Palaeotempestology: Lake sediment records

Digging in to sediment records

With continued debate among scientists on exactly how future climate change will affect storm frequency and severity, it seems logical to see if we can find out more about variability in storm activity from the past.

Lake sediments are extremely useful in studying past climates, for which we have no observational record (through conventional weather recording equipment). They provide a slice through time to look at the changes in lake chemistry and environmental activity affecting the make up of suspended particles in the lake that eventually settle at the bottom.

Radiocarbon dating, thickness of layers of different sediments, analysis of diatoms and inference from the occasional break in the record (a hiatus, perhaps due to the drying out of a lake), are various ways in which lake sediments can give us clues about the past.

Within this range of different approaches there are a few ways in which sediments from lakes can be used to look at past storm events. In my previous blog, I highlighted a paper by Dr Jeff Donnelly et al. in 2015 entitled “Climate forcing of unprecedented intense-hurricane activity in the last 2000 years”. It presents a history of storm events over the past two thousand years, using an analysis of sediment grain size in their collected samples, with a resolution of around 1 year. The work uses evidence gathered from field work during the project (and previous studies) to determine the presence of two distinct periods of higher activity in severe hurricanes for the west North Atlantic coastline of North America: one between 1400 and 1675 C.E.; and another period of high frequency storms further back in time between 250 and 1150 C.E.

The study location is a place called Salt Pond, in Massachusetts. It has a tidal inlet linking it to the ocean, making it full of brackish waters. This proximity to the ocean means that the pond is exposed to ‘overwash’ during storm surge events associated with large storms heading northwards along the Eastern seaboard of the United States. These salt water incursions occur when the storm surge level is higher than any natural or man-made defences. This ‘overwash’ leads to ‘coarse grain event beds’, and so these can be used as an indicator of severe storm activity. This process is vaidated using known hurricanes landfalls, which are represented in the sediment records and act as ‘anchors’ to verify that the samples are valid.

The study builds on a number of papers that were produced after the convening of a workshop on Altlantic palaeohurricane reconstructions in 2001 at the University of South Carolina. The workshop aimed to identify new opportunities in the field of palaeotempestology. A summary of the workshop can be found here. Dr Jeff Donnelly and colleagues studied a number of lakes in the Northeast of the US, in the states of New Jersey and New England, and so to learn a bit about the methodology, I dug into some of the papers in some more depth.

Getting your hands dirty

It seems the only way to get at clues available from sediment records is to get your hands dirty. I found an earlier paper by Donnelly at al. from 2001 which built a 700 hundred year sediment record of severe storms in New England. This paper (and a couple more in Boldt et al. 2010, Liuand Fearn, 2000) started to show me that each project strategy is subtly different. 

Various schemes are planned based on the conditions of the study sites, to find the best locations for sampling overwash areas in a consistent manner. The aim is to try to consistently capture the process by which more intense storms erode more sand from the coastal beach and bring this coarse sediment into the brackish lakes and ponds, larger storms being assumed to produce wider fans of overwash sand deposits, being thicker near the shore and thinner near the centre of the study lake. A range of
samples should be taken to try to represent the range of possible characteristics of past intense storms. Figure 1 (below) is a hypothetical diagram from Liu and Fearn (2000) to show various patterns of deposition. Note the radial patterns associated with the various directions of storm approach, with the larger fans associated with more intense storms.

Figure 1: Hypothetical coarse grain deposition fans in severe storm surge events. Source: Liu and Fearn, 2000 

The coarse sand creates a layer over the more usual organic-based deposits that settle on the bottom of a lake as a stratified layer. This happens most effectively in anoxic lake beds (lacking dissolved oxygen) since any mixing from plant of animal life will be minimal.

Having never been in the field to collect sediment samples, I found it interesting to see how Donnelly et al. (and other teams) maintained a consistent chronology in the sediment records. They took multiple samples and use the variety of methods above to build their chronology.

Markers in time

Isotopic radio carbon dating and stratigraphic markers used to mark certain control points to validate the data. Pollution horizons are useful in this respect, for example lead concentrations mark the beginning of the industrial revolution as it quickly made it's way into the water systems and lakes and then 'fixed' by anoxic sediments. The presence of lead pollution is an indicator of the late 1800's (Donnelly et al. 2001) and then another change occurs when lead was removed from gasoline in the 1970's and 1980's. This is a good example as it shows how these markers are useful for calibrating sediment records, in a way that is easily understood and recognised.

Pollen records can also mark certain points in history, for example the European colonisation of the eastern U.S. led to large scale clearance of the vegetation for farmland meaning that the pollen composition changes drastically (Russell et al. 1993).

Once these markers are established, previous storms are used to calibrate storm events, and then previous coarse grain even layers are identified and carbon dated.

Clear as mud?

So having learned a lot more about sediment analysis in relation to palaeotempestology, I now have a greater respect for what these cores of old mud and sand can tell us about the past. However, it does seem to me that there is still a large degree of uncertainty in the data when trying to discern an idea about individual storms. For example, what if two storm occur in quick succession as a cluster, before a sediment layer has had a chance to settle and ‘lock in’ the information? This may end up looking look like one larger or more intense storm, when actually it is the frequency of storms in that season which is varing. Donnelly et al. 2001 give an example from their study location of a lack of agreement between historical accounts of two intense storms in 1635 and 1638 which likely created overwash signatures, but in the sediment proxy data, only one event was indicated. This means that the estimated frequencies may have significant uncertainty.

Also, responses of lake or pond to overwash events may change over time due to changes in natural or man-made barriers. However, even with these uncertainties in mind, it is still clear that there is great value in understanding the past clues left behind by storms in our coastal lake sediments. 

Without any alternative information, the best that we can do is to piece together palaeotempestological proxies and glean snippets of information to build a longer record of storms.

It also provides grounds for comparison in using climate models to try to understand past variability,
another subject I intend to explore in a future blog.

For now, I’ll leave you with an informational video by Ocean Today in conjunction with the Smithsonian Institution and NOAA, just after Hurricane Sandy in 2012 which will hopefully make a clear demonstration of what overwash looks like and how the coastal beach material can be dragged in across to end up in lakes or ponds that lay close to the ocean to give us these markers of past events.

My next blog will be on the evidence that can be derived from coral cores.

The calm *blog* before the storm...

The standard advice about starting any writing endeavour is to write about what you know. Well, I’m here to use this blog to learn more about the complex issue of climate change, so that leaves me somewhat limited, but since no journey starts without a first step my starting point will be storms.

Now, I’m not talking about your everyday ‘it’s a bit blowy outside’ kind of storms, but instead I mean the big ones, the rare beasts that wreak havoc to those in their path, with impacts that spread way beyond those who are unlucky enough to be directly in the firing line.

In a changing climate, extreme storms are expected to also change in terms of their frequency and severity, but there is much uncertainty. It’s not just the storms themselves that are likely to change. In the future our exposure and vulnerability to their impacts will also evolve. How these factors interlink, and exactly how extreme storms are likely to change, are subjects of much debate and a key question in the field of climate change science. A warmer world is expected to change how often we see storms like Hurricanes Katrina which ravaged an ill-prepared New Orleans, or Hurricane Sandy which flooded much of the east coast of the U.S. including New York  (comparison before and after images on this link - worth a look). Then there are clusters of windstorms like Lothar and Martin that hit Europe just after Christmas in 1999 causing billions of Euros in insured losses, let alone the amount of uninsured damage, which may also change in frequency in the future (the subject of a future blog I'm sure). Storms like Super Typhoon Haiyan caused immeasurable suffering and loss in the Philippines in 2013. Tragic loss of life reached into the thousands, and the huge amount of damage was largely uninsured with government responses barely able to help those affected and international aid being a large source of the relief. This highlights the global differences in vulnerability to these extreme events. However, in the context of climate change, the often-asked questions are: “Can ‘storm x’ be attributed to climate change?” or “Are we likely to see more storms similar to ‘storm x’?” Not easy questions to address, but so in this blog I’ll be searching for the answers that are out there (if any), and look at various sides of the science in a quest to respond to these very reasonable questions.

Hurricane Katrina on its way towards New Orleans. Source: NASA

With those sobering thoughts in mind, I intend to use this blog as an exploration into what we know about tropical cyclones and their relationships to climate in the past, present and future. From palaeotempestology looking back thousands of years, to high resolution climate modelling out to 2100, I’ll be bringing together interesting facts, reviewing papers, and hopefully bringing out some entertaining and engaging moments along the way.

Perhaps just a quick word about me before we get going, for the sake of good manners. I’ve just started a part-time Masters degree in Climate Change (hence this blog), but also work full time coordinating and leading research projects for a large insurance/reinsurance broker. My background is in meteorology having worked for over a decade as a weather forecaster in the UK and in Bermuda (I wanted to forecast hurricanes).

I also have an interest in the communication of science having worked in the media side of forecasting at the start of my career, and studied at art college before moving on to a BSc in Environmental Science at the turn of the millennium. So basically, I’m a self-confessed weather geek, who likes paintings and films. I've contributed to my company's blog a few times, but never run my own site, so I'll hopefully build some good content on this page as I go along.

I'm always happy to receive comments about the weather and climate, especially severe storms, so feel free to reply on any posts and I'll respond as soon as I can.