Field Activity 4:Sandbox Survery;Creation of a Digital Elevation Surface using critical thinking skills and improvised survey techniques

Introduction

Waldo Tobler`s First Law of Geography states that; “Everything is related to everything else, but near things are more related than distant things” (Esri GIS Dictionary). This statement is the basis of sampling; data collection of a representative population or sample area used to investigate a whole population. Well-chosen spatial samples can be used to create an acceptable description of the Earth`s surface. There are many factors that must be taken into consideration in order to create a well-chosen spatial sample. Some of these considerations include sample size; larger samples tend to be more representative of the whole, how to minimize bias when sampling, and which sampling technique is most appropriate for the area being sampled. A common scheme for spatial sampling is based on points within a grid framework and there are three primary sampling techniques; random, systematic and stratified. Each sampling technique has benefits and disadvantages. Both random and systematic sampling can be sub-classified into point, line and area sampling. Point sampling involves data collection at x, y intersections or in the center of the grid correlating to an x, y intersection, line sampling involves data collection at points along a line, and area sampling involves data collection within grid squares. Random sampling removes bias but can result in a poor representative population due to clustering of sample points. Systematic sampling involves evenly distributed points however, bias can lead to over or under-representation of a certain pattern. Some study areas have a know proportion of specific subdivisions, in these cases stratified sampling would be the best technique. Stratified sampling would evenly distribute sample points taken in proportion to each subdivision and the samples taken in each subdivision could be taken randomly or systematically. The goal of this activity is to create a terrain in an approximately one square meter area and then determine the most fitting sampling scheme and technique to obtain sufficient data for the creation of an accurate digital elevation surface of the terrain.

Methods

                It was decided that systematic point sampling, recording elevation (Z) at each  X, Y intersection was the most appropriate for this project; as the whole study area was only slightly larger than a square meter and the entire study area could be sampled efficiently in a relatively short time frame. A similar alternative to this approach would have been to record Z at the center of each grid square, however in some grid squares a sharp change in elevation near the center would make it difficult to determine which measurement to take. The stream box used was 114cm x 114cm, from these measurements it was determined that a grid with 6cm x 6cm squares would be ideal to capture sufficient data points for mapping. These divisions would result in 361 squares measuring 36 cm² each.  Meter sticks were used as guides and thumbtacks were used to mark each 6cm point on all four sides (figure 1) and string was wrapped around them to form a physical grid (figure 2) to ensure accurate location of the Z points recorded. The top of the stream box was made to be sea level, since the string grid was at the same level as the top of the stream box and the grid could also serve as a guide for sea level during data collection. Where the terrain exceeded sea level, the string was pulled into the terrain without damaging the structures so that all string sat at sea level. After the grid system had been completed over the terrain an origin was set at the southeast corner of the stream box. To facilitate data collection a similar grid of smaller proportions was drawn on paper and the Z value for each X, Y coordinate recorded in its corresponding box. In cases where the terrain in an individual grid square changed sharply, two measurements were taken and depending on which plane the change happened along the X or Y value was given a decimal value ending in .5 and then a whole number. The collected data was then entered into an excel spreadsheet, beginning at the origin and working up each Y column along the X axis. 

figure 2

figure 1

Results/Discussion

                After grid squares with sharp contrasts in elevation were split into two separate Z values the final number of data points collected was 434. The Z value`s minimum was -13, maximum was 10, the mean was ~-2.151, and standard deviation was ~4.146. The X and Y values both ranged from 1 to 19. The sampling method chosen at the beginning of this project and ultimately utilized for data collection appears to have been the best method however fewer points could have been collected. The primary issue turned out to be the grid squares with sharp changes in elevation; this was corrected through the collection of two Z values within those grids and depending on which plane the change happened along the X or Y value was given a decimal value ending in .5 and then a whole number. Another issue was the quantity of data points to be collected. The process was expedited through the delegation of tasks; one person held the measuring stick, another read aloud the value, and a third recorded the data.

Conclusions

                The sampling technique used in this project was slightly tedious but a majority of the points taken represent a 36 cm² section of the whole study area and the even distribution of points will give an unbiased, more accurate representation of the entire study area. Sampling is an effective way to create a reasonably accurate portrayal of spatial features on the Earth`s surface because surfaces that are closer together tend to be related to one another and data about a surface can be used to predict the nature of another surface that is close by. The activity of sampling a surface created in a small area provides an experience that would help with spatial sampling on a larger scale; it touches on the types of problems that could be encountered and the methods that can be used to solve these problems. After learning about the First Law of Geography, it seems the resulting dataset from this project was slightly excessive and fewer data points could have been collected in order to recreate the surface accurately. At the same time, the collection of data point was not costly and should result in a more accurate portrayal of the terrain that was created.