Figure 2.38 Trends in surface wind speed for 1988–2010. Shown in the top row are data sets based on the satellite wind observations: (a) Cross-Calibrated Multi-Platform wind product (CCMP; Atlas et al., 2011); (b) wind speed from the Objectively Analyzed Air-Sea Heat Fluxes data set, release 3 (OAFlux); (c) Blended Sea Winds (BSW; Zhang et al., 2006); in the middle row are data sets based on surface observations: (d) ERA-Interim; (e) NCEP-NCAR, v.1 (NNR); (f) 20th Century Reanalysis (20CR, Compo et al., 2011), and in the bottom row are surface wind speeds from atmospheric reanalyses: (g) wind speed from the Surface Flux Data set, v.2, from NOC, Southampton, UK (Berry and Kent, 2009); (h) Wave- and Anemometer-based Sea Surface Wind (WASWind; Tokinaga and Xie, 2011a)); and (i) Surface Winds on the Land (Vautard et al., 2010). Wind speeds correspond to 10 m heights in all products. Land station winds (panel f) are also for 10 m (but anemometer height is not always reported) except for the Australian data where they correspond to 2 m height. To improve readability of plots, all data sets (including land station data) were averaged to the 4° × 4° uniform longitude-latitude grid. Trends were computed for the annually averaged timeseries of 4° × 4° cells. For all data sets except land station data, an annual mean was considered available only if monthly means for no less than eight months were available in that calendar year. Trend values were computed only if no less than 17 years had values and at least 1 year was available among the first and last 3 years of the period. White areas indicate incomplete or missing data. Black plus signs (+) indicate grid boxes where trends are significant (i.e., a trend of zero lies outside the 90% confidence interval).
In AR4, large interannual variability of the Hadley and Walker circulation was highlighted, as well as the difficulty in addressing changes in these features in the light of discrepancies between data sets. AR4 also found that rainfall in many monsoon systems exhibits decadal changes, but that data uncertainties restrict confidence in trends. SREX also attributed low confidence to observed trends in monsoons. Observational evidence for trends and variability in the strength of the Hadley and Walker circulations (Annex III: Glossary), the monsoons, and the width of the tropical belt is based on radiosonde and reanalyses data (Box 2.3). In addition, changes in the tropical circulation imprint on other fields that are observed from space (e.g., total ozone, outgoing longwave radiation). Changes in the average state of the tropical circulation are constrained to some extent by changes in the water cycle (Held and Soden, 2006; Schneider et al., 2010). Changes in the
monsoon systems are expressed through altered circulation, moisture transport and convergence, and precipitation. Only a few monsoon studies address circulation changes, while most work focuses on precipitation.
Several studies report a weakening of the global monsoon circulations as well as a decrease of global land monsoon rainfall or of the number of precipitation days over the past 40 to 50 years (Zhou et al., 2008, see also SREX; Liu et al., 2011). Concerning the East Asian Monsoon, a year-round decrease is reported for wind speeds over China at the surface and in the lower troposphere based on surface observations and radiosonde data (Guo et al., 2010; Jiang et al., 2010; Vautard et al., 2010; Xu et al., 2010). The changes in wind speed are concomitant with changes in pressure centres such as a westward extension of the Western Pacific Subtropical High (Gong and Ho, 2002; Zhou et al., 2009b). A weakening of the East Asian summer monsoon since the 1920s is also found in SLP gradients (Zhou et al., 2009a). However, trends derived from wind observations and circulation trends from reanalysis data carry large uncertainties (Figure 2.38), and monsoon rainfall trends depend, for example, on the definition of the monsoon area (Hsu et al., 2011). For instance, using a new definition of monsoon area, an increase in northern hemispheric and global summer monsoon (land and ocean) precipitation is reported from 1979 to 2008 (Hsu et al., 2011; Wang et al., 2012a).
Figure 2.39 (a) Indices of the strength of the northern Hadley circulation in December to March (Ψmax is the maximum of the meridional mass stream function at 500 hPa between the equator and 40°N). (b) Indices of the strength of the Pacific Walker circulation in September to January (Δω is the difference in the vertical velocity between [10°S to 10°N, 180°W to 100°W] and [10°S to 10°N, 100°E to 150°E] as in Oort and Yienger (1996), Δc is the difference in cloud cover between [6°N to 12°S, 165°E to 149°W] and [18°N to 6°N, 165°E to 149°W] as in Deser et al. (2010a), vE is the effective wind index from SSM/I satellite data, updated from Sohn and Park (2010), u is the zonal wind at 10 m averaged in the region [10°S to 10°N, 160°E to 160°W], ΔSLP is the SLP difference between [5°S to 5°N, 160°W to 80°W] and [5°S to 5°N, 80°E to 160°E] as in Vecchi et al. (2006)). Reanalysis data sets include 20CR, NCEP/NCAR, ERA-Interim, JRA-25, MERRA, and CFSR, except for the zonal wind at 10 m (20CR, NCEP/NCAR, ERA-Interim), where available until January 2013. ERA-40 and NCEP2 are not shown as they are outliers with respect to the strength trend of the northern Hadley circulation (Mitas and Clement, 2005; Song and Zhang, 2007; Hu et al., 2011; Stachnik and Schumacher, 2011). Observation data sets include HadSLP2 (Section 2.7.1), ICOADS (Section 2.7.2; only 1957–2009 data are shown) and WASWIND (Section 2.7.2), reconstructions are from Brönnimann et al. (2009). Where more than one time series was available, anomalies from the 1980/1981 to 2009/2010 mean values of each series are shown.
The additional data sets that became available since AR4 confirm the large interannual variability of the Hadley and Walker circulation. The strength of the northern Hadley circulation (Figure 2.39) in boreal winter and of the Pacific Walker circulation in boreal fall and winter is largely related to the ENSO (Box 2.5). This association dominates interannual variability and affects trends. Data sets do not agree well with respect to trends in the Hadley circulation (Figure 2.39). Two widely used reanalysis data sets, NNR and ERA-40, both have demonstrated shortcomings with respect to tropical circulation; hence their increases in the Hadley circulation strength since the 1970s might be artificial (Mitas and Clement, 2005; Song and Zhang, 2007; Hu et al., 2011; Stachnik and Schumacher, 2011). Later generation reanalysis data sets including ERA-Interim (Brönnimann et al., 2009; Nguyen et al., 2013) as well as satellite humidity data (Sohn and Park, 2010) also suggest a strengthening from the mid 1970s to present, but the magnitude is strongly data set dependent.
Figure 2.22 Trends in surface temperature from NCDC MLOST for three nonconsectutive shorter periods (1911–1940; 1951–1980; 1981–2012). White areas indicate incomplete or missing data. Trends and significance have been calculated as in Figure 2.21.
Box 2.5, Figure 1 Some indices of climate variability, as defined in Box 2.5, Table 1, plotted in the 1870–2012 interval. Where ‘HadISST1’, ‘HadSLP2r’, or ‘NNR’ are indicated, the indices were computed from the sea surface temperature (SST) or sea level pressure (SLP) values of the former two data sets or from 500 or 850 hPa geopotential height fields from the NNR. Data set references given in the panel titles apply to all indices shown in that panel. Where no data set is specified, a publicly available version of an index from the authors of a primary reference given in Box 2.5, Table 1 was used. All indices were standardized with regard to 1971–2000 period except for NIÑO3.4 (centralized for 1971–2000) and AMO indices (centralized for 1901–1970). Indices marked as “detrended” had their linear trend for 1870–2012 removed. All indices are shown as 12-month running means except when the temporal resolution is explicitly indicated (e.g., ‘DJFM’ for December-to-March averages) or smoothing level (e.g., 11-year LPF for a low-pass filter with half-power at 11 years).
Consistent changes in different observed variables suggest a weakening of the Pacific Walker circulation during much of the 20th century that has been largely offset by a recent strengthening. A weakening is indicated by trends in the zonal SLP gradient across the equatorial Pacific (Section 2.7.1, Table 2.14) from 1861 to 1992 (Vecchi et al., 2006), or from 1901 to 2004 (Power and Kociuba, 2011b). Boreal spring and summer contribute most strongly to the centennial trend (Nicholls, 2008; Karnauskas et al., 2009), as well as to the trend in the second half of the 20th century (Tokinaga et al., 2012). For boreal fall and winter, when the circulation is strongest, no trend is found in the Pacific Walker circulation based on the vertical velocity at 500 hPa from reanalyses (Compo et al., 2011), equatorial Pacific 10 m zonal winds, or SLP in Darwin (Nicholls, 2008; Figure 2.39). However, there are inconsistencies between ERA-40 and NNR (Chen et al., 2008). Deser et al. (2010a) find changes in marine air temperature and cloud cover over the Pacific that are consistent with a weakening of the Walker circulation during most of the 20th century (Section 2.5.7.1 and Yu and Zwiers, 2010). Tokinaga et al. (2012) find robust evidence for a weakening of the Walker circulation (most notably over the Indian Ocean) from 1950 to 2008 based on observations of cloud cover, surface wind, and SLP. Since the 1980s or 1990s, however, trends in the Pacific Walker circulation have reversed (Figure 2.39; Luo et al., 2012). This is evident from changes in SLP (see equatorial Southern Oscillation Index (SOI) trends in Table 2.14 and Box 2.5, Figure 1), vertical velocity (Compo et al., 2011), water vapour flux from satellite and reanalysis data (Sohn and Park, 2010), or sea level height (Merrifield, 2011). It is also consistent with the SST trend pattern since 1979 (Meng et al., 2012; see also Figure 2.22).
Observed changes in several atmospheric parameters suggest that the width of the tropical belt has increased at least since 1979 (Seidel et al., 2008; Forster et al., 2011; Hu et al., 2011). Since AR4, wind, temperature, radiation, and ozone information from radiosondes, satellites, and reanalyses had been used to diagnose the tropical belt width and estimate their trends. Annual mean time series of the tropical belt width from various sources are shown in Figure 2.40.
Since 1979 the region of low column ozone values typical of the tropics has expanded in the NH (Hudson et al., 2006; Hudson, 2012). Based on radiosonde observations and reanalyses, the region of the high tropical tropopause has expanded since 1979, and possibly since 1960 (Seidel and Randel, 2007; Birner, 2010; Lucas et al., 2012), although widening estimates from different reanalyses and using different methodologies show a range of magnitudes (Seidel and Randel, 2007; Birner, 2010). Several lines of evidence indicate that climate features at the edges of the Hadley cell have also moved poleward since 1979. Subtropical jet metrics from reanalysis zonal winds (Strong and Davis, 2007, 2008; Archer and Caldeira, 2008b, 2008a) and layer-average satellite temperatures (Fu et al., 2006; Fu and Lin, 2011) also indicate widening, although 1979–2009 wind-based trends (Davis and Rosenlof, 2011) are not statistically significant. Changes in subtropical outgoing longwave radiation, a surrogate for high cloud, also suggest widening (Hu and Fu, 2007), but the methodology and results are disputed (Davis and Rosenlof, 2011). Widening of the tropical belt is also found in precipitation patterns (Hu and Fu, 2007; Davis and Rosenlof, 2011; Hu et al., 2011; Kang et al., 2011; Zhou et al., 2011), including in SH regions (Cai et al., 2012).
The qualitative consistency of these observed changes in independent data sets suggests a widening of the tropical belt between at least 1979 and 2005 (Seidel et al., 2008), and possibly longer. Widening estimates range between around 0° and 3° latitude per decade, but their uncertainties have been only partially explored (Birner, 2010; Davis and Rosenlof, 2011).
In summary, large interannual-to-decadal variability is found in the strength of the Hadley and Walker circulation. The confidence in trends in the strength of the Hadley circulation is low due to uncertainties in reanalysis data sets. Recent strengthening of the Pacific Walker circulation has largely offset the weakening trend from the 19th century to the 1990s (high confidence). Several lines of independent evidence indicate a widening of the tropical belt since the 1970s. The suggested weakening of the East Asian monsoon has low confidence, given the nature and quality of the evidence.