"There are uncertainties how wind, drought and regional temperature patterns will change in response to increasing greenhouse gas concentrations," said Michael Mann of Pennsylvania State University in the Science podcast, "because uncertainties remain in the climate models that we use to make those projections and they don't all agree on how the El Niño phenomenon will change in the future. We can address those types of question from the palaeoclimate record as well."
Mann and colleagues used proxy data from tree rings, ice cores, coral, sediments and other records of past climate. By calibrating this proxy data with instrumention data for the period from 1850 to 1995, the team was able to reconstruct surface-temperature fields, in a so-called climate-field reconstruction approach.
The researchers focused on spatial patterns during the Medieval Climate Anomaly – a period between the 9th and 13th centuries when average northern-hemisphere temperatures were relatively warm compared with subsequent centuries – and during the Little Ice Age between the 17th and 19th centuries, when temperatures averaged over the northern hemisphere were relatively cool compared to the medieval period.
"The Medieval period, even though relatively warm compared to the Little Ice Age, compares on a global sense at most with the level of warming we saw in the mid-20th century," said Mann. "It doesn't reach the levels of warmth that we've seen in the most recent decades, at least globally."
Mann and colleagues compared their temperature reconstructions for these two time periods with outputs from two climate model – the US National Center for Atmospheric Research (NCAR) Climate System Model 1.4 and NASA Goddard Institute for Space Studies-ER (GISS-ER). They drove the NCAR model with estimated solar and volcanic forcings, and the GISS model only with solar forcing.
"This gave us a sampling of two different physical representations of the way the climate system responds to these different factors," said Mann. "We could look at both of those models, see what they predicted and see how it compared to what we actually observed."
The results reveal that the tropical Pacific Ocean may have acted as a "thermostat" during the Medieval period.
"We observed that the surface-temperature patterns during the Medieval era counter-intuitively look like the La Niña phemonomon, when it's cold in the eastern and central tropical Pacific," said Mann. "This is a period of time when solar output was relatively high and there was relatively little explosive volcanic activity, so there was a net heating of global climate. Yet those same factors appear to have driven the climate into the cold La Niña state of the tropical Pacific."
The team's results indicate that the tropical Pacific Ocean was actually colder during the Medieval Climate Anomaly than it was during the Little Ice Age.
Neither the NCAR- nor GISS-coupled climate models show such a tropical thermostat. "There is only a relatively small subset of coupled climate models that exhibit this response," said Mann. "Most of the models do the opposite – when you warm the tropical Pacific the models respond with an El Niño-like state, and when you cool the tropical Pacific they respond with a La Niña-like state."
In fact, the majority of the models used in the most recent IPCC projections do not show this thermostat response, instead favouring an El Niño-like response to the surface warming currently being caused by increased greenhouse-gas concentrations in the atmosphere.
"However the palaeoclimate record of the past 1000 years in our analysis suggests that, at least with respect to the response that the climate has exhibited to natural factors, the response appears to be that thermostat – it appears to be the opposite of what most of the IPCC projection models project," said Mann.
If the tropical Pacific Ocean does respond in this way under future climate change, there will be serious implications for future drought patterns.
"In the next IPCC assessment it may very well be the case that there will be a new section that deals with the question of how palaeoclimate data can inform our understanding of some of these fairly complex dynamical responses of the climate," said Mann. "While I don't see our results as being explicitly made part of an IPCC projection, I do see them as potentially informing our assessment of the extent to which the current generation models are, or are not, capturing some of the regional mechanisms that may be important in making regional climate change assessments."
The researchers, from Pennsylvania State University; Roger Williams University; the University of Massachusetts, Amherst; the University of Arizona; NASA Goddard Institute for Space Studies; and the US National Center for Atmospheric Research, reported their work in Science.