ArcNews Online

Spring 2010
Search ArcNews

E-mail to a Friend
Central Business District of Adelaide, South Australia

Locating, Appraising, and Optimizing Urban Storm Water Harvesting Sites

By Matthew D. Shipton and Sekhar V. C. Somenahalli


  • GIS is helping make South Australia more drought resilient.
  • GIS is used in the creation of strategic water resource plans.
  • Specific storm water harvesting benefits are identified and maximized with GIS.

Storm water is surface water runoff generated from rainfall. It is created when rainfall intensity exceeds infiltration capacity. This often occurs when rain falls on hard, impermeable surfaces, such as footpaths, car parks, and roads. In urban areas, this water then typically flows untreated via storm water drains into local watercourses.

Storm water harvesting areas modeled in proximity to public roads.

Harvesting collects and stores storm water before it enters natural environments so that it can be reused at a later time as an alternative to water from water mains. Before it is reused, captured storm water may be treated to improve its quality. Recycled storm water is typically used for nonpotable demands that do not require a high level of treatment, for example, irrigation, toilet flushing, car washing, and certain industrial and commercial uses.

Harvesting storm water reduces the detrimental impacts that urban development can have on rivers. It reestablishes more natural river flow regimes and improves the quality of the water in rivers. In addition to the environmental benefits, storm water harvesting can also be financially attractive to both its suppliers and users. Harvesting storm water reduces demand for water from water mains, delays the need for major new water resource infrastructure, increases water security, and has low pumping costs since the source is often close to the point of use (i.e., it is generally harvested and used in the same urban area).

Storm Water Harvesting and GIS

Though beginning as a master's research project, the study described below was extended by the University of South Australia, through its Esri site license, to include detailed hydrological modeling of individual storm water schemes. The team performed this analysis using the Environmental Protection Agency specialist storm water modeling GIS (EPA SWMM). The objective was to verify claims made in the South Australian government's Water for Good plan, with results to be published later in 2010.

This study demonstrates the value of GIS to storm water managers through the assessment of harvesting opportunities in the central business district of Adelaide, South Australia. Adelaide is the largest city in the driest state in the world's driest inhabited continent.

Data used in the study was obtained from various sources, including the Australian Bureau of Meteorology; the South Australian Department for Environment and Heritage; the South Australian Department of Planning and Local Government; and the South Australian Department of Water, Land and Biodiversity Conservation.

The team members identified areas suitable for storm water harvesting in ArcGIS Desktop by analyzing land cover, land use, and topography. They defined suitable areas as having

  • Predominantly impermeable surfaces, such as concrete or asphalt pavement
  • An appropriate land use (that is, one that did not pollute the quality of the runoff from that area)
  • A natural drainage pattern that facilitated the collection of storm water from a large area without significant earthworks

First, the team defined each part of the study area as permeable, semipermeable, or impermeable. This categorization was derived from a reclassification of detailed land-use data based on assumptions regarding the average permeability of each land use. The output was then refined using local site-specific information, Normalized Difference Vegetation Index (NDVI) data, georeferenced aerial photography, and the Digital Cadastral Database (a source of legal land parcel data that, among other things, contains data on roof materials). The refinement allowed corrections, such as the alteration of the botanic gardens from being defined as impermeable—because its land use was public institution—to being permeable, based on its high NDVI and land cover as determined by aerial photography.

Next, the team identified all the roads and the areas in proximity to them in the study area. It was assumed that storm water runoff in these areas would be particularly polluted by contaminants from cars. These contaminants would likely include heavy metals, hydrocarbons, particulates from vehicle exhausts, debris from tires and brake linings, ultrafine platinum from worn catalytic converters, and lead salts from batteries. Most of the roads in Adelaide's central business district are laid out in blocks. Each side of the main blocks is 150 meters long. Areas in proximity to roads were defined as being less than 60 meters from a public road, since this is slightly less than half of 150 meters. The areas identified by the buffer that were more than 60 meters from a road were then intersected with the previously described impermeable areas layer. This procedure identified 245 separate areas that were impermeable and not close to a road. From these potentially suitable areas, sites that were less than 100 square meters were excluded on the grounds that they would not be economically feasible for harvesting storm water. This reduced the number of potentially suitable sites to 28.

Then, the team spatially joined information from a digital elevation model (DEM) with a 3-meter resolution to the polygons representing the 28 remaining potentially suitable sites. This enabled the average slope across each site to be calculated. Sites with a greater slope would generate more runoff per unit area per unit of time. The DEM was also used to define the main catchments and subcatchments in the study area (using the Arc Hydro extension). The team members calculated the number of subcatchments that each of the 28 potentially suitable sites covered. The number ranged between one and four. Those covering one subcatchment were deemed better suited to storm water harvesting since surface water runoff in them would drain to a single point.

The team ranked the suitability of the 28 remaining sites and visited the 10 most suitable sites to confirm if harvesting could be done within the available space; would not negatively impact existing land uses; and would be practical, given the local topography and existing drainage infrastructure. This process led to the removal of sites, such as cemeteries, that would be inappropriate for storm water harvesting.

Among the remaining potentially suitable sites, there were two that were highly suitable. These were the grounds of Adelaide Festival Centre and the land occupied by the University of South Australia's City East Campus and the main Adelaide University City Campus. These two areas are not highly permeable, less than 100 square meters, close to roads, or flat.

The next task using GIS was to examine if and how the most suitable harvesting sites could be connected by storm water infrastructure to form one storm water harvesting scheme. This would have economies of scale, since the project would then only require one treatment and storage solution. This task was completed using EPA SWMM.

Once the infrastructure layout of the optimal scheme had been determined, the team used EPA SWMM to model each of the harvesting sites' vulnerability to flooding after heavy rainfall. EPA SWMM was also used to determine the likely impact of climate change (reduced rainfall) and urbanization (increased impermeable area) on harvesting yields at each of the sites.


In South Australia, storm water harvesting is increasingly regarded as an untapped, sustainable water resource. This study demonstrates how GIS can be used to plan storm water harvesting schemes at both a strategic level by way of options appraisal and at a design level via hydraulic simulation. GIS provides decision makers with information that enables them to maximize system performance (storm water yield), economic feasibility (payback period), and public acceptability of harvesting schemes while reducing environmental degradation (caused by polluted storm water) and risks to public safety.

The GIS modeling undertaken with ArcGIS Desktop, Arc Hydro, and EPA SWMM in this study identified a number of potentially suitable areas in Adelaide's central business district that could be used as storm water harvesting catchments. A simulation of a harvesting scheme that used four of the most appropriate catchments found that 330 ML of reusable storm water could be collected every year. This finding should be interpreted with due regard for both the untreated quality of the harvested storm water and the irregular timing at which it would be available.

About the Authors

Matthew Shipton has experience working in the Hydrological Department of the Atlantic Rainforest Research Centre (Brazil). Shipton is completing a master's in water resources management at Adelaide University, Australia. Dr. Sekhar V. C. Somenahalli teaches courses related to GIS and its applications to environmental and planning disciplines in the School of Natural and Built Environments, University of South Australia, Adelaide.

More Information

For more information, contact Matthew Shipton (e-mail: or Sekhar V. C. Somenahalli (e-mail:

Contact Us | Privacy | Legal | Site Map