Dredge volume dispute between pre- and post-dredge surveys

Do the arithmetic first, because it decides the whole dispute.

Channel area = 120 x 800 = 96,000 square metres. A systematic vertical offset of 0.15 m over that area gives a volume error of 96,000 x 0.15 = 14,400 cubic metres. The apparent over-dredge is 61,400 - 48,000 = 13,400 cubic metres. So a single unmodelled systematic bias is the same size as the entire disputed quantity. That does not prove the contractor is right, but it means the volume cannot be relied on until the bias is dealt with.

Structure the computation:

1. Common vertical datum. Reduce both surveys to LAT using a consistent, verified method - a local gauge or tide corrector at the channel, not a gauge 2.1 km away with a known 0.15 m residual. Alternatively use GNSS-RTK tide. Apply the same method to both surveys.
2. Consistent acquisition. The same vessel and configuration for both surveys is strongly preferred, or at minimum the same sound velocity strategy (casts at the start, middle and end, or at a defined interval), the same draft and offset values, and check lines in both.
3. Difference surface, not two separate volumes. Compute post minus pre as a surface and derive volume from it. Systematic errors common to both surveys partly cancel; taking two independent volumes does not give you that benefit.
4. Uncertainty band. The plus or minus 0.10 m at 95 percent is a random uncertainty, so over 96,000 square metres it produces a volume band, not a bias, and the difference of two surveys increases the variance. Report the volume with that band rather than as a single number.
5. Filtering. Apply the same filter and the same gridding to both surfaces. A filter applied only to the post-dredge surface is itself a bias.

Noise and bias behave differently and the difference surface tells you which one you have.

• Noise is random and roughly symmetric. Over a large area it averages towards zero and should not produce a systematic volume. If the difference surface has a mean offset significantly away from zero, that is bias, not noise - datum, tide reduction, sound velocity, transducer draft, or vessel offset.
• Bias is uniform and directional. Plot the difference surface statistics: mean, standard deviation, and the distribution. Then look at the spatial pattern - a constant offset across the whole channel points to a vertical datum or draft error; a gradient across the channel points to a tide or sound velocity problem; a pattern near the edges points to a boundary or side-slope modelling problem.

Other tests:

1. Check line comparison - cross lines in both surveys, and compare the two surveys against each other on the check lines rather than the main lines.
2. Soundings density and line spacing - sparse or irregular coverage, especially near the channel edges, will corrupt a prism or grid volume regardless of vertical accuracy.
3. Volume method - check whether the volume was taken to the design surface or the pre-dredge surface, whether side slopes were included, and whether the boundary polygon is identical for both surfaces. Most volume disputes are boundary and side-slope errors, not sounding errors.
4. Independent evidence - the dredger's production records (hopper loads, dredging time, draghead logs) give a sanity check on the order of magnitude.

The conclusion should be stated with the uncertainty band: for example, 'the computed volume is X with a plausible range of plus or minus Y, and the pre-dredge tide reduction biases the result by up to Z'. That is a defensible position; a bare 61,400 is not.

Write the requirements around the error sources that actually caused this one.

• Vertical datum control: tide reduction using a gauge at or adjacent to the works, or GNSS-RTK tide, with a stated maximum residual and a verification against a benchmark. Specify the datum (LAT/chart datum) and the reduction method explicitly.
• Sound velocity: casts at defined intervals and whenever conditions change, with the cast locations and results recorded and submitted. Require the same strategy for pre- and post-dredge surveys.
• Equipment and configuration: same vessel class and transducer configuration for both surveys where practicable, with draft, vessel offsets and patch test results recorded and submitted.
• Coverage: minimum sounding density, maximum line spacing, and a specified number of cross check lines, with the requirement that coverage is achieved over the full dredge polygon including side slopes.
• Independent verification: a specified percentage of the pre-dredge survey independently verified by a second system or a second surveyor.
• Volume computation method: define the surface method, the boundary polygon, the treatment of side slopes, the filter and gridding parameters, and require both surfaces to be computed identically. Also define whether payment is on the design prism or the surveyed difference.
• Uncertainty: require the survey to state vertical uncertainty and to report the volume with an uncertainty band derived from it, so that a discrepancy within the band is a commercial conversation rather than an argument about whose survey is wrong.