Spatial Analysis Based on Raster Data Processing - Introducing Geographic Information Systems with ArcGIS - page 506

You would, of course, have to rethink the maximum and other values to form conclusions as to what was
ultimately desirable and what wasn't.
3. (Optional) Investigate creating a map in which the land use variable is twice as important as
the other variables. Or choose some other combination of rasters to produce a map that shows
varying degrees of suitability. Or go on to Exercise 8-4 where you will learn about making rasters
that represent surfaces that are based on values as specific points.
Exercise 8-4 (Demonstration)
Making Surfaces with IDW, Spline, Trend, Nearest Neighbor,
and Kriging
I will now introduce a very big subject with a very small exercise. The subject is basically this: Given a
finite set of points with known values, say, “z” values (e.g., altitudes, pressures, pollution levels), what
values could reasonably be assigned to the remaining (infinite number) of points that are in the same
area? That is, how would one interpolate between the known points?
This is an important area, because all we can really measure with a high degree of precision are
conditions at a point. Think about altitude, temperature . . . . We can be very sure what the situation is
at a given point (in space and time). We have to infer what conditions exist in the vicinity of that point. So
what follows is a quick and very cursory look at the tools in ArcGIS that let you do that. An entire course
could be built around the techniques these tools represent and the statistics involved. The Help files are
some help, but to make real use of these tools, you need to do a lot of reading, or engage a statistician.
(You have, of course, seen one approach to developing a continuous surface from a set of points: TIN.
What is different about these surfaces you are about to explore is that they have no sharp line breaks—
that is, they are differentiable, in the calculus use of the word—and they are represented by continuous
mathematical functions.)
1. Start ArcCatalog. In ___IGIS-Arc_ YourInitials make a folder named Experiment_with_
Interpolation. Make a folder connection. (This might also be a good time to clear up previous
folder connections that you have accumulated but no longer need.) Within this folder make a
file geodatabase named Surfaces.gdb.
2. From ___IGIS-Arc_ YourInitials \Trivial_GIS_Datasets, copy over, to the Experiment_with_
Interpolation folder, three shapefiles:
Known_Altitudes.shp
Known_Populations.shp
Only_points.shp
3. Start ArcMap with a Blank map and add as data Known_Altitudes.shp.
4. Consider the shapefile Known_Altitudes.shp: Make the map units and the display units Meters.
Look at the attribute table. Each number is the altitude in hundreds of feet of the particular