Lawrence Berkeley National Laboratory

The proposed Monterey Peninsula Water Supply Project (MPWSP) involves desalinating water produced from slant wells completed in sand aquifers along the coast of Monterey Bay in Marina, California. Aquifers in the adjacent Salinas Valley are used heavily for groundwater for agricultural irrigation, and seawater intrusion has been a longstanding problem in the area. As part of the CEQA process, a team led by the CPUC carried out groundwater modeling to determine the impacts of the MPWSP on groundwater in the surrounding aquifers. Following a change in leadership of the groundwater modeling effort, the CPUC requested LBNL hydrogeologists to carry out an independent and objective peer review of the original groundwater modeling that was used to support the Draft EIR published in April 2015. In our review, we re-created the workflow used by the original modeling team, reviewed conceptual models of the shallow subsurface in the Marina area, re-ran models using data files and executable codes provided by the CPUC, and compared the outputs of our modeling results against those presented in Appendix E2 of the Draft EIR. We found that the computer simulations carried out by the modeling team can be replicated using the input and executable codes provided to us. Agreement between the original output and our re-run results was mostly excellent (agreed exactly or differences were very small). Differences in simulation results can probably be attributed to machine round-off and cancellation errors. We also found that the groundwater model results may not represent the detailed response of the actual system because the conceptual model used for groundwater modeling of the shallow sands at Marina neglected to include an aquitard present in the subsurface (the Fort Ord Salinas Valley Aquitard, or FO-SVA). We recommend that future groundwater modeling include the FO-SVA. Finally, we found the initial and calibrated hydraulic conductivities in the simulation were higher by one to two orders of magnitude and the Dune Sand aquifer storativity* was low compared to values derived from nearby field data. This may be because the lack of FO-SVA in the model resulted in higher horizontal to vertical conductivity ratios in the aquifers than is typical and indicated by the field data. We recommend using results from surrounding field data to initialize the model in those areas.