Read the floor before you repair it.

Part 1 explained the summer problem: a cool room above a humid crawlspace can leave the underside of a raised wood floor wetter than the top.⁶ This part asks the next question: once the floor has changed shape, what should a homeowner do before someone offers to sand it flat?

Cupping: edges high, center low

The terms are not interchangeable: cupping is concave (trough-like); crowning is convex (center high).¹¹ Here we explain solid-wood boards, including oak and pine; engineered products can behave differently.

What actually bends the board

In normal service, wood swells as its cell walls take up water and shrinks as they dry, mostly across the grain.¹¹ Unequal swelling bends it; vapor does not physically push up the edges.

Picture upper and lower wood as bonded strips. If the bottom gets wetter, it wants to grow wider, but cannot slide freely past the top.¹¹ It follows the longer outside of the bend; the top follows the shorter inside. That makes a trough, with edges higher than the center.

Which way is the vapor moving?

Vapor diffusion goes from higher to lower water-vapor pressure, not automatically upward or downward.¹⁰ In the Louisiana twelve-house study, crawlspace vapor pressure exceeded indoor pressure from May through October, driving moisture inward and upward through the raised floors.⁹

Cooling raises local relative humidity at a given vapor pressure, tending to raise wood moisture; less-permeable coverings can slow drying toward the room.⁹ Thus the underside can become wetter without a puddle.

Before you sand a cupped floor

A sander removes the high places. On a cupped board, those high places are the edges; if the board later dries and relaxes, the missing wood cannot return, so the center is left higher than the edges.⁸

This is a machined shape left by past distortion, not necessarily a board currently bending because its top is wetter. A crowned floor can therefore have normal moisture readings today and still bear the consequences of sanding too soon.⁸

The warning comes before the sander

A floor that is still cupped, still changing moisture content, or still has a moisture difference through its thickness is a warning against immediate leveling.⁸ Crowning that appears after refinishing and subsequent drying is a retrospective clue to that sequence, not proof of a new leak above the floor.

Engineered floors need a separate diagnosis

Some engineered floors dry-cup when the face layer loses moisture and shrinks across its width more than the supporting layers allow; the stress pulls the edges upward.^8,11 Identify the product before treating every cup as a crawlspace-wetting problem.

Crowning: center high, edges low

A solid board can crown from an active moisture imbalance: the upper portion is wetter than the underside.^8,11 Moisture gained near the top, drying near the bottom, or both can produce that difference.

Why the curve reverses

The wetter upper wood wants to expand farther across the grain than the drier lower wood.¹¹ Because the portions remain connected, the upper portion follows the longer outside of an upward arch. The surface becomes convex: center high, edges relatively low. It does not require something underneath to push the center up or a force above to press the edges down.

Does that mean vapor is moving downward?

Only if vapor pressure is higher above than below. Under that boundary condition, diffusion is downward.¹⁰ High indoor humidity combined with a dry substrate can create a top-wetter imbalance; NWFA also identifies heat from below as a contributor.⁸ Compare temperature and humidity together, not RH percentages alone.¹¹

But water left on the surface, including wet-mopping water, can wet the top directly.¹² That is liquid wetting, not evidence that the whole assembly’s vapor drive has reversed. Conversely, water entering from above can pass through joints and wet the underside, causing cupping instead.¹² The entry point and the eventual moisture distribution are different questions.

The warning can come before the curve

A leading indicator is an early change that gives you a reason to investigate before obvious damage. Here, it means a warning condition or developing moisture imbalance, not a reliable prediction that a particular floor will fail.

Before moisture-driven cupping

Notice the exposure: new water beneath the house, a plumbing leak, or persistently wetter crawlspace conditions can increase moisture available to the floor.⁸ A colder floor above that humid space can also raise local relative humidity at the wood.⁹ Neither one humid afternoon nor a low thermostat setting alone establishes the cause.

Look for a trend in the wood: rising moisture readings in any separate subfloor are a warning about the assembly. A developing wetter-bottom/drier-top difference within the floorboard more directly supports the bending mechanism.^8,11 These are different measurements; one does not prove the other.

Watch the exception: if edges rise while gaps open, rapid drying of the upper surface may be contributing.⁸ That combination is a reason to investigate drying, not automatically conclude that more water is entering from below.

Before moisture-driven crowning

Notice the exposure: repeated wet cleaning, a spill left on the surface, or a sustained rise in indoor humidity over a relatively dry substrate can favor a wetter upper portion.^8,12 Heat or increased drying below can also contribute to the imbalance.⁸ Record changes in cleaning, leaks, and HVAC operation.

Check which portion is changing: depth-specific measurements can show whether the top is gaining moisture, the bottom is drying, or both.¹¹ An already-raised center is a symptom, not an advance warning. If readings no longer show an imbalance, earlier wetting or premature sanding may explain the remaining shape.⁸

Buckling: the floor lifts away

The illustration assumes two wood layers: finish boards over a separate supporting deck, called the subfloor. In that assembly, buckling means the finish flooring has lifted away from the subfloor, often with swelling and cupping nearby.¹

Some historic floors instead have boards nailed directly to joists or sleepers, without that separate deck.⁷ PHP field judgment: inspect the actual board-and-support arrangement; do not assume every raised house has a hidden subfloor.

Moisture-driven expansion can create the pressure, while insufficient expansion space, poor fastening, adhesive problems, or inadequate subfloor construction can aggravate the failure.¹ A ridge alone cannot tell you which combination occurred. A floating floor that tents also needs its own product-specific investigation rather than a nail-down repair prescription.

Gapping: spaces between boards

Wood usually shrinks across its width as it loses moisture, so spaces can open between neighboring boards.^1,5 Some gaps open during drier conditions and narrow again when conditions become more humid; this seasonal movement is expected in solid wood, although its size varies.¹

A permanent gap needs a different explanation. Flooring installed too wet can shrink after installation; wood severely compressed during a wet period can remain narrower after drying, leaving gaps even when the original moisture conditions return.^1,2

There is no universal rule that gaps under 1/16 inch are normal or gaps over 1/8 inch prove failure. Board width, species, grain, moisture change, installation history, and seasonal closure all matter.^1,5

Do not fill first and investigate later

NWFA calls for eliminating the cause and stabilizing the floor before restoring abnormal gaps.¹ PHP field judgment: do not respond to gaps by automatically adding a humidifier to a Gulf Coast house. Measure first, then decide whether humidity adjustment is warranted.

Build a moisture history

Moisture content (MC) is the amount of water in wood relative to the wood’s oven-dry weight.⁵ Relative humidity (RH) describes the air, and it helps explain where wood moisture is headed; it is not a substitute for measuring the wood itself.^4,5

PHP recommends a dated record, not a single meter number. Ask the investigator to document:

• The floor: solid or engineered, species if known, width, thickness, fastening method, prior sanding, and whether there is a separate subfloor. These affect both moisture interpretation and repair options.^1,2,4
• The history: onset, seasonality, and preceding changes such as leaks, A/C settings, vacancy, refinishing, or underfloor work.^1,2
• The environment: room temperature and RH, thermostat setting, and safely obtained underfloor observations of water entry and moisture sources.^1,3
• The wood: readings in affected and unaffected flooring and any separate subfloor, labeled by location and material. Document unusual readings rather than averaging them away.^1,4
• The method and trend: meter model, settings, depth, and repeat readings at mapped locations, using the manufacturer’s calibration checks and corrections.^4,5

A reading from above is not necessarily a “top” reading

Pinless meters scan a volume below their sensor. Depending on the meter and flooring thickness, that volume can include the flooring, adhesive, and substrate; different materials within the sensing depth can affect the result.⁴ Pressing the meter on the finished surface does not isolate the top fibers of a board.

Where a through-board moisture difference matters, a trained investigator can use appropriate insulated-pin equipment to compare depths because only the exposed pin tips make the measurement.^4,5 Species, wood temperature, instrument settings, and the particular assembly must be accounted for as the meter manufacturer directs.⁴

When is it safe to sand?

Not “after two weeks,” not “when it feels dry,” and not automatically when one reading falls below 12%. NWFA describes drying and flattening over weeks, months, or even a heating season, and LSU advises waiting for stabilization before refinishing.^1,3 The condition of this floor must decide the schedule.

Use the following as PHP’s discussion checklist with the flooring professional, not as a repair specification:

• The moisture source is addressed: the leak, drainage defect, or ongoing moisture condition has been identified and corrected, not merely hidden by a fresh finish.^1,3
• The house is operating normally: establish the temperature and humidity conditions the home will actually maintain, within the flooring manufacturer’s requirements. Expected in-use conditions matter more than a universal MC target.⁴
• Measurements support stability: record the flooring and any separate subfloor over time, and investigate any remaining moisture gradient within the boards. Do not treat one surface scan as proof that drying is complete.^1,4,5
• The shape has stopped changing: allow the boards to flatten as much as they will. Some deformation can be permanent, so drying is not a promise that every floor will become flat.^1,2,3
• The proposed repair fits the floor: ask the flooring professional to verify attachment, the supporting assembly, remaining sandable thickness, and product limits before choosing sanding, localized replacement, or a larger repair.

Read the gap pattern before filling

A few larger gaps between groups of boards can warrant investigation of boards sticking together, including finish-related sidebonding; that is different from every board moving independently.¹ We will examine finish layers in Part 3.

Why wider boards show more movement

This example estimates width change in an unrestrained solid board. It does not calculate cup height, buckling force, or the gap you will necessarily find in an installed floor. The coefficient comes from the USDA Wood Handbook’s approximate method for moisture changes within 6% to 14% MC.⁵

Width reduction ≈ starting width × 0.00369 × 4.

2.25 inches × 0.00369 × 4 = 0.0332 inch, about 1/32 inch.
5 inches × 0.00369 × 4 = 0.0738 inch, a little over 1/16 inch.

The wider board’s estimated movement is about 2.2 times the narrower board’s because its starting width is about 2.2 times greater. That does not make the wider board defective; it means width belongs in the diagnosis.

Different species and cuts have different coefficients, and commercial boards are not perfectly flatsawn or quartersawn.⁵ Restraint from neighboring boards and attachment changes how movement appears in an installed floor.² Do not multiply these examples across a room and call the result a predicted buckle.

The assumed four-point moisture change is a change over time, not a difference between the top and bottom faces. This arithmetic explains potential width movement, not a floor’s three-dimensional shape.

A flat floor is not the whole repair.

Cupping, crowning, buckling, and gapping are reasons to investigate, not interchangeable instructions to sand, fill, or replace. The same visible condition can have more than one cause, and an old moisture problem can leave a shape that outlasts the water.^1,2

For the homeowner, the practical question is not just “Can you make it look flat?” It is “What changed, what evidence explains it, and what must be corrected before we spend money on the surface?”

PHP’s recommendation: before approving refinishing, ask for a short written explanation of the cause, the measurements supporting it, and the conditions the floor must reach before work begins. A repair proposal should explain both how the surface will be restored and why the same problem should not return.

The next question is about the layers

Part 1, Cold house, wet floor, explains how summer humidity and a cold A/C setting can place a raised floor between two different climates.⁶ Part 2 adds the diagnostic sequence: read the shape, then verify what the wood and its supporting assembly are doing.

Part 3 will look at what lies above and below the boards: urethane finishes, other coatings, underlayments, and vapor-control layers. The question will not be whether a floor has a finish or a barrier. It will be how that particular combination affects wetting and drying.

Before you pay to make the floor look right, understand what it needs to stay right. That is the handoff from reading the damage to understanding the assembly.

Sources & footnotes