Final Project

Keywords: suitability analysis, slope, polygon width, bluespot analysis, sun shade volume, ArcMap, ArcScene

1.0 Introduction & Background

The development of urban areas effects natural cycles and processes. Two cycles in particular, the hydrological cycle and the carbon cycle, are affected by the removal of vegetation that occurs during urban development. The hydrological cycle is directly impacted by the changes in groundcover that come with urban development, such as the conversion of vegetative cover to parking lots, roads, and rooftops (EPA, 2015). In a natural, undeveloped watershed, rainfall soaks into vegetation covered soils and is filtered and cleaned through natural processes. In contrast, in a conventional urban setting rainfall is typically unable to infiltrate through pavement or cannot reach the ground due to rooftops. This results in an altered hydrological cycle with the effect of increased flooding and runoff. In addition, water passing over these surfaces picks up pollutants that are then carried into nearby waterbodies, or into the ground water. The removal of vegetation during development also impacts the carbon cycle. Trees and other vegetation naturally uptake carbon dioxide (CO₂) and produce oxygen (O₂) necessary for life on earth to continue (Bonan, 2008). They remediate pollution from air, water, and soil through their life cycle, contribute to reduction of the heat island effect, improve storm water management, provide habitats for animals, and deliver myriad other ecosystem services (Mitchell, 2012).

As our population continues to increase, urban development puts increasing pressure on the remaining natural terrain. But is there a way to minimize the negative effects? Low Impact Development (LID) is a development paradigm that seeks to mimic the natural functions of a watershed, interrupting hydrological and carbon cycles as little as possible through maintenance of vegetation and the retention of water, encouraging it to infiltrate into the ground rather than allowing it to run off. The benefits of LID are social, environmental, and economic, and include improved water quality, reduction of flooding events, restoration of aquatic habitats, improved groundwater recharge, and enhancement of the urban landscape (EPA, 2015). There are many methods employed to this end. The following proposal describes a project focused on two of these – the maximization of permeable surfaces, and the support of healthy vegetative growth.

Though most urban plans include trees, the urban setting is not ideal for most large plant life. A natural tree growth pattern is shown in Figure 1. This figure demonstrates that a large portion of tree root growth is shallow and lateral. In addition, for healthy growth, a typical tree requires 120 ft³ of soil when the crown has a 10ft diameter, 500 ft³ when the crown is 21ft diameter, and 1000 ft³ with a 30 ft crown diameter, as illustrated in Figure 2 (Mitchell, 2012).

Lateral root growth and the required soil volume are not supported by a conventional tree pit, shown in Figures 3-6. In a conventional tree pit, the roots have a small radius in which they can freely penetrate the soil. However, the surrounding soil tends to be so compacted that roots are not able to move through the soil, so the tree cannot anchor itself or access enough water and nutrients to grow to its full height. This often contributes to unhealthy growth, early death, or trees collapsing under wind force due to insufficient anchoring. It can also lead trees to seek space by growing roots upwards, leading to costly infrastructure damage. Research indicates that a tree with a diameter at breast height (DBH) of 30 inches removes 70 times the pollution as a tree with a 3-inch diameter (Nowak, 1995). In addition, a mature tree can increase the life of the pavement under its canopy by more than 10 years (Stolte, 2016). This suggests that directing resources into ensuring that urban trees reach maturity is a worthwhile cause.

The current project focuses on a specific LID method designed specifically to support the growth of mature trees in an urban setting: the bio-infiltration tree pit. This technology supports healthy tree growth, while providing passive filtration for rainfall that would otherwise contribute to urban runoff. In these systems, shown in Figure 7, pavement is suspended on top of interlocking cells which are filled with quality soil. Tree roots are able to freely pass between the spaces in the cells, spreading out in a way that mimics unfettered root growth in nature. Many municipalities are recognizing the value of these systems as a way to support healthy tree growth and the myriad accompanying ecological services (Stolte, 2016).

2.0 Context

Dr. Grant Clark’s Ecological Engineering Lab at Macdonald Campus is engaged in a LID project, in which suspended pavement tree-pits are being built in Montreal’s Hochelaga neighborhood. The lab wishes to identify potential locations for bio-infiltration tree-pits in the neighborhood, based on a GIS analysis. The analysis will include a slope analysis, an identification of blue spots (low-lying areas where water tends to gather), and the identification of sidewalk segments that are wide enough to support the tree pit installation. A preliminary assessment of the typical shade patterns during a summers day is performed, based on developed 3D models of Hochelaga buildings and ArcScene tools. The shade assessment will be based on an existing study by Norris (2014), “Made in the shade: using GIS to model pedestrian shade in Austin, Texas”.

3.0 Project objectives and questions

The three main questions to be answered through the project are presented below:

1. Which areas in Hochelaga are suitable for the construction of suspended pavement tree-pits?

2. What infiltration rates can be expected, based on the soil types under the identified suitable areas?**

3. How much sunlight can be expected in suitable areas? Can this information be employed to choose the best tree species for each area?

4.0 Location

Hochelaga, Montreal, QC. Coordinates: 45.57525°N 73.53325°W

The study takes place in the Hochelaga neighborhood of Montreal (see Figure 9). The neighborhood was named for an Iroquoian village, as historians believed an ancient village may have been located there (Trigger, 1976). Hochelaga is an urban neighborhood in the east of the city, part of the Mercier-Hochelaga-Maisonneuve borough. Hochelaga is one of the five most populated boroughs in Montreal. Approximately 129,000 Montrealers, 7.1% of Montreal’s population, live in the region; a mix of working-class Quebecois, students, and recent immigrants (Hochelaga.ca, 2019). Economically, the region paints a complicated picture. In the past the region was an industrial hub. Periods of booms and busts resulted in a neighborhood characterized by one of the most low-income populations in Montreal, while more recent gentrification is changing the landscape. Currently, Hochelaga is considered a lively and up-and-coming neighborhood, and is a good candidate for enhancement with LID development installations (Hochelaga.ca, 2019).

5.0 Data

The base data used for this project can be grouped into three main categories. Elevation data, infrastructure data, and geographical data.

1. Elevation Data --> HRDEM Can Elevation Series - 1m (Open Canada, 2018)

a. DTM - Digital Terrain Model

b. DSM - Digital Surface Model

2. Infrastructure Data

a. Sidewalk Data --> Roadway - T12 - active Sidewalk (Montreal Open Data Portal, 2019)

b. Building Footprints --> Basic Cartographic Compilation 2017 (Montreal Open Data Portal, 2017)

c. Montreal Road Shapefiles --> NRN (National Road Network) (Open Canada, 2019)

3. Geographical Data

a. Hochelaga Boundary File --> ArcGIS Online - Montreal Borough Boundaries (ArcGIS.com, 2016).

6.0 Methodology

In order to answer the study questions, two main explorations were carried out: 1) Suitability Analysis, 2) Shade Analysis. Before these tasks could be executed, data pre-processing was done, consisting of projecting files to UTM 18N, and clipping them to the Hochelaga boundary. A detailed methodology is provided below.

6.1 Data Pre-processing

1. Project all files to UTM 18N
2. Clip files to boundary

6.2 Find Suitable Location for Tree Pits

6.2.1 Calculate Slope

Though the sidewalks themselves were assumed to be relatively level, sidewalks adjacent to areas with steeper slope were considered preferable as the steeper slope would encourage water accumulation in the tree pits, enhancing their impact in the LID context. Slopes of 20-70 degrees were considered optimal.

1. Slope Tool
   • Input(s): DTM
   • Output(s): Slope Raster

6.2.2 Find Sidewalk Segment Widths

The bioinfiltration tree pits require an area of 2m (length) x 1.5 m (width) . Given that sidewalks must be wide enough to support the installation of the tree pit while remaining functional for pedestrians, only sidewalks with widths of of 3m or more were considered suitable for installation. It is assumed that this will allow enough space for foot-traffic. The following steps were performed to determine sidewalk width, and create a shapefile of suitable sidewalks. A snapshot of sidewalks, ranked by width, is shown in Figure 11.

1. Make centrelines (used Polygon to Centerline Tool from Dilts, T.E. 2015) --> 20 hours to process!
2. Convert polygon to line file
   • Input(s): Sidewalk polygon file
   • Output(s): Sidewalk polygons as lines
3. Near tool - to calculate the distance from the centerline to the nearest edge of the sidewalk line file
   • Input(s) centerline, sidewalk line file
   • Output(s): centerline file with added attributes, including near distance, and nearest polygon FID
4. Calculate Width
   • Add new field to centerline attribute table --> Width
   • Field Calculator --> Width = 2*near_distance (since near_distance is distance from centerline to nearest edge of polygon, 2*near_distance is a suitable estimate of the width of the polygon)
   • Output(s): centerline file with Width field
5. Select segments with suitable widths
   • Select by attribute: calculator --> width >= 3.00
   • Output(s): segments of centerline contained within sidewalk polygons with widths >= 3.00m
6. Select by location
   • Source Layer: Centerline contained within sidewalk polygons with widths >= 3.00m
   • Target Layer: Sidewalk File
   • Match Option: contain source layer feature
   • Output(s): sidewalk shapefile comprised of only sidewalk polygon segments >= 3.00 m in width
7. Spatial Join
   • Target Features: Output from Select by Location (sidewalk shapefile comprised of only sidewalk polygon segments >= 3.00 m in width)
   • Join Features: Output of Select by Attribute (Centreline file comprised of centreline segments contained within polygons of widths >= 3.00 m)
   • Match Option: Contains (since centerline segments are contained within polygons)
   • Join: One to Many
   • Output(s): sidewalk shapefile containing only sidewalk segments with widths >= 3.00 m

6.2.3 Identify Low Lying Areas (Blue-spots)

The bioinfiltration tree pits are designed to support the natural hydrological cycle by allowing for infiltration of rainwater and runoff. Placing the pits in areas likely to accumulate rainwater enhances their effectiveness. As such, bluespots were identified based on the DEM of the study area. Bluespots are areas that will accumulate water, and, depending on the permeability and saturation of the surface, may be at risk for flooding. Sidewalks that intersected bluespots were considered optimal for tree-pit installation. The Bluespot Identification workflow was adapted from an ESRI tutorial, Find Areas at Risk of Flooding in a Cloudburst (ESRI, 2017).

1. Fill
   • This tool was run twice
     • Run 1: A conditioned DEM was created by removing possible data errors, represented by sinks. Filling sinks does not distort the landscape, but provides a more accurate hydrologic model by smoothing over values that are less than the vertical accuracy of the data (ESRI, 2017).
       • Input(s): DTM
       • Output(s): Conditioned DTM
     • Run 2: A filled DEM is produced, by filling all remaining sinks to their pour point
       • Input(s): Conditioned DTM
       • Output(s): Filled DTM

2. Minus

• Subtracts values in the conditioned DTM from the original DTM to give a bluespot raster showing the locations and depths of low-lying areas that are likely to accumulate water
  • Input(s): Conditioned DTM, Filled DTM
  • Output(s): Bluespot depths (cell by cell) raster

3. Con

• Used to create a two-value raster. Bluespot cells (Value >0) are given a value of 1, and non-bluespot cells (Value < 0) are given a value of NoData.
  • Input(s): Bluespot raster
  • Output(s): Preliminary Bluespots

4. Region Group

• Groups bluespot cells into separate regions by location - if cells are directly next to each other or diagonal, they are considered to be part of the same region.
  • Input(s): Preliminary Bluespots (Raster)
  • Output(s): Bluespots with unique IDs (Raster)

5. Raster to Polygon

• Input(s): Bluespots with unique IDs (Raster)
• Output(s): Bluespots Polygon

6. Dissolve

• Input(s): Bluespots Polygon
• Output(s): Bluespots

7. Flow Direction

• Calculates direction of water flow on the surface
  • Input(s): Conditioned DTM
  • Output(s): Flow Direction Raster

8. Watershed

• Identifies areas that contribute flow to a particular bluespot
  • Input(s): Flow Direction Raster, Bluespots with unique IDs
  • Output(s): Watersheds

These outputs are joined, and a bluespot volume field is calculated to determine the amount of water that a particular bluespot will likely accumulate. This final feature class, shown in Figure 13, is used in the following suitability analysis.

6.2.4 Compile outputs to identify suitable sites

This final step integrates the results from the slope, sidewalk width, and bluespot identification steps to identify segments of sidewalks that are suitable for bioinfiltration tree pit installation.

1. Polygon to Raster
   • Convert sidewalk shapefile of widths > 3m (output from Spatial Join) to raster using the Polygon to Raster tool
2. Reclass Rasters
   • For analysis with Weighted Overlay tool, rasters must have the same number of classes. The slope raster was reclassified from 1-10 based on natural jenks. The sidewalk raster was reclassified from 1-10 on an equal interval basis. The Bluespot shapefile was subjected to the Euclidean Distance tool, and the output raster was manually reclassified into 10 classes ranging from 0-2m, 2-5m, 5-10m, 10-25m, 25-50m, 50-75m, 75-100m, 100-1000m, and 1000-the last value.
3. Weighted Overlay

Input 1: Reclassified slope raster -- > 40% influence

• slope values from 20-70 are given a value of 10, 0-10 are ranked as 4, 10-20 are ranked as 5, and values above 70 are given a rank of 8.

Input 2: Reclassified sidewalk raster --> since the sidewalks with suitable widths have already been selected, width is considered to have some, but not much bearing on the outcome --> 10% influence

• sidewalks from 3-4 m in width ranked as 9, above 4m ranked as 10.

Input 3: Reclassified Blue-spot Euclidean Distance Raster --> 50% influence

Output(s): Suitable Areas Raster (see Figure 14).

6.3 Shade Analysis

6.3.1 Creating the 3D Landscape

ArcScene was used to create a 3D landscape for the shade analysis.

1. First, a TIN was created using the DEMs listed above. The TIN contains information about surface texture that can be used to create a realistic 3D representation of the scene.

2. Next, the sidewalk and road shapefiles were draped over the TIN image, to show the relationship between these features and the terrain. The results of this action are presented in Figure 15.

3. In order to translate the building footprint shapefile to the 3D landscape, a height element had to be added to the building footprint attribute table. This was done using the following method, adapted from Obtaining elevation information for building footprints (ArcGIS, 2019a):

3.1 Create Random Points:

• to generate accurate height data, several sample points were created for each building footprint using the Create Random Points tool, with the Building Footprint Shapefile as the input. Up to 20 points were generated for each footprint, and saved in a new Sample Points file.

3.2 Use DSM data to define the height of the buildings:

• the DSM (Digital Surface Model) data contains information about the earths surface and all objects on it, including buildings and trees. In the current project, the elevation of the buildings was extracted from the DSM files, and joined to the building footprint attribute table
• using the Add Surface Information tool the Z value, or elevation value from the DSM layer was added to the Sample Points file.
• an average of the data from each sample point was developed to provide accurate height information. This was done with the Summary Statistics geoprocessing tool. This tool generates a table containing the given statistics, associated with the fields selected from the input file. The Sample Points file was used as the Input table, and the Case Field CID was selected, so the table could be joined with the Building Footprints file.
• the output table was joined to the Building Footprints file using Join - the estimated height of each building was now a field in the attribute table

4. Import into ArcScene

1. Once the building footprints were associated with their relative elevations, they could be imported into ArcScene for the 3D analysis. The imported buildings were draped over the TIN file, which is an important part of 3D analysis. This creates a base surface for the Building Footprints, and is taken into account during the 3D extrusion process, so that the effects of terrain are negated and the building heights are extruded accurately in relation to the terrain.

5. Creating 3D Buildings

The method used to extrude the Building Footprints is as follows:

5.1. In Layer Properties , under Extrusion, the option to Extrude features in layer is checked. The extrusion value or expression is calculated as [Mean_Z] (the field name of our height attribute). The option to Apply extrusion by using at a value that features are extruded to ensures that the buildings are extruded to their proper height relative to the terrain, and are not artificially tall. **note: ArcScene operates in meters, so the default assumption is that all input data is in meters. Data inputs in other units must be converted. For example if [Mean_Z] was in feet, the extrusion expression would be [Mean_Z]*3.28, to convert feet to meters.**

Figure 16 shows a closeup of the 3D buildings extruded using the above method. The heights of several buildings were checked using the information tool, and the values were, on average, as expected. Validation through obtaining height information for various buildings in the area would improve the robustness of the results. Note that some of the buildings appear to have been merged during the extrusion process. This was not corrected, due to the large number of buildings in the study area, but this will have bearing on the results, and should be considered. It is possible to separate buildings manually - this could be done on a smaller study area once the ideal locations have been identified through the suitability analysis.

2. The extruded buildings were converted to 3D Multipatch Objects using the Layer 3D to Featureclass tool.

6.3.2 Sun Shadow Volume

The Sun Shadow Volume tool requires extensive processing power and takes several hours to run with the given 3D Building Feature Class as an input. The following section reports on the process, and the challenges encountered when working with this tool and in interpreting the results.

1. Setting parameters

The Sun Shadow Volume tool was used to estimate the shadow impact on the study area for a single day, from sunrise to sunset. In theory, the tool could be run for the entire growing season, which would allow for an estimate of the seasonal shadow impact in the suitable areas, to promote the choice of trees with required sun or shade tolerance. However, the processing power required for this approach proved to be too substantial, and the program crashed several times.

The first day of summer (June 21st) was chosen to illustrate the utility of the tool. On June 21st, sunrise is at 5:05 AM, and sunset is at 7:35 pm. The tool was configured to account for daylight savings time, and to give shadows at 4 hour intervals between sunrise and sunset. The time zone was set for UTC 7:00 Mountain Time.

Output

The tool, as configured, produced output for 5AM, 9AM, 1PM, and 5PM. The initial output of the Sun Shadow Analysis are shown in Figures 17 - 20.

2. Analysis

In order to analyze the amount of shade produced by the shadow volumes, this data was manipulated to a form that could be imported into ArcMap. The study “Made in the shade: using GIS to model pedestrian shade in Austin, Texas” (Norris, 2014) was followed to perform this task, though some adaptations were made for the current project.

1. Produce 3D version of study area (sidewalks)

1.1 In order to work with sidewalks in 3D, the original polygon file first had to be simplified. The Dissolve tool was employed to remove unnecessary lines, to improve functionality within the 3D ArcScene environment.

1.2 Sidewalks were extruded by 1.0 m and converted to a multipatch using the Layer to 3D Feature Class tool in order to provide a layer on which to sum the shadow volumes. The Enclose tool was run on the sidewalk multipatch to ensure that the feature would be compatible with the tools used in the following steps.

2. Prepare the shapefiles for shadow calculations

2.1 A field called shadow was added to the original 2D sidewalk shapefile in its extruded form and calculated as zero using the field calculator.

2.2 This extruded shapefile was converted to a raster using the Feature to Raster tool. This raster will eventually hold the shadow values, and is called Zero Raster.

2.3 Values for each hour of the Sun Shadow Volume output are selected in the attribute table, and exported as individual feature classes, one for 5:00AM shadows, one for 9:00AM shadows, another for 1:00PM shadows, and the last for the 5:00PM shadows.

3. Create the hourly shadow maps

3.1 The Intersect 3D tool is used on each shadow map to determine the intersection between the 3D sidewalk shapefile and the 3D shadow volume.

3.2 The Multipatch Footprint tool is used to obtain a 2D rendering of the intersection --> these polygon shapefiles are called fiveAM_Footprint, nineAM_Footprint, etc.**

3.3. A field called "shadow" is added to the Multipatch Footprint outputs, and set to 1 using the Field Calculator.

3.4 Feature to Raster is used to convert the Footprint shapefiles to a Footprint raster that contains values of 1 for shaded areas, and then the Con tool is applied using the Zero Raster as the Input Conditional Raster and the Input True Raster, and the Footprint raster as the Input False Raster. This is done to create a common extent on which to layer the shadows. The output of this tool is the Shadow Raster Con.

3.5 Using the Is Null tool on Shadow Raster Con, all NODATA values are converted to zero. This output is called Shadow Raster Is Null.

3.6 Again the Con tool is used with the Shadow Raster Is Null as the Input Conditional Raster, Zero Raster as the Input True Raster, and Shadow Raster Con as the Input False Raster. This output represents the shadow map, and can be imported to ArcMap and visualized in 2D.

3.7 (Optional) The raster can be converted to a polygon shapefile for ease of manipulation in ArcMap 2D environment using the Raster to Polygon tool.

The above process can be repeated for each of the hourly shadows, to produce the shadow map results.

7.0 Results

7.1 Sidewalk Suitability Analysis

The results of the sidewalk suitability analysis, which took into account sidewalk width, slope, and areas with high potential to accumulate water, are presented in Figure 21. Figure 22 shows the streets adjacent to the suitable sidewalks.

7.2 Shade Analysis

Due to insufficient computer processing power, it was not possible to produce output for each of the four hours results of the Sun Shadow Volume tool. In the end, it was only possible to produce outputs for the 5AM analysis. I believe that this was due, primarily, to the amount of detail present in the sidewalk shapefile, and RAM limitations. Tool run-times in ArcScene often exceeded 5 hours, even on the computers available at the GIC (Geographic Information Center), and often terminated early, without completing. It was difficult to obtain results, especially if an error occurred, given the GICs opening hours. However, the 5AM shadow map can be used as proof of concept - given more time and increased access to high power computers, shadow maps can be produced for any hour of the day, any day of the year, using the method detailed above. Figure 23 shows the results of the method for the 5AM shadow volume, transferred from the 3D ArcScene environment to the 2D ArcMap environment, and configured for viewing. Yellow patches reflect areas where no shadow intercepts the sidewalk at 5AM, while grey patches reflect shadow at this time.

8.0 Discussion

8.1 Suitability Analysis

The majority of the suitable sidewalks are clustered in the North and South residential areas of Hochelaga. The South-East tip has a high concentration of suitable areas. Based on these results, it is recommended that the identification of installation locations be focused in this area. A closeup of this region is shown in Figure 24. Saint Catherine East, between Avenue Bourbonniere and Avenue William-David, and Rue Adam, between Rue Nicole and Avenue Desjardines appear to be the most suitable streets. These results should be validated by a site visit to support their applicability.

8.2 Shade Analysis

Ideally, the data shown in Figure 23 above would have been produced for each of the Sun Shade Volume tool outputs, so that relative shading maps were available for each hour of the day. These maps would have been overlayed, and the relative shade of the sidewalks in the areas identified as suitable would have been calculated using field and raster calculators. This would have supported planing, as trees with more or less sun tolerance could have been selected depending on the estimate of the shadows effect at a particular location. With a more powerful computer, or a simplified study area, this analysis would have been possible. In retrospect, using the Suitability Analysis to identify the Optimal Planting Areas, then extracting ONLY the Optimal Area to use in the Sun Shade Analysis would have been a more intelligent way to approach this problem, given the struggles encountered when processing data in ArcScene (see 9.0 Limitations and Challenges for further discussion)

9.0 Limitations and Challenges

9.1 Working with extremely detailed files

Sidewalks were central to this project. The sidewalk shapefile from Open Montreal was extremely detailed and valuable. However, the amount of detail contained in the file presented problems when running certain tools. It was difficult to determine sidewalk width, as tools that would normally be employed to find the width of roads and rivers did not run properly when provided with the sidewalk polygon shapefiles, some of which were less than 0.5 m wide. In the 3D environment, the processing power required to analyze the relatively small sidewalk segments was huge -- many tools took more than 3 hours to complete.

9.2 Working in 3D

While the tools used for the initial Suitability Analysis were familiar, ArcScene was an entirely new environment. Working with 3D data has a special set of challenges. Though I haven't had extensive experience with other types of data in 3D, I believe that working at high resolution made navigation in 3D especially difficult. For example, some of the sidewalk polygons used had boundaries of less than 0.5 m. Edges that don't meet and uneven boundaries make the application of many tools difficult. The building footprints occasionally did not line up with the sidewalk shapefiles in ArcScene (see Figure 25). Some buildings, when extruded, merged into single long buildings (see Figure 26). These errors, when compounded, effected the validity and accuracy of the model results. In addition, several pre-processing steps were required to ready data for use with the 3D tools, and even so, tool run-times were extremely long for each tool, which contributed to delays in project advancement.

Tip 1: When working in 3D in ArcScene or ArcGlobe, complex polygons can cause serious issues. Using the dissolve tool when possible to remove unnecessary polygon boundaries can improve processing speeds in these environments. This also reduces errors when extruding in 3D. Using the Enclose tool on extruded 3D multipatch files is also important, before running tools like 3D Intersect. Enclose reduces multipatch edges and encloses open ends of 3D objects to minimize errors when running 3D tools. After applying both Dissolve and Enclose on my sidewalk shapefile, my results improved.

Tip 2: It would have saved a considerable amount of time if I had identified a smaller area to work with in ArcMap, and clipped my files to a smaller extent before transferring my data to ArcScene. In the future, I will try to work with the minimum extent in ArcScene to improve processing times.

9.3 Data Validation

Several project steps, including width calculations and extruding the heights of the buildings in the study area, were completed using assumptions, estimations, and simplifications. The heights of the buildings and the widths of the sidewalks were verified with the information and measurement tools, and seemed realistic. However, more direct validation, through a site visit, for example, is recommended, before applying the results of the model generated in this project. A site visit could be used to provide information for further adjustment of the model, to improve its accuracy.

9.4 Naming conventions

Some tools (though not all) do not allow for output tables to be created if a number is the first character of a name. Figure 27 shows an example of the 3D Intersect Tool that failed after 2 hrs and 53 minutes because I had chosen to name the output table 5PMIntersect. The same tool was executed successfully with an output specified as FivePMIntersect. You can save a lot of time if you stick to a standard naming convention that avoids the use of numbers as the first character, along with spaces (as other classmates have mentioned unpredictable tool results when using names with spaces), or longer than 13 characters (an ArcGIS limit).

10.0 Conclusion

The results from this study suggest that bio-infiltration tree pit installation would have the greatest impact if the installations were focused in the South-East region of Hochelaga, on Saint-Catherines Street East, and on Rue Adam. Recommended next steps are to visit the identified sites and determine if other obstacles exist that may make installation difficult.

11.0 References

12.0 Initial proposal

Here's a link to the proposal of this project.

Created for Advanced GIS for Natural Resource Management, in the McGill University Department of Natural Resource Sciences, Professor Jeffrey Cardille