Tutorial: Trade-offs between policy objectives

Fishing policy exploration for Anchovy Bay

We will explore trade-offs between fisheries rent, social benefits, mandated rebuilding of species, ecosystem structure and diversity. The methodology was originally described by Christensen and Walters (2004)^[1], but has since been modified to include species diversity (by default using the Shannon diversity index).

1. Use the Anchovy Bay model in a version that is fitted to time series (or download). Once loaded, load the Ecosim scenario, (but not the time series), check that you have vulnerability multipliers saved, not just default values (Ecosim > Input > Vulnerabilities). Make a run to check model.

Open the fishing policy search (Ecosim > Tools > Fishing policy search), select Initialize using Random F’s, and set Number of runs to 3. Click Objectives on the tab at the bottom part of the form. Set the search objective for Net economic value to 1. This will result in three searches, all seeking to optimize the net economic value.

1. Do a search. Examine and evaluate the results – at the end of the search the estimated fishing efforts are transferred to the Ecosim > Input > Fleet effort form, so you can simple run Ecosim to explore the results from the optimizations.
2. Are the three runs similar?

1. Try a similar search for optimizing for social considerations (jobs).
   1. Do a run and display the absolute value of the landings (Data to plot > Value (absolute)
   2. Reset the fleet effort for all, run again, and check what the original values were.
   3. What has the optimization done?
2. The policy search works by fleet, and in this case the trawlers catch both cod and whiting. It is very likely that if we split the trawlers in two fleets, one for cod and one for whiting, that the optimizations would keep whiting down in order to rebuild cod. You can try this on a copy of the database.
3. Try to optimize for Net economic value (1), Social value (1), and Ecosystem structure (1)
   1. Is this a more balanced outcome?
   2. Using the same weight on different objectives does not in any way guarantee that all count the same. What matters for this is how easy it is to change the objectives, and the weights may have to be changed to reflect that. There is no rule of thumb for this, you will have to use trial and error combined with evaluation of runs.