Atmospheric blocking is a recurrent weather pattern characterized by a quasi-stationary persistent large-scale high-pressure system which ‘blocks’ the westerly flow, typically at the exit of the storm tracks. Blocking has elicited the attention of scientists since the early 1950s, due to its inherent dynamical complexity and its impacts on mid-latitude weather, which may lead to cold spells in winter and to heatwaves in summer.
In addition to being a weather phenomenon only partially understood by the scientific community, blocking has always been a difficult problem in numerical simulations, from both the weather and the climate point of view. However, blocking has been rarely investigated at seasonal timescale. A new study has reviewed the representation of blocking in ECMWF seasonal prediction systems, tested its sensitivity to key model characteristics, explored predictive skill and compared it with climate simulations of the ECMWF model.
Climatology and sensitivities
Blocking representation in ECMWF seasonal prediction systems shows considerable improvements from System 3, to System 4, and to SEAS5, as shown by the blocking frequency climatology in the figure. Reduced biases are observed over the North Pacific and Greenland, even though over Europe – a region where numerical models often struggle the most – a non-negligible underestimation of blocking frequency is still present in SEAS5.
Making use of a series of complementary hindcast experiments, it has been possible to assess that increasing atmospheric and oceanic resolution improves considerably the statistics of blocking. On the other hand, the simulated blocking frequency remains largely insensitive to coupled model sea-surface temperature (SST) errors, suggesting that atmosphere–ocean coupling might not be the key to correctly simulating blocking as long as the mean oceanic state does not show large biases. The implementation of stochastic parametrizations, which are operationally included in SEAS5, tends to slightly displace blocking activity equatorward, possibly pointing to a change in the mean climate.
Variability and prediction
SEAS5 predictive skill of blocking frequency and corresponding signal-to-noise ratios remain globally low, but interesting positive results are found at low latitudes, especially over the Pacific sector. Moreover, positive skill is also obtained over parts of western and central Europe, achieving moderate but significant values (ensemble mean correlation skill of about 0.3) over an extended region. Most importantly, the areas of significant skill are larger and more robust than in previous seasonal prediction systems, pointing not only to an improvement of the mean climate but also of the prediction capabilities.
Interestingly, SEAS5 blocking interannual variability is underestimated too, and its bias is proportional to the climatological frequency. This shows that a negative bias in the blocking frequency implies a negative bias in the interannual variance. As a result, it seems plausible to expect that eventual improvement in the blocking mean state would be reflected by improvement in its interannual variability. It might also influence positively the overall skill of mid-latitude geopotential height which, for SEAS5 over Europe, remains lower than for blocking.
Comparison with climate runs
Seasonal hindcasts have been compared against a set of climate runs of the same model configuration. It has been shown that SST errors have a larger impact on blocking bias in climate runs than in seasonal runs, and that increased ocean model resolution contributes to improved blocking more effectively in climate runs than in seasonal forecasts. Making use of the same methodology, it has been found that the largest contributors to the long-standing underestimation of blocking are persistent errors in the atmospheric model. Seasonal forecasts can thus be considered a suitable seamless test-bed for model development targeting blocking improvement in climate models.
Further information can be found in an article by Paolo Davini and co-authors, Quarterly Journal of Meteorological Society, https://doi.org/10.1002/qj.3974.