Eye-tracking analysis of the influence of relief shading on finding labels on tourist maps

is paper evaluates the influence of relief shading in finding labels on tourist maps. Two types of maps were compared, one in which terrain was depicted with contour lines and spot heights, and the other enhanced with relief shading. -e task was to find specific hills and villages. Two aspects were investigated – whether shading helped users find hills, and whether shading made it more difficult to find villages. -e eye-tracking method was used for this study. -e results indicate that respondents prefer shaded maps from an aesthetic point of view. Pairwise comparison of individual stimuli pairs and groups of stimuli was performed with the use of five eye-tracking metrics. Most of the eye-tracking metrics were significantly different for most of the stimuli. -e results of the experiment show that shaded maps are less suitable for finding hills and villages. -e least effective result was observed in finding villages on a shaded map. key words cartography – maps – eye-tracking – relief shading – visual perception – map reading popelka, s. (2018): Eye-tracking analysis of the influence of relief shading on finding labels on tourist maps. Geografie, 123, 3, 353–378. Received November 2017, accepted June 2018. © Česká geografická společnost, z. s., 2018 354 geografie 123/3 (2018) / s. popelka 1. State of the art Contouring is the most frequently used technique to provide map readers with elevation information (Dušek, Miřijovský 2009). Contours are isarithms – lines connecting points of equal elevation. Most casual map readers cannot imagine landforms indicated by contour lines. As Castner and Wheate (1979, p. 78) stated: “Relief depiction with contours is not usually immediately interpretable or imaginable, especially for inexperienced users.” Instead, most people recognize shapes primarily by the interplay of light and dark. -is method of portraying the landsurface form is called shading (Robinson et al. 1995). Shaded reliefs are already found in early manuscript maps from the seventeenth century. With the invention of lithography (1798), it became possible to print half-tones. Cartographers began to combine relief shading with other means of terrain display. A_er World War II, many reliefs were drawn by airbrush. Since the end of the twentieth century, relief shading has mostly been generated from digital elevation models (Jenny, Räber 2015). -e angle of lighting must be defined when using shading in maps. Conventional lighting comes from the upper le_ corner of the map – from the northwest in maps of the northern hemisphere (Bernabé-Poveda, Sánchez-Ortega, Çöltekin 2011). A recent publication (Biland, Çöltekin 2016) suggested that NNW illumination is better than NW. Imhof (2007) states that this may be due to people writing from le_ to right, light being on the le_ hand side when holding a pen while the right being in shadow. Even though this situation cannot occur in reality (in the northern hemisphere), this light is psychologically the most effective in perceiving terrain plasticity (Imhof 2007). Most users are accustomed to the light from the northwest, therefore a map illuminated from the south will be perceived negatively (Imhof 2007). -is was demonstrated with the use of eye-tracking in the author’s dissertation (Popelka 2015). -e main aim of relief shading in maps is to provide information about height, but it also has an aesthetic function. Ortag (2009) published a chapter focusing on the variables of aesthetics in maps. In this chapter, the results of focused interviews with more than 150 participants were summarized. Relief or 3D impression was mentioned as a the third most important reason for describing a map as beautiful (a_er color and readability). Usability evaluation of visual representations have gained much attention in recent cartographic and visual analytics research (Coltekin et al. 2009, Fabrikant et al. 2008). “Usability evaluation allows us to obtain data, o_en quantitative, about aspects of a system or user performance with that system which can be used to identify aspects that are problematic for users and highlight potential fixes. -ese methods can also be used for comparative purposes, for example, against established benchmarks or alternative designs or products in order to identify eye-tracking analysis of the influence of relief shading... 355 which is easier to use or identify their relative advantages and disadvantages.” (Fuhrmann et al. 2005, p. 559) -e evaluation of shading on tourist maps is important because it is not known whether shading helps map readers or not. Both map types are available on the market – shaded and non-shaded – but the study of usability of these types of relief visualization has not been previously performed. -e motivation for this case study was to learn whether shading affects search performance when looking for a specific object on a tourist map. -e research questions that guided this study focus on the basic search tasks that can be done with tourist maps. -e first question investigates the search for relief-related objects (hills), the second, the search for non-relief-related objects (villages). -e research questions are: 1. Does shading help participants find hills, as participants need only scan the darkest areas (representing hills), not the whole map? 2. Is searching for villages slower on shaded maps, as these maps are darker by design than maps without shading and may negatively impact its legibility? Maps that include a comprehensive representation of terrain together with a landform relief (topographic or tourist maps) have not been studied from a user perspective as deeply as other types of geovisualizations (i.e., city maps or urban plans; Burian, Šťávová 2009), although their importance remains high in many common tasks, necessitating a high-level understanding of terrain (Putto et al. 2014). In the 1970s, experiments with terrain visualization methods concentrated on legibility studies of different methods of representing topographic information (Chang, Antes, Lenzen 1985). Phillips, Lucia, and Skelton (1975) performed a questionnaire study testing four different types of relief maps (contours, contours with hill shading, layer tints, and spot height maps). In most of the questions, statistically significant differences were found, but no single map type was the best for all 13 map reading questions. Visualization using contour lines with hill shading was an advantage only in questions requiring visualization of the landscape (e.g., visibility, finding the steepest slope). -is result corroborates the study of DeLucia (1972), who found a significant increase in time necessary when extracting required information from a map with hill shading. Similar visualization techniques were investigated in the study of Potash, Farrell, and Jeffrey (1978), who analyzed contour maps and contour maps supplemented by layer tints and shading. -e results of the study showed that layer tints increased reading speed, whereas shading did not and caused a decrease in accuracy. In the study of Castner and Wheate (1979), contour maps and shaded relief were analyzed. -e results showed that in tasks where a search target was associated with a topographic situation, shaded relief was an advantage. 356 geografie 123/3 (2018) / s. popelka In a more recent study, Petrovič and Mašera (2004) created a questionnaire to find out how different 3D cartographic representations of terrain could fulfil a user’s needs. Savage, Wiebe, and Devine (2004) compared performance between 2D and 3D topographic representations in solving different tasks. -e authors learned that “there was no apparent advantage in 3D maps for tasks requiring elevation information, nor was there a disadvantage for integrated tasks which did not require elevation information”. Schobesberger and Patterson (2007) conducted a study comparing conventional (2D) and perspective (3D) trail maps of an outdoor area of the Zion National Park in Utah. Respondents generally agreed that 3D maps depict reality better than conventional maps. Wilkening and Fabrikant (2011) investigated how varying time constraints and different map types influence people’s visuospatial decision making. According to the scheme introduced by Rohrer (2014), evaluation methods can be divided into behavioral (objective) and attitudinal (subjective) methods. In many previous studies, methods such as questionnaires or interviews have been used. -ese methods are attitudinal methods because they show “what people say”. By contrast, eye-tracking can be considered a behavioral (objective) method, because it shows “what people do”. Without an objective evaluation method, it is not possible to reveal the true efficiency of shading in tourist maps. -e advantage of an eye-tracking study over a questionnaire is that not only can a respondent’s answers in solving a task can be analyzed, but also their strategy. Eye-tracking has been used in the field of cognitive cartography since the 1970s, but work with devices of the time was time consuming and expensive. With cheaper technology, eye-tracking has become more widely used in cartographic studies (Brychtová et al. 2013; Coltekin et al. 2014; Fuhrmann, Komogortsev, Tamir 2009; Kubíček et al. 2017a; Kubíček et al. 2017b; Popelka, Brychtová 2013). One of the first eye-tracking studies dealing with terrain visualization was performed by Chang, Antes, Lenzen (1985), who analyzed the effect of experience on reading topographic relief maps. Putto et al. (2014) performed an eye-tracking study evaluating three different terrain visualizations (contour lines, relief shading, and oblique view) in solving three types of spatial tasks (visual search, area selection, and route planning). -e results showed that performance on contour line visualization and shaded relief were comparable (oblique view was the slowest). Popelka and Brychtová (2013) investigated the differences in reading 2D maps with contour lines and p