Additional Considerations for the Development of the Siloxanes Monitoring Program
Scott D. Dyer, Principal Scientist
Scott E. Belanger, Research Fellow
Environmental Stewardship Organization
The Procter & Gamble Company

For the 
Personal Care Products Council

1 August 2012

In order to successfully complete a monitoring program, it is imperative to agree on the goals of the effort.  In the specific case of D4/D5 cyclic siloxanes there are two competing goals.  These are loosely integrated in a lengthy "charge" statement addressed to the Siloxane Technical Working Group - POTW Monitoring date 24 July 2012.

Goal 1:  Develop a science-based, rational proposal for monitoring of Publicly Owned Treatment Works (POTW) sites to evaluate sources of and pathways to the environment for D4 and D5 discharged from POTWs.

Goal 2:  Develop a science-based, rational proposal for monitoring of Publicly Owned Treatment Works (POTW) sites to develop a national risk assessment for D4 and D5.

Stated as above, these goals are still very broad and do not extend to other aspects of the D4/D5 monitoring. Other possible goals are certainly in discussion and would include understanding releases from manufacturing sites/scenarios.  Also, while the goals are stated in terms of POTWs, they also necessarily include understanding pathways and exposure in the context of the receiving water and its biota.  Goal 1 and 2 are not totally incompatible, but each requires a different strategy.  With the understanding that Goal 1 was the original stated goal of the siloxanes monitoring program under discussion with SEHSC and USEPA, we would offer some thoughts given the long experience that P&G has in conducting both types of monitoring programs.  This includes an evaluation of ideas brought forth by USEPA in the documents "Major Factors Impacting Environmental Levels of Siloxanes (D4/D5) at POTWs without Industrial Siloxane Sources and Selection of Sites for Monitoring" and "Results from EPA's Proposed D-Optimal Design of Experiment Analysis".

It appears from both of the USEPA documents that there are a number of unknowns being considered by USEPA that are actually much more well understood than believed.  The choice of critical factors that could control environmental levels of siloxanes to be included in the design could be streamlined and improved.
   * Population Served by POTW  -  this factor is unlikely to be critical.  The use of consumer products, for example containing D4/D5, in a region is not dependent on population size associated with a given POTW.  While this may increase the total load to a treatment facility, the per capita use of water will also increase and result in fairly equivalent levels of a compound in the influent.  High population-served communities (urban areas) typically have storm water issues and poorer habitat quality due to hydrological modifications  -  leading to poorer biological data. Of course, we recognize there is noise in these factors, but population is not a sound surrogate by itself.
   * Treatment Type  -  this factor is likely to be critical.  Treatment types vary in efficiency.  A frequency analysis by Procter & Gamble (unpublished) on the 2000 CWNS showed that there were nearly 200 treatment train combinations.  Twenty treatment processes accounted for 89% of all WWTPs in the dataset.  The top five treatment types were (in order of frequency):  activated sludge (AS), stabilization pond (SP), aerated lagoon (AL), oxidation ditch (OD), trickling filter (TF).  These comprised 63% of the dataset.  The remaining proportion of plants did not contain data describing secondary treatment processes or was found to be a combination of two or more treatment processes (e.g., AS/TF  -  activated sludge with trickling filter).  
   * Total Suspended Solids  -  this factor is unlikely to be critical if the site selection is constrained to well operating treatment plants dominated by municipal (non-industrial) inputs.  TSS is a common reporting parameter for POTWs and this is indicative of performance.  There is also a relationship between treatment type and TSS (see Appendix 2), therefore these are known co-variates and would be covered by Treatment Type. Note also that Trickling Filters often have less stringent permitted values  -  as they are low hydrological retention periods, thus treat with less efficiency.  An examination of water quality values is a good way to make sure that that treatment plant is working appropriately.  In this context, "high TSS" for TF systems is contextually different than "high TSS" for AS plants (along with many other factors).
   * Geography  -  this factor is unlikely to be critical relative to the other factors.  Previous monitoring campaigns for D4/D5 (Hope et al. 2012; Wang et al. in press) in colder climates (colder than the southern half of the US) indicate this is not likely to be a factor.  Other monitoring campaigns P&G has conducted on volatile/semi-volatile fragrance chemicals (e.g., Federle et al. 2010) did not find geographic location to be a factor (sites range from California to Texas, the Carolinas upper mid-west to the northeastern states).
   * Receiving water body flow  -  in practical terms this term is actually not clear, but we are interpreting this to mean the relative level of dilution at a site, not velocity or total river discharge.  This is unlikely to be an influence depending on the goals of the work.  Again, in practical terms, the nationwide summer time low flow dilution which usually represents worst case, is very close to 1.  P&G has termed this Practical Dilution Factor and this is not the same as simple as instantaneous Effluent:River Discharge volume.  Since many treatment facilities exist along water bodies, their discharge also contributes to the condition downstream.  When considering these factors, actual dilution in-stream is close to 1 nationally.  Dyer and colleagues have discussed this concept recently with US EPA as part of broader discussions involving the use of the ACI iSTREEM river model.  This is the only model currently available, including USEPA national models, which can address upstream municipal sources within a watershed-based context.  Discharge into lakes from wastewater facilities is also less likely to be an issue in most of the combinations of factors involved (for example, high populations being served by a POTW in the south with low organic matter and high suspended solids release will virtually be impossible to find).
   * Organic content of receiving water  -  in order for this to be important, organic matter in receiving waters must be more competitive for D4/D5 than the effluent solids to which they are already associated (noting that finding "biologically free" D4/D5 in wastewater is highly unlikely given present monitoring efforts.  Organic matter in wastewater will dominate this phenomenon and will already be covered by assessing a variety of wastewater treatment plant types which will likely different solids loading to the environment.

In our view, the SEHSC proposal correctly identified that the dominant variable of concern is the type of wastewater treatment under the assumption that monitoring is intended to only be conducted at well operated POTWs as sites. We also believe that to rely on precise definition of projected α and β in modeling numbers of sites (either SEHSC or the USEPA proposals) suggest a level of knowledge and distributions associated with the individual factors that is either imperfect or simply not present.  For example, distribution of environmental parameters such as organic carbon, dilution, and TSS are decidedly non-normal and are not independent.  The use of best professional judgment at some point is a necessity.  Based on the underlying context of the parameters chosen by USEPA as important to factor in, it appears the underlying assumption would be that sediment would be the primary compartment of interest:
   * TSS as a concern being a vector to sediment
   * Low flow as a concern being associated with the likelihood of having depositional regions
   * Organic matter as a concern being associated with deposition as well.


What is missing?
   * Assessment of the percent contribution of industrial discharge to sites.  This is likely a critical factor for deselecting sites.
   * Assessment of the permit compliance situations for sites.  This is another critical factor for deselecting sites.
   * Any monitoring effort involving POTWs requires the POTW as a complicit cooperating party.  The breadth and intent of the monitoring effort (by either proposal) will not be welcome at many POTWs without full protection for anonymity which will be virtually impossible in this instance.  This does constrain what can be done and where.
   * Assessment of state level information that can be used to inform what biological condition exists at potential sites.  In fact, we would strongly encourage a discussion that does not de-couple POTW or manufacturing site monitoring from the in-stream biological and chemical monitoring effort.  These are intimately linked to the goals.  
   * The need to understand the relationships of physical factors, such as habitat, when working to assess the potential environmental effects of chemicals is well articulated by G. Allen Burton et al (2012, Environmental Toxicology & Chemistry, 31:  1-10).  Dr. Burton is a frequent participant in USEPA SAB and US NAS panels and is the current Editor in Chief of the journal Environmental Toxicology and Chemistry.
   * A clear acknowledgement about what is already known from a physical-chemistry, environmental monitoring, and biological condition point of view that can inform the appropriate path forward.  This would be supplemented by what is known from experiences in constructing similar monitoring efforts in the very recent past.  If all sites proposed using D-Optimal Design are to include biology this would become one of the largest such efforts, if not the largest, ever undertaken to relate environmental exposure pathways to biological effects.

We would recommend using the WERF Trace Organic Contaminants (TOrCs) Program as a template for site selection and prioritization as this represents the state-of-the-science.  Their recent report, CEC5R08, has been recognized as a basis for discussions in other venues, such as the Great Lakes Contaminants of Emerging Concern discussions held under the International Joint Commission with USEPA, Environment Canada, and scientific stakeholders.  Importantly, the approach can accommodate both site-specific, watershed, and national scale types of assessments.  The consumer products industry has used the tools outlined therein for developing watershed risk assessments for a large array of consumer product chemicals discharged from wastewater in the Trinity River, Texas (Atkinson et al., 2009).  The tools were also applied in an a very complex evaluation of aliphatic alcohols in POTWs, water, sediment, and biota to understand microbially generated alcohol from that associated with consumer products in several watersheds across the US (DeLeo et al. 2011,  Mudge et al. 2010, 2012).

Finally, in reiteration, we share the opinion that 5 sites are sufficient to achieve the goals of the program if stated as Goal 1 above as this is built upon much previously developed knowledge of the environmental fate, effects, and behavior of D4 and D5.

Literature

Burton GA, De Zwart D, Diamond J, Dyer S, Kapo K, Liess M, Posthuma L.  2012.  Making ecosystem reality checks the status quo.  Environ. Toxicol. Chem.  31:  1-10
Degao Wang et al. (In Press). Concentrations of cyclic volatile methyl siloxanes (cVMS) in various environmental media from Southern Ontario and Southern Quebec. Water Science and Technology Directorate Environment Canada.

DeLeo P, Mudge SM, Dyer SD 2011. Use of market forensics to estimate the environmental load of ingredients from consumer products.  Environ Forensics 12:349-356.

Federle TW, Sun P, Dyer SD, and Kiel B. 2010. Aquatic and terrestrial risk assessments for polycyclic musks based on U. S. distributions of measured concentrations in effluent and sludge.  Platform presentation at the Annual Meeting of the Society of Environmental Toxicology and Chemistry-North America, Portland, Oregon.

Hope BK, Pillsbury L, Boling B. 2012. A state-wide survey in Oregon (USA) of trace metals and organic chemicals in municipal effluent. Sci. Total Environ. 417/418:263-272.

Mudge SM, DeLeo PC, Dyer SD. 2012. Quantifying the anthropogenic fraction of fatty alcohols in a terrestrial environment. Environ. Toxicol. Chem. 31:1209-1222.

Mudge SM, Eadsforth C, DeLeo P, Meier-Augenstein W. 2010. What contribution do detergent fatty alcohols make to sewage discharges and the marine environment? J. Environ. Monit. 12:1846 - 1856.

 



 
Appendix 1

Important Data Sources Used for Site Selection
   * iSTREEM
         o National watershed model for 48 coterminous United States.  Built off of USEPA's Enhanced Reach File 1 with WWTP inputs from Clean Water Needs Survey and Permit Compliance System (location).  Includes raw drinking water intakes from Safe Drinking Water Information System.
         o Considers upstream contributions for mean and 7Q10 low flow conditions, hence provides bounding conditions for realistic exposures.  Upstream conditions are always a potential confounding factor as WWTPs are distributed along a river. 
         o Can provide WWTP influent, effluent, receiving water concentrations and raw drinking water concentrations based on per capita use.
         o Useful for determining the relationships of dilution with treatment type.
         o Is the only model currently available, including USEPA national models, which can address upstream municipal sources within a watershed-based context.
   * Clean Water Needs Survey (CWNS)
         o Core dataset for iSTREEM
         o Provides per capita served per WWTP as well as percent industry (proportion of flow)
               # The type of industry is not identified
         o Provides information on the treatment trains
               # For example, AS/TF is an activated sludge plant with a trickling filter unit 
               # A frequency analysis by Procter & Gamble (unpublished) on the 2000 CWNS showed that there were nearly 200 treatment train combinations.  Twenty treatment processes accounted for 89% of all WWTPs in the dataset.  The top five treatment types were (in order of frequency):  activated sludge (AS), stabilization pond (SP), aerated lagoon (AL), oxidation ditch (OD), trickling filter (TF).  These comprised 63% of the dataset.  The remaining proportion of plants did not contain data describing secondary treatment processes or was found to be a combination of two or more treatment processes (e.g., AS/TF  -  activated sludge with trickling filter).  
   * Permit Compliance System (upgraded to ICES-NPDES)
         o Core database for compliance of all dischargers
         o Provides monthly water quality data (ammonia, BOD, COD, TSS, TDS, etc...) as well as compliance metrics (codes for various types of permit violations)
               # Procter & Gamble has used violation metrics as a method of deselecting sites for monitoring of consumer product chemicals in STPs and receiving waters.  Significant Non-compliance Errors (SNCE), violations, are typically water quality exceedances (e.g., ammonia, TSS, BOD, or toxicity).   These exceedances may also be associated with inadequate treatment capacity at a site.  If the goal is to understand the distribution of siloxane, these are confounding factors that can be excluded by limiting use of sites where treatment plants operate within their legal limits.  
                     * In another Procter & Gamble Study (unpublished), it was found that violation codes denoted in PCS as "E90" (effluent parameter exceeding permitted limits) have been used as a general descriptor of plant operating performance.  Significant Non-compliance Errors (SNCE) describing: 1) permit scheduling/paperwork, 2) significant effluent permit exceedences, and 3) non-receipt of influent/effluent monthly/quarterly measurements were have also been used to determine STPs to avoid in monitoring programs.
               # WERF's Trace Organics Program has used PCS data to derive risk scenarios (see attached WERF Report CEC5R08C and Executive Summary-Eco Workshop 4-29-08)
                     * Note:  while the WERF program was devoted to a mixture of down the drain chemicals (e.g., contaminants of emerging concern, ammonia, others), the process of identifying potential sites where causality may be addressed is very pertinent to hypothetical risks and verification needs regarding siloxanes.  Note also that this program was born out of Trace Organics workshop in which USEPA, USGS, USDA and state environmental agencies participated.  Some scientists may consider siloxanes as CECs    
         o Used to screen for historical adequacy of treatment
   * State EPA/DNRs
         o State agencies are responsible for understanding biological status.  Instream sampling is used to verify sufficient biological condition.  Habitat quality is also a critical determinant in understanding potential impacts to instream biology.
   * On-the-ground Verification
         o All screened sites need to be verified via phone calls (i.e., make sure treatment types are accurate, dilution, presence of biota, etc.) and then via direct observation (i.e. what a treatment plant operator is often biased, hence a set of the same people should verify each potential location before final selection) 

Appendix 2

Analysis of the relationships between treatment, % design flow and water quality (total ammonia, BOD and TSS), and PCS violations (P&G, unpublished).  The Figures below describe these relationships.

  
Summary:  Stabilization Ponds and Aerated Lagoons provide the poorest treatment, particularly for TSS, Ammonia and BOD.  This is due to their purpose to grow bacteria and algae which are often discharged in the effluent.  There are significantly more Significant Non-compliance Errors (SNCE) for these treatment types.  STPs within 75-100% design capacity do best in removing contaminants.  Oxidation Ditch and Activate Sludge plants appear to be roughly equivalent in treating for conventional pollutants.




                  
