Round It Up and Sign

Is it real? Positional uncertainty at 1:12,000 over a 450-ft perimeter is about 450/12,000 = 0.0375 ft. Area uncertainty ≈ (P/2) x δ = 225 x 0.0375 ≈ 8.4 sf. The shortfall is 154 sf - roughly 18 times the uncertainty. It is real. Rounding it up would be fabricating 154 sf of land.

Compliant fix - move the common line into Lot 8, which has surplus.

• Lot 7 is 100.0 ft wide; to gain 154 sf with a line roughly 198 ft deep, shift the common side line about 1.0 ft toward Lot 8: Lot 7 gains ~198 sf to about 20,044 sf; Lot 8 goes from 110.0 to 109.0 ft wide and from 21,500 to about 21,309 sf.
• Check everything the shift touches: Lot 8 still exceeds 20,000 sf and 100.0 ft width; frontages unchanged or still over 60 ft; no new encroachment into the wetland buffer; the relocated line does not clip either building envelope at required side yards.
• Do not design to 20,001 sf. A 44-sf margin is inside a re-survey's variation; leave visible margin so the plat survives as-built review.

If no shift works, the answer is one fewer lot, not a variance the director has already refused.

Tolerance argument fails. Width is the difference between two side lines, so uncertainty is about 0.06 ft at 1:12,000; the shortfall is 0.6 ft, roughly ten times that. It is a real 0.6 ft, not instrument noise.

Geometric fix: on the bulb the side lines are radial, so the width at the setback line is controlled by the included angle between those radials, not by a simple offset.

• Rotate the common radial line with Lot 17 so the arc between radials grows enough to put 100.0 ft on the chord at the 30-ft setback line - a small rotation, since 0.6 ft at ~40 ft from the center is only a couple of degrees.
• Verify the donor side: Lot 17 must still hold ≥100.0 ft width at its setback line and ≥60 ft of arc frontage.
• Verify Lot 18 still fits the required house footprint plus side yards inside the buildable envelope - a lot can pass on width and still fail because the house will not fit.
• Check that the rotated line does not make the Lot 18 driveway exceed the maximum curb-cut width on the bulb.

If the rotation cannot be made without breaking Lot 17, the bulb geometry is wrong and the layout changes.

Decline, in writing, and say why.

• Your seal certifies the geometry shown. A plat that prints 20,000 sf where you measured 19,846 sf, or 100.0 ft where you measured 99.4 ft, is a false statement made by you, not by the developer. 'We will fix it later' means the sealed document is wrong now, and the plat is what the lots are sold from.
• The engineer's tolerance argument is a category error: measurement uncertainty describes how well you know a dimension, not what value you may write down. If your uncertainty is ±0.04 ft and the dimension is 99.4 ft, the lot is short.

What you do instead:

• Show measured values at the precision your survey supports (0.01 ft, 0.01 ac) and note the minimum-lot-area and minimum-width standards on the face of the plat.
• Revise the layout - the Lot 8 line shift and the Lot 18 radial rotation - and resubmit. That is a one-day fix and it makes the plat approvable without a variance.
• Give the developer and the reviewing planner your computations and the correspondence in writing, so the file shows you identified and reported both deficiencies rather than being asked to hide them.

Certifying noncompliance as compliance is not a business decision you get to make for the client.