Road cracking

Why do roads crack in regular patterns even where no vehicle ever drives?

Walk along the hard shoulder of a motorway and you find cracks. Not random ones either — often a series of transverse lines running across the surface at surprisingly even intervals. If traffic breaks roads, what broke this? And if it was not traffic, why the regularity?

Start with what a road is. Asphalt is stone bound with bitumen, and bitumen is not a solid in the way steel is: it is viscoelastic, flowing on warm days and behaving as a brittle glass on cold ones. Its stiffness changes by orders of magnitude across the temperature range an ordinary year contains.

Now cool that layer. It wants to contract. Can it? It is bonded to the base beneath and continues for miles in each direction, so it is restrained — and a restrained material that wants to shrink develops tension instead. On a cold night, in a material that has gone brittle exactly when the tension is highest, that stress can exceed what the bitumen carries. It cracks, right across the width, with no vehicle involved.

Here is where the regularity comes from. Once one crack opens, the material immediately either side is free to move and the tension there is relieved. Further away, stress rebuilds through friction and bond with the layer below, over a characteristic distance. So the next crack cannot form nearby; it forms out where the stress has climbed back to critical, and every subsequent crack obeys the same rule. The resulting spacing reflects material properties and restraint, not traffic. Concrete pavements make the point in reverse: engineers saw joints at regular intervals precisely because unjointed slabs will crack at their own natural spacing, and it is better to choose where.

So the cracks arrive without loading. What does the loading do? Two things, and they come after.

First, the crack is now a doorway. Water enters, reaches the unbound layers beneath, and softens them. If it freezes there it expands by roughly a tenth of its volume, prising the crack wider and, over cycles, breaking the surrounding material apart. Where traffic passes, trapped water gets pumped — pressurised and squirted out with fines from the base, hollowing out support beneath a surface that then has nothing to bear on.

Second, repetition. A single axle flexes the pavement and does no visible harm; the stresses are well below what would break it in one go. But each flexure grows microscopic damage at the bottom of the bound layer, and damage accumulates until cracks initiate and propagate, typically as the interconnected polygons called alligator or fatigue cracking. The road does not fail when something too heavy crosses it. It fails after enough things of ordinary weight have.

The arithmetic of "enough" is where intuition fails hardest. The AASHO Road Test, run in Illinois around 1960, drove controlled traffic over instrumented sections until they failed. Fitting damage to axle load gave the result that pavement damage rises roughly as the fourth power of axle load: double the axle weight, do about sixteen times the damage. A loaded truck axle against a car's is a ratio of several times over, raised to the fourth — so one heavy axle can be worth thousands of car passes.

Treat that as a rule of thumb, not a law. The exponent came from one test programme on particular pavement designs and climates, and later work finds it varies with pavement type, thickness, temperature and which mode of distress is measured — values from roughly three to five are quoted, and for some distresses the power law describes things poorly. What survives the caveats is the qualitative point: the relationship is strongly non-linear, and cars are almost irrelevant to structural pavement damage.

A few refinements to hold together.

Cracking is not one phenomenon with one cause. Regularly spaced transverse cracks usually indicate thermal contraction; interconnected polygons in the wheel paths indicate load fatigue; a crack tracing a line from below indicates reflection from a joint in an older layer; longitudinal cracks along a lane edge often mean a construction joint or a failing shoulder. The pattern is diagnostic, which is why inspectors read it rather than merely count.

The causes compound rather than add. Bitumen oxidises and hardens as it ages, so an older road cracks thermally at temperatures a new one tolerated. Those cracks admit water, which weakens the base, which increases deflection under each axle, which accelerates fatigue. Each mechanism makes the next one cheaper.

One correction worth stating plainly: cracks in an untrafficked shoulder are not evidence of poor workmanship or of "hidden" traffic. They are the expected behaviour of a restrained viscoelastic material in a climate with seasons. Traffic determines how fast a crack becomes a pothole, not whether cracks appear.

• How mechanistic-empirical design methods replaced the AASHO equations while keeping their empirical spirit.
• Why polymer-modified binders widen the temperature range over which bitumen is neither soft nor brittle.

Every term the collection defines is gathered in the glossary.