IPCC Fourth Assessment Report: Climate Change 2007
Climate Change 2007: Working Group I: The Physical Science Basis

3.8.5 Summary

Even though the archived data sets are not yet sufficient for determining long-term trends in extremes, there are new findings on observed changes for different types of extremes. The definitions of the phenomena are summarised in Table 3.7. A summary of the changes in extremes by phenomena, region and time is given in Table 3.8 along with an assessment of the confidence in these changes.

New analyses since the TAR confirm the picture of a gradual reduction of the number of frost days over most of the mid-latitudes in recent decades. In agreement with this warming trend, the number of warm nights increased between 1951 and 2003, cold nights decreased, and trends in the number of cold and warm days are also consistent with warming, but are less marked than at night.

For precipitation, analysis of updated trends and results for regions that were missing at the time of the TAR show increases in heavy events for the majority of observation stations, with some increase in flooding. This result applies both for areas where total precipitation has increased and for areas where total precipitation has even decreased. Increasing trends are also reported for more rare precipitation events, although results for such extremes are available only for a few areas. Mainly because of lack of data, it remains difficult to draw a consistent picture of changes in extreme precipitation for the tropics and subtropics.

Tropical cyclones, hurricanes and typhoons exhibit large variability from year to year and limitations in the quality of data compromise evaluations of trends. Nonetheless, clear evidence exists for increases in category 4 and 5 storms globally since 1970 along with increases in the PDI due to increases in intensity and duration of storms. The 2005 season in the North Atlantic broke many records. The global view of tropical storm activity highlights the important role of ENSO in all basins, and the most active year was 1997, when a very strong El Niño began, suggesting that observed record high SSTs played a key role.

For extratropical cyclones, positive trends in storm frequency and intensity dominate for recent decades in most regional studies performed. Longer records for the northeastern Atlantic suggest that the recent extreme period may be similar in level to that of the late 19th century.

As noted in Section 3.3.4, the PDSI shows a large drying trend over NH land areas since the mid-1950s and a drying trend in the SH from 1974 to 1998. Decreases in land precipitation, especially since the early 1980s are the main cause for the drying trends, although large surface warming during the last two to three decades has also likely contributed to the drying.

Table 3.7. Definition of phenomena used to assess extremes in Table 3.8

PHENOMENON  Definition 
Low-temperature days/ nights and frost days  Percentage of days with temperature (maximum for days, minimumfor nights) not exceeding some threshold, either fixed (frost days) or varying regionally (cold days/cold nights), based on the 10th percentile of the daily distribution in the reference period (1961–1990). 
High-temperature days/nights  See low-temperature days/nights, but now exceeding the 90thpercentile. 
Cold spells/snaps  Episode of several consecutive low-temperature days/nights. 
Warm spells (heat waves)  Episode of several consecutive high-temperature days/nights. 
Cool seasons/warm seasons  Seasonal averages (rather than daily temperatures) exceedingsome threshold.  
Heavy precipitation events (events that occur every year)  Percentage of days (or daily precipitation amount) with precipitation exceeding some threshold, either fixed or varying regionally, basedon the 95th or 99th percentile of the daily distribution in thereference period (1961–1990). 
Rare precipitation events (with return periods >~10 yr)  As for heavy precipitation events, but for extremes further into the tail of the distribution. 
Drought (season/year)  Precipitation deficit; or based on the PDSI (see Box 3.1). 
Tropical cyclones (frequency, intensity, track, peak wind, peak precipitation)  Tropical storm with thresholds crossed in terms of estimated wind speed and organisation. Hurricanes in categories 1 to 5, according tothe Saffir-Simpson scale, are defined as storms with wind speedsof 33 to 42 m s–1, 43 to 49 m s–1, 50 to 58 m s–1, 59 to 69 m s–1, and >70 m s–1, respectively. NOAA’s ACE index is a measure ofthe total seasonal activity that accounts for the collective intensityand duration of tropical storms and hurricanes during agiven tropical cyclone season. 
Extreme extratropical storms (frequency, intensity, track, surface wind, wave height)  Intense low-pressure systems that occur throughout the mid-latitudes of both hemispheres fueled by temperature gradients and actingto reduce them. 
Small-scale severe weather phenomena Extreme events, such as tornadoes, hail, thunderstorms, dust stormsand other severe local weather.  

Table 3.8. Change in extremes for phenomena over the specified region and period, with the level of confidence and section where the phenomenon is discussed in detail.

PHENOMENON  Change  Region  Period  Confidence  Section 
Low-temperature days/nights and frost days  Decrease, more so for nights than days  Over 70% of global land area  1951–2003 (last 150 yearsfor Europeand China)  Very likely 
High-temperature days/nights  Increase, more so for nights than days  Over 70% of global land area 1951–2003  Very likely 
Cold spells/snaps (episodes of several days)  Insufficient studies,but daily temperature changes imply a decrease          
Warm spells (heat waves) (episodes of several days)  Increase: implicit evidence from changes of daily temperatures Global  1951–2003  Likely  FAQ 3.3 
Cool seasons/ warm seasons (seasonal averages)  Some new evidence for changes in inter-seasonal variability Central Europe  1961–2004  Likely  
Heavy precipitation events (that occur every year)  Increase, generally beyond that expected from changes in the mean (disproportionate)  Many mid-latitude regions (even where reduction in total precipitation) 1951–2003  Likely 
Rare precipitation events (with return periods > ~10 yr)  Increase  Only a few regions have sufficient data for reliable trends (e.g., UK and USA)  Various since 1893  Likely (consistent with changes inferred for more robust statistics) 
Drought (season/year)  Increase in total area affected  Many land regions of the world  Since 1970s  Likely  3.3.4 and FAQ 3.3 
Tropical cyclones  Trends towards longer lifetimes and greater storm intensity, but no trend in frequency Tropics  Since 1970s  Likely; more confidence in frequency and intensity  3.8.3 andFAQ 3.3 
Extreme extratropical storms  Net increase in frequency/intensity and poleward shift in track NH land  Since about 1950  Likely  3.8.4, 3.5,and FAQ 3.3 
Small-scale severe weather phenomena  Insufficient studies for assessment