High-resolution insights into future European winters

Most—roughly 70%—of Europe’s winter rainfall is brought by extratropical storms, which are steered our way by the westerly North Atlantic jet stream. The wettest winters, such as 2013/14 (Figure 1), are often when Europe is at the receiving end of a veritable convoy of storm systems, the human and economic impacts of which are felt far and wide.

huntingford_2014_fig1
Figure 1 | Observed UK rainfall anomaly, expressed as a percentage of 1981–2010 monthly average, for (a) December 2013, (b) January 2014, and (c) February 2014. Figure from Huntingford et al. (2014). © Nature Publishing Group

What can we say about future winters? How will anthropogenic climate change impact the jet and, in turn, Europe’s wintertime hydroclimate? To find answers, we turn to global climate models that simulate large-scale atmospheric circulation and synoptic-scale weather. In principle, these models are no different from those that produce our everyday weather forecasts. They’re based on the same physics and dynamics that engender weather. Running these models for many decades into the future requires enormous computing resources. To make this feasible, compromises must be made, one of which is reducing model resolution.

An analogy: the sharpness of a digital photograph depends on how many pixels comprise it—in other words, on the camera’s resolution. The more pixels; the higher the resolution; the sharper the image. Climate models break down Earth’s atmosphere into three-dimensional pixels called grid cells. With a relatively low number of large grid cells (each typically 100–200 km wide), simulated weather systems not only appear pixelated when visualised, but are actually not all that realistic; their flows of air, heat and moisture don’t properly resemble those in the real world. This limits our confidence in using these models to make predictions. However, developments in high-performance computing have enabled the size of a climate model’s grid cells to be shrunk (to roughly 25 km in our case), thereby increasing their number and enabling the simulation of air flows over mountainous terrain, weather processes, and other aspects of atmospheric variability in more detail.

In a recent paper published in Journal of Climate, we address two questions. Does increasing a model’s atmospheric resolution improve the fidelity of simulated European winter hydroclimate? How do high-resolution future projections differ, if at all, from those at low-resolution? We compared simulations with low- and high-resolution versions of the same climate model—Met Office’s Hadley Centre Global Environmental Model (version 3) Global Atmosphere 3.0 (hereafter HadGEM3-GA3; Walters et al. 2011)—to establish exactly what the impact of resolution is on the North Atlantic jet and on downstream storm activity and precipitation. At the lowest resolution (‘N96’), the latitude-longitude grid is made up of grid cells each 135 km. At mid- (‘N216’) and high-resolution (‘N512’), grid cells are each 60 and 25 km, respectively. We’ll focus here on how the North Atlantic jet behaves at different model resolutions.

The wintertime low-level jet is not fixed in place, but instead varies in latitude. It tends to occupy one of three positions: southern (~35–40 °N), central (~42–58 °N) and northern (~53–60 °N). The jet’s position is important for downstream weather over Europe because it steers mid-latitude storms. Many previous-generation climate models with relatively low resolutions of 100–200 km did not reproduce this so-called trimodal behaviour, but HadGEM3-GA3 more or less does, and does so more accurately at 25-km resolution. Interestingly, the high-resolution simulations show a double peak in the northern regime that is also apparent in the latest reanalysis, ERA5, which is of a comparable resolution (Figure 2, upper panel). High resolution reveals the impact of Greenland’s topography on the westerly flow, which is not captured by lower-resolution reanalyses (ERA-Interim and NCEP-CFSR) and model simulations.

jet_upscale
Figure 2 | Frequency of North Atlantic eddy-driven jet latitude in reanalyses and HadGEM3-GA3 under historical climate (upper panel) and the projected future change (lower panel). We use ‘N’ notation to describe resolutions: ‘N96’ (135 km), ‘N216’ (60 km) and ‘N512’ (25 km). Figure adapted from Baker et al. (2019).

What about future projections? Under climate change (RCP 8.5), at all model resolutions, southern jet occurrences decrease but northern jet occurrences increase (Figure 2, lower panel). The upshot of this is fewer storms making landfall over southern Europe and more across northern Europe towards the end of the twenty-first century. These consequences of climate change are significantly enhanced by increased resolution. Crucially, this reveals the extent to which lower-resolution models may have previously underestimated aspects of the jet’s response to climate change—and thereby changes in winter storms and their associated hazards. There is much more work to do: further studies investigating other models and climate change scenarios are needed, but our study offers insight into how high resolution might bring the picture of Europe’s future winters into sharper focus. ♦

Author’s note

This post originally appeared on Weather and Climate @ Reading and this research was supported by the European Commission-funded PRIMAVERA project.

References

Baker, A. J. et al., 2019. Enhanced climate change response of wintertime North Atlantic circulation, cyclonic activity and precipitation in a 25-km-resolution global atmospheric model. Journal of Climate 32, 7763–7781.

Huntingford, C. et al., 2014. Potential influences on the United Kingdom’s floods of winter 2013/14. Nature Climate Change 4, 769–777.

Walters, D. N. et al., 2011. The Met Office Unified Model Global Atmosphere 3.0/3.1 and JULES Global Land 3.0/3.1 configurations. Geoscientific Model Development 4, 919–941.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s