eling results and over-dependence on these expensive models for new and changing problems. Conversely, simpler models are quicker and cheaper to develop, more easily understood (in terms of strengths and weaknesses), and could provide better simulations beyond the training data. The Hub organization, with the engagement of capable experts, could help guide the right balance between simplicity and complexity in model selection and development, as well as identification and prioritization of modeling improvements. This could improve the transparency and robustness of decision making that uses models.

Many of the activities listed above are done independently by individual organizations. The community of agencies (and their missions) would benefit from more collaborative activities across the board. The various modeling groups are primarily focused on the objectives of a single agency or department within an agency. There is little time or direction for these experienced modelers to interact with other agencies or even across programs within the same organization particularly when planning or developing innovations or the next generation of a model or method (refer to section on Siloing). The Hub could provide a forum and resources for key individuals to engage across boundaries through mechanisms such as dedicated Sea Grant Fellows and postdoctoral researchers to work within modeling teams, enhanced communications such as collaborative online communication tools, and CWEMF or other in-person activities. Some activities likely to reap the largest benefits from coordination across institutions include routine syntheses of existing data and research; identifying and testing new monitoring approaches; stewardship of existing models and identifying and supporting new modeling initiatives; and planning and executing integrated scientific studies that could include field, laboratory, and modeling initiatives focusing on major confounding issues such as those highlighted in Chapters 2–4.

A small core staff, which might include a program manager and technical staff, could manage the Hub. The technical staff would have skills in areas such as data science and modeling and would be able to work with experts across the Bay-Delta watershed on key issues such as model interoperability or new monitoring approaches. Some staff could be assigned on a project basis, either full-time or part-time, from contributing organizations (including agencies, tribes, nongovernmental organizations [NGOs], consultants, and academia). Specific project leadership would be identified by agencies or entities with particular interest in the action(s) that will be informed by the science or study outcome.

The Hub would need a physical location and a virtual presence, a modest funding base to support these assets and staff, and access to computational resources (e.g., by agreement with agencies and/or universities). Specific project activities coordinated by the Hub could access funding from additional sources, such as agencies, research programs, foundations, the private sector, and other entities. Projects could be funded as directed research or through a competitive grant process with clear articulation of the contribution to management actions. This process could be facilitated through existing avenues such as the Delta Science Program and California Sea Grant but would need more involvement from other major projects and regulatory agencies.

The Hub’s priorities could be developed from the Delta Science Plan (which includes the watershed and connections to the Bay), previous work by the Coordinated Science and Adaptive Management Program, the Sacramento River Science Partnership, agency/tribal priorities, and other sources. The Hub would also have flexibility to coordinate and mobilize scientific activities in response to episodic events such as an earthquake, unanticipated levee break, contaminant spill, or a newly identified invasive species determined to significantly influence the ecosystem and Project operations. An effective governance structure is critical if the concept of a Hub, however it is constituted, is to be successful.

Feasibility Study

The Committee recommends a multi-agency feasibility study to explore improvements to the management of science for water and ecosystems in the Central Valley, including specifically the concept of a Hub. The feasibility study should consider earlier assessments and concepts and describe the following:

1. The basic organizational structure of the Hub: What, Why, How, and Who.
2. A governance structure to ensure that the Hub is agency-led and enhances the in-house agency expertise and mission of agencies. Careful consideration should be given to a governance structure that ensures the

primary users of the science (including the State Water Resources Control Board, CDWR, and USBR) receive the information they need, develop shared scientific knowledge, and support complementary collaborative tools. The Delta Science Program is widely respected for facilitating, coordinating, reviewing, and working across agency and interest boundaries to identify common science priorities. However, this role does not currently extend to having responsibility to coordinate or implement a comprehensive interagency science program. Similarly, effective engagement with California’s academic institutions should also be explored within this feasibility assessment.

3. How priority activities would be selected and governed (expanding from the efforts of the Delta Plan Interagency Implementation Committee and the Delta Science Plan).
4. The (a) staffing, (b) how experts would be engaged (particularly those representing underserved communities, tribes, NGOs, academia, and groups with limited resources), (c) how computational resources would be accessed, (d) how data would be accessed and managed from different repositories, and (e) the tools and expertise that would be available to investigative teams.
5. A base budget for sustaining the Hub, and source of base funding support.
6. Additional funding mechanisms for supporting specific projects.
7. How problems to be addressed would be selected, scoped, prioritized, and implemented.
8. How this entity would ensure meaningful engagement of experts from interested parties, and how results are communicated.
9. Performance metrics, review process, and renewal cycle.
10. A potential charter, which would also examine any need for legislative establishment.

A well-scoped feasibility study of the Hub concept would provide agency leaders and policymakers with a pragmatic, structured framework to evaluate this opportunity. A feasibility study could be completed in less than 12 months, and the approach could be piloted by pursuing two or three demonstration problems with a timeline, adequate funding and other resources, and clearly specified deliverables. This approach would allow the concept to evolve, building on what has worked. The science hub is not a new concept, and the feasibility study could build on past ideas and include a review of other frameworks and their effectiveness, successes, and failures—including assessments done by the National Center for Ecological Analysis and Synthesis, the National Socio-Environmental Synthesis Center, the Southern California Coastal Water Research Project, and the John Wesley Powell Center for Analysis and Synthesis.

MODELING

California water management draws upon a broad array of numerical models, and the Committee is impressed with the quality and sophistication of the agencies’ modeling work. Nevertheless, the models have been developed by different agencies and different groups within agencies, academia, and other organizations. Some models have been repurposed for uses other than those for which they originally were developed, and some models were originally developed when computational science was less advanced. Most models have their origins in periods when California’s climate, water management issues and technologies, and water demands were significantly different than they are today or will be in the future. All of these limitations suggest that better integration of new data into models (rather than reliance on the historic record), better integration among models, development of new models that can outperform their predecessors, and better management of the modeling enterprise as a whole are key to better linkage of scientific knowledge and management decisions.

Computer modeling of aquatic systems and water management for the Central Valley has developed since the 1970s to be among the best in the world. Yet, the Central Valley and its modeling enterprise face new challenges from the diversification and intensification of water demands including higher evaporative demand; competition between native species that have adapted to a historical climate and non-native species, some of which are better adapted to current conditions; and the targeted end of groundwater overdraft by 2040, which currently supports a large proportion of water use. The following sections specifically discuss higher-level issues with CalSim applicable to all three actions reviewed in the previous chapters (and other actions).

or derivatives of CalSim simulations include peak winter flows, spring recession peak and duration, fall pulse duration and peak, ramping rates, and accurate minimum instream flows within reaches (CEFWG, 2021; Poff et al., 1997; Yarnell et al., 2015). These are examples of hydrologic output that cannot currently be generated from CalSim’s monthly average inflow. Reducing the CalSim time step would help encompass these metrics.

Moving to a weekly time step version of CalSim may seem major, but it need not happen all at once. This will require more detailed data sets including hydrologic inflows, and reformulation of the operational description of some parts of the system. Different approaches to providing finer time step results for ecological and flood management purposes could be explored.

Transparency

CalSim3 is an important planning tool, although there is much room for improvement. Although CalSim3 has been reasonably well documented (CDWR and USBR, 2002), some details on its usage and inner workings need better transparency—a difficult chore given the many agencies, consultants, model developers, and researchers that have provided upgrades for CalSim3. In addition, resources to train modelers across the wide swath of CalSim3 operations and continuous training of a new generation of CalSim3 modelers should be a priority. The CWEMF/CDWR efforts to run training sessions for CalSim3 for the modeling community are steps in the right direction.

Additional Processes

In recent years, higher-resolution hydrodynamic models and innovations in particle tracking have provided greater insight into many ecological processes. Examples include the work of Holleman et al. (2022), who combined telemetry of tagged hatchery fish with a high-resolution hydrodynamic model to quantify the range of swimming behavior of emigrating salmon smolts at a tidal junction. They found that smolts moved through the area more slowly than the mean flow velocity, which may have important implications for flow management in the Delta. Similarly, movement behavior for Delta smelt was investigated using behavior movement models that included high-resolution hydrodynamics and population models by Korman et al. (2021) to explore proportion entrainment loss. They found that behavioral rules based on concepts such as tidal surfing, movement toward more turbid water, or movement toward less saline water did not, on their own, explain the temporal and spatial variability in trawl catches or the temporal variation in salvage. Similar types of analyses that seek to understand species movement in response to management action using high-resolution hydrodynamic models have examined longfin smelt (Kimmerer and Gross, 2022) and the movement of copepods (Hassrick et al., 2023). Numerical experiments using high-resolution coupled models to test actions can reveal insights about local processes or short-term effects that are not captured in ecological models driven with DSM2 outputs. Such three-dimensional water flow, water quality, and ecological modeling capability can also expand the horizons of Delta management actions to include modifications of tidal flows in the Delta to support biological productivity and transport of food and organisms to desirable locations, and reduce movement of native species into unsuitable parts of the Delta.

Both CalSim3 and DSM2 lack turbidity and sediment transport capabilities that could enhance their use in OMR flow management. This deficiency precludes calculations pertinent to turbidity bridge avoidance, requiring use of measured turbidities at specific locations for OMR management. Scour is another major issue at some flow junctions (divergence regions), and sediment transport models coupled with hydrodynamic models such as DSM2-HYDRO or DSM2-PTM could simulate when excessive scour could harm fish (Bowen et al., 2009). Delta sediment/bed erosion models have been in development over past decades: for example, DSM2-General Transport Model for real-time operations and long short-term memory ANN suitable for CalSim3-based planning (Abrishamchi and Nam, 2019; Kim et al., 2022). Their operational or planning use awaits further developments. To this end, development of parsimonious models such as ANN salinity-flow relationships in CalSim3 would be useful.

CalSim’s origin as an operation planning model for the SWP and CVP has hindered its ability to represent the coordinated operations of many local resources and decisions within the larger California water system. The model misses much of the more sophisticated water banking, market, and exchange decisions made locally and regionally to integrate a broader portfolio of local, regional, and groundwater supplies, and many water demand

management activities are largely absent from state and federal modeling. These local and regional decisions are increasingly relied on statewide and greatly reduce the costs of shortages, droughts, Sustainable Groundwater Management Act compliance, and climate adaptation.

Siloing

Developing and upgrading models in an operational setting with many participants and stakeholders is difficult. There is a high degree of inertia in advancing modeling for the Bay-Delta system. Some of this inertia is because of the convenience of using models (and monitoring networks) already accepted and more usable in one’s agency or program (because of staffing, accessibility, ease of internal communication, or agency missions). Some is perhaps due to the fear, expense, or risks of an agency having to master an additional model or modify existing models for additional missions outside their normal operations. Furthermore, creating new monitoring to support new models can be expensive. This can be compounded by separation, within agencies, between model developers, system operators, and decision makers. All of these factors can lead to institutional modeling “silos” that provide comfort and expertise within individual agencies. However, such “silos” of modeling also impede communications, collaboration, and the development of common explorations and understandings of problems and solutions, which increasingly span several state and federal agencies. Modeling should help to bring together agencies and interests for common understanding, knowledge, and perhaps trust, both among agencies and across the entire stakeholder community. However, modeling currently often reinforces separation of agencies (and programs within agencies) and creates barriers to common analyses, discussions, and knowledge. The proposed Hub could help break down these silos and lead in the coordinated development and strategic application of models (new or existing) to controversial problems involving multiple agencies.

WATER MANAGEMENT IN A CHANGING CLIMATE

Each of the previous three chapters has discussed how various climate impact drivers (CIDs) will affect the three actions reviewed by the Committee. This section provides an overview of the CIDs, reiterates their effects on the actions, and recommends ways to improve the integration of climate-related impacts into modeling and project operations and planning. CDWR and USBR have adopted two different approaches. Increased coordination and a detailed comparison of these approaches would offer additional insights about uncertainties and avoid confusion or conflicting planning decisions resulting from divergent methodologies and predictions.

CIDs are defined as physical conditions of the climate system—such as average states, events, or extremes—that affect human or natural systems (IPCC, 2021). The eight drivers considered by the Committee are described below, along with their relevance to the three actions. This information is also summarized in Table 5-1, which includes three rankings (A, B, or C) that reflect whether the impact is direct or indirect, whether observed changes are statistically significant, whether there is consensus among models on projected changes, and whether the impact is positive or negative. Furthermore, because changes in CIDs can involve different characteristics—such as magnitude, intensity, frequency, duration, or timing—Table 5-1 specifies which characteristic is relevant for each driver. For example, streamflow changes refer specifically to shifts in the timing of streamflow, not its magnitude. For additional information on observed and projected changes associated with these CIDs, see Appendix A and Chapters 2, 3, and 4.

Annual Precipitation

Annual precipitation in this context refers to the long-term average of total yearly precipitation, typically calculated over a 30-year period or longer. Changes in annual precipitation are considered in terms of magnitude only. Observations indicate that California’s annual precipitation has decreased over the period 2002–2021 compared to the long-term average; however, future projections are highly uncertain, with models suggesting either a slight increase or decrease in annual precipitation (see Appendix A). The rankings in Table 5-1 are based on recent decreasing trends (2002-2021) in annual precipitation and their potential impacts on the three CVP actions.

Annual and Seasonal Temperature

Annual and seasonal temperatures in this context refer to long-term average air temperatures, calculated over a 30-year period or more. Only changes in magnitude are considered. Both observational data and model temperature projections show an upward trend across California (see Appendix A, Figure A-10). Table 5-1 shows that, along with drought, this CID is critical for all three actions, mainly due to the impacts of higher water temperatures on fish health. Rising temperatures alone could jeopardize the existence of wild coldwater species in the Delta and the Central Valley (Brown et al., 2016).

Snowpack

Changes in snowpack involve both decreases in snow depth and snow water equivalent (magnitude), as well as shifts in the timing of peak snowpack and snowmelt events (timing), both of which are discussed in detail in Appendix A. A decrease in snowpack is a critical CID only for the Shasta Coldwater Pool Management Action due to its impacts on formation and maintenance of the coldwater pool.

by creating a collaborative environment where new approaches can be developed and tested, where monitoring can be better coordinated, and where cross-agency training can be implemented. It would also facilitate better coordination across the science that supports CVP and SWP activities.

Recommendation 5-2: Finer-scale models are needed for water and ecosystem management in the Delta and its watershed.

In particular CalSim (which is now widely used to evaluate ecosystem impacts) would benefit from a finer time step (perhaps one week) coupled with explicit or empirical disaggregation to still finer ecologically relevant time steps. With such capability, CalSim (or some other model) could then provide systematic hydrologic information for the more fully coupled hydrodynamic and population models needed for ecosystem management. Innovative ecosystem-wide approaches and quantitative life-cycle models could help to identify management actions that can contribute to species recovery. Monitoring should be tailored and managed to provide data at the spatial and temporal scales needed by the models.

Recommendation 5-3: CDWR and USBR should coordinate more closely and develop shared standards and protocols for modeling climate change impacts on the CVP and SWP.

Consistent protocols are essential for minimizing uncertainty in projected climate change impacts, particularly those arising from two sources: future greenhouse gas emission scenarios and model-related uncertainties (e.g., climate sensitivity). Protocols will help to explain differences between projections and avoid confusion among policymakers and the public around different representations of climate change. Moreover, CDWR and USBR should regularly update their modeling protocols to ensure alignment with current best practices in climate impact assessment. Finally, given the increasing evidence of more frequent compound events (multiple extreme events occurring simultaneously) and cascading events (extreme events occurring in sequence), there is a pressing need for more detailed modeling studies to assess how such events may affect the Projects.

REFERENCES

Delta Stewardship Council. 2024. The Delta Plan and 2024 Five-Year Review. Appendix 1A- Best Available Science. https://deltacouncil.ca.gov/delta-plan/.

Hall, K. L., D. Stokols, R. P. Moser, et al. 2012. “The Collaboration Readiness of Transdisciplinary Research Teams and Centers: Findings from the National Cancer Institute’s TREC Year-One Evaluation Study.” American Journal of Preventive Medicine 42(2):117–123. https://doi.org/10.1016/j.amepre.2011.10.011.

Hassrick, J. L., J. Korman, W. J. Kimmerer, E. S. Gross, L. F. Grimaldo, C. Lee, and A. A. Schultz. 2023. “Freshwater Flow Affects Subsidies of a Copepod (Pseudodiaptomus forbesi) to Low-Salinity Food Webs in the Upper San Francisco Estuary.” Estuaries and Coasts 46(2):450–462. https://doi.org/10.1007/s12237-022-01142-1.

Hausfather, Z., K. Marvel, G. A. Schmidt, J. W. Nielsen-Gammon, and M. Zelinka. 2022. “Climate Simulations: Recognize the ‘Hot Model’ Problem.” Nature 605:26–29. https://www.nature.com/articles/d41586-022-01192-2.

Herrgesell, P. L. 2012. A Historical Perspective of the Interagency Ecological Program: Bridging Multi-Agency Studies into Ecological Understanding of the Sacramento-San Joaquin Delta and Estuary for 40 Years. California Department of Fish and Game.

Holleman, R. C., E. S. Gross, M. J. Thomas, A. L. Rypel, and N. A. Fangue. 2022. “Swimming Behavior of Emigrating Chinook Salmon Smolts.” PLoS One 17(3):e0263972. https://doi.org/10.1371/journal.pone.0263972.

IPCC (International Panel on Climate Change). 2021. Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change, edited by V. Masson-Delmotte, P. Zhai, A. Pirani, et al. Cambridge, UK and New York: Cambridge University Press. https://doi.org/10.1017/9781009157896.

Kim, H. S., M. He, and P. Sandhu. 2022. “Suspended Sediment Concentration Estimation in the Sacramento-San Joaquin Delta of California Using Long Short-Term Memory Networks.” Hydrological Processes 36 (10):e14694. https://doi.org/10.1002/hyp.14694.

Kimmerer, W., and E. Gross. 2022. “Population Abundance and Diversion Losses in a Threatened Estuarine Pelagic Fish.” Estuaries and Coasts 45(8):2728–2745. https://doi.org/10.1007/s12237-022-01101-w.

Kirchhoff, C. J., M. C. Lemos, and S. Dessai. 2013. “Actionable Knowledge for Environmental Decision Making: Broadening the Usability of Climate Science.” Annual Review of Environment and Resources 38(1):393–414. https://doi.org/10.1146/annurev-environ-022112-112828.

Korman, J., E. S. Gross, and L. F. Grimaldo. 2021. “Statistical Evaluation of Behavior and Population Dynamics Models Predicting Movement and Proportional Entrainment Loss of Adult Delta Smelt in the Sacramento–San Joaquin River Delta.” San Francisco Estuary & Watershed Science 19(1). https://doi.org/10.15447/sfews.2021v19iss1art1.

Luoma, S. N., C. N. Dahm, M. Healey, and J. N. Moore. 2015. “Challenges Facing the Sacramento-San Joaquin Delta: Complex, Chaotic or Simply Cantankerous?” San Francisco Estuary & Watershed Science 13(3). http://dx.doi.org/10.15447/sfews.2015v13iss3art7.

Lyall, C., and I. Fletcher. 2013. “Experiments in Interdisciplinary Capacity-building: The Successes and Challenges of Large-scale Interdisciplinary Investments.” Science and Public Policy 40(1):1-7. https://doi.org/10.1093/scipol/scs113.

Mahardja, B., S. Bashevkin, C. Pien, et al. 2025. “Heatwaves and Rising Temperatures in the Upper San Francisco Estuary: Trends and Impacts on Ecosystems and Humans.” San Francisco Estuary & Watershed Science 23(1). https://doi.org/10.15447/sfews.2025v23iss1art4.

OPC (California Ocean Protection Council). 2024. State of California Sea-Level Rise Guidance: 2024 Update. Sacramento, CA. https://opc.ca.gov/wp-content/uploads/2024/01/SLR-Guidance-DRAFT-Jan-2024-508.pdf.

Pierce, D. W., S. Rahimi, S. Iacobellis, D. R. Cayan, and J. Kalansky. 2023. LOCA2 Training Data. Data Justification Memo for EPC-20-006 Development of Climate Projections for California and Identification of General Use Projections. https://www.energy.ca.gov/sites/default/files/2024-03/01_BiasCorrectioninWRF_and_LOCA2Projections_DataJustificationMemo_Pierce_Adopted_ada.pdf.

Poff, N. J., J. D. Allan, M. B. Bain, J. R. Karr, K. L. Prestegaard, B. D. Richter, R. E. Sparks, and J. C. Stromberg. 1997. “The Natural Flow Regime: A Paradigm for River Conservation and Restoration.” Bioscience 47(13)769–784. https://doi.org/10.2307/1313099.

Ramón, C. L., M. Acosta, and F. J. Rueda. “Hydrodynamic Drivers of Juvenile-Salmon Out-Migration in the Sacramento River: Secondary Circulation.” Journal of Hydraulic Engineering (2018): n. pag. https://doi.org/10.1061/(ASCE)HY.1943-7900.0001484.

Rose, K., H. Jager, N. Monsen, Z. Bai, and E. Howe. 2024. Peer Review of the Fish and Aquatic Effects Analysis for the Long-Term Operations of the Central Valley Project and State Water Project. A report to the Delta Science Program. https://deltacouncil.ca.gov/pdf/science-program/reports/2024-04-23-long-term-operations-fish-and-aquatic-effects-analysis-final-report.pdf.

Schwarz, A. 2024. DWR Approach to Climate Analysis. Presentation to the NASEM Committee at Meeting 4, Davis, CA, August 12, 2024.

Sommer, T., C. Armor, R. Baxter, et al. 2007. “The Collapse of Pelagic Fishes in the Upper San Francisco Estuary.” Fisheries 270–277. https://doi.org/10.1577/1548-8446(2007)32[270:TCOPFI]2.0.CO;2.

Sridharan, V. K., S. G. Monismith, O. B. Fringer, and D. A. Fong. 2018. “Evaluation of the Delta Simulation Model-2 in Computing Tidally Driven Flows in the Sacramento-San Joaquin Delta.” San Francisco Estuary & Watershed Science. https://doi.org/10.15447/sfews.2018v16iss2art6.

Stokols, D., S. Misra, R. P. Moser, K. L. Hall, and B. K. Taylor. 2013. “The Ecology of Team Science: Understanding Contextual Influences on Transdisciplinary Collaboration.” American Journal of Preventive Medicine 42(2):97–105. https://doi.org/10.1016/j.amepre.2011.10.013.

USBR (U.S. Bureau of Reclamation). 2021. 2021 WestWide Climate and Hydrology Assessment (Version 1.2): SECURE Water Act Report Technical Report. Washington, DC: U.S. Department of the Interior.

Yarnell, S. M., G. E. Petts, J. C. Schmidt, A. A. Whipple, E. E. Beller, C. N. Dahm, P. Goodwin, and J. H. Viers. 2015. “Functional Flows in Modified Riverscapes: Hydrographs, Habitats and Opportunities.” BioScience 65(10):963–972. https://doi.org/10.1093/biosci/biv102.