Machine-Learning Emulator for Multispecies Biological Reference Points in the Eastern Bering Sea Groundfish Complex

Friday, August 21, 2026 at 1:00pm to 2:00pm

Online - Zoom
Dr. Gavin Fay
gfay@umassd.edu
https://umassd.zoom.us/j/99402231312?pwd=NWBAMQDRf7GCRokw1gEbUeyj4uHHyF.1

Thesis advisor: Dr. Gavin Fay - Fisheries and Oceanography, SMAST

Committee members: 
Dr. Yuchou Chang - Computer & Information Science. 
Dr. Leander Hohne - Fisheries and Oceanography, SMAST

Date and time: Friday, August 21st 2026 at 01:00 pm 

Location: SMAST East 101-103, 836 S. Rodney French Boulevard, New Bedford MA 02744.

Link: Join Zoom Meeting
https://umassd.zoom.us/j/99402231312?pwd=NWBAMQDRf7GCRokw1gEbUeyj4uHHyF.1

Meeting ID: 994 0223 1312
Passcode: 125997

Join instructions
https://umassd.zoom.us/meetings/99402231312/invitations?signature=IwGtdDUXSdtTFwwD7wLeTk8UizBoqmyfri87TyMXGAs

Abstract: 

Fisheries management increasingly relies on biological reference points, target fishing mortality and biomass levels, to set sustainable catch limits. For single-species stocks these reference points can often be computed from a fitted assessment model, but when a stock's mortality is driven by multispecies predation rather, the reference points have no closed form and depend on the assumed status of every species in the system; obtaining them requires repeatedly re-projecting a multispecies model forward, at every assessment year of simulations, and ignoring this predation-driven mortality process is known to bias both stock assessment outputs and the reference points used for management advice. This computational cost is one of the bottlenecks in evaluating harvest strategies against multispecies operating models, including one built here for the Eastern Bering Sea groundfish complex (walleye pollock, Pacific cod, arrowtooth flounder) using the Rceattle implementation of CEATTLE. Here we show that a gradient-boosted tree model, guided by age-structured natural-mortality and trained on terminal-year assessment output from 500 closed-loop management strategy evaluation simulations, can emulate these multispecies reference points, without re-running the dynamic projection, achieving cross-validated R² of 0.80 - 0.95 across six targets. Feature-importance analysis reveals that a single pollock-specific natural-mortality feature dominates prediction for four of the six targets, even for cod's and arrowtooth flounder's own reference points, a machine-learned signature of the predation coupling built into the operating model. This demonstrates that these emulator models can substitute for costly re-projection, offering a path to rapid screening of harvest-strategy across the simulation ensembles that multispecies management strategy evaluations increasingly require.

For further information please contact Dr. Gavin Fay at gfay@umassd.edu.