Using GIS to Prioritize Green Infrastructure Installation Strategies in

an Urbanized Watershed

Lauren Owen

1, Daniel R. Hitchcock

2, David L. White

3, Gene W. Eidson

4

AUTHORS: 1 Graduate Research Assistant, Biosystems Engineering, Clemson University, Clemson, SC 29634, USA. 2Associate Professor, Baruch Institute of Coastal Ecology and Forest Science, Clemson University, Georgetown, SC

29440, USA. 3Clemson Computing and Information Technology, Clemson, SC 29631, USA. 4Director, Institute of

Computational Ecology, Clemson University, Clemson, SC 29634 and Urban Ecology Center, Aiken, SC 29801, USA.

REFERENCE: Proceedings of the 2014 South Carolina Water Resources Conference, held October 15-16, 2014 at the

Columbia Metropolitan Convention Center.

ABSTRACT. A Geographic Information System

(GIS) was utilized to create a digital elevation model

(DEM) and a stormwater systems map in order to

perform spatial hydrologic analyses on watershed

discharges from urban Aiken, SC. Arc Hydro® was used

to create and analyze hydro networks, to determine flow

accumulation, and for subwatershed delineation. The

HEC-HMS preprocessor HEC-GeoHMS was then used

to transform the ArcMap model into a format that can be

imported into watershed model HEC-HMS to create

hydrographs for peak flow and runoff in 10

subwatersheds within the entire urban watershed. Model

calibration is underway based on isolated storm event

data to determine stormwater flow and volume

contributions from respective subwatersheds. By

modeling storm events, hydrologically informed

decisions can be made related to the addition of green

infrastructure-based stormwater control measures

(SCMs) in specific subwatersheds. This modeling effort

is also essential for understanding the infrastructure

connections within the existing city stormwater system.

INTRODUCTION

Urbanized watersheds, specifically those with high

percentages of impervious cover, create increased runoff

volumes, flows, and energy that can result in downstream

erosion and other water quality impacts. These highly

impervious urban landscapes could benefit from the use

of sustainable restoration strategies based on green

infrastructure principles that mimic natural hydrological

and ecological processes. Effective sustainable land use

strategies for highly developed areas require an

optimization at varying spatial and temporal scales: not

only at the watershed level, but also at the scale of the

urban centers and residential neighborhoods. This is

especially the case with respect to selection and design

criteria for stormwater control measures (SCMs) as

related to their individual and collective performance

within an urban watershed as well as to their integration

into the existing landscape. In this paper, we share

results from investigation into a flow simulation

methodology for urban stormwater systems.

PROJECT DESCRIPTION

This project is part of the Institute of Computational

Ecology's Intelligent River® Research Enterprise

monitoring and visualization effort along with being part

of an ongoing project located in an urban center being

retrofit to manage stormwater quantity – the Aiken Green

Infrastructure project – while improving downstream

water quality at the Sand River headwaters through

upstream stormwater volume reduction. The urban

stormwater infrastructure of downtown Aiken, SC (33o

33’ 38”N, 81o43’10”W), has recently been retrofit with

pervious paving materials, bioretention cells, and a

cistern, offering an opportunity for the evaluation of

these SCMs. The modeling exercise reported here will

guide priority locations for future SCM installations.

This project is part of the statewide Intelligent River™

monitoring and data visualization effort, which offers

science-based knowledge to guide decision-making with

respect to water resources management.

A major concern exists at a ten-foot diameter pipe that

discharges most of the urban stormwater into the Sand

River, serving as its headwaters - high energy at the

outlet has produced a large canyon due to erosion. In an

attempt to reduce high energy discharges and thus bank

erosion, green infrastructure based SCMs have been

installed within the upstream urban watershed, with

future new projects being proposed. The next step –

through preliminary work presented here – is to

determine the best locations (subwatersheds) for new

flow and volume reduction strategies based on

hydrologic monitoring and modeling.

METHODS

A major issue with urban stormwater characterization

is that these systems typically consist of underground

infrastructure, therefore difficult to recognize throughout

the Digital Elevation Model (DEM) when creating the

representation. Many studies have been conducted on

methods of displaying the pipe system within a

watershed model, and the most common method

demonstrated is a technique called “burning” in the

streams to the DEM (Gironas et al. 2010, Paz 2007,

Maidment 1996, Maidment 2002, Brock 2006). In this

method, a constant depth is subtracted at all cells that

include a known conduit that is not resolved by the DEM

(Gironas, 2010). Once this is completed, the stormwater

network is accepted as the natural stream element of the

DEM and can be used to delineate watersheds based on

the piping and extract various characteristics. The

ArcHydro extension (Maidment, 2002) of ArcMap 10.1

is used to create a drainage network and catchments

through various tools.

Once the drainage network is completed, the

preprocessor for HEC-HMS, HEC-GeoHMS, can be used

to transform the ArcMap model into a format that HEC-

HMS accepts. This is done in several studies including,

but not limited to, Verma 2009, Beighley 2003, Du 2012,

Emerson 2003, McColl 2007, Knebl 2005, Al-Abed

2005, Chen at al. 2009, and Ali et al. 2011. The drainage

network and catchments created in using ArcHydro are

then used to create subbasins and a river network that is

accepted by HEC-HMS. It executes this function through

a series of steps collectively known as terrain

preprocessing, implying the utilization of the surface

topography as the origin of the stream network (Knebl et

al., 2005). It is also used to derive several basin and river

characteristics using the input data from the DEM,

drainage network, soil data, and land use data. HEC-

GeoHMS outputs a basin model, meteorological model,

and river model ready to use in HEC-HMS. These main

components focus on determining runoff hydrographs

from sub-basins and routing the hydrographs through

channels to the study area outlet (Beighley, 2003). This

output data can then be used to compare to the volume

and flow data from the Sontek IQ Pipe sensors and

calibrate the modeling. The model can then be used to

predict subwatershed discharges during future storms and

in what subwatershed the storm will have the greatest

impact.

The first step in creating the watershed model in

ArcMap 10.1 was creating a high definition DEM from

Light Detection and Ranging (LiDAR) data. It is

demonstrated that a LiDAR-derived DEM with high

accuracy and high resolution offers the capability of

improving the quality of hydrological features extracted

from DEMs (X, 2005). The DEM was created using

various toolboxes inside ArcMap 10.1 to convert the

LiDAR file to a raster. After the DEM is created, the

stormwater pipe system is “burned” into the DEM

following Method 2 of Gironas et al. of burning all pipes

to the same artificial elevation below the DEM. This

creates the illusion that the pipe system is the natural

stream element of the watershed. Terrain preprocessing is

then completed and the DEM is filled and the flow

direction and flow accumulation grids are created

following Maidment’s methods in ArcHydro: GIS for

Water Resources. Catchments, an adjoint catchment, and

a watershed for the 10 foot pipe outlet, along with

subwatersheds for each monitoring location are created.

Following Merwade’s tutorial from Purdue

University, the HEC-GeoHMS toolbar is used to convert

the ArcMap 10.1 model into an acceptable format for

input into HEC-HMS. This is conducted by inputting the

output files from the ArcHydro toolbox including the

RawDEM (burned DEM), the HydroDEM (filled DEM),

flow direction, flow accumulation, stream grid, stream

link grid, catchment grid, catchment polygon, drainage

line polygon, and adjoint catchment polygon to output a

subbasin shapefile and a river shapefile. All subbasins

were merged to create subbasins that matched the

subwatersheds derived within ArcMap 10.1 using the

basin merge tool in HEC-GeoHMS. Basin processing is

completed to extract subbasin characteristics necessary

for input into HEC-HMS. An HMS schematic is then

created using the HMS schematic tool and inputting the

project point, centroid, river, and subbasin files to output

HMSlink and HMSnode files. This shows the schematic

that will be demonstrated in HEC-HMS along with the

subbasin background shapefile. It is important to note

that an impervious surface grid was clipped to the area of

study and soil groups and land use data were merged to

create a curve number (CN) lookup table within ArcMap

10.1 for use in HEC-GeoHMS. The CN lookup table and

the soil/land use polygon were then used to create a CN

grid, which was input into HEC-GeoHMS along with the

impervious surface grid to calculate subbasin

characteristics such as CN, lag time, and impervious

percent per subbasin. SSURGO soil data from NRCS

was used along with NLCD land cover and NLCD

percent impervious data, along with Microsoft Access,

and input into ArcMap 10.1.

Once opened in HEC-HMS, the only thing that has not

been previously calculated in HEC-GeoHMS is lag time

(minutes) per reach. This was calculated by extracting

elevations and lengths to calculate slope from ArcMap

10.1 and inputting them into equation 1:

( ) (

)

(hours)

with L representing hydraulic length (km), CN

representing the average CN, and Y representing percent

slope (Costache, 2014). The SCS storm method was

chosen for the meteorological model and the storm depth

in inches is input into HEC-HMS. A simulation run is

then generated and computed to output a global summary

of volumes and flow rates per reach and subbasin along

with hydrographs for each option. These outputs can then

be used to compare to the sensor data calculated volumes

and flow rates to calibrate and improve accuracy of the

hydrologic model. The model can then be used to predict

the outcome of a given storm of a certain depth of

rainfall.

Ten Sontek® IQ Pipe™ sensors have been installed to

monitor the stormwater and baseflow volumes and flows

travelling through key stormwater pipes throughout the

city and surrounding watershed. These data will be

compared to predicted subwatershed stormwater flows

and volumes.

RESULTS

The DEM generated from the LiDAR data with

elevations ranging from 219 feet to the highest elevation

of 566 feet, shown in Figure 1 with monitoring locations.

Figure 1: DEM created from LiDAR data and

monitoring locations.

The most important outputs from the ArcHydro toolbox

are the watershed and subwatershed polygons, along with

the drainage line, catchment and adjoint catchment

polygon layers which are demonstrated in Figures 2-5

respectively. Resulting watershed and subwatershed

areas are also provided in Table 1. The delineated total

watershed area as represented by a sum of these

subwatershed areas is 1079 areas while the total area

determined by GIS at the 10-ft. pipe discharge (Station 6)

is 1070 acres.

Figure 2: Subwatersheds per monitoring location.

Total 10 foot pipe watershed area includes all

subwatersheds except that draining at Station 11.

Table 1. Subwatershed output areas. Subwatershed 6

is the entire drainage area at the 10-ft pipe (1079

acres) excluding Subwatershed 11.

Figure 3: Drainage line indicating the natural streams

and inclusion of the stormwater system.

Figure 4: The catchment polygon layer dividing at

each stream segment.

Figure 5: Adjoint catchment polygon layer, in pale

blue, demonstrating the same boundary definition as

the watershed layer.

Figure 6: The river and subbasin file generated in

HEC_GeoHMS for input into HEC-HMS.

Once input into HEC-GeoHMS, these layers along with

several others are transformed into a river shapefile and a

subbasin shapefile as shown in Figure 6.

Table 2 provides the average CN, impervious

percentage (Xian, 2011), and lag time (hours) per

subbasin that are input into HEC-HMS and automatically

calculated through HEC-GeoHMS using the impervious

surface grid, the soil group data (Soil Survey Staff,

2014), and the land use polygon (Jin, 2011).

Table 3 provides sample calculations from a July 19th

storm and demonstrates volume (gal) and peak flow rate

(cms) per subwatershed. Runoff ratios (or the percentage

of rainfall generated as watershed discharge) ranged from

0.37 (Subwatershed 5) to 6.68 (Subwatershed 8).

Table 2: Subbasin characteristics extracted and

calculated using HEC-GeoHMS.

Table 3: HEC-HMS volume and peak flow rate per

watershed for a July 19th

storm of 0.51 inches along

with the rainfall runoff ratio.

DISCUSSION

As mentioned before, the point of “burning” the

stormwater pipe system into the DEM is to force ArcMap

10.1 to accept them as the natural stream element of the

study area. In order to prove this, the watershed polygon

and the adjoint catchment polygon must illustrate the

same boundary outline due to their definitions. A

catchment is defined as a tessellation or subdivision of a

basin into elementary drainage areas defined by a

consistent set of physical rules, while a watershed is

defined as a tessellation or subdivision of a basin into

drainage areas selected for a particular hydrologic

purpose (Maidment, 2002). In other words, a catchment

is defined by the natural hydrologic elements while a

watershed is set specifically by an outlet. This means that

if the stormwater system was accepted by the model as

the natural stream element, then the adjoint catchment

and watershed boundaries should be the same as

demonstrated by Figure 2 showing the total watershed

and Figure 5 which considers adjoints. It is also proven

in Figure 3 which illustrates the drainage line, where the

pipe system is clearly included along with the natural

streams of the area in the shapefile.

The output calculations including percent impervious,

average CN, and lag time per subbasin all demonstrate

reasonable numbers for HEC-HMS input. HEC-HMS is

currently being calibrated to compare with sensor-

collected volume data and flow rate calculations;

however, the model is functioning properly and

generating output hydrographs and volume and flow

calculations. The rainfall runoff ratios were then

calculated, which should be between 0 and 1.

Subwatersheds with ratios greater than 1.0 could be due

to rainfall variability over the entire watershed as well as

baseflow contributions from groundwater, irrigation

within the watershed, and potential water leaks in the

system that were not considered in flow simulations. The

model can then be used to decide which subwatersheds

are contributing the most flow to the stormwater system

and where additional green infrastructure will be the

most efficient.

This work shows that urban stormwater systems can be

modeled and analyzed in various modeling formats

including: ArcMap 10.1, HEC-GeoHMS, and HEC-

HMS. We also demonstrate how these various programs

can be used in unison to create a hydrologic model for

urban stormwater analysis. Other urban or developing

areas can use the skeletal model of this strategy to solve

their existing stormwater runoff issues, as well as, and

more importantly, to prevent these problems from arising

in the first place.

ACKNOWLEDGMENTS

This project is funded through a City of Aiken, SC

research grant to the Institute of Computational Ecology.

I would like to give a huge thank you to the City for their

funding and support. Special thanks to all who have

made amazing efforts to make this project what it is

today including but not limited to: Kelly Kruzner, Alex

Pellett, Ben Smith, and Ron Mitchell.

LITERATURE CITED

Al-Abed, N., F. Abdulla, and A. Abu Khyarah. "GIS:

Hydrological Models for Managing Water Resources

in the Zarqa River Basin." Environmental Geology

47.3 (2005): 405-11. Web.

Ali, Muhammad, Sher Jamal Khan, Irfan Aslam, and

Zahiruddin Khan. "Simulation of the Impacts of

Land-use Change on Surface Runoff of Lai Nullah

Basin in Islamabad, Pakistan." Landscape and Urban

Planning 102.4 (2011): 271-79. Web.

Beighley, R. Edward, John M. Melack, and Thomas

Dunne. "Impacts of California's Climatic Regimes

and Coastal Land Use Change on Streamflow

Characteristics." Journal of the American Water

Resources Association 39.6 (2003): 1419-433. Web.

Brock, Larissa Mason. Water Quality, Nutrient

Dynamics, Phytoplankton Ecology and Land Uses

within Defined Watersheds Surrounding Six

Detention Ponds on Kiawah Island, South Carolina.

Thesis. The Graduate School of the College of

Charleston, 2006. N.p.: n.p., n.d. Print.

Chen, Ying, Youpeng Xu, and Yixing Yin. "Impacts of

Land Use Change Scenarios on Storm-runoff

Generation in Xitiaoxi Basin, China." Quaternary

International 208 (2009): 121-28. Print.

Costache, Romulus. "USING GIS TECHNIQUES FOR

ASSESSING LAG TIME AND CONCENTRATION

TIME IN SMALL RIVER BASINS. CASE STUDY:

PECINEAGA RIVER BASIN, ROMANIA.”

Geographia Technica 9.1 (2014): 31-38. Web.

Du, Jinkang, Li Qian, Hanyi Rui, Tianhui Zuo, Dapeng

Zheng, Youpeng Xu, and C.-Y. Xu. "Assessing the

Effects of Urbanization on Annual Runoff and Flood

Events Using an Integrated Hydrological Modeling

System for Qinhuai River Basin, China." Journal of Hydrology 464-465 (2012): 127-39. Web.

Emerson, Clay H., Caire Welty, and Robert G. Traver.

"Application of HEC-HMS to Model the Additive

Effects of Multiple Detention Basins over a Range of

Measured Storm Volumes." World Water and

Environmental Resources Congress (2003): n. pag.

Web.

Giron s, orge, effrey D. Niemann, arry A. Roesner,

Fabrice Rodrigue , and Herv Andrieu. Evaluation

of Methods for Representing Urban Terrain in Storm-

Water Modeling." Journal of Hydrologic Engineering

15.1 (2010): 1. Web.

Jin, S., Yang, L., Danielson, P., Homer, C., Fry, J., and

Xian, G. 2013. A comprehensive change detection

method for updating the National Land Cover

Database to circa 2011. Remote Sensing of

Environment, 132: 159 – 175.

Knebl, M.r., Z.-L. Yang, K. Hutchison, and D.r.

Maidment. "Regional Scale Flood Modeling Using

NEXRAD Rainfall, GIS, and HEC-HMS/RAS: A

Case Study for the San Antonio River Basin Summer

2002 Storm Event." Journal of Environmental

Management 75.4 (2005): 325-36. Web.

Maidment, David R. Arc Hydro: GIS for Water

Resources. Redlands, CA: ESRI, 2002. Print.

Maidment, David R. "GIS and Hydrologic Modeling-an

Assessment of Progress." GIS and Environmental

Modeling 3 (1996): n. pag. Web.

Mccoll, Chris, and Graeme Aggett. "Land-use

Forecasting and Hydrologic Model Integration for

Improved Land-use Decision Support." Journal of Environmental Management 84.4 (2007): 494-512.

Web.

Merwade, Venkatesh. "Creating SCS Curve Number

Grid Using HEC-GeoHMS." Purdue University

(2012): n. pag. Web.

Merwade, Venkatesh. "Downloading SSURGO Soil Data

from Internet." Purdue University (2012): n. pag.

Web.

Merwade, Venkatesh. "Terrain Processing and HMS-

Model Development Using GeoHMS." Purdue

University (2012): n. pag. Web.

Paz, Adriano Rolim, and Walter Collischonn. "River

Reach Length and Slope Estimates for Large-scale

Hydrological Models Based on a Relatively High-

resolution Digital Elevation Model." Journal of

Hydrology 343.3-4 (2007): 127-39. Web.

Soil Survey Staff, Natural Resources Conservation

Service, United States Department of Agriculture.

Soil Survey Geographic (SSURGO) Database for

[Aiken, SC].Available online at

http://www.arcgis.com/apps/OnePane/basicviewer/ind

ex.html?appid=a23eb436f6ec4ad6982000dbaddea5ea.

Accessed [August, 2014].

USACE, 2003. Geospatial Hydrologic Modeling

Extension. HECGeoHMS: User’s Manual. Version

1.1.. USACE, Davis, CA, USA.

Verma, Arbind K., Madan K. Jha, and Rajesh K.

Mahana. "Evaluation of HEC-HMS and WEPP for

Simulating Watershed Runoff Using Remote Sensing

and Geographical Information System." Paddy Water

Environ 8 (2010): 131-44. Web.

X, Liu, J. Peterson, and Z. Zhang. "High-Resolution

DEM Generated from LiDAR Data for Water

Resource Management." (2005): 1402-408. Web.

Xian, G., Homer, C., Dewitz, J., Fry, J., Hossain, N., and

Wickham, J., 2011. The change of impervious surface

area between 2001 and 2006 in the conterminous

United States. Photogrammetric Engineering and Remote Sensing, Vol. 77(8): 758-762.