Wood density and specific gravity: definitions, ranges and why they matter

Why this matters in the sawmill

Density is the cheapest single number you can attach to a piece of wood, and it quietly explains a lot of daily sawmill life: how heavy a truckload really is, how much a board will shrink in the kiln, how stiff and strong it is likely to be, how hard it is on saw teeth, and how much energy the offcuts deliver as fuel. Yet the word "density" is used loosely. One person means the weight of a freshly sawn board, another the figure in a species datasheet, a third the value used in a design standard. These can differ by a factor of two without anyone being wrong, because they refer to different moisture states.

This article sets out the definitions, shows how to convert between them with a calculator, gives typical values for common European and North American species with the source of each figure, explains where variation comes from, and ends with what a mill can do with the number.

Definitions: density, specific gravity and the moisture problem

Density is mass divided by volume, usually in kg/m³ (1 g/cm³ equals 1000 kg/m³). Specific gravity is density divided by the density of water, so it is a pure number: 0.50 means half the density of water on the stated basis.

Wood is hygroscopic, so both its mass and its volume change with moisture content (MC). A density or specific gravity is meaningless unless the moisture state is given. The USDA Forest Products Laboratory Wood Handbook handles this with several bases. Specific gravity always uses the oven-dry mass; what changes is the volume:

• Oven-dry specific gravity (G0): oven-dry mass over oven-dry volume. Some North American design specifications use this basis.
• Basic specific gravity (Gb): oven-dry mass over green volume. Green volume, at or above fibre saturation, is the largest and most repeatable volume, so Gb is the steadiest reference and is common in forestry and fibre work.
• Specific gravity at 12 % MC (G12): oven-dry mass over volume at 12 % MC. Many strength tables use it because 12 % is a typical air-dry state.

Density proper uses the actual mass at the stated MC. MC itself is defined on the oven-dry basis: MC (%) = (wet mass − oven-dry mass) / oven-dry mass × 100, so it can exceed 100 %.

The fibre saturation point is the MC below which cell walls start losing bound water and the wood begins to shrink. The Handbook says it averages about 30 % with several points of scatter between species and pieces, and uses 30 % as a working approximation.

How to calculate and convert

All of this follows from the definitions. Use 1000 kg/m³ for water.

Density at a given MC. If G12 is oven-dry mass over the 12 % volume, the density with water included is ρ12 = G12 × 1000 × 1.12. The Handbook's worked example is white ash with G12 of about 0.605, giving 678 kg/m³; the arithmetic (605 × 1.12 = 677.6) confirms it.

Green density from basic specific gravity. A green piece at moisture content MC has a mass of Gb × 1000 × (1 + MC/100) per cubic metre of green volume. A spruce with Gb 0.40 and sapwood at 120 % MC is therefore about 880 kg/m³. This is my own arithmetic from the definitions, and it is the number that matters for truck loading and debarker feed.

From Gb to G12. Below fibre saturation, shrinkage is roughly proportional to the moisture lost, so volumetric shrinkage at MC x is about Sx = S0 × (1 − x/30), where S0 is total green-to-oven-dry shrinkage. Then Gx = Gb / (1 − Sx/100). For white ash the Handbook lists Gb 0.55 and S0 13.3 %. At 12 % MC, Sx is about 8.0 %, so G12 is about 0.55 / 0.920 = 0.598, close to the 0.60 in the Handbook's clear-wood table. Treat results as good to a few hundredths.

Maximum moisture content. The Handbook takes the specific gravity of the cell wall as 1.54 (I could not read its equation 4-3 in the copy I opened). My own calculation of the most water a piece can hold, (1/Gb − 1/cell-wall SG) × 100, gives about 267 % at Gb 0.30 and 44 % at Gb 0.90 with a wall SG of 1.5, matching the figures the Handbook quotes, while 1.54 would give about 268 % and 46 %; the small difference is unresolved. With 1.5 the result at Gb 0.40 is about 183 % (my arithmetic, not a published figure).

Sanity checks. For the same wood, Gb < G12 < G0 because volume is largest when green. G12 and ρ12 differ by the factor (1 + MC/100), so never compare them directly. If someone quotes "500 kg/m³ for spruce" without an MC, ask which one.

Typical values for European and North American species

Published figures differ in basis, origin and sample size, so read these as indicative ranges and not as specifications. In the Handbook's clear-wood table the specific gravity is based on oven-dry weight and the volume at green or at 12 % MC, which is the basis used in the next table.

North American species. The Wood Handbook's clear-wood table gives G12. I converted to density at 12 % MC (multiply by 1.12, rounded):

The Handbook also states that oven-dry density of most commercial woods falls between about 320 and 720 kg/m³, with balsa near 160 and some imported woods above 1040 kg/m³.

European species, measured at 12 % MC (Czech Republic). Studies from Mendel University in Brno, published in the Journal of Forest Science, report directly measured densities (for the conifers the abstract states the 12 % basis for the near-pith values; we assume it applies to the outer values too):

In a Polish study of 27 butt-end Scots pine boards, one per tree, from one 120-year-old stand, apparent density at 12 % MC ranged from 469 to 686 kg/m³ (Mirski et al. 2021): a spread of roughly 45 % between lightest and heaviest board in one stand (our calculation).

Two cautions. The oak comparison is confounded, because each species was sampled in the vegetation zone where it usually grows, so species and site cannot be separated. And a conifer species mean blends light juvenile wood with denser outer wood in proportions that depend on log diameter and growth history.

Variation within a tree and between trees

Inside the annual ring. In 85 Norway spruce from central and south-eastern Finland (Jyske, Mäkinen and Saranpää 2008), 49–80 % of total density variation lay within individual rings, between earlywood and latewood. Differences between rings accounted for 11–27 %, and differences between heights in the stem for only 3–6 %. Density is as much a texture property as a bulk one.

Pith to bark. In conifers density generally rises outward. The Finnish study relates the outward rise to increasing latewood density and latewood percentage (a relation, not a proven cause), with earlywood density rising only slightly. The Brno figures show the same direction, and pine had the largest gap (161 kg/m³). Beech in the Brno study showed a mild opposite trend, rising slightly toward the stem centre. Oak depends on species: English oak was more uniform, sessile oak had a marked gradient, and ring width explained only 12–13 % of density variance.

Along the stem. In the Finnish spruce, at equal cambial age density rose with height, but for rings formed in the same calendar year it fell with height. A single disc at breast height is therefore only a rough proxy for the whole tree.

Between trees. The Handbook gives a coefficient of variation of about 10 % for oven-dry density within a species and about 15 % for transverse and volumetric shrinkage. With a normal distribution and a 10 % coefficient of variation, roughly one piece in three lies more than 10 % from the mean (my statistical illustration, not a Handbook statement). The sources I used do not support a simple rule such as "fast-grown always means light"; test your own material.

Compression wood, tension wood and knots also change local density, so a whole-board measurement includes defects. That suits a structural question and not a clear-wood one.

What density says about strength, hardness, shrinkage and fuel

Strength and stiffness. Across species, higher density usually means higher strength, but the link to stiffness looks weaker (my reading of the three-species comparison below, North American clear wood). In the Handbook's 12 % table, Sitka spruce (G12 0.40) has a modulus of rupture of about 70 MPa and a modulus of elasticity of about 10.8 GPa; Coast Douglas-fir (0.48) about 85 MPa and 13.4 GPa; white oak (0.68) about 105 MPa and 12.3 GPa. From spruce to oak, density rises 70 % and bending strength about 50 %, but stiffness only about 14 %, and Douglas-fir is stiffer than oak at lower density. Within one species the scatter is larger still: the Polish pine boards had a density spread of 469–686 kg/m³ but a stiffness spread of 7.25–16.15 GPa in static bending (EN 408), more than a factor of two. That paper deals with knot distribution; I add grain angle and juvenile wood as other likely influences, which is general knowledge not tested in the sources. This is why machine grading combines density with other measurements.

Hardness. Resistance to indentation is generally expected to follow density (general knowledge; I did not verify this in a cited page). The Handbook's side hardness test embeds an 11.28 mm ball to half its diameter, with load applied perpendicular to the grain.

Shrinkage. The Handbook states that greater shrinkage is associated with greater density. Green-to-oven-dry shrinkage (radial / tangential / volumetric, %): Atlantic white cedar 2.9 / 5.4 / 8.8; ponderosa pine 3.9 / 6.2 / 9.7; white ash 4.9 / 7.8 / 13.3; white oak 5.6 / 10.5 / 16.3; mockernut hickory 7.7 / 11.0 / 17.8. Tangential shrinkage is roughly 1.4 to 1.9 times radial in these five species (my division), while longitudinal is typically only 0.1–0.2 %. All values are North American species. (The source tables do not state a common density basis, so the pairing with density is indicative.)

Fuel value. Per kilogram, dry wood is almost uniform. NIBIO (Norwegian Institute of Bioeconomy Research) gives a lower calorific value of 19.1 MJ/kg for softwood and 18.9 MJ/kg for hardwood, stated on dry and ash-free material and for wood without bark, leaves or needles; the page gives no moisture figures or energy per volume. My inference is that differences in energy per cubic metre of solid dry wood come almost entirely from density, and that water in the wood reduces the useful heat per tonne.

Sawing and drying. From general experience, not from a figure in these sources: light species need less feed power and wear tools less, dense hardwoods dry more slowly and are more prone to checking, and light woods can hold very large amounts of free water (267 % MC at Gb 0.30). Schedules still have to be developed by species and thickness.

Why a sawmill should measure it

• Buying by weight, selling by volume. The conversion depends on density and MC; a systematic 5 % supplier difference shifts revenue.
• Sorting. Light and heavy fractions suit different products.
• Kiln loading. Water to remove per cubic metre of green wood is Gb × 1000 × MC/100 kg.
• Grading. Density is one input of machine strength grading.
• Raw material tracking. Trends by supplier, stand or diameter reveal juvenile-wood share.

A simple routine: cut a clear sample, weigh it, measure its dimensions (or find its volume by water displacement), dry it in a ventilated oven until its mass stops changing, and weigh again. That gives MC, density and, with the green volume, Gb. Always record the MC at which dimensions were measured.

Practical rules of thumb for sawmills

• Attach an MC or a stated basis to every density number; otherwise it is unusable.
• Density at 12 % MC (kg/m³) is about 1120 × G12; green density is 1000 × Gb × (1 + MC/100).
• Expect at least ±10 % within a species and more within a stand; sample several dozen pieces per batch.
• Sample the same position on every log and note whether the pith is included.
• Use density to rank strength and hardness roughly across species, stiffness only within a species.

What is uncertain or varies

• Bases are not always stated. Secondary sources reproduce Handbook tables but blur the basis. I used only the clear-wood table that states the 12 % basis.
• Cell-wall density. The Handbook states a cell-wall specific gravity of 1.54; the 1.5 that reproduces its maximum-MC values is my back-calculation, so a small discrepancy remains.
• Species means hide distributions. Sampling methods changed over decades, and European studies cover single regions.
• Conversions are approximations. Linear shrinkage and a 30 % fibre saturation point are conventions; the Handbook says the fibre saturation point averages about 30 % and varies by several percentage points between species and pieces.
• Relations to strength, drying speed and tool wear are statistical tendencies, not guarantees.

Key points

• Specific gravity is oven-dry mass over the mass of an equal volume of water; the volume basis (green, 12 %, oven-dry) must always be stated.
• Basic specific gravity is the most repeatable reference; density at 12 % MC is the usual trade figure.
• Czech measurements at 12 % MC give conifer means of roughly 390–650 kg/m³ (near-pith to outer wood) and averages of roughly 670–750 kg/m³ for beech and oak, with large scatter within species (single pine boards in a Polish stand: 469–686).
• Most density variation within a spruce stem is inside the annual ring; the pith-to-bark rise in the cited conifers is roughly 80–160 kg/m³.
• Density predicts strength and hardness fairly across species, stiffness poorly, and shrinkage and fuel energy per volume closely.
• A few weighed and oven-dried samples per batch give cheap control over yield, drying and product allocation.