IPCC Report.pdf - Adam Curry

Changes in Climate Extremes and their Impacts on the Natural Physical EnvironmentChapter 3The AR4 (Trenberth et al., 2007, based on Alexander et al., 2006) reporteda statistically significant increase in the numbers of warm nights and astatistically significant reduction in the numbers of cold nights for 70 to75% of the land regions with data (for the spatial coverage of theunderlying data set and the definition of warm/cold days and nights, seeSection 3.2.1 and Box 3-1, respectively). Changes in the numbers ofwarm days and cold days also showed warming, but less markedthan for nights, with about 40 to 50% of the area with data showingstatistically significant changes consistent with warming (Alexander etal., 2006). Less than 1% of the area with data showed statisticallysignificant trends in cold/warm days and nights that were consistent withcooling (Alexander et al., 2006). Trenberth et al. (2007) also reported,based on Vose et al. (2005), that from 1950 to 2004, the annual trendsin minimum and maximum land-surface air temperature averaged overregions with data were 0.20°C per decade and 0.14°C per decade,respectively, and that for 1979 to 2004, the corresponding linear trendsfor the land areas with data were 0.29°C per decade for both maximumand minimum temperature. Based on this evidence, the IPCC AR4 (SPM;IPCC, 2007b) assessed that it was very likely that there had been trendstoward warmer and more frequent warm days and warm nights, andwarmer and less frequent cold days and cold nights in most land areas.Regions that were found to depart from this overall behavior towardmore warm days and nights and fewer cold days and nights inAlexander et al. (2006) were mostly central North America, the easternUnited States, southern Greenland (increase in cold days and decreasesin warm days), and the southern half of South America (decrease inwarm days; no data available for the northern half of the continent). Incentral North America and the eastern United States this partial tendencyfor a negative trend in extremes is also consistent with a reported meannegative trend in temperatures, mostly in the spring to summer season(also termed ‘warming hole’, e.g., Pan et al., 2004; Portmann et al.,2009). Several explanations have been suggested for this behavior,which seems partly associated with a change in the hydrological cycle,possibly linked to soil moisture and/or aerosol feedbacks (Pan et al.,2004; Portmann et al., 2009).More recent analyses available since the AR4 include a global study (forannual extremes) by Brown et al. (2008) based on the data set fromCaesar et al. (2006), and regional studies for North America (Peterson etal., 2008a; Meehl et al., 2009c), Central-Western Europe (since 1880;Della-Marta et al., 2007a), central and eastern Europe (Bartholy andPongracz, 2007; Kürbis et al., 2009), the eastern Mediterranean regionincluding Turkey (Kuglitsch et al., 2010), western Central Africa, GuineaConakry and Zimbabwe (Aguilar et al., 2009), the Tibetan Plateau (Youet al., 2008) and China (You et al., 2011), Uruguay (Rusticucci andRenom, 2008), and Australia (Alexander and Arblaster, 2009). Furtherreferences can also be found in Table 3-2. Overall, these studies areconsistent with the assessment of an increase in warm days and nightsand a reduction in cold days and nights on the global basis, althoughthey do not necessarily consider trends in all four variables, and a fewsingle studies report trends that are not statistically significant or eventrends opposite to the global tendencies in some extremes, subregions,seasons, or decades. For instance, Rusticucci and Renom (2008) foundin Uruguay a reduction of cold nights, a positive but a statisticallyinsignificant trend in warm nights, statistically insignificant decreases incold days at most investigated stations, and inconsistent trends inwarm days. Together with the previous results from Alexander et al.(2006) for southern South America (see above) and further regionalstudies (Table 3-2), this suggests a less consistent warming tendency inSouth America compared to other continents. Another notable feature isthat studies for central and southeastern Europe display a markedchange point in trends in temperature extremes at the end of the1970s/beginning of 1980s (Table 3-2), which for some extremes canlead to very small and/or statistically not significant overall trends sincethe 1960s (e.g., Bartholy and Pongracz, 2007).There are fewer studies available investigating changes in characteristicsof cold spells and warm spells, or cold waves and heat waves, comparedwith studies of the intensity or frequency of warm and cold days ornights. Alexander et al. (2006) provided an analysis of trends in warmspells [based on the Warm Spell Duration Index (WSDI); see Table 3-2and Box 3-1] mostly in the mid- and high-latitudes of the NorthernHemisphere. The analysis displays a tendency toward a higher length ornumber of warm spells (increase in number of days belonging to warmspells) in much of the region, with the exception of the southeasternUnited States and eastern Canada. Regional studies on trends in warmspells or heat waves are also listed in Table 3-2. Kunkel et al. (2008)found that the United States has experienced a general decline in coldwaves over the 20th century, with a spike of more cold waves in the1980s. Further, they report a strong increase in heat waves since 1960,although the heat waves of the 1930s associated with extreme droughtconditions still dominate the 1895-2005 time series. Kuglitsch et al.(2009) reported an increase in heat wave intensity, number, and lengthin summer over the 1960-2006 time period in the eastern Mediterraneanregion. Ding et al. (2010) reported increasing numbers of heat wavesover most of China for the 1961-2007 period. The record-breaking heatwave over western and central Europe in the summer of 2003 is anexample of an exceptional recent extreme (Beniston, 2004; Schär andJendritzky, 2004). That summer (June to August) was the hottest sincecomparable instrumental records began around 1780 and perhaps thehottest since at least 1500 (Luterbacher et al., 2004). Other examples ofrecent extreme heat waves include the 2006 heat wave in Europe(Rebetez et al., 2008), the 2007 heat wave in southeastern Europe(Founda and Giannakopoulos, 2009), the 2009 heat wave in southeasternAustralia (National Climate Centre, 2009), and the 2010 heat wave inRussia (Barriopedro et al., 2011). Both the 2003 European heat wave(Andersen et al., 2005; Ciais et al., 2005) and the 2009 southeasternAustralian heat wave were also associated with drought conditions,which can strongly enhance temperature extremes during heat waves insome regions (see also Section 3.1.4).Some recent analyses have led to revisions of previously reportedtrends. For instance, Della-Marta et al. (2007a) found that mean summermaximum temperature change over Europe was +1.6 ± 0.4°C during1880 to 2005, a somewhat greater increase than reported in earlier134