April 27, 2018

Sacramento, CA

Participants

Ben Geske, Jim Peterson, Dave Mooney, Josh Israel, Erin McCreless, Mike Hendrick, Mike Urkov, Anna Allison, Chuck Hanson, Griffin Hill, Denise Reed , Jim Hobbs, Sally Rudd, Corey Phillis, Brad Cavallo, Gabrielle Boisrame, Rainer Hoenicke, Ken Kundargi, Will Smith, Bruce DiGennaro, Lauren Hastings, You Chen Chao, Victor Pacheco, Scott Hamilton, John Callaway, Dennis Murphy, Matt Nobriga, Rene Henry

Presentation by Sally Rudd, Compass Resource Management LLC

Structured Decision Making for Delta Smelt

Compass specializes in facilitating decision-making processes and using collaborations to solve environmental management decisions. The work Sally described was a demonstration project to show how SDM can be applied to prioritize management actions for Delta Smelt in the Bay-Delta. The project evaluated the 13 Resiliency Strategy actions identified by the state to improve the status of delta smelt,with a goal of determining which actions should be implemented over the next few years.

Compass established a Technical Working Group to populate a consequence table and define build-out scenarios evaluating how each action would likely affect smelt. The next step was to define clear objectives and connect these to each action through means-ends diagrams. Many of the actions had common effect pathways, meaning they influenced the same means objectives. The group also considered other ecological and socioeconomic values that would be affected by the actions. The expected consequences of each action for delta smelt were scored on a scale of -3 to 3; this relatively coarse approach was taken due to time constraints and because the team decided it was sufficient to give a general sense of the effects and tradeoffs of the actions.

The A partial life cycle model developed around the Rose (2013) bioenergetics model originally embedded in the Rose et al. (2013)individual-based life cycle model, was used to simulate the outcomes of each action relative to baseline conditions. This model was chosen because the spatial and temporal scales matched well with those of the proposed management actions. Sally gave a few detailed examples of how the model allowed the use of quantitative metrics to compare different actions and how they would change delta smelt biomass and survival relative to baseline conditions. This was done by breaking down different mechanisms that might be influenced by an action (i.e., effect pathways). The modeling process did not include people other than natural resource managers, e.g., engineers, because the goal of the process was to determine whether different actions would have benefits, but not to investigate whether or how to implement each action. The next step over the coming months will be to work with CSAMP (Collaborative Science and Adaptive Management Program) parties to identify and evaluate the comprehensive set of actions to benefit delta smelt.

Discussion of this group's SDM process for Delta Smelt

This meeting was structured similarly to the last one, in which the group built on the CVPIA SIT process focused on delivering Chinook to the Delta. This meeting was devoted to undertaking a similar process for delta smelt. There was discussion of whether to use the same type of model developed at the last meeting for Chinook, or to develop a new model for delta smelt. The group also discussed using multiple models, and comparing their results. There was also a discussion about how best to break the Delta up into sub-regions to better compare between each model.

Comments from attendees

• We need a full life-cycle model; Fish and Wildlife Service has been working on this for years so it's not practical for this group to reinvent the wheel
• It would be good to look at Delta models for salmon and look for overlaps for structuring the salmon and smelt models
• SDM is useful for considering needs of multiple species
• Should create models that use the actions identified in salmon-focused meetings because many of the same actions will benefit smelt

Details about the FWS life cycle model

• It's a state-space model
• Survival across four life stages
• Accounts for observation and process errors
• Scaling of model may not match the scale of management actions
• Will be good at quantifying broad / whole-system effects, but less good at fine-scale effects
• Relevant time steps likely depend on life stage

Discussion of temporal resolution needed for smelt SDM

• Mortality rate slows as fish age; so for young fish short term actions would be more effective, less so for older fish
• General consensus to use monthly time step
• This has the benefit of lining up with the timeframe of the Chinook model
• Some inputs can be broken down to fractions of a month

Discussion of spatial resolution needed for smelt SDM

• Jim showed how regions within the Delta will be divided for Chinook model based on Russ Perry's maps; are these reasonable for smelt?
• Probably not because Chinook model accounts for movement migration through the Delta toward the ocean whereas smelt are mostly sedentaryresident in the estuary
• In its BiOp for CA Water Fix, FWShas considered dividingdivided the estuary into regions that are permanently, seasonally, or rarely occupied by smelt; FWS is using this model as well
• Decision to use the Rose et al. (2013) map but with an extra region in the north (circled in blue below)

Development of conceptual models for each life stage transition – see network diagrams below

• Larvae to juvenile transition: discussion of whether food is a limiting factor and thus whether food addition would be beneficial. Suggestion that temperature and stress are the most important factors driving this transition
• Discussion of whether Sub-adult to Adult transition should be included in model
• Discussion of entrainment as an important driver of life stage transition

Discussion of outflow effects on smelt

• This wasn't included in the Rose bioenergeticsquantitative part of Compass' decision model but Sally Rudd's work did some outside modeling and then integrated it into their model

Discussion of hatchery origin and releases as management strategies

• Discussion about how many hatchery fish could/should be released
• One problem is that hatchery fish won't know how to feed or escape predators
• Alternatively, could release eggs; this is cheaper but riskier than releasing adults
• If adults were released, should probably happen in late fall when water has cooled
• A starting point for addressing this is to assess how many releases would be needed to get to a specified level of population growth; should releases take place during certain critical periods
• Ultimately the goal is to determine whether this is a good use of funds compared to other potential actions

Next steps and actions

• Modified Rose 2013 model is an R code so should be easily accessible; Will might have to explain it
• Scott Hamilton's model: When candidate actions have been identified, he can plug them into his model; the model is data intensive so requires actions that have high quality data
• Sally Rudd's work is well documented and organized and she is happy to share
• Mike Urkov: needs a list of data inputs that are needed; he can make a list of what exists
• Engineers who know about flow manipulation should be identified and attend the next meeting
• Influence diagrams for the different actions for Chinook and smelt will be sent around to the group for input

Map regions for Delta Smelt model (all same as color-coded with additional northern region highlighted in blue)

Delta smelt conceptual models for each life stage transition