Assessment of simulations of a polar low with the Canadian Regional Climate Model

The simulations 24d12h and 24d18h have a small negative ME during the mature and dissipa- tion stages of the PL, whereas the other simulations show a positive ME during the whole lifetime of the PL (Fig 8A). The ME of the latter steadily increases from around 1100 UTC until around 1600 UTC, which is likely due not only to the deepening of the PL, but also to the fact that its cen- tre is getting closer to the surface stations. The time series of the RMSE of these simulations shows a similar pattern to that of the ME (Fig 8B). The main difference is that 24d12h and 24d18h also show an increase in the RMSE during the deepening of the PL, although it remains much lower than that of the other simulations. This decrease in accuracy, together with the nega- tive mean bias, confirms the previous finding that the simulated PL in both simulations is deeper than the observed one. All simulations except for 24d12h and 24d18h show a significant decrease in the Spearman correlation coefficient from 1300 UTC until 1500 UTC, when it reaches a mini- mum, which corresponds to the time when the SLP minimum is observed (Fig 8C). Then, the correlation coefficient increases until the PL has dissipated. The simulation 24d12h shows a simi- lar pattern, but the decrease in the correlation coefficient is much less pronounced. In contrast, 24d18h shows a small decrease in the correlation coefficient. ERA5 shows better skill at representing SLP than the simulations, even than 24d18h. Although the absolute ME of ERA5 and 24d18h are both small (Table 2, Fig 8A), the

Fig 7. Scatterplots displaying, for each simulation, the relationship between the simulated SLP and the SLP measured at surface stations and drifting buoys. The scatterplot on the bottom displays the relationship between the ERA5 SLP and the observed SLP. The red line represents the regression line that would correspond to a perfect match between the values. https://doi.org/10.1371/journal.pone.0292250.g007 PLOS ONE Assessment of simulations of a polar low with the canadian regional climate model PLOS ONE | https://doi.org/10.1371/journal.pone.0292250 October 5, 2023 12 / 28