Climate variability on Fit for 55 European power systems

variations of wind and solar generation (Fig 3). We observe how the correlation in the north- ern regions (mostly wind-driven) shows a cluster, albeit less evident during summer. The southern regions (Iberia, Italy and SEE) are weakly correlated with each other and show a neg- ative correlation with the northern regions. The strongest negative correlation is between Ibe- ria and Northern, -0.49 (CI 95% [-0.84, -0.13]) during winter, and between CEE and Iberia, -0.45 (CI 95% [-0.78, -0.12]) throughout the year. This first step offers a glimpse of the spatial relationships among the European regions regarding the characteristics of their power systems. As a next step, we explore in more detail the relationship between the other aspects of the simulated power systems (Fig 4). Solar and wind generation show correlation across multiple zones, suggesting that the meteorological drivers substantially affecting solar and wind are large-scale. This is consistent with Fig 3 as well as with other results found in scientific literature [47, 48]. For wind genera- tion, we can observe negative correlations between the two clusters: for example, Iberian and Northern show a correlation of -0.53 (CI 95% [-0.83, -0.22]), Italy and Northern -0.50 (CI 95% [-0.84, -0.16]) and Italy and UK & Ireland -0.47 (CI 95% [-0.77, -0.16]). We further observe that solar power also shows a positive correlation among multiple regions but it is, in general, weaker than wind. In terms of other indicators, emission intensity is positively linked in many regions, in par-

ticular in UK & Ireland and CWE (rs = 0.87, CI 95% [0.78, 0.95]) and CWE with CEE (rs = 0.90, CI 95% [0.83, 0.98]). Indeed, with emission intensity directly connected to the availability of low carbon/renewables sources, the availability of renewables is correlated across most of the regions. Situations with higher/lower RES generation thus lead to lower/higher emissions and can span multiple regions, if not the entire continent. Finally, the panel on net export (i.e., the difference between export and import at annual level) shows two things: a) Some regions (often neighbours) can simultaneously experience an increase/decrease of electricity exports, meaning that their supply tends to vary year-over-year with a similar pattern b) The negative correlation can be seen as a natural implication of electricity trading, but it may also suggest that when a region tends to have a deficit of low-cost electricity another one has a surplus and vice versa. Fig 3. Year-over-year Spearman correlation of wind and solar generation for the entire year or for summer/winter. The colours and size of the edges between the nodes is based on the value of the Spearman correlation coefficient. We show only values above 0.4 and below -0.4. https://doi.org/10.1371/journal.pone.0289705.g003 PLOS ONE Climate variability on Fit for 55 European power systems PLOS ONE | https://doi.org/10.1371/journal.pone.0289705 December 20, 2023 7 / 16