flood risk and flood hazard maps â€“ visualisation of hydrological

IOP Conference Series: Earth and Environmental Science

OPEN ACCESS

Flood Risk and Flood hazard maps – Visualisationof hydrological risksTo cite this article: Karl Spachinger et al 2008 IOP Conf. Ser.: Earth Environ. Sci. 4 012043

View the article online for updates and enhancements.

You may also likeFlood risk analysis in Sg. Sam, Kuala Krai,Kelantan using remote sensing and GIStechniqueWani Sofia Udin and Naimah Abd Malek

-

A Parallel Computation and WebVisualization Framework for Rapid Large-scale Flood Risk MappingMin Wang, Ruixun Lai, Runliang Xia et al.

-

The necessity of flood risk maps on TimiRiverGeogr Ctlin Aldescu

-

Recent citationsA.J. Kettner et al-

Land Use and Land Cover ChangeAssessment in the Context of FloodHazard in Lagos State, NigeriaDorcas Idowu and Wendy Zhou

-

Incorporating Uncertainty of the SystemBehavior in Flood Risk Assessment—SavaRiver Case StudyToni Kekez et al

-

This content was downloaded from IP address 126.203.212.39 on 26/10/2021 at 11:46

https://doi.org/10.1088/1755-1307/4/1/012043

https://iopscience.iop.org/article/10.1088/1755-1315/169/1/012060








https://doi.org/10.1016/B978-0-12-819412-6.00007-9

https://doi.org/10.1016/B978-0-12-819412-6.00007-9

https://doi.org/10.3390/w13081105

https://doi.org/10.3390/w13081105

https://doi.org/10.3390/w13081105

https://doi.org/10.3390/w12102676

https://doi.org/10.3390/w12102676

https://doi.org/10.3390/w12102676

Flood Risk and Flood Hazard Maps – Visualisation of Hydrological Risks

Karl Spachinger1, Wolfgang Dorner1, Rudolf Metzka1, Kamal Serrhini2, Sven Fuchs3

1 University of Applied Sciences Deggendorf, Germany2 Université de Technologie de Compiègne, Génie des Systèmes Urbains, France, and Université François Rabelais, Unité Mixte de Recherche, Tours, France3 Institute of Mountain Risk Engineering, University of Natural Resources and Applied Life Sciences, Vienna, Austria

E-mail: [email protected]

Abstract. Hydrological models are an important basis of flood forecasting and early warning systems. They provide significant data on hydrological risks. In combination with other modelling techniques, such as hydrodynamic models, they can be used to assess the extent and impact of hydrological events. The new European Flood Directive forces all member states to evaluate flood risk on a catchment scale, to compile maps of flood hazard and flood risk for prone areas, and to inform on a local level about these risks. Flood hazard and flood risk maps are important tools to communicate flood risk to different target groups. They provide compiled information to relevant public bodies such as water management authorities, municipalities, or civil protection agencies, but also to the broader public. For almost each section of a river basin, run-off and water levels can be defined based on the likelihood of annual recurrence, using a combination of hydrological and hydrodynamic models, supplemented by an analysis of historical records and mappings. In combination with data related to the vulnerability of a region risk maps can be derived. The project RISKCATCH addressed these issues of hydrological risk and vulnerability assessment focusing on the flood risk management process. Flood hazard maps and flood risk maps were compiled for Austrian and German test sites taking into account existing national and international guidelines. These maps were evaluated by eye-tracking using experimental graphic semiology. Sets of small-scale as well as large-scale risk maps were presented to test persons in order to (1) study reading behaviour as well as understanding and (2) deduce the most attractive components that are essential for target-oriented risk communication. A cognitive survey asking for negative and positive aspects and complexity of each single map complemented the experimental graphic semiology. The results indicate how risk maps can be improved to fit the needs of different user groups. Recommendations were developed of how to provide stakeholder-oriented information on hydrological risks.

1. IntroductionOn the European level, a lot of research institutes work in the field of flood risk assessment in different projects, such as EUROFlood, EUROTAS, FLOODAWARE or FLOODsite. In the field of interregional cooperation (Interreg), risk maps have been developed and tested as a first element of risk management in different projects, e.g. IRMA and TIMIS for the Rhine catchment. Complex

XXIVth Conference of the Danubian Countries IOP PublishingIOP Conf. Series: Earth and Environmental Science 4 (2008) 012043 doi:10.1088/1755-1307/4/1/012043

c© 2008 IOP Publishing Ltd 1

considerations and assessments of the flood process chain, linking different processes leading to flood disasters have been carried out, from precipitation to runoff generation and concentration in the catchment towards flood routing in the river network. Possible failures of flood protection measures, inundation within certain areas and the economic damage were evaluated through different methods[1]. A number of methods were integrated to repeat random sampling and to compute reliable results, such as Monte Carlo approaches. Such approaches are most suitable for computational use because of their reliance on repeated computation and random or pseudo-random numbers, and tend to be applied when it is infeasible or impossible to calculate an exact result by a deterministic algorithm – e.g. if probabilities have to be assessed in the field of risk analyses. With respect to the application of Monte Carlo simulations, a considerable amount of research results is currently available – on a local, regional, national and international level, and also related to practitioners’ needs. However, with respect to flood risk mapping, harmonisation is still outstanding. In 2006 the European Union published the drafts of the Directive on the assessment and management of flood risk to harmonise the research activities and the final results in the different European countries, which were put into force in 2007 [2].

The project RISKCATCH, funded under the CRUE ERA-Net initiative by the German Federal Ministry of Education and Research and the French Ministry of Ecology and Sustainable Development, addressed the issue of harmonisation in the assessment of hydrological risks and vulnerability assessments focusing on the flood risk management process. The RISKCATCH project aimed to deliver new, practical and viable solutions for an integrated risk-assessment-based management of natural hazards in Alpine environments and related forelands with a special focus on the interaction of technical and non-technical measures during recent flood events in catchment areas. By evaluating already available methods and results from other projects and experiences from other countries, the status quo and historical development of risk was compared and possible future developments were deduced based on scenarios. Necessary action was derived and efficiency of non-technical measures of flood defence was evaluated. Based on the assessment of historical and possible future development of the hazard, the values at risk and the associated vulnerability, non-technical measures were evaluated from an economic and technical point of view. Regarding different scenarios of the temporal development of risk, maps were generated and assessed. These maps provided the basis for the study on risk perception using the method of experimental graphic semiology.

2. MethodologyThe main focus of the project, also presented in this paper, was a user-oriented transfer of the results to different stakeholder groups, e.g. outputs of hydrological and hydrodynamic models, necessary modifications and final output in exemplary maps, which were finally evaluated by test persons from target groups. Focusing on medium-scale information on risk, test sites in the Alpine foreland were chosen. The project area in Bavaria is situated at the two rivers Vils and Rott. The Vils river has a catchment size of about 1450 km² and is flowing into the Danube river after 120 km. The catchment of the Rott river is 1200 km², and its length is about 100 km. Both catchments are part of the tertiary hilly landscape, an intensively used agricultural landscape. Rural structures, settlements with scattered buildings, the intensive agricultural use are determining factors. However, during the last two decades there was a remarkable increase of industrial and commercial use of land in potential inundated areas.


2

Figure 1. Project areas situated in Bavaria, Germany

2.1. HydrologyThere is a consensus that hazards resulting from hydrological extremes are on the increase (e.g., [3]). This fact is confirmed by evidence both for recent changes in the frequency and severity of floods as well as droughts and for outputs from climate models which predict increases in hydrological variability[4]. Because of major flood events in Europe in 2002 and 2005 there is much concern about the levels of protection provided by existing flood defences and their consequences.

During the last decades the characteristics of floods and land use patterns in the flood plains changed fundamentally. These changes are affected by climatic developments, land use changes, river development and training and also by measures of flood defence, enabling building in former flood plains and reducing the natural detention and profile for flood runoff. The main question is how manyupcoming impacts and actual changes will influence the future runoff scenario and change the characteristics of floods and extreme discharges.

In combination with land use maps, development plans and historical maps describing the changes of land use and damage potential in the flood plain, hydraulic and hydrological data sets were used to analyse the developments of flood-related problems for the status quo (ex-post perspective), i.e. flood risk. In combination with data for existing technical facilities of flood defence the effects (levels of protection and restrictions) of classical flood defence were evaluated and compared with non-technical measures. Future developments and underlying scenarios were derived and assessed with respect to risk management.

2.2. Hydrodynamic ModellingHydrodynamic models delivered the basic data for the compilation of hazard information. These data were derived from models, pre-processed and compiled in maps. HydroAS 2D in combination with SMS as a standard software model was used to simulate flood events; consequently, data were derived from this software. In a further step, these data were pre-processed in SMS, evaluated and visualised using ArcGIS 9 to derive the relevant maps (inundated area, water depth and flow velocity).


3

2.3. Hazard and RiskAs shown by Dorner et al. [5], maps indicating different information about floods and related risks have considerable relevance to visualise hydrological data in a spatial context (see Fig. 2). Regarding hydrological aspects especially the visualisation of probabilities is bound to restrictions e.g. if run-off is influenced by controlled detention structures.

a) b)

c) d)

e) f)

Figure 2. a) an intensity-probability diagram or risk index enables the combination of risk components in one map; b) explains the derivation of the different risk levels; c) and d) show the

extent of flood with different recurrence probabilities; e) and f) indicate the extent of extreme events (HQ100) on a large scale to provide an overview e.g. for the broad public.


4

In the area of natural hazards, risk is defined as a function of probability of occurrence and extent of damage. Extent of damage is constituted by the two factors damage potential and vulnerability. In general, this function has gained acceptance in accordance with the definition of the United Nations (e.g., [6],[7]) and is indicated in the form of a product as follows:

SiOjSiOjOjSi vpApR ,, (1)

where R = riskpSi = probability of scenario iAOj = value at risk of object jpOj, Si = probability of exposure of object j to scenario ivOj, Si = vulnerability of object j, dependent on scenario i

In Alpine countries, mainly in Austria and Switzerland, the procedure of hazard assessment is methodologically reliable in determining the hazard potential and the related probability of occurrence (pSi) by mapping, examining, modelling, and assessing individual processes and defined design events (see e.g. [8], [9], [10], [11]; eq. 1). So far, little attention has been given to the damage potential (AOj) affected by hazardous processes, particularly concerning spatial patterns and temporal shifts [12, 13]. Studies related to the probability of exposure of an object (pOj, Si) to a defined scenario and the appropriate vulnerability of the object (vOj, Si) have predominantly been carried out so far in terms of expert’s reports. Only few approaches and conceptual proposals determine the risk of property and human life (e.g.,[14], [15], [16]).

The different approaches of several German federal states concerning hazard maps showed some similarities. The approach of Switzerland (e.g. [17]; [18]) to indicate hazard uses the two parameters intensity and probability. It combines them in a matrix (probability-intensity-matrix) to indicate different hazard levels. This approach was extended by a third dimension to include also the vulnerability and receive a scale of color indicators to visualise risk [18].

In a further step the hazard maps were developed towards a risk map by taking further geographical data but also statistical data on elements at risk into account. Modifying the approach of Switzerland, a 3-dimensional cube was developed representing three input values, (1) probability, (2) intensity, and (3) vulnerability (see Fig. 3). Depending on the intended use of the map, different additional parameters were added.

The European Flood Directive recommends the creation of flood risk maps with different hazard criteria, (1) flood events with a high probability (HQ 10), (2) flood events with a medium probability (HQ 100), and (3) flood events with a low probability (extreme event). The content should include information on water depth and velocity as well as areas with embankment erosion and sediment deposit. Furthermore, the number of potentially affected inhabitants and the economic as well as environmental damage should be depicted. Apart from these considerably complex content required, it is necessary to keep the basis of these maps very flexible in order to allow for a computer-aided automatised creational processes due to the necessary update of these risk maps with a six-year frequency.


5

Figure 3. Risk cube including 3 dimensions: probability, intensity and vulnerability

The hazard intensity is reflecting two different factors: water depth and velocity of the water body. Critical water depths and velocities had been explored and were derived from hydrodynamic models. The values evaluated for the hazard maps were in an additional step used for the creation of risk maps. Therefore, hazard intensities were grouped into three different levels in order to fulfil the criteria of the risk cube, (1) 0-0.5 m, (2) 0.5-1.2 m, and (3) >1,2m. According to the European Flood Directive three relevant flood events were defined: (1) HQ10, (2) HQ100, and (3) HQ1000 (extreme event).

Regarding the extent of flood plains in the test sites, it was necessary to use cumulated data to efficiently consider vulnerability in the risk equation. In order to meet the requirements of a small-scale analysis, an economic approach was chosen to assess vulnerability. In doing so, vulnerability was calculated on an object basis for different types of buildings using damage functions depending on the water depth. Since for the test sites no record of historical or previous flood damage was available, design values and design loss functions were applied according to suggestions made by Meyer & Mai[19]; this procedure was possible due to comparable economic settings in the test sites. Due to the large amount of buildings situated in flood-prone areas, these damage functions were not applied to individual buildings but to groups of buildings and homogenous settled areas by aggregation. The spatially representation of vulnerability turned out to be an appropriate method for a small-scale analysis.

This procedure is analogously equivalent to the small-scale approach described in the Swiss guidelines [16] to provide an area-wide overview on susceptibility to hazard processes. Hence, vulnerability values varied between low (such as agricultural areas and individual farm estates), medium (such as dispersed settlements and small villages), and high values (city centres and industrial zones).

2.4. Experimental Graphic SemiologySince quality is an important aspect when looking at the European Directive on the Assessment and Management of Flood Risks [2], two major questions have to be addressed during the compilation of flood risk maps, (1) How to evaluate the quality of a map, and (2) How to produce a map adapted to the issue of communication and especially with respect to the public recipient concerned?

To quantify risk perception, the output maps were presented to a group of stakeholders from different European countries. The method used is based on the approach of experimental graphic semiology, reversing the traditional communication pattern from transmitter to receiver by a feedback loop allowing to integrate different perception patterns of multiple end-users (Fig. 4). Starting from


6

receiver, the maps were presented to the test persons using an ophthalmic device for the record of the eye movements during picture reading. The test was accompanied by a specific survey; hence, the cognitive perception of risk maps was evaluated. All maps were presented to the test persons for a relatively short time period to identify the most attractive components of each map. The eye movements were subsequently statistically analysed in order to assess patterns of visual perception for each map and to study the reading behaviour for text elements included in the maps. The visual strategies of each test person were quantified.The approach of experimental graphic semiology followed a three-stage procedure:

1. Compilation of the maps and the experimental protocol: Exchanges between the RISKCATCH partners allowed the development of reference maps while varying various elements of graphic semiology. Examples of significant maps in terms of information of the public and risk management of flood were considered. The RISKCATCH partners carried out a first proposal of maps to be used as support for the ocular pre-tests. Based on these pre-tests, draft amendments were made. These amendments were the origin for the choice which variables to be tested (position of the title and the legend, level of detail of the legend, merges of maps, etc) and for the subsequent development of 17 maps to be used as reference for the ocular tests and the cognitive study. The creation of a form for the cognitive investigation (in order to specify the results of the ocular recordings) and the identification of the various candidates also belonged to this preliminary stage.

2. Realisation of experimental graphic semiology approach: Recording the ocular movements of the selected candidates and undertaking the cognitive survey. Furthermore, the data were extracted and prepared for the statistic, static, and dynamic analyses.

3. Analysis of the recordings and interpretation of the results obtained: The results resulting from the statistical processing (calculation of the averages by map and by group of candidates), from the recording of the ocular movements (static, spatial and dynamic) and from the cognitive investigation were analysed and interpreted. Furthermore, recommendations for the compilations of risk maps were derived.

Figure 4. Principle of the cyclic model of experimental graphic semiology


7

3. Results

3.1. Historical DevelopmentAnalysing the temporal development of risk by using historical and actual risk maps resulted inremarkable changes in land use patterns, especially during the last 20-50 years. In Figure 5, the development of residential areas affected by the HQ100 and HQ1000 is presented for the Lower Vils region, based on the analysis of a multi-temporal series of land register maps. It was shown that land use increased considerably in recent decades, an associated increase in risk was proven.

Figure 5. Concerned developed areas of the Lower Vils region within the time period of 1850-2000

When comparing the different time periods, a tenfold increase of the developed area affected by the inundated area of a one hundred year flood resulted. Considering the thousand year flood, the affected area in the year 2000 was twenty times larger as it had been in 1850. Results from the GIS-based analysis of the entire study area supported these findings (Fig. 6).

Figure 6. Comparison of affected areas (left: year 1850, right: year 2000)


8

3.2. Hydrological IssuesWithin the catchments of Vils and Rott two major storage reservoirs are available for flood control. The storage reservoir situated in the middle of the Rott catchment was built between 1968 and 1971 and has a total capacity of 13.9 Mio m³ and a flood retention capacity of 12.75 Mio m³. The main purposes for the erection were (1) the protection of fertile agricultural areas, (2) flood defence for settlements in the downstream riparian areas, as well as (3) recreational purpose. The hydrological calculations undertaken within the RISKCATCH project showed that the estimation for the design floods used for the construction of the reservoir were underestimated. The evaluation of the control strategy of the reservoir based on actual data and a new hydrological model for the river Rott provedthat the expectations about the detention effect were considerably optimistic. For the storage reservoir an innovative control strategy was suggested, including (1) an increase of the flood discharge without any detention and (2) a transition to a flexible control strategy for higher discharges.

At the Rott already minor events result in the inundation of large areas of the intensively used flood plain. Modelling a discharge of 40 m³/s at the outlet of the reservoir, the first bridge and agricultural sites were affected, and during a discharge of 120 m³/s, first buildings and settlements were concerned. Thus, at the moment the basic discharge is set to a maximum of 40 m³/s. An increase up to 80 m³/s would be possible without considerable harm to buildings or infrastructure. However, the results had shown that even with an improved steering concept the reservoir, full protection for all assets in the underflow cannot be guaranteed due to the limited capacity (Figure 7, [4]).

Figure 7. Probability map with different detention pond steering scenarios (HQ1 and HQ5)

3.3. Risk MapsThe initial approach to use the population density for risk mapping is presented in Figure 8 for a section of the Lower Rott test site. In this Figure, the amount of affected inhabitants is shown fordifferent flood events (HQ10, HQ100, HQ1000). This initial and quite simple approach turned out to be effective, in particular considering precision and accuracy of data available and level of detail versus costs for data acquisition and compilation.


9

Figure 8. Detail of a risk map (variant 1: inhabitants)

In a second set of calculations, this approach was modified in order to achieve another type of risk information to be tested for communication purpose. Hence, the distribution of inhabitants provided the basis for risk information, and a separation of housing structures and industrial and commercial structures was performed on the basis of land use data. Simultaneously the number of employees of each single community was attributed to the commercial areas within the test site (Fig. 9).

Figure 9. Extraction of a risk map (Variant 2.1: inhabitants and employees)

In a further step the principle of the risk cube was applied to all risk maps (Fig. 10) in order to provide aggregated risk values to the group of test persons.


10

Figure 10. Extraction of a risk map based on the risk cube (Variant 2.2: inhabitants and employees)

To indicate the spatial distribution of risk, three to seven classes of risk were used based on either individual objects or the aggregation of values, e.g., settlement areas, industrial areas, agricultural areas, and forestry.

Linking process intensities, vulnerabilities and values at risk, small-scale risk maps were compiled for the test sites (Fig. 11). In doing so the focus was not only on different parameters of how to express risk, but also on the different communicative purposes with respect to the presentation of these maps to different stakeholder groups. All maps were compiled using a landscape layout in DIN-A1 format. The sets of maps were modified according to several variables in order to test the different accessibility and readability, and thus understanding by different stakeholders.

Considering that scale and consequently level of detail varied due to different size of the test sites, the concept of visualising risk was remarkable successful using spatial signatures already at scales ≤ 1:10000. This required the recalculation of values at risk for individual objects in order to adapt the values to larger areas.

Figure 11. Extraction of a risk map (economical damage)


11

3.4. Experimental Graphic SemiologyThe results of the eye tracking using experimental graphic semiology resulted in several findings related to the preferred layout of risk maps.

3.4.1. Statistic Analysis. The average number of fixations per map provided information concerning the extent of visual exploration; however, the 17 maps did not induce the same visual impact. Maps that included more graphic information, e.g., an infrared or coloured orthophoto representing the overall setting of the situation depicted, resulted in significantly more fixations than those maps that contained only a black and white land register plan for background information instead. Hence, the visual accessibility of maps is highly dependent on minimising information where possible and highlighting only necessary information since the eye visually fixes the most outstanding elements successively.

However, the average number of fixations per map is not only dependent on the map content, but also on (1) the effect of habituation, in particular the control group of laypersons, that gradually get accustomed to map reading, (2) the effect of tiredness, since the recording process spanned a period of 25 minutes per person, and (3) the repetitiveness character of certain maps with only marginal differences in depiction and thus visual distinction.

Those risk maps that followed general rules of design with respect to natural hazards, i.e., red and yellow hazard zones and only sketchy depiction of the overall setting, resulted in a considerable number of fixations within the group of specialists. On the other hand, a strongly coloured map background seemed to induce a noticeable attractiveness to the group of laypersons; the visual exploration was stimulated in particular for those maps that were based on infrared and coloured orthophotos. The group of people sensitized for map production or flood risk showed rather variable visual behaviour.

As a result of the statistic analysis, specialists in risk mapping do not scatter their attention over the map (less fixations per map) but do more thoroughly access the map content (more time for each fixation), while laypersons spend most of the time in a random access of the map content with no particular focus except from a certain interest in coloured information. A correlation between the number of fixations (attractive zones visually fixed) and the number of saccades (ocular movements that enable the passage from one zone of the picture to another) was observable.

3.4.2. Static Analysis. The static analysis was used to identify systematic and regular spatial patterns in map exploration. First, based on the analysis of video recordings, certain repetitive elements were identified that attracted the gaze of the entire sample of test persons. Second, certain specific elements of the maps could particularly be assigned to individual groups of test persons. In general, the focus of visual perception was on textual and coloured information, and areas with clear contrast; 90 % of the fixations were concentrated in these areas. There was a clear tendency that the gaze of all test persons followed along those information that was arranged in a vertical or diagonal order, e.g. along the river that was depicted on some of the maps.

The map title provided the first major information concerning the displayed content. If the title was located at the upper part of the legend section, the whole group of test persons looked at it during their first ten fixations. If the tile was located at the bottom, this pattern was not as clearly observable, in particular if the title was additionally written inverse in bright colours with a dark coloured box. Therefore, the title seems to fulfil its informative function best when placed above the legend and emphasized by good contrast (preferably black coloured text on a bright coloured background). This was also confirmed by the cognitive survey that accompanied the study.


12

The map legend is fundamental information for the understanding of the map content and enables the comprehension of the graphic symbolic system used. The test persons devoted between three and seven seconds out of 15 to the reading of the legend, which equals to approximately 20-50 % of the total exposure time. Placed either on the right or on the left side of the map, the legend attracted the eye; and vertical or quasi-vertical saccades clearly represented the process of vertical reading. However, the most coloured elements of the legend were the areas of major concentration of fixations. This tendency had an explicitly strong appearance if the legend was highly complex and contained a large amount of different information. A simple legend, containing up to five classes and two sets of information resulted in one set of ocular movements, while the visual strategy necessary for a detailed and complex legend generated two or even three sets of ocular movements.

The central element of every map is the cartographic information depicted. This element in general drew the most attention and the test persons spent 65-80 % of the exposure time of 15 seconds for analysing information. In doing so, the ocular movements were concentrated systematically on the most coloured areas, which represented the ‘colour effect’ [18]. In addition, the more a specific information was visually distinguishable –namely through the contrast created by the overlapping of various colours used and the map background – the more fixations focused on this information. Conversely, the less specific information appeared in contrast to the map background the more fixations seemed to be dispersed.

3.4.3. Dynamic Analysis. The dynamic analysis was based on the assessment of eye movements for all 21 test persons on individual sectors of the maps. It could be shown that the gazes were drawn to the main elements of the map, the legend and the figure part.

Maps with various centres of strong visual contrast resulting from a considerable spread of cartographic information lead to an accessible delivering of information. Conversely, for low-contrast maps it was shown by the dynamic analysis that the location and focus of the gaze scattered considerably. By the end of the time intervals during the test, these maps were almost entirely covered by the gaze due to little visual contrast between the main elements and the background. During statistical analysis, this was proven by a large number of short fixations.

Moreover, the sectoral analysis for the whole sample had shown that approximately two third of the observation time is devoted to less than one fourth of the map surface. While maps with dark background and only little visual contrast between the various elements depicted were entirely covered by the ocular movement of the test persons, maps with a clear distinction between central element and background showed less disperse fixations.

4. ConclusionRisk management for natural hazards is based on risk assessment techniques, including methods to determine the hazard potential and procedures to analyse and evaluate the damage potential exposed. For these management issues, risk maps provide the basis (1) for any planning and implementation of mitigation measures by public authorities as well as for the prioritisation of these measures, and (2) for any activities concerning regional development, land-use and construction engineering. Since maps are changing their function from final products to value-added geoservices, and the development of geoservices is more user-driven than purpose-driven, the overall aim of risk mapping includes (1) the delineation of areas endangered by defined risk thresholds, (2) the assessment of exposure levels in such areas, and (3) the communication of risk to various stakeholders, e.g. politicians, residents and other people concerned. It is axiomatic that, for communication to be successful, the message must be received, understood, accepted and acted upon. Therefore, the impact of information has to be


13

assessed in order to understand the attitudes of the target groups and to provide these issues appropriately.

Using the method of experimental graphic semiology had shown that the structure of maps influences the visual strategy of the readers; therefore, map perception is iconographic. The more accessible visual information is the more effective it will be in terms of visual transmission of information. Moreover, particular reading behaviour of specialists, sensitised people and laypersons led to the conclusion that perception is anthropic. Hence, risk maps should be compiled according to these different needs, in particular bearing in mind that approximately 65 % of the observation time of subjects is devoted to less than 25 % of the map surface.

To summarise, the spatial and dynamic analysis highlighted certain aspects that were identified as being important for an efficient design of risk maps, and will contribute to professional risk communication:

Coloured zones and written information concentrated approximately 90 % of the fixations. The concentration of information in the legend needs to be visible (contrast and colour used)

and accessible (limited number of information), to attract the eye and deliver information. The spatial localisation of information considerably influences the perception by the reader.

Even if the study is based on some restrictions and constraints, above all the limited number of 21 participating test persons, a number of general conclusions originating from visual strategies resulted (Fig. 12). Specific elements of semiology for a cartographic representation of risk include

a map background in bright colour to increase the contrast to informative elements, and to avoid an overload of information;

a sufficiently large legend, preferably on the right side of the central element of the map, with a limited number of information (five classes of discretisation) comprised from one range in colour only and arranged in decreasing values; and

a sufficiently large scale that the elements of the map are recognisable sufficiently rapid.A risk map compiled according to these conclusions would result in a visual strategy that is

composed from three clear sets of ocular movements (Fig. 13). Starting from the centre, the eye moves to the title of the map (1), following a vertical axis downwards the legend section (2) and returning back to the central element of the map (3). If there is sufficient time, the additional peripheral elements of the figurative part are explored subsequently.

It had clearly been shown that if risk maps will be designed according to certain guidelines, the information could be delivered in a visually efficient manner. Specific elements of semiology that have to be taken into account when designing risk maps include the contrast, the level of discretisation and the colour range and hue. Consequently, if risk maps are adjusted to these findings, risk communication will be enhanced, and awareness-building of the public will be increased. In particular concerning the European Flood Risk Directive, but also with respect to the overall aim of building hazard-resilient communities, future studies might include the applicability of risk maps within flood risk management plans.


14

Figure 12. Conceptual map suitable for efficient risk communication to different stakeholder groups

Figure 13. Representation of the visual strategy generated by the application of Fig. 12


15

References[1] Merz B and Apel H 2004 Risiken durch Naturgefahren in Deutschland. Abschlussbericht des

BMBF-Verbundprojektes Deutsches Forschungsnetz Naturkatastrophen (DFNK). GeoForschungsZentrum Potsdam Scientific Technical Report 04/01 (Potsdam)

[2] Commission of the European Communities 2007 Directive 2007/60/EC of the European Parliament and of the Council of 23 October 2007 on the assessment and management of flood risksDirective 2007/60/EC on the assessment and management of flood risks. http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:L:2007:288:0027:0034:EN :PDF(14 August 2008)

[3] Munich Re 2008 Topics Geo. Natural catastrophes 2007 (München: Munich Re.)[4] IPCC 2007 Climate Change 2007 Synthesis report – An assessment of the Intergovernmental

Panel on Climate Change (Cambridge: Cambridge University Press)[5] Dorner W, Schrenk C, Spachinger K and Metzka R 2006 Flood risk management plan -

Relevance of the hydrological behaviour of a catchment. In ed Bruck S and Petkovic T XXIII Conference of the Danube Countries on the hydrological forecasting and hydrological bases of water management CD-ROM

[6] Varnes D 1984 Landslide hazard zonation: A review of principles and practice (Paris: UNESCO)

[7] UN/ISDR 2004 Living with risk. A global review of disaster reduction initiatives (Geneva:United Nations)

[8] Kienholz H 1977 Kombinierte geomorphologische Gefahrenkarte 1:10.000 von Grindelwald(Bern: Geographica Bernensia G4)

[9] Grunder M 1984 Ein Beitrag zur Beurteilung von Naturgefahren im Hinblick auf die Erstellung von mittelmassstäbigen Gefahrenhinweiskarten (mit Beispielen aus dem Berner Oberland und der Landschaft Davos) (Bern: Geographica Bernensia G23)

[10] Bollinger D, Noverraz F and Tripet J P 1992 Die kartographische Darstellung der Bodenbewegungsgefahren als Planungsgrundlage: Ein Beitrag zur Vereinheitlichung der Gefahrenaufnahmen in der Schweiz. Proc. Int. Symp. Interpraevent 1992 (3): 191-200

[11] Kienholz H and Krummenacher B 1995 Symbolbaukasten zur Kartierung der Phänomene. Bundesamt für Wasserwirtschaft (BWW) (Bern: Bundesamt für Umwelt, Wald und Landschaft (BUWAL))

[12] Fuchs S, Keiler M, Zischg A and Bründl M 2005 The long-term development of avalanche risk in settlements considering the temporal variability of damage potential Natural Hazards and Earth System Sciences 5 (6) 893-901

[13] Keiler M, Sailer R, Jörg P, Weber C, Fuchs S, Zischg A and Sauermoser S 2006 Avalanche risk assessment – A multi-temporal approach, results from Galtür, Austria Natural Hazards and Earth System Sciences 6 (4) 637-51 Borter 1999 Risikoanalyse bei gravitiativen Naturgefahren BUWAL Umweltmaterialien 107/I,II Bern

[14] Wilhelm C 1997 Wirtschaftlichkeit im Lawinenschutz Mitt. Eidgenössisches Institut für Schnee-und Lawinenforschung 54 Davos

[15] Heinimann H, Hollenstein K, Kienholz H, Krummenacher B and Mani P 1998 Methoden zur Analyse und Bewertung von Naturgefahren. BUWAL Umweltmaterialien 85 Bern

[16] Borter 1999 Risikoanalyse bei gravitiativen Naturgefahren BUWAL Umweltmaterialien 107/I,IIBern

[17] BWW, BRP and BUWAL 1997 Berücksichtigung der Hochwassergefahren bei raumwirksamen Tätigkeiten. Bundesamt für Wasserwirtschaft (BWW), Bundesamt für Raumplanung (BRP), Bundesamt für Umwelt, Wald und Landschaft (BUWAL) Bern und Biel

[18] Serrhini K, Rochman J, Fuchs S, Dorner W and Spachinger K 2008 Sémiologie graphique expérimentale et cartographie du risque d’inondation. In Pinet, F & Miralles, A. (eds.) Actes de l'Atelier «Systèmes d'Information et de Décision pour l'Environnement». Proc. XXVIème


16

Congrès INFORSID – Fontainebleau (27.-30. Mai). Clermont Ferrand and Montpellier: Cemagref : 31-40 Casale R and Samuels P 1998 Hydrological risks: Analysis of recent results from EU research and technological development actions (Brussels: European Commission, Directorate-General Science, Research and Development)

[19] Meyer V and Mai S 2004 Überflutungsschäden im Küstenhinterland nach Deichbruch. Wasserwirtschaft 11 23-24

[20] IPCC 2001 Climate Change 2001: Impacts, Adaptation and Vulnerability Contribution of Working Group II to the Third Assessment Report of the Intergovernmental Panel on Climate Change (IPCC) (Cambridge, Cambridge University Press)

[21] Kienholz H, Krummenacher B, Kipfer A, Perret S 2004 Aspects of integral risk management in practice – Considerations with respect to mountain hazards in Switzerland Österreichische Wasser- und Abfallwirtschaft 56 43-50

[22] Casale R. and Samuels P 1998 Hydrological risks: Analysis of recent results from EU research and technological development actions (Brussels: European Commission, Directorate-General Science, Research and Development)


17

flood risk and flood hazard maps â€“ visualisation of hydrological

Documents