Conference Call

May 8, 2019 Slides


Bruce McLaughlin, Patricia Bratcher, Cesar Blanco, Mike Berry, Michael Prowatzke, Carl Dealy, Mike Urkov, Corey Phillis, Felipe Carillo, JD Wikert, Chester Lindley, Rod Wittler, Sadie Gill, Erica Meyers, John Kelly, Mike Hendrick, Mike Beakes, Tanya Sheya, Robyn Bilski, Shelly Hatleberg, Adam Duarte


  1. Update on Science Coordinator (Cesar)

    In HR's hands right now, but they are backlogged. It has been identified as a key position for the program so just waiting to hear back from HR so they can announce they are hiring.

    Chester Lindley is leaving for CDFW. His last day is this Friday.

  2. Update on habitat decay proposal (Rod)

    Working on looking at flows in Keswick. Also, a 2016 gravel augmentation (lasted 3 years). Talked to Scott Kennedy and what the discharge was when gravel mobilized (about 13,000 cfs). Looking at release flows from Keswick, specifically above 13,000 cfs. Plotted to 1991 through March 2019. Spikes on plot are magnitude/duration of flows above 13,000 cfs. Then he will work with John Hutchings and John Hannon to include records of gravel placement to place on the plot and look for patterns (# cfs-days and see if there is a constant number relating to gravel replenishment). Should have this done by the next SIT call. If it works, we could ascertain what movement is per site and it would be easy in the model to accumulate days and identify when gravel may need to be replenished. Adam – is the idea to identify when gravel is needed or how much gravel is already there? Rod – want to identify when we would need to replenish gravel, not looking at area of available habitat. Adam – wondering how to tie to DSM inputs. Rod – as the # cfs-days limit is reached, the value would go from 1 to 0. Let Rod know if any other ideas. Tricia – high flows are great from a geomorphic perspective, however, some high flows have pushed gravel up on the banks and where it is deposited and no longer usable. As we move forward, it would be helpful to see if/how lower flow events could be used to move gravel further downstream. Higher flows seems to 'perch' gravel in areas and then it's not usable. Rod – please send that in a paragraph to him. Mike Berry – gravel placement is upstream of tribs so release data is probably pretty good, but some smaller tribs flow in when Keswick closed. Typical summer flows is at 13,000 cfs, so no real movement until you get above that. If 13,000 occurs too frequently, maybe look at 13,500 cfs instead. Rod – gravel may be placed at other times (when money, permits available) so there are other factors besides flows. Adam – this is very important to get resolved for use in the model.

  3. Update on status of watershed meetings (Mike U.)

    Will start scheduling as soon as we have updated version of model. Then it will take about a month to work up presentation to share with groups. Probably two months from when results are seen.

  4. Update, description and uses of the Strategy Builder Shiny App (Sadie)

    Creating a way to specify inputs. Which strategies result in outcomes. Model will send back results of scenarios. Don't have new model yet, but will be ready to show something by end of the month. Adam – this is a good tool; need to think about tools and constraints (5-year timeframes).

  5. Using decision analysis to develop state-dependent strategies (Adam)

    The SIT has spent a lot of time developing models, but we need to maximize how to use. See Adam's presentation (attached).

    Make sure we're talking about the same thing

    • Evaluate trade-offs in management actions – we can look at current conditions and look at optimizing and developing strategies

    • Incorporate known issues/uncertainty

    • Identify key uncertainties

    • Forward simulation – evaluating strategies moving forward

    Kevin McDonnell's dissertation work (not available yet). Life-cycle model (used FY16 priorities). Course resolution (watershed scale at monthly intervals). Looked at three different decision alternatives (add gravel, excavate floodplain, do nothing). Identified key uncertainties and which uncertainties matter for simulation results. Used DSM to develop a transition probability matrix to evaluate trade-offs in management actions (gravel addition vs. floodplain excavation).

    Mike Beakes – can run optimization routine through model with constraints correct? Adam – yes, you can.

    Adam – can break up by different tributary types. Kevin also broke out by diversity group.

    Can look at spawning, rearing habitat and how tribs work based on this. Can expand based on what SIT would like to consider. Can then do forward simulation and look at what tribs have what (spawning/rearing habitat, redds) and decide what optimal decision should be considered (what strategy results in highest cumulative utility). Can start comparing strategies as a result as well and look at returns on investments.

    Mike Beakes – SIT has four or five fundamental objectives. Adam – can weight all equally or differently in order to combine values to maximize.

    Possible next steps – use current SIT submodels and use range of DSM inputs to develop matrices. Solve the model using stochastic dynamic programming. Forward simulate optimal policies.

    Things to consider – watershed/trib types (mainstem vs trib vs others?). Constraints (how many actions can be done every year/ every five years? Penalty constraint?)? System states (amount of habitat for redds, juvenile survival)? Others? Rod – constraints are really hard questions right now due to politics. We should pick questions that illuminate decision possibilities. Adam – you don't have to agree on a single strategy constraint. Spreadsheet with different strategies (Corey's) could be used to start? Rod – that was a straw dog, but it could be further developed.

    Tricia – would like to somehow acknowledge listed species needs/considerations a little more. Adam – good suggestion. Could do something like this for each species. And then when combined. Need to not focus on a single taxa or run.

    Tricia – floodplain excavation is a bit unnerving. Is it defined somewhere? Adam – SIT not currently looking at, but it's something from Kevin's paper. Tricia – floodplain availability? Rod – floodplain excavation is very important as a management action (CVO/BiOps). Tricia – doesn't promote this treatment on tribs without dams. Would promote more flows instead. Adam – SIT looks at things like increasing floodplain habitat, but not the mechanism to achieve that.

    Adam – Rod or Shelly send Excel spreadsheet (straw dog) to folks to help come up with strategies. Rod will send to Adam.

  6. Update on survival and movement model (Adam)

    Looking at Kevin's model, we see that juvenile survival matters. FY17 priorities, survival matters there too. If high survival, want them to stay around longer (not moving downstream). Need to know what that means. Not easy to estimate. Went on tangent about how to do. Integrated different data sets (carcass surveys, escapement, etc.) and can get rough estimate of juveniles per trib. Used multistate model. Wanted to make sure actual possible. Simulated low/high amounts of different parameters. Found that modeled survival pretty well. Raw data was obtained from Felipe (CAMP) but output information varies greatly. Worked with Chester and got estimates of juveniles, etc. Used CAMP estimates directly. Don't need to try and replicate. When fitting model, looking at Grandtab, CAMP, pre-spawn survival (carcass data), habitat, fecundity and egg to fry survival (Hammersmark's American River data).

    Had to make call what year/months included and what tribs. Clear Creek, Feather River (missing temp data), American River, Moke River (missing 2005 catch data), Stan River. Model for Stan River still running. Other models completed this morning. Model parameters are feeding off each other. Clear Creek graphics shared. Chris working with Flowwest on American River data (strange outcomes). Moke River also needs work (Adam to check with Sadie on data).

    Next steps – finish Stan model; make temp a covariate for survival; might consider making floodplain covariate; work with Flowwest to update habitat estimates; use these empirical estimates and new DSM and inputs when calibrating DSMs.

    Chester – where is temp data from? Most RSTs collect temp data as well. Particularly Feather River. Can get in CAMP platform (will try and get before leaving, otherwise Felipe can provide)

  7. Other issues or topics?


  8. Next meeting Call-in Meeting on Wednesday, May 22 @ 10:00