How Do You Find a Boundary Forty Kilometres Down? Time the Earthquakes

Andrija Mohorovicic was examining records of an earthquake near Zagreb in 1909 and noticed that stations at a certain distance recorded two separate arrivals of the same kind of wave rather than one. He worked out that one had travelled directly through the shallow rock while the other had travelled down into a deeper layer where waves move faster, along it, and back up, arriving first despite the longer path because of the higher speed. That required a sharp boundary at depth with faster material beneath it, and calculating when the two arrivals swap order gives the depth of that boundary.

The layer beneath differs from the crust in composition rather than in state:

• •Wave speed jumps sharply across it, which is how it is detected
• •Crust above is lighter rock, rich in silicon and aluminium
• •Mantle below is denser rock rich in magnesium and iron
• •Both are solid, so this is not the boundary with anything molten
• •Depth runs from about five kilometres under oceans
• •To seventy kilometres or more under mountain ranges

Reaching the boundary has been attempted repeatedly for over sixty years and has never been achieved. An American project in the early 1960s aimed to drill through thin ocean crust and was cancelled in 1966 after funding and management problems. A Soviet borehole on the Kola Peninsula reached over twelve kilometres by 1989 through thick continental crust, which is the deepest hole ever drilled and is nowhere near the boundary there. The obstacles are temperature, which softens the rock and the drill, pressure, which closes the hole, and the practical difficulty of working from a ship in deep water. Projects continue and the target remains unmet.

The boundary can be examined without drilling to it, because slices of ocean floor have occasionally been thrust up onto land intact during continental collisions. Those sequences preserve the whole structure from deep mantle rock through the crust to the sediments that lay on the sea floor, exposed where geologists can walk across it, and the best known are in Oman, Cyprus and Newfoundland. Studying them established what the boundary actually looks like in rock rather than in seismic data, and revealed that it is frequently a transition zone tens or hundreds of metres thick rather than the sharp surface the wave speeds imply.

The attempts produced results that had nothing to do with the original objective and changed geology anyway. The Kola hole found no transition to a different rock type where seismic data had suggested one, and instead found that the change in wave speed at that depth came from cracks filled with water under pressure, which nobody expected at such depths. It found water chemically bound in rock far deeper than thought possible, and microscopic fossils in rock at six kilometres. Temperature rose far faster than predicted, reaching a hundred and eighty degrees at the bottom, which is what finally stopped it and which mattered for estimates of the Earth's heat budget.

Two arrivals of the same wave at stations a certain distance from a 1909 earthquake revealed a sharp boundary with faster material beneath, and the distance at which they swap order gives its depth. The crust above and mantle below are both solid and differ in composition. Sixty years of attempts have not reached it, and the deepest hole ever drilled found unexpected water and fossils instead.