GIS & GPS APPLICATIONS IN EARTH SCIENCE Determination of an Optimal Bike Path Based on Slope & Distance Data S Daniel Zafar Fall 2011Geo 327G S Daniel Zafar 1. Introduction Many people bike in areas close to campus. Many of these are student who live nearby and use bikes as a quick and cheap method for travel in the otherwise cluttered campus areas. Hills in Austin can be a tenacious annoyance, providing more arduous journeys than in Houston, per se, where Qal is the topmost geologic layer for the city‟s full expanse and hills are not seen. The significance of this is compounded since single-speed fixed-gear street-bikes have come into vogue in recent years. These bicycles don‟t provide the extra assistance for traversing hills as other bicycles with multiple speeds would. Many bikers try to determine the best routes which avoid hills. The author finds himself in search of the optimal biking routes as well. One trip frequently made is between Bellevue Pl. and Harris Park Rd. (the abode of the author) and Genard St. and Leralynn St. (the abode of a very close friend). 2. Question The purpose of this project is to determine the most optimal route from the abode of the author to that of the close friend. This must take into account not only a minimized distance, but a minimized slope as well. 3. Method The only data needed are a DEM of the central Austin area and a shapefile consisting of streets. Using the DEM, a slope raster is generated which will give a higher number for a higher slope. An aspect raster is also generated, and then used to produce a new raster with the sloperaster which will account for slope in direction of movement. Next, this is combined with a rasterized version of the streets shapefile. The result is a slope raster which only contains data for cells in roadways. It will be used as a cost surface. This cost surface can be fed into the „Cost Distance‟ and „Cost Backlink‟ tools. These tools are needed to use the final tool, the „Cost Path‟ tool. This tool will produce a raster displaying the path which has the least accumulative cost associated with south-facing slope and distance. 4. Data Collection & Preprocessing The DEM dataset was collected from the USGS National Map Seamless Server Viewer. While the DEM “cut out” did require the area between the two points, the size of the DEM was much larger and included much of central Austin. The associated metadata was used for datum/projection data solely. The file was unzipped using windows and extracted into a folder within the project‟s folder. The URL for the DEM is: http://seamless.usgs.gov/website/seamless/viewer.htm The City of Austin GIS Data Sets webpage was used to download „Street Centerlines‟ and the associated metadata. The metadata was not used in this project and was not needed expect for basic datum/projection data. If the problem was able to be solved by „Network Analyst‟, the metadata could have been used to create step-by-step directions for the biker. Again, the file was unzipped using windows and extracted into a folder within the project‟s folder. The URL for the centerlines & metadata is: ftp://ftp.ci.austin.tx.us/GIS-Data/Regional/transportation/STREETS.zip5. Data Processing Within the project‟s folder use ArcCatalog to create a new personal geodatabase called “Bike_Austin”. Import the unzipped street shapefile as a feature class. Define the following feature classes (New > Feature Class): - “Boundary”, which stores polygons, - “Destination”, which stores points, and - “Starting_Point”, which stores points Make sure all feature classes utilize UTM 14N for special reference and WGS 1984 for vertical reference. Also alter the DEM to reflect the same special reference. Open a blank ArcMap document. Add the contents of the “Bike_Austin” geodatabase to the TOC. Also add the saved DEM to the TOC. Afterwards, change the data frame‟s coordinate system to reflect the feature classes. In the „General‟ tab, change units to meters. First the data had to be clipped. As the boundry class is already in the TOC, draw a box just inside the borders of the DEM. The symbology is specified as „Hollow‟ and purple was chosen. The boundary should match:Figure 1: Placement of the Boundary for Clipping Clip both the raster and the street shapefile to the bounding. For the shape file, use „Clip‟ under Analysis > Extent. For the DEM, use „Extract by Polygon‟. This will give us a nice clean set of data to work with that will not slow down the processing to a great extent. Many rasters will be generated at this size. The result should reflect this:Figure 2: Clipped streets and DEM Uncheck the boundary layer and the other two unclipped layers, so that only the clipped street layer and the DEM are shown. These can be removed from the TOC to avoid clutter. Change the DEM symbology to precipitation and check the inverse box. This display symbology will be used exclusively when changing the symbology of a raster. Now we will create point features for the starting and ending location. Turn on editing and select the „Destination‟ feature class to edit. Uncheck the DEM and turn on snapping. Place the destination at the intersection of Leralyn St. and Genard St. Use Google maps to find the location. In a similar manner, create a point for „Starting_Point” raster at Bellevue Pl. Harris Park Rd. This is where the trail will run. Next, we must create a cost surface raster. This is no quick task. To do this, we will combine the DEM raster with the streets dataset. To do this, go to tools and under „Conversion‟ select the „Feature to Raster‟tool. Use this to create a new raster. Select the field to be road_class, though this will be arbitrary. The raster will resemble a road but will have some different colors. We want all the values to be „1‟. Open the Raster Calculator and input the new raster. We want to make a „1‟ for every value and leave everything else as „no data‟. This is the calculator: Figure 3: Showing Raster Calculator Input The output should look like this (with the starting and ending locations displayed):Figure 4: Streets Raster with 'Starting_Point' and 'Destination'. We now have the streets portion of our cost surface. Now we must overlay slope information. Next, a raster displaying slope will be generated from the DEM. This will give us a higher number for more intense