Grantee Research Project Results

2001 Progress Report: Municipal Sewers as Sources of Hazardous Air Pollutants

EPA Grant Number: R827930
Title: Municipal Sewers as Sources of Hazardous Air Pollutants
Investigators: Corsi, Richard L.
Institution: The University of Texas at Austin
EPA Project Officer: Chung, Serena
Project Period: January 2, 2000 through December 31, 2002
Project Period Covered by this Report: January 2, 2001 through December 31, 2002
Project Amount: $298,798
RFA: Urban Air Toxics (1999) RFA Text |  Recipients Lists
Research Category: Air

Objective:

The U.S. Environmental Protection Agency (EPA) has developed National Emission Standards for Hazardous Air Pollutants (NESHAPs) for Publicly Owned Treatment Works (POTWs). However, despite an acknowledgement that municipal sewers "have been identified as significant sources of hazardous air pollutant (HAP) emissions from certain POTWs," sewers were omitted from the NESHAP because "little information currently is available to the EPA regarding these emissions." The primary objectives of this research project are to assess whether municipal sewers are significant area sources of HAPs, and whether such emissions can lead to localized "hot spots" that should be considered for future NESHAPs related to POTWs. Specific objectives include: (1) development of a database that includes measured stripping efficiencies for a wide range of volatile chemicals in municipal sewers; (2) estimation of HAP emissions from a large urban sewer network; and (3) comparison of such emissions with other known sources of HAPs.

Progress Summary:

A series of field experiments to assess removal of volatile tracers from operating municipal sewers have been completed. We also used tracer data to evaluate existing models for the prediction of volatile organic compound (VOC) removal from sewers, and developed a novel modeling approach to predict emissions of VOCs from entire municipal sewer networks.

Field experiments were initiated during the previous year, and were completed during this reporting period. Three sewer reaches were used in the City of Austin (North Austin Outfall [NAO = 3.6 km], Waller Creek Line [WCL = 2.9 km], Lower Shoal Main [LSM = 2.7 km]), and a total of five experiments were completed: two each in WCL and LSM, and one in the NAO. The reaches spanned a wide range of conditions, particularly as related to channel slope, wastewater flow rate, and other relevant features (lengths of uniform channel, representative drop structures, etc.). Experimental tracers included (Henry's law constant at 25°C and units of (mg/L)gas/(mg/L)liq provided in parentheses) 1,2-dibromoethane (0.029), dibromomethane (0.036), 1,3,5-trimethylbenzene (0.14), 1,4-xylene (0.32), cis-1,3-dichloropropene (0.63), trans-1,3-dichloropropene (0.72), and cyclohexane (7.9).

An example result is shown below in Figure 1 for the LSM. In this example, tracer removal efficiencies (fraction of mass removed from wastewater along the 2.7 km reach) ranged from 60 percent (1,2-dibrimoethane) to 100 percent (cyclohexane). A summary of average stripping efficiencies for each experimental reach is provided in Figure 2. Note that 1,3,5-trimethylbenzene (1,3,5-TMB) and p-xylene values are low for the WCL. This reach extended to the University of Texas, and there was evidence of chemical discharge into the reach during experiments (i.e., increasing levels downstream of tracer injection point). The lowest removal efficiencies were consistently observed for the WCL, which also was characterized by the mildest channel slopes. The VOCs, with relatively low Henry's law constants (1,2-dibromoethane and dibromomethane), were effectively stripped from wastewater along the NAO and LSM; both of which were characterized by short sections of steep channel slopes (> 10 percent). The NAO also included a pump station and a section of pressurized pipe.

Figure 1. Example Result (Lower Shoal Main - Experiment 5).

During this reporting period, the field experiments' results were used to evaluate an existing mathematical model for estimating VOC emissions from sewer reaches. For individual reaches and compounds, model predictions were generally within a factor of two of experimental results. However, the model was more favorable when compared with a series of experimental sections (entire reach), with relative differences generally less than 20 percent between predicted and experimental removal efficiencies.

Figure 2. Average Removal (Stripping) Efficiencies for Each Experimental Reach and Tracer

We originally intended to use the model described above to simulate VOC emissions from a specific sewer network to determine the significance of volatile HAP emissions from municipal sewers and to identify "hot spots" for HAP emissions. However, we encountered a major problem with respect to the unwillingness of municipalities to participate in this research project. In response, we developed a novel approach for estimating system-wide volatile HAP emissions from municipal sewer systems, and used NPDES and industrial pretreatment data submitted by municipalities to the U.S. EPA as a means of "back-calculating" emissions from sewers. The modeling approach is based on Monte Carlo simulations of VOC discharges and transport through sewer systems with reasonable distributions of pipe sizes, channel slopes, and other operating conditions. The model described above was used to develop probabilistic distributions for VOC removal efficiency for a wide range of VOCs, types of sewers, and drop structures. These distributions were sampled 100,000 times to develop system-wide stripping efficiency distributions for individual VOCs.

When coupled with reported VOC mass loadings into individual wastewater treatment plants, the calculated stripping distributions allow for a back-calculation of VOC mass emission distributions. Tenth, 50th, and 90th percentile values were calculated for mass emissions of individual HAPs from individual sewer systems. These now are being compared with reported Toxic Release Inventory (TRI) emissions data from major industrial sources. For example, in one city with a population of approximately 500,000, we predict benzene, ethylbenzene, and naphthalene emissions from sewers to exceed those from the largest industrial emitters listed in TRI. Toluene emissions (50th percentile) were predicted to be 11 metric tons per year from the sewer system of that same city. We predicted that sewers serving a 40 million gallon per day (mgd) plant in another city would emit between 55 (10th percentile) and 350 (90th percentile) metric tons of benzene each year. These are potentially significant amounts that exceed industries reporting to the TRI in that city. We also predicted that sewers serving an 11 mgd plant in yet another city would emit between 55 and 310 metric tons per year of chlorobenzene.

Future Activities:

For the remainder of this study, we plan to expand and complete the system-wide modeling assessment described above and to engage in significant technology transfer through journal submissions and Web site development.

Journal Articles:

No journal articles submitted with this report: View all 7 publications for this project

Supplemental Keywords:

wastewater, VOCs, toxics, emissions, ambient air, Texas, TX, EPA Region 6., Air, Scientific Discipline, Geographic Area, Water, Toxics, EPA Region, State, Chemistry, HAPS, Environmental Chemistry, 33/50, Wastewater, Hydrology, air toxics, hazardous air pollutants, Benzene (including benzene from gasoline), Region 6, municipal sewer emissions, emissions, atmospheric chemistry, MTBE, benzene, Tetrachloroethylene, chemical composition, wastewater tracer studies, Toluene, effluents, air pollutants, Chloroform, Methyl tert butyl ether, Texas, municipal sewers, POTWs, Volatile Organic Compounds (VOCs), ambient air quality, Xylenes, Ethyl benzene, TX, acute toxicity, POTW, Methyl chloride (Chloromethane)

Relevant Websites:

http://www.ce.utexas.edu/prof/corsi/ Exit

Progress and Final Reports:

Original Abstract

2000 Progress Report

Final Report