Global and regional climate in 2009


Global climate

The global average temperature in 2009 was 0.44±0.10 degC above the 1961–1990 average (Figure 1(a)), making 2009 one of the ten warmest years in the 160-year HadCRUT3 series (Brohan et al., 2006). Other estimates of global temperature produced independently by NASA's Goddard Institute for Space Studies (Hansen et al., 2001) and the National Oceanic and Atmospheric Administration (Smith et al., 2008) also rank 2009 among the ten warmest years, although the exact ranking varies due to differences in analysis technique. The HadCRUT3 analysis uses observations from a network of more than 4000 land stations, of which 1200 provide regular monthly updates. It also incorporates historic sea-surface temperature (SST) measurements made principally by ships and buoys, taken from the International Comprehensive Ocean Atmosphere Data Set (Worley et al., 2005) and regularly updated using observations reported on the Global Telecommunication System. Around 16 million SST observations were gathered in 2009. The annual average temperature anomaly for the Northern Hemisphere was 0.56±0.10 degC, the seventh highest on record. For the Southern Hemisphere the anomaly was 0.32±0.13 degC (fourth highest) (Figures 1(b) and 1(c)).

Figure 1(a)–(c).

Annual combined land-surface air and sea-surface temperature anomalies ( degC, blue bars) and uncertainty range for 1850–2009. (a) Globe, (b) Northern Hemisphere, (c) Southern Hemisphere. The red line shows the annual values after smoothing with a 21-point binomial filter and highlights interdecadal variations. The grey areas in panel (a) show the average temperature and 95% confidence range (scale offset by -0.6 degC) for each non-overlapping 10-year period starting 1850–1859. Data are an update of Brohan et al. (2006).

Figure 1(d)–(f).

As for Figure 1(a)–(c) but for the (d) Tropics 20°N–20°S; (e) Northern Hemisphere north of 20°N; (f) Southern Hemisphere south of 20°S.

Figure 1(g).

Three-month (January to March, April to June, July to September, and October to December) average sea-surface temperatures from 1953 to 2009 (highlighted in green) for four regions in the tropical Pacific: Niño 1+2 (80°–90°W, 0°–10°S), Niño 3 (90–150°W, 5°S–5°N), Niño 3.4 (170°W–120°W, 5°S–5°N), and Niño 4 (160°E–150°W, 5°S–5°N). Data are an update of Rayner et al. (2006).

The energetic background of year-to-year natural variability can temporarily augment, or diminish, the slower changes expected from external forcings of the climate system, such as warming caused by anthropogenic emissions of CO2 (Knight et al., 2009). By considering longer averaging periods, the effects of year-to-year variability are reduced. It is noteworthy that the decade 2000 to 2009 was the warmest in 160 years, significantly warmer than the 1990s which were warmer in turn than all earlier decades. The warmth of this decade would probably have been greater but for a decline in stratospheric water vapour (Solomon et al., 2010) which absorbs terrestrial radiation and so contributes to the overall greenhouse effect.

2009 was approximately 0.1 degC warmer than 2008, reflecting the shift from La Niña to El Niño in the tropical Pacific (Figures 2 and 3). The year started cooler than average in much of the tropical Pacific (Figure 1(g)), but temperatures rose throughout the year, exceeding El Niño thresholds in the summer and peaking in late December. Temperatures in the eastern Pacific reached levels similar to those of other El Niños this century, but were well below the levels seen in 1998. However, temperatures in the central and western Pacific exceeded any seen since at least 1960.

Figure 2.

(a) Land-surface air and sea-surface temperature anomalies ( degC relative to 1961–1990) for 2009. Data are an update of Brohan et al. (2006). (b) As part (a) but expressed as percentiles of the 1961–1990 distribution of annual temperatures calculated using the method in (Horton et al.,2001). Crosses indicate that 2009 was the warmest year on record in that 5° pixel, dashes that 2009 was the coldest. In some pixels there are too few data in the climatology period to calculate accurate percentiles. As a result, there are more missing data points in Figure2(b) than in Figure2(a).

Figure 3.

Land surface air and sea surface temperature anomalies ( degC, relative to 1961–1990) for December 2008 to February 2009, March to May 2009, June to August 2009, and September to November 2009. Data are an update of Brohan et al. (2006). The crosses and dashes show the extreme anomalies expressed as percentiles of the 1961–1990 distribution of seasonal temperatures calculated using the method of Horton et al. (2001).

The effects of El Niño were felt throughout the tropics (Figures 2 and 3), with significant warmth – temperatures exceeding the ninetieth percentile – over large areas between 30°S and 30°N. The significance of the anomalies (Figure 2(b)) depends not only on the magnitude of the anomaly (Figure 2(a)), but also on the short-term variability at that location. In the Tropics, short-term variability is small so a smaller anomaly is needed to exceed the ninetieth percentile than at higher latitudes. The annual average temperature anomaly in the Tropics was 0.52±0.02 degC, the second highest on record (Figure 1(d)). The global effects of El Niño often lag changes in the Pacific by a few months (Trenberth et al., 2002), so more widespread warming associated with the 2009/2010 El Niño is likely to be more fully realised in 2010.

Western Europe and the North Atlantic were also significantly warmer than average. Large areas of the North Atlantic have been unusually warm (relative to the 1961–1990 period) since the mid-1990s. Part of this warmth is probably due to cyclical changes in the strength of ocean currents known as the Atlantic Multidecadal Oscillation. Routine monitoring of the strength of the currents has only recently become possible and shows that the year-to-year variability is much larger than previously thought (Cunningham et al., 2007). Observations in the Arctic and Antarctic are relatively sparse, but both areas recorded significantly higher than average temperatures in 2009.

Areas of North America, Russia, and the Southern Ocean were colder than average, although only limited areas fell below the tenth percentile. It is difficult to judge the significance of the cold anomalies in the Southern Ocean because consistent monitoring of the area by drifting buoys has only taken place in the last few years.

During the Northern Hemisphere (NH) winter, temperatures (Figure 3) were significantly above average over India and southeast Asia, central Africa, and the southern United States and Mexico. Limited areas of northern Australia and northeast Siberia were significantly colder than average. Many of the same areas of significant anomalous warmth persisted into the NH spring. In addition, many areas of Europe experienced significantly above average temperatures. Areas experiencing significantly cold weather were once again limited. During the summer, El Niño began to take hold, with significant warmth over large areas of the tropical Pacific. Other areas of significant warmth were southern and southeast Asia, central and northern Africa, Europe, Australia, Central America and northeast Canada, and neighbouring seas. The northern United States and southern Canada, on the other hand, experienced significantly below average temperatures during the summer months. By the autumn, El Niño was well established with significant warmth across large areas of the Tropics. Western Eurasia and Canada also experienced significant warmth.

Figure 4 shows series and trends in lower-troposphere and lower-stratosphere temperatures since 1958 (radiosonde era) and since 1978 (the satellite era). They are compared with surface temperature trends for reference and illustrate the uncertainty arising from imperfections in observing systems and analysis techniques. The warming associated with the El Niño is evident in temperatures in the troposphere.

Figure 4.

Global seasonal average lower stratospheric (upper panel) and tropospheric (lower panel) temperature anomalies from satellites and weather balloons. The trends in the series are shown in the panels to the right where the tropospheric trends are also compared to temperature changes at the surface over the periods 1958–2009 and 1979–2009. RSS (Mears and Wentz,2009a, 2009b) and UAH (Christy et al., 2003) are based on satellite data. HadAT2 (Thorne et al.,2005) is based on weather balloon data, which have been vertically weighted to resemble the satellite data.

The sharp peaks in the stratospheric temperature are the result of absorption of solar and terrestrial radiation by volcanic aerosols. Temperatures in the troposphere and stratosphere are not correlated, because transport by convection is inhibited by the static stability at and above the tropopause. Globally-averaged temperatures in the lower stratosphere in 2009 were similar to those in recent years. Thompson and Solomon (2009) compared temperatures in the lower stratosphere worldwide with total column ozone amount, most of which is contributed by ozone in the lower stratosphere. They found that the stratosphere was cooler after than before the volcanic eruptions of El Chichón and Pinatubo (as seen in Figure 4), in line with observed ozone losses through chemical reactions involving the volcanic material in the stratosphere. The weak stratospheric warming between 1995 and 2002 was consistent with the observed recovery of ozone, and the subsequent slight cooling is consistent with the effects of increasing greenhouse gases. Thompson and Solomon (2009) also explained the contrast between the worldwide stratospheric cooling trend and the polar-focused declining trend in ozone amount as a consequence of a strengthening of the ‘Brewer-Dobson’ stratospheric circulation, in which air rises near the Equator, moves poleward, sinks near the Poles, then returns equatorward.

Regional and local climate

In Europe, 2009 was warmer than average in most places including the UK (Figure 2(a)). The annual average land-surface air temperature anomaly based on CRUTEM3 data (Brohan et al., 2006) was +1.05±0.14 degC for the region 35°N–75°N, 10°W–30°E (Figure 6(a)). 2009 was between the third and tenth warmest year for Europe since 1850. The decadal fluctuations in European temperature are similar to those seen in the Central England Temperature record (Figure 6(a)). The decade 2000–2009 was the warmest on record for Europe (1850–2009) and Central England (1659–2009) and significantly warmer than the 1990s. The most significant temperature anomalies for the year as a whole (Figures 2(a) and 2(b)) were observed over southern and western Europe, though the magnitudes of the anomalies were greater further east where natural variability is higher.

Figures 5(a) and 5(b) show seasonal average temperature and total precipitation anomalies. Daily mean Central England Temperature is shown in Figure 6(b) and monthly and annual values of Central England Temperatures and England and Wales Precipitation are given in Table 1.

Figure 5.

(a) Surface air-temperature anomalies for Europe and sea-surface temperature anomalies for neighbouring waters (degC, relative to 1961–1990) for December 2008 to February 2009, March to May 2009, June to August 2009, and September to November 2009. Data are an update of Brohan et al. (2006). The crosses and dashes show the extreme anomalies expressed as percentiles of the 1961–1990 distribution of seasonal temperatures calculated using the method of Horton et al. (2001). (b) Seasonal precipitation totals expressed as a percentage of the 1961–1990 average for December 2008 to February 2009, March to May 2009, June to August 2009, and September to November 2009. Data are from Rudolf and Rubel (2005), Rudolf and Schneider (2005), Schneider et al. (2008) and Fuchs and Schneider (2008).

Figure 6.

(a) Annual average mean Central England (Manley, 1974; Parker and Horton, 2005, updated) and European (Brohan et al.,2006, updated) temperature anomalies (degC relative to 1961–1990, blue bars) from 1880 to 2009 and two-standard-deviation uncertainties (fine black lines). The red lines show the annual values smoothed with a 21-point binomial filter and highlight the interdecadal variations. The dashed portions of the red line show where the smoothed values are liable to change as new data are added to the end of the series. (b) Mean daily Central England Temperature (°C, CET) for 2009 (dark blue line). The heavy black line shows the normal for 1961–1990 after smoothing, and the red lines are the corresponding 10th and 90th percentiles for each day of the year. The yellow band is the interval between the 5th and 95th percentiles. The light grey lines represent the highest and lowest values of mean CET in the daily record since 1772 (Parker et al., 1992).

Table 1. Monthly and annual mean Central England Temperature (Parker et al., 1992), total England and Wales Precipitation (Alexander and Jones, 2000) and UK temperature for 2009 (Perry and Hollis, 2005a; 2005b).
 CET 2009°C (anomaly, degC)EWP 2009 mm (anomaly, %)UK 2009°C (anomaly, degC)
Jan3.0 (−0.8)93.5 (103)2.8 (−0.2)
Feb4.1 (0.4)63.1 (97)3.7 (0.7)
Mar7.0 (1.3)42.1 (57)6.1 (1.4)
Apr10.0 (2.1)47.5 (77)8.9 (2.2)
May12.1 (0.9)560 (86)10.9 (1.1)
Jun14.8 (0.6)55.8 (86)13.7 (1.0)
Jul16.1 (0.1)136.0 (218)15.2 (0.8)
Aug16.6 (0.8)63.9 (83) 15.4 (1.1)
Sep14.2 (0.6)36.0 (46)13.2 (1.0)
Oct11.6 (0.9)78.8 (90)10.4 (0.9)
Nov8.7 (2.1)196.0 (212)7.3 (1.8)
Dec3.1 (−1.5)111.0 (116)2.1 (−1.8)
Annual10.11 (0.6)979.6 (107)9.2 (0.8)

Winter 2008/2009 was generally colder and drier than average in the west of Europe. It was the coldest winter in the UK and Ireland since 1996/1997 although daily temperatures in Central England did not often drop below the fifth percentile. In France, it was the third coldest winter in 20 years. Severe cold waves during January and February brought temperatures of –25°C and below to Germany and Poland. Further east, however, the winter was predominantly mild, with areas of significant warmth in northern Scandinavia and in the eastern Mediterranean.

Spring was warmer than average across Eurasia except for a small area between the Black Sea and the Caspian Sea which was cooler than average, though not significantly so. In Europe, the most significant anomalies were observed in central areas and monthly anomalies exceeding 5 degC were recorded in Germany, Austria and the Czech Republic in April. In the Central England area, daily mean temperatures dropped below the 1961–1990 average on only a handful of days during the spring.

The summer of 2009 bore a number of similarities to the previous two summers, with the jet stream displaced south of its climatological average latitude. Wetter than average conditions extended east from the UK through southern Scandinavia into the eastern Baltic States. Rainfall over the UK was frequent, particularly in July which was the wettest on record (in a series from 1910) but only eighteenth wettest in the much longer England and Wales precipitation series which begins in 1766. However, not all the summer months were wetter than average in the UK – June was mostly drier than average, and August was drier, too, in the south, but very wet in the north. The UK also had some hot weather in late June and early July, with 32°C recorded at Hampton in Greater London on 1 July. At the same time, unusually high temperatures were recorded in southern Europe. During July, temperatures exceeded 45°C in parts of Italy, and parts of the Iberian Peninsula also experienced several heat waves.

Autumn was warmer than average almost everywhere, with the most significant deviations in the west of Europe. In the Central England area, autumn 2009 (11.5°C) was one of the ten warmest on record – with daily mean temperature reaching near record levels on four occasions – however, it was more than a degree cooler than the warmest autumn, that of 2006 (12.6°C). In November, a series of low-pressure systems brought warm, moist maritime air into the continent, dropping record amounts of rain across the UK and Ireland. For the England and Wales region it was the fourth wettest November since 1766. Prolonged and heavy rainfall on 18–20 November led to severe flooding across the Lake District. Some areas of high ground received more than 400mm in 72 hours. Seathwaite in Borrowdale recorded 316.4mm in 24 hours, setting a new UK record. Many rivers in the Lake District exceeded their previous maximum flows by a wide margin. Several road bridges collapsed or were damaged and some 1500 properties were flooded, many in the town of Cockermouth. The November rainfall totals were the highest on record at most Irish stations, including Valentia Observatory where the total of 360mm was the highest of any month in records that extend back to 1866 (more at

December 2009 saw an extreme negative phase of the Arctic Oscillation, namely much higher than average pressure over the North Pole and much lower than average pressure at lower latitudes. This led to much lower than average temperatures over northern Eurasia and the UK (Figure 6(b)) and above average temperatures further south, a pattern that persisted through the whole winter. It has been suggested that there is a link between El Niño events and the negative phase of the Arctic Oscillation during winter (Ineson and Scaife, 2008), with some evidence that the link is stronger during solar minima (Kryjov and Park, 2007).

Sea-ice extent – the area of ocean covered by sea ice at a concentration of 15% or higher – has been routinely monitored using microwave detectors on satellites since the late 1970s. These records have been combined with earlier records from aerial reconnaissance and ship reports to give a homogeneous view of changes in Arctic sea-ice extent over the past 55 years. The annual minimum sea-ice extent in the Arctic typically occurs during September (Figure 7). In 2009, the September Arctic sea-ice extent was the third lowest on record, falling to 5.28×106 km2: only in 2007 and 2008 was it lower. In 2007 and 2009, low sea-ice extent was associated with a high pressure anomaly over the Beaufort Sea (Ogi et al., 2010). In 2007, however, the high pressure combined with below average pressure over Siberia to advect warm air over the Chukchi Sea. Although 2009 saw higher pressure anomalies over the Beaufort Sea than 2007 in June and July – the Arctic Oscillation Index was strongly negative in both months (L'Heureux et al., 2010) – the exact conditions, particularly the absence of lower than average pressure over Siberia, were less conducive to melting of the sea ice. Cloudy conditions later in the summer also slowed the rate of melting. Temperatures at high northern latitudes were well above normal, which is consistent with the low ice extents.

Figure 7.

Arctic sea-ice extent (106 km2) for September from 1953 to 2009. Sea-ice extent is obtained by summing the area covered by all 1° latitude×1° longitude elements which have a sea-ice concentration of 15% or above. The analysis is an update of Rayner et al. (2003).

More detail on the climate in 2009 is to be found in the World Meteorological Organization's (WMO's) statement on the status of global climate in 2009 on and in the State of theClimate report published in the July issue of the Bulletin of the American Meteorological Society. A fuller account of European climate will be published by WMO in its Annual Bulletin on the Climate in WMO Region VI – Europe and the Middle East, 2009. Selected global and UK data sets can be accessed from A wider background for our results can be found in IPCC (2007), in particular Chapter 3, available from


We were supported by the Joint DECC and Defra Integrated Climate Programme – DECC/Defra (GA01101). We would like to thank Phil Jones at the University of East Anglia who contributed to the land-surface temperature analysis and John Prior, Mike Kendon and Tim Legg of the National Climate Information Centre who assisted with the UK information.