Arthur Oldeman

The Earth is warming, this is caused by human activities, it has life-threatening consequences, and there are many things we can do to stop this. Future climate depends on future emissions, but also on the internal dynamics of the Earth system itself. We inform ourselves about possible futures by simulating the climate under different scenarios using climate models.

Coordinated scientific efforts have improved climate models and have led to an advanced understanding of climate change. Still, uncertainties on future climate projections exist, related to climate models and to internal climate variability. The most prominent source of year-to-year climate variability is El Niño - Southern Oscillation (ENSO), an internal climate oscillation caused by interactions between the ocean and atmosphere. Extreme weather events related to El Niño variability can lead to substantial damages worldwide. Understanding El Niño variability and its behaviour in the future is therefore of large societal relevance.

In the past decades, we have been able to verify simulated climate variability with observations, including from satellites and local measurement stations. However, we cannot observe the future. Fortunately, we can turn to the geological past. The Earth’s history includes many periods that were warmer than today, and of which we have indirect observations in the form of proxy reconstructions.

The most recent period with similar temperatures to near-future climates was the mid-Pliocene. It occurred around 3 million years ago and saw similar CO2 concentrations as the present. Because of these similarities, it has been called the ‘best analogue’ to near-future climates. Compared to the present, the mid-Pliocene was a warmer climate, which makes it a relevant period for studying climate variability in warm climate conditions. Apart from elevated atmospheric CO2 concentrations, the mid-Pliocene also saw strongly reduced ice sheets and closed Arctic sea-ways.

In this thesis, we investigate climate variability in the mid-Pliocene. We answer the following specific research questions:
1. How is interannual climate variability, specifically ENSO variability and atmospheric winter variability in the Northern Hemisphere, different in the mid-Pliocene compared to the pre-industrial?
2. Are there changes in the modes of variability and their interactions with the mean background climate, in the mid-Pliocene compared to the pre-industrial?
3. What is the role of elevated CO2 versus other climate forcing differences in past climate variability?
4. Can we propose guidelines for assessing climate analogy between the future and the past?
We answer these questions using simulations of the mid-Pliocene from global coupled climate models, and comparing to simulations of a pre-industrial reference climate. By answering these research questions, we have been able to draw some relevant conclusions on mid-Pliocene climate variability, and its potential as an analogue to the climate of the future.

We find a robust suppression of El Niño variability in the mid-Pliocene, as simulated by the Pliocene Model Intercomparison Project phase 2, PlioMIP2. Results on variability as well as on annual-mean sea-surface temperatures are in reasonable agreement with proxy reconstructions available. The suppression of ENSO variability is found to be related to shifts in the tropical mean climate, which includes changes to atmospheric circulation patterns.

Atmospheric variability in the North Pacific, which is partially driven by ENSO variability, shows a range of responses in the mid-Pliocene depending on the climate model used. We are able to attribute this range in responses to the particular sensitivity of climate models to the different past climate conditions. In fact, models that simulate intensified North Pacific atmospheric variability are more sensitive to elevated atmospheric CO2 concentrations. On the other hand, models that simulate a suppressed variability are generally more sensitive to the other mid-Pliocene conditions, including reduced ice sheets and closed Arctic sea-ways. Most of these conditions are not expected in the (near-) future, which makes comparisons with future climate variability difficult.

An important conclusion is that the physical connections between the mean background climate and climate variability are found to be quite robust. Both the link between ENSO variability and the North Pacific climate, as well as the link between North Pacific variability and atmospheric jet stream behaviour, are still present even in simulations that show large changes in the mid-Pliocene climate.

Finally, we present a practical framework which can be used to assess analogy between past and future climates. We employ it to re-evaluate the mid-Pliocene as potential analogue to the future. While mid-Pliocene surface temperatures are similar to those projected in 2100 under a medium-warming scenario, precipitation patterns are not comparable to any projected future. Furthermore, mid-Pliocene climate variability is generally not comparable to that simulated for the future. The variability changes in the mid-Pliocene are largely not driven by elevated CO2, suggesting that the mid-Pliocene climate is not suitable as an analogue for future climate variability.