GPS and GIS Useful Tools in an Academic
Setting
Andrew Honaman, Mike Kunzmann, and Wolfgang Grunberg
ABSTRACT
To help facilitate the
integration of GPS and GIS technologies in an academic setting a course
was designed to expose students to the issues involved with landscape-level
decision-making in a digital environment.A
major emphasis of the course was to collect real world GPS data and incorporate
them into a validated GIS database.GPS
and GIS tools are useful in exposing students to mapping techniques, landscape
models, and cartographic procedures.To
illustrate the value of this course we will discuss an example of a student-based
project that incorporates many of the tools and techniques learned to produce
a decision surface, which could be used to validate the elevational gradient
of a standard football field.
INTRODUCTION
It is often customary
for students in GIS classes to be presented educational data sets, tailor
made by the faculty to aid in the conceptual content of the course.
This
method is valid under some circumstances, but does not provide any educational
training with regard to the collection, creation and presentation of real
world data required by student researchers.
To
provide a practical hands-on learning environment to students a GPS/GIS
class was created at The University of
Arizona
where students were offered the opportunity to design a problem where original
field GPS data would need to be colleted from which a final ArcView project
would be created.
The course has
been offered four times.
Each year
four or eight projects are produced.
Examples
of student projects range from monitoring fire burn patterns in the Catalina
Mountains to mapping the locations of. private wells in Arivaca, Arizona.
The
majority of the discussion to follow will next focus on the steps and methods
taken by one project group.
The objective
of the group was to evaluate GPS data using a standard shape and easily
identifiable image with known dimensions and surface geometry.
METHODS
Determination of field
site:
Most of the work in GIS
using GPS field data is concentrated on the x (longitudinal) and y (latitudinal)
location of objects and their related accuracy.
In
this project attention was turned to the z value (elevation) accuracy of
the GPS data.
To facilitate the evaluation
of the project objectives it was determined that the ideal location for
the field site would be relatively smooth, flat, and free from obstructions.
Based
on these criteria the collegiate football field at The University of Arizona
was selected as the
field site.
Field Data Collection:
Students
were provided laptop computers with installed copies of
Geolink®
and 8 channel Motorola receivers.
One
advantage of using the football field as the field site is that the yard
markers chalked onto the field in 5 yard intervals helped create a horizontal
grid pattern (in the x direction) without extra effort by the group members.
These
yard markers along with the end zone and sideline lines provided the bases
of our grid.
Static points were collected
at
six locations: the four corners, at both intersections
the of 50 yard marker, and the sideline (
pentagons).
Besides
these six new static points, two existing GPS locations were utilized to
validate static data: 1) USGS brass cap named UNIVERSITY, by the NGS with
a PID CZ1836 located approximately 270 meters (886 feet) NNW from the football
field (
yellow triangle); and 2) our base station
identified as BSE located on the roof of a university building approximately
291 meters (955 feet) to the NWN from the football field (
red
triangle).
Static
points:
The
static points were collected using static, differential Carrier Phase (L1
1.57542 GHz) settings.Each of the
six static points constituted a separate file and consisted of 20-minutes
worth of data.The measurement at
the published USGS brass cap was collected twice; once before collecting
the data from the field site and a second time after collecting the data
from the field site.The duration
and setting for these secessions were identical to those used for collecting
the static points on the football field.The
BSE base station ran continuously (total of 4.5 hours) being started before
taking the USGS brass cap measurements, and ending after recording the
second 20-minute USGS brass cap measurement.These
six new static data sets were very useful for providing spot location values
for accuracy assessment but provided a very crude distribution from which
to produce a surface model of the football field.To
achieve a higher density distribution of data points, kinematic (or moving)
data collection mode was deemed best.
Kinematic
points:
In an
effort to produce a highly accurate continuous surface elevation model
representing the football field, GPS data was colleted using the kinematic
mode.One of the group's goals was
to try to collect data in the z direction (elevation) using Carrier phase
+ C/A differential kinematic post-processed mode that would yield positional
accuracies around 10 cm to 20 cm.The
semi grid pattern created by the white chalk lines of the football field,
namely the goal line and the 10, 20, 30, 40, and 50-yard lines came into
play for this method of data collection.A
bicycle equipped with a GPS antenna mounted atop a pole was pushed along
the chalk lines.It was hoped that
the use of a bicycle would help to minimize the vertical fluctuations that
might arise from one's walking gait and aided in producing a relative constant
velocity of about 2 meters per second along the 13 interconnected transects.These
transects took about 45 minutes and all points were collected into a single
file.This method allowed for a near
continuous array of data representing the
surface
of the field along the x direction.This
data set contained 1598 individual points.
Remote
sensing:
With a
little investigation, some aerial photography and remote sensing data that
included the study site was located.The
university had contracted a third party to provide elevation data for the
entire campus.This remote sensing
data proved to be an invaluable addition to the GPS data set collected
by the group.Not only did the remote
sensing data set represent a third set of elevation data with which to
compare the accuracy of the field collected GPS data, (and vice versa)
but it also provided a fairly concentrated set of points collocated with
the field site making it possible to create an elevation surface on its
own merits.The term masspoints
was given to this data set.
DATA PROCESSING
Data Processing:
The
GPS static and kinematic data files were post-processed using differential
processing with the BSE data record as the base station.
The
software application GravNav ver 4.2 was utilized for the
post
processing calculations to produce corrected x, y, and z coordinates.
The
results were checked against the USGS brass cap as well as the remote sensing
elevation data.
The newly corrected
coordinates provided point accuracy levels that approached the tens of
centimeters range for the x and y, as well as the critical z (elevation)
coordinates.
The x, y, and z coordinates
were written to an ASCII text file.
These
were imported into ArcView (ver 3.1) with Spatial Analyst (ver 1.1) as
event themes, and saved as shape files.
These
shape files were in turn imported into ArcInfo (ver 7.2.1) using the <shapearc>
command to create a point coverage.
The
point coverage was projected into UTM (units of meters) coordinates from
geographic (units of decimal degrees) coordinates.
Once
in ArcInfo, the topogridtool AML was invoked to produce an elevation surface.
These
steps were used on the
kinematic data set as well
as the masspoints data set.
As a
comparison to the interpolated surfaces created by topogridtool, TINs were
also created for the
masspoints and the
kinematic
points.
The cell size for the TINs
was set to 5 meter, the distance from yard marker to yard marker.
The
previous screen captures illustrate how ArcView3D (ver 1.0) extension provides
a powerful visual aid for interpret ting elevation surfaces.
All
of the 3D images utilized a Z factor of 10.
DISCUSSION
During the course of
post-processing the GPS data points the group found the previously reported
elevation
for the base station was in fact too low by around 20 meters (19.72 meters
to be exact).
Had the group not looked
for and found the remote sensing data set to evaluate their field GPS data,
they may not have found this error.
Even
with this discrepancy accounted for, the
difference
surface of masspoints minus kinematic points revealed some rather large
deviations amounting to almost 50 centimeters for points in the southeast
corner of the field.
Although the
group had initially thought the football field would have provided unobstructed
views of the sky, they came to realize that due to the geometry of the
satellites at 30
° northern latitude, the group encountered
some problems with data quality while in the southeast corner of the stadium
while in kinematic mode.
This was
believed to be the result of the "canyon effect" from the stadium grandstands,
and once clear of the southeast corner the rest of the data collection
process was uneventful.
The poor
quality of the data is indicated by the points colored
red
indicating a 4 or 5 for data quality (1 being excellent data quality) ranking.
With
the relatively larger number of data points available in the
kinematic
data set as compared to the
masspoints data set,
the influence of the weighted average TIN interpolation method is clearly
evident.
Unexpected though was the
behavior of the drainage enforcement algorithm in the topogridtool GUI
in ArcInfo.
The group found that
the enforce option (the default setting) indicating that drainage would
be enforced on the surface created some geomorphic artifacts which field
inspection clearly proved were not there.
Besides these types of
technical issues, other logistical hurdles to overcome were constantly
looming in the background.Database
management issues such as data definition language compatibility from the
GPS application not transferring over gracefully into ArcView/ArcInfo were
an example of one such item.ArcInfo
is not capable of having a number as the leading character of a field name,
but this is perfectly acceptable for Geolink.Significant
figures are also another issue of confusion when translating between databases
modules.Group members also found
that maintaining application version consistency and file data management
on the eight or nine field laptop computers with the desktop computers
in the research lab also required attention.
This course provided
graduate research students some practical exposure to the often cumbersome
process of GPS data collection, and the necessary tools for producing a
finished GIS project.
CONCLUSIONS
Perhaps one outcome of
this group's project will be that they will never look at a football field
in quite the same way again.However,
there are other lessons to be learned from this and other group projects.Some
examples would be to investigate other sources for possible data that may
involve field sites.Or, don't necessarily
trust other data sources without attempting to verify the data provided.In
addition, the well known adage that there is no substitute for good fieldwork
is also applicable here.
As a summary, below is
a top 10 list by students of the most common non-technical problems encountered
by the student groups, which were usually the most troublesome to overcome
or correct.
10
|
Remember to take the
matching hardware key for the appropriate software application.
|
9
|
Secure the necessary
trespassing/research permits before heading out into the field.
|
8
|
If you are a foreign
student working close to the Arizona border take your passport with you.
|
7
|
Don't mistake the witness
USGS markers for the GPS geodetic control station marker.
|
6
|
Take a reconnaissance
trip to locate the GPS markers before hand. You
may not be able to find them even with good x
and y coordinates.
|
5
|
Review relevant NANUs
before you head out to collect data for the day.
|
4
|
Keep the desired data
attributes simple.
|
3
|
Remember Occam's razor
when trouble shooting equipment in the field.
|
2
|
Take plenty of batteries,
and then some more.
|
1
|
Your field data is probably
more accurate than most of your current base maps you will work with.
|
The group learned how
to collect field data, process the data into an appropriate format and
present the data in a visual format, resulting in a GPS derived GIS elevation
model based on a standard football field.
ACKNOWLEDGMENTS
We would like to express our sincere thanks to
the following organizations for their willingness to contribute information,
or time: The College of Agriculture, The University of Arizona Advanced
Resources Technology group (ART), the USGS Sonoran Desert Field Station,
The University of Arizona Campus and Facilities Planning, and the many
University of Arizona graduate students that worked on different projects
over the years.
REFERENCES
The National Geodetic Survey
Products CD-ROM Data Sheets South Central States April 1996
AUTHOR INFORMATION
Andrew Honaman
Graduate Student Watershed Management
The University of Arizona
School of Renewable Natural Resources
Michael Kunzmann
Michael R. Kunzmann is an Ecologist at the USGS Sonoran
Desert Field Station which is located within the School of Renewable Natural
Resources. The School of Renewable Natural Resources is in the College
of Agriculture and is centrally located on The University of Arizona campus.
Correspondence may be addressed: Michael R. Kunzmann, USGS Sonoran Desert
Field Station, The University of Arizona, 125 Biological Sciences East,
Tucson, Arizona, 85721. Mr. Kunzmann may also be reached by telephone at
(520) 621-7282 or by email:
mrsk@sherpa.srnr.arizona.edu
Wolfgang Grunberg
Research Assistant
Graduate Student Renewable Natural Resources
The University of Arizona
School of Renewable Natural Resources
email wolfgang@nexus.srnr.arizona.edu