Shoal Lake Aqueduct
The Shoal Lake Aqueduct is a 155 km concrete conduit that delivers water from the Shoal Lake/Lake of the Woods watersheds to a reservoir in The City of Winnipeg. Like the aqueducts built by the early Romans, it is powered entirely by gravity. Built between 1914 and 1918, the conduit conveys water primarily as a covered open-channel flow conduit. There are seven river crossings where it is operated under pressure as an inverted siphon. The construction of the conduit and the post-construction operation of the system was accomplished by an adjoining railway built specifically for these purposes.
The aqueduct is known variously as the Shoal Lake Aqueduct and the Greater Winnipeg Water District (GWWD) Aqueduct. The GWWD was an inter-municipal corporation owned by eight, but not all, of the civic entities in the areas around the Red and Assiniboine Rivers. While the project was spearheaded by The City of Winnipeg, the corporation’s Board of Directors was not dominated by Winnipeg. The project impetus, as stated in the campaign promise of the 1913 elected mayor of Winnipeg, Thomas Deacon (an engineer), was to provide “at once for the people of Winnipeg an ample and permanent supply of pure soft water which will forever remove the menace now hanging over Winnipeg of a water famine”.
This article is confined to the engineering involved in the design and implementation of the aqueduct and not the social and political processes involved in the lead up to the approval of the project.
- 1 Initial Investigations
- 2 Major Features of the Aqueduct
- 3 Engineering & Design
- 4 Conclusion
- 5 Key Players
- 6 Supplemental Information Sources
- 7 References
The Shoal Lake Area
The source of water for the Winnipeg Aqueduct is Shoal Lake, a tributary of Lake of the Woods. The intake location on Shoal Lake is on Indian Bay on the lake’s western edge – about 3km west of the Manitoba-Ontario boundary (see Figure 1). The longer dimension of Indian Bay is east-west. On the south side is an east-west oriented promontory of land. The settlement of the members of the Ojibway First Nation that occupy Shoal Lake Indian Reserve No. 40 is located on that promontory. On its south side is another bay of Shoal Lake known as Snowshoe Bay. The narrowest portion of the land between the two bays is about 840m and is close to the western shore of Indian Bay. A stream known as the Falcon River discharges into Indian Bay immediately south of where the water for the aqueduct is withdrawn. The Falcon River is the outlet of Falcon Lake, some 10 km to the northwest, and also drains much of the muskeg area in between. The height of land forming the western boundary of Shoal Lake along the route of the aqueduct is 6km west of Indian Bay. The general layout of the area is shown in Figure 1. Note that the scheme involves a channel that diverts the water of the Falcon River away from Indian Bay thereby leaving the indigenous community of Shoal Lake Band 40 on an island with its only vehicular access being seasonal service by a ferry to the east side of the bay and an ice road in the winter months. Years of advocacy by the community protesting that isolation has attracted the attention of three levels of government (Winnipeg, Manitoba, Canada) and in 2016 there was agreement by the three to fund a 27 km all-season road, with three bridges, that will connect the community “to Canada” at the Trans-Canada Highway North-West of the intake.
In the run up to the decision to build the aqueduct The City of Winnipeg commissioned a 1913 report from a ‘Board’ of New York based consulting engineers, R. Herring, F. P. Stearns and J. H. Fuertes “on a water supply from Shoal Lake for the Greater Winnipeg Water District”. The report relied upon survey work undertaken by the City Engineer H.N. Ruttan in the winter of 1913 and provided a conceptual design and an estimate of the costs ($13,045,600). The report from Ruttan’s survey included topographical information about both a possible route and for the area around Indian Bay and Snowshoe Bay. Soundings were also taken in both bays. It also included for the first time a precise figure for the difference in elevation between Shoal Lake and the McPhillips Reservoir, namely 293.19 ft. (89.42 m).
As noted earlier, Shoal Lake is a tributary of Lake of the Woods. The feasibility of using Shoal Lake as a water source was enhanced by earlier hydrological interventions on Lake of the Woods. A history of the modifications to the water levels of Lake of the Woods as provided by the Shoal Lake Watershed Working Group reads as follows:
"Shoal Lake is connected to Lake of the Woods at a location known as Ash Rapids. Construction of a control dam at the outlet of Lake of the Woods in the 1880s raised the level of the lake by about a metre above its natural condition. In turn, this brought water levels in Shoal Lake into approximate balance with levels in the much larger Lake of the Woods, at least over an extended portion of the year. The channel at Ash Rapids was deepened and widened from its natural state, through blasting, around the turn of the century . This was reportedly done to provide a water based transportation route to serve both timber and mining operations in the Shoal Lake area."
An aerial perspective of the area surrounding the aqueduct intake is shown in Figure 1, and hydraulic data on Shoal Lake is provided in Table 1.
In testimony by three engineers and one hydrologist at the 1914 International Joint Commission hearing on the project, agreed that withdraw of the equivalent of years volume at 85,000,000 imperial gallons per day in a single day, i.e. with no inflow, the draw down of the level of Lake of the Woods would be less than 1.50 inches.
|Lake of the Woods||Shoal Lake|
|Aspect||sq. Miles (km2)||sq. Miles (km2)|
|Drainage Area||27 000 (69 000)||360 (930)|
|Surface Area||1250 (3200)||107 (280)|
|Ratio of Drainage to Surface Area||21.6||3.5|
Major Features of the Aqueduct
The overall scheme of the Winnipeg Aqueduct, following in the direction of flow, entailed:
- a soft water source that required no treatment for potability, colour, or hardness,
- a 2.4 km dike across a portion of Indian Bay and a 840 m channel excavated between Indian Bay and Snowshoe Bay to divert the water of the Falcon River,
- an intake structure on the edge of Indian Bay,
- a 155 km gravity fed enclosed conduit that conveys water, primarily in an unconfined channel, but with some portions under pressure, from the inlet to The City of Winnipeg’s McPhillips Street water reservoir,
- provision for an equalizing and storage reservoir (Deacon) approximately 21 km east of the McPhillips reservoir,
- metering facilities for the measurement of the volume of water flowing at vital points, and
- a railway that facilitated the initial construction and the on-going operation and maintenance of the system.
Features of the enclosed conduit include:
- a design capacity of 85,000,000 gpd (386,400,000 L/d) per day,
- cutting through the height of land that forms the boundary of the Shoal Lake watershed,
- provision for delivery of water into a future second conduit that could increase the combined design capacity to at least 100,000,000 gpd (454,600,00 L/d),
- seven rivers crossings by means of inverted siphons,
- a system for water and air pressure relief during operations,
- a means for inspection during partial operation and for isolation and dewatering of sections for maintenance, and
- maintained the integrity of local surface drainage systems.
Engineering & Design
Field Studies for Final Design
The GWWD had an in-house engineering department. The Chief Engineer was W.G. Chace, a McGill graduate who had also worked on the Pointe du Bois generating station.
When the October 1913 vote authorizing the project to proceed was settled, the GWWD engineers set about the detailed design aided by J. H. Fuertes (a co-author of the Board of Consulting Engineers report) as an ongoing consultant. The first order of business was for the GWWD engineers to determine the final route selection so that the right of way could be established and the railway started. Doing so required more precise and extensive survey information than had been provided by Ruttan. Survey parties were dispatched with one important task, that being to establish a precise set of benchmarks. During the winter of 1913-1914, the survey parties accomplished the following:
- 95 miles of precise levels,
- 362 miles of transit lines,
- 1,317 miles of levels, and
- 380 square miles of topography.
Additionally, some 12,000 ft of bore holes were made to assess the foundation conditions and determine the depths of muskeg. Anyone who has surveyed during a Manitoba winter with survey instruments of that era will recognize the accomplishment and appreciate the ordeal that those surveyors endured.
Billy Von Koen in his book Definition of the Engineering Method defines it as “the strategy for causing the best change in a poorly understood or uncertain situation within available resources.” There is little doubt that those engineers responsible for the implementation of the Winnipeg Aqueduct, in seeking to cause the best change, were faced with uncertain situations and finite resources. While they had access to the experiences of other aqueduct designers, they also had to deal with factors that were specific to the locality of the project and that were not well understood. There were two in particular. One was the issue of selecting the most all-round economical route for the conduit. The other was developing a design for the concrete mixture for use in the conduit utilizing the available local aggregates that would meet the necessary compressive strength, permeability, and durability requirements.
Designing for the Terrain
|Slope of Aqueduct||Dimension of Section|
|inches per 100 ft||(Height x Width)|
|0.11||10’-9” x 9’-0’’|
|0.279||10’-9” x 9’-0”|
|0.300||8’-9” x 7’-45/8’’|
|0.382||8’-31/4” x 7’-0’’|
|0.480||7’-111/2” x 6’-81/2’’|
|0.600||7’-71/2” x 6’-51/4’’|
|0.684||7’-51/2” x 6’-31/2’’|
|0.744||7’-4” x 6’-21/2’’|
|1.290||6’-7” x 5’-61/8’’|
|1.537||6’-41/2” x 5’-41/4’’|
The terrain between Shoal Lake and the prairie country just east of Winnipeg is treed, has a number of rivers, and contains some 80 km2 of muskeg or swamp. It is also for the most part uninhabited and did not lend itself to access or communication.
In approaching the route selection task, the GWWD engineers had available the preliminary design that was provided as part of the report from the Board of Consulting Engineers (Hering et al, 1913) that included details of typical conduit arch sections. With that information, and on-going refinements of the sections, curves were developed showing cost variations for typical aqueduct cross-sections based on depths of excavation and for a range of slopes. With that information available, field staff could make on-site decisions in choosing an alignment that would minimize the costs and the line length while striving to maintain the average slope of 0.57 ft per 1,000 ft of length. By that process, an alignment was established by March of 1914 on which over 30% of its length was very close to the average slope. The final average gradient was 0.62 ft per 1,000 ft at a length that was only 8% longer than the straight-line distance. Table 2 provides a listing of the slopes for each aqueduct cross section.
An early start to the railway was essential to the project, and once the route was decided, the right-of-way could be established and the railway construction could get underway. The right-of-way selected was generally 300 ft (91 m) wide with the railway located 40 ft (12 m) from the south boundary. At the easterly end, which had more construction challenges such as the depth of excavation, the width was increased to 500 ft (152 m).
Concrete Mix Design
The GWWD engineers were keenly aware of the effect that the amount of Portland cement used in a cubic yard of concrete would have on the cost of the project. Their awareness would have been heightened by the knowledge that cement in Canada cost 45% to 50% more than in the US or Great Britain. The cement component eventually selected for the aqueduct concrete, based on their testing program, was 430.5 lbs/cy (255 kg/m3) vs. 549.5 in the mixtures commonly used and recommended for watertight work by other authorities of the time. Chief Engineer Chace reported that the savings achieved by relying on the recommendations of the GWWD engineers were projected to be $350,000 for 400,000 cy (306,000 m3) of concrete. That savings figure was based on a 1915 cement cost of $0.0079 per pound. To put that in 2010 dollars, with cement then at $255 per metric tonne ($0.1158 per pound), the saving would be $5,500,000.
In achieving that economy, the GWWD engineers relied on their own methods. Knowing the sources of granular material available as due to the ongoing 1914 railway construction, and other exploratory work, they undertook an extensive program of analysis and testing of materials from those sources. The program consisted of five tests:
- a mechanical analysis of the aggregates from the available natural deposits,
- the weight per cubic foot of the various gradations and combinations of materials,
- volumetric tests of the materials for density,
- tests for compression and tension of various sand-cement mortar mixtures,
- tests for both permeability and compression of concrete with various mix proportions of stone and sand when selected by using the data derived from the other tests.
The testing program involved two lots of Portland cement, with the major difference between them being the time interval to final set after mixing. The cement used was manufactured locally by the Canada Cement Company Limited, which began production in Winnipeg in 1913. The specimens for the program were made from 28 distinct concrete mixes. There were 29 tests of permeability and 47 in compression.
The compression testing program, with specimens 8" (20 cm) in diameter and 16" (40 cm) long, seems to have followed standard procedures. However, the permeability test, if not unique, was at a minimum innovative. The concrete specimens were 13" (33 cm) in diameter and 14.5" (37 cm) long and cast with a small internal chamber connected to a metal injection pipe with an external water-stop. The test apparatus forced water into the chamber at a constant pressure with a gauge to measure the water entering the specimen with a means of measuring the amounts passing through the concrete, and a separate measure of any leakage from around the pipe used for the injection. A photo of the apparatus used in the test is shown in Figure 2.
In a 1917 paper, the GWWD concluded from their testing program, “that with lean mixtures and the gravel materials available, the addition of fine sand would give the work contemplated impenetrable concrete of the desired strength.” Their opinion was borne out by tests of the performance of the completed conduit. Chace also reported in that mixes adopted on the basis of those tests would develop “a strength of 2,800 pounds per square inch and a six-inch wall of concrete will be watertight against a hydrostatic pressure of 200 feet of head.”
Cross Section Design
While the work on the aggregate supply and the mix details was ongoing, the other step was to finalize the shape and dimensions of the conduit sections. As part of that process, test sections were built and loaded as shown in Figure 3. It is of interest to note that W.M. Scott, the contractor for the test sections, seems to have been the same W.M. Scott (an engineer) who sometime later became the Chief Commissioner of the GWWD. He was also later the President of the first Council (provisional) of the Association of Professional Engineers of the Province of Manitoba in 1920.
The cost significance of the volume of concrete to be incorporated in the project was also a priority. It was estimated that a one-inch increase in the sectional thickness would have cost $400,000.
The engineers were also cognizant of the durability of reinforcing steel should it become exposed to water through cracking of the concrete. With that in mind, they elected to use the unreinforced self-supporting arch type cross-section for the cut-and-cover portions that comprised most of the aqueduct. The arch rested on the edges of a previously cast invert section as seen in Figure 4. The invert had a circular concave upper surface with the radius of the curve somewhat greater than the height of the arch. As an example, the radius for a section with a 2.25 m interior height was 3.4 m. The term “invert” refers to the lowest point in the internal cross-section of an artificial channel, and is thought to have originated from describing an inverted arch. Ruttan et al comment on the function of an invert in the chosen system as follows:
"Although, as actually constructed, the side walls of the arch rest upon the ends of the invert, the greater part of the load, in a section like the one shown, may be assumed to be borne by those portions of the trench bottom directly beneath the side walls, because the central part of this relatively thin unreinforced type of invert is not considered to provide much in the way of support, but rather to serve merely as a firm water-tight bottom to the aqueduct."
These typically unreinforced arch sections were configured such that, under all loading conditions, the concrete in the arch would be in compression and thereby resistant to cracking. The situations where reinforcing steel was required in the arch section are as follows:
- at road crossings and undeveloped road allowances,
- for railway crossings, and
- in the arches where the weight of the backfill material was so light that there was a risk of deformation from outward ground water pressure and separation from the invert.
In each of those cases the concrete thickness was increased to protect the steel.
When the conduit crossed a river as a siphon, it was under internal pressure, and as the concrete would then be in hoop tension, those sections were reinforced with steel and the wall sections thickened for protection purposes. The same treatment was used for the 6.7 km cast-in-place pressure section east of the Deacon Reservoir location. The circular sections were also built in two castings with the edges of the invert section thicker than the nominal thickness in the upper portion. The reinforcing steel was circular and crossed the construction joint where the steel was lapped. Figure 5 shows one end of a circular section.
Aside from its cost, the use of reinforcing steel was an issue because at the time it did not come prefabricated. It was manufactured and delivered as a straight twisted square bar that had to be bent on site. Figure 6 shows an example of the bending process.
Falcon River Dike and Diversion
As previously noted the Falcon River drains much of the muskeg area west of Indian Bay, and had a brown colour. The dike and channel that diverted the water of the Falcon River to Snowshoe Bay was built to dilute that water with the much clearer, greater Shoal Lake water (see Figure 1). In that way before the diverted water could reach the intake of the aqueduct, it would have to make its way around the promontory and back into Indian Bay – a distance of 14 km. No doubt the opportunity of the diversion scheme was recognized because of the topographical work and soundings undertaken by City Engineer Ruttan’s staff in 1912 and early 1913. The alternative to the diversion would have been to extend the aqueduct considerably further into Shoal Lake so that it accessed unaffected water. The District’s cost for the dike and diversion work was $147,000. In a paper, Fuertes indicated that the cost to extend the aqueduct a further 8 km would have been $1,000,000. Figure 7 provides an indication of the scope of the dike construction.
The dike was built using the scow and bridge method. Significantly, as shown by the date on the photo, it was built before the GWWD railway reached Indian Bay. Figure 8 shows excavation of the diversion channel in 1915.
The designers located the intake structure in a rock outcrop on the shore of Indian Bay adjacent to the north end of the dike. Gathering dikes extend into the lake and a concrete structure in the rock cut controls the water entering the aqueduct. The structure includes the usual gates, trash screens, and stop log provisions that one might expect. Two features are noteworthy. The first is that the designers provided dual entrance chambers, each with its own screens and stop log facilities. In that way, one can be isolated for maintenance while the other was providing water to the aqueduct. The second feature had to do with preventing cold air from entering the system in the winter. The lower edge of the front wall of the structure, which is the top of the water opening, was constructed 1.9 m below the lake’s lowest level. In that way, it was below the bottom of the ice and cold air could not enter the system, thereby preventing the formation of ice in the intake.
The capacity of the intake structure was 85,000,000 gpd (386,400,000 L/d) at low water level. Chace suggested that at the high water level of Lake of the Woods established by the IJC in 1917, the intake could accommodate 100,000,000 gpd (454,600,000 L/d). Figure 9 shows the completed intake structure before flooding. The operating water level would be below the letters in the photo at a distance of approximately one and one-half times the height of those letters. To put the capacities of the intake and the aqueduct in perspective, the peak levels of water ever used by Winnipeg was 300,000,000 L/d (66,000,000 gpd) in 1988. That was just less than 500 litres per person per day. Since then, through the City’s water conservation program, that figure has been reduced, and in 2000, it was approximately 380 litres (84 gallons) per person per day. There is still plenty of unused capacity in the Winnipeg Aqueduct.
Red River Valley Siphon
From a point approximately 27 km east of Winnipeg (Mile 17) to the McPhillips Reservoir, the aqueduct is designed as an inverted siphon, which means that the entire length of conduit is under pressure.
This siphon was by no means of a constant configuration. A significant change point was at the site of the future Deacon Reservoir. The other change point was at the crossing of the Red River. The section from the east end, “Mile 17”, to the Deacon Reservoir is an 8' (2.4 m) diameter round cast-in-place reinforced concrete pipe. The section from the Deacon reservoir to the Red River is a 5'-6" (1.7 m) “Lock Joint” precast concrete pipe. The crossing of the Red River is a 5ft (1.5 m) diameter cast iron pipe, and the section from the Red River to the McPhillips Reservoir is a 4 ft (1.2 m) diameter Lock Joint precast concrete pipe. The section between Deacon and the Red River incorporated the service connections to supply some of the partners in the GWWD: Transcona, St. Vital, St. Boniface, and Fort Garry. The section west of the Red River included a service connection to supply the James Avenue High Pressure Pumping Station, eliminating the need for Red River water. While there was an overflow provision at the east end of the siphon (Mile 17), the only pressure relief facility in the entire siphon section itself, as initially constructed, was a surge tank with a weir on the east side of the Red River.
Red River Crossing
The crossing of the Red River is a siphon within a siphon. The conduit crosses the river in the limestone bedrock some 24m below the banks and 6m below the river bottom. The core drilling that established the depth and nature of the rock was one of the earlier contracts tendered by the District. One presumes that the engineers were comfortable with the resulting information as the reports of the Chief Engineer in the minutes of the Administration Board meetings do not mention concerns.
As noted, the conduit is a cast iron pipe. The configuration is a vertical section on each river bank built in a shaft and a horizontal section built in a tunnel in the rock. The 16m vertical shafts were 5 m in diameter and lined with a 600 mm reinforced concrete wall. The upper portion serves as housing for the valving system. The annular space for the portion below the valve house floor down to the bedrock was backfilled with gravel. The tunnel for the horizontal section was nominally 3 m by 3 m and the pipe was centred on that opening. The cast iron pipe sections were specially fabricated so that they could be caulked from within. The material used for the caulking was hemp and lead. Once the cast iron pipe had been finished the space between the rock and the pipe was filled with concrete. Remarkably, this section has functioned since 1918 without ever being dewatered.
Red River Siphon and Surge Tank
Perhaps the most vital component of the Red River siphon system is the surge tank located on the east side of the Red River adjacent to the river crossing on the corner of Tache Avenue and Rue Messager. It is also the most visible in that it stands the equivalent of a four-story building above the ground level. Figure 10 provides a view of its external structure in 2010.
As noted, the design of the tank provided the only pressure relief and overflow facility on the Red River siphon. That is significant because the inlet to the McPhillips Reservoir was controlled by valves which could have been inadvertently closed. Chace noted that “it must be kept in mind that there is (water) flowing constantly west of Mile 17 at considerable velocity a solid volume of water of huge weight. It is a serious matter to suddenly disturb the rate of flow of such a body of water.” As such, since the rate of flow out of the McPhillips reservoir would vary many times during the day, any excess arriving at the reservoir had to be accommodated. The engineers’ solution was that reinforced concrete surge tank designed to serve two purposes.
They were, overflow to relieve pressure caused by too much water entering the siphon at Mile 17, and to spill the excess that might be created at the entrance of the McPhillips Reservoir. To do that, a closed circular structure was built with a concentric internal circular weir. The lip of the weir was at a fixed elevation. In that way, it spilled the excess water from either or both of the two causes.
The supply line from Deacon enters the base of the structure in a chamber at the bottom of the internal weir and a separate line leaves the chamber to bring water to the cast iron line that crosses the Red River. Both these lines are under pressure with the maximum head determined by the lip elevation of the weir. That elevation was about 9m above the ground line or 14m above the centre line elevation of the two pipelines. To collect the water that spilled over the weir, a second concentric wall of the same height was built outside of the weir wall leaving an annular space of 760mm. The excess water collected in that space was then taken away by a drainage line that discharged into the nearby Red River. Primarily for protecting the system from frost and secondarily for aesthetic reasons, the wall was faced with brick as is shown in Figure 10. Significantly, there was no valve between the 1.7 m incoming line and the surge tank. In that way, the pressure in the siphon to the east of the Red River could not be inadvertently increased to the point where it would damage the line. The system was operated so that there was always some water flowing over the weir.
The structural features are also notable. Given the need to ensure that the pressure relief system would never have to be taken out of service, structural integrity was important. Unlike with the conduit, it seems to have been designed on the principle of “no risk.” The base of the structure was supported on a series of caissons under the walls, excavated to the bedrock. On top of the caissons were grillages made of steel beams and four steel beams spanned from grillage to grillage around the base. The entire system was then encased in concrete. The concrete too was heavily reinforced once again using twisted square bars like those in some of the main aqueduct sections. This time, however, the drawings included bending diagrams for the reinforcing steel.
The engineers, administrators, and contractors on the Winnipeg Aqueduct executed a unique project that is remarkable for its scope and its lasting ability to serve the needs of The City of Winnipeg. In so doing they were confronted with and overcame unique physical and environmental conditions using creative design, testing and construction processes.
The project was completed on time and quite close to budget by December of 1918. However, because of a concern that there might be an adverse effect on the industrial boilers in the City due to the change from hard to soft water during the height of the heating season, the changeover was delayed. Water started to flow into the McPhillips Reservoir on March 29, 1919.
- Charles S. Slichter – consulting engineer
- James H. Fuertes – consulting engineer (New York)
- Hugh A. Robson (Judge) – Public Works Commissioner, Province of Manitoba
- James H. Ashdown (Mayor of Winnipeg)
- Thomas Russell Deacon (Mayor of Winnipeg)
- William G. Chace – Chief Engineer, GWWD
- Douglas L. McLean – Assistant to Chief Engineer, GWWD
- Henry N. Ruttan – Winnipeg City Engineer – later a consultant; also General and Commanding Officer of Military District NO. 10
- John G. Sullivan – CPR – also a consultant
Supplemental Information Sources
The lead up to the concept and the political process that brought the aqueduct about is documented in the thesis “Developing a domestic water supply for Winnipeg from Shoal Lake and Lake of the Woods: the Greater Winnipeg Water District Aqueduct, 1905 – 1919” http://mspace.lib.umanitoba.ca/handle/1993/4457, and in the Manitoba Historical Society’s Journal article Pressure to Act: The Shoal Lake Aqueduct and the Greater Winnipeg Water District (No. 72, Spring-Summer 2013) http://www.mhs.mb.ca/docs/mb_history/72/aqueduct.shtml.
Additional and more detailed engineering information is provided at http://mspace.lib.umanitoba.ca/handle/1993/4457 and in the Manitoba Historical Society’s Journal article Not All Down Hill From There: The Shoal Lake Aqueduct and the Greater Winnipeg Water District (No. 75 Summer 2014) http://www.mhs.mb.ca/docs/mb_history/75/aqueduct.shtml
- Canadian Public Works Association Manitoba Chapter (CPWA). 2000. Greater Winnipeg Water District’s Shoal Lake Aqueduct Nomination for the American Public Works Association’s Top Ten Public Works Projects of the 20th Century. Winnipeg MB: Self Published.
- Chace W.G., M.V. Sauer. 1917a. The aqueduct for the Greater Winnipeg Water District. Transactions of the Canadian Society of Civil Engineers. Volume XXX Part II: 375-395.
- Chace W.G., D.L. McLean. 1917b. Studies regarding concrete mixtures employed in the work: Supplement to the aqueduct for the Greater Winnipeg Water District. Transactions of the Canadian Society of Civil Engineers. Volume XXX Part II: 396-424.
- Chace, W.G. 1917c. Sands and consistency of concrete. The Canadian Engineer 33: 282
- Chace, W.G. Chace. 1920a. Construction features of the water works of the Greater Winnipeg Water District. Journal of the American Water Works Association, Volume VII: 931-949.
- Chace, W.G. 1920b. Notes and Instructions for the Guidance of the Staff Responsible for the Care and Operation of Aqueduct Supplying Water to the Communities Comprising the Greater Winnipeg Water District. Winnipeg MB: The City of Winnipeg Water and Waste Department.
- Cherney, B. 2009. 90th anniversary of Shoal Lake aqueduct — first phase was construction of GWWD Railway. Winnipeg Real Estate News, October 16, 2009. www.winnipegrealestatenews.com/Editorials.aspx?id=968 (2010/11/29)
- Fuertes, James H. 1920. The basic principles used in the designs for the new water supply works of Winnipeg Manitoba. Journal of the American Water Works Association, Volume VII: 693-748.
- Greater Winnipeg Water District. 1918. Aqueduct Construction Scheme: What it Is What It Means. Winnipeg MB: Telegram Job Printers Ltd.
- Herring, R., F. P. Stearns and J. H. Fuertes. 1913. Report on a water supply from Shoal Lake for the Greater Winnipeg Water District. New York NY: Stovel Company Limited Winnipeg.
- Landon, C.S. 1918. The Greater Winnipeg Water District. The Canadian Engineer. 34: 290-301, 315-317.
- Library and Archives Canada, Canada. n.d. Mining Lands & Yukon Branch of the Department of the Interior, File 116619
- Mclean, D.L. 1919. Earth Slides in Winnipeg Aqueduct Construction. The Canadian Engineer. 37: 469-471.
- Prodan C.S. 1979. The building of the Winnipeg Aqueduct. Manitoba Pageant. www.mhs.mb.ca/docs/pageant/24/winnipegaqueduct.shtml
- Ruttan, H.N. 1909. Winnipeg’s water works: Visit to the City water works High Pressure Plant and artesian well system. British Association for the Advancement of Science, 1-8. Winnipeg, MB.
- Ruttan, H.N. 1913. Untitled report to the Mayor and City Council of Winnipeg dated May 8, 1913. Winnipeg MB: The City of Winnipeg Water and Waste Department.
- Ruttan, H.N., J.G. Sullivan and R.S. Lea. 1916. Report on the Aqueduct of the Greater Winnipeg Water District by the Special Board of Consulting Engineers. Winnipeg MB: The City of Winnipeg Water and Waste Department.
- Scott, W. M. 1938. The Water Supply of Greater Winnipeg. Journal of the American Water Works Association, Vol. 30 No. 11: 1872-1885. Winnipeg MB: The City of Winnipeg Water and Waste Department.
- Scott, W. M. n.d. The Water Supply of The Greater Winnipeg Water District. Unbound paper labelled as presented at a meeting of the American Water Works Association in March, 1938. Winnipeg MB: The City of Winnipeg Water and Waste Department.
- Shoal Lake Watershed Working Group. 2002. Shoal Lake Watershed Management Plan. Manitoba Water Stewardship, Winnipeg MB. www.gov.mb.ca/waterstewardship/water_quality/quality/shoal_lake_wsmp.html (2010/12/28)
- Shropshire, L. 1994 (January 7). Mayor Battles Critics. Winnipeg Real Estate News. Winnipeg MB: self published.
- Siamandas, G. n.d. Winnipeg’s Shoal Lake Aqueduct. timemachine.siamandas.com/PAGES/winnipeg_stories/SHOAL_LAKE_AQUEDUCT.htm (2011/02/7)
- Slichter, C.S. 1912. The Water Supply of the City of Winnipeg to Public Utilities Commissioner, Province of Manitoba, Winnipeg, MB.
- The Canadian Engineer. 1917. Winnipeg Aqueduct Excavation. The Canadian Engineer. 32: 149-151.
- University of Manitoba – MSpace - Developing a domestic water supply for Winnipeg from Shoal Lake and Lake of the Woods: the Greater Winnipeg Water District Aqueduct, 1905 – 1919 http://mspace.lib.umanitoba.ca/handle/1993/4457 Accessed December 17, 2016.
- Manitoba History No. 72, Spring-Summer 2013. Pressure to Act: The Shoal Lake Aqueduct and the Greater Winnipeg Water District http://www.mhs.mb.ca/docs/mb_history/72/aqueduct.shtml Accessed December 19, 2016
- Manitoba History No. 75, Summer 2014. Not All Down Hill From There: The Shoal Lake Aqueduct and the Greater Winnipeg Water District http://www.mhs.mb.ca/docs/mb_history/75/aqueduct.shtml Accessed December 19, 2016
- Aqueduct: Colonialism, Resources, and the Histories We Remember, Adele Perry; ARP Books, Winnipeg Manitoba. http://arpbooks.org/books/detail/aqueduct Accessed December 19, 2016
- City of Winnipeg - http://www.winnipeg.ca/waterandwaste/water/shoalLake.stm Accessed December 18, 2016
- City of Winnipeg - http://www.winnipeg.ca/waterandwaste/dept/railway.stm Accessed December 18, 2016
- City of Winnipeg - http://winnipeg.ca/waterandwaste/water/freedomRoad.stm
- The Canadian Society for Civil Engineering (through Ryerson University) http://www.ryerson.ca/~amackenz/history/national/w.html Accessed December 17, 2016.