My GIS project was the attempt to create an accurate viewshed of the First National Bank Tower in Omaha. It is the tallest and most visible feature in the Omaha vicinity. I was interested to see the visual extent of the tower on the landscape. I also chose to work with this project because it seamed challenging and it included additional material that was not covered in class. The goal of my project was not to portray what could be seen from the top of the Bank Tower, but rather where you can see the tower from in the city. A problem with my line of site analysis is that big obtrusions like trees and buildings will not be figured with the calculations. However, I felt I could get a general idea of where the tower is visible.
The information I needed to conduct my project was a Digital Elevation Model and roads for the area. The DEM was used to analyze the surface terrain for the line of site portion and the roads helped me locate the tower. For my project I wanted an area large enough to include Douglas, Sarpy and Pottawattamie counties but smaller than their full aerial extent. I was able to achieve this by retrieving National Elevation Data from the USGS Seamless Data Distribution website. This website includes an interactive map for downloading. I was able to personalize the aerial extent of my data using a drag box. Obtaining my street files was more difficult. I was particularly interested in getting street data that had string addresses encoded in its attribute table because I wanted to geocode using Address LA, the same way we geocoded in our class lab. I did locate US State street .EDGE files in the ESRI Data folder but I was unable to extract the roads for my project from the files. I opened ArcToolbox and surfed around in the conversion tools looking for ways to change the format of the .EDGE street files. I ended up converting the geodatabase .EDGE files into shape files and then successfully clipped off the roads that were outside my DEM coverage.
My initial thought out method of creating the viewshed was to change the value of a rastor cell to the elevation of the bank tower. Then place the observation point for the view shed at this location. I used a query station in the ArcGIS software to help. I was able to locate a rastor editor toolbar that could be added into the user workstation in ArcMap. My first major problem arose when I tried to edit my raster DEM. The pixel type of my data was invalid so I tried on a different DEM from a different provider and received the same invalid result. Once again I looked through ArcToolbox but was unable to find a conversion that helped. I enlisted the help of a classmate who was also working with DEMs and viewsheds. He explained to me a way that dealt with modifying a Triangulated Irregular Network.
This method created a point with elevation and added it to an existing TIN. This made a new TIN with the points elevation triangulated into its surface. At this time I started creating a TIN from my DEM. I did this by using Spatial Analyst to make a contour map of my region. Next, I created a TIN from that contour map using 3D Analyst. Then I put a point in a shape file on the location of the tower. This was done by creating a shape file in ArcCatalog and then using the Editor Toolbar to make a point. I opened the attribute table of the tower location point and then added a field to the table. I named the field HEIGHT and gave it a value of 510. This value represented the meters above sea level that the tower reached. I calculated this value by adding the towers height to the ground elevation value of the tower location. Finally I used 3D Analysis to modify my existing TIN. Height was the field I used for the source of elevation in the modifying triangulation process.
The methodology proved to work but a significant problem existed with my results. A close up view showed that my modified TIN did have a very steep pyramid looking figure (picture 1 below). This figure reached up in elevation, but their was only a pie shaped visible area extended out into the distance from my observation point (picture 2 below). I was very puzzled that the viewshed had something obstructing most of the visibility. I repeated the process and I continually got the same kinds of results. The viewshed was incorrect and I could not figure out why.
Picture1 Picture 2
I determined that one of the possible problems could be that I created a slope that was too steep and that all sides arose to a very fine point (the point I was modifying). This was somehow causing difficulties with the results in my viewshed. To solve this I tried to come up with ways to make a more gradual slope up and then place more than one observation point at and around the top. My first method was placing several points that gained in elevation towards the center. This was ineffective because between the incremental gains of elevation, the triangulation of the modifying TIN led elevations down to the base values. Next, I constructed polylines that incrementally gained in elevation to a peak rim (picture 3 below). This did produce a more gradual slope but the original contours were still cutting through my feature (picture 4 below).
Picture 3 Picture 4
I was left with polygon features to modify my base TIN with (picture 5 below). As the idea played out in my head it seemed logical. With polygons I was giving all the points and lines in that specified region an elevation value as apposed to just one point or polyline among others. There were also some new triangulation methods that became available when modifying a TIN with a polygon feature. The two that interested me was the “soft replace” and “hard replace” methods. Up to this time in my project I only used the “mass point” or “line” methods of triangulation for the TIN modification because they were the only ones available. To add the polygon feature in my project I used the “soft replace” method. When I viewed the feature up close, I could see that this was creating a more gradual slope up to my desired elevation without any interruptions (picture 6 below). After that, I was able to place my observation points on the figure I created to make the viewshed.
Picture 5 Picture 6
RESULTS AND DISCUSSIONS
On this line of site analysis of the polygon figure, the area that was visible extended throughout the whole map. At first I thought that the area visible was too much. Then I figured that since the trees and buildings were not calculated, the results may be accurate. The Missouri River flood plain really stands out on the viewshed. I studied the map and found that visibility did not decrease with the increase of distance from the tower, like I had expected . After reviewing the results I did not see a major fault in the patterns so I decided to use this for my final output (map 1 below).
I made this map again and added a layer with roads (map 2 below). This shows where the built up parts of the city are located in the coverage area. I also thought the roads helped give reference to where the visible areas are located through out the line of site analysis.
I made a comparison map of the viewshed and the DEM (map 3 below). The dendritic pattern of the river flowing in the DEM is repeated in the viewshed analysis. This is because the red (not visible) areas are located heavily in lower elevations along the backside slopes of the river valleys. I also correlated the visible regions with higher elevations in the coverage.
There was not enough time for an elaborate field analysis but I wanted to check some of the visible areas represented in the results. I focused on distant places that were surrounded by agricultural land. This way I could check my work without the problems of trees and buildings in my view. I went out on a route that was mostly around the fringe of the Omaha suburbs. I was able to document 20 points (map 4 below). I still had to maneuver around objects that were on the closer hilltops but I had no problems finding the tower. At my farthest distance (204th St. approx. 20 miles) the tower was fading into the gray of the background. At these points I was unable to see the pulsating light on the tower with the naked eye. When I checked with binoculars the blinking light would resume visibility. I was satisfied with the distance because on that particular day the atmospheric conditions for field site analysis were unsatisfactory. On a different day I believe I could have traveled closer to the limits of my map region before having these sorts of problems. I had to cut the field analysis short because of time. I did not get to verify locations in the red (not visible) or more of the coverage area like Council Bluffs.
I made this map below (map 5) for a closer view at the sites I located. As I stated earlier most of the sites are in the periphery of the city. I also created a comparison of the DEM and viewshed up close (map 6 below map 5) to show that my verification sites are on higher elevations in the hilly region surveyed. This backs up the assumptions, to a minimal degree, made in map 3 earlier.
Overall, after the field analysis and reviewing my map I believe the viewshed is accurate. When I was driving the route I realized that the tower is commonly visible throughout the city. Most of the time concentration is spent looking at the road so the view is passed. Sure a lot of the visible areas can be nullified due to trees and buildings but in those regions you can shift around to see the tower without obstructions. I am not happy with the extent of my surveys in the field but I think that the goal I set out for was accomplished. I believe my maps display high probable areas where the tower is visible from the city.
I became much more familiar with the ArcGIS software in my time spent on the computer investigating how to reach my goal. Most of the time was spent figuring how to manipulate the elevation of the bank tower into my data. That left me with only basic operations that helped me conclude my output was logical. I also believe these two areas were the most important for learning more about the ESRI ArcGIS software.
Submitted by Paul Hunt 12/12/03