Skip to main content

Stormwater Management

Effects of Roof Surface, Slope, and Substrate Depth on Stormwater Retention

Two studies were conducted that quantified the differences in water retention among (1) roof vegetation types and (2) among combinations of green roof slopes and substrate depths. Fifteen simulated roof platforms (ChristenDetroit, Detroit, Mich.) with overall dimensions of 2.44 m x 2.44 m (8 ft x 8 ft) were utilized at the Michigan State University Horticulture Teaching and Research Center (East Lansing, Mich.). Each platform simulated an actual commercial extensive green roof and included an insulation layer, protective layers, and waterproofing membrane. Three of the platforms were divided into three sections measuring 0.67 m x 2.44 m (2.2 ft x 8 ft) and were used for the vegetation study. The other twelve platforms were utilized for the slope/depth study. Aluminum sheet metal troughs were attached on the low end of the platforms to capture stormwater exiting the platform. For the vegetation type study the troughs are divided into three separate sections corresponding to the three divided sections. For the vegetation type study, platforms were set at a 2% slope with the top edge of the high end 0.9 m (3.0 ft) above ground level. Platforms used in the slope/depth study were placed at 2% or 6.5% slope. All platforms were oriented with the low end of the slope facing south to maximize sun exposure.

All platforms except for one of the three self-contained sections on each of the platforms utilized for the vegetation type study contained XeroFlor drainage layers, water retention fabric, and substrate carriers. The growing substrate consisted of heat-expanded slate (PermaTill®, Carolina Stalite Company, Salisbury, N.C.), USGA (United States Golf Association) grade sand, Michigan Peat, 5% Dolomite, composted yard waste, and composted turkey litter. For the vegetation type study, each platform section with the  XeroFlor drainage system installed received planting substrate to a depth of 2.5 cm (1.0 in). The twelve platforms utilized in the slope/depth study received 2.5, 4.0, or 6.0 cm (1.0, 1.6, or 2.4 in) of substrate on top of the substrate carrier. The four treatments were a 2% slope with either 2.5 or 4.0 cm of substrate and the 6.5% slope with 4.0 or 6.0 cm of substrate.

Three roof types were tested in the roof vegetation type study: an extensive green roof with vegetation, an extensive green roof with substrate only, and a conventional commercial roof with a 2 cm (0.8 in) deep gravel ballast. Treatments were arranged in a randomized complete block design with each platform corresponding to one block. Vegetation, substrate only, or gravel ballast were randomly assigned within sections of each platform. For the slope/depth study, all platforms were covered with vegetation. For both studies, 100% coverage was achieved on the vegetated section prior to the initiation of data collection. Taxa included several species of Sedum planted from seed (Jelitto Staudensamen, GmbH, Schwarmstedt, Germany) at a rate of 1.3 g•m-2 (0.004 oz•ft-2) for each species.

Model TE525WS tipping bucket rain gauges (Campbell Scientific, Inc., Logan, Utah) were mounted under the drain of each platform section to quantify stormwater runoff. An additional tipping bucket was mounted in each gravel section to record actual rainfall on the vegetation type platforms. Also, a Campbell Scientific’s CM6 tripod weather station was installed on the research site to aid in determining meteorological variables. The weather station included a temperature and relative humidity probe covered by a 6-plate gill radiation shield. In addition, instruments to measure wind speed and direction were installed as well as a pyranometer. Data from the tipping bucket rain gauges and tripod weather station were collected every five minutes, 24 hours a day using a Campbell Scientific CR10X datalogger with peripheral multiplexers, switch closure modules, and storage module. A rainfall event was defined by two rainfalls separated by a period of six hours or more, with an easily distinguishable flow hydrograph.

Over the 14-month period the study was conducted, the vegetated roof treatments retained 60.6% of  rainfall from 83 measured rain compared to 50.4% and 27.2% for the substrate-only and conventional gravel ballast roofs, respectively.  In the slope/depth study, platforms at 2% slope with a 4 cm substrate depth had the greatest mean retention, although the difference from the other treatments was minimal.   The combination of reduced slope and deeper substrate clearly reduced the total quantity of runoff.  For both studies, vegetated green roof systems not only reduced the amount of stormwater runoff, they also extended its duration over a period of time beyond the actual rain event. 

Complete results are published in:

VanWoert, N.D, D.B. Rowe, J.A. Andresen, C.L. Rugh, R.T. Fernandez, and L. Xiao.  2005. Green roof stormwater retention: Effects of roof surface, slope, and media depth.  J. Environ. Quality 34(3):1036-1044.

A platform divided into 3 sections
Tipping buckets used to measure the volume and rate of runoff from roof platforms

Quantifying the Effect of Slope on Stormwater Retention

Runoff was analyzed from twelve extensive green roof platforms constructed at four slopes (2%, 7%, 15%, and 25%).  Rain events were categorized as light (<2.0 mm), medium (2.0 – 10.0 mm), or heavy (>10.0 mm).  Data demonstrated an average retention value of 80.8%.  Mean retention was least at the 25% slope (76.4%) and greatest at the 2% slope (85.6%).  In addition, runoff that did occur was delayed and distributed over a long period of time for all slopes.  Curve numbers, a common method used by engineers to estimate stormwater runoff for an area, ranged from 84 to 90, and are all lower than a conventional roof curve number of 98, indicating that these greened slopes reduced runoff compared to traditional roofs.

These findings are applicable to the midwestern United States and other geographical areas with similar climates.  The Michigan State University campus covers 21.0 km2 (5,200 acres) and has 1.1 km2 (12 million ft2) of flat roof surface.  If all of these roofs were greened similar to the roof platforms in this study, then based on a mean retention of 80.2%, these roofs would have retained 377,041 m3 (99,603,827 gallons or 13,315,095 ft3) during 2005.  Of course, retention on any roof depends on rainfall distribution throughout the year, the intensity of each event, ambient air temperatures, plant selection, and the influence of local environmental conditions on evapotranspiration.

Complete results are published in:

Getter, K.L., D.B. Rowe, and J.A. Andresen.  2007.  Quantifying the effect of slope on extensive green roof stormwater retention.  Ecological Engineering 31:225-231.

An arial view of many research platforms - organized in long rows; varying in slope.
Measuring differences in runoff at various substrate depths and roof slopes

A weather-proof box with electronics inside; used for data collection.


Comparison of Stormwater Runoff Quantity and Quality from Sedum, Native Prairie, and Vegetable Producing Green Roofs

Extensive green roofs used for crop production will differ from more traditional Sedum and prairie covered roofs in plant density and water use efficiency, but the impact on stormwater retention has not been well studied.  Three vegetation types (unfertilized Sedum and native prairie species mixes, and a fertilized vegetable and herb species mix) were compared for stormwater runoff quantity over three growing seasons, and stormwater runoff quality during one growing season. Under the fertilization regimen used for this study, there was no increase in runoff NO3- concentrations in the vegetable green roof over the other, more traditional green roof types, but there was an increase in P. NO3- concentrations decreased over the course of the growing season and runoff phosphorus (P) concentrations also decreased over time in the Sedum and prairie treatments, which were lower than P concentration from the vegetable crop system throughout the growing season. This is likely a result of the difference between amounts of NO3- and P applied to the vegetable treatment and the needs of the crop plants in that treatment. The similarities in water retention and water quality between vegetable producing extensive green roofs and Sedum green roofs suggest that vegetable production with careful nutrient management will not have a negative impact on the ability of extensive green roofs to retain stormwater or manage stormwater quality.

Complete results have been published in:

Whittinghill, L.J., D.B. Rowe, B.M. Cregg, and J.A. Andresen. 2014. Comparison of stormwater runoff from sedum, native prairie, and vegetable producing green roofs.  Urban Ecosystems (In press) (

A row of research platforms.
Collection of runoff from sedum, prairie, and vegetable producing green roofs

Retention on Steep Sloped Extensive Green Roofs

The study was conducted on raised roof platforms (1.22 m x 2.44 m) on the Old Mission Peninsula, Grand Traverse County, MI. Each platform was partitioned into three sections (0.76 m x 1.22 m) and set at a 4:12 pitch (33% slope) to replicate a typical sloped (4:12 pitch) roof on a single family home in Michigan. Platforms were built on a post structure to a height of 0.91 m, which allowed room for monitoring equipment to be installed under the drains of each partition. Five treatments (conventional steel roofing, conventional fiberglass/asphalt shingles, engineered substrate only (15.2 cm of an engineered substrate containing 80% heat-expanded slate, 15% sand, and 5% organic matter), engineered substrate plus native grasses (Buchloe dactyloides and Bouteloua gracilis), and engineered substrate plus sedum(Sedum acre, S. album, S. ellacombianum, S. kamtshaticum, and S. reflexum)) each replicated three times were tested to quantify the difference in stormwater retention.

Runoff was directed into a downspout that ran directly into TB-4 tipping bucket rain gauges (Campbell Scientific Inc.) and recorded continuously on a CR-1000 datalogger. Accuracy of the rain gages, as reported by the manufacturer, are ±1%, +0 and -2.5%, and +0 and -3.5% for rainfalls of <25.4 mm•h-1, 25.4 to 50.8 mm•h-1, and 50.8 to 76.2 mm•h-1. Independent rain events were defined as precipitation events that were separated by six or more hours.  In the event runoff was still occurring six hours after the initial rainfall event, the two events were combined.  A review of annual NOAA rain events suggests that rain events in this part of Michigan could be categorized as light (< 5mm), medium (5 - 14 mm), or heavy (>14 mm); this was an arbitrary categorization set by the author to facilitate data analysis. 

The ANOVA model showed rain category and roof treatment, as well as the interaction of both, to be significant. Over the 2 year study period; there was 701.54 mm of cumulative rainfall from the 63 measured rainfall events.  The fiberglass asphalt shingle roof treatment retained 99.18 mm (14.1%), the steel roof treatment retained 68.72 mm (9.8%), the engineered substrate only treatment retained 179.33 mm (25.6%), the engineered substrate plus sedum roof treatment retained 291.80 mm (41.6%), and the engineered substrate plus grass roof treatment retained 281.40 mm (40.5%).  This study is not yet complete.

Effect of Vegetation Selection on Water Runoff Quality for Steep Sloped Green Roof Systems

The sloped green roof platforms in Traverse City described above were also used to monitor runoff water quality.  Three green roof coverage types (sedum vegetation, grass vegetation, and substrate only) are being compared to conventional shingles and steel roofing with respect to water runoff quality. Runoff samples have been collected for a period of three years and are being analyzed for nitrogen (ammonia, nitrate, and nitrite), phosphorous (phosphate), and turbidity.  This study is currently in progress.

3 tipping buckets connected to a platform - white in color.
Tipping buckets and collection of runoff samples