Rain gardens are a popular green infrastructure choice. But until recently, not much was known about their actual performance. That knowledge could help public and private entities decide where to use them, how to design and maintain them, and how to budget for them.
Setting out to address that challenge, Hey and Associates of Chicago, IL–in conjunction with the Center for Neighborhood Technology (CNT) in Chicago–has conducted numerous monitoring studies of rain gardens and developed a simplified method the firm believes municipalities can use and afford.
CNT’s mission is to show urban communities in Chicago and nationwide how to develop more sustainably through researching, promoting, and implementing solutions to improve the economy and the environment through use of existing resources, community assets, and the restoration of the health of natural systems. Chicago, with its large areas of impermeable surface, is a blueprint for other communities for green infrastructure approaches.
In a report titled “Monitoring and Documenting the Performance of Stormwater Best Management Practices” prepared for the Illinois Sustainable Technology Center, CNT and Hey and Associates concluded that rain gardens are effective under certain circumstances.
Rain gardens are no more important than any other stormwater best management practice (BMP) but are one of several tools, says Tim Pollowy, RLA, ASLA, senior landscape architect for Hey and Associates.
“It’s still an emerging technology,” he says. “With interest in all things green and sustainable development, a lot of times marketing has blurred the science and engineering behind these things.”
Pollowy says he’s attended conferences where he’s asked about research relating to native plants without receiving a satisfactory answer.
“Show me where somebody did infiltration tests in the exact same soil under the exact same rainfall conditions,” he says. “Everyone wants to build these rain gardens, but there’s not a lot of information out there about how they’re performing.”
Hey and Associates has been involved in multiple rain garden projects in Aurora, IL. One, through the Chicago Metropolitan Agency for Planning (CMAP), bundles smaller projects and works with several municipalities to implement Section 319 grants from the Illinois EPA.
One of those projects, the Spring Street rain gardens, was featured in the July/August 2011 issue of Stormwater magazine. A second project entailed a rain garden detention basin at the Lincoln Avenue Metra Commuter Rail parking lot. A third project on the east side of town near the Fox River involved the construction of a diversion structure in a storm sewer to divert low flows into a constructed rain garden.
Following those three projects, Aurora obtained another grant with American Recovery and Reinvestment Act stimulus funds to create a project similar to Spring Street at McCarty Park in 2010.
Aurora–Illinois’ second-largest city–also is an older city with some of the older neighborhoods still having combined sewers.
“Aurora is in the process of trying to separate those to the extent they can, so both Spring Street and McCarty Park rain gardens were done in association with sewer decombination projects,” says Pollowy. “Not only did they take the stormwater out of the combined sewer and put it into a separate storm sewer, they also built rain gardens to keep stormwater from getting into the storm sewer in the first place. It has multiple benefits on multiple levels.”
After the McCarty Park project, Aurora received one of the first Illinois Green Infrastructure Grants (IGIG) through a program operated by the Illinois EPA. The project centers on Downer Place in downtown Aurora. Downer Place is an older street between two bridges that are being replaced. Rain gardens will be part of a streetscape enhancement project to make the area look more attractive for businesses, Pollowy says.
He notes that rain gardens are typically used in conjunction with traditional infrastructure rather than replacing it. “You can often build a rain garden for half of what it costs to build a storm sewer,” he says. “Doing it in addition to traditional infrastructure gets the value-added component. Maybe research will help, but generally you don’t get a lot of stormwater volume credit for building BMPs. It’s not as if building a rain garden necessarily means your detention basin or storm sewers can be half the size. Sometimes you can get some credit for it. But generally it’s done more for the perceived water-quality benefits.
“If possible, you try to get some water infiltrating into the ground to keep it out of the storm sewers as a qualitative flood control measure,” he continues. “It’s hard to quantify how much is actually going into the ground instead of going into the pipe. The feeling is even if some goes into the ground instead of a pipe, that’s better, and it’s helping in flood-prone areas. The big issue is the water quality.”
The genesis for engaging in monitoring rain gardens occurred in 2007 when CNT asked Hey and Associates to develop a plan and a strategy to help get a handle on whether rain gardens work, how they work, and when they need maintenance, says Deanna Doohaluk, a water resources planner with Hey and Associates.
She says the questions led to more questions: “What is the limiting factor in rain garden performance? Is it topsoil? Is it the engineered soil? Is it the surface storage capacity of the rain garden? How do they function? From there, we developed a monitoring program.”
Jeff Wickenkamp, vice president of Hey and Associates, echoes Doohaluk’s sentiments that rain garden performance poses many interesting questions. One of the areas under examination was volume control–reducing the volume of water that reaches sewers.
“Between developing the city of Chicago’s stormwater volume control requirements, advising other communities on their stormwater regulations, and various technical advisory committees we were on, there was always a curiosity about knowing as much as we could about rain gardens,” he says.
“Some of the things we were trying to write into the regulations were how far away does seasonal groundwater need to be from the bottom of the facility before it reduces the effectiveness?” says Wickenkamp. “What is the minimum porosity of the underlying soils? Is effectiveness reduced at the native subsoils? On the top, are engineered soils required, and how helpful are they?”
In the Chicago region, a typically specified engineered soil mix is a combination of sand, topsoil, and compost for the topmost layer to help with infiltration.
“We wanted to understand how those facilities were performing and in particular how to grant credit for them,” says Wickenkamp. “If some of the storage is underground, some of the storage is in the porosity of the soil, and some is in the surface storage volume, should you give credit for all of those? Should you give them credit for only the surface storage and assume that it later migrates into the porous space? That’s what the original tests were designed to investigate.”
Wickenkamp says what his team found in the monitored facilities was that engineered soils that were properly placed did not limit the infiltration capacity. Water very readily moved downward into the underlying basins.
Those who install rain gardens do so on the basic premise that by creating a depression in the landscape, water will be retained and this will benefit the area’s overall stormwater management by reducing peak flow and total volume and by encouraging infiltration, Wickenkamp says.
“There are several different reasons why people might put in rain gardens,” he adds. “People say they want to put in a rain garden to encourage infiltration. By that, do they mean they wish to recharge groundwater, or simply reduce the peak rate and volume of stormwater runoff?
“It’s an intellectually curious question about infiltration. Are these devices really infiltrating water and leading to groundwater recharge, or are they just reducing stormwater runoff? We more typically come at it from the approach of reduced stormwmater runoff, even though people have multiple goals for improving water quality, the performance of the sewer system, restoring natural hydrology of the stream, or trying to preserve some of the natural hydrologic characteristics of a waterway during development.”
Most people have trusted the common sense and basic math on rain gardens, Wickencamp says. “But questions remain: Are these really better than a centralized facility? Are they easier to manage or maintain going forward? Should they primarily be encouraged as an add-on to enhance the overall performance of the traditional stormwater management methods? These are critical questions for stormwater managers.”
Developing a Monitoring Approach
In developing its monitoring approach, Hey and Associates first started with high-intensity data collection with expensive equipment at only a few sites. The firm’s first monitoring effort, at the request of CNT, centered on two churches that had been experiencing recurring flooding and at which rain gardens were constructed. The first site is Our Lady Gate of Heaven Parish, where contractor Joseph Jackson battered through 18 inches of steel slag to build a bioswale.
At the other site, St. Margaret Mary Parish, the bioswale and rain gardens were placed at the front door of the rectory. Students at the Catholic school worked to plant and care for the bioswale. Permeable pavement patches relieved the church properties of decades-long frequent flooding.
At the St. Margaret Mary site, the bioswale measures 900 square feet with a drainage area of 8,625 square feet. The depth of excavation is 2.25 feet, the depth of amended soil is 1 to 2 feet, and vegetation includes 412 plants of 16 species.
The east rain garden measures 96 square feet. The drainage area of the roof is 600 square feet. The depth of excavation of the rain garden is 1 foot, the depth of amended soil is up to 1 foot, and vegetation includes 54 plants of six species.
The west rain garden measures 96 square feet; the drainage area of the roof is 600 square feet. The depth of excavation is 2 feet, the depth of amended soil is 1 to 2 feet, and the vegetation consists of 54 plants of six species.
The bioswale and two rain gardens at St. Margaret Mary church and school were monitored to measure the ability of green infrastructure to reduce stormwater runoff in urban neighborhoods. The monitoring system was designed to measure the volume of water that was captured in each BMP and the volume that overflowed from each.
The goal was to assess the performance of the bioswale and two rain gardens, including the surface and subsoil infiltration rates, storage and infiltration capacity, and expected annual performance of the BMPs. A quality assurance project plan was developed for the sampling program at St. Margaret Mary and provides a comprehensive discussion on the sampling methods, equipment used, and quality assurance/quality control practices associated with the monitoring program.
A Hobo Weather Station Range Gauge Smart Sensor from Onset was mounted to a weather mast on the roof of the school building to obtain site-specific rain data on a five-minute interval. Data were downloaded from the rain gauge at least once every 30 days between November 29, 2007, and June 16, 2008, and between September 18, 2008, to October 31, 2008, into a Microsoft Excel spreadsheet. The lapse in data collection from June 16, 2008, and September 18, 2008, was caused by water damage to a connector in the wire between the rain gauge to the data recorder. Tests were conducted in that time period to identify the cause of the problem.
The site-specific rainfall data, in conjunction with the tributary area computed from the design plans, field observation, and field survey, were used to develop the inflow portion of a water budget.
Also in the bioswale and two rain gardens, two groundwater-monitoring wells equipped with a water-level meter were installed. Well #1 is 36 inches long and installed to a depth of 18 inches below ground surface. It was slotted from approximately 3 inches below ground surface to approximately 6 inches above ground surface. The water-level meter was installed just below ground level and was used to measure the amount of water stored above ground surface in the BMP.
Well #2 is approximately 66 inches long and was installed to a depth of 48 inches below ground surface. Well #2 was slotted from approximately 12 inches below ground surface to approximately 48 inches below ground surface. A water-level meter was installed within the slotted area and used to measure the amount of water stored in the engineered soils and subgrade portions of the BMP. With the exception of Well #1 in the west rain garden, water-level data were collected between November 29, 2007, and December 7, 2007, and between April 4, 2008, and October 23, 2008. Water-level data were collected from Well #1 in the west rain garden between November 29, 2007, and December 7, 2007, and between April 4, 2008, and September 18, 2008. Afterward, the meter was removed from the rain garden after it was hit by a vehicle.
The water-level meters were configured to record data on five-minute intervals, and the data were downloaded to Microsoft Excel at least once every 30 days.
The BMPs were surveyed to develop topographic contours over the surface of the bioswale and rain gardens and to obtain elevations of the monitoring wells.
A Hobo Weather S-SMA Station Soil Moisture Smart Sensor was installed at a depth of approximately 4 inches in the engineered soil portion of each of the BMPs, configured to record data on a five-minute interval. Soil moisture data were collected between November 29, 2007, and October 31, 2007. All soil moisture data were managed using Microsoft Excel.
The data were used to determine the pre-rain soil saturation conditions of the BMPs and to evaluate the in situ porosity of the surface soils. The monitoring equipment was installed in November 2007, but no storm events occurred after the date of installation. Following winter, monitoring resumed on April 4, 2008, and concluded on October 23, 2008, for the bioswale and east rain garden.
To determine the water-quantity performance of the rain gardens, the monitoring data were broken into discrete events. The initial screening of rainfall data looked at events that exceeded 0.05 inch. Twenty-nine events were identified during the monitoring period for the bioswale and east rain garden, and 23 events were identified during the monitoring period for the west rain garden.
For each of the BMPs, the volume of total runoff was computed by multiplying the rainfall by the tributary area to the BMP.
The tributary area to the bioswale is the area of the adjacent parking lot that drains to the bioswale. The tributary area to each of the rain gardens is a portion of the rectory’s roof. The monitoring system was designed to account for the fate of the runoff through measurements made with the soil moisture meter and the two groundwater-monitoring wells. The soil moisture data were used to determine the porosity of the soil. The data acquired from the soil moisture meter indicate that the typical porosity of the soil in the bioswale and each of the rain gardens is 29%.
A simple water budget was constructed for each measured rain event. The cumulative results over many rain events allow each BMP’s performance to be characterized.
The water budget calculations can be described as follows:
- An initial abstraction of 0.05 inch of rainfall was made at the start of each storm to account for rainfall that was trapped in pavement cracks and did not flow into the BMP.
- For each five-minute interval during a storm, the runoff is calculated: Accumulated rainfall (feet) x (8,325 square feet of pavement + 933 square feet of bioswale) = cubic feet of inflow.
- For the same five-minute interval, the water stored in the bioswale is calculated: Change in water level in above surface storage x 933 square feet + change in water level within the bioswale x 933 square feet x porosity = change in storage.
The results have led to several findings. Using the monthly full-depth charts, it was determined that the BMPs at St. Margaret Mary did not have any overflow events during the monitoring period. Water-level meters did not indicate any storage of water above the ground surface during any rain event. Additionally, there was only one event, on September 13, 2008, during which the subsurface water-level meter showed any significant storage of stormwater in the BMPs.
Although the team did not have site-specific rainfall data for the September 13 event, nearby rain gauges recorded rainfall ranging from 5 to 7 inches. The data indicate that the infiltration rates of the engineered soils and subsurface soils are extremely high, and the BMPs at St. Margaret Mary are deemed virtually 100% effective in capturing and infiltrating the stormwater that runs off the parking lots and roof of the rectory.
Anecdotal evidence of success of the BMPs at St. Margaret Mary was also obtained from staff members, who indicated that previously, water from the parking lot was able to seep into the basement during flooding. Even with extreme rain event in September 2008, the school has reported no basement flooding since the installation of the BMPs.
Based on the water budgets constructed for the bioswale, the water budget data indicate that runoff captured in the bioswale was far less than predicted by the runoff calculations. One potential reason is the amount of parking lot runoff that reaches the bioswale may be less than measured. The grades are very flat and, because of asphalt surface variations, the sheet flow patterns may not result in the full delivery of parking lot runoff to the bioswale. In addition, the highly porous soils in the bioswale don’t allow for runoff to pond and generate a saturated water surface. The monitoring system was dependent on the development of a saturated water surface that could be measured. If the runoff is in constant flux, quickly infiltrating to the subsoils, then the static water-level meters are inadequate for this measurement.
The team nonetheless concluded the bioswale was capable of capturing–storing and infiltrating–all of the runoff generated from the storm events over the course of the monitoring project.
Scaling Up
Relating activities at varying geographic scales is one key to the proposed approach. To effectively plan and measure green infrastructure performance requires connecting local site-based conditions with broader scales of cumulative water-quality impact.
Putting green infrastructure to work means adapting to existing site conditions, including the location and concentration of built infrastructure, such as buildings, roads, and other impermeable surfaces, as well as the open space surrounding them. The monitoring team suggested follow-up modeling studies to plan strategic deployment of green infrastructure practices and measure the extent to which they work in the affected urban watersheds.
After this first phase, CNT was still interested in learning more about rain gardens and infiltration facilities, Doohaluk says. “This is when we started to develop the list of fundamental questions regarding rain gardens, their operation, and needs for maintenance,” she says. “Are there any differences between native vegetation or turfgrass planting? Any differences in operation throughout the region? When do they need maintenance? Do they always function? Do newer ones function similar to ones that are seven years old? Are there any benefits to using engineered soils over basic topsoil or native soils?
“We were exploring various monitoring strategies when Stormwater magazine showed up on our desk with an article about the University of Minnesota’s methodologies [“The Four Levels,” September 2008 issue]. We used their methodology and further streamlined it, making even simpler.”
Doohaluk says efforts were focused on developing a method that could be used to gauge the performance of rain gardens by almost anyone who owns or maintains a rain garden.
The University of Minnesota was using a modified Philip-Dunne (MPD) infiltrometer that is relatively easy to construct, Doohaluk says. “Ours was constructed using parts obtained at a retail home improvement center,” she says. The infiltrometer was constructed by Hey and Associates staff from a PVC pipe with a height of 61 centimeters and an inner diameter of 10.2 centimeters.
In its study, the University of Minnesota did extensive data collection, Doohaluk says. “They were taking 20 to 40 measurements in the rain garden. That’s a challenging work load without the student help.”
In January 2009, CNT was provided a grant by the Illinois Sustainable Technology Center (ISTC) for monitoring and documenting the performance of stormwater BMPs. The ISTC project continued the monitoring of the bioswale at Our Lady Gate of Heaven Parish and the bioswale and two rain gardens at St. Margaret Mary Parish.
In December 2009, CNT requested a modification to the project for 2010, which would have extensive benefits for the state and the region. CNT and Hey and Associates proposed to conduct an inventory of the available rain gardens and permeable pavement installations in the Chicago region and to select a number of them for monitoring in 2010. In conducting an inventory of rain gardens as well as permeable pavement in the Chicago region, the project team surveyed a range of sources to identify as many BMPs in the region as possible.
The team was encouraged by the results of the 2009 monitoring, as well as the work done by the University of Minnesota, where a procedure using three levels of assessment has been successfully employed and standardized. The three levels of alternative approaches used as part of this project included visual inspection, infiltration rate testing, and synthetic drawdown testing.
For the visual assessment, every rain garden and other example of green infrastructure that had been identified prior to snowfalls in December 2010 was visited. The team selected approximately 70 sites for a detailed visual inspection, assessing each using a checklist developed by the University of Minnesota. The checklist included hydrologic problems, vegetation, and soils. A typical visit, including taking several photographs, took about 20 to 45 minutes.
The results of the visual assessment:
- The age of most features is not readily available, but age may not be a major factor after a year or two.
- There is little or no standing water after a day or so unless the rain garden is designed to retain a permanent water level. It is not essential to know details about recent rainfall.
- None appeared to have saturated soils unless there was supposed to be standing water.
- None of the soils appeared to be compacted (the soil profile was not examined).
- Many of the features did not have well-defined inlet structures, and very few had inlet structures that indicated malfunctions. A few had minor erosion or sedimentation but not enough to affect performance.
- No features showed signs of water pollution.
- Vegetation was widely divergent. All features contained native plants, but the distribution of native and other plants was not determined. Some features were well maintained with healthy and primarily native plants. Others needed maintenance. Only a small fraction had inadequate plant coverage; only recently planted vegetation had inadequate plant coverage. The time of year affects the distribution of predominant species and the look of the feature.
- Even the most poorly maintained features at smaller and urban sites looked as if they could be restored with routine weeding or a prescribed burn for larger sites.
- Bank erosion was identified in very few sites.
For the infiltration rate testing, 15 rain gardens were selected based on permission from the owner or operator, geographic location, and design of the rain garden. The goal was to select rain gardens that had a variety of runoff sources–roof, street, parking areas–and differing design elements, such as engineered soils, native soils, and sediment forebays.
The infiltration rate testing of a rain garden includes the use of an infiltrometer to determine near-surface saturated hydraulic conductivity (Ksat) at a number of locations throughout the garden. For this project, an infiltrometer based on the MPD infiltrometer was used. It was selected because of the minimal volume of water needed to run the test, its portability, and its low cost.
Use of the MPD infiltrometer can easily track rain garden performance over time and assist in the development and implementation of maintenance plans to ensure long-term success of the rain gardens. The infiltrometer used was constructed by Hey and Associates staff from a PVC pipe. A transparent piezometer tube was attached to the outside of the infiltrometer next to a measurement tape for making water-level readings.
MPD infiltrometer measurements were taken at a number of locations throughout each rain garden. The number of samples per rain garden was determined by the size of the rain garden. The individual testing locations in each rain garden were selected to provide a representative infiltration rate of the various parts of the rain garden.
In most cases, an infiltration test was conducted near the inlet, center or deepest section, and outlet of each rain garden. In rain gardens where more than three infiltration rate tests were conducted, the additional test sites were selected based on the configuration of the rain garden and included areas such as near dense or sparse vegetation and areas of known ponding.
Infiltration testing locations also avoided bushes, tr