National Standard 1 Guidelines Use of SPR reference

National Standard 1 Guidelines, Use of SPR reference points, and Incorporating uncertainty Grant Thompson Alaska Fisheries Science Center

Part 1: NS 1 Guidelines • National Standards in MFCMA since 1976 – In 1996, SFA changed MFCMA to MSFCMA • • • NS 1: prevent overfishing, achieve OY NS guidelines required by M(S)FCMA First published in 1977 Revised in 1983, 1989, and 1998 New proposed revision published in 2005 Final rule unlikely this year

What’s old: MSY • Largest long-term average catch possible under prevailing environmental conditions • MSY control rule – Constant catch – Constant fishing mortality rate – Constant escapement – More complicated functional forms • Must incorporate consideration of risk • Should be re-estimated when warranted

What’s old: OY • Optimum yield specification – Provides greatest overall benefit to Nation – Takes protection of ecosystem into account – MSY reduced by economic, social, ecological – Rebuilds any overfished stocks to MSY level • “Precautionary approach” urged – Targets set safely below limits – F should be lowered if B<Bmsy – F should vary inversely with uncertainty

What’s old: SDC • Two status determination criteria required – If F>MFMT, overfishing is occurring – If B<MSST, stock is overfished • MFMT must be ≤ MSY control rule • If possible, MSST should equal greater of: – ½ Bmsy – Minimum stock size at which stock can be rebuilt to Bmsy in 10 years given F=MFMT

What’s old: Rebuilding • Rebuilding required if F>MFMT, B<MSST • Rebuilding time period required – Tmin = rebuilding time given F=0 – Tmax = 10 yr if Tmin < 10 yr – Tmax = Tmin + 1 generation if Tmin ≥ 10 yr • Progress monitored at least every 2 years • No guidance on updating, extending plans

What might be new (1 of 4) • Nomenclatural changes – Thresholds (MFMT, MSST) become limits (Flim, Blim) – Overfished becomes depleted • Specification of core stocks and assemblages – Core = principal target stocks; may also include historical targets, bycatch stocks, vulnerable stocks – Assemblage = stocks that co-occur with similar productivity, but insufficient data for individual SDC – Assemblage should contain at least 1 core stock – Manage assemblages more conservatively than cores

What might be new (2 of 4) • OY control rules required – For each core stock (and assemblage? ) – Must be less than Flim – Should quantify social, economic, ecological factors • Blim definition “simplified” – 50% BMSY is new default – Range of natural fluctuations can justify other values • Exceptions to required specification of Blim – If OY control rule is as conservative as rebuilding plan – If data are insufficient

What might be new (3 of 4) • New formula for Tmax – Tmax = 10 yr if Tmin + 1 gen < 10 yr – Tmax = Tmin + 1 gen if Tmin + 1 gen ≥ 10 yr • Revisions to rebuilding plans – Treb can increase to Tmax if F targets met – F targets can increase if E(B(Treb))=Bmsy – Provision for new estimates of reference points – F cannot increase if stock not rebuilt at t=Tmax

What might be new (4 of 4) • Possible restrictions in present flexibility regarding timing of status determinations – Not in proposed NS 1 guidelines revision • “Is overfishing occurring? ” may change to “Was overfishing occurring in year X? ” • Possible scenario: – Set OFL for Y 2 based on assessment in Y 1 – Conduct Y 2 fishery so as to achieve catch<OFL – Set new Y 2 OFL based on Y 3 assessment – Compare Y 2 catch to new Y 2 OFL

Part 2: SPR • %SPR commonly used as MSY proxy – S-R data are usually noisy enough to give large errors in Fmsy, Bmsy estimates – However, %SPR is independent of S-R curve, because recruits in numerator, denom. cancel • Proxy values depend on assumptions – Clark (1991, 1993) showed F 35%-F 40% was a safe proxy for MSY in “typical” groundfish – Some recent studies of west coast rockfish suggest F 50%-F 60% better for those stocks

Part 3: Incorporating Uncertainty • Method 1: Set target so as to achieve constant probability of exceeding limit – Implies harvest rate dichotomy, with all F>Fmsy equally bad – Target has no optimality properties • Method 2: Set target so as to maximize expected value of risk-averse utility function – Implies harvest rate continuum, with any given overharvest worse than an equivalent underharvest – Target has clear optimality properties

Uncertainty: Examples of Method 2 • Alaska groundfish Tier 1 – Limit F = arithmetic mean Fmsy – “Target” F = harmonic mean Fmsy • Alaska groundfish PSEIS Alternative 3 b – Reverse engineer SRR from F 35%, B 35% – Closed-form solution exists, assuming linear dynamics, normal process & assessment error

Uncertainty: Alaska groundfish Tier 1