Variations in the behaviour of the North African summer monsoon were highlighted in IPCC (1990). Moron (1997) demonstrated that long-term variations of Sahel annual rainfall, particularly the wet 1950s and the dry 1970 to 1980s, are seen over the Guinea coast area, although trends are strongest in the Sahel. The significant decrease in Guinea coast rainfall (Ward, 1998) is present in both the first and second rainy seasons, but is strongest in the second. Janicot et al. (1996) and Moron (1997) demonstrated that the moderate influence of ENSO (towards drier conditions) has increased since 1960, with warm events associated more strongly with large-scale anomalous dry conditions over the Guinea and Sahel belts. Ward et al. (1999) show that the Sahel has become moderately wetter since 1987, despite the increased drying influence of ENSO events, a trend that continued to 1999 (Parker and Horton, 2000). This recent behaviour may be related to a quasi-hemispheric variation of SST (e.g., Enfield and Mestas-Nuñez, 1999) shown to be related to Sahel rainfall by Folland et al. (1986), and which may be related to the recent strong increase in North Atlantic SST mentioned in Section 2.2.2.2. Many other parts of tropical Africa are influenced by ENSO towards either drier or wetter conditions than normal, sometimes modulated by regional SST anomalies near Africa (e.g., Nicholson and Kim, 1997; Nicholson, 1997; Indeje et al., 2000), but few trends can be discerned.

Multi-decadal and decadal variations of the Indian monsoon have been widely noted (e.g., Pant and Kumar, 1997) but links with El Niño do not now seem straightforward (Slingo et al., 1999). However, despite the recent strong El Niño episodes, the inverse relationship between the ENSO and the Indian summer monsoon (weak monsoon arising from an ENSO event) has broken down in the recent two decades (Kumar et al., 1999a). This link operated on multi-decadal time-scales

with NINO 3 SST until at least 1970. Kumar et al. suggest that persistently increased surface temperatures over Eurasia in winter and spring (Figure 2.10) have favoured an enhanced land-ocean thermal gradient conducive to stronger monsoons; they also observe a shift away from India in the sinking node of the Walker circulation in El Niño. Changes have also occurred in relationships with Indian monsoon precursors (Kumar et al., 1999b). One possibility is that warming over the Indian Ocean (Figures 2.9, 2.10) may have increased moisture and rainfall for a given state of the atmospheric circulation (Kitoh et al., 1997). There may be a link to multi-decadal variations in Pacific SST, but this remains to be investigated, together with other monsoon indices (e.g., Goswami et al., 1997).

It has been known for some time that the position of the western North Pacific sub-tropical high affects the East Asian monsoon. Gong and Wang (1999a) showed that summer (June to August) precipitation over central and eastern China near 30°N is positively correlated with the intensity of the high, with negative correlations to the north and south. A location of the sub-tropical high further south than normal is conducive to heavy summer rainfall in this region. Time-series of the sub-tropical high show an increase in areal extent in the 1920s, then another increase from the mid-1970s to 1998, giving frequent wet summers in this region in recent years. The north-east winter monsoon has also shown low-frequency variations. Thus the strength of the Siberian high increased to a peak around 1968, and then weakened to a minimum around 1990 (Gong and Wang, 1999b), in phase with the increased frequency of the positive phase of the NAO (Wallace, 2000 and next section). This is likely to have contributed to strong recent winter warming in China shown in Figure 2.10.

IPCC Homepage