How to Calculate Wood Movement: One Formula for Tabletops, Breadboards, and Panel Grooves
How to Calculate Wood Movement: One Formula for Tabletops, Breadboards, and Panel Grooves - A step into the workshop.
Stop Wasting Wood
Generate clear cut diagrams instantly with our new Cutlist Optimizer. Plan your cuts to make your next build easier.
Picture the glue-up. Thirty inches of black walnut, four boards jointed straight, clamps open and waiting. Before the glue comes out, I want a number from you. Not "wood moves, leave some room." An actual number: how much wider is this top in a Minneapolis January than in a Phoenix July?
About ten sixteenths. Five-eighths of an inch. And you can get there with one multiplication on the back of the lumber receipt. If that surprises you, it should change how you cut your next breadboard mortise. And if you think the kiln-dried stamp or three coats of finish excuses you from the arithmetic, I've got news from every swampy August I've sweated through in Austin: the wood doesn't read labels.
One Multiplication, Three Numbers
The USDA Forest Products Laboratory's Wood Handbook, the closest thing this trade has to a rulebook, gives the estimating equation in Chapter 13. Dimension change equals starting dimension, times a dimensional change coefficient for the species, times the moisture content change in percentage points. That's the whole formula. The Handbook says it estimates satisfactorily between 6% and 14% moisture content, which covers most furniture living out its life indoors.
The coefficient comes from the Handbook's Table 13-5, in two flavors. Flatsawn width moves tangentially, along the growth rings; quartersawn width moves radially, across them; and tangential movement runs about twice radial. For black walnut, that's 0.00274 flatsawn and 0.00190 quartersawn. Length you ignore entirely. Longitudinal shrinkage is 0.1% to 0.2% from green to ovendry for most species, which is nothing.
So: width is a tape measure, the coefficient is a table lookup, and the only number that takes real thought is the moisture swing.
Phoenix Dry, Minneapolis Damp
Wood settles toward an equilibrium moisture content set by the air around it, and the climate work is already done. William Simpson at the Forest Products Laboratory computed average monthly EMC for 262 U.S. locations from roughly thirty years of NOAA data back in 1998. Phil Mitchell at NC State reran the numbers in 2016 with newer 1981-2010 climate normals. They barely disagree. Phoenix in July sits at 6.2% in both. Minneapolis in January reads 13.7% in the old table, 14.0% in the new one.
Two details worth keeping. First, June, not July, is Phoenix's basement. Mitchell has June at 4.4%, because the July monsoon pushes humidity, and EMC with it, up almost two points. Second, the national spread is wilder than people think: Mitchell's update has Las Vegas holding the lowest annual average in the country at 5.9%, Spokane swinging a brutal 10.3 points from August to December, and Houston barely moving at all, 0.8 points all year. A Houston shop doesn't have seasons. It has soup.
Run It Hot
Take Mitchell's numbers. Phoenix July at 6.2%, Minneapolis January at 14.0%. That's a 7.8-point swing.
30 in. x 0.00274 x 7.8 = 0.64 in.
Ten sixteenths, give or take. Quartersawn walnut runs 30 x 0.00190 x 7.8 = 0.44 in., about 7/16". Quartersawn buys back roughly 3/16" on this one top, which is the right-tool-for-the-job argument in a single line. Use Simpson's 1998 figures instead and the answers shift by less than 1/32". Pick either dataset; the wood can't tell.
Same 30" flatsawn top, same 7.8-point swing, other species, coefficients straight from Table 13-5:
| Species | Coefficient | Movement |
|---|---|---|
| Red oak | 0.00369 | 0.86" |
| White oak | 0.00365 | 0.85" |
| Sugar maple | 0.00353 | 0.83" |
| Black walnut | 0.00274 | 0.64" |
| Black cherry | 0.00248 | 0.58" |
| Mahogany | 0.00238 | 0.56" |
| Teak | 0.00186 | 0.44" |
Flatsawn teak moves about half as much as red oak, which puts it dead even with quartersawn walnut. And walnut is the calm one among the domestics for a reason: the Handbook's shrinkage table gives it a tangential-to-radial ratio around 1.4, where sugar maple's is 2.1, so walnut cups and distorts less as it cycles.
The true worst case, Phoenix June at 4.4% to Minneapolis December at 14.9%, pencils out to 0.86". But both ends of that trip fall outside the Handbook's 6-14% window, so treat it as a rough outer bound, not gospel.
Your Furnace Flips the Script
Now the trap in the premise. Those EMC tables describe wood in outdoor air: sheltered from rain and sun, but breathing the weather. That's your garage shop, a shipping container, a three-season porch, the box truck the table rides north in. A heated house is a different planet. The Wood Handbook's Table 13-2 says interior wood in most of the U.S. should run about 8% average, individual pieces between 6% and 10%, with 6% in the dry Southwest and 11% in damp coastal country. And a heated Minneapolis house in January is desert-dry inside; the Handbook notes framing lumber in heated houses in cold climates can drop to 6% or 7%.
Follow that through. A top glued up in a Phoenix garage in July at around 6.2% and delivered into a heated Minneapolis living room in January may barely move at first. It's already close. The hit lands the following summer, when the furnace shuts off, indoor EMC climbs, and the top swells. The dramatic 5/8" belongs to unconditioned space and transit. A realistic in-home swing across Table 13-2's 6-10% band is 4 points: 30 x 0.00274 x 4 = 0.33". Call it 5/16". Smaller, sure. Still plenty to wreck a top that looked fine at delivery and lets go by the time football season starts. And kiln-dried is no pardon. Furniture hardwood leaves the kiln at 6-8%, per the Handbook, which just means it starts near indoor service. It chases the air forever after.
What the Gap Buys at the Joint
Breadboard ends. Marc Spagnuolo at The Wood Whisperer names the failure plainly: glue and screw a breadboard on rigid, and as the top cycles, "tension builds up and eventually the top will crack." Paul Anthony, writing in Woodcraft Magazine, puts numbers on the fix: glue only the center 6" or so of the joint, elongate the outermost pin holes, keep individual tenons to about 4" wide, cut mortises 1/16" deeper than the tenons, and leave slack at the ends of the outer mortises. Our number sizes that slack. With the center pinned, the 30" top's 0.64" full swing splits into about 5/16" per side, so each outer slot needs at least that much travel. The breadboard will also sit proud or shy of the top's edge by that much depending on season. That's not sloppy work. That's the joint doing its job.
Frame-and-panel. Run a 12" flatsawn walnut panel through the same math: 12 x 0.00274 x 7.8 = 0.26", which lands right at the top of Anthony's rule of thumb that seasoned plainsawn domestics move 1/8" to 1/4" per foot of width. His seasonal rule is the part worth taping to the wall: build in winter, when the wood is at its driest, and leave at least 1/8" per side at the groove bottoms; build in summer and seat the panel nearly home with 1/16" of insurance. Pin the panel centered so it moves evenly both ways, and never glue a solid panel in its grooves. Plywood is the one exception you're allowed to glue.
Cutting boards. Spagnuolo's autopsy of a split board is the same physics at kitchen scale. A long-grain decorative strip glued into an end-grain field is a breadboard end nobody meant to build: the end-grain squares push across both plan dimensions, the strip holds still along its length, and the board splits. Keep the grain running one direction, and make the decorative strips from end grain too. The magnitude is real. An 18" edge-grain walnut board over a kitchen's 4-point swing moves 18 x 0.00274 x 4, about 3/16", and some part of the build has to absorb it.
Error Bars and Eleven-Dollar Insurance
The estimate has honest limits, and the Handbook lists them itself. Commercial boards are rarely dead flatsawn or dead quartersawn, so the method tends to overestimate width movement and underestimate thickness movement. Board-to-board shrinkage varies with a coefficient of variation around 15%; the Handbook says predicting an individual board accurately is impossible, though the average of a stack predicts well. And on finishes it is blunt: "Protective coatings can retard dimensional changes in wood but do not prevent them." Film finish buys time. It does not buy immunity.
If you'd rather not push the numbers yourself, the tools are cheap. Lee Valley sells a dial-style Wood Movement Reference Guide covering 73 species for $11.50. WoodBin's Shrinkulator is a free online calculator built on Wood Handbook data; note it defaults to a 28% fiber saturation point where the FPL tables assume 30%. Woodworkers Source, the Phoenix hardwood outfit that's been at this since 1978, publishes a free EMC calculator and repeats the rule that matters most: match your lumber's moisture content to its service conditions as closely as practical. And if you want to measure instead of assume, Wagner's Orion 910 pinless meter runs $430 with a calibrator and a 7-year warranty. Real money, but it ends arguments.
Wood movement isn't a mystery and it isn't bad luck. It's three numbers and a multiplication that takes less time than opening the glue bottle. Run it, size your slots and grooves to the answer, and the seasons can do whatever they want to your table. Skip it, and a Minneapolis January will run the numbers for you, straight through the middle of your top.