How to Perform a Draft Survey: The Full Procedure

A draft survey does not measure cargo. It measures how much the ship weighs, twice, and attributes the difference to cargo. Everything that makes the procedure demanding follows from that one fact: every tonne of water, fuel and unrecorded weight that changes between the two surveys lands in the cargo figure unless somebody accounts for it.

This article sets out the procedure in the order it is actually carried out, with the reasoning behind each step and a worked example that reconciles.

What the survey establishes

Two displacements, measured at two points in time:

• Initial survey — before loading, or before discharge
• Final survey — after loading, or after discharge

Each displacement is reduced to a net displacement by deducting everything on board that is not cargo. The difference between the two net displacements is the cargo quantity.

The consequence is worth stating plainly, because it explains why surveyors care about things that look peripheral. A ballast tank that is 20 tonnes emptier than the sounding suggests does not produce a 20-tonne ballast error. It produces a 20-tonne cargo error, and cargo is what the invoice is based on.

Before reading anything

The measurements are only as good as the information they are entered against, so the preparation matters.

Obtain and check:

• The vessel’s hydrostatic particulars or deadweight scale, and confirm they are the approved set for the vessel as she is now, not a sister ship and not a superseded booklet.
• Tank calibration tables for every ballast, fresh water and fuel tank, with any trim and list correction tables.
• The light ship weight and the constant, and the basis on which the constant was established.
• The distances of the draft marks from the perpendiculars, forward, midship and aft. Marks are rarely exactly on the perpendiculars, and on some hull forms the difference is significant.
• LBP, and the position of the LCF, TPC and MCTC at the drafts expected.

Confirm before the first reading that the vessel is upright or that the list is measured, that no ballast or bunker transfer is in progress, that no water is being taken on, and that the crew understand that nothing is to move until the survey is complete. Agree this with the chief officer rather than assuming it.

Step 1 — Read the six drafts

Read forward, midship and aft, port and starboard: six readings.

Read from a launch or from the quay at as near to eye level with the waterline as can be managed. Reading down from the deck introduces a parallax error that is systematically in one direction, which is worse than a random error because it does not average out.

In any sea or swell, watch a full cycle and record the mean of the crest and trough, not a single glance. Take several observations and average them. The swell is the dominant source of error in exposed berths, and an honest survey report records the sea state that the readings were taken in.

Where a mark is obscured, damaged or repainted out of position, do not estimate it silently. Record what was done and note it in the report.

Step 2 — Correct the observed drafts to the perpendiculars

Draft marks are positioned where the shell plating allows, not where the naval architect would like them. The hydrostatic tables, however, are built on drafts at the forward and after perpendiculars. So each observed draft is corrected for the distance between the mark and its perpendicular, using the apparent trim.

For each station:

Correction = (apparent trim / LBP) × distance of mark from perpendicular

Apply the correction in the direction that the trim carries the hull at that point. The sign is where errors creep in, so the useful discipline is a sanity check rather than a memorised rule: after correction, a vessel trimmed by the stern must have a larger corrected after draft and a smaller corrected forward draft than she did before, and the corrected trim must be larger than the apparent trim when the marks lie inside the perpendiculars.

The corrected drafts are then averaged port and starboard to give a single forward, midship and after draft.

Step 3 — Calculate the quarter mean

A hull is not a rigid beam. Loaded, it sags; light or unevenly loaded, it hogs. The midship drafts are the only measurement of that deflection, and the quarter mean is how it is taken into account.

Mean of forward and aft  =  (F + A) / 2
Mean of means (MoM)      =  (F + A + 2M) / 4
Quarter mean (QM)        =  (MoM + M) / 2  =  (F + A + 6M) / 8

The quarter mean is the draft entered into the hydrostatic tables. Not the simple mean of forward and aft, and not the midship draft.

The weighting is not arbitrary: it approximates the actual distribution of the deflected waterplane far better than a simple mean, which is why it is the accepted figure in the trade. On a vessel with significant deflection, the difference between the simple mean and the quarter mean is easily worth several hundred tonnes on a Panamax.

If the midship drafts differ from the mean of forward and aft by an implausible amount, stop and re-read them before calculating anything. A misread midship draft is multiplied by six in the quarter mean, so it is the single most damaging reading to get wrong.

Step 4 — Take displacement from the hydrostatic tables

Enter the hydrostatic particulars at the quarter mean and read off displacement, TPC, MCTC and the position of the LCF, interpolating between tabulated drafts where necessary.

Read TPC, MCTC and LCF at the same draft as the displacement. Taking one of them from a different draft is a common and entirely avoidable error.

Step 5 — Apply the trim corrections

Two corrections, for two different effects. They are covered in detail in trim correction explained; in summary:

First trim correction allows for the fact that the LCF is not at amidships, so the draft at the LCF is not the quarter mean:

First trim correction (t) = (Trim × LCF distance from amidships × TPC × 100) / LBP

with trim and LCF distance in metres, TPC in tonnes per centimetre and LBP in metres.

Second trim correction allows for the change in the shape of the waterplane as the vessel trims, which the first correction does not capture:

Second trim correction (t) = (Trim² × ΔMCTC × 50) / LBP

where ΔMCTC is the difference in MCTC between 0.5 m above and 0.5 m below the quarter mean. It is small at modest trim and grows with the square of trim, so it becomes significant precisely when a vessel is trimmed hard by the stern — which is a normal loaded condition, not an unusual one.

Step 6 — Correct for water density

The hydrostatic tables are computed for a stated density, conventionally 1.025 t/m³ for salt water. Dock water is almost never exactly that.

Density corrected displacement = Displacement × (observed density / table density)

Take the density with a calibrated hydrometer, from water drawn at a depth representative of the hull, not skimmed off the surface. In a river berth or anywhere with fresh water inflow the water is often stratified, and a surface sample can be materially lighter than the water the ship is actually floating in. Take samples at more than one point along the length and, in a stratified berth, at more than one depth.

Record whether the hydrometer reads density or specific gravity, and whether in air or in vacuum. Mixing these up is a small error repeated on a large number, which is the definition of an expensive mistake.

Step 7 — Deduct everything that is not cargo

From the corrected displacement, deduct:

Item	How it is established
Ballast water	Sounded or gauged, corrected for trim and list, density confirmed
Fresh water	Sounded or gauged
Fuel oil, diesel, lubricating oil	Sounded, corrected for temperature
Sludge, bilge and slop	Sounded; often neglected, and should not be
Light ship	From the approved particulars
Constant	From the vessel’s own figure, verified where possible

Sound the ballast tanks yourself, or witness them being sounded. A tank reported as empty and a tank that is actually empty are different things, and the difference is frequently tens of tonnes in a wing tank with a slack bottom.

Confirm that no tank was pressed up or partially emptied between the initial and final surveys, and record the ballast condition in both surveys so that the comparison is visible in the report.

The constant

The constant is the difference between the vessel’s actual light displacement and the light ship figure in the particulars. It accumulates over the vessel’s life: mud in ballast tanks, unrecorded stores and spares, additional structure, accumulated sludge.

It is also the most frequently disputed item in a draft survey, for a simple reason. A constant supplied by the vessel without evidence, and used without challenge, transfers its entire error into the cargo figure. Where a constant is materially different from the figure obtained on a previous survey of the same vessel, ask for the basis of it, and record both the figure used and its source in the report.

Step 8 — Calculate the cargo

Net displacement = Corrected displacement − total deductibles

Cargo = Final net displacement − Initial net displacement

For a discharge survey the subtraction runs the other way. The figure is then compared with the shore figure or the bill of lading quantity.

Worked example

A Supramax loading a bulk cargo. Illustrative particulars, LBP 190.00 m, table density 1.025 t/m³.

Final survey, corrected drafts

Station	Corrected draft
Forward	11.820 m
Midship	11.960 m
Aft	12.640 m

F + A                    = 11.820 + 12.640  = 24.460
Mean of F and A          = 24.460 / 2       = 12.230 m
Mean of means            = (24.460 + 2 × 11.960) / 4 = 12.095 m
Quarter mean             = (12.095 + 11.960) / 2     = 12.0275 m
Trim                     = 12.640 − 11.820           = 0.820 m by the stern

The midship draft is 0.270 m less than the mean of forward and aft, so the hull is hogging. That is why the quarter mean (12.0275 m) sits well below the simple mean (12.230 m): a difference of 0.2025 m, which at 62 t/cm is worth about 1,255 tonnes of apparent displacement. This single step is the reason the midship drafts are read.

From the hydrostatic tables at 12.0275 m

Quantity	Value
Displacement	56,480 t
TPC	62.0 t/cm
LCF	3.10 m aft of amidships
MCTC at 12.5275 m	688 t·m/cm
MCTC at 11.5275 m	672 t·m/cm

First trim correction

= (0.820 × 3.10 × 62.0 × 100) / 190.00
= 15,760.4 / 190.00
= +82.9 t

Positive here: the vessel is trimmed by the stern and the LCF is aft of amidships, so the draft at the LCF is greater than the quarter mean and the tables under-read.

Second trim correction

ΔMCTC = 688 − 672 = 16 t·m/cm
= (0.820² × 16 × 50) / 190.00
= (0.6724 × 800) / 190.00
= +2.8 t

Small at 0.82 m of trim. At 3.00 m of trim, all else equal, the same expression gives roughly +37.9 t — the square term is what makes it matter on a heavily trimmed vessel.

Corrected for trim

56,480 + 82.9 + 2.8 = 56,565.7 t

Density correction, observed dock water density 1.0192 t/m³:

56,565.7 × (1.0192 / 1.025) = 56,245.7 t

Deductibles

Item	Tonnes
Ballast	480.0
Fresh water	165.0
Fuel oil and diesel	1,240.0
Lubricating oil and sludge	58.0
Light ship	9,860.0
Constant	210.0
Total	12,013.0

Final net displacement = 56,245.7 − 12,013.0 = 44,232.7 t

If the initial survey gave a net displacement of 1,684.5 t, then:

Cargo = 44,232.7 − 1,684.5 = 42,548.2 t

Note how the deductibles dominate the arithmetic. The trim corrections together moved the figure by 85.7 t; the density correction moved it by 320 t; the constant alone is 210 t, and it was supplied by the vessel. The precision of the draft readings is not the limiting factor in most surveys — the quality of the deductible information is.

Where surveys go wrong

In rough order of how much they typically cost:

The constant is accepted without question. An unverified constant is a direct, unbounded transfer of error into the cargo figure. It is also the easiest thing in the survey to challenge politely and document.

Ballast is taken from the vessel’s figures. A tank reported empty with a slack bottom, a tank where the sounding pipe does not reach the aftermost point, a trim correction not applied to the tank table. Sound them.

Density is taken at the surface. In a stratified berth the surface sample can be materially lighter than the water supporting the hull. The error is systematic and applies to the whole displacement.

Midship drafts are misread. Weighted six times in the quarter mean, so a 2 cm error becomes a 1.5 cm error in the quarter mean — which on a Panamax is worth well over a hundred tonnes.

Drafts are read from the deck in a swell. Parallax plus swell, both systematic. Read at the waterline, observe full cycles.

Something moves during the survey. A ballast pump run, a bunker transfer, fresh water made. Establish at the outset that nothing moves, and confirm it again before closing the survey.

TPC, MCTC and LCF are taken at different drafts. Read them all at the quarter mean.

The wrong hydrostatic booklet. A superseded set, or a sister ship’s. Check the vessel’s name and the approval on the front of the particulars.

Practical considerations

Do the calculation on board, and do it before leaving. A figure that will not reconcile is far easier to investigate while the vessel is still at the berth and the ballast can be re-sounded.

Compare your figure with the shore figure before signing anything. A difference of the order of half a per cent is commonly treated in the trade as within the expected band; a larger one should be investigated and, if it cannot be resolved, documented in a letter of protest. That threshold is commercial practice rather than any rule, and the applicable charter party or sale contract may set a different one.

Record the conditions. Sea state, list, whether the vessel was upright, where the density samples were drawn, how the ballast was sounded, whose constant was used and on what basis. A draft survey that is challenged six months later is defended on the quality of its record, not on the recollection of the surveyor.

Applicable standards and basis

There is no single international convention that prescribes draft survey procedure. That is a point worth being clear about, because it is often assumed otherwise.

• The method is established industry practice, set out in recognised technical texts and in the guidance published by P&I clubs and cargo superintendents’ organisations.
• The obligation to carry out a draft survey, and whose figure governs, is contractual — it comes from the charter party or the sale contract, not from regulation.
• The vessel’s hydrostatic particulars and stability information are approved by the flag administration or a recognised organisation on its behalf, and the survey depends on them being the current approved set.

Where a trade association’s rules apply to the cargo, they may impose requirements on measurement and documentation. Those are contractual obligations on the parties to that contract.