On the nature of time information in the fossil record
What is the fossil record, you wonder? It is a facebook of life history and it has stories to tell. It knows which animals socialized with which millions of years ago, what they ate for their dinner and how they got into various troubles.
Fossils are organisms from the past imprinted in rock, palaeontologists find them and keep the record. The term fossil record refers to everything that is known, yet the actual record is scattered over many sources. Just like there is no single global social network, there is no single fossil database, there are many. My closest to home fossil database records which fossil mammal species lived where in the world over the last 50 or so millions of years. Data data everywhere.
Along with global trends in science, palaeontology research is becoming more and more computational. The frontier of palaeobiology these days is associated with computational modelling of evolutionary processes and patterns. Conventional palaeontological studies of fossil material emphasize computational placement of organisms into ancestry trees. And while the world is changing more rapidly than ever, more and more researchers from biosciences and climate sciences turn to the fossil record hopeful to infer from the past data what’s coming in the future.
So how do we know how old a fossil is? This article is meant to help computationally inclined researchers entering the fossil world to understand the nature of time information and uncertainties associated with it. No background knowledge in stratigraphy, geology or palaeontology is assumed. Specialists are likely to find this account oversimplified and lacking details. So it is.
Another one bites the dust
Most non-specialists know that fossils as well as archaeological artefacts are found in the ground. Coming from a medieval town of crooked streets, battles and saints I learned early in my life that nearly every construction in the old town unearths artefacts from earlier cities. Walking near open pits waiting for pipe reconstructions to complete one could see remains of medieval streets meters below the current level.
As a kid I thought that items or even streets slowly sink over time. I was, of course, promptly explained that this is not the case, and that rather dust piles up on top. How terrible, I thought, that hundreds of years from now only top floors of our houses will be operational.
Talking about fossils we usually talk about millions rather than hundreds of years. During so long times not only dust accumulates, but rock layers of Earth move around, mountains rise and erode, land turns into seas, seas become land and fossils come along.
Fossils are remains or traces of organisms that lived in the past and later turned into rock. For any individual organism fossilization is extremely unlikely¹, but over millions of years and myriads of organisms it happens often enough to leave a record. Fossilization is more likely if an organism quickly gets buried limiting exposure to oxygen or scavengers that would destroy the remains. Thus, the fossil record is by and large a record of misfortunes — somebody fell into a narrow cave, drowned in a lake or got washed away by floods.
Here I always remember a story of an unfortunate indricothere told by Donald Prothero in his book. Indricotheres, relatives of rhinoceroses, were the largest land mammals ever lived², about twice the height of an elephant. An expedition of palaeontologists in Mongolia stumbled upon a huge fossilized leg fragment standing upright. It must have been about the size of a human torso. Soon three more legs were unearthed nearby, all upright. That is highly unusual. The team rapidly concluded that there was only one plausible explanation — the poor creature must have sunk in quicksand perhaps while attempting to drink from a pool. When one of the remote expedition leaders asked why only legs were recovered, nothing else, another replied:
“if you have brought us here thirty-five thousand years earlier, before the hill weathered away, I would have the whole skeleton for you!”
Natural museum collections often display well-preserved fossil skulls or skeletons, but those are in fact very rare. The majority of the fossil record consists of fragments — pieces of bones, shells, isolated teeth.
Time what is time
The fossil record as we, analysts, know it usually comes in a form of flat table listing items with their geographic location, age, taxonomic affiliation³ and optionally other descriptive attributes. One way or the other data analyses of the fossil record are usually about extracting and comparing descriptive patterns at different places and times, as well as analyzing how those patterns change over time. The time is always part of such analyses. And while affiliation and descriptive attributes can be inferred directly from the fossil itself, age (thus time) is usually a property of the place⁴ where the fossil was found. Thus, fossil time is static, it is effectively frozen in rock.
So how to infer the age of a fossil? The most straightforward way is to ask what is the age of the rock layer in which the fossil was found. That is called absolute dating. The bad news is that usually absolute dating is not possible.
Absolute dating is based on decay rates of radioactive elements (e.g. carbon, potassium, uranium). Radioactive elements are not stable, they are eager to turn into something else. When formed, a radioactive element is in its full known concentration and this concentration declines over time at known rates. When a fossil is found one can sample the rock layer nearby, measure the remaining concentration of a radioactive element and how long it must have taken for the element to decay to this level. The problem is that this gives the time when the chemical element was formed, not when the layer was formed. Most of the layers are formed of recycled elements, that is grains of rock that existed already when they were transported here by water or air.
Only certain types of rock layers form their elements at the time of their layering, one of such is volcanic ashes, another is magma. Thus, when volcanos erupt and lay around their ashes or magma, those layers can be dated. Such layers are rare, thin and usually contain no fossils⁵, though. But that is where vertical positioning of layers comes into play.
Until it sleeps
If it wasn’t for the current pandemic and countless online sessions from the kitchen office, I would have never found how instructive are my tablecloths for explaining geological time.
One of the key principles in geology is superposition, coined by Nicolas Steno a Danish scientist and a Catholic bishop from the XVII century. The principle says that rock layers that are on top are younger than those that are in the bottom. This is because new layers always pile up on top⁶. When those layers form, fossils form along and stay within their host layers for millions of years until those layers get exposed on the surface of Earth.
Exposure happens due to movement of large plates that make up Earth’s crust. Even though they move very slowly, at about the speed that human nails grow, over millions of years many things can happen. Plate movement causes stress, as a result rock layers on the surface crack exposing older layers. Over time due to rains, winds and other forces the elevated parts erode (=crumble away) exposing more old layers to the surface, which opportunities to find fossilized remains of organisms from those times appear. Old fossils can potentially be accessed climbing the slope.
Layers do not always stay vertically stacked. When the large plates of Earth collide, near the point of collision mountains lift up carrying along the rock layers possibly with fossils in them. Over time mountains erode and old rock layers potentially carrying fossils get exposed.
The whole process of mountain uplift and erosion is very slow, it takes tens of millions of years. Yet similar processes in action can be observed on lakes in winter, climate permitting.
Correlation does not imply causation
Back to time, how do we know the time?
The tablecloth example shows a very long sequence of layers, but typically only short snapshots of Earth’s history are exposed on the surface at different places of Earth. In lucky situations the exposed layer contains volcanic ashes or rock formed of cooled down magma, which can be dated. If a fossil is somewhere in between of two volcanic layers, due to superposition it is clear that the fossil cannot be older than the volcanic layer below and cannot be younger than the volcanic layer above it. Those layers can be dated quite precisely and give the minimum and the maximum age limits. One can even refine this further by the position of the fossil within those layers — if it is close to the top, one can refine the age bracket further, for example, say that the fossil is in the younger half of the computed age range. One needs to be cautious though, since the speed of vertical accumulation of sediment may be changing over times.
Complications start when there are no volcanic layers in the exposure and fossils still need to be dated. That is where relative dating methods come along. The idea is to find places of reference with known age (e.g. from volcanic dating) that are similar to the place of interest where a fossil was found. Finding similarities between sites is called correlation, thus when palaeontologists say “we correlated those two sites” it means that an agreement was reached that the sites in question are of similar age.
The prime option for relative dating is based on Earth’s magnetism. Currently our compasses point to the north, but during the history of Earth the magnetic field has flipped towards the south and back many times. This happened at irregular time intervals, roughly every few hundreds of thousands of years. The times of reversals are relatively precisely dated, there is a global reference scale, a snapshot of which is shown below. Black means the magnetic pole is in the north (like now), white means it is in the south. The names are just names people gave to the periods.
At fossil sites magnetic polarity of a layer can be inferred from orientation of small magnetic particles within the sediment. Roughly speaking, when sediments were brought to this place along water or as dust, those small magnetic particles oriented themselves to align with Earth’s magnetic field.
When the layer solidified the particles could no longer move, thus they remained pointing towards the magnetic pole of the past. With luck an exposure where fossils are found may contain several layers pointing to different magnetic directions from which one can extract a sequence of magnetic polarities, which works like a bar-code. Since magnetic reversals happened at the same time on Earth one can then match this bar-code to the global scale, and, infer ages from there.
The longer the bar-code from the fossil site, the more reversals it contains, the easier is to find a unique match with the global scale. But if finding a unique match solely based on the bar-code is too difficult, additional information can be used to narrow down the time around which to look for matches on the global scale. Such information can be, for instance, presence of indicative species of animals that are known to have lived at certain times.
When a unique match is found, one can infer upper and lower age bounds of the fossil site looking up the global reference ages of that magnetic bar. Potentially one can then attempt to compute a more precise numeric age based on the vertical position of fossils within that magnetic bar.
In practice, unfortunately, magnetic dating is not feasible very often. Only fine-grained layers, like clay, can be analyzed for magnetic polarities. If the grain size was too large upon arrival back in the days then particles did not have a chance to align themselves pointing to the magnetic pole. Even if the gran-size is suitable, the retrieved bar-code may contain too few or no magnetic reversals, thus it would not be possible to find a unique match. And even if there are enough of reversals within a site, due to interruptions in the deposition processes there may be too much uncertainty in the widths of the bars for them to be informative.
The zoo book
If all else fails, the most uncertain but still operational way of assigning age is by similarity to real or artificially constructed reference sets of fossil species. The reasoning goes metaphorically as follows. Suppose elephant, giraffe and lion fossils were found at a site of an unknown age. A known fossil site Meteora also has an elephant, a giraffe and a lion. The new site thus must be of a similar age to Meteora. The age of Meteora is known from volcanic dating, thus the same age can be assigned to the new site.
I can already hear objections. Yes, this is uncertain and complicated. An extra complication is that often there are no reference sites close enough that would have precise age. The reference sets are thus artificially constructed and approximate consensus ages are assigned to those reference units from any hints available.
Differently from magnetic polarity, those age boundaries are fuzzy. The same fossil species might have lived at different parts of the world at somewhat different times. Reference units represent implied average or consensus time and the community adjusts those reference times regularly. Age assignment by reference units is the most uncertain of all dating methods, but also the most widely used globally since often there are no better options.
- Some of the section titles are borrowed from songs by Metallica, Queen, Blind Guardian.
¹For an accessible introduction to fossilization, I highly recommend “I shall return: an intrepid reporter with dinosaur-sized ambitions discovers just how hard it is to become a good fossil” by Steve Mirsky published in Earth in 1998, it’s an illustrated story of a one-person quest to become a fossil.
² Dinosaurs were not mammals and whales are marine mammals.
³ Who is a relative of whom, and optionally who is the ancestor of whom.
⁴ That is not always true. In archaeology where time spans thousands or tens of thousands of years organic matter from the object itself can be used for dating. It is based on radioactive carbon decay rates. One can survey what proportion of radioactive carbon is left in a fossil and from known decay rates reverse engineer how long the decay took. This does not work over millions of years, unfortunately, because by then all the relevant radioactive carbon is effectively gone.
⁵ Yes, Pompeii preserving a snapshot of an ancient city is a rare exception.
⁶ This is not completely true, occasionally breaches may mess things up.
