Creation of a Digital Elevation Surface using critical thinking skills and improvised survey techniques

Introduction:

This lab was designed to learn how to take a sample of x,y,z data and turn that into an elevation model in Arcmap.  This was done by creating a landscape in a sandbox and taking points to be mapped at a later time.  One important aspect of this lab was deciding on what kind of sampling method we were going to use.  Since we knew the box was going to be 114x114 cm long we decided for the sake of this project a systematic sampling technique was used.  This means we took Z-data every 6 cm in the X and Y direction.  This would give the best spatial representation of what we were making.  The other sampling techniques we could have used are random sample and stratified.  Random sample is done by generation random points and taking the elevation of those points.  This is a good way to collect data but we felt that using random sample in this situation might not yield the best results.  Stratified sampling is when points are taken from certain areas in the study area and are suppose to be representative of the whole.  We decided this technique would also not yield the best results.  The objective of this lab was to use the best technique to collect the most accurate data we could.  This data will be imported into Arc and turned into a surface model.

Figure 1

Methods:

We chose to use the systematic sampling technique because we felt that taking the Z-value every 6 cm in both directions would yield the most accurate data that was representative of the whole.  Systematic and random sample did not quite fit what we wanted to do.  We built our landscape in a sandbox created by Dr. J. Hupy located in an open area directly across Roosevelt Ave from Phillips Hall at the University of Wisconsin - Eau Claire.  To create the landscape we used very advanced shaping tools called our hands.  We were sure to include a ridge, hill, depression, valley and plain.  Once our landscape was created we created a grid using thumb tacks and string.  The thumbtacks were set into the edge of the sandbox every 6 cm.  String was then attached to the tacks and wrapped around all the tacks until the grid (figure 1) was created.  One corner was chosen to be our 0,0. To determine our 0 elevation we measured to the dirt at the bottom of the box to the string.  When taking our Z measurements we measured from the string to the sand and entered this number into the Excel table we created.  The Excel table was set up before the lab with all the X,Y points we were going to collect.  After all the Z measurements were taken a calculation was done to get the true elevation.  This was done by doing 14-measurement.  This gave us the height of the sand from the zero elevation we determined before collecting any data.  The data was entered into Excel by using the Excel app on an Iphone.  The sheet was created on a computer and exported to the phone.  This allowed for easy data entry in the field.  All in all the whole data collection process took about 2 hours.

Figure 2

Results/Discussion:

We collected a total of 401 points using the systematic method. Here are some statistics on our data:

Minimum: -.5 cm

Maximum: 23 cm

Mean: 8.7 cm

Standard Deviation: 4.43 cm

We felt that the systematic sampling method will somewhat accurately represent our sandbox.  We felt that this method maybe lacked some detail in the areas that have big changes in elevation but for the most part it represents the sandbox very well.  We found ourselves making up measurement for areas that we really wanted to emphasize like the "mountain peak" or the "plain".  We realized we were doing this and it was skewing our data so we stopped doing it.  We also had areas that the sand was going above our string so it was hard to get accurate measurements.  These are the areas we found ourselves emphasizing the Z values.   We realized we were doing this and went back to taking data the way we were.  

Conclusion:

Our sampling was a great example of systematic sampling.  We took samples from throughout the area of study.  This way of sampling is hard to do when looking at an area that is not 114 cm x 114 cm.  When the area is 10 miles big, is it much harder to do a systematic sample.  This is when a random sample or a stratified sample is useful.  Sampling is hard to do in a spatial situation because land and space is always changing and ever different.  A sample will never be truly representative of the whole.  We feel that the data we have collected will effectively represent our sandbox.  To get an even better sample the measurement could have been taken every 3 cm but that would take way too long!