about EART60071
1 · Exam Mapcourse.json · brief
Final: timed two-hour online assessment on Canvas, under examination conditions, during the examination period; 10 questions, same format as the formative assessment. Generative AI is prohibited.
Official rule “The assessment is open book: you may bring and consult your own notes during the assessment.” (EART60071 syllabus, 2026/27 1st semester)
| Likely task | Answer spine |
|---|---|
| Calculate | formula → SI units → substitution → unit + sense-check |
| Interpret output | changed input → plotted response → physical/numerical mechanism |
| Compare models | question → family/frame → explicit/omitted process → cost/limit |
2 · Model ChooserC1 topic map
| Frame | Variables | Module use |
|---|---|---|
| Lagrangian | move with parcel/path | adiabatic parcel |
| Eulerian | fixed grid; material moves | plume, column, shallow water |
Controlled comparison: state control, change one input, hold all else fixed, compare the named output, explain mechanism. Complexity is useful only when its assumptions and detail fit the question.
Practical workflow
| Stage | Action + evidence to retain |
|---|---|
| connect | SSH to remote server; check hostname and pwd |
| obtain/edit | clone supplied code; change the labelled input only |
| run | compile where required; record settings and control run |
| retrieve | SFTP from remote to the directory shown by lpwd; rename output descriptively |
| interpret | compare plots/numbers; identify mechanism, artefact and omission |
SSH ≠ SFTP: SSH runs commands remotely; SFTP moves files between remote and local directories. A repeated remote filename can overwrite the prior run, so preserve each output under a meaningful local name.
3 · Forward EulerC2
Nₙ₊₁=Nₙ(1−λΔt)
- Write differential equation.
- Replace derivative by forward difference.
- Compute the one-step factor before iterating.
- Reduce Δt and check convergence.
| Signature | Diagnosis |
|---|---|
| Δt↓ → exact curve | convergence |
| 1−λΔt<0 | unphysical sign flip |
| oscillator amplitude grows | method error; exact amplitude is constant |
| plot interval changes only | sampling, not numerical solution |
vₙ₊₁=vₙ−ω²yₙΔt
amplitude factor/step = √(1+ω²Δt²) > 1
4 · Advection + CFLC3
Cᵢⁿ⁺¹=Cᵢⁿ−𝒞(Cᵢⁿ−Cᵢ₋₁ⁿ)
𝒞=|u|Δt/Δx ≤ 1
Stable ≠ accurate. For 0<𝒞<1, repeated cell blending lowers/widens peaks: numerical diffusion. At 𝒞=1, constant 1-D flow shifts one cell per step. At 𝒞>1, growing ripples/negative values signal instability. Fix by reducing Δt or increasing Δx; refining Δx also requires a smaller Δt.
| Spike [0,4,0,0] | After one step | Read |
|---|---|---|
| 𝒞=.5 | [0,2,2,0] | conserved but smeared |
| 𝒞=1 | [0,0,4,0] | clean one-cell shift |
| 𝒞=1.2 | [0,−.8,4.8,0] | overshoot + negative |
5 · State + Lapse RatesC4
ρa=P/(RaT); Ra≈287 J kg⁻¹ K⁻¹
| Need | Method | Trap |
|---|---|---|
| moles | n=PV/(RgasT) | hPa→Pa |
| molecules | N=n·6.022×10²³ | mol ≠ molecules |
| density | ρ=P/(RaT) | T in K |
| parcel T | T₂=T₁−ΓΔz | km with K km⁻¹ |
Unsaturated parcel: ≈10 K km⁻¹ (9.8; rounded in exercises). Saturated: representative 6 K km⁻¹ because condensation releases ≈2.5×10⁶ J kg⁻¹. The environmental lapse rate is measured surroundings; the parcel rate belongs to the moving parcel.
6 · Humidity ChainC4 notes
es=610.7 exp[(Lv/Rv)(1/273.15−1/T)]
RH=e/es
rv=0.622e/(P−e) ≈ 0.622e/P
- Convert vapour density to kg m⁻³.
- Find e, then es(T).
- RH<1 unsaturated; RH=1 saturated.
- Use total pressure in Pa for mixing ratio.
Say precisely: saturation vapour pressure rises steeply with temperature; warm air does not “hold” a fixed reservoir of water.
Quick numeric anchors
n=95000×5/(8.314×285)=200.5 mol
N=200.5×6.022×10²³=1.21×10²⁶ molecules
ρ=95000/(287×285)=1.16 kg m⁻³
At 500 hPa and 250 K, ρ=50000/(287×250)≈0.70 kg m⁻³: lower pressure dominates despite colder air. Cooling humid air without removing vapour leaves e tied to vapour density and temperature while es falls steeply; saturation occurs near the intersection.
7 · Köhler MethodC5
| Factor | Effect | Exam language |
|---|---|---|
| nw/(nw+ins) | <1 | solute lowers equilibrium RH |
| exp(curvature/r) | >1 | small drops need higher RH |
| larger wet r | curvature weakens | activation becomes easier |
- Dry mass: ms=(π/6)ρsDdry³.
- Moles salt ns=ms/Ms; water nw from stated ratio.
- Water volume nwMw/ρw; add dry volume.
- Wet r=[3V/(4π)]1/3; evaluate both factors.
Kelvin-only screen: ev/e₀=exp[2σw/(RvTρlr)]. At 290 K the exponent is 1.12×10⁻⁹/r (r in m). Observed ratios rarely exceed ≈1.02, so very small particles rarely activate.
8 · Visibility + PMC5
Vvis=K/σext
Cmass=Np(π/6)ρpD³
| Class | Count | Boundary |
|---|---|---|
| PM10 | sum mass of eligible populations | aerodynamic diameter ≤10 µm |
| PM2.5 | subset of PM10 | aerodynamic diameter ≤2.5 µm |
Compute each size population separately, convert kg m⁻³ to µg m⁻³ (×10⁹), then classify and sum. Do not infer number concentration from mass without diameter and density.
Classification drill
=2.51×10⁻⁶ kg m⁻³=2513 µg m⁻³
2 µm population=1508 µg m⁻³
PM10=4021; PM2.5=1508 µg m⁻³
9 · Parcel ModelC6
Frame/family: Lagrangian ODE model following one adiabatic parcel; supplied code integrates seven coupled quantities through the ascent. Assumes parcel properties match the atmosphere at the same height.
| Controlled change | Prediction + mechanism |
|---|---|
| w: 0.3→3.0 m s⁻¹ | cooling outruns condensation; higher S activates smaller particles |
| initial RH ↑ | saturation reached lower/earlier |
| sea salt added | large soluble particles activate readily |
| w: 0.3→0.03 | condensation removes S; fewer activate |
Kelvin filter: at 290 K, r=50 nm requires ≈2.3% supersaturation; r=0.5 µm requires ≈0.2%. At 1%, only the larger drop grows.
10 · StabilityC7
| Lifted parcel | Class | Plume |
|---|---|---|
| colder/denser; returns | stable | fanning; least mixing |
| same T | neutral | coning |
| warmer; rises | unstable | looping; fastest mixing |
Force + balance bridge
|ug|=|aPGF|/|f|; wind ∥ isobars
low: v²/R+fv−aPGF=0 → v<ug
high: v²/R−fv+aPGF=0 → near branch v>ug
Pressure gradient points high→low. Coriolis turns motion right in the Northern Hemisphere and left in the Southern, is zero at the equator and does no work. For a high, a negative discriminant means no anticyclonic solution.
11 · Gaussian PlumeC7
Empirical, Eulerian, analytical: evaluated on a 3-D fixed grid. Simplifications: no time dependence; transport only downwind; no along-wind mixing. Meandering/looping becomes a time average through wider Gaussian profiles.
half peak: |y|=σy√(2 ln2)=1.177σy
|y|=2σy → factor=e⁻²=0.135
| Plot change | Mechanism |
|---|---|
| wind direction changes | plume axis rotates |
| variable wind | time-mean footprint broadens |
| more sources | contributions superpose |
| less stable | σy, σz wider; more dispersion |
| RH ↑, PM mass ↑; shape same | hygroscopic D³ mass; widths unchanged |
12 · Column → DynamicsC8–C10
| Quantity/method | Rule |
|---|---|
| potential T | advected because conserved during pressure change; latent heating in process step |
| warm rain | collision/coalescence bridges condensation’s tens-of-days timescale to ≈20 min |
| more aerosol | same water / more drops → smaller drops → slower collection → rain suppressed |
| ice | rare ice-nucleating particles (e.g. desert dust) open cold-rain pathway |
| shallow wave | c=√(gh) |
| vorticity | ζ=∂v/∂x−∂u/∂y; absolute=ζ+f |
Transport anchor: with w=1.5 m s⁻¹, vapour/cloud rise at 1.5, snow falls at 1 but still rises at 0.5, and rain falls at 6 so descends at 4.5 m s⁻¹. Use the largest |net speed| when checking 𝒞.