Blog‎ > ‎

Identification of Sanitary Sewer Overflow Hotspots in Maryland

posted Oct 27, 2015, 11:43 AM by Erika Howder

Lori A. Lilly, Environmental Restoration Specialist; Nick Chamberlain, Freelance Software Developer; and Benjamin Sigrist, GIS Consultant

From January 1, 2005 – April 17, 2015, 9,579 sanitary sewer overflows (SSOs) discharged approximately 900 million gallons of raw, untreated sewage into Maryland streams (Maryland Department of the Environment (MDE) Combined Sewer Overflow/Sanitary Sewer Overflow Master database) (Figure 1). Of these SSOs, 40% were caused by blockages related to grease, rags, trash and other inappropriate material placed into the sanitary sewer system, resulting in nearly 16 million gallons of untreated sewage discharges from 2005-20014 (Figure 2). Many large cities in the Chesapeake Bay watershed are working under Consent Decrees to eliminate sanitary sewer overflows (SSOs) associated with rainfall and inflow/infiltration. While this is an important effort that will require significant investment in sanitary sewer infrastructure improvements, dry weather SSOs are more damaging because, unlike those SSOs caused by precipitation events, the raw sewage is not diluted by billions of gallons of associated stormwater. These dry weather SSOs represent a completely preventable pollution source as they are based primarily on individual behavior and, theoretically, should be able to be rectified through an effective social marketing campaign to change behavior.

SSO_Photo

Figure 1. Sanitary sewer overflow caused by a blockage.

SSO_Graph

Figure 2.  Number of SSOs caused by blockages from 2005-2014.  Data from Maryland Department of the Environment.

The direct cost for clearing a clogged sewer is approximately $4,000, which does not include addressing any resulting property or structural damage. For the approximate 3,757 MDE reported SSOs caused by blockages, this suggests that a conservative cost estimate of $15,028,000 has been spent in MD since 2005 to address this correctable problem. The direct costs associated with this issue, and the indirect costs in terms of degraded water quality and risks to public health, could be significantly reduced if consumers would follow proper disposal procedures. A study from the University of North Carolina Charlotte indicates that the likelihood of changing behavior on this issue are high. This research revealed that: “key publics reflect a willingness, even an eagerness to comply with proper grease disposal procedures when they are made aware of the risks of improper disposal. As a result, the study’s plan calls for ongoing, targeted awareness efforts (based on traditional and emerging mass media channels) coupled with targeted interpersonal communication efforts to move the public from awareness to personal interest to positive behavior change.”

This project will develop a social marketing plan to eliminate SSOs caused by the improper disposal of materials into the sanitary system. In order to best identify the target audiences, the location of SSOs across the state needed to be located. “Hotspots” of SSO activity were mapped based on density and recurrence interval using the following process.

SSO data was downloaded from MDE’s Combined Sewer Overflow/Sanitary Sewer Overflow Master database . This data was accessed on April 17, 2015 for this analysis. Of the 9,579 records in the database, 3,757 records were extracted based on their relation to blockages from grease, rags or trash.
A command-line application was written to geocode each record’s address into a list of latitudes and longitudes that could be imported into ArcGIS. Since the locations were within the Maryland State boundary, a publicly available RESTful web service provided by MD Department of Information Technology’s MD iMap program (http://imap.maryland.gov/Pages/mdimap-2.0-data-services.aspx) was utilized. The application was written in Golang (https://golang.org/) due to its ease of use and portability, as well as the potential for it to be improved to use a concurrent programming model in the future. The command-line utility can also be used in both Windows and Mac OSX environments since Golang compiles to a binary executable for various environments.

The command-line application, called go_geocoder (https://github.com/heynickc/go_geocoder), allows the user to input a .csv file with a column representing the MD addresses to be geocoded. It will then parse these records and send each address to the geocoding service. For each record, the response from the geocoding service is a list of addresses ranked by similarity, just as ArcMap’s geocoding tools do natively. The utility will then choose the best address match from the responses and write the corresponding latitude and longitude coordinates back to the .csv file as new columns. The application is currently not production-ready, but does show the advantages of using ArcGIS Server’s RESTful web services in combination with a cross-platform programming language to create a portable client application that is decoupled from specific environments or vendors.

The application successfully geo-located the majority of records. It was necessary to evaluate a small portion of the records manually to either correct the address or discard if the record was unable to be located. Of the 3757 original records, 3167 were successfully geolocated. Kernel density was used to identify geographic hotspots of SSO. Since the goal of the project was to identify hotspots within one city block, a 250’ search area was utilized to best capture recurrences in the immediate area. The kernel density raster was further refined to only show the hotspots, which were then turned into vector polygons. The geolocated SSO points were spatially joined to the polygons and then exported into a separate dataset. Temporal recurrence was identified by summarizing the dataset by its geocoded address, sorting the attribute table from high to low, and assessing the number of times an address was reported to have a SSO. In this manner, SSO hotspots from blockages were identified across the State of Maryland (Figure 3).

SSO_BaseMap

Figure 3. SSO hotspot locations in Maryland.

In the Chesapeake Bay watershed, nutrients are heavily regulated as a water quality pollutant. In addition, many waterways are impaired for bacteria. Sanitary sewage overflows and leaks are big contributors to nutrient and bacteria pollution problems. This project aims to improve the quality of streams and rivers by changing behavior to prevent dry weather SSOs. The first step has been to identify the locations of SSO hotspot activity. Armed with this information, the next step of target audience identification and crafting appropriate messages to reach those audiences can be undertaken much more effectively.

This project was generously funded by the Chesapeake Bay Trust (http://www.cbtrust.org/site/c.miJPKXPCJnH/b.5368633/k.BDEA/Home.htm) and is being completed in partnership with Ridge to Reefs (http://www.ridgetoreefs.org/). For more information, please contact Lori Lilly at lorililly@gmail.com.