Mapping 'Bikeability' for the St. Johns River-to-Sea Loop Alliance

MAPPING ‘BIKEABILITY’ FOR THE ST JOHNS RIVER-TO-SEA LOOP ALLIANCE A PAPER SUBMITTED FOR COMPLETION OF SENIOR RESEARCH FOR THE COLLEGE OF ARTS AND SCIENCES STETSON UNIVERSITY BY Kelsy Michely IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF BACHELOR OF SCIENCE ENVIRONMENTAL SCIENCE ADVISORS Dr. Wendy Anderson, Ph.D. Dr. Jason M. Evans, Ph.D. AUGUST 2017 ii Table of Contents Table of Contents (p. ii) List of Illustrations (p. iii) List of Tables (p. iv) Acknowledgements (p. v) Abstract (p. vi) Introduction (p. 1) Study Area (p. 1) Literature Review (p. 1-4) Methodology (p. 4-9) Observations and Data Analysis (p. 9) Conclusions and Discussion (p. 10) Figures and Tables (p. 11-26) Works Cited (p. 26-27) iii List of Illustrations Figure 1 – p. 11 Figure 2 – p. 12 Figure 3 – p. 13 Figure 4 – p. 14 Figure 5 – p. 15 Figure 6 – p. 16 Figure 7 – p. 17 Figure 8 – p. 18 Figure 9 – p. 19 Figure 10 – p. 19 Figure 11 – p. 20 Figure 12 – p. 21 Figure 13 – p. 22 iv List of Tables Table 1 – p. 23 Table 2 – p. 23-25 Table 3 – p. 25 v Acknowledgements I thank the St Johns River-to-Sea Loop Alliance for providing me an internship with abundant opportunities to learn, and I specifically thank Marguerite Ardito for working closely with me. I also thank Dr. Wendy Anderson and Dr. Jason Evans for their mentorship, and also my peers at Stetson University for their support. vi Abstract The St Johns River-to-Sea Loop (the Loop) is a 260-mile long multi-use trail that spans five counties, yet cyclists do not know what parts are complete to safely travel. I predict that more people will get involved with the Loop when they know what parts they can safely cycle and what parks or points of interest are around the Loop. There are studies that show that higher knowledge of the destination results in greater use, so people who have an idea of where they are going on the Loop will use it more, whether it is for cycling, running, or other recreational forms. The bicycle level of traffic stress (LTS) system is a way to score stress of riding on roadways from 1-4, 1 being the lowest level of stress and 4 the highest stress level. With the Google Maps, LTS Map, Parks Map, and the ESRI Story Map, SJR2C users have more information on what parts of the Loop are completed, how stressful the cycling conditions are for FDOT managed roads around the Loop, and what points of interest they can visit throughout their trip. Introduction The St Johns River-to-Sea Loop (the Loop) is a 260-mile long multi-use trail that spans five counties, yet cyclists do not know what parts are complete to safely travel. Local businesses are not reaping the benefits of increased cycle tourism. I predict that more people will get involved with the Loop when they know what parts they can safely cycle and what parks or points of interest are around the Loop. Study Area The study area includes the five counties along the Loop: St Johns, Flagler, Putnam, Volusia, and Brevard. These counties are relevant to my research because the Loop spans them and most likely, people will be cycling from nearby roads in the five counties to get to the Loop. Analysis was conducted using ArcGIS, ArcGIS Online, and Google MyMaps, using Stetson University computers, as well as my personal computer with my ArcGIS Student License. Literature Review Assuming that cyclists’ use of trails increases when there are clear maps showing where they can go and what there is to do around the trail, it is important to examine previous studies. A study based on tracking visitors using GPS monitoring systems shows that higher knowledge of the destination “results in greater use distribution regardless of activity type,” so people who have an idea of where they are going on the loop will use it more, whether it is for cycling, running, or other recreational forms (Beeco and Hallo 2014). Frequent visitors “with motivations to exercise do not stay as long as other visitors,” while larger groups and those seeking to develop skill stay longer (Beeco and Hallo 2014). People wanting to develop skill are those who want to get better at cycling, running, or another recreation form. Becco and Hallo (2014) also suggest that the more familiar users are, the further they travel, encountering more zones, and increasing distance from starting points. When people know where they are going, 2 they travel further and through more areas. Understanding travel behaviors helps trail managers improve the trails as well as the experiences of the users. With this, the trail gets the most use and these users are traveling through more areas, with higher chances of contributing to those economies. A 2013 study by the Institute for Tourism and Recreation Research (ITRR) found that bicyclists are supporting Montana communities since it is a slow-paced travel form, and spend around $100 per person per day (Polovitz et al. 2014). Since cyclists are travelers, they spend more nights in different towns, supporting each community. Cyclists want wholesome food and various accommodation options where they stop, so communities wanting to keep cyclists in town should consider cyclist-friendly restaurants, breweries, and lodging. When the cyclists have a positive experience with a local business, they tend to share and recommend the business to others (Polovitz et al. 2014). This places the importance of investment in local businesses catering to cyclists to get a return of investment. Creating gap detours on bikeable roads is a key concept to improve the Loop’s connectivity. Without gap detours, use of the trail is discouraged because cyclists do not know where to go. A preference survey among tourists in Dublin gave results that cycling tourists are willing to increase their travel time by 100% to cycle on a facility fully isolated from traffic rather than on a road without cycling infrastructure (Deeniham and Caulfield 2015). They are also willing to increase their travel time 40-50% to cycle on a road with a bike lane rather than one without (Deeniham and Caulfield 2015). When planning out gap detours, it is important to put priority on routes that have a bike trail separated from traffic, and if there is not one, then the next priority should be roads with bike lanes. If a road with a bike lane is unavailable, considering a road’s level of stress is the next thing to do. 3 The bicycle level of traffic stress (LTS) system is a way to score stress of riding on roadways from 1-4, 1 being the lowest level of stress and 4 the highest stress level (Mekuria et al. 2012). The LTS decision chart from the Mineta Transportation Institute helps to determine bicycle stress levels for any roadway (Mekuria et al. 2012). The LTS rankings range from 1 to 4, with “LTS 1” as a tolerable level for most children, “LTS 2” as a tolerable level for most adults, “LTS 3” as a level tolerated by experienced cyclists who are confident yet still prefer a designated riding area, and “LTS 4” as a tolerable level for the fearless cyclists (Mekuria et al. 2012). Full descriptions are detailed in Table 1. Focusing on bikeable detours improves the Loop’s connectivity, so people will be more inclined to take the detour. Other studies can be considered in the map-making process when creating gap detours. For example, a study on automatic bicycle routing shows that datasets from mobile sports tracking applications can be used to provide automatic route suggestions for cyclists (Bergman and Oksanen 2016). Fitness applications can be used to provide data for planning (Jestico et al. 2016). If there is data on routes that cyclists frequent, it might prove useful in determining bikeability for a gap detour. Mapping bikeability has been shown to promote bicycle travel (Winters et al. 2013). Once there are maps, funding is another thing to focus on. For grant-writing purposes, I had the idea to incorporate educational signs along the Loop. Florida Department of Transportation (FDOT) funds the construction of the ground on which the Loop is on, but funding of trail signs and other trail attributes are up to the St Johns River to Sea Loop Alliance. Crossing signs are necessary shown in a study of bicycle and passenger car collisions, with 78% of collisions occurring in crossing situations (Isaksson- Hellman and Werneke 2017). To keep cyclist collision risk down, a great number of signs are 4 needed along the 260-mile Loop, leaving little money for anything else (Yiannakoulias et al. 2012). Grants that the Alliance apply for focus on the recreational importance of the Loop, but when another aspect is brought into the application, like education, the Loop will reap greater benefits. Educational signs posted along different points of the Loop for users to examine while taking a break could serve as a tool to obtain funding as well as to educate people on different things associated with recreation on the Loop like health and conservation. One option for educational signs are those that focus on the attributes of a certain part of the Loop that foster human well-being. With this, the importance of natural ecosystems in wildland-urban interface (WUI) areas is highlighted because they “provide biodiversity, improved water and air quality, protection of wildlife habitat, and recreation/tourism opportunities,” all of which foster human well-being (Kil et al. 2012). Recreation can be used as a conservation tool in endangered species’ habitats by increasing environmental literacy, mitigating the negative effect of species decline that recreation has on public lands (Cornelisse and Duane 2013). Methodology ArcMap to Google Map I downloaded Florida Counties 1M from “St Johns River Water Management (SJRWMD) Open Data.” Then, I opened Florida Counties 1M shapefile in ArcMap, selected attributes for Brevard, Flagler, Putnam, St Johns, and Volusia counties, created layer from selected features, and projected into NAD_1983_StatePlane_Florida_East_FIPS_0901_Feet (Table 2). I opened “Jason’s reference list” in the “SJR2C GIS Interns” folder in Google Drive, and opened “Map showing Loop Status based on SUNTrail data from May 18, 2016 using Google 5 mymaps with 5 county boundaries” (Table 2). Export as KML. I started ArcMap and found the conversion tool titled “From KML” to add this map KML and save. This layer shows trail status 0 as yellow, 1 as green, 2 as red, and 3 as purple. To change symbology from layer properties, I set value field as SymbolID and add all values. I changed 0 to purple and label “GAP NO ACQUISITION,” 1 to green and label “EXISTING,” 2 to red and label “GAP REQUIRING ACQUISITION,” and 3 to yellow and label “PROGRAMMED/FUNDED.” After changing symbology, I saved map as Project_Status. I used the conversion tool “To KML” to convert the 4 color project status SJRTrail_7 layer to KML. Next, I created a new map through ArcMap and saved as “Completed_vs_Incomplete.” Add SJRTrail_7 layer. I opened symbology from layer properties, set value field as SymbolID and added all values. Next, I set 1 as green and label “EXISTING,” and set 0, 2, and 3 as orange and label “NOT COMPLETE.” Find conversion tool “To KML” and convert the SJRTrail_7 layer to KML. Save in Layer_to_KML folder. Then, I created a new Google My Map titled “SJR2C Trail Schedule.” To begin, I imported the 4 color project status SJRTrail_7 layer. Layer comes in as many placemarks with colors representing trail statuses. With this, I had to rename each placemark to the trail status that correlates with the color. After that, I styled by “STATUS.” Colors changed but placemarks were grouped into the four project status categories, so I edited the color to correlate with the project status labeled. I made 6 copies of the map, titling each “North East Segment,” “North West Segment,” “Central East Segment,” “Central West Segment,” “South East Segment,” and “South West Segment.” For each segment map, I created detours for that segment by using the “Draw a Line” button, select “add biking route,” place the starting point at the end of an existing segment and place the end point at the beginning of the next existing 6 segment. I renamed detour to the area like, “Palatka Gap.” From “http://www.instantstreetview.com/” I copied and pasted the addresses of each detour start and end point and evaluated the conditions of the road. I made sure to edit start and end points to access better road conditions, especially if original road was dirt/gravel/etc. Key components from the Google Map work flow are detailed in a work flow diagram (Figure 2). Links to Google Maps are detailed in Table 3. LTS Map Moving on to the LTS Map, I downloaded data from http://www.fdot.gov/planning/statistics/gis/roaddata.shtm. The files were “FDOT Maximum Speed,” “FDOT Median Type,” “FDOT Median Width,” “FDOT Number of Lanes,” and “FDOT Bike Lanes” (Table 2). The Maximum Speed file has the maximum speed limits for all Florida roads. I needed that data for the initial speed limit questions in the LTS decision chart like, “Is the speed limit less than or equal to 40 MPH?” (Figures 4-6). I downloaded the Median Type and Median Width files to have a file for analysis purposes that would help me understand if there was a median on a certain roadway. The Number of Lanes file has data on the number of lanes for all of the roadways. I needed that data for the diagram questions such as, “Are there 6+ lanes?,” or “Are there 4+ lanes with no median?” (Figures 4-6). The Bike Lane file has bike lanes so I needed that data for questions asking if there is a bike lane on the segment (Figures 4- 6). I added these layers into ArcMap and clipped them down to a five county boundary of St Johns, Flagler, Putnam, Volusia, and Brevard. I needed to determine the bike lanes with and without alongside parking, so I converted the bike lane clip to Excel to add a column on alongside parking, with attributes as ‘YES’ or 7 ‘NO.’ With the satellite world imagery map as the basemap, I went through and selected each attribute to determine if there was alongside parking. I recorded my findings in Excel and converted from Excel to table in ArcMap. I joined the bike lane clip to the new bike lane table based on roadway. From there, I selected attributes in the bike lane clip "bike_lane_table:ALONGSIDE_" = 'NO' and created a layer from selected features. I cleared the selection and then went back again to select "bike_lane_table:ALONGSIDE_" = 'YES’ and created a layer from selected features. Next, I completed a series of joins that helped me select features that fit the requirements of certain parts of the LTS decision chart. I was able to move on through all pages of the decision chart, creating new layers from selected features and renaming them to the LTS number that they corresponded with; however, I was not able to finish page one of the LTS decision chart (Figure 4). I did not have any data on center yellow lines, so I created layers for the question right before the chart asks if there is a center yellow line, “Is speed limit > 25 MPH?” I selected attributes with speeds greater than 25 MPH, created a new layer, and then switched selection to create a new layer for the speeds less than 25 MPH. I exported these two new layer’s tables as a DBF file because exporting as an Excel file failed. I opened the DBF files through Excel and added a column in each for center yellow lines. I went through and selected each attribute in each of the layers to determine if there was a center yellow line. I recorded my findings as ‘YES’ or ‘NO’ in excel and converted from excel to table in ArcMap. From there, I could select attributes with ‘YES’ or ‘NO,’ create layers from the selections, and rename the new layers with the corresponding LTS number. Parks Map 8 To create a map showing parks within three miles of the Loop, I started by navigating to http://www.fgdl.org/metadataexplorer/explorer.jsp. I downloaded “FDOT DISTRICT 2 - PARKS,” and “FDOT DISTRICT 5 - PARK BOUNDARIES” (Table 2). Next, I add the shapefile layers into a blank ArcMap and saved as Parks_POI Map. I merged “parks_d2_jun06” and “parkbnd_d5_jan07” and saved layer as “parks_merge.” Then, I added SJRTrail_7 layer to map to have SJR2C Loop layer. I used the buffer tool to create a 3-mile parks buffer from the loop layer and save layer as parks_3. I intersected “parks_3” with “parks_merge,” and saved as “parks_intersect_3.” This was the final layer I left visible, along with the Loop layer. To get a basemap, I added “World Light Gray Canvas Base” because the contrast of the gray to the colors of map layers enhanced visibility of park boundaries. Finally, I inserted necessary map elements like legend, north arrow, scale bar, Stetson University logo, and title. ESRI Story Map I created an ESRI Shortlist Story Map on ArcGIS Online, titled Bicycle Emergencies. Within the My Content page, I navigated to the Web Map Application and created tabs for Attractions, Parks, Food, Lodging, and Bicycle Shops (Figures 8-12). I added points of interest into each tab, including picture, address, phone number, website, and description. After adding points of interest, I downloaded the “SUN Trail Data” layer and the “Blue Spring Gap Detour” layer as a KML from the SJR2C Current Complete Google My Map. This must be done through a Google Chrome web browser or function will fail. I directed myself to ArcGIS Online, then to the My Content page, and clicked on “Add From My Computer,” to search for the SUN Trail Data KML and then repeat to add the Blue Spring Gap Detour KML. Then from the My Content page, I clicked on the Sun Trail Data KML link to see more information. I scrolled to the bottom 9 right of the page and copied URL. I went back and opened the web map for Bicycle Emergencies, clicked on “Add,” “Add layer from Web,” then “Add KML,” and finally pasted the copied URL link. I repeated for Blue Spring Gap Detour KML. I created a work flow for Maggie Ardito so that if the Loop completion status updates, Google Maps update, or she wants to change layers on the story map, she can follow this work flow (Figure 12). Story map is accessible at: https://arcg.is/10m85K (Figures 8-12). Observations and Data Analysis My final result maps that stem from data analysis are the Parks Map and the LTS Map (Figures 3 and 7). The Parks Map shows parks within 3 miles of the Loop. It also includes a four color trail status of the Loop. The LTS Map details Bicycle Level of Traffic Stress (LTS) for FDOT managed roads in the study area. Using Excel to analyze the mileage in each trail status, I found that 37% of the Loop is existing, while 36% is in the Gap Requiring Acquisition status, 14% is in the Programmed/Funded status, and 13% is in the Gap No Acquisition status. Overall, 37% of the Loop is complete and 63% is incomplete. There are 94.2 completed miles of the Loop for users to cycle. When the next round of researchers and interns work on the Google Maps, my hope is that the view count analysis will help determine if there is a significant difference in people getting involved with the Loop. My LTS analysis helps contribute to existing knowledge because this process has been completed in few locations. Also, the ESRI Story Map includes the general public in using ArcGIS Online. 10 Conclusions and Discussion With the Google Maps, LTS map (Figure 7), Parks map (Figure 3), and the ESRI Story Map (Figures 8-12), SJR2C users have more information on what parts of the Loop are completed, how stressful the cycling conditions are for FDOT managed roads around the Loop, and what points of interest they can visit throughout their trip. My hope was that the Google Maps could be published on the SJR2C website, but it cannot be done at this time. The Alliance and I had the mission to cycle each gap detour around the Loop to evaluate the conditions and quality of the detour; however, this was not accomplished due to summer weather conditions and conflicting schedules. Since we were not able to get out and test the detour, we could not publish them for the public. Furthermore, I was not able to monitor the view count for the Google Maps to analyze any view trends for each segment because they were not published. My recommendation for future SJR2C interns is to focus on vetting the selected detours and to publish the Google Maps on the SJR2C website as soon as possible so that cyclists can get from one existing segment to another. I presented my idea of incorporating both a recreational and educational trail signs to Marguerite Ardito, and she is taking this in stride to use in the application process for grants. 11 Figures and Tables Figure 1. Map showing study area of Florida counties: St Johns, Flagler, Putnam, Volusia, and Brevard. Data from “Florida Counties 1M” shapefile obtained from "St Johns River Water Management District (SJRWMD) Open Data.” 12 Figure 2. Google Map work flow diagram detailing the key components of creating trail status maps with gap detours. Color coded based on how trail status data comes in. 13 Figure 3. St Johns River-to-Sea Loop trail statuses including Gap No Acquisition, Existing, Gap Requiring Acquisition, and Programmed/Funded, as well as Parks within 3 Miles of Loop. 14 World Light Gray Canvas Base used as basemap. Scale bar and north arrow included for reference. Figure 4. Level of Traffic Stress (LTS) Decision Chart (Mekuria et al. 2012). Page 1 of decision chart to follow if there is mixed traffic without a bike lane (Mekuria et al. 2012). 15 Figure 5. Level of Traffic Stress (LTS) Decision Chart (Mekuria et al. 2012). Page 2 of decision chart to follow if there is a bike lane and no alongside parking (Mekuria et al. 2012). 16 Figure 6. Level of Traffic Stress (LTS) Decision Chart (Mekuria et al. 2012). Page 1 of decision chart to follow if there is a bike lane with alongside parking (Mekuria et al. 2012). 17 18 Figure 7. Bicycle Level of Traffic Stress (LTS) for Florida Department of Transportation (FDOT) roads in the counties of St Johns, Flagler, Putnam, Volusia, and Brevard. Legend includes LTS 1 through 4 as well as the St Johns River-to-Sea Loop and the five County Boundaries. OpenStreetMap used as basemap. Scale bar and north arrow included for reference. Figure 8. View of “Attractions” tab in St Johns River-to-Sea Loop ESRI Short List Story Map. Each item features a photograph and description including address, phone number, website link, and a brief description. 19 Figure 9. View of “Parks” tab in St Johns River-to-Sea Loop ESRI Short List Story Map. Each item features a photograph and description including address, phone number, website link, and a brief description. Figure 10. View of “Food” tab in St Johns River-to-Sea Loop ESRI Short List Story Map. 20 Figure 11. View of “Lodging” tab in St Johns River-to-Sea Loop ESRI Short List Story Map. 21 Figure 12. View of “Bicycle Shops” tab in St Johns River-to-Sea Loop ESRI Short List Story Map. 22 23 Figure 13. Work Flow created to use when changing the Loop layer that is visible in the ESRI Story Map. Table 1. Levels of Traffic Stress (LTS) descriptions for each level (Mekuria et al. 2012). Name of Source Source Data Map Link (if applicable) Map showing Loop Status based on SUNTrail data from May 18, 2016 http://www.fdot.gov/planning/system s/SUNTrail/SUN_Trail_Network_Sh apefiles.zip http://bicycledaytona.org/SJR2CLO OP/maps/SUNTrail_SJRCLoop_map box_legend.html SJR2C Loop https://www.google.com/maps/d/u/0/ https://www.google.com/maps/d/u/0/ 24 May 2017 viewer?ll=29.080535614365612%2 C- 80.95225133154298&z=9&mid=1w xbO7t7yCbNHD82pmfQ2yJLYZX4 viewer?ll=29.080535614365612%2 C- 80.95225133154298&z=9&mid=1w xbO7t7yCbNHD82pmfQ2yJLYZX4 Florida Counties 1M http://data-floridaswater. opendata.arcgis.com/da tasets/1fe947e33f484d7fabd7d22930 83eb96_7 https://arcg.is/0qaXj4 FDOT DISTRICT 2 - PARKS http://www.fgdl.org/metadataexplore r/explorer.jsp FDOT DISTRICT 5 - PARK BOUNDARIES http://www.fgdl.org/metadataexplore r/explorer.jsp FDOT Maximum Speed http://www.fdot.gov/planning/statisti cs/gis/roaddata.shtm FDOT Median Type http://www.fdot.gov/planning/statisti cs/gis/roaddata.shtm FDOT Median Width http://www.fdot.gov/planning/statisti cs/gis/roaddata.shtm FDOT Number of Lanes http://www.fdot.gov/planning/statisti cs/gis/roaddata.shtm 25 FDOT Bike Lanes http://www.fdot.gov/planning/statisti cs/gis/roaddata.shtm Table 2. Sources with source name, source data link, and a map link (if applicable). Google Maps Google Map Link North East Segment https://drive.google.com/open?id=1OGTo28CkcTO-­‐ uA3EdnxieV5m7LY&usp=sharing North West Segment https://drive.google.com/open?id=1YNosGNaB9OCEetVUy0EKxU3to8Q &usp=sharing Central East Segment https://drive.google.com/open?id=1Io1H_Ryp4OfFeEcE9R3kaku7Kss&us p=sharing Central West Segment https://drive.google.com/open?id=1y_j221w3hs_QRV2SSMslYawzanE&u sp=sharing South East Segment https://drive.google.com/open?id=1P2F192De1WEibA3Hy-­‐ 5xZjmLRr0&usp=sharing South West Segment https://drive.google.com/open?id=1ViNzHc_gUMUHgtatArR6OLFTTY4& usp=sharing SJR2C Trail Schedule https://drive.google.com/open?id=1rua5z84R_ohMg3W0VhEz2w5bD44 &usp=sharing SJR2C Current Complete https://drive.google.com/open?id=1JVctls_s9Rmeo63vp-­‐ DxUg24Pqw&usp=sharing Table 3. Google Maps for each segment of the Loop with links to each map. 26 Works Cited Beeco, J. A. and J. C. Hallo. 2014. GPS Tracking of Visitor Use: Factors Influencing Visitor Spatial Behavior on a Complex Trail System. Journal of Park & Recreation Administration 32:43- 61. Bergman, C. and J. Oksanen. 2016. Conflation of OpenStreetMap and Mobile Sports Tracking Data for Automatic Bicycle Routing. Transactions in GIS 20:848-868. Cornelisse, T. M. and T.P Duane. 2013. Effects of knowledge of an endangered species on recreationists' attitudes and stated behaviors and the significance of management compliance for ohlone tiger beetle conservation. Conservation Biology: The Journal Of The Society For Conservation Biology 27:1449-1457. Deenihan, G. and B. Caulfield. 2015. Do tourists value different levels of cycling infrastructure? Tourism Management 46:92-101. Hochmair, H. H., D. Zielstra, and P. Neis. 2015. Assessing the Completeness of Bicycle Trail and Lane Features in OpenStreetMap for the United States. Transactions in GIS 19:63- 81. Isaksson-Hellman, I. and J. Werneke. 2017. Special Issue Article: Cycling Safety: Detailed description of bicycle and passenger car collisions based on insurance claims. Safety Science 92:330-337. Jestico, B., T. Nelson, and M. Winters. 2016. Mapping ridership using crowdsourced cycling data. Journal of Transport Geography 52:90-97. Kil, N., T. V. Stein, S. M. Holland, and D. H. Anderson. 2012. Understanding place meanings in planning and managing the wildland–urban interface: The case of Florida trail hikers. Landscape and Urban Planning 107:370-379. 27 Krizek, K. J. and P.J. Johnson. 2006. Proximity to Trails and Retail: Effects on Urban Cycling and Walking. Journal of the American Planning Association 72:33-42. Polovitz Nickerson, N., J. Jorgenson, M. Berry, J. Kwenye, D. Kozel, and J. Schutz. 2014. Bicycle Tourism. Montana Business Quarterly 52:2-7. Mekuria, M., P. Furth, and H. Nixon 2012. Low Stress Bicycling Network Connectivity. Minnesota Transportation Institute 11-19. Wang, H., M. Palm, C. Chen, R. Vogt, and Y. Wang. 2016. Does bicycle network level of traffic stress (LTS) explain bicycle travel behavior? Mixed results from an Oregon case study. Journal of Transport Geography 57:8-18. Winters, M., M. Brauer, E. M. Setton, and K. Teschke. 2013. Mapping bikeability: a spatial tool to support sustainable travel. Environment & Planning B: Planning & Design 40:865- 883. Yiannakoulias, N., S. A. Bennet, and D. M. Scott. 2012. Mapping commuter cycling risk in urban areas. Accident Analysis and Prevention 45:164-172.