RMP Emerging Contaminant Workgroup Meeting

SPLWG Minutes

May 14th, 2008

Sources, Pathways and Loading Workgroup Meeting

May 14th, 2008

San Francisco Estuary Institute

Meeting Minutes

In attendance:

1. Introductions and Review of Agenda, Lester McKee

2. Update: Small Tributaries Loads Study #2 – Z4LA, Hayward

2a. Year 1 Draft report / modeling development

Summary:

Lester presented an update on the Zone 4 Line A Year 1 report. His presentation covered the general watershed and sampling location characteristics, as well as the field methods and hydrology for Year 1. The report is two-thirds complete and will be emailed out for workgroup review soon.

Comments/Questions: None except general disappointment that the data report was not complete (see Item #2c below)

2b(1). Year 2 Draft data to-date, Rand Eads

Summary:

Rand Eads presented a summary of HY 2008 preliminary data collected from Zone 4 Line A. He described sampling station equipment setup and showed pictures of the bed-mounted sampling boom. He explained that the sampling location had to be moved due to construction at prior sampling site. While the sampling site was only moved from the upstream side of the culvert to the downstream side, unfortunately, between the two sampling points, there was an active stormdrain connection (which delivered “foul” water during storms). The end result is that HY 2007 and HY 2008 data from Zone 4 Line A cannot be directly compared. Rand also discussed data collection limitations, namely the sample volume limitation of the automated equipment (once the 24 water sample bottles are filled, automated sampling ceases until the bottles are replaced). He showed hydrographs and turbidity levels from individual storm events, and pointed out the how turbidity tends to peaks right before water levels peak.

Comments: Interesting but it would be nicer to focus more on interpretation and less on data dump.

2b(2). Year 2 Draft data to-date, Alicia Gilbreath

Summary:

Alicia Gilbreath presented an update to the group about the data gathered to date at Zone 4 Line A for this year of monitoring (WY 2008, Year 2 of the study). She noted the late start to the season due to the construction work on the channel at the site. Of the Hg and PCB samples allotted for this year of study (32 each), 27 samples have been gathered. She showed the relationship between the SSC and Hg, and SSC and PCB, sampling. Both relationships were strong and lined up with the previous year’s data.

Comments/Questions: The question was raised by Arleen Feng and Mike Stenstrom of how the size of the particles affects this relationship. Lester – commented that during higher flows we see sand particles in the sampling bottles. Jim Kuwabara asked about dissolved methyl mercury. Lester said it constitutes up to 30% but that on average it is less.

The point was made that high flow events could confound this relationship if buried uncontaminated sediment is scoured during event.

2b(3). Year 1 Modeling Development, Michelle Lent

Summary:

Michelle Lent presented preliminary results from Zone 4 Line A watershed modeling. She described the watershed modeling software, HSPF, as well as the various inputs to the model including local land use data and meteorological data. She has a basic, uncalibrated hydrology model working, and later waterborne sediment and contaminants will be added into the model. Other future steps include adding in storm drains to expand the hydrology network beyond open channels and calibrating the model to observed water flow data (though calibration expected to be minimal since not many parameters to adjust in such a developed area - basically a concrete basin).

Comments:

Arleen Feng noted that there is error associated with the ABAG land cover dataset used in the modeling and that Michelle may want to consider other land cover datasets that go down to parcel level. Later in the discussion Mike Stenstrom noted that by downscaling/simplifying ABAG’s 110 land use categories down to 7, estimates of impervious surfaces could be changed up to 30-40%. Chris Sommers proposed digitizing aerial photographs for improved land use identification. A few comments/questions went out around the room in regards to “how much effort can we really afford to put into this?”

There was discussion about why the modeled hydrograph was different (generally lower) than the observed hydrograph. Arleen Feng suggested that the uncalibrated hydrograph could at least be scaled based on expected differences in precipitation between the modeled watershed and the precipitation gage used (in Oakland). Michelle supported the idea. Lester McKee noted that the difference in MAR is probably not the reason for the lower predicted modeled values, since scaling the hydrograph would only lower it even further because the modeled watershed typically receives less rainfall than the Oakland rain station. Jim Kuwabara asked how the model handles infiltration? Arleen suggested that the watershed be broken into subwatersheds of 150 acres.

Michelle said a more local rain gage is available, but the data comes in variable intervals making it difficult to work with. Arleen suggested looking at an Aquaterra report showing possible methods to mimic, including dividing watershed into subcatchments after adding storm drains. Eric Stein suggested eventually taking the time-step down to 1 minute intervals and Lester McKee reminded the group that such involved analysis would require a larger budget. Arleen and Eric commented on the land use data and deficiency and Eric suggested using landsat data (he has a published paper on this). Jon Konnan commented on the inconsistencies between AlamedaCounty and other counties. Arleen suggested that the rainfall data could be adjusted (scaled) up or down when gauges are not available in the watershed. Don and Arleen commented that the one day time step is too long

2c. Action Item: Year 3 - proposed analytes, Lester McKee

Summary:

Lester McKee presented the proposed analytes lists for Year 3 at Zone 4 Line A. He presented 2 lists. The “Option A” list shows the same analytes as done in Year 2. “Option B” proposes to drop OC pesticides and PAHs, reduce the number of MeHg and trace metals analyses, and add dioxins and perflourinated. The sets of analytes in the two lists would cost approximately the same for lab analysis.

Comments:

General discussion around the room about whether or not dioxins should be studied next year as some committees are moving towards studying dioxins while BASMAA has not yet been included in this broader discussion. After some suggestions about which analytes might be dropped and why, general consensus is attained at the end of the discussion that there is not yet enough data to support dropping analytes, although agreement exists that there are datagaps pertaining to dioxins. Two suggestions were proposed towards resolving the issue:

1) Perhaps a workgroup meeting should be scheduled to present the Year 1 results sooner than the regularly scheduled next meeting, and then the workgroup may be able to consider dropping some of the Years 1 and 2 analytes to make room for the proposed new analytes, and

2) Excess sampling budget from Year 1 and 2 could be used to add the new proposed analytes without dropping any of the others. Lester: No budget left from year 1 and 2. Year 1 budget used up. Year 2 budget used to complete some dry weather flow sampling since spring this year was very dry and we had samples left to apply to this question.

Action: Lester and Alicia to complete Y1 report and email to group for review.

Action: WG recommended same analyte list for year 3 with the exception of potentially adding dioxins pending upper management decisions.

3. Update: Modeling Sediment and Contaminant Loads

3a. Suspended sediment Loads from Bay Area Small Tributaries, Mik

Lewicki

Summary:

Mik Lewicki presented an update about the modeling work he is doing of suspended sediment loads from small tributaries in the Bay Area. Using data from the 29 USGS sediment gages in the Bay Area, Mik is attempting to find statistical relationships between a watershed characteristic (drainage area, annual discharge, annual peak discharge, land use, geology, slope, other) and suspended load. The best relationship he has found so far is the unit annual peak discharge correlated with unit sediment load. He has gone on to run these correlations for each of the counties, different types of geology, and different land use types. Remaining work includes correlating unit annual peak discharge with unit sediment load for different median watershed slopes, for areas with high number of construction sites, for mature urban watersheds, and for watersheds with land slides or forest fires.

Comments:

Arleen noted that ABAG has developed charts of development over time, and that documentation for the illustrations is in the text documents associated with the ABAG datasets.

Someone raised a question to Mik about whether or not he is using a multi-parameter approach and he responds that the datasets may be too constrained for such an approach.

Arleen noted bias between using data for new construction on uncontaminated sediments versus data for areas of redevelopment, assuming that the end goal is to use suspended sediment loads as a surrogate for contaminant loads. Phil also notes the how suspended sediment loads tend to balance themselves and therefore one should not go backwards from using measured suspended sediments to indicate land use correlations. Lester agreed and reassured him that we are not doing this.

Eric and Chris raised the concern about covariance between the parameters in the model. Eric also commented about the data outliers having a large impact on the slope of equations. Use “cooks D” to justify removing the outliers. Chris asked how the provinces were done. Mik answered that they were based on the work by Rantz but will clarify that in the report. Arleen asked about the Geology and how it was mapped. Mik responded that he did it using expert opinion based on Sarah Pearce’s (SFEI) experiences classifying it on a watershed scale as dominated by a particular class (most eroding to least eroding). Arleen asked about reservoirs. Mik said they are taken into account but again he will clarify in the report. Eric suggested for Mik to try a multidimensional analysis and to try to take into account co-variation. Arleen recommended that population may not be a good proxy for development and mentioned some ABAG documents describing development might be better.

General concerns were raised regarding this project’s purpose and methods as related to how data generated by project will be used for answering management questions. Arleen commented that the project is important because it will help BASMAA to decide where to sample in the future. Chris agreed saying it is one type of data that will help with the sampling questions.

3b. Guadalupe R. Load Model Year 1 – Hydrology and Sediment, John Oram

Summary:

John Oram presented initial results from Guadalupe watershed modeling. The motivation for the study is to better understand the distribution of PCBs in the watershed which result in a higher concentration of PCBs as a function of SSC during the rising stage of a storm event than during the falling stage. He described the watershed characteristics as mostly residential (58%) with some commercial (13%) & industrial (13%) areas and some undeveloped area (16%). There are numerous reservoirs in the higher elevation southern portion of the watershed, and so, for modeling purposes, he divided the watershed into a reservoir section and a non-reservoir section. He is focusing on the more developed non-reservoir section (the lowlands) but will use the reservoir releases as inputs into the model.

John showed initial uncalibrated predicted flow from a model run using just one set of precipitation data for the whole watershed and not including reservoir flow. The model clearly underestimates baseflow, which would be provided by reservoir releases. Then he showed reservoir release data, which looks approximately like the missing baseline flow, suggesting that the model will better fit observed data once reservoir flows are added.

The next steps are to 1) add in more precipitation data sets which will be associated with specific locations in the watershed, 2) add in reservoir flows, 3) calibrate and validate, and 4) run scenarios to determine lag times and long-term water budget.

Comments:

Arleen Feng noted that irrigation will be a water source in this watershed, and should be incorporated in the model. Chris Sommers commented that if reservoir data is not available for model hindcast, John needs to consider reservoir management changes over time. John replied that he has about eight years of reservoir data from SCVWD. Arleen advised not spending too long on hydrological calibration since often re-calibration of hydrology is needed when fitting model to contaminant data (citing AquaTerra). Trish commented about the number of pump stations in the watershed for ground water leading urban slobber unrelated to rainfall. Mike Stenstrom talked about goals? To refine loads estimates. To lean about how management can change loads. Jim K talked about the need to define the modeling in terms of biological targets. Chris responded by saying the model could be helped to inform fish sampling in various areas. Jim Kuwabara responded saying that there is an assumption that loading relates to biological targets and this assumption may be flowed. Chris responded saying that other parts of the RMP are presently exploring this.

4. Collaboration Opportunity: USGS NAWQA Contaminant Trends in LakeSediments, Barbara Mahler

Summary:

Barbara presents some of the study results to date (1996-2007) and future sampling sites for the USGS NAWQA Contaminant Trends in LakeSediments. This study involves drawing sediment cores from lakes throughout the nation with the goals of being able to identify trends in metals and organic compounds in U.S. surface waters, characterize relations between the trends and changes in land use and environmental regulations, improve our understanding of transport processes and fate of contaminants in aquatic sediment, and identify and quantify major urban sources of contaminants transported to lakes.

The results to date indicate that DDT, lead and chromium are showing decreasing trends, while mercury and zinc are showing mixed signals across the nation, and PAHs appear to be on the rise. This study also shows differences between contaminants in lakes with differing types of airsheds and fluvial inputs.

Barbara and her team are planning to sample in California in 2009 and are currently funded for 2 reference lakes, one in the Sierras and one in the Bay Area, with the potential to add on a reference lake and an urban lake in the Bay Area, provided funding from the workgroup. She asks the group for their local expertise on feasible lakes in the area and presents the criteria for selection in her PowerPoint.

Comments:

Jay Davis asked about whether or not the study plans to analyze pharmaceuticals. Barbara responded that they do expect pharmaceuticals to be accumulating in some lakes and that they do plan to analyze for these in lakes receiving wastewater effluent – more testing will be done in a Texas study. Eric asked which metals. Barbara said all the typical ICP-MS analysis. Jon Konnan asked about PCB congeners – Barbara said yes. Chris asked for a clarification on “reference”. Barbara said no mining, no roads, no current logging – often the watersheds are for water supply reservoirs. What about Cu treatment. Barbara said not a reason to eliminate but she needs to know. Chris asked if it includes lakes and reservoirs. Barbara said yes and ponds too. Jon Konnan asked about the differences between Asphalt seal coat versus coal tar seal coat. Barbara said coal tar is higher PAH concentration. Eric S mentioned their experiences suggested that PAHs were more associated with pyrogenesis (air sources). Arleen commented that part of the reason is that urban areas have more efficient hydrological transport.

Action Item:

Barbara / Lester to send out list of potential lakes for sampling along with the criteria and introductory material for the study (background info and time line) so that work group can make suggestions about which local lakes to sample. The workgroup will review the material and get their recommendations back to her within 2 weeks. A follow-up phone conference will be scheduled.