The atmospheric circulation in the mid-latitudes of both hemispheres is usually dominated by westerly winds and by planetary-scale and shorter-scale synoptic waves, moving mostly from west to east. A remarkable and frequent exception to this “usual” behavior is atmospheric blocking. Blocking occurs when the usual zonal flow is hindered by the establishment of a large-amplitude, quasi-stationary, high-pressure meridional circulation structure which “blocks” the flow of the westerlies and the progression of the atmospheric waves and disturbances embedded in them. Such blocking structures can have lifetimes varying from a few days to several weeks in the most extreme cases. Their presence can strongly affect the weather of large portions of the mid-latitudes, leading to the establishment of anomalous meteorological conditions. These can take the form of strong precipitation episodes or persistent anticyclonic regimes, leading in turn to floods, extreme cold spells, heat waves, or short-lived droughts. Even air quality can be strongly influenced by the establishment of atmospheric blocking, with episodes of high concentrations of low-level ozone in summer and of particulate matter and other air pollutants in winter, particularly in highly populated urban areas. Atmospheric blocking has the tendency to occur more often in winter and in certain longitudinal quadrants, notably the Euro-Atlantic and the Pacific sectors of the Northern Hemisphere. In the Southern Hemisphere, blocking episodes are generally less frequent, and the longitudinal localization is less pronounced than in the Northern Hemisphere. Blocking has aroused the interest of atmospheric scientists since the middle of the last century, with the pioneering observational works of Berggren, Bolin, Rossby, and Rex, and has become the subject of innumerable observational and theoretical studies. The purpose of such studies was originally to find a commonly accepted structural and phenomenological definition of atmospheric blocking. The investigations went on to study blocking climatology in terms of the geographical distribution of its frequency of occurrence and the associated seasonal and inter-annual variability. Well into the second half of the 20th century, a large number of theoretical dynamic works on blocking formation and maintenance started appearing in the literature. Such theoretical studies explored a wide range of possible dynamic mechanisms, including large-amplitude planetary-scale wave dynamics, including Rossby wave breaking, multiple equilibria circulation regimes, large-scale forcing of anticyclones by synoptic-scale eddies, finite-amplitude non-linear instability theory, and influence of sea surface temperature anomalies, to name but a few. However, to date no unique theoretical model of atmospheric blocking has been formulated that can account for all of its observational characteristics. When numerical, global short- and medium-range weather predictions started being produced operationally, and with the establishment, in the late 1970s and early 1980s, of the European Centre for Medium-Range Weather Forecasts, it quickly became of relevance to assess the capability of numerical models to predict blocking with the correct space-time characteristics (e.g., location, time of onset, life span, and decay). Early studies showed that models had difficulties in correctly representing blocking as well as in connection with their large systematic (mean) errors. Despite enormous improvements in the ability of numerical models to represent atmospheric dynamics, blocking remains a challenge for global weather prediction and climate simulation models. Such modeling deficiencies have negative consequences not only for our ability to represent the observed climate but also for the possibility of producing high-quality seasonal-to-decadal predictions. For such predictions, representing the correct space-time statistics of blocking occurrence is, especially for certain geographical areas, extremely important.
Stefano Tibaldi and Franco Molteni
Since the dawn of the digital computing age in the mid-20th century, computers have been used as virtual laboratories for the study of atmospheric phenomena. The first simulations of thunderstorms captured only their gross features, yet required the most advanced computing hardware of the time. The following decades saw exponential growth in computational power that was, and continues to be, exploited by scientists seeking to answer fundamental questions about the internal workings of thunderstorms, the most devastating of which cause substantial loss of life and property throughout the world every year. By the mid-1970s, the most powerful computers available to scientists contained, for the first time, enough memory and computing power to represent the atmosphere containing a thunderstorm in three dimensions. Prior to this time, thunderstorms were represented primarily in two dimensions, which implicitly assumed an infinitely long cloud in the missing dimension. These earliest state-of-the-art, fully three-dimensional simulations revealed fundamental properties of thunderstorms, such as the structure of updrafts and downdrafts and the evolution of precipitation, while still only roughly approximating the flow of an actual storm due computing limitations. In the decades that followed these pioneering three-dimensional thunderstorm simulations, new modeling approaches were developed that included more accurate ways of representing winds, temperature, pressure, friction, and the complex microphysical processes involving solid, liquid, and gaseous forms of water within the storm. Further, these models also were able to be run at a resolution higher than that of previous studies due to the steady growth of available computational resources described by Moore’s law, which observed that computing power doubled roughly every two years. The resolution of thunderstorm models was able to be increased to the point where features on the order of a couple hundred meters could be resolved, allowing small but intense features such as downbursts and tornadoes to be simulated within the parent thunderstorm. As model resolution increased further, so did the amount of data produced by the models, which presented a significant challenge to scientists trying to compare their simulated thunderstorms to observed thunderstorms. Visualization and analysis software was developed and refined in tandem with improved modeling and computing hardware, allowing the simulated data to be brought to life and allowing direct comparison to observed storms. In 2019, the highest resolution simulations of violent thunderstorms are able to capture processes such as tornado formation and evolution which are found to include the aggregation of many small, weak vortices with diameters of dozens of meters, features which simply cannot not be simulated at lower resolution.