UCD-led study shows sea-level rise is accelerating - Dr Andrew Parnell
Rising sea levels are a threat to economics, population and infrastructure, but how fast is that tide encroaching? And how do rates of change in recent times compare to rates stretching back through millennia? A new analysis led by UCD researchers brings those answers - and the rapid rates of change - into context and stark focus.
“We have shown that sea level rise on the US east coast has been much less than 1mm per year for the entire period 0AD to 1800AD and since then it’s skyrocketed,” says principal investigator Dr Andrew Parnell, a lecturer in statistics at UCD School of Mathematical Sciences. “We found the rate of sea level rise on the east coast of the US is the highest its been for at least 2000 years, and the rate of global sea level rise is above the 1.7mm per year stated by the International Panel on Climate Change.”
Taking the long view
The study used existing measurements stretching back hundreds and even thousands of years, and applied statistical techniques to take uncertainty in the datasets into account. “We looked at these longer records to compare with rates of sea-level rise that we are seeing now,” explains researcher Niamh Cahill, a graduate student at UCD School of Mathematical Science and lead author on the study in the Annals of Applied Statistics. “We are seeing rates increasing now, but unless we look back at the past, we don’t know how unusual those rates are.”
Cahill, Dr Parnell and colleagues Professor Andrew Kemp from Tufts University and Professor Benjamin Horton from Rutgers University and Nanyang Technological University focused on two types of data - ‘instrumental’ measurements from tide gauges around the world dating back to the 1880s, and even longer-term ‘proxy’ measurements taken from cores of sediments on the east coast of the United States that offer clues about sea levels over the last 2,000 years.
In both cases, there was uncertainty in the datasets, explains Cahill. “For the the instrumental data from tide gauges, the further back in time you go the fewer tide gauges there were around the world,” she notes. “So the more recent part of the record has a lot more certainty in it as there are a lot more tide gauges, but if you go a little further back in time there is more uncertainty about the sea level because there are fewer measurements.”
The proxies from cores of sediment taken in North Carolina also have their limitations - by analysing the remains of tiny biological fossils in the cores it’s possible to reconstruct tidal patterns across 2,000 years, but these ‘proxy’ records bring with them uncertainty not just in sea-level measurements but also in time.
More ‘honest’ findings
To tune up the findings from the datasets, the UCD statisticians put the datasets through their paces. “We wanted to develop a model that would take account of the fact that there is a lot of uncertainty in these datasets,” explains Cahill. “We wanted it to be as honest as we could make it.”
Applying a statistical method known as errors-in-variables integrated Gaussian processes, the researchers crunched the numbers from the datasets and confirmed that sea-level increases have sped up vastly in recent times.
Their analysis of the global tide-gauge record showed that the rate of sea-level rise increased continuously from 1.13 mm per year in 1880AD to 1.92 mm per year in 2009AD. Meanwhile, the mean rate of rise in North Carolina at 2000AD was 2.44 mm per year - the fastest rate of rise in the 2000-year long reconstruction - the difference in rates between North Carolina and the global tide gauges illustrates the role of local processes such as the land subsiding, explains Cahill.
Pulling the story from the noise
“They are noisy datasets and we wanted to pull the story out of the noise,” she says. “If you eyeball the data most people will say that looks like it is increasing, but we are showing it with the stats behind it. We are taking all the uncertainty into account and we are still coming to the same overall conclusions.”
Trawling back through the proxy data from North Carolina sediments, the UCD researchers’ findings also showed some variations in the rates of sea-level change over time. “There is an increase in the rate around the time period known as the ‘Medieval Climate Anomaly’ and there is a small decrease in the rate consistent with the ‘Little Ice Age’, when things are slowing down,” says Cahill. “Then at the end of the 19th century, which is after industrial revolution has started, sea-level rise increases rapidly.”
“Fiddly but necessary”
Taking account of the uncertainties to produce meaningful results is “fiddly but necessary” according to Dr Parnell. “I believe this study is the first time a group of statisticians have had a really close examination of sea-level data, and have properly taken account of the uncertainty in the data points,” he says. “When people [previously] calculated rates or accelerations of sea-level change they were always quite hesitant to talk about their findings as they knew that they were ignoring parts of the data, such as the age uncertainties, or using techniques which didn’t really stand up to scrutiny. In general this study should point to a way forward where people can confidently quantify sea-level rates for data sets at their own study sites.”
The findings should also feed into better models for future sea-level predictions, notes Cahill. “When you can access the underlying variability in the sea-level trend, then you have your uncertainty around that underlying variability,” she says. “You can then use that information to constrain the models that are going to predict into the future, so you should be better able to model more meaningful predictions.”
And while she believes there is a general public awareness of sea-level rise, she argues that the findings of this UCD-led study could help drive home the pace of change. “It is not going to come as a shock, we all know sea levels are rising,” she says. “By using these statistical methods we are showing that this rise is accelerating over time. This needs to be taken into account for decisions about coastal defences.