org/sites/default/files/asset/document/2017_USAID ATLAS_Climate Change Risk Profile_Honduras.pdf. USGS, 2002a: Honduras Flood Mapping Publications and Products. https://wa.water.usgs.gov/projects/honduras/publications.htm (Accessed May 11, 2018). USGS, 2002b: Flood-Hazard Mapping in Honduras in Response to Hurricane Mitch. https://pubs.usgs.gov/wri/wri014277/pdf/WRIR01-4277.pdf. Victoria State Government, 2016: Victorian Flood Data and Mapping Guidelines. Victoria, Australia, https://www.water.vic.gov.au/__data/assets/pdf_file/0027/53757/Victorian-Flood-Data- and-Mapping-FOR-WEB.pdf. Villanón, M., 2003: Guía para la Elaboración de Inventarios de Elementos en Riesgo en Áreas Vulnerables a Inundaciones y Sismos en Costa Rica Individual Final Assignment Riesgo en Áreas Vulnerables a Inundaciones y Sismos. http://www.itc.nl/external/unesco-rapca/Informes estudiantes ITC/Analisis de riesgo en Canas Costa Rica/villalon_semanat.PDF. Westen, C. Van, and N. Kingma, Guide book Session 5 : Vulnerability assessment. 1–33. World Meteorological Organization, 2013: Flood Mapping. Integr. Flood Manag. Tools Ser., 88. www.apfm.info.
82