BEGIN:VCALENDAR VERSION:2.0 X-WR-CALNAME:EventsCalendar PRODID:-//hacksw/handcal//NONSGML v1.0//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:America/New_York LAST-MODIFIED:20240422T053451Z TZURL:https://www.tzurl.org/zoneinfo-outlook/America/New_York X-LIC-LOCATION:America/New_York BEGIN:DAYLIGHT TZNAME:EDT TZOFFSETFROM:-0500 TZOFFSETTO:-0400 DTSTART:19700308T020000 RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU END:DAYLIGHT BEGIN:STANDARD TZNAME:EST TZOFFSETFROM:-0400 TZOFFSETTO:-0500 DTSTART:19701101T020000 RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU END:STANDARD END:VTIMEZONE BEGIN:VEVENT CATEGORIES:College of Arts and Sciences,College of Engineering,Graduate Stu dies,Lectures and Seminars,Thesis/Dissertations DESCRIPTION:Thesis advisor: Dr. Gavin Fay - Fisheries and Oceanography, SMA ST Committee members:ÌýDr. Yuchou Chang - Computer & Information Science. ÌýDr. Leander Hohne - Fisheries and Oceanography, SMAST Date and time: Fri day, August 21st 2026 at 01:00 pmÌý Location: SMAST East 101-103, 836 S. R odney French Boulevard, New Bedford MA 02744. Link: Join Zoom Meetinghttps ://umassd.zoom.us/j/99402231312?pwd=NWBAMQDRf7GCRokw1gEbUeyj4uHHyF.1 Meeti ng ID: 994 0223 1312Passcode: 125997 Join instructionshttps://umassd.zoom. us/meetings/99402231312/invitations?signature=IwGtdDUXSdtTFwwD7wLeTk8UizBo qmyfri87TyMXGAs Abstract:Ìý Fisheries management increasingly relies on bi ological reference points, target fishing mortality and biomass levels, to set sustainable catch limits. For single-species stocks these reference p oints can often be computed from a fitted assessment model, but when a sto ck's mortality is driven by multispecies predation rather, the reference p oints have no closed form and depend on the assumed status of every specie s in the system; obtaining them requires repeatedly re-projecting a multis pecies model forward, at every assessment year of simulations, and ignorin g this predation-driven mortality process is known to bias both stock asse ssment outputs and the reference points used for management advice. This c omputational cost is one of the bottlenecks in evaluating harvest strategi es against multispecies operating models, including one built here for the Eastern Bering Sea groundfish complex (walleye pollock, Pacific cod, arro wtooth flounder) using the Rceattle implementation of CEATTLE. Here we sho w that a gradient-boosted tree model, guided by age-structured natural-mor tality and trained on terminal-year assessment output from 500 closed-loop management strategy evaluation simulations, can emulate these multispecie s reference points, without re-running the dynamic projection, achieving c ross-validated R² of 0.80 - 0.95 across six targets. Feature-importance a nalysis reveals that a single pollock-specific natural-mortality feature d ominates prediction for four of the six targets, even for cod's and arrowt ooth flounder's own reference points, a machine-learned signature of the p redation coupling built into the operating model. This demonstrates that t hese emulator models can substitute for costly re-projection, offering a p ath to rapid screening of harvest-strategy across the simulation ensembles that multispecies management strategy evaluations increasingly require. F or further information please contact Dr. Gavin Fay at gfay@umassd.edu.Ìý Ìý\nEvent page: /events/cms/20260821-machine-learnin g-emulator-for-multispecies.php\nEvent link: https://umassd.zoom.us/j/9940 2231312?pwd=NWBAMQDRf7GCRokw1gEbUeyj4uHHyF.1 X-ALT-DESC;FMTTYPE=text/html:

Â鶹¹ÙÍø

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

\n

Commi ttee members:Ìý
Dr. Yuchou Chang - Computer & Information Sc ience.Ìý
Dr. Leander Hohne - Fisheries and Oceanography\, SMAST

\n

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

\n

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

\n

Link: Join Zoom Meeting
https://umassd.zoom.us/j/99402231312?pwd=NWBAMQDRf7GCRokw1gEbUeyj4u HHyF.1

\n

Meeting ID: 994 0223 1312
Passcode: 125 997

\n

Join instructions

\n

Abstrac t:Ìý

\n

Fisheries management increasingly relies on biologic al reference points\, target fishing mortality and biomass levels\, to set sustainable catch limits. For single-species stocks these reference point s can often be computed from a fitted assessment model\, but when a stock' s mortality is driven by multispecies predation rather\, the reference poi nts have no closed form and depend on the assumed status of every species in the system\; obtaining them requires repeatedly re-projecting a multisp ecies model forward\, at every assessment year of simulations\, and ignori ng this predation-driven mortality process is known to bias both stock ass essment outputs and the reference points used for management advice. This computational cost is one of the bottlenecks in evaluating harvest strateg ies against multispecies operating models\, including one built here for t he 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 natura l-mortality and trained on terminal-year assessment output from 500 closed -loop management strategy evaluation simulations\, can emulate these multi species reference points\, without re-running the dynamic projection\, ach ieving cross-validated R² of 0.80 - 0.95 across six targets. Feature-impo rtance analysis reveals that a single pollock-specific natural-mortality f eature dominates prediction for four of the six targets\, even for cod's a nd arrowtooth flounder's own reference points\, a machine-learned signatur e of the predation coupling built into the operating model. This demonstra tes that these emulator models can substitute for costly re-projection\, o ffering a path to rapid screening of harvest-strategy across the simulatio n ensembles that multispecies management strategy evaluations increasingly require.

\n

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

\n

Ìý

E vent page: /events/cms/202 60821-machine-learning-emulator-for-multispecies.php
Event link:

DTSTAMP:20260805T205318 DTSTART;TZID=America/New_York:20260821T130000 DTEND;TZID=America/New_York:20260821T140000 LOCATION:Online - Zoom SUMMARY;LANGUAGE=en-us:Machine-Learning Emulator for Multispecies Biologica l Reference Points in the Eastern Bering Sea Groundfish Complex UID:8f06e5cdbe7778278d2e94e8a3c7db31@www.umassd.edu END:VEVENT END:VCALENDAR