Selected issues of cartographic communication optimization for emergency centers

Abstract

Cartographic communication and support within emergency management (EM) are complicated issues with changing demands according to the incident extent and phase of the EM cycle. Keeping in mind the specifics of each purpose, it is obvious that spatial data used for maps preparation and production must be differently visualized even for the same type of emergency incident (traffic accident, fire, and natural disaster). Context-based cartography is a promising methodology to deal with the changing demands of an operational EM center.

An overview of cartographic communication is presented within the context of an operational EM center, activities of particular actors, and map use supporting the incident elimination. The authors of the paper respond to a series of questions, for example: what is the current cartographic support of operational EM in the Czech Republic in Digital Earth conditions? What possibilities are there to improve the cartographic communication? How can contextual cartographic services be implemented in a Web environment and how can the usability of results be tested? The paper gives several examples of the usage of cartographic technologies in map creation for various emergency situations.

Keywords:

• map use
• cartographic communication
• context-based cartography
• Digital Earth
• operational emergency management

1. Introduction

The idea of Digital Earth in a scientific, technological, and social context came after the successful and hopeful introduction and realization of the idea of the National Spatial Data Infrastructure, which was signed as one of the first Executive Orders (The White House 1994) by President Clinton. The framework of Digital Earth was formulated by Gore (1992 and mainly 1998), Vice President of the USA, at his memorable speech in Los Angeles. Gore defined Digital Earth as: ‘a multi-resolution, three-dimensional representation of the planet, into which we can embed vast quantities of geo-referenced data.’

His idea was followed by many other activities in different parts of the world (see also Foresman 2008, Goodchild 2008, Shupeng and van Genderen 2008). In Europe, the realization of the Digital Earth idea was going on successfully under a wide range of activities of e-Europe, e-Commerce, newly Digital Europe, or Digital countries (e.g. the UK, Germany), and many others which have been more oriented on e-Government and e-Governance design than on geospatial–oriented approaches. The latest initiatives of the European Union, Global Monitoring for Environment and Security (http://www.gmes.info), and especially the Infrastructure for Spatial Information in Europe (INSPIRE: http://inspire.jrc.ec.europa.eu) directive are rapidly and positively changing situation, and several countries have started discussing and designing spatial or localization strategies (such as the UK, Germany, and the Czech Republic). INSPIRE is based on three requirements: data must be available, accessible, and follow corresponding legal conditions. The vision of INSPIRE is to build a distributed network of databases on local, national, and European levels; each database will be managed in such a way that it will provide information and services required by both individual countries and the European Union. The databases will respect common standards and protocols, which will provide their interoperability and compatibility (Hrebicek and Konecny 2007, Konecny 2008).

The new generation of spatial data infrastructures (SDIs) and Digital Earth efforts fills up many requests and the current task is based on the effort to install a ‘spatial’ aspect in all governmental, societal, business, and other activities including the area of emergency management (EM). The idea of a ‘spatially enabled society’ coming from Australia and accompanied by support from many parts of the Earth is being formulated and developed.

Rajabifard (2007) proposes the following definition: ‘society or a government can be regarded as spatially enabled when location and spatial information are regarded as common goods made available to citizens and businesses to encourage creativity and product development. Spatial enablement uses the concept of place and location to organize information and processes and is now a ubiquitous part of e-Government and broader government ICT strategies. It is also defined as an innovator and enabler across society and a promoter of e-Democracy. As a result of this, we are potentially on the verge of the most dramatic change in the use of spatial information in our lifetime’ (Rajabifard 2007, p. 1).

The development of the above mentioned ideas with the enhancement of the societal context of Digital Earth and its benefits, costs, policies, and regulations, together with efforts to also create Digital Earth strategies at local and regional levels, allows realization of new ICT-based approaches in EM. These are more user-oriented and their design is based on detailed investigation of real users' expectations and users' profiles. In cartography, new forms of cartographic communication are proposed.

2. The role of cartographic communication in emergency management (EM) decision support

2.1. Background

A decision-making process within EM involves a significant amount of communication related to spatial information. Especially in a case of spatial information, visual communication is considered one of the most effective types of communication, proved by millennia of maps usage. The saying ‘A good sketch is better than a long speech,’ attributed to Napoleon Bonaparte, has many variations that reflect if not its truthfulness then at least its popularity. This commonly accepted fact was also proved by several empirical studies like Bartram (1980), Bell and Johnson (1992), and Glenberg and Langston (1992). Burkhard and Meier (2005) summarized the scope of the advantages of visual communication: to illustrate relations, identify patterns, present an overview and details, improve understanding and problem solving, and communicate different types of knowledge. In this paper, the form of visual communication where a map is the focal point is addressed. In order to distinguish it from other types of visual communication, the term ‘cartographic communication’ will be used.

In relation to crisis management, as mentioned above, spatial information plays a significant but not exclusive role. Moreover, this spatial information is not at all exclusively communicated through maps. Our interest is to ascertain why maps are used in crisis management in a specific way and fashion and if possible to make improvements in crisis management with the aid of more effective cartographic communication. Obviously, cartographic communication is inseparable from the rest of the communication processes. But according to practice, the improvement of crisis management is an iterative evolutionary process. It is logical for us to predominantly focus on cartographic issues of the decision-making process with necessary links to other tools.

2.2. Cartographic communication during emergency situations

Usage of maps within crisis management has been frequently discussed especially in the context of geographic information systems (GIS) involvement in EM support (Grönlund 2005). Current developments within the information systems for emergency and crisis management focus on several geoinformation-relevant aspects. Interactive map technologies including mashups for spatiotemporal analysis using Google maps were introduced by Liu and Palen (2009). Other aspects include rapid geoinformation dissemination via immediate use of geo-referenced images (Skinnemoen et al. 2009) and geo-collaborative techniques within EM using interactive cartographic tools (MacEachren and Cai 2006) and live video streaming from the field (Bergstrand and Landgren 2009). However, only limited attention has been paid to particular cartographic aspects which need to be further discussed.

It was documented that EM is a relatively conservative environment relying on paper maps. Therefore, proper implementation of GIS or computer cartography tools is not an easy task. The adoption of new technologies must be justified by a clear advantage over traditional tools. Such an advantage is achievable through a detailed understanding of the role of spatial information tools during the EM process. It is clear that GIS itself brings some significant improvements. Especially improved is the accessibility to information, amount of information, and topicality of information. But aspects of visual communication are often worsened by these advantages. The large amount of information combined with visual composition controlled by the user makes the processing of the information difficult. Graphical incapability (limited symbology, small display, etc.) also raises complications leading to a lingering preference for paper maps. In Sheppard and Blatchford (1995), the main benefits of visual information for EM are described:

• Speed of information access for the user.

• Provides images (maps, photographs, video, diagrams, etc.) which allow people unfamiliar with the location to obtain a fuller more realistic picture.

• Crosses cultural, language, and professional discipline boundaries.

• Provides a rich source of quantitative data and chronological documentation or updates.

It is clear that the speed of understanding information and a full picture of the situation are closely related. In fact proper visualization should even help people familiar with a location. It is also obvious that although maps are quite culturally and professionally dependent, they can make communication between people with different backgrounds much easier. The last point is more an issue of visual analytics than one of information visualization. In any case the potential of all these benefits can be exploited only through effective forms of cartographic visualization.

There are several domains in the above mentioned principles that are suitable for use in the improvement of cartographic communication effectiveness within the crisis management process:

• To reorganize communication flows in order to use redundancy of cartographic communication. Obtaining the same information from various channels can improve understanding and in some cases extend awareness of relations between various parts of information. But there is a necessity to care about the time sensitivity of information transfer within crisis management. So it is desirable to remove information channel loops and equivalent information resources. To release the information channel it is also necessary to remove from cartographic communication all parts, which do not need to be communicated in this way. In accordance with Burkhard and Meier (2005), the advantages of cartographic visualization are in the transfer of spatial relations, patterns, parallel display overview, and detail. If the knowledge requested is of a different kind from that mentioned above, it is not necessary to use cartographic communication for its transfer.

• To improve cartographic representation. It is possible to divide this task into two segments; the first is manipulation with the content and the second is manipulation with symbolization. In this case, overloading is a crucial complication of cartographic communication. Merely identifying and removing unnecessary features is not enough (in fact, according to users' opinions there are nothing as thematic features at all; in general everything is important). The solution is to separate cartographic representations into smaller parts tailored to particular tasks, which would include only features in the necessary amount and granularity to make a decision. This approach, based on context, is demonstrated in Staněk et al. (2007) and Mulíčková et al. (2009). Symbol handling presumes simplicity of drawing, familiarity and clarity of the symbol in relation to meaning, and clear identification of the symbol's importance. As already mentioned, topographic reference is most open to modification of representations.

• To improve the user interface. Here there are several issues for improvement:

  1. Context-based switching, which basically means minimizing the modification of map content by changing the item representing the solved task.

  2. Making movement inside the map face easy through named locations and active map features.

  3. Enabling geo-collaboration at least on the level of sharing of locations.

In the following text, an analysis of the current conditions of cartographic communication in KOPIS (EM regional operational and informational center) is presented, followed by discussion of possible modifications of these conditions in the framework of communication processes, content, symbology, and the user interface, along with reflection on contemporary cartographic research and empirical experiences.

3. Map use within operational emergency management (EM)

The organization of EM within the Czech Republic was described in detail by Kubicek and Stanek (2006) and Mulíčková et al. (2009). Only the regional dispatch center responsible for the operational decisions (KOPIS in Czech terminology) is taken into account in the rest of the analysis. In case of any extraordinary emergency situation or incident the regional dispatch center is alerted through a 112 distress phone call (similar to 911 emergency calls in the USA). The dispatcher in the regional center mobilizes the required emergency service: fire brigade, police, or ambulance, which together form the basic parts of the integrated rescue system (IRS). During relatively small incidents each IRS part works independently, although they consult at the place of incident. This situation corresponds with the incident analysis done by Diehl et al. (2006) and Snoeren et al. (2007).

Step-by-step analysis was performed in order to formalize the needs and current status of the cartographic support. The first step was the description of information processes within the dispatch center activities including brief characteristics and actors. We used the operational management analysis done by T-mapy (2004) and tested the real activities performance during the field surveys in KOPIS.

In the process of communication during the operational emergency situation, it is possible to distinguish several segments and roles. While the call operator only communicates with the emergency caller and his or her role is over after recording the accident, the incident operator is responsible for the successful accomplishment of the emergency situation/incident:

• Communication between the incident operator and call operator: the call operator provides the localization of the incident and creates a record of the call which is transmitted to the responsible operator.

• Communication between incident operator and spatial data source: the incident operator evaluates the situation and looks for appropriate information sources.

• Communication between incident operators: according to various responsibilities, the operator communicates with other operators either to acquire additional knowledge or to transfer subtasks to other operators – this is mostly verbal communication.

• Communication between incident operators and units in the field: finally the operator sends commands to field units and at the same time serves as the knowledge resource for this unit – this is verbal communication partially based on visual perception of the cartographic product.

Within these segments mostly mixed verbal and cartographic communications are used.

Cartographic support within the particular information processes is the second step of the analysis. Despite the advent of digital geographic databases, many traditional paper-based cartographic tools are still in use and so relevant printed cartographic sources are also mentioned. Both steps are described in detail in Tables 1 and 2, respectively.

Table 1. Activities within the operational emergency management demanding spatial operations.

Activity	No.	Characteristics	Stakeholders	Cooperation (type)
Emergency call localization	1	Spatial localization according to the call identification (telecommunication operator cells or fixed line address) or direct interview with the calling person	Call operator (112)	Vocational interaction with emergency caller, semiautomatic localization. Human–human
Incident classification	2	Identification and classification of the incident according to the standardized listing	Call operator (112)	Incident classification. Human–machine
Information processing and handover	3	XML data clause definition and handover to the responsible IRS units	Call operator (112)	Data clause definition. Human–machine and machine–machine
Dispatch decision about forces and means	4	If relevant for the fire brigade, the dispatch officer makes a decision about the necessary means and forces to be used for incident intervention including the best routing and information support for the intervention commander	Dispatch unit (150)	Dispatch center communicates with the intervention units. Human–human. Bi-directional feedback about the current situation at the incident
Alarm sound	5	Raising the alarm for intervention units of the IRS via phone or fire horn	Dispatch unit (150)	Human–human and machine–human
Processing of document for incident intervention	6	Intervention order and incident localization map printout. Optional: incident object fire protection details including detailed plans	Dispatch unit (150)	Human–human and machine–human
Navigation to and cooperation at the place of intervention/incident	7	IRS units' harmonization and control in order to solve the incident with the optimized amount of means and forces	Dispatch unit (150)	Human–human and machine–human
Intervention command	8	Optional, in the case when the intervention demands controlled cooperation with the regional headquarters	Dispatch unit (150)	Human–human and machine–human

Table 2. A map handling overview.

Cartographic documents	Document type	Timeliness and update	Geodata structure	Activity link	Interaction type
Paper maps	Road atlas	Commercial product	Reduced topographic map with extended road structure and specific POIs (park places, gas stations)	4	Visual contact
	Wall maps	Commercial product, often replenished by local updates (hand drawn)	Administrative units, IES regions, and local responsibility	4	Visual contact
	Operational plans	Available for selected buildings and areas. Regular update	Fire intervention large-scale plans for important buildings and areas with fire relevant thematic data (hydrants, rescue and evacuation routes)	5, 7, 8	Visual contact at place of incident
	Map sheets generated from digital geodata	Direct localization of incident generated on demand for intervention unit	Basic topography, routing, place of incident, fire and rescue intervention relevant thematic data (if available)	4, 6, 7	Localization of incident and orientation support at the place of incident
Web maps	Thin web client	Batch update from the central database	Only region relevant data available, often extended by region-specific or local data (POI, municipal transportation stops, detailed electricity infrastructure). WMS ready, but not used so far. Real-time information about weather conditions (CHMI) and traffic situation (ìSD) planned in the near future	4, 5	Not in use in the analyzed KOPIS center.
	Mapy.seznam.cz	Predefined map content from the information provider	Basic topography, POI, street network, and address points	4, 5	Query on demand for unknown address points or POI
Desktop GIS	ArcGIS	DTTO GiSel	DTTO	6, 7, 8	Spatial analysis and modeling of new spatial layers to be used in GiSel
	GiSel IZS	Only region relevant data available, often extended by region-specific or local data (POI, municipal transportation stops, detailed electricity infrastructure). Basic editing of local layers with a permanent update (brushwood burning)	Reference data: topography, administrative units, address points, economic subjects registry, road and street network. Thematic data: water management and network, rail and road segmentation network, forest inventory and management	4, 5, 6, 7, 8	Visual analysis and basic cartographic decision support
Mobile GIS	GiSel GPS (planed)	Online cooperation with intervention units	To be developed	7, 8	Update of situation from the field: incident place

Members of the fire brigade GIS commission play an active role in the development of an integrated emergency system (IES) GIS operational module, which is interconnected with the whole robust national system. Its infrastructure also serves as the technological platform for the whole operational management. Four basic types of cartographic outputs are used within the operational EM process: paper maps, Web maps, desktop GIS, and mobile GIS. Their particular output type, timeliness, geodata structure, potential links to the EM operational processes, and type of interaction are described in Table 2.

While the paper maps and selected Web sources are used only for specific purposes, the key role within the cartographic support is played by the GISel IZS desktop application (for details see Skrivanek et al. 2009). Both the geodata types and cartographic visualization have been changed and extended in the last years. However, in the description of modifications is the following peculiar statement:

Current map project left the cartographic correctness of the map, uses non-contrast pastel colours for depiction of basic map (e.g. arable land is not depicted at all, see Figure 1), whilst important depictions created by annotations, territorial division and other points that are necessary for prompt and right decision of the operator, are highlighted and SQL inquiries were substituted by spatially generalized data kits. (Skrivanek et al. 2009)

Even if what is meant by cartographic correctness is not comprehensible from the statement, the ‘new’ map is visibly graphically improved from the cartographer's point of view. The main problems from which the original map suffers, overload or thematic universalism, remain nevertheless untouched.

Figure 1.  Original (left) and current (right) types of visualization used in GiSel IZS (from Skrivanek et al. 2009).

4. Possibilities for improvement in cartographic communication

Following the analysis of map use in the operation center, modifications for improvements were proposed. These improvements are based on a combination of empiricism, available studies, and recommendations from the areas of map use and psychology. As already stated, improvements were categorized according to their relation to information flows, map content, map symbology, and user interface. Nevertheless, issues of symbology are mutually related to information flows and map content. Symbology within all categories will therefore be discussed.

4.1. Information flow arrangement

When analysing the arrangement of information flows, it was possible to identify three issues:

1. Redundancy in symbol design.

2. Redundancy of maps.

3. Information loops.

4.1.1. Redundancy in symbol design

In the case of the arrangement of information flows, the first identifiable issue is the redundancy of attributes expressed by symbol. In principle, due to the redundancy's role in the cognition processes, avoidance of the redundancy is not completely desirable. Repetition of information has an important role in fixing patterns into the long-term memory (Alkon 1989). Even more frequent use of redundancy in cartography is the case when the same information is expressed by several graphical means. The goal is to accent either the importance of the feature or the main theme of the map (MacEachren 1995). But there is also the reverse side of the redundancy in the graphical encoding. The information capacity of the mapped area is obviously strongly decreased in the case of excessive redundancy. This consequently either limits the communication of multiple attributes related to the feature or increases graphical overloading and consequently makes reading of the map difficult (Figure 2).

Figure 2.  Multi-parametric symbol with (left) and without multiple attributes (right).

There are two hypotheses concerning improving usability:

1. Avoiding graphical variables redundancy for vital features (this is obvious in the case of background features). The full-size map usually has no place for graphical variables redundancy due to its complexity.

2. Using graphical redundancy for underlining the most important vital features. It is necessary to understand the possibilities and restrictions in combining various graphical variables in relationship to the feature they are representing. For example, changing the shape of the symbol together with the color can obscure the affiliation of the symbol with the group of symbols it belongs to. The shape and color are strongly bound to quality. To make the symbol easier to spot, it is better to use a graphical variable assigned to quantity, such as intensity or size.

4.1.2. Redundancy of maps

Another redundancy issue is offering the same vital information through several maps. Such a procedure is counterproductive for making an effective decision in limited time. Users spend the same effort on cognition to gain the same or similar information. Besides, processing another cartographic representation with different concepts, but the same meaning, decreases efficiency. A clear signal of the map's dysfunction is the necessity for the map user to consult another map for information included in the map originally used (Figure 3). When communicating over the map, it is undeniably advantageous for both sides to have identical map environments covering all the necessary information.

Figure 3.  Finding the way: by road atlas (part of the common knowledge; left), in an emergency management-oriented environment (expert knowledge; right).

The hypothesis for improvement is the map environment should be ‘all in one’; that is, the need for operators to consult different map systems has to be eliminated.

4.1.3. Information loops

The last case of flows organization issues is information loops. An information loop is caused by the necessity to gain again the same knowledge that has already been discovered. This issue is strongly related to the design of the user interface. An information loop is the consequence of the impossibility of making a graphical comment or bookmark tagging the already discovered knowledge inside the working environment. An information loop is also the consequence of usage of multiple disconnected cartographic environments.

The hypotheses tested here are:

1. Incorporation of ‘redlining,’ a simple comment layer with or without navigation over comments. Redlining comments can be realized by predefined symbols with distinct visual parameters.

2. Enabling named extents: a user can save the selected extent with a short explanation.

4.2. Map content and graphical load

It is possible to look at elements included in a map from at least three different angles:

1. Map content.

2. Graphical load.

3. Layering.

Map content is topically focused on the quality, graphical load is focused on the amount of elements on the map, and layering is focused on the visual order of objects. All of them have an impact on information gain.

4.2.1. Map content

Map content is the scope of cartographic features (classes) present on the map. In the case of analogous maps, the content is fixed on the medium (paper, but it can also be a PDF file) and as such is visible from the first moment. The map content of electronic maps can be due to their ability to dynamically change width, and is then in the concrete moment visible on the screen. For map perception, it is generally convenient to minimize the content actually displayed.

It is possible to divide the features on a map into:

• Background features, which play the role of spatial reference.

• Vital or active features, which express the function of the map.

• In some cases, complementary features, which link the map function to the broader context (some concurrent theme, features persistent through the task, or custom features).

Possible improvements related to the background features are:

1. From the topographic dataset, features which are not necessary for localization are omitted and the granularity of the remaining features is decreased by thematic aggregation. The visual appearance of the background is suppressed. The reduction cannot be uniform: in this case, space was divided into urban, suburban, and rural areas with different contents derived from user testing (for comparison, see Figure 4).

2. The granularity of the background features is reduced and their visual appearance is suppressed, but there is no omission of features.

3. Only suppression of the visual appearance of the background is carried out (the motivation here is to keep the standard topographic background familiar to users).

A special case is to use one of the virtual globes (Google Earth, NASA WorldWind, or Marble) for the visualization environment. All of the mentioned virtual globes support the technology of Web Map Service (WMS) layers, which is used for the compilation of vital features at the context service. The following possibilities for how to use these environments as emergency center cartographic tools exist:

1. Using a standard topographic background of a virtual globe service and using a very simple representation of vital features with semi-opacity.

2. Masking the area of interest with a gray semi-opaque rectangle for visual suppression of the standard background.

3. Masking the area of interest with a proprietary topographic layer with similar parameters as discussed in the previous paragraph.

For vital features, the approach based on contextual cartography was chosen as the best method of improvement. The context is a set of determinants identifying the particular cartographic representation. If something happens around the map device, its context is changed and the appropriate visual representation is selected. The selection of attributes defining context is strongly related to the overall purpose of the map. In the case of the operational center, the core of the context definition lies in the identification of user solved tasks, which in effect divides vital features into themed groups of features related to the tasks. In Table 3, identified tasks based on the KOPIS procedures analysis are listed.

Figure 4.  Comparison of the map with full-scale topography (left) and with strongly reduced features (right).

Table 3. Incident classification for emergency calls.

Incident type	Incident categories
Fire	Low buildings
	High buildings
	Industrial, agricultural, and storage buildings
	Assembly points
	Underground spaces, tunnels
	Electricity distribution point
	Transportation vehicles
	Dustbins, waste, dump sites
Traffic accident	Persons extrication
	Vacation of road, towing
	Road cleaning
	Railway accident
	Aeroplane accident
Natural disaster	Flood, torrential rain, rain
	Snow, hoarfrost
	Gale
	Landslide
	Other
Technical assistance	Cooperation with Integrated Rescue System
	Building destruction
	Opening of locked space
	Obstacles elimination
	Elimination of dangerous circumstances
	Water pumping
	Monitoring and measurement
Persons and animal rescue	From height or depth
	Closed spaces, elevator
	Smothered
	From water
Other extraordinary events	Radiation accident
	Evacuation of inhabitants
	Other
Undefined	Undefined
	Technological testing

For the content change, there are two triggers: the users need to change the solved task and the level of detail (scale). The change of scale is less significant than the change of task, as most emergency situation solving happens at the operational scale only. The use of contextual cartography is the core of visualization improvement (Figure 5).

Figure 5.  Example of two different contexts (visualization for transport of dangerous cargo: monitoring (left) and incident (right)).

4.2.2. Graphical load

Graphical load is tied to the number of objects currently present in the mapped area and is usually represented by the percentage of the mapped area completely covered by graphical elements. Nevertheless, graphical load is hard to quantify, as not all the visualized elements are included in the final value of the coverage. The included areas do not allow another layer of the graphical object to be placed over them. This usually means graphical objects or their parts in black or dark colors, also including annotations. Hojovec (1987) determined the average graphical load of a map as 20% and the maximum as 36% on the basis of various thumbnail rules.

It is necessary to mention that Hojovec (1987) determined the graphical load for analogous maps. In the case of electronic maps, it is not advisable to try for higher values of the graphical load. On the other side, an ‘empty’ map can be the source of users' disturbance or dissatisfaction with the amount of information obtained.

4.2.3. Layering

The term ‘layering’ implies the visual order of objects where the upper ones are usually considered as more significant than the lower ones. But the first rule that the ‘thematic’ layering has to abide by is graphical correctness. All the objects on the map must be visible. Layering can be supported by opacity and shadow, but there is graphical impact on the coloring and the shape of the objects. Bertin's (1974) rules are followed as there is no space to reduce the number of layers.

4.3. Symbology

The graphical part of symbology in cartography deals with manipulation using graphical variables. Bertin (1974) specified seven graphical variables: size, shape, color, orientation, texture, and position. These variables are dealt with pre-attentively in cognitive processing and are therefore quickly processed (Healey et al. 1995) and play a significant role in cartographic legibility.

Legibility can be defined as the amount of effort one needs to detect and discriminate objects symbolized on the map (this means there is a clear difference in graphical variables and the ability of the symbolized object to be separated from the surroundings). There is limited possibility to express multiple objects’ attributes by graphical variables. Every added variable has negative consequences for legibility.

Readability is, analogically to legibility, defined by the amount of effort one needs to make to correctly interpret a map. As already mentioned, map reading involves several issues, such as the meaning of symbols, patterns, and relations. Improper usage of graphical variables can support misleading patterns.

Most of the symbology issues can be fixed by the application of standard cartographic visualization approaches (for example usage of complementary colors). In cases where possible alternative approaches exist, it is necessary to evaluate the best option for the concrete situation. For immediate testing, the following set of hypotheses concerning symbol design was selected as the most common in reading issues:

1. Enforcement of an important feature by:

   • Symbol size: the size of the symbol is ruled by its immediate importance in accordance with the concrete situation and is excluded from scaling.

   • Outline: the size of the symbol does not change, but there is an additional outline in a different color, width, or both.

   • Patch: the symbol is underlined by a colored area or shading. The effect can be very similar to the usage of outline.

2. Supporting understanding by association with:

   • Geometric shapes: these are easy to discriminate, but usually the associative power is very low.

   • Schematic pictograms: relatively thorough previous knowledge of the symbol system is assumed. This approach combines a good level of association with low impact on the graphical load of the map.

   • Complex pictograms: despite the near reality of the objects’ depiction, there is a possibility of confusion when representing an abstract object (e.g. a picture of a horse can represent stables but also a racecourse).

3. Enforced layering by:

   • Shade: simulated 3D effects.

   • Color: the enforced theme is represented by coloring the symbols with a specific tint.

   • Shape: the enforced theme is represented by uniform symbols.

4. Symbol hierarchy and articulation: it is possible to decompose the symbol. Individual parts of the symbol have specific meanings according to the object's attributes.

   • Color

   • Shape

   • Structure

5. Use of alerts and selections: similar to enforcement.

   • Highlighting by color.

   • Highlighting by size.

One of the most significant influences on symbol reading is bias. Users are accustomed to reading existing maps with defined legends and expect similar symbols to have similar meanings. Uniformity of symbol systems is also helpful for geo-collaboration. There are a few attempts to create standards at national and international levels. Nevertheless in the Czech Republic and surrounding countries such standards do not exist. In this aspect, we can only use experiences from United States Federal Geographic Data Committee – Homeland Security Working Group (FGDC HSWG 2005; Dymon 2003) and partially from risk management-oriented Dutch standards (van den Worm 2006). A more detailed description of the relationship of our project to existing standards was already described by Friedmannova (2009).

The proposed methods of possible improvements in this case are:

• implementing the selected international standard (such as FGDC HSWG);

• keeping those symbols that are already in use as far as possible; and

• defining a system based on common conventions (for example, road signs of which there is common awareness).

4.4. Annotations

Special types of cartographic features are annotations. They are usually attached to objects bringing a name, additional attribute, or information about some significant aspect of the object to which they are attached. There exist independent annotations serving fuzzy definition of areas which are usually part of the topographic background (mountain ranges, regions without explicit boundaries; the annotation is not attached to any of the map objects, but stand in the object's place). In the case of vital features, a policy of avoiding annotations altogether and using just data tips is adopted. So there is only space to modify annotations in the background.

4.5. User interface

A user interface is nowadays an integral part of an electronic map. A user-friendly environment for map manipulation is crucial for its adoption by users. We can distinguish between a basic and an extended set of functions. All graphic tools contain a set of similar functions, which are anticipated by experienced users when moving from one environment to another. The functions included in this set are the basic ones. Based on longstanding research (for example You et al. 2007, Haklay and Zafiri 2008), the functions belonging there are:

• Zoom and pan within the map face.

• Feature visibility.

• Switch on/off.

• Simple full-text search.

The enrichment of the user interface lies solely in the area of extended features by:

1. Involvement of parametric queries, which extend the possibility to search for objects inside the map. Because the user usually does not know the parameters of the objects, there are lists of classes of parameters or intelligent features (type of query by example).

2. Involvement of a dynamic map key: a map key is bound with displayed features in contrast to so-called structured map keys (what is seen in the specific moment on the map is also seen in the map key).

3. Usage of data tips: predefined, dynamic by relevance, or user defined.

Ergonomic issues such as the following also belong in the frame of the user interface problematic:

1. Minimization of the number of clicks needed to accomplish a desired goal.

2. Preserving overall orientation when moving between details.

3. Preserving continuity of the work logic during map use (steps back and forward, tracing of multiple solved objects, and visual situation indicators).

All the above mentioned mechanisms are objects of map usability testing. Due to the sensitivity of the working environment in the EM, it is virtually impossible to carry out the testing in situ.

5. Testing of the map adjustments

5.1. Background of the map usability testing

Issues of the map's effectiveness can be generally addressed by map usability research (for example van Elzakker et al. 2004, Wachowicz et al. 2005). This kind of research follows longstanding study of map usage with the implementation of usability testing within the software engineering. This fusion is the result of modern computer cartography development. An electronic map becomes a software product as both a GIS front-end and an independent form. A graphical user interface is now an integral part of map design and in some senses the map itself becomes a user interface to GIS. Within usability testing, it is not possible to test just the speed of knowledge processing, because such results are difficult to interpret. It is necessary to decompose part of the knowledge acquisition for them to be more related to various parts of the map design. In relation to Larkin and Simon (1987) and Burkhard and Meier (2005), the parts of the process of knowledge acquisition are as follows:

• Attention: visual attractiveness, readability, and legibility; there are significant perception issues, or to put it differently, the visual representation is easy for the user to follow.

• Overview: an understanding of the general situation, relations, and patterns; this kind of knowledge is the ‘raison d’être’ of cartographic communication.

• Detail: what is discoverable and what can be identified; despite the previous statement, it is also necessary to realize particular objects and their attributes on the map.

• Metaphor: understanding the concept of how this particular map works: how the map generalizes reality and represents it by symbols. Full understanding is achieved if users can create a similar map. Another proof of concept understanding is parallel orientation in different maps.

• Motivation: the ability to discover map possibilities and extend their use in case of another decision.

The map effectiveness or usability is not exactly equal to its quality. The issues of map quality involve consideration of resource data quality, which is beyond the strictly cartographical part of map processing. Evaluation by map users plays a key role in usability testing. As mentioned in Nivala et al. (2007), the way to construct a system by continuous verification by users is very extensive and hardly feasible for such projects as cartographic support for crisis management. In this case, combined perception testing, user surveys, and cartographic experiences are used to establish a set of principles to create alternatives which are then subjects of usability testing.

5.2. Testing

For the purposes of usability testing of proposed map improvements, an interactive Web-based testing tool was developed. The tool was designed in order to test a wide variety of inputs from isolated cartographic symbols or symbol sets to both static and interactive complex map composition. The main goal is to test different cognitive functions from long-term memory to immediate attention, perception, and decision functions. Test results are stored in a database and automatically evaluated for further statistical processing.

Basic test functionality includes the test person identification and pre-test calibration of individual computer and cartographic abilities. Within the test environment, three basic types of tasks/scenes exist: forms with pull-down menus with predefined testing answers; visual choice scenes, where the test person is forced to choose one or more possibilities among visual variables; and localization tasks, where the test person must place the symbol in the right position or draw a line or polygon of a certain task (Figure 6). Both reaction times and positional accuracy are stored and processed further.

Figure 6.  The testing tool – pre-calibration – task tests the ability of the test person to draw a line within the Web environment and to keep the edited line inside the bounding tolerance buffer (left); the real test task: choosing the evacuation route (right).

The influence of acute stress on cognitive processes and map reading was studied by Stachoň and Šašinka (2009), who concentrated on understanding the stress mechanism and its influence on particular cognitive functions, such as short-term and long-term memory, attention, perception, and decision making. They further explored the possibilities and methods of testing the usability of maps and validated the results by statistical analysis tools using a combination of quantitative and qualitative methods.

Ongoing tests have more of a supporting role in the evaluation of the proposed thesis of map design discussed in Section 4. To make exact statements according to general map design is very difficult due to the size of tested samples and the holistic nature of cartographic mechanisms. Significant differences in map use can disappear with an increasing number of tested subjects (which is quite difficult). Every change of cartographic method interacts with the particular content of the map and it rarely makes sense to change just one graphical variable to achieve improvement. It is hard to identify one positive factor and even a complex of interacting factors cannot be easily appointed as a principal map improvement agent. So we are often limited in comparison of two maps and the empirical evaluation of causes.

With the above mentioned testing environment, we collected some results with the limitations mentioned in the previous paragraph. In a PhD thesis (Zboril 2010), existing maps used contemporarily in EM were compared with part of the project proposals on improvement. The following belong to the interesting results:

• significantly better perception of ideograms over annotated symbols;

• simple pictograms are better perceived than complex logo related pictograms;

• changes in the font parameters used do not change the search effort;

• a higher amount of visually enhanced features decreases the effectiveness of searching the map; and

• a higher density of topographic background objects decreases the effectiveness of thematic feature searches.

In a Master's thesis related to orientation on a map in relation to different topographic backgrounds (Svobodova 2009), unambiguous results were not achieved, but a significant role of sufficient color contrast in map reading was identified.

A larger testing experiment accomplished in Autumn 2009 dealt with alternative base maps and the efficiency of use perceived by military users and geography students. Preliminary results have shown significant differences between the user groups, proving that the level of map use skills influences the speed and durability of task accomplishment. Besides that, simplified topographic base maps have proven higher suitability for localization than orthophoto base maps. Details of the test including descriptions of particular tasks are given in Konecny et al. (in press).

5.3. Prototype implementation

In order to realize the process of contextual cartographic visualization, a contextual map service called Sissi has been developed (Kozel and Štampach 2009). Sissi is a Web-based server application intended to provide contextual (or context aware) maps for map clients (Figure 7). Communication between Sissi and map clients is ensured by an extension of the WMS specification (Open Geospatial Consortium (OGC) 2008). The extension has been designed in order to provide contextual abilities so that any WMS client can request any contextual map. A more detailed description of context types and Sissi implementation can be found in Kozel and Štampach (2009). One of the advantages of such technology has already been discussed. It is possible to combine WMS layers with virtual globes and therefore to incorporate a contextual map service into the Digital Earth framework.

Figure 7.  Sissi contextual Web map service interface.

6. Conclusion and future work

As most of the maps used in EM are only modifications of already existing maps, overload of the maps is easily reached. The GIS and visualization tools used, for the most part, emulate analogous maps and propagate their disadvantages, such as inconsistent symbology (created by adding new symbols when needed without their integration into the existing graphical schema) or overload. Although there are clear improvements in the system over time, the changes are mostly on the surface and the structure of the system remains untouched. In the area of cartographic communication, there exists large space for improvements.

As there is a large quantity of studies on human perception, it was decided to adopt their results when possible and to focus on cognitive issues. Unfortunately, higher levels of cognition are strongly individual and the results of studies are mostly not transferable. This is probably also the reason for their rarity.

Due to map complexity, a simple change in the design of one map symbol has a deep impact on the psychological effect of the map. This means that it is much easier to evaluate map usability than to trace relations between the symbol change and the map use cognition process. We propose to involve more measurement of map complexity in results; nevertheless map improvements will follow the conceptual possibilities discussed in Section 4.

Testing results were very close to the assumptions based on empirical observations and expert-based knowledge. Nevertheless, it is necessary to state that the tested group was limited in numbers and did not include operators. Testing on incident operators is planned for the final stages of the project, when the proposed symbol systems will be completed.

To ensure the validity of testing, we propose to increase the number of tested persons to obtain sets exceeding 50 participants per task. We consider testing of the system of operations of EM dealing with maps to be even more important. A context-based approach leads to lighter maps which are supposed to be better and are likely to be more efficient than existing overloaded multipurpose maps. The aim is to prove that the chosen approach is significantly better in the whole operation process incorporating multiple uses of maps and collaboration between incident operators.

Notes on contributors

Karel Staněk has worked as an assistant professor at the Department of Geography, Masaryk University in Brno, since 1998, where he specializes in analytical and computer cartography.

Lucie Friedmannová has worked at the Department of Geography, Masaryk University in Brno, since 2000, where she specializes in cartographic visualization, graphical design, map use, and the relationship between art and cartography.

Petr Kubíček is Vice Chairman of the Czech Association for Geoinformation. Since 2005, he has worked at the Department of Geography, Masaryk University in Brno, where he specializes in digital cartography and geospatial solutions in crisis management.

Milan Konečný is a professor of Cartography at the Department of Geography, Masaryk University in Brno, where he specialize in GI studies and interoperability and is also the Head of the Laboratory on Geoinformatics and Cartography. He is also former President of ICA and Chairman of the ICA Working Group on Cartography for Early Warning and Crisis Management, Vice President of ISDE, and Member of the Task Force Group for Early Warning of JB GIS.

Acknowledgements

The project (No. MSM0021622418) is supported by the Ministry of Education, Youth and Sports of the Czech Republic.