LIGHT COVERAGE ON COLBY COLLEGE CAMPUS

Meghan Cornwall ’11 and Francis Gassert ’11

ES212: Introduction to GIS and Remote Sensing

Environmental Studies Program, Colby College

Abstract

Outdoor lighting is an essential component for nighttime safety on college campuses. Outdoor lighting on Colby College campus is not uniform, leaving some areas with minimal light. The purpose of this project is to analytically evaluate if and where major walkways on Colby College have inadequate lighting. While we were unable to define inadequate lighting, we found that most paths on campus do appear to be well lit, while in general open spaces on campus have the lowest light levels.

Introduction

Campus safety is a major concern to all members of a college community. Lighting is a critical aspect of nighttime safety, which is of highest importance on commonly used walkways. However, there are numerous costs to lighting, including electrical consumption and light pollution. In addition, low light areas may be strategically placed to constrain movement. The Colby Physical Plant Department designed and manages the outdoor light scheme on Colby College. Every lamp type has an array of qualities to best fit its purpose. Each of these factors affects the decision of the placement and type of lamps on campus. Because of the complexity of designing a lighting system, this project focuses on the current outdoor light levels.

More specifically, the purpose of this project was to ask if and where major walkways on Colby College campus are inadequately lit. We used GIS to analyze outdoor light patterns on campus, including buildings and elevation differences that may block light.

Methods

We used ESRI’s ArcGIS software to perform the analysis. Our analysis has two main questions: first, where does each lamp’s light hit; and second, how much is the ground illuminated? To answer the first question, we used a viewshed analysis. Viewshed uses an elevation map to determine what can be seen by an observer at a given location. We combined the digital elevation maps of Colby College with a map of the buildings on campus, retrieved from the Maine Office of GIS and Physical Plant Department, respectively. We used their most current map of Colby. However, the recent renovations to the Cotter Union, Pierce and Perkins-Wilson buildings were not included in the data so we added them manually from aerial photographs. The building data did not include elevations. Yet, because no lamps in our analysis are higher than buildings, the actual building elevation does not matter so long as they act as blocking structures for the viewshed. Therefore, we assumed the buildings to all be an arbitrary height of 60 feet. We then retrieved an AutoCAD drawing of all of the outdoor lamps on campus from the Physical Plant Department, and used this data to perform the viewshed analysis. Figure 1 demonstrates the output of a single lamp’s viewshed.

Next, we determined how much each lamp lights up the area around it. Illuminance, the amount of light hitting a flat surface from a single light source, is determined by the following formula:

^[1]

Where i is illuminance, l is luminous intensity of the light source, d is the distance from the light source and q is the difference between the normal angle of the surface and the angle of the light. Since the surfaces of people and objects are complex, not flat, and surface properties affect the diffusion of light, we chose to eliminate the angle coefficient. Assuming that any lit surface is normal to the light source (q =0) effectively simplified the equation to:

.

Luminous intensity is measured in candela; however, the Physical Plant Department data only stated each lamp type, bulb type, and power. We used this information to look up the light output for an industry equivalent bulb of the given type and power. Unfortunately industry standards only reports lumens. Lumens measure the total light output whereas candela measures the output in any given direction. To simplify the conversion between lumens and candela we made two assumptions. First, we assumed that the light bulbs emit light evenly in all directions; second, we assumed that lamp housings are completely reflective, such that light is emitted evenly in a cone. If this is true, we can divide lumens by the measure of the cone of emittance in steradians to get candela. The measure of steradians in a cone of apex angle 2q is determined with the formula^[1]. Then we determined distance using a three dimensional distance formula (demonstrated in Figure 2). This gives us how much light the lamp spreads at any distance (Figure 3).

Because we needed a viewshed and illuminance analysis for each lamp, we decided to use a Python module to perform the analysis rather than manually processing all 322 lamps. The module performed a viewshed for each lamp. Next, it calculated illuminance for all areas hit by the lamp by taking the distance from each lamp and then applying the formula discussed above. The module then sums the illuminance outputs resulting in an aggregate light map of campus.

Finally, we used an aerial photograph of campus to highlight the major walkways. We took the light map outputted by the python module, and overlaid it with our path data to compare the light levels and major walkways.

Results

Figure 4 shows the darkest 10% of paths on campus, and highlights areas below an equivalent illuminance threshold of 2.76 Lux. From the resulting light map we identified the five focus areas over the dark paths, labeled a-e. Figure 5 includes an aerial photograph of campus.

Discussion

This study shows the illuminance on campus based on outdoor light posts only. Much of the campus is also lit by building mounted lamps or indoor lighting. Furthermore, this study does not include trees and bushes. Therefore, when interpreting the light map, one can expect that areas near buildings are actually brighter than portrayed, while other areas may be darker. Additionally, when estimating light’s affect on safety, open clearings require less light. Area a, the academic quad, is minimally lit, but its light level may still be considered safe because it is an open field. Area b and d are dimly lit in our analysis, but they are actually additionally lit by lights mounted on the surrounding buildings. Area c is also partially lit by building mounted lights, though it did not appear as bright through empirical observation. Area e leads to the athletic fields, so when the fields are in use, the field lights light up the pathway. Most of the other highlighted paths were also near buildings. Furthermore, we were unable to determine a definition for inadequate and adequate light levels; the map merely shows what is lighter and what is darker. It is important to take all these factors into consideration when reading the map.

We did not include trees and building mounted lights primarily because of data availability. In addition, modeling trees’ affect on light is difficult due to their complex shape and seasonal variation. Nonetheless, the model is built in such a way that additional light sources can be easily included, and if a simple model for trees was created, trees could be included as additional blocking structures.

Taking into consideration its assumptions and limitations, this model can be used to assist in placement of future lights. In addition to identifying which areas on campus are the darkest, it can be used to identify areas where there may be more light than necessary.

Conclusion

Although we were unable to define inadequate lighting, we successfully identified lightest and darkest areas on campus based on our available data. From this we identified five main dark areas, usually in open spaces. In addition, many other dark areas were near buildings where mounted lights and/or indoor lights would shine. Because of the complexity of designing a lighting system, this project focuses on the current outdoor light levels. It is not intended to dictate lighting design rather to be used as a reference tool.

References

^[1]Angelo, Joseph A. The Dictionary of Space technology, 2^nd edition. Facts on File: New York. 1999.

^[2]Buylighting.com <www.buylighting.com>. 2009. 4/12/2009.

^[3]Colby College Physical Plant Department, Personal Interview. 4/3/2009.

Acknowledgements

We would like to acknowledge the following people. First of all we would like to thank Gordon Cheesman and Andy Gockel with the Physical Plant Department at Colby for providing the lamp data including the AutoCAD drawings and lamp attributes. We would like to thank Manny Gimond and Philip Nyhus for all of their guidance with the GIS. We would also like to thank the Oak Foundation.

Figures:

Figures 1, 2, and 3. Demonstrations of viewshed, distance, and illuminance calculations on a single lamp, respectively.

Figure 4. Illuminance map highlighting the 10% darkest paths on campus and areas below an equivalent illuminance threshold (2.76 Lux). The five focus areas are shown.

Figure 5. Outdoor light levels overlaying an aerial photograph.