Geography Reference
In-Depth Information
Goodchild, 2000; Lucieer and Kraak, 2004; Deitrick and Edsall, 2006; Devillers and Jean-
soulin, 2006; Goovaerts, 2006), current GIS technology and practice have done little to
provide support for incorporating information about the quality of spatial data. Goodchild
(2006) calls for the GIScience community to insist that GIS incorporate scientific standards
and principles, including the reporting of results with a precision that does not exceed their
accuracy or the precision of their source data. He notes a mismatch between responsible
GIScience theory and practical GIS use concerning uncertainty. This mismatch, he writes,
may stem from the comfort that GIScientists have with the necessary compromise between
the conflicting objectives of representing the world with both accuracy
and
visual or com-
putational simplicity, generalization and clarity. Users, on the other hand, are less likely to
understand that GIS data and analyses are not necessarily accurate and reliable. In addition,
many uncertainty models and concepts are based on difficult mathematics. The complexity
of uncertainty added to the conflicting goals and differential training between GIScientists
and the diverse community of users introduces the potential for misinterpretation of these
concepts and the relationships they are meant to explain (Goodchild, 2006).
Geographic uncertainty in GIScience - and the importance of its communication - thus
arises from a mismatch between the needs and goals of the user and those of the producer
of geographic information. These needs include the match between the specifics of a user's
request for data or information, the data format, the use of that data, and the phenomenon
being investigated (Gottsegen, Montello and Goodchild, 1999).
Several recent efforts in cartography, visualization and GIScience research have sought to
bridge this gap through the integration of uncertainty information into geographic represen-
tations. Researchers have sought the most appropriate and effective means of representing
uncertainty to map readers, carrying out experiments comparing representational tech-
niques. A common approach is to begin with the adaptation of Bertin's (1983) visual vari-
ables for the representation of uncertainty. In addition to these variables, additional graphic
variables, such as transparency, saturation and clarity have also been proposed (MacEachren,
1992; Slocum
et al
., 2004) specifically for uncertainty representation. Gershon (1998) pro-
posed two general categories of representation strategy: intrinsic and extrinsic.
Intrinsic
techniques integrate uncertainty in the display by varying an existing object's appearance
to show associated uncertainty, while
extrinsic
techniques rely on the addition of geometric
objects to highlight uncertain information. Cliburn
et al
. (2002) utilized this distinction in
an interactive environment which depicted the results of a water balance model along with
the associated uncertainty of these results (Figure 14.1). Interactive computer environments
have opened new possibilities for uncertainty representation, including the development of
interfaces that allow users to manipulate the display of uncertainty by deciding how and
when to display uncertainty information (Fisher, 1994; Davis and Keller, 1997; Ehlschlaeger,
Shortridge and Goodchild, 1997; Cliburn et al., 2002; Aerts, Clarke and Keuper, 2003).
In most studies of uncertainty visualization in GIScience, when the importance of po-
tential differences in users has been acknowledged, it is often included as an ancillary study,
and not as the explicit and main focus of the study. The primary focus of most experiments
has been on representational design. MacEachren, Brewer and Pickle (1998) developed and
tested a pair of intrinsic methods for depicting 'reliability' of data on choropleth maps used
in epidemiology. Newman and Lee (2004) evaluated both extrinsic and intrinsic techniques
for the visualization of uncertainty in volumetric data comparing glyph-based techniques,
such as cylinders and cones, with colour-mapping and transparency adjustments. Lucieer
and Kraak (2004) developed an interactive environment for exploring the uncertainty of
Search WWH ::
Custom Search