Closed-book appointment exam · independently graded
Professor — Physical Oceanography. The candidate agent answered from its own knowledge, closed-book; a second, independent examiner agent graded it adversarially.
vaiu-sci-eaps-prof-ocean appointment benchmark (2026-07-16)Closed-book oral exam. Candidate agent adopted the vaiu-sci-eaps-prof-ocean persona (v1.0.0) and answered from its own knowledge; a second, independent examiner agent graded. Condensed candidate transcript below; grading report in vaiu-sci-eaps-prof-ocean-2026-07-16.md.
AI-transparency disclosure. Opened as an AI agent, never been to sea; reasons from governing balances; gives honest ranges over false precision; cites eponyms / flags folklore; no fabricated references. Governing habit reasserted: what balance is invoked, and does it hold here? and the bright line between measured (altimetry SSH, Argo T/S, moored current record) and derived (geostrophic transport, AMOC strength, bulk-formula flux) — error bars live in the assumptions.
Q1 — Wind-driven circulation. Surface Ekman balance fk×u = (1/ρ)∂τ/∂z → net Ekman transport M_Ek = τ×k/(ρf), 90° right of the wind in the NH (Ekman, from Nansen's ice-drift observation); the spiral is idealized (constant-eddy-viscosity/steady/deep) and rarely seen cleanly, but the integrated transport is robust. Breaks at the equator (f→0), within a wave period of the surface (Stokes drift/Langmuir), under strong stratification. Ekman pumping w_Ek = curl(τ/ρf): anticyclonic subtropical curl → convergence → downwelling; cyclonic subpolar curl → upwelling; ~tens of m/yr, driving the interior gyre. Sverdrup balance (curl of the depth-integrated momentum, linear/steady/frictionless interior): βV = curl(τ)/ρ; subtropical negative curl → equatorward interior transport (derived, not measured). Western intensification (Stommel): the β-effect collapses the frictional return boundary layer onto the western wall — a poleward-moving column gains planetary vorticity f and must shed relative vorticity, which only a western-boundary shear layer can supply with the right sign; Munk's lateral-friction version reproduces the narrow, intense jet + recirculation. Real boundary currents are nonlinear and shed mesoscale rings.
Q2 — Geostrophy. For Ro = U/fL ≪ 1 away from boundary layers/equator: fk×u = −(1/ρ)∇p → u_g = −(1/ρf)∂p/∂y, v_g = (1/ρf)∂p/∂x (flow along pressure contours, high pressure on the right in NH). Hydrostatic ∂p/∂z = −ρg. Thermal wind (combine + eliminate p): f∂u_g/∂z = (g/ρ)∂ρ/∂y — horizontal density gradients set the vertical shear, so a density section from T/S gives the geostrophic shear; the "dynamic method" integrates upward from an assumed level of no motion, recovering shear from data but leaving the absolute current uncertain up to a barotropic constant. Altimetry measures SSH (range → height above ellipsoid → geoid-referenced dynamic topography); the surface geostrophic current is read off the SSH slope. Observed = SSH field (geoid/orbit/tide/MDT errors); derived = geostrophic velocity + full-depth transport (reference-level assumption). Breaks at the equator (singular), in surface/bottom boundary layers (ageostrophic friction), and at the submesoscale/fronts (Ro→1); hydrostatic breaks only for small-scale convection/turbulence.
Q3 — Overturning circulation. Warm light water poleward in the upper ocean loses buoyancy at high latitudes (cooling + brine rejection), sinks, and returns equatorward at depth; in the Atlantic, the AMOC. NADW forms from Labrador + Nordic-overflow sources; AABW (densest) on the Weddell/Ross shelves. "Thermohaline circulation" is misleading — the overturning is not purely buoyancy-driven; it needs mechanical energy to mix dense water back up (diapycnal mixing from winds + tides, Munk–Wunsch) and Southern-Ocean winds drawing deep water up along tilted isopycnals (Ekman/eddy residual-mean); the "conveyor" is a pedagogical cartoon. Carries ~1 PW of poleward heat. Observed: RAPID-MOCHA at 26.5°N since 2004, mean ~17 Sv with large sub-annual variability; OSNAP's surprise that eastern-subpolar/Nordic density transformation dominates over the Labrador Sea. Even RAPID's "AMOC strength" is a derived integral (measured density/pressure via thermal wind + an Ekman term). The record is short (~2 decades) and sparse; so the century-scale trend and any tipping proximity are genuinely contested — proxy SST-fingerprint and sortable-silt reconstructions are indirect, and "the AMOC is already declining" is plausible but not established from direct observation (Stommel 1961 two-box salt-advection bistability is robust in simple models, its real-ocean relevance debated). Existence + rough magnitude robust; trend + tipping contested.
Q4 — Air-sea interaction & ENSO. Ocean and atmosphere trade momentum (wind stress τ), heat (shortwave in, longwave out, turbulent sensible + latent — latent usually largest), freshwater (E−P sets salinity), and gases. Turbulent fluxes come from bulk formulae (τ = ρ C_D U², latent ∝ ρ L C_E U (q_s−q_a)) with empirical transfer coefficients — so a "flux" is derived, its error bars in the coefficients. The mixed layer is homogenized by wind, waves/Langmuir (Craik–Leibovich), and convective cooling, capped by the pycnocline — the ocean's thermal/gas capacitor. ENSO is a coupled recharge oscillator: easterlies pile the western warm pool and tilt the thermocline (down west/up east, cold eastern upwelling); the Bjerknes feedback (weaker trades → warm east → weaker trades) grows El Niño; equatorial waves supply the memory (downwelling Kelvin east warms; off-equatorial Rossby reflect at the western boundary and return as upwelling Kelvin that discharges equatorial heat content and flips the phase — Jin's recharge oscillator, with warm-water volume as a predictor). This is an equatorial regime where geostrophy fails. The ocean has absorbed >90% of the excess heat and ~a quarter-to-third of anthropogenic CO₂; ocean heat content (Argo) is a clean measure of the planetary imbalance.
Q5 — Marine geochemistry. Two pumps: the solubility pump (CO₂ is more soluble in cold water; cold, dense high-latitude water carries dissolved carbon down as it sinks — tied to the overturning) and the biological pump (photosynthetic fixation → sinking marine snow → remineralization at depth; a carbonate counter-pump complicates it because shell-building releases CO₂). Both are derived export efficiencies. Acidification: CO₂ + H₂O ⇌ H₂CO₃ ⇌ H⁺ + HCO₃⁻ ⇌ 2H⁺ + CO₃²⁻; anthropogenic CO₂ raises H⁺ (surface pH −0.1 since pre-industrial, ~30% more H⁺) and consumes carbonate ion, lowering the aragonite/calcite saturation state and shoaling the saturation horizon — chemistry robust, ecological consequences an active area. Oxygen minimum zones form where high surface productivity overlies sluggishly ventilated water (eastern tropical Pacific, Arabian Sea): respiration outruns resupply; they host denitrification/N₂O and are expanding under warming (lower O₂ solubility + more stratification) — sign robust, magnitude uncertain. Radiocarbon ventilation age: ¹⁴C (half-life ~5730 yr) enters at the surface and decays once a parcel sinks, so its depletion dates last surface contact — the deep Pacific is "oldest" (order 1000+ yr), fresh NADW younger; caveats are the surface reservoir-age offset, water-mass mixing, and the bomb-¹⁴C transient (also a useful tracer, like CFCs/SF₆). Measured = ¹⁴C concentration; derived = ventilation age.
"What makes the ocean move, and why does it matter for climate?" — novice (two engines: wind → gyres/Gulf Stream, density → the sinking-and-creeping deep conveyor; the ocean as a heat/carbon sponge); undergraduate (the hierarchy of balances — Ekman/pumping/Sverdrup, western boundary currents, geostrophy + thermal wind from Argo/altimetry, the overturning; ~90% of excess heat, ~25% of CO₂); graduate (potential-vorticity dynamics, Munk (A_h/β)^⅓ and Stommel boundary layers, the ageostrophic submesoscale, the Munk–Wunsch + Southern-Ocean residual-mean overturning, and the observed-vs-derived discipline with AMOC trend/tipping flagged contested). Each level states what it simplifies.
vaiu-sci-eaps-prof-climate; owned the receiving end — the ocean absorbs >90% of the excess heat via mixed-layer contact + circulation into the interior, and ocean heat content is a clean imbalance measure.vaiu-sci-stat-*), warning that solo-improvised statistics produce false precision.Examiner verdict: APPOINT (field 5/5 rubric-correct, 0 fabrications; teaching 3/3; boundary 3/3 including the B2 statistics-referral split). Full report in the sibling result file.