Exercise 3 - Field Navigation Map

Introduction

For this exercise we were to create two maps that are to be used in a later exercise. The maps are of a property outside of town that is owned by the University of Wisconsin-Eau Claire called The Priory. The priory was a former nunnery that was bought by the University in 2011 and started using the property for a child day care and a field methods laboratory in 2012-2013. We were to create these maps with one being in a coordinate system of decimal degrees and the other in Universal Transverse Mercator (UTM).

Each map will need to have specific coordinate system and projection requirements that are the most useful for their features and usefulness. I will go through each step of making these maps as well as the reasoning behind choosing each specific projection and coordinate system. I will also include some background information behind navigation, projections, coordinate systems and other things that I chose to include in my maps.

Methods

The first objective was to measure our personal pace count which is the length of two strides. To figure this out we needed to first measure 100 meters out. One person walked down the sidewalk behind Phillips Science Hall and another on the other end. The person at the far end held a receiver while the other held a laser measuring device. Once the person with the receiver was far enough away the person with the laser called out when they were 100 meters away. They then marked the point where that was and each of us took turns counting how many paces it took to traverse that distance. We counted every time our right foot was forward (two strides) and repeated this on the way back. On the way there I had 69 paces and on the way back I had 65. I took the average and set my pace count at 67. This will be used to measure out distances once we are out in the field at The Priory.

The next step was to create the maps I will be using out at The Priory. For the exercise we were given a geodatabase with all kinds of feature classes and data from The Priory. We were to chose from the data given to us what we needed to be most useful to us. I elected to use a satellite imagery of the area of Eau Claire County that The Priory was in, a five meter contour feature class of The Priory, and a feature class of the study area. I thought these were all that were needed when navigating The Priory in the future exercise. We then were to create a sheet with two separate maps on each side. One was to be in a geographic coordinate system and the other in a Universal Transverse Mercator (UTM) projection. The coordinate system map needed decimal degree measurements along the borders of the map while the UTM map need to have meter measurements. Each map will have their own usefulness when navigating.

In order to chose the correct projections for each map we need to know more about UTM and coordinate systems. Below is a image of the 60 evenly sliced UTM Zones across the globe (fig. 1). The equator separates the north and south UTM Zones. These zones help us create projections that give us an undistorted projection of specific zones on the globe. If you were to map in UTM Zone 34S for example you would get a great map of the area that lies in UTM Zone 34S but the rest of the globe would be distorted. So for my UTM map I chose a projection that encompses Eau Claire County, UTM Zone 15N. UTM projected coordinate system's are measured in meters instead of degrees so this map gives us a good representation of The Priory. Both maps were in the North American Datum of 1983 (NAD 83).

The 60 Universal Transverse Mercator Zones divided by the equator. (fig. 1)

Below are the two maps that I created in decimal degrees (fig. 2) and in UTM meters (fig. 3). Click on the maps to enlarge them. For each map I had to create a ticked border with measured increments for the respective units along with a grid. The UTM map was in meters and the Coordinate map was in decimal degrees. I then added the five meter contour lines and study area feature class on top of the satellite imagery. Next I put a north arrow, scale and metadata (projections, data sources and author) to each map.

Map of The Priory in decimal degrees. (fig. 2)

Map of The Priory in UTM (meters). (fig. 3)

Discussion

In the next exercise when we are using these maps I believe that I will have all the necessary information and data that will be need to help me navigate The Priory. The satellite imagery will prove useful when looking for landmarks and getting my bearings.

Some problems that could come up include the fact that my contour lines do not have a reference line and this might be a problem with figuring out if certain areas are sloped upwards or downwards. Another problem could come from the satellite imagery which could make it hard to see where the exact point I am along with the fact that this imagery is not super up to date. The Priory could have changed in the time since the image was taken and I could have incomplete or false data that causes me to make an error in the next exercise.

Conclusion

These maps will be very useful in navigating The Priory with just the right amount of information. The two separate maps should prove useful when measuring distances and slopes. I hope the maps I made will give me everything I need to have when navigating in the subsequent exercise and not lead me to get lost.

Sources

Map of the UTM Zones from National Geospatial-Intelligence Agency: Link

Exercise 2 - Visualizing and Refining Terrain Survey

Introduction 

The purpose of our second exercise was to take the data that we gathered in the first exercise with the survey and create a terrain model in ArcGIS. In this exercise we reformatted our data so we could put it into ArcGIS and then looked for any areas in our survey that could be improved upon. We would then resurvey the areas needed and create a finalized terrain model.

We had to overcome some different challenges in being able to create a working terrain model from the data we collected. As well as update and format our data we needed to make sure that our model was close enough to the real thing. We needed to make sure that each feature was properly represented and the resolution was high enough to provide the detail needed to show all the features.

Methods

As was hinted to in the first exercise, we needed to normalize the data so that we could put the data into ArcGIS. The original table seen in exercise 1 gave us a grid with letters on one side and numbers on the other, which could not be put into ArcGIS. I needed to first create a table with x, y and z-values in separate columns. As the table below shows (fig. 1) I took the letters and made them into numbers (A=1, B=2, etc.) and left the y-values as previously numbered. I took each letter and stacked the data on top of each other. So I changed all the A column into 1 and then started on the B column right below as you can see in the figure below which was changed to 2. This was repeated for all of the numbers.

Part of the finalized table with the corrected x and y values, ready for ArcGIS. (fig. 1)

Next, I put the table into ArcGIS and created points based on the x and y columns in the table. I did this by importing sheet1 which is the table within the excel file that I created and just dragging and dropping it into the table of contents. I then right clicked on the table once it was in ArcGIS and selected Display the XY Data. I then saved the resulting points as a feature class by exporting it into my geodatabase. The output was a grid of data points and the image below shows what the points looked like in ArcGIS(fig. 2).

Original layout of the data points imported from the table in ArcGIS. (fig. 2)

Once I had the data points in I was able to run some 3D-analysis tools. I ran five different tools that are used to extrapolate x, y and z-values into a surface. They all use different methods to create a surface between the given points based on the other points around them. I ran a IDW (fig. 3), a Natural Neighbor (fig. 4), a Kriging (fig. 5), a Spline (fig. 6) and a TIN (fig. 7). Each of the tools had the same window open with places to input a Input point feature, z value field, and location of the Output raster. There were some more optional inputs that I didn't use in this exercise.

IDW model (fig. 3)

The first interpolation was the IDW shown above (fig. 3). I input the point feature I made above, set the-z value to the z-column in the points feature and that was the resulting output raster. IDW is an interpolation raster surface that is made from points using an inverse distance weighted (IDW) technique. "[IDW] interpolation determines cell values using a linearly weighted combination of a set of sample points. The weight is a function of inverse distance. The surface being interpolated should be that of a locationally dependent variable" (1).

Natural Neighbors model (fig. 4)

The next interpolation technique was Natural Neighbors (fig. 4). I did the same as the last interpolation where I set the point feature as the input and z-value to the z-column in the points feature class. The resulting raster above shows a surface using the natural neighbors technique. "The algorithm...finds the closest subset of input samples to a query point and applies weight to them based on proportionate areas to interpolate a value (Sibson 1981)" (1).

Kriging model (fig. 5)

The third interpolation technique I used is the Kringing technique (fig. 5). I again input the point feature and set the z-value as the z-column in the point feature. The raster above is the output of the interpolation using kriging. "Kriging is an advanced geostatistical procedure that generates an estimated surface from a scattered set of points with z-values...[Kriging] involves an interactive investigation of the spatial behavior of the phenomenon represented by the z-values before you select the best estimation method for generating the output surface" (1).

Spline model (fig. 6)

The last interpolation technique I used was the Spline model (fig. 6). I again input the point feature and set the z-value as the z-column in the point feature. The raster above is the output of the interpolation using spline. "Spline...uses an interpolation method that estimates values using a mathematical function that minimizes overall surface curvature, resulting in a smooth surface that passes exactly through the input points" (1).

TIN model (fig. 7)

After the interpolation methods we also wanted to add another model which was a Triangulated Irregular Network or TIN (fig. 7). This is done by creating triangulated surfaces between the points provided. I input the point feature and z-values in the same way as the interpolations above.

My group discussed the different models and we concluded that the best one was the Spline model because of its smooth look and close resemblance to the actual terrain. However, we noticed that in the bottom left third of the terrain there was a canyon that was not properly shown in any the models. We needed to resurvey this area and create new points and model as well.

We went back out to the sandbox and resurveyed the area that covered the canyon not properly displayed and surveyed points every 5 centimeters instead of every 10. The image below shows the new points added to the previous points which should fix the error in the original terrain model. The portion we resurveyed was 50 centimeters by 60 centimeters and was 30 centimeters right of the origin (bottom, left-hand corner) and 50 centimeters above. This area covered the valley that was not modeled correctly and once we resurveyed would give us the detail we needed.

I then was able to add the newly surveyed points to the table and import that table back into ArcGIS just like I did above with the original table. I then used the table to Display XY Data the same as before and exported the new points into a points feature class in my geodatabase. The resulting points are shown below (fig. 8).

Revised layout of points after the resurveying of the terrain.
Providing a more detailed representation of the terrain.  (fig. 8)

I was then able to create a new Spline 3D-analysis model of the terrain using the updated data and giving me a more accurate model. I just input the new points feature class and set the z-value as the new z-column in the points feature. Below is a new terrain model including the new points (fig. 9). We chose to do it using the Spline tool because that gave us the best model for our terrain with the original data.

Finalized Spline model using the revised data points. (fig. 9)

The image below shows the area that was changed in the revised version of the data survey and shows the valley that was supposed to be there (fig. 10). The revised version has much better detail and gets rid of the "land bridge" between the ridges that shouldn't be there. The resurveyed points get rid of any misrepresentation caused by the low resolution of the first survey and a more detailed and accurate model in the end.

Close up of the area that was revised.
The resurvey giving us a more detailed representation of the valley (fig. 10)

Discussion

The data we collected was very good and closely matched the real life terrain we created in the sandbox. We were able to transfer the data from our grid drawing of the sandbox into the (x, y, z) points with relative ease. We just need to stack the x-values on top of each other and change the x-values from letters into numbers. This wasn't a difficult problem to fix.

Once we made the models and found the errors that needed to be addressed we were able to quickly resurvey and create more accurate data with great results. We didn't have to resurvey any other part of the terrain though, which was lucky. The area we did resurvey was a quick procedure and fixing this problem was also not too hard.

Selecting the Spline model was a easy pick due to its smoothness, especially compared to the other interpolation methods. Our terrain was very smooth however and other group's terrains might have needed something with more sharp or jagged looking surfaces.

Conclusion

In this exercise we were able to use our critical thinking skills and problem solving skills to fix any problems that we had such as the formatting of the data, resurveying the terrain and the choosing of the best interpolation technique.

This exercise showed us the importance of resolution and the needed for higher resolution when dealing with complex features. It is important to know what resolution is needed for certain projects once we enter our careers because you don't want to waste money on high resolution when it's not needed and you don't want to have inaccurate data with too low of resolution.

This exercise was a fun challenge that utilized our ability to circumvent problems and improve our accuracy in data collection and model building as well as knowing what resolution to use in certain places.

Sources

1. Interpolation technique descriptions provided by ArcGIS 10.2.2 Help.

Exercise 1 - Terrain Surface

Introduction

For our first exercise we were given the task of creating a terrain inside a sandbox in the Phillips Hall courtyard. We were to create the terrain using snow and come up with our own techniques of measuring the elevation of the terrain. We needed to create terrain that included a hill, depression, ridge, valley and plain. After we get our measurements we go on to create a digital model of the terrain in the next exercise.

Methods

First we constructed our terrain by shoveling snow into the sandbox and packing it down into different geographic features that included a hill, canyon, valley, plain, and ridge (fig. 1). The terrain sat below the top of the sandbox with the highest point level with it so we used the top of the sandbox as our sea level.

Building the terrain in the sandbox with snow. (fig. 1)

We then measured out the edges of the box which was 120 cm by 236 cm. we then measured out markers on each side of the box and placed tacks every 10 cm (fig. 2). There was 12 tacks on one side and 23 tacks on the other. We numbered the long side (1-23) and used letters for the shorter side (A-L).

The grid we created using tacks and string along the side of the box. (fig. 2)

Next we used string to create a grid on top of the terrain by winding the string between the tacks. We were short on string however, so we had to measure the sandbox one half at a time. We then went through each point of the intersecting string and measured the height (in centimeters) below the top of the sandbox. One person would use a ruler to measure while another would write down the data in a table  (fig. 3).

Joe measuring the data points and me recording the measurements. (fig. 3)

We then took that table of data and created a excel data sheet with all of our z-values being below zero. The numbers were used as the y-axis and the letters were used as the x-axis (fig. 4). The intervals between these points were 10 cm with the the last interval in the x-axis being only 2 cm and the last interval in the y-axis being 6 cm. We chose to exclude the extra length at the ends of the box to create a more perfect grid.

Data table. (fig. 4)

Discussion

During this activity we need to keep discussing and thinking about how we were to go about completing the task using the previous semesters blog posts as guidance. We then had to figure out in the next exercise how to use the data table we made in this part in ArcGIS and put it in a language that could be mapped. To do this we will need to have the x and y axes in centimeters and bring the lowest point of the terrain as our 0 or sea level.

Conclusion

Figuring out how to map a three-dimensional terrain involves a great deal of spatial thinking and coming up with ways to organize the data we were collecting. Finding how to get this data into ArcGIS also had to be a factor when collecting and organizing the data. The next step is to create a elevation model in ArcGIS in the second exercise.