Field Activity 5: Visualizing and Refining Terrain Data

Introduction

Table 1: Snip from the Excel file used to input
elevation. Shows normalized data

The previous lab involved the creation of a landscape and the collection of the landscape’s elevation data. the purpose of the lab was to gain an understanding of sampling and the various methods of sampling. The data collected was put into an excel file which was normalized to be compatible with ArcMap. This meant there was only a X, Y, and Z column (Table 1).


The data points of the landscape were collected every 5 cm for the areas that had a lot of elevation change (hill, depression, ridge, valley) and every 10 cm for the areas of little change (plain) (Figure 1). These points will be plotted in ArcMap to be interpolated and transferred into a 3D visualizing software, ArcScene. Interpolation is the calculation method used to create new values for points within a known range. This allows for a more detailed data set and therefore high accuracy. The purpose of the lab is to create a DEM model of the landscape and compare the various interpolation methods including inverse distance weighted (IDW), Natural Neighbors, Kriging, Spline, TIN. 

Figure 1: The landscape created in the sandbox during the previous lab.
The needed features have been identified.

Methods

Figure 2: The data points after being added into ArcMap.
        There has been axes and the starting points
added to provide correct orientation. 

Before beginning the processing on the data points a folder was created along with a geodatabase (“sandbox”), both of which were dedicated to this individual lab to keep the files organized. The excel file created in the previous lab (refer back to Table 1) was imported into the sandbox geodatabase. To add the data points into ArcMap, the XY data was added via the imported excel file. The layer was exported as a point feature class (Figure 2).

Interpolation must be done in order create an accurate 3D visualization. In order to do so, the extensions had to be set to 3D Analyst. The feature class was then added as the input file for the first interpolation tool, IDW. This process was repeated with the remaining interpolation methods (Table 2). 

Once the interpolation was finished, the data was then exported to ArcScene for each individual feature class.  The scale had to be adjusted to the calculated range in order to portray the small area as a more accurate representation of the landscape. The orientation of the imagery is with the (0,0) point being in the lower left corner of the landscape (Figure 4). The various results are discussed further in the “Results” section. 

Table 2: The table provides a brief description of the various interpolation methods used to create DEMs of the landscape.

Results/Discussion

The overall results of the interpolation methods used: IDW, Natural Neighbors, Kriging, Spline, TIN; show a range of advantages and disadvantages. Some of the methods resulted in similar representations of the landscape created in the previous lab. Each individual method will be analyzed further.

Inverse Distance Weighted (IDW)

Figure 3: IDW interpolation result

The IDW interpolation result shows the landscape and the features needed (Figure 3). However, when analyzing the image, it is clearly interpolated by creating clusters of data points weighted heavier at certain points. However, if you compare Figure 3 to Figure 1 it is easy to point out the landscape features. A disadvantage of the IDW interpolation method is that it is not a smooth topographic representation of the landscape created in the previous lab. 

Natural Neighbors

Figure 4: Natural Neighbors interpolation result

The results of the Natural Neighbors interpolation method provided an accurate representation of the landscape created (refer to Figure 1) (Figure 4). An advantage of this interpolation method is the fact that creates a smooth surface when compared to the previous method, IDW. The extreme elevations are very clear within the image and therefore the landscape features are easy to identify. The valley leading into the low area surrounding the hill is easy to identify. The area near the starting point (0,0) is higher than the plain, could be referred to as an unintentional plateau. However, in the areas with little elevation change it is difficult to see the detail of the landscape. 

Kriging

Figure 5: Kriging interpolation result

The Kriging method of interpolation resulted in a very similar image to the previous method of Nearest Neighbor. The landscape shows the surface features of the grid (Figure 5). The ridge near the upper left portion of the landscape appears to be very broken up and not of high elevation as expected. The feature is not as defined as it appeared to be in the actual landscape (refer back to Figure 1).  An advantage of the method is a smoothed surface and transition. A disadvantage of this method is that there is not much detail in the image in the plain areas. The plain, though it had little change, still had some elevation change which did not appear in the Kriging method results. 

Spline

Figure 6: spline interpolation result

The Spline method of interpolation had a large effect on the intensity of the elevation change (Figure 6). This is a major advantage of using this interpolation method. One main aspect that is better with this method is the ability to see even the small elevation changed in the areas of the plain and plateau. It also represents the highest elevation well. The transitions from one elevation range to another is very smooth which allows for a more appealing DEM. 

Tin

Figure 7: TIN interpolation result

TIN interpolation was the one method that visually stood apart from the other (Figure 7). An advantage of the TIN method is that it shows the elevation extremes. The highest points, at the peak of the hill and another hill in the lower right area of the grid, are a light grey (almost white) which is easy to identify. The lowest elevation within the depression and valleys are also a lighter color than the higher surrounding landscape. 

Summary/Conclusion

After analyzing the results of the various interpolation methods it is concluded that Spline does the best at representing all aspects of the landscapes topography. The Spline method showed smooth transitions between the elevation ranges but also intensified the little changes in the flatter areas. This method’s results looked most realistic to the original landscape and did the best to represent the sandbox landscape.

 The overall results of the images proved that there had more than enough data points input into ArcMap to accurately represent the landscape that had been created previously. This means that there did not need to be a repeat of the lab to collect more points. The only errors that may have influenced the data is human error while relaying data vocally.

This survey provided experience with the advantages of sampling but also making sure to collect enough data to create a full picture of the survey area. This was the most important aspect of this lab when compared to others thus far. However, it is always not needed to collect many points in certain regions. As seen in this lab, with areas of little change there was not much need to collect many points because they were all similar and could be represented as so with few points. That being stated, areas of high amounts of change results in the need for more points for detail.