Lab 1: Creation of a Digital Elevation Surface

Introduction

 Sampling is a method by which scientists obtain data on a small part of a whole area to make inferences for the entire area (Royal Geographical Society).  Sampling is used because often there is not enough time, money, or other resources to survey an entire geographic area.  There are three main types of sampling:

• Random: samples are collected in a random manner to reduce bias
• Systematic: samples are gathered regularly throughout the terrain which also avoids bias
• Stratified: the area is separated into subgroups and the number of samples in each subgroup is in proportion to its size in the whole area

 The objective for this lab was to design and then build a terrain in a sandbox to practice surveying techniques.  The terrain had to include a ridge, hill, depression, valley, and plain in any design of the students' choosing within the parameters of the approximate one square meter sandbox. This assignment was completed on February 1st, 2018 at 3:30 pm.

Methods

The location of our sample plot (sandbox) was the University of Wisconsin-Eau Claire campus:

105 Garfield Avenue

Eau Claire, WI 54702

Within the campus the plot was located across Roosevelt Avenue on the eastern side of Phillips Science Hall next to a small shed in an open plot of land.   

The group used the following materials to complete the survey:

• Sand 
• meter stick
• string
• thumbtacks
• pen 
• field notebook
• paper
• computer 

The group chose to use the systematic sampling method.  This method was selected because of its simplicity and complete coverage of the sample plot.  Sample points were taken at regular intervals so the entire terrain could be sampled consistently and therefore include all parts of the terrain (ridge, hill, depression, valley, and plain).  In addition, the consistent sampling method meant that a grid could be created with the string and thumbtacks to create evenly distributed intersections for sampling points.  Random sampling was not used because the inconsistent sampling would leave out certain aspects of the terrain.  Stratified sampling was not used because of the increased difficulty of determining sample points for each section of the terrain and creating a way to systematically sample within created subgroups.  Therefore systematic sampling was chosen to create the most complete data set.

Before creating a sampling scheme, the group chose a design to create out of the sand incorporating the 5 required aspects of the terrain listed in the paragraph above (Figure 1).  Then the group moved outside to create the design in the sand with their hands.  Figure 2 shows the completed design in the sandbox.

Figure 1. Model of the design incorporating a plain, ridge, depression, hill, and valley.

Figure 2. Completed design in the outdoor sandbox with captions of the features of the terrain.

After completing the design, the group created a grid to determine sampling points.  The sandbox measured roughly 114 by 114 centimeters (cm), so thumbtacks were inserted at 6 cm intervals on each side.  String was then wrapped around the thumbtacks for part of the grid and the remaining portion of the grid was created by tacking down individual pieces of string from one side to the other. The resulting grid was 19 by 19 boxes. with 18 string intersections. Figure 3 shows the grid sketched out in a field notebook and Figure 4 is the completed grid over the sand design.

Figure 3. Grid sketched in field notebook.

Figure 4. Completed grid and sand design.

Samples were taken at the intersection of the strings starting in the lower left corner (Figure 3) and moving up the first row.  The next set of points was taken from the bottom moving up.  Figure 3 has intersection numbers to illustrate the sampling movement.  A meter stick was held vertically at each intersection while another group member read off the measurement to a third group member who recorded the values with a pen and field notebook. Towards the end of the survey recording outside became impossible so the group member recording relocated inside and values were read over a phone call. The data was recorded in four columns: the number of the point, the x value, the y value, and the z (elevation) value.  Sea level (zero elevation) was considered to be the top of the sandbox, so all z values were negative.  Afterwards a formula in excel was used to convert the values to positive numbers based on the lowest z value recorded.  This method of entry was used because the cold weather prevented electronic entry and by using a numbered point system, x and y values could later be recorded to prevent prolonged exposure to the cold. In addition, minimal pictures were taken during this exercise due to the extremely cold weather which prevented the group from using electronic devices for a long period of time.

Results/Discussion

The total number of points sampled was 288.  After corrections, the minimum value was 0 cm, the maximum value was 8.5 cm, the mean was 3.8 cm, the median was 3.6 cm, and the standard deviation was 1.7 cm. The proximity of the mean and median values suggests that there were no outliers, or in this case extremely high or low points on the terrain that would skew the mean.  The group's sampling method directly related to the method of sampling choosen and was the best option for the objective of the lab.  The sampling technique did not change over the survey which created results that were expected.

There were some problems faced during sampling.  The string used to make the grid was very convoluted, so halfway through the grid-making process the group switched to tacking individual pieces of string across the sandbox instead of wrapping string around the thumbtacks.

The temperature was between -15 and -20 degrees Fahrenheit with wind chill during surveying, so multiple trips were taken inside to warm up.  This resulted in two columns being missed during the surveying process. Columns 10 and 11 are the columns that were likely missed as they were in the middle of the grid and the group paused somewhere around that location to warm up.  Figure 5 shows the data entry location where columns 10 and 11 were most likely missed.

Figure 5. Data entry including point, x, y, and z columns. The dashes with numbers represent the column numbers. The two column missed were most likely 10 and 11 thus the 10 and 11 recorded here would be columns 12 and 13.

This is an unfortunate possibility with field work.  Due to the unknown location of the missed columns the group will combine the known columns of data and process the resulting data.

Conclusion

Our sampling relates to the definition of sampling because we took a series of evenly spaced points from the entire area to determine trends of the whole terrain.  This is the exact definition of systematic sampling. Sampling should be used in spatial situations because there is not enough time, money, energy, or resources to completely sample every area on earth.  This activity relates to sampling spatial data over larger areas because the sandbox is a realistic representation of a terrain on earth and the need to choose a sampling method in the exercise is also true of scientists who want to obtain data about a terrain on the earth's surface. Systematic sampling is also a sampling technique that is utilized in real surveying projects.  Overall the group's sampling method did a decent job of sampling the area we were asked to sample.  The missing two rows however removed sample points that were intended to be included leaving the survey with 36 missing points and only 289 total.  To refine the survey, all data points would be obtained and samples would be taken around the perimeter of the sandbox as well.  The number of samples could be increased as well, either by decreasing the space between the grid lines or using stratified sampling to get more data points around the terrain features.

Sources

Royal Geographical Society. Retrieved February 2, 2018 from http://www.rgs.org/OurWork/Schools/Fieldwork+and+local+learning/Fieldwork+techniques/Sampling+techniques.htm
Hupy, Joseph. "Understanding Survey Grids." Handout. University of Wisconsin-Eau Claire. Eau Claire. n.d.
.