Skip to content

Using sedimentation rates to infer long-term global climate change

June 28, 2010

ResearchBlogging.orgOver geologic time scales, the Earth naturally captures carbon dioxide from the atmosphere through weathering of silicate rocks and sequesters it via the production of carbonate rocks. Ultimately, subduction can return these rocks to the Earth’s interior and carbon dioxide is once again emitted into the atmosphere by volcanism. Thus, understanding the history of erosion and, by association, weathering of our planet’s surface will help us understand this important geochemical cycle and its relationship to climate.

Oxygen isotope curve for Cenozoic (Zachos et al., 2001; Science)

Paleoclimate records for the past ~50 million years show overall global cooling with higher-frequency fluctuations (see plot of δ18O measurements, a proxy for temperature, at left). The annotation on the right side of the figure denotes when researchers think continental glaciation was initiated — approximately 30 million years ago for Antarctica and 5-8 million years ago for the northern hemisphere.

What caused this long-term global cooling?

Some researchers have hypothesized that increased tectonic uplift (especially the Himalayas) and the increased weathering that came with it is one of the primary causes of this long-term global cooling*. That is, as more carbon dioxide was withdrawn from the atmosphere and naturally sequestered, the Earth cooled. One line of evidence geologists have used to infer changes in rates of erosion and weathering is the depositional record.

Sediment-dispersal systems transfer eroded material from mountainous uplands to sites of sediment accumulation in alluvial lowlands, coastal deltas, the continental shelf, and the deep sea. Because erosion in one location is balanced by deposition in another location, some geologists are using changes in deposition rates to infer changes in erosion rates. This seems like an elegant solution but, as usual, it’s not quite so simple.

I was delighted to see this paper by Willenbring and von Blanckenburg about late Cenozoic erosion/weathering rates come out in Nature last month. There’s a lot to say about this paper and its implications, too much to cover in this one blog post. I recommend this great review article by Yves Godderis, which highlights the global paleoclimate implications of the study nicely. Here, I will focus on one aspect I find fascinating: how erosion and deposition is recorded in the stratigraphic record.

Willenbring and von Blanckenburg revisit an inherent measurement bias demonstrated by Peter Sadler in his 1981 seminal paper “Sediment accumulation rates and the completeness of stratigraphic sections.” I’ve written about the ‘Sadler Effect’ before on this blog (read this post from 2007) but to make a long story short — the longer the measured time interval the smaller the sedimentation rates and vice versa.

Sediment accumulation rates and erosion rates as functions of geological time (Figure 2 from Willenbring & von Blanckenburg, 2010, Nature 465; doi:10.1038/nature09044

This phenomenon arises because the accumulation of sediment is unsteady — that is, at one location it varies over time. As longer durations are measured in a succession, more hiatuses, or periods of no deposition, are captured and, thus, the rate (thickness/time) is lower. Conversely, rates measured from very short durations can be quite high. Think about it — if you go out to a river mouth when it’s flooding and measure the thickness of sediment that piled up over a day and converted it to a yearly rate by simply multiplying by 365 you’d get a huge rate that wildly overestimates the actual yearly rate (unless, of course, that river is flooding every single day of the year). This is intuitive and few would attempt to make such a conversion, but this is essentially what is happening when comparing sedimentation rates that were measured from different time scales. Similarly, geomorphologists have discussed the unsteadiness of processes such as erosion and uplift^.

Willenbring and von Blanckenburg show this by plotting various process rates against time. The plots at right show, from top to bottom, (a) ocean-basin sediment accumulation rates, (b) volumetric erosion rates, (c) sediment accumulation rates in Asian offshore basins, and (d) global denudation rates. Note how all of them show a huge uptick in rates during younger time intervals. The point here is that this apparent increase in rates is an artifact of measuring unsteady processes — it isn’t real. The inset log-log plots of the same data show the Sadler Effect clear as day.

So, does this mean we can’t use measured rates of deposition to say something about the variability of processes in Earth history? I think we can but we need to significantly increase our understanding of rates of processes at multiple time scales. Careful documentation of sediment-dispersal systems with ever-improving geochronological tools is one approach to unraveling these complex temporal relationships. Numerical/physical experimental methodologies can be designed to address process rate questions as well. Ultimately, the integration of multiple approaches will lead to a better understanding of the complex temporal relationships of sediment erosion, transfer, and deposition and how we can utilize the stratigraphic record to reconstruct Earth history.

Willenbring, J., & von Blanckenburg, F. (2010). Long-term stability of global erosion rates and weathering during late-Cenozoic cooling Nature, 465 (7295), 211-214 DOI: 10.1038/nature09044

* just one example is this Raymo and Ruddiman paper from 1992

^ Gardner et al. (1987); Geology

9 Comments leave one →
  1. June 28, 2010 8:34 am

    Very nice post Brian. I’ve always found the ramping up of deposition rates towards the present day rather suspicious, and now I understand why!

  2. June 29, 2010 4:10 am

    So if erosion of old crust doesn’t increase, what causes the cenozoic change in seawater Sr isotopic ratio?

  3. RenatoCSfonsecA permalink
    June 29, 2010 9:27 am


    For a crude aproximation of depositional/denudation rates, perhaps the area of the graphs that shows a plateau could be considered. What do you think?

  4. June 29, 2010 4:08 pm

    Lab Lemming — I don’t think what’s discussed above necessarily refutes the idea that there could be increased erosion. Point is that it’s tricky to use sed accum rates as evidence for it, that’s all.

    In their paper they mention that: “… isotopes such as Sr from ocean sed time series are good proxies for processes related to denudation style and source-rock isotope composition but are not necessarily good indicators of the weathering flux magnitude.” I’m not an expert in these matters … I might defer to you for the validity of that comment.

    Renato — I’m not sure what you mean — to get an average?

  5. June 29, 2010 10:09 pm

    Increased Sr ratios come from dissolution of old, (K-rich) continental crust. so it tells you nothing about rates from erosion of juvenile crust (e.g. arcs), but is thought to ba a good proxy for cratonic erosion. The conventional wisdom is that the himalayan uplift increased erosion of old indian craton rocks with high Sr isotopic ratios, and that is the main recent forcing on ocanic values.

    I have no idea what the relative sediment fluxes are from the Himalayan rivers vs. juvenile arc drainages (e.g. the andes, PNG, Japan)

  6. RenatoCSfonsecA permalink
    June 30, 2010 11:28 am


    This is what I mean: If we get a plateau over a significant
    portion of the graphs, couldn´t that reflect the best estimation
    of the long-term rate of deposition/denudation for a particular area? Over relatively short-time periods, maybe one possible estimation could come from actual counting of the number of laminae and bedding contacts, in “continuous” successions (no major breaks in the sedimentary record. Arbritary, but reasonable durations, could be attributed to these “minor” hiatuses. Continuously cored intervals could provide the means to do this kind of work. This problem is analogous to trying to determine the pluviometry of an area by measuring the volume of water poured out in a single event, not taking into account that it doesn´t rain all the time, etc.

  7. Julien permalink
    July 11, 2010 6:27 pm

    I’m sorry I’ll be a bit out of the topic but I’m trying to find the latest (updated) climate and sea-level data available on the Neroproterozoic to Cambrian.
    I’m working on a carbonate-to-clastic filled graben where I’d like to know if 2nd orders cycles are more influenced by local subsidence (margin backsteps in diapirism context) rather than “global” changes
    Being a former damned Quaternary geologist, I have no ideas of which work will be the most accurate for that. And I’m pretty sure you guys know solid new refs.

  8. July 12, 2010 6:27 am


    I don’t know much about Neoproterozoic-Cambrian … you might try Chris Rowan from the blog Highly Allocthonous, I know he’s worked on those time periods before:

  9. Passerby permalink
    July 13, 2010 11:30 am

    There is a confounding factor in erosion-depositional rate processes: human-induced change in landscape at local, regional and global scale.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: