Florida Watershed Journal — Low-Impact Development
Change Language:
Infiltration Gallery: Conventional Design, Contemporary Performance Evaluation

Mike Gregory, P.E.
AECOM Kitchener,


The Alton Mill is located in Caledon, Canada, on a tributary of the Credit River that discharges into Lake Ontario. Although located in a cold weather climate, the concepts and analytical methods described in this paper are relevant and applicable to locations in Florida where soil and groundwater conditions are conducive to infiltration. The Alton Mill Initiative is aimed at converting a century-old plus heritage stone mill and pond complex into an artisan-oriented tourism destination. In 2006, The Alton Development Inc. submitted rezoning and site plan applications for this initiative, proposing future building additions, parking areas, and driveway alterations to service adjacent property owners.

A stormwater management servicing plan was developed to lessen the impact of this development on the adjacent properties and receiving watercourse. The servicing plan featured an infiltration gallery and was included as part of the submittal for site plan approval and permitting in October 2006. Construction of the stormwater facilities was completed in September 2008. While no formal postconstruction monitoring was required as part of the permitting process, additional analysis was undertaken in January 2010, including continuous hydrologic simulation to validate the long-term expected performance of the infiltration gallery.


The site is located in an industrial land use zone and has a total area of 1.07 ha (2.64 ac). Figure 1 shows the proposed stormwater management system layout in the southern portion of the property.The receiving watercourse, Shaws Creek, runs to the east (left to right) just below the bottom of Figure 1.

Roof runoff from a portion of the main building is collected by a roof drain that is discharged to the mill race within the building. The mill race discharges to a culvert outlet in Shaws Creek at the eastern end of the building. Before development, stormwater runoff from the remainder of the site discharged by overland sheet flow to the southeast directly into Shaws Creek. The Alton Mill Initiative added new impervious areas and re-grading of the existing terrain. Proposed stormwater management facilities were sized to accommodate runoff for ultimate build-out conditions in future phases, as shown in Figure 1.

A recent version of the StormWater Management Model (SWMM) was used to represent the hydrology and hydraulics of the site and stormwater facilities (U.S. EPA, 2009). The design objective was to preserve the natural infiltration features of the underlying soil, facilitated by the collection and conveyance of surface runoff into a central infiltration gallery. The gallery was designed to temporarily store surface runoff for subsequent percolation and groundwater recharge.

Development of the stormwater servicing plan followed a traditional design approach to achieve water quantity control objectives using event-based hydrology. That is, facilities were designed such that postdevelopment peak flows, flood stages, and velocities did not exceed predevelopment conditions for a range of prescribed design storm events.

A system of swales, catchbasin inlets, and storm sewer pipes was designed to direct surface water runoff into the infiltration gallery.The configuration and sizes of pipes, manholes, and catchbasins (as shown in Figure 1) were determined according to the municipal design standards (Town of Caledon, 2006). Storm sewer pipes represented in the model ranged in size from 250 mm (10 in.) To 300 mm (12 in.) In diameter with a total length of 122 m (400 ft). Past soil Investigation studies by others in July 1998 and June 2006 were based on boreholes and test pits drilled on the site and confirmed that native subsoils are suitable for infiltration facilities.

Profile views of the infiltration gallery design are shown in Figure

2. Proposed gallery dimensions were 14 m (46 ft) by 14 m (46 ft) at the bottom (i.e., 407.3 m elevation) with 1:1 sideslopes to a top gallery elevation of 408.8 m. Runoff is directed into the gallery via 300 mm (12 in.) Header pipes at each end and distributed evenly through a manifold of seven 150 mm (6 in.) Perforated pipes. The gallery was filled with a continuous layer of 50 mm (2 in.) Clear stone, wrapped in filter cloth, and covered with native soil material. A void ratio of 30% was used to determine the available gallery storage volume.

Surface storage was represented in the model in order to track surface flooding depths for the larger rainfall events. Further, surface overflow channels were represented in the model to account for overflows when the conveyance capacity of the storm sewer pipes was exceeded and the computed hydraulic gradeline exceeded the ground elevation at any location within the collection system. These overflows were designed to be discharged via overland sheet flow into Shaws Creek within acceptable tolerances of depth and velocity specified by the local permitting agencies.

During construction of the gallery in June 2008, bedrock was discovered at an approximate elevation of 407.0 m (note: the bottom of the gallery was initially proposed to be 406.5 m). Coincidently, five test pits were dug in the vicinity of the proposed gallery location during the 2006 soil investigation, and none revealed the presence of bedrock. Due to the bedrock, the size and elevation of the gallery was redesigned with the final orientation as shown in Figure 1.

Based on the 2006 soil investigation, a saturated hydraulic conductivity of 2.1×10-4 m/s (30 in/hr) was used in the design. This results in a peak infiltration rate of 0.025 m3/s (0.9 cfs) through the sides of the gallery. Infiltration through the bottom of the gallery was assumed to be zero due to the underlying bedrock.

The following rating curve was used in the model to represent the rate of infiltration discharged from the gallery (with linear interpolation assumed between values):

• No infiltration when the gallery is dry (i.e., hydraulic gradeline is at or below the gallery bottom elevation of 407.3 m);

• Two-thirds of the peak infiltration capacity, 0.017 m3/s (0.6 cfs) when the hydraulic gradeline reaches an elevation of 407.8 m; and

• Peak infiltration capacity of 0.025 m3/s (0.9 cfs) when the hydraulic gradeline exceeds the gallery top elevation of 408.8 m. Modeling results for proposed site conditions indicated that stormwater runoff for all design storms up to and including the 25-year return period event will be contained onsite and dispersed through the proposed infiltration gallery without surface discharge to the receiving watercourse. Further, the peak hydraulic gradeline remained below the top of the infiltration gallery (i.e., elevation 408.8 m) for all design storms up to and including the 10-year return period event.

A detailed analysis of the structural impact of increased groundwater levels on basement floor slabs in adjacent buildings was not conducted. Peak stage results were compared to the critical threshold elevation of the main building floor slab (408.37 m), located approximately 30 m (100 ft) from the infiltration gallery. It was also confirmed that the peak hydraulic gradeline remained below basement floor slab elevations in all adjacent buildings for all design storms up to the 5-year return period event. Nevertheless, it was decided to install floor drains in the main building basement as a precaution.


Construction of the stormwater facilities began in May 2008 and was completed in September 2008. Figure 3 includes photos taken during construction of the infiltration gallery in June and July 2008.The top left photo shows the assembly of the header manifold and perforated distribution pipes. The top right photo shows the installation of the porous media on top of the distribution pipes. The bottom left photo shows the placement of the filter cloth on top of the infiltration gallery. The bottom left photo shows the placement of native soil material over the filter cloth.

The servicing plan included temporary erosion and sedimentation controls during the construction period to protect the porous media in the infiltration gallery as well as limit the impact on adjacent downstream properties and the receiving watercourse. Given the high cost of replacing the porous media in the gallery, it was critical to incorporate permanent sediment/silt controls. Structures CB2, CB3, CB5, DICB6, and CBMH7 were equipped with 600 mm (24 in.) Sumps for sediment capture. Further, hooded outlet covers were installed in these structures to restrict sediment and silt from entering the pipes.

Although these controls were designed to limit the entry of fine sediment into the infiltration gallery, this facility essentially acts as a filter and will eventually become clogged. A scenario was simulated that represented a failure of the infiltration gallery under frozen or clogged media conditions, in which there was no infiltration. Further, the available storage within the gallery was greatly reduced in this failure scenario (i.e., 5% void ratio in the porous media). Model results confirmed the resulting overflow depths and velocities were within acceptable limits for the 100-year design storm event.

LONG-TERM PERFORMANCE EVALUATION With larger stormwater facilities, permitting agencies typically require a formal post-construction monitoring program to confirm the facility achieves its design objectives. Such a monitoring program was not required as part of the permitting process for this project, given the relatively small size and limited impact of future land use activities. No stormwater overflows from the collection system have been observed since the gallery was constructed over two years ago. In spite of this, a long-term performance evaluation was conducted following construction as a matter of due diligence.

As noted earlier, stormwater facilities were designed using a traditional event-based hydrology approach, which is most appropriate for flood control where the design objective is to collect and dispose of runoff in a manner that protects public safety and minimizes property damage during extreme events. Water conservation was an additional goal for the Alton Mill Initiative, where the design objective is to capture and retain runoff in a manner that preserves the natural hydrologic balance onsite. Water balance controls are implicit in Low-Impact Development (LID) design, intended to manage runoff volumes in an attempt to maintain predevelopment hydrologic conditions.

Continuous hydrologic simulation requires long-term meteorological input that encompasses a range of historical rainfall events as well as the dry weather periods in between, not just selected design storms that characterize extreme events. Continuous simulation is therefore most appropriate for LID facilities, such as the Alton Mill infiltration gallery. Other design objectives such as maintaining baseflows in receiving watercourses, groundwater recharge/discharge, water supply/wellhead protection, and natural habitat impacts also require a long-term water balance approach using long-term (i.e., multiple year) continuous simulation.

For this project, additional local rainfall data was compiled and continuous hydrologic simulation applied as a means of validating the long-term expected performance of the infiltration gallery. The new rainfall record included 19 years of tipping bucket rainfall data from a more hydrologically representative gage location (in a rural area closer to the site) compared to the urban airport weather station that was used to derive the local standard design storm events used in the original design model.

The performance evaluation included a variety of hydraulic indicators to compare against pre-development conditions, including:

• Rate of surface runoff;

• Volumetric proportions of various hydrologic components (i.e., evapotranspiration, infiltration, and surface runoff);

• Peak stage in relation to various threshold elevations (i.e., ground surface and basement floor slab in adjacent buildings);

• Peak routed flow rates, particularly surface overflows to the receiving watercourse (i.e., Shaws Creek) and groundwater infiltration discharged from the gallery;

• Peak routed runoff volumes, particularly surface overflows and infiltration from the gallery;

• Flow frequency curves quantifying discharge into the receiving watercourse; and

• Peak depth and velocity of surface overflows in relation to a variety of flood damage/risk envelopes.

It was beyond the scope of this paper to present results for all of the performance indicators listed above. Results that highlight differences between event-based and continuous simulation are described.

Figure 4 shows a comparison between peak routed flow rates, representing discharge to Shaws Creek under proposed conditions. To present a fair comparison, design storm hyetographs were developed from the same 19- year rainfall record that was used for continuous simulation. The peak intensities and total volumes for each design storm (i.e., 2-, 5-, 10-, and 25-year return period events) were significantly higher (greater than 25%) than the corresponding values in the design storm events used in the original design model. As a result, the apparent performance of the infiltration gallery using the design storm events prescribed in the local design standards was much better than that predicted using design storm events derived from a more hydrologically representative rain gage. This is evidenced by the surface overflows shown in Figure 4. In the original design model, runoff from the 25-year return period event was contained on-site.However, with the new design event hyetographs, discharge into the receiving watercourse was noted for the 2-year event (0.08 m3/s [2.8 cfs] as plotted in Figure 4).

With continuous simulation, it is possible to plot the complete flow frequency curve and this is included in Figure 4. The majority of Runoff producing events (over 2,220 in total) featured flow rates less than 0.02 m3/s (0.7 cfs), representing runoff from the small subcatchment that directly discharges into Shaws Creek. For the 19-year period of record, there were only 6 events that exceeded a flow rate of 0.02 m3/s (0.7 cfs), representing surface overflows when the capacity of the ollection system or infiltration gallery was exceeded.This uneven distribution of events explains the smooth frequency curve below a 2-year return period and the less elegant curve at higher return periods.

Comparing the event-based and continuous simulation results in Figure 4 suggests a wide discrepancy in inferences about the performance of the infiltration gallery. While the 5-year return period event closely matches continuous simulation results, the 2-year return period event significantly overpredicts flows and the 10-year event significantly underpredicts flows.

Event-based and continuous simulation results were also compared with respect to the distribution of surface runoff disposal volumes. The chart in Figure 5 shows results for the 5-year return period event, where approximately one-third of the surface runoff volume was discharged into the receiving watercourse and two-thirds to groundwater via the infiltration gallery. The chart in Figure 6 shows continuous simulation results and indicates a much improved performance in terms of onsite volume control, with over 90% of the surface water discharged through the infiltration gallery.


The sizing of stormwater management facilities across North America largely relies on event-based hydrologic modeling methods, which continue to be prescribed in local stormwater design guidelines.These traditional methods originated in times when flood control and peak flow management was the sole design objective. With the increasing popularity in LID and green infrastructure initiatives, contemporary stormwater management is now concerned with water balance as an additional design objective that would maximize surface runoff volume reductions onsite. Event-based methods cannot be used to sufficiently assess the long-term volume control performance of such facilities.

This conclusion extends beyond small-scale source controls and can be applied to stormwater management facilities in general.
Continuous simulation offers a greater diagnostic tool for assessing hydraulic performance compared to event-based modeling, since it can describe the full range of runoff response characteristics, versus a snapshot of selected individual return period events. As a consequence, in lieu of monitoring, only continuous simulation can be used to assess the long-term average performance of a stormwater facility. Further, results describing the duration of runoff (e.g., flood inundation) or quantifying the various hydrologic components in a water balance calculation are more meaningful with continuous simulation. With event-based simulation, the assessment of duration is only useful as a relative measure.

Ultimately however, the key consideration in stormwater design is minimizing the risks to public safety and property damage. It is common practice to assess the hydraulic performance of stormwater facilities based on controlling peak flow rates, and in some cases runoff volume, to pre-development conditions. Neither one of these variables can accurately quantify the associated flood risk or property damage potential: the depth and velocity of flow are more appropriate indicators. A true assessment of risk can only be achieved by integrating these risk-based indicators with the full range of flow frequencies computed using continuous simulation.


• Town of Caledon (2006). Development Standards, Policies & Guidelines document (Version 3). Public Works & Engineering Department.

• U.S. Environmental Protection Agency (2009). Storm Water Management Model User’s Manual, Version 5.0. National Risk Management Research Laboratory, Office of Research and Development; Publication EPA/600/R-05/040. Cincinnati, OH.