GI Analysis:
Stormwater Terraces
St Clair Street, which was part of the 2020 street reconstruction project that took place in the study area, had stormwater terrraces, pervious pavement, and rock cribs being added throughout the street. Data is being collected from storm sewers near the study area to analyze the effects from multiple GI practices added, but there is still too few data because of the installation being relatively new. We decided to take a different approach and model only one practice that was added: Stormwater Terraces.
St Clair Street
Existing Conditions and Runoff Modeling Parameter
In order to accurately model runoff in the area, a couple of parameters had to be found in order to model the existing conditions. From the NRCS Soil Survey, hydrologic soil groups (HSG) were found for each property/lot across that section of street that was being reconstructed (seen below). After finding soil groups, cross slopes of properties averaged around 0.1% and the slope of the road from west to east averaged around 3.5%. Because properties typically have just grass on their lawns and are relatively flat, each property was modeled as sheet flow with short grass determining the Manning's roughness (n = 0.015). regarding the time to concentration. The lots averaged out at about 0.2-0.3 acres per lot, so 1/4 residential lots for the respective HSG were used in finding the curve number for each property at average moisture conditions. Because pervious pavers were not being modeled, the contribution of runoff from the pavement was not included as it would have been the same for each model scenario.
Soil Types on St Clair Street
The larger section on the left is labeled as MdC2 and is known as McHenry Silt Loam. The larger section on the right is labeled as DnB and is known as Dodge Silt Loam. This image was generated using the NRCS Soil Survey.
Soil Types by Parcel
For modeling, each of the lots were separated into the different soil groups based on the soil the majority of the property was on. Blue lots are MdC2 (HSG B) and orange lots are DnB (HSB C).
In modeling the existing conditions, each property was modeled as a sub-catchment in HydroCAD with sheet flow on short grass to calculate time of concentration. Curve numbers for runoff were determined using the HSG for each property along with the average moisture condition curve number for "1/4-acre residential lots". The street was modeled as a shallow rectangular reach for different portions of the street in order to show that different houses have runoff at different parts of the street. Eventually the connected reaches would outlet at the storm sewer (shown by the red box in the image above).
​
In modeling added GI, the stormwater terraces were modeled as small, shallow detention basins that spanned across a 50-ft portion of the terrace. The only stormwater terraces that were modeled were those that were proposed and installed during the 2020 street reconstruction and are shown in the figures below as blue triangles. Scenarios were run for the 1-yr, 10-yr, and 100-yr recurrence interval storm events as found on NOAA Atlas-14.
HydroCAD
Modeling Setup
Pre-GI HydroCAD Setup
Post-GI HydroCAD Setup
Results
Do stormwater terraces decrease peak runoff from storms?
After running all the scenarios, results showed that this section of street experienced up to a 27% decrease in runoff entering the storm sewer. When talking about larger storms, such as the 100-yr recurrence interval storm, only so much water can be abstracted using this form of GI so there may be smaller decreases comparatively to lower recurrence interval storms. The stormwater terraces can only hold so much storage and their abstraction of water is eventually limited by the saturated infiltration rate (~saturated hydraulic conductivity). Because there is more water associated with these larger storms, the water needs somewhere to go once those stormwater terraces have reached their capacity. Overall, these terraces still showed a reduction in runoff entering the storm sewer, and if more streets/neighborhoods implemented these practices, we may see some more promising results in the future.
Example Model Outputs at the Storm Sewer
Pre-GI
10-yr Storm
POST-GI
10-yr Storm
Modeling Limitations and Error
There are some limitations to what can be modeled and how it can be modeled. There were many assumptions in having each property be a sub-catchment and areas of the road being modeled as stream reaches. It also is limited by elevations and slopes of the existing terrain, because all that could really be modeled was left to right and up or down, but there may have been other associated slopes that may have changed the direction of runoff from going straight into the street to going onto other properties. Overall, this was a pretty simple approach at looking at peak runoff reductions from GI throughout a street in the study area, and shows that GI may be what we need.