Geography Reference
In-Depth Information
Existing but Inappropriate Data
While it is true that considerable data of the types important to this discussion have been collected,
many of them are not directly usable in a spatial database. This occurs principally because these datasets,
collected by groups or agencies with specific missions to serve, have been assembled in nonuniform
categories or have been interpreted in a specific manner for a particular purpose. For example, early
soils data categories may not contain the necessary information that will enable measurement of some
environmental effects of land use activities.
Effort Required for Development
The data in our base won't develop as a natural consequence of some already ongoing process. Other
database developers are more fortunate. As a clerk processes applications for auto license tags, he or
she may type the pertinent information about the car, owner, and tag directly into a database. Thus, the
database develops as a result of the tag-selling process that must occur anyway. Spatial databases about
the environment have not evolved as consequences of other processes; the work starts almost from
scratch in most cases.
The Changing Environment
One cannot get an entire spatial database of any size developed before part of it is incorrect because
some of the values in the real world will have changed over time. Land use is an example of a variable
whose data values are changing in many places on a daily basis. Houses are built. Roads are paved.
Even such stable phenomena as topography change drastically over time. For example, the Mississippi
River was about 1300 miles long when LaSalle floated down in his canoe. When Mark Twain wrote
about it 200 years later, it was less than 1000 miles in length. Not only that, very little of what was
wet in LaSalle's day was still river in Twain's. And to further illustrate the futility of any attempt at a
“permanent” spatial database, Ole Man River has moved at least two towns from one state to another
by its meanderings. The moral is that some data values of all variables in a spatial database are going
to change over time. Some procedure for updating the base must be developed or the value of the base
will be degraded by time. Further, different variables are of different value to the analysis and decision-
making process and, of course, change at different rates. In some cases, the efficient thing to do is to note
changes as they occur; in other cases, replacement of all data related to a particular variable is in order.
Either way, there are difficulties and costs.
Multiple Paradigms for Storing Geographic Data
Ingenious ways of taking the continuous, virtually infinite environment and storing its important
features in a discrete computer have been developed. Spatial data stored in one scheme are not easily
converted to another, and one almost always loses information in such a transfer.
For example, take the matter of representing the elevations above sea level of a geographic area. To
begin with, one is dealing, in theory, with an infinite number of values. If you establish the elevation
of a given point, then, depending on the precision of your measurements, a nearby point, say 1 meter
to the north, will have a different elevation. Elevations of an area may be thought of as a continuous
surface, potentially different at every latitude and longitude position. Unless we can model this surface
with a mathematical equation (and, with the average mountain or cow pasture, this is usually out of the
question), we are stuck with having to select a set of specific points, determine their elevation, and make
assumptions about the elevations between those points.
Search WWH ::
Custom Search