Changes to the RMP Sampling Design Through 2014
Changes to the RMP Sampling Design Through 2014
There have been numerous changes over the years to the RMP in order to better address management questions and adapt to changing regulatory and scientific information needs.
Summary of Changes to the Sampling Design for Water and Sediment
2014 was the twelfth year of the probabilistic sampling design for long-term water and sediment monitoring, which employs the EPA’s Generalized Random Tessellation Stratified (GRTS) sample design (Stevens, 1997; Stevens and Olsen, 1999; Stevens and Olsen, 2000). This type of design is more appropriate for addressing the RMP’s overarching goals to collect data and communicate information about water quality in the San Francisco Estuary in support of management decisions. An important advantage of random station selection is that estimates of regional condition derived from a probabilistic survey will have a known level of uncertainty associated with them. Prior to 2003, a targeted sampling design was used. The targeted stations were purposefully located along the central axis of the Estuary as far from anthropogenic sources as possible to monitor ‘background’ concentrations of pollutants of concern. A subset of those historic water and sediment stations were retained from the original RMP monitoring design, established in 1993, to provide continuity in the long-term monitoring program.
The RMP water and sediment monitoring stations are located in six hydrographic regions of the Estuary. Random design stations are located in five of those regions: Suisun Bay, San Pablo Bay, Central Bay, South Bay, and Lower South Bay. Historic stations are also located in each of those five regions, and additionally at the confluence of the Sacramento and San Joaquin Rivers in the freshwater Rivers region of the Estuary. The sampling frames for water and sediment monitoring (the area within which stations were allocated), are the three-foot and one-foot contours of the Estuary at mean lower low water, respectively (based on NOAA’s NAD-83 bathymetry coverage). About seventy-two random water and sediment stations were allocated into the hydrographic regions. Biennially, a subset of the water stations are sampled in sequential order, increasing the spatial density of monitoring over time. For sediment, a station re-visit schedule was incorporated into the design to better evaluate trends over time.
The number of random design sites sampled in each region can change based on management decisions. The initial number of sites sampled in 2002 was based on a power analysis using existing, targeted site data and Water Board management priorities. A power analysis is generally used to evaluate the number of samples needed to detect a change in contaminant concentrations over time with a known level of statistical confidence. The initial random design recommended that 26 water and 40 sediment sites be monitored while maintaining a subset of 5 historic water sites and 7 historic sediment sites (a total of 31 water and 47 sediment sites). A second power analysis was conducted in 2006 using the random design data (Melwani et al. 2008). Based on those results for key contaminants of current concern and discussions with the RMP oversight committees, which include Water Board staff, the number of water sites was reduced from 31 sites to 22 sites per year beginning in 2007, while the number of sediment sites was maintained at 47 sites per year.
In 2007/2008, a new redesign review was undertaken by the TRC. After a statistical review and consultation with the RMP participants, the RMP decided to add wet weather sediment sampling back into the Status and Trends program and recommended that wet weather sediment sampling alternates with dry weather sampling. The addition of wet weather sampling (typically done in February) will provide monitoring of contaminants that have higher ambient concentrations during the winter when runoff increases. In 2014, Dry season sampling was reduced from eight random sites per region (n = 40) to four random sites per region (n=20), the same number of sites sampled during Wet season sampling. Sampling of the historic stations will not change, and samples from these sites will continue to be collected during each sampling event (maintaining one station per region plus the two Rivers stations (n = 7)). The change in design necessitated an update from a five-year repeat sampling cycle to a six-year repeat sampling cycle to allow for balanced alternating season sampling. See the Memorandum on our web page for more details. Sites sampled in 2013 and 2014 are listed in Appendix 2 for water, sediment and bivalve sampling.
For more information on the Status and Trends monitoring design, refer to the following articles and technical reports: Power Analysis and Optimization of the RMP Status and Trends Program (Melwani et al., 2008), Re-design Process of the San Francisco Regional Monitoring Program for Trace Substances (RMP) Status and Trends Monitoring Component for Water and Sediment (Lowe et al., 2005), and the 2000 Pulse of the Estuary. Additionally, the 2012 Annual Monitoring Report provides a detailed description all changes in water sampling design and methodology.
Summary of Changes to the Sampling Design for Bivalve Bioaccumulation Monitoring
The previous monitoring design includes the analysis of trace organics in bivalves biennially, and the analysis of trace metals every 5 years. In 2014, the SC and TRC decided to reduce the number of parameters analyzed in bivalves; PAHs, PBDEs, and Selenium are now monitored biennially and PCBs are included every eight years. The RMP no longer includes the analysis of trace metals in bivalves.
Other RMP Monitoring
The bivalve bioaccumulation sample design remains a fixed sample design because deployment of caged bivalves requires secure moorings. Based on the findings from a series of special studies between 2000 – 2005 intended to redesign and improve technical aspects of the deployed bivalve bioaccumulation monitoring component of the RMP, several changes were made in 2003. These included:
- Dropping three sites in the northern Estuary: Napa River (BD50) , Petaluma River (BD15) , and Horseshoe Bay (BC21) because only two to three sites were required per region to track long-term changes in contaminant concentrations.
- Deploying only one bivalve species (Mytilus californianus). Because of the reduced salinity range of the study area due to the dropped sites, the program was able to deploy one, fairly salinity tolerant bivalve species, which makes comparing bioaccumulation results between regions possible.
- Deploying bivalves in cages, rather than mesh bags, reduces the loss of organisms through predation.
- Discontinuing the bivalve maintenance cruise. This was discontinued in 2006 after a study conducted from 2002-2005 showed no significant difference in survival of bivalves in maintained and non-maintained cages.
In 2014, two stations were removed from the sampling design, Red Rock (BC61) and Davis Point (BD40). The monitoring design includes sampling for bivalves at 6 stations in the Bay and back-up samples will be deployed at three stations, BD20 (San Pablo Bay), BB71 (Hunter's Point/Alameda), and BA30 (Dumbarton).
Changes in Parameter Reporting
PCBs are available through the RMP web tool, Contaminant Data Display and Download (CD3). The Sum of 40 PCBs include the 40 historic target PCBs for the RMP. The Sum of 208 PCBs provides an index of the PCBs present in Aroclor mixtures. PCB 11 is excluded; it is abundant in some matrices but is derived from pigments and not Aroclors. PCB 11 does not have dioxin-like potency and has different sources than Aroclors. The Sum of 209 PCBs is provided solely for comparison to other studies that include all 209 congeners. SFEI does not recommend using this sum for comparison to any Aroclor-based thresholds (the TMDL target, OEHHA thresholds, etc.) - the Sum of 208 PCBs is better for that purpose.
To view a summary of Changes to the RMP as a PDF, please refer to Appendix 8 of the 2012 AMR, linked here.
Programs and Focus Areas:
Clean Water Program