| 英文摘要: | There has been an ostensibly large number of extreme weather events in the Northern Hemisphere mid-latitudes during the past decade1. An open question that is critically important for scientists and policy makers is whether any such increase in weather extremes is natural or anthropogenic in origin2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13. One mechanism proposed to explain the increased frequency of extreme weather events is the amplification of mid-latitude atmospheric planetary waves14, 15, 16, 17. Disproportionately large warming in the northern polar regions compared with mid-latitudes—and associated weakening of the north–south temperature gradient—may favour larger amplitude planetary waves14, 15, 16, 17, although observational evidence for this remains inconclusive18, 19, 20, 21. A better understanding of the role of planetary waves in causing mid-latitude weather extremes is essential for assessing the potential environmental and socio-economic impacts of future planetary wave changes. Here we show that months of extreme weather over mid-latitudes are commonly accompanied by significantly amplified quasi-stationary mid-tropospheric planetary waves. Conversely, months of near-average weather over mid-latitudes are often accompanied by significantly attenuated waves. Depending on geographical region, certain types of extreme weather (for example, hot, cold, wet, dry) are more strongly related to wave amplitude changes than others. The findings suggest that amplification of quasi-stationary waves preferentially increases the probabilities of heat waves in western North America and central Asia, cold outbreaks in eastern North America, droughts in central North America, Europe and central Asia, and wet spells in western Asia. A series of weather extremes have hit the Northern Hemisphere mid-latitudes in recent years1, such as the European heat wave in summer 20038, cold and snowy winters in 2009/10, 2010/11 and 2013/14 in the northeast United States6, 16, 21, the Russian heat wave in summer 20102, 3, 4, 5, the Texas drought of 20116, and the summer 2012 and winter 2013/14 floods in the United Kingdom7, 10; all have had significant socio-economic impacts. There is increasing scientific evidence1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13 and a growing public perception22 that extreme weather events are occurring more frequently. However, the mechanisms that drive weather extremes and through which climate change may influence climate variability are poorly understood. A potential cause of increased weather extremes is the amplification of atmospheric planetary waves14, 15, 16, 17. It has been proposed that the weakening north–south temperature gradient—a key characteristic of anthropogenic climate change23, 24—may cause larger amplitude planetary waves14, 15, 16, 17, although the observational evidence for this and the dynamical mechanism have been questioned18, 19, 20, 21. It is further proposed that high-amplitude planetary waves favour the occurrence of extreme weather14, 15, 16, 17. It is this hypothesis that we examine here (and not wave amplitude trends). First, it is necessary to define precisely ‘extreme weather’ for this application. We are concerned with persistent anomalies in land surface temperature (TL) and land precipitation (PL), such as heat waves, cold spells, droughts and prolonged wet periods, which are evident on monthly timescales and large spatial scales (Methods). Initially we focus on absolute (that is, irrespective of their sign) TL and PL anomalies (denoted |TL′| and |PL′|). This is appropriate because planetary waves tend to induce positive temperature (and perhaps precipitation) anomalies at some longitudes and negative anomalies at other longitudes. Figure 1a, b shows normalized time series for monthly |TL′| and |PL′|, respectively, area-averaged over northern mid-latitudes (35°–60° N; area shown in Fig. 2). The 40 months with largest values (approx. 10% of cases) are highlighted by circles and labelled on the lower x axis, and are hereafter referred to as months of extreme temperature/precipitation. The months of extreme temperature and the months of extreme precipitation lie relatively evenly through the 34-year period, and there is no long-term trend. A full discussion of 34-year trends is provided in Supplementary Discussion 1.
|
