Site Map - Home : Issues and Expectations : Five-D GIS
The simple definition of a Five-D spatial reference system is as follows:
- the first three D’s refer to the Cartesian coordinates used to define physical space
- the fourth D refers to time when some content in the 3D space changes with time
- the fifth D refers to attributes associated with the space or time defined by the first four D’s
Prior to the advent of computers most 3D objects were represented by two or more 2D drawings. These drawings, which were often orthographic projections, or normal views, of a 3D object, referenced two of the three Cartesian coordinates (which two depending on the view) and were frequently made or printed on paper (a 2D space) and were thus often referred to as drawings in paper space.
AutoCAD, the popular CAD program by Autodesk, which was originally developed as a 2D program, now refers to model space (3D) and paper space (2D) … using paper space for recording the actual 2D working drawings used by the various trades responsible for constructing the building or structure described by the drawings.
Pseudo 3D GIS
It is both interesting and important to note that much of today’s GIS technology is predicated on the use of a two-dimensional spatial reference system (similar to paper space) for representing, analyzing and visualizing geographic information. This is due to the fact that when GIS got started in the early ‘70s geography was a 2D subject. In fact, even today, when most people think of geography they typically think in terms of a 2D map … usually a 2D paper map.
The advent of 3D in GIS, or what was then call 3D, first emerged as a relief map, where for every x-y in planar space there was a single z value in vertical space. Triangulated irregular networks (TINs) were used to represent these “3D” surfaces. TINs were later used to represent buildings by incorporating the exterior form of the building as part of the surface. The problem was that TINs could not represent a vertical surface or anything with multiple z’s on a single x-y. Nor could they be used to represent any surface past vertical, such as an overhang.
GIS developers also explored the ability to assign 3D symbols to points on the surface. The 3D symbols could symbolize anything, such as a particular type of building (house, school, fire station, etc.) or a particular type of tree (conifer, deciduous, etc.). The use of 3D symbols gave users the resulting ability to create crude visual representations of geographic environments.
The next step in the evolution of 3D GIS was to add the ability to vertically extrude polygons in planar space, such as those representing jurisdictional boundaries or building footprints, and, in so doing, providing the user with a simple, yet relatively effective way, to display 3D information, such as population by county or building heights.
This also led to the ability to position layers in vertical space by assigning separate z-values to each layer in a stack of layers. This was sometimes used to show how thematic layers could be combined (over-laid) to produce a composite layer. It was also used to represent the floors, or floor plans, in multi-story buildings giving the user the ability view GIS enabled floor plans in 3D space and to do 3D suitability analysis.
The ability to extrude 2D building footprints to create simple renditions of 3D buildings, coupled with the ability to show floor plans in 3D space, led to the desire to display the exterior geometry of buildings. This, in turn, led to the ability to import 3D models created in other programs, such as 3D Studio, Form Z, SketchUp and others. Since these programs had the ability to create any type of 3D object (buildings, vehicles, streets, trees, etc. and even people) users were able to use GIS to display fully rendered, or nearly fully rendered, environments.
While most of the 3D editing tools (programs) have ability to create geometry defining both the interior and exterior of buildings, most users have focused on creating models of the building’s exterior. The structures and activity spaces inside a building are probably more important and of greater interest to GIS users, particularly those interested in doing spatial analysis (e.g. 3D suitability and vulnerability analysis, evacuation studies, and disaster simulations).
Most of today’s spatial technologies (GIS, CAD, BIM, and 3D editors) are now interoperable, at least to some degree, giving the user a relatively rich suite of tools for representing, analyzing and rendering both rural and urban environments, with the focus moving more toward the urban than the rural. This is particularly true when one begins to think of the interoperability potential GIS and the modern BIM programs, such as Archicad by Graphisoft (popular in Europe), and Revit by Autodesk (popular in the United States).
Real 3D GIS
A problem remains, however, GIS is still predicated on 2D and 2.5D representations of geographic space. For GIS, geographic space is still a surface, a very rich surface, but a surface non-the-less. As such, GIS has, for the most part, limited its application to terrestrial conditions, neglecting those occurring below the surface of the earth, as well as those occurring above the surface, all of which are best modeled in 3D space.
While CAD has migrated from 2D (the original AutoCAD) to 3D (Revit) GIS has remained tethered to its view of the world as a surface. The world, however, is more than a surface. The world, or that part of it important to life, includes everything below, on, and above the surface of the earth that supports life … what some are now calling the planet’s life zone.
GIS needs to be able to represent, analyze and visualize conditions in the planet’s life zone, including those below the surface of the earth (geology and sub-surface hydrology), on the surface of the earth (as it is presently doing), and conditions above the surface of the earth (pertaining to the earth’s atmosphere and oceans).
It also needs to support the full integration of 3D objects (buildings, structures, vegetation, etc.) created by other programs, including those created in CAD, the traditional 3D editors (e.g. 3D Studio) , and perhaps more importantly some of the newer 3D modeling tools, such as Rhinoceros (coupled with Grasshopper) by Robert McNeel and Associates, and CityEngine by Esri.
This is not to diminish the value of GIS but to only say it needs to broaden its definition of geographic space and to extend its capabilities into true 3D space. As we will soon see, it will also need to expand its capabilities with respect to time and the creation and utilization of spatially dependent variables.
4D – Time
Time dependent data or temporal data (such as, data describing changing weather conditions) is supported by most GIS systems through the use of animation tools. Little or no support is provided for data analysis or for making projections.
Growth models (urban growth, policy impact models, the spread of a forest fire, etc.) are supported, at least in a limited way, through the use of 2D geoprocessing tools. Full support, including the utilization of feed-back and feed-forward functions, is left to the user using generic programming tools.
The simulation of movement systems (evacuation events, the kinetic behavior of transit systems, etc.) are supported programmatically, but without the tools normally associated with discrete event-based simulation tools.
Most GIS systems do not support the idea of clocks. That is, they do not provide the user with the time management tools for monitoring nested events, such as might be required to manage and monitor a multi-modal transit system or simulate the events associate with a major disaster.
Future GIS systems will need to support a full range of tools for managing temporal data and for modeling time-based movement systems.
5D – Spatial Variables
Spatial variables are scalar measurements pertaining to some aspect of one or more of the spatial entities referenced in a spatial domain. For example, a Land Use Map (a spatial domain) might be queried to yield the total area (a measurement) of all land allocated to commercial uses. In this case, the total area is a scalar value derived by summing up the individual areas associated with each of the individual polygons of the type commercial.
One can think of the area of an individual polygon as a source variable (one that is normally maintained by the spatial domain) and the total area for commercial as a derived variable (one that is calculated apart from, or outside of, the spatial domain).
Users can be interested in monitoring a wide range of spatial variables (including both source variables and derived variables), either as individual values or in comparison to other variables. For example, users might want to monitor the total area allocated to commercial as compared to the total area allocated to residential. Additionally, they might want to display this comparison graphically using some type of statistical widget, such as a pie chart.
Given a multi-layered, multi-functional, geographic information system it is not hard to imagine users wanting to make use of a wide array of spatial variables, with sub-sets of these variables arranged and displayed using various types of dashboards. Some of these dashboards might be designed to serve the needs of the team responsible for creating and maintaining the spatial domains (map layers), others designed to meet the needs of stakeholders, and still others designed to meet the information needs of the general public.
All of this points to the need for a dashboard authoring system. Most GIS systems provide their users with the ability to harvest source variables and even to use these source variables as the basis for calculating derived variables, at lease in a limited way. Few, however, provide their users with a well-developed dashboard authoring system.
While not seen as a dashboard authoring system, Microsoft’s Excel is a good example of just such a system. It provides the user with a rich set of mathematical functions that can be used to calculate just about any type of derivative value one might imagine. It provides the user with an extended suite of statistical widgets (pie charts, bar charts, status dials, etc.) for displaying the results of those calculations. Plus, it allows the user to layout those widgets, along with other content, to create dashboards fully customized to meet the needs of the user’s user community. Additionally, and this is very powerful, it gives the user the ability to specify conditional formatting. For example, when a variable reaches a user specified value it can trigger a change in how the various elements in the dashboard are rendered.
Spatial variables can also be used to render the geographic information displayed on a map. GIS maps, or map layers, can portray a realistic view of the displayed geography, as is the case with an air-photo, or a symbolic view, as is the case with a zoning map. Spatial variables can be used to condition the symbolic information displayed in on either type of map.
It should be clear by now that 5D really stands for 5+D, where the + sign represents the fact that there can be any number of spatial variables of interest to a user. These variables of interest are often called key performance indicators (KPIs) and often serve as the primary conveyors of geo-spatial information.
Great Minds and Small Thoughts
The idea of 5D GIS is just now beginning to emerge in the minds of GIS professionals and, as a consequence, the phrase is becoming a part of normal GIS conversations. The implication is that today’s GIS supports 5D. The truth is, it supports parts of 5D.
Today’s systems support two and a half of the first three Ds, (terrestrial surfaces). They provide the illusion of full support for the first three Ds via the ability to import 3D models created in other programs. The forth D (time) is supported programmatically through the use of user generated programs and scripts, and the fifth D is partially supported by providing limited access to spatial variables and the subsequent display of those variables.
Today’s GIS systems do not support the representation of 3D fields, such as would be required to model sub-surface conditions, our atmosphere, or the earth’s oceans. While they nominally support the representation of time dependent data they lack the ability to create and monitor time-based simulations. They also lack the ability to fully support the creation and display of key performance indicators, referring here mostly to the absence of a dashboard authoring system.
One of the problems lies in the fact that development agendas typically focus on the incremental extension of existing capabilities as opposed to the development of fully capable systems. The development of GIS has proved to be of no exception to this rule. GIS technology has been, almost since its inception, predicated on a two-dimensional view of the world. The various data models, geoprocessing tools, and user interface environments have all been predicated on this two-dimensional view of the world.
The thought of developing a fully capable truly-5D GIS system would face a mountain of impedance. New conceptual models, data structures, and analytical procedures would need to be developed, as well as new visualization strategies and tools for navigating to and viewing data imbedded in 3D fields. Strategies for connecting 2D and 2.5D surfaces to 3D fields and 3D fields to surfaces would need to be developed, as well as procedures for modeling movement and time dependent data in 3D space, not to mention the number of user interface (UI) and user experience (UX) strategies that would need to be established.
Similar conditions were encountered when Autodesk faced the idea of moving from CAD (2D drawing tools) to BIM (3D modeling tools).Basically, they had to start over and re-invent the technology relevant to the needs of their users. Such would be the case for the future developers of GIS. They would have to give up their small thoughts, however cleaver, and give their attention to the big idea … the idea of developing a fully capable truly-5D GIS.
They might even have to give up the idea of a Geographic (2D) Information System (GIS) and embrace the idea of a Spatial (5D) Information System (SIS). This mental/emotional shift would, in and of itself, represent a move not without significant impedance.
Need for 2D, 3D, 4D, and 5D GIS
Regardless of the moniker, GIS or SIS or something else, the need for a fully functional 5D Geo-Spatial Information System exists. In general, the system should include the following capabilities:
- 3D – The ability to represent, analyze and visualize the four principle domains of 3D geographic space (terrestrial surfaces, sub-surface geology and hydrology, the earth’s atmosphere, and oceans).
- 3D – The ability to incorporate, analyze and display 3D objects in that space.
- 4D – The ability to incorporate time dependent data and the simulation of movement systems.
- 5D – The ability to monitor spatial variables coupled with a dashboard authoring system.
Additionally, the system should be both conceptually and architecturally well organized and relatively easy to use.
Principal Authors: Bill Miller