The National Spatial Data Infrastructure: A Societal Information System
Presented November 8, 2000 by Paul Cote as a guest lecture for
ENR-100.
Professor Clark has asked me to prepare a lecture, some readings, and
laboratory exercise for you dealing with the subject of GIS and Remote
Sensing as they relate to biodiversity. I will admit from the start
that I am not an expert on biodiversity; but I have done a lot of
thinking about GIS and a little about remote sensing. I know that
investigations of biodiversity, and analysis of policy impacts upon it
are very information intensive enterprises, and many if not most
of the important questions related to biodiversity have an important
spatial component. Therefore, if we are interested in understanding
biodiversity and how our actions may affect it, we really need to
have an understanding of the information systems that permit us
to look at and model these complicated spatial relationships.
Information policy is a very important piece of environmental policy.
- The way institutional leaders plan and share their information resources
is a very critical dimension of environmental policy
now and in the future.
- The Federal Government's Spatial Data Infrastructure is a good
example of a concerted attempt to create a societal information
system in which the aggregate is far more valuable than the sum of the
individual parts.
- Leaders in administration and policy research are irresponsible if
they ignore the growing information infrastructure; both in terms
of their use of it, and their potential contributions to it.
- Geographic information systems are an important tool for integrating
pieces of the growing information infrastructure. While it is not necessarry
for each and every policy maker to understand the details of GIS
software, there is a certain level of understanding required in
order to responsibly interpret the results of GIS analysis, and
to assure that the information resources management in our institutions
makes appropriate accomodation for GIS.
In my class on GIS, I find it useful to demystify information
systems by beginning with the fundamentals: The essential
functions of an information system consist in systematically storing
and retrieving representations, and modeling systematic
associations
among these elements. One important corollary of this definition
is that the effective differences among various information systems
can be boiled down to the following questions:
- What is the nature of the representations that may be stored?
- What is the nature of the associations that may be made?
Given this theoretical framework for evaluating information systems,
lets look at the common information systems that are important to us,
and how they function to help people expand our mental models of
the world:
- The World Wide Web permits creation of documents including text,
images, audio and video; and facilitates associations among
documents created by others anywhwere on the internet. Web browsers
permit others to examine any of the documents and their imbedded
associations.
- Relational Database Management Systems -- systems
of tables such as those containing inventories of Trees, Fixtures
and Parks, not only facilitate documentation of inventories, with
queries and reports. properly designed RDBMS may also be used for
making useful associations and analyses.
- Vector-based
Geographic Information Systems are an extension of RDBMS.
Vector GIS permit the storage of geometric information with
relational tables. These permit users to examine spatial
associations among the things that are represented.
In addition
to the typical relational associative operators:
- Equal-To (or Not)
- Less Than
- Greater Than
Vector GIS permit us to associate entities based on spatial
associations:
- Occupying the Same Location, Intersecting, Containing, or Is
Within.
- Is Near-by (or not)
example (by Amy Cupples Rubiano)
- Raster-based Geographic Information Systems facilitate the storage
of information about discrete locations. With a Raster GIS users
can very systematically model relationships and associations among
locations accross and among many types of surfaces (e.g. land cover,
elevation, soils.) Typical applications of Raster GIS involve
derivation and recombination of various information surfaces in a
process called 'Cartographic Modelling." Raster GIS methods are
also useful in the digesting of remotely-sensed images (such as
those collected from space) into land cover maps.
The fundamental capacities of information systems when reduced to
their elements, as above, appear symplistic. Information systems have
many intersting properties when complicated sequences of data input
and processing are combined. The web is an interesting case of an
information system that promotes societal intellegence -- as the
structure of the web itself is not designed, but the individual
compnents are dynamically accessed and associated in such a way that
new information is generated by the sum of the contributions. In the
next few topics we will look how GIS is beginning to behave in a
similar way.
From Project-Based GIS to Information Infrastructure
An
interesting evolution is occurring in information systems.
In the early stages of information systems and GIS, most
databases were single-purpose. More recently, organizations that have
learned to manage information are planning I.S. strategies to leverage
information resources in virtual Data Warehouses that permit
people to collect and inter-associate information collected for
different purposes within the organization. (Dangermond, 1999)
Such information infrastructure-building efforts require strategic
information-system planning at the very top of the organization.
From the perspective of GIS, one such strategic information
infrastructure plan has been undertaken by the Federal Government of
the United States. The
National Spatial Data Infrastructure was mandated by an executive
order of President Clinton.
The National Spatial Data Infrastructure (NSDI) recognizes that:
- Left to their own resources, government agencies will create
costly, redundant information resources.
- Standardization of
digital base-map material can help agencies share information and
coordinate policies.
- Sharing standardized information
resources with the public creates societal benefits:
- For more efficient spatial planning in the public, and
commercial sectors at many different levels.
- Overall, more
efficient use of resources (e.g. through fuel savings or better
logging practices)
- Better accountability and public
participation in policy making and analysis.
These justifications make sense merely at face value; but in a short
time, we will see that freely available government information is
creating elemental building blocks for a much more intellegent
society.
A small sampling of federal data components of the NSDI available for
free or at no charge with minimal or no copyright restrictions:
...and many, many
more
At the state level (in Massachusetts, and most other states) similar
strategic data sgharing policies are in effect. Some examples of data
available to the public over the web, at no cost, include:
...and many,
many more
At the local city level, administrators have not been quite as quick
to recognize the benefits of making data free to the public or easy to
share. These information resources are some of the most useful for
planning and understanding environents, however, the public Boston and
Cambridge public agencies (for example) attempt to fallaciosly
"recover costs" of database production by charging well over the costs
of reproduction for these databases.
Property Parcel Data from the Tax
Assessor | 
Engineering-Scale information
regarding Buildings, Streets and Terrain |
Societal GIS
As far as I know, the term societal GIS was
coined by Jack Dangermond (GSD MLA, 1969.) It is interesting to think
of government officials as stewards of a physical public
infrastructure as they are increasingly creators and stewards of an
infrastructure of representations -- an information infrastructure.
Every application that begins with government data sources, builds on
this infrastructure. The spatial data infrastructure provides a basis
for development, just as the physical infrastructure does.
Consider the
following example (by Amy Cupples Rubiano, a former student in
GSD6322, Fundamentals of GIS) which takes advantage of several
independently published data sources:
- A person starts with a listing of events referenced with street
addresses. (in this case it is the EPA's Toxic Releases
Inventory.)
- Using a vector GIS, the person associates the
Toxic Release addresses with street locations from US Census TIGER
Line files.
- Another GIS association defines the areas within
half a mile from the Toxic Release locations.
- Information
about how many people live in the area is refined with land use
data from the MassGIS.
- The refined population map is then
used to estimate the population and ethnic composition in the
areas near toxic relaeases.
Phenomena that involve individuals and groups creating, by their
collective actions, systems that are more than collections of
individual parts, can be termed societal phenomena. As time goes on,
datasets become more detailed, and more useful. At the same time,
people are sharing data more, and using one data layer as a base to
make yet another useful database. Understanding this, it is easy to
predict that whereas GIS is currently a very important means of
storing, studying and sharing spatial information, that its use and
usefulness is at the beginning of a very steep logarithmic ascent.
Growing Pains on the GIS Frontier
It should be clear from
the arguments made above, that for people interested in the
arrangement of things in space, that GIS and information
infrastructure are important things to understand. I believe that it
is useful to discuss three levels of GIS-Literacy:
- Formal Capacities of Information Systems:
In order to plan
or think critically about GIS applications, one should have the most
general knowledge about how "Ideal" information systems behave.
These formal capacities can be understood without regard to specific
data or software or any specific application:
- What are the capacities of a particular information system to
store and associate representations of things?
- How can
the capacities of a particular I.S. be applied to represent
entities and relationships that are of interest to us?
- Practical Problems of Implementation:
These are the issues
that come up as soon as one attempts to find or create real data to
build information models.
- How are databases designed such that formal queries and
associations return valid results?
- What are the most
appropriate data sources, how can one assess relative quality of
geographic data sources?
- What sorts of spatial coordinate
systems is one likely to encounter? Which one is most appropriate
for a particular application in a given geographic area.
- What
sort of problems is one likely to encounter in data collection,
conversion, and assimilation of various datasets into a useful
assemblage of information resources for a model?
- Technical Problems of Implementation:
These are often
software-specific issues of how-to-do-it:
- How do the buttons and menus of this application represent the
formal capacities of the information system?
- How should
information resources be organized for an effective information
system implementation?
- How can I write a program or
customize the software to facilitate a particular process?
- Critical understanding of I.S. Models:
No representation is
complete. In an assemblage of representations, many practical
and technical choices are made, each one has an effect on the
quality of the result. Formal, practical and technical
understanding of GIS are prerequisites for evaluating:
- Is there an appropriate fit between the formal capacities of
information systems employed and the types of entities and
relationships that are being modeled?
- Are the
databases designed in such a way that queries and
associations employed are returning valid results?
- What sort of errors should one expect from a given
assemblage of imperfect information resources?
- How
would evaluate whether an GIS representation is "good
enough" for a given purpose?
- How could a given GIS
technical implementation be improved?
- Understanding of the Institutional Concerns related to Information Infrastructure
As leaders, you may choose to distance yourselves from "nerdy" issues such as
database administration. I hope that you now have a sense of the importance
of institutional information and GIS, and that you will recognize the importance
of proper planning, stewardship and sharing of institutional information. As
Bill Clinton demonstrated with his Executive Order, these issues often require mandates
from the highest levels.
