Print This Post

What is an SDI?


The most often quoted definition of a Spatial Data Infrastructure (see is that it is “a framework of spatial data, metadata, users and tools that are interactively connected in order to use spatial data in an efficient and flexible way. Another definition is the technology, policies, standards, human resources, and related activities necessary to acquire, process, distribute, use, maintain, and preserve spatial data “[1].

The difficulty with this definition is that it provides little insight as to what the software component of the infrastructure might amount to, or what the benefits and objectives of the infrastructure might be. 

If one looks at most SDI’s today they heavily emphasize a librarian view of the software infrastructure part (metadata collection and discovery are the primary points of focus), and as a result are inappropriate to support many of the use cases for which such an infrastructure might be deployed.

Why an SDI?

To address “what is an SDI” we need to think more on the function – the “why” of an SDI. Only then can we think more correctly about what such a thing might entail.  Without being to grandiose, I think there are a number of reasons why government and the private sector might think to invest and build an SDI.

These include:

  • It enables a local, regional, national or even a planetary accounting system on the state of the environment, by making visible the distribution and temporal variation in key environmental parameters from carbon emissions to deforestation.
  • It enables the local, regional, national or even planetary response to emergency events whether caused by people or natural forces.
  • It enables the more efficient interaction of corporations, government and citizens in the collaborative (and competitive) development of the built environment, from urban construction to mining, agriculture and energy production.   
  • It provides a vehicle by which government can meaningfully interact with its citizens at all levels on planned projects, programs and policies; enabling both the presentation of what is planned, as well as enabling direct feedback from the citizenry.

For many of you, much of this will sound like e-Government, and in fact one can simply look at SDI as no more than spatially enabling or geographically enabling e-Government.  I prefer to think of it as e-Society, since while Government has a key role, so does the private sector and the agents of civil society.

What do we build?

With the objectives as above, what does that imply we should be building?   Are the national and pan-national initiatives to build SDI such as the Canadian CGDI, the European INSPIRE, or the US NSDI being designed to address these objectives?  Furthermore, with the rise of big search engines as players in the geo-space, what aspects should government and the private sector focus on in developing an SDI?

If we look at the “accounting system” view of SDI, we see that the key thing is to make the state of the environment locally, regionally, nationally, and globally, visible to all of us.  This means it applies to a very wide range of data types over a wide range of scales. Most of this information is in some way spatially related, but this spatial relationship may be very implicit and not part of the actual data itself.  So SDI must be concerned with access to all kinds of data that are only nominally spatial.  Since much of this sort of data is to be found in government and private sector databases, an SDI must deal with data access for a wide range of databases and not just the nominally spatial ones.  Furthermore, if we are to do the sort of “roll up” of information that is required in any sort of accounting system, we need access to the actual data, and not portrayals or presentations. Visualizations are important of course, but only if driven directly from the data, and only if available at various levels of the society and government.  Finally, to make any sort of “accounting system” feasible, we will need appropriate models of the parameters that we are to report and measure – meaning we need public, shared schemas that deal with everything from forest cover, to floods, carbon emissions and water pollutants.  This is a major undertaking, but so was the construction of a national highway system, a national broadcasting system or a national air transportation system.

If we look at the SDI as supporting the emergency response, we encounter similar issues, but with the additional requirements of tighter information security, and temporal responsiveness.  Even more than in the case of the SDI as “accounting system”, we must see SDI as enabling the pushing of data change events to data consumers.  A forest fire manager will of course want to request information, but that is of little use if first responders on the ground or in the air cannot supply new information “as it happens”. An SDI focused primarily on metadata and discovery cannot hope to support such activities.

SDI, as supporting emergency response, also entails the interaction of many levels of government and the private sector, including both national and international agencies.  Recent events in Myranmar and Indonesia highlighted the need for quick and effective integration of information.  This implies that information must be sourced in real time from a wide variety of locations, at a wide variety of scales, from citizens as well as professionals, and integrated on as needs basis.   Again this cries out for common data models that are shareable across regions, provinces/states and nations.  Some regions such as in the European INSPIRE are making encouraging moves in this direction.  After all the parameters of a forest fire, flood or earthquake of interest to responders can readily be generalized over regions or nations. This is again not a small problem, but with costs for natural disasters in the hundreds of billions, and yearly losses in human life exceeding 100,000 people, the amount of money currently expended on SDI is still vanishingly small.

SDI, as supporting efficient collaboration of government, the private sector and civil society in the constructing the built environment, also calls for a re-thinking of traditional SDI initiatives.  Fortunately, we are starting to see movement in this direction through the emergence and gradual adoption of Building Information Models (BIM).  While nominally a “standard”, the key thing about BIM is a process that encourages early and continuous collaboration amongst all of the “actors” that contribute to the development of the built environment, including ordinary citizen, investors, construction companies, developers, owners, engineers (of many kinds), and architects.  While most of these efforts have been focused on building design, there is no reason to draw such an artificial barrier.  One can equally apply these ideas to the development of a highway, a mine, an airport or an entire city.  SDI, as collaboration, again implies dealing with a wide range of information scales, spatial and non-spatial information, the need to support information push as well as information pull, and the development of common models for buildings, transportation systems, and entire cities.  Big strides have been made in these areas by groups like the Smart Building Alliance, the German cityGML consortium and others.  We need to see these initiatives as simply a part of the broader notion of SDI, and to think of how we provide direct support for things like projects, collaboration “spaces” and the integration of design information (buildings, transportation infrastructure etc) and what we nominally call “geography”.  Governments and the private sector will spend close to $ 5 trillion dollars on construction in 2008.  Even ignoring the immense soft benefits of spending this money in more environmentally sustainable and human friendly fashion, any improvements in efficiency must be measured in the billions of dollars annually.

SDI, as supporting communication with and amongst our citizens in relation to the actions of government, also requires a re-think of what SDI means.  Many of the issues raised above are equally true for this use case.  The notion of collaboration cannot be stressed enough.  Government can use SDI to communicate new policies, new laws (e.g. changes in zoning or construction), and new projects to its citizens, and citizens can use that same SDI to make counterproposals, comments and suggestions.  Again only some of this information need be explicitly spatial, but the spatial component is often the key to understanding how this information relates to you as an individual.

Many of you may look at this discussion and realize that these objectives are being better met today by the likes of Google Earth, Facebook and wide area games like Second Life.  You may think that there is really no role, or only a diminishing role for government in the process, and I have had that view expressed to me by several members of national government agencies.  I think this is not the case.  Governments, the private sector and agencies of civil society have it their interest to work together to build information infrastructures that work for the betterment of the world.  This cannot be left to just information companies as their ultimate interests must inevitably lie elsewhere. At the same time, we must recognize the immense contribution of these companies, and build them and their technologies into the SDI process itself.

Spatial information Infrastructures can make an important contribution to the security of our nations and the environment.  We need to start thinking in the right direction.