16.2.2.2 Observed trends

Temperature

New observations and reanalyses of temperatures averaged over land and ocean surfaces since the TAR show consistent warming trends in all small-island regions over the 1901 to 2004 period (Trenberth et al., 2007). However, the trends are not linear. Recent studies show that annual and seasonal ocean surface and island air temperatures have increased by 0.6 to 1.0°C since 1910 throughout a large part of the South Pacific, south-west of the SPCZ. Decadal increases of 0.3 to 0.5°C in annual temperatures have been widely seen only since the 1970s, preceded by some cooling after the 1940s, which is the beginning of the record, to the north-east of the SPCZ (Salinger, 2001; Folland et al., 2003).

For the Caribbean, Indian Ocean and Mediterranean regions, analyses shows warming ranged from 0 to 0.5°C per decade for the 1971 to 2004 period (Trenberth et al., 2007). Some high-latitude regions, including the western Canadian Arctic Archipelago, have experienced warming more rapid than the global mean (McBean et al., 2005).

Trends in extreme temperature across the South Pacific for the period 1961 to 2003 show increases in the annual number of hot days and warm nights, with decreases in the annual number of cool days and cold nights, particularly in the years after the onset of El Niño (Manton et al., 2001; Griffiths et al., 2003). In the Caribbean, the percentage of days having very warm maximum or minimum temperatures has increased considerably since the 1950s, while the percentage of days with cold temperatures has decreased (Peterson et al., 2002).

Precipitation

Analyses of trends in extreme daily rainfall across the South Pacific for the period 1961 to 2003 show extreme rainfall trends which are generally less spatially coherent than those of extreme temperatures (Manton et al., 2001; Griffiths et al., 2003). In the Caribbean, the maximum number of consecutive dry days is decreasing and the number of heavy rainfall events is increasing. These changes were found to be similar to the changes reported from global analysis (Trenberth et al., 2007).

Tropical and extra-tropical cyclones

Variations in tropical and extra-tropical cyclones, hurricanes and typhoons in many small-island regions are dominated by ENSO and decadal variability which result in a redistribution of tropical storms and their tracks, so that increases in one basin are often compensated by decreases in other basins. For example, during an El Niño event, the incidence of tropical storms typically decreases in the Atlantic and far-western Pacific and the Australian regions, but increases in the central and eastern Pacific, and vice versa. Clear evidence exists that the number of storms reaching categories 4 and 5 globally have increased since 1970, along with increases in the Power Dissipation Index (Emanuel, 2005) due to increases in their intensity and duration (Trenberth et al., 2007). The total number of cyclones and cyclone days decreased slightly in most basins. The largest increase was in the North Pacific, Indian and South-West Pacific oceans. The global view of tropical storm activity highlights the important role of ENSO in all basins. The most active year was 1997, when a very strong El Niño began, suggesting that the observed record sea surface temperatures (SSTs) played a key role (Trenberth et al., 2007). For extra-tropical cyclones, positive trends in storm frequency and intensity dominate during recent decades in most regional studies performed. Longer records for the North Atlantic suggest that the recent extreme period may be similar in level to that of the late 19th century (Trenberth et al., 2007).

In the tropical South Pacific, small islands to the east of the dateline are highly likely to receive a higher number of tropical storms during an El Niño event compared with a La Niña event and vice versa (Brazdil et al., 2002). Observed tropical cyclone activity in the South Pacific east of 160°E indicates an increase in level of activity, with the most active years associated with El Niño events, especially during the strong 1982/1983 and 1997/1998 events (Levinson, 2005). Webster et al. (2005) found more than a doubling in the number of category 4 and 5 storms in the South-West Pacific from the period 1975–1989 to the period 1990–2004. In the 2005/2006 season, La Niña influences shifted tropical storm activity away from the South Pacific region to the Australian region and, in March and April 2006, four category 5 typhoons occurred (Trenberth et al., 2007).

In the Caribbean, hurricane activity was greater from the 1930s to the 1960s, in comparison with the 1970s and 1980s and the first half of the 1990s. Beginning with 1995, all but two Atlantic hurricane seasons have been above normal (relative to the 1981-2000 baseline). The exceptions are the two El Niño years of 1997 and 2002. El Niño acts to reduce activity and La Niña acts to increase activity in the North Atlantic. The increase contrasts sharply with the generally below-normal seasons observed during the previous 25-year period, 1975 to 1994. These multi-decadal fluctuations in hurricane activity result almost entirely from differences in the number of hurricanes and major hurricanes forming from tropical storms first named in the tropical Atlantic and Caribbean Sea.

In the Indian Ocean, tropical storm activity (May to December) in the northern Indian Ocean has been near normal in recent years. For the southern Indian Ocean, the tropical cyclone season is normally active from December to April. A lack of historical record-keeping severely hinders trend analysis (Trenberth et al., 2007).

Sea level

Analyses of the longest available sea-level records, which have at least 25 years of hourly data from 27 stations installed around the Pacific basin, show the overall average mean relative sea-level rise around the whole region is +0.77 mm/yr (Mitchell et al., 2001). Rates of relative sea level have also been calculated for the SEAFRAME stations in the Pacific. Using these results and focusing only on the island stations with more than 50 years of data (only four locations), the average rate of sea-level rise (relative to the Earth’s crust) is 1.6 mm/yr (Bindoff et al., 2007). Church et al. (2004) used TOPEX/Poseidon altimeter data, combined with historical tide gauge data, to estimate monthly distributions of large-scale sea-level variability and change over the period 1950 to 2000. Church et al. (2004) observed the maximum rate of rise in the central and eastern Pacific, spreading north and south around the sub-tropical gyres of the Pacific Ocean near 90°E, mostly between 2 and 2.5 mm/yr but peaking at over 3 mm/yr. This maximum was split by a minimum rate of rise, less than 1.5 mm/yr, along the equator in the eastern Pacific, linking to the western Pacific just west of 180° (Christensen et al., 2007).

The Caribbean region experienced, on average, a mean relative sea-level rise of 1 mm/yr during the 20th century. Considerable regional variations in sea level were observed in the records; these were due to large-scale oceanographic phenomena such as El Niño coupled with volcanic and tectonic crustal motions of the Caribbean Basin rim, which affect the land levels on which the tide gauges are located. Similarly, recent variations in sea level on the west Trinidad coast indicate that sea level in the north is rising at a rate of about 1 mm/yr, while in the south the rate is about 4 mm/yr; the difference being a response to tectonic movements (Miller, 2005).

In the Indian Ocean, reconstructed sea levels based on tide gauge data and TOPEX/Poseidon altimeter records for the 1950 to 2001 period give rates of relative sea-level rise of 1.5, 1.3 and 1.5 mm/yr (with error estimates of about 0.5 mm/yr) at Port Louis, Rodrigues, and Cocos Islands, respectively (Church et al., 2006). In the equatorial band, both the Male and Gan sea-level sites in the Maldives show trends of about 4 mm/yr (Khan et al., 2002), with the range from three tidal stations over the 1990s being from 3.2 to 6.5 mm/yr (Woodworth et al., 2002). Church et al. (2006) note that the Maldives has short records and that there is high variability between sites, and their 52-year reconstruction suggests a common rate of rise of 1.0 to 1.2 mm/yr.

Some high-latitude islands are in regions of continuing postglacial isostatic uplift, including parts of the Baltic, Hudson Bay, and the Canadian Arctic Archipelago (CAA). Others along the Siberian coast and the eastern and western margins of the CAA are subsiding. Although few long tide-gauge records exist in the region, relative sea-level trends are known to range from negative (falling relative sea level) in the central CAA and Hudson Bay to rates as high as 3 mm/yr or more in the Beaufort Sea (Manson et al., 2005). Available data from the Siberian sector of the Arctic Ocean indicate that late 20th century sea-level rise was comparable to the global mean (Proshutinsky et al., 2004).