|
FREC 480 -- GIS in Natural Resource Management A geographic information system (GIS) stores, manipulates and analyzes heterogeneous data elements representing geographic features. These data elements are referenced in space and (optionally) time or other dimensions representing attributes of interest. The core of a GIS is computer software that handles these data in digital form. But any GIS also involves hardware (computers) to run the software, store data, print maps, etc., as well wetware (people) to design and execute the analyses. Course Objectives This course provides hands-on training in the the use of GIS for natural resource management. At the end of this course you should have a working familiarity with GIS data structures, data sources, spatial logic and basic vector and raster functions. The course projects are intended to develop understanding of how GIS analyses can be applied real-world resource management problems. This course will use ESRI's ArcGIS software (including raster, 3D and other functions), but the course will reference other GIS software and is not intended to be software-specific. The GIS skills you develop in this class should be readily transferable to other problems and other GIS packages. Overview of the course work environment The graded coursework is a series of hands-on GIS projects which are to be completed, submitted as web pages rather than hardcopy, and presented to the class. In the process of doing your GIS projects, you will generate various maps and other graphics, saved as digital images, as well as various statistical results. This course will also teach you how to write the HTML and edit your graphics so that you can communicate your results via the Web very efficiently. Once you have completed this course, you should have an on-line portfolio of your GIS work that you would be proud to show any prospective employer. HTML
stands for HyperText Markup Language, the plain-text format tags used
to write Web
pages. I will insist that you write your first few web pages in
raw HTML with a plain text editor like Notepad before you become
dependent on a higher-end HTML editor; these editors generally don't
give you exactly what you want
unless you know how to fine-tune the raw code. This is a lot to learn in a single term. We will introduce all of
these
elements in the first few class sessions, and you may feel overwhelmed
at first. Don't panic. It takes two or three weeks to get comfortable
with
this work environment. Raster vs. vector data zone: 18
The first six lines specify the "header" information, identifying the north, south, east and west geographic boundaries of the map. In this case these are UTM (NAD 1983, Zone 18N) coordinates expressed in meters. Taking the difference of east and west edges, the map spans exactly 600 meters in 20 columns of "pixels;" The north-south span is 810 meters in 27 rows. Consequenly, the dimensions of each pixel are 30 x 30 meters. The rest of the file is simply an array of pixel values--in this case, elevation values in feet for each pixel. We are really modeling a 3D surface here, and these are just the Z-values; the GIS software keeps track of the X and Y coordinates based on the header information. If we assign a palette specifying a display color for each pixel value, we can display the elevation map (left panel, below). Here, the palette progression or "color ramp" for lowest-to-highest elevation values, is yellow-green-cyan-blue-magenta-red. Note the obvious pixel artifacts in this map. This is a very small file, containing only 540 Z-values. If we had sampled elevations more densely, say at 5 x 5 meter intervals, the pixels would be barely noticeable, although we would then have 36 times as many Z-values.
The right panel in the image above represents a vector map of 10-foot-interval elevation contours for the same part of Chestnut Hill. Vector GIS derived from CAD (computer-aided drafting) technologies, which manipulate points, lines and polylines/polygons referenced in two or three dimensions. These elevation contours are piecewise-linear arcs, defined by a series of XY coordinates.. The elevations between contours aren't specified, but can be interpolated. In this resolution the contour lines look pretty smooth, because the shape of each contour is defined by a reasonably large number of points, so the resulting file is pretty large. Here is the beginning of the (big!) vector datafile in one of many possible ASCII formats: ZONE: 18
Each arc (contour) is specified by an identifier line
followed
by a string of X-Y coordinates. Each identifier line
specifies
that the feature is a line (as opposed to a point or area), specifies
the
number of verticies that follow, and specifies the elevation value for
the
contour. As with the raster example, the X-Y coordinates are
specified in meters based
on the same projection. Geographic features can be classified as points, lines or regions. Point features might include well locations, houses, or bird
nesting sites--any feature small enough that its actual dimensions
(relative to the map scale) don't matter. A vector GIS represents
a point feature as a single (dimensionless)
coordinate pair. A raster GIS would represent a point feature as a
single
grid cell. . On
a map of the entire US you could represent entire cities as points,
although on a map of a housing development you might represent the
individual house footprints as polygons. Line features might include roads or streams or elevation
contours. A vector GIS represents
each line feature as a series of geographic coordinates. A raster
GIS represents
it as a continuous string of grid cells connected at their edges or
corners. When line features such as streams and roads form
networks, the GIS has to keep track of the connectivities between them:
one feature's ending node must be identical to a connecting feature's
starting node. A line feature that closes on itself (its end node
is identical to its own starting node) is termed a "polyline."
Technically a polyline defines a perimeter only, and has length but
no area. (Some GIS's don't bother distinguishing between
polylines and polygons.) Region or Polygon features might include wetland areas or
parks or Census tracts.
A vector GIS represents a region or polygon feature as a closed string
of coordinate pairs defining the region by its perimeter. A raster GIS
represents regions as clumps of grid cells. A region feature has both
length
(perimeter) and area. Note that region features can be complex: a
single
region feature like the state of Hawaii includes multiple islands which
are
separate pieces
of the same feature. Other features have interior "lakes" that
are
not considered part of them, like the Vatican and San Marino inside
Italy. Topology (consistencies in feature
connectivity) In the case of the vector elevation contour lines above, the contours don't intersect, and each line is specified separately. But things get more complicated when vector features do connect or intersect. The usual strategy is to model such features as multiple connected segments. For example, a stream network has sequential junctions where tributaries merge into higher-order streams, so we would insure that end nodes of the two tributary segments match to the start node of the receiving segment. A city avenue is a sequence of segments between each pair of crossing streets.
Today's GIS's are built on specialized database engines that manage these topology issues efficiently. You should be aware of them, but they shouldn't create problems for you in this course. |
A segmentation
strategy is also used for insuring boundary consistency between
contiguous polygon featurs. If you simply
digitize adjacent regions as complete polygons, you are digitizing
each shared border twice, which can generate inconsistencies, i.e.,
gaps
assigned
to neither region and/or overlap slivers assigned to both
regions. A better strategy is to construct edge-consistent
polygons piecewise from unique arcs representing boundary
segments.