Measuring years from ice cores

A reader asked the question in one of my posts about Antarctica:

DPF, while you're there talking to these learned scientists, can you ask them how they work out how old the ice is. When they take an ice core sample how are they to know what year each level of snow comes from. Especially as there is continual snowing and melting over the ages. Its a little different to counting rings of a tree or looking at layers of sediment.

I duly asked the scientists and got this reply back:

You are correct to observe that there could be missing layers in ice cores – i.e. when there was little snow fall during that year, or when strong storms might have removed a whole year's worth of snow or more.

Saying that our is first of all based on counting annual layers. We normally drill ice cores from the surface (at times with overlapping cores) and since we know the age of the surface (‘today'), we can start counting layers back in time. This can be done through geochemical indicators such as isotopes (show the seasonal temperature variations so we can count warm summer, cold winter as couplets of years) or water chemistry – such as changes in sea salt concentrations which comes from the seasonal changes in sea ice for example. There many, many different indicators that we can use and we like to use as many of those as possible to give us confidence.

But all of those would more or less suffer from the same issue that some year's no snow remains or some years show signs of double peaks which we could count as two years. For this reason we also use absolute age benchmarks. These are often volcanic eruptions (which we can see as actual deposit of the ashes or though geochemical changes in the water chemistry). Other possibility are radioactive age benchmarks such as nuclear testing in the mid 1950 and 1960s or natural changes in the production of some of these isotopes (such as 10Be). These age benchmarks helps us to determine how many years might be missing between those benchmarks. Furthermore we use the greenhouse gas record to correlate between ice cores (this is for older ice – where annual layer counting might become difficult). Since greenhouse gases in the atmosphere mix very well across the world, when we have a really well dated ice cores (such as from Greenland where high snow fall makes annual layer counting much easier than in Antarctica) we can use the correlation between their greenhouse gas records to determine the age of another ice core. We also employ models to help us date old ice – such as the Vostok and EPICA ice cores which span 400,000 and 800,000 years each.

Quite fascinating stuff.