Changes in climate extremes

The last section looked at the likely regional patterns of climate change. Can anything be said about likely changes in the frequency or intensity of climate extremes in the future? It is, after all, not the changes in average climate that are generally noticeable, but the extremes of climate - droughts, floods, storms and ts

(a)

climate /

Less / cold weather

-^ i-^—

/ « New \ / climate \

More hot weather

\ More \ record hot weather

Average

Cold

Previous climate

Average

Increase in variance

Previous climate

Increase in variance

Cold

Average

More More cold record weather cold weather

Cold

Average

Increase in mean and variance extremes of temperature in very cold or very Increase in mean warm periods - which provide the largest impact on our lives (see Chapter 1).23

The most obvious change we can expect in extremes is a large increase in the number of extremely warm days and heatwaves (Figure 6.8) coupled with a decrease in the number of extremely cold days. Many continental land areas are experiencing substantial increases in maximum temperature and more heatwaves. An outstanding example is the heatwave in central Europe in 2003 (see box on page 215). Model projections indicate, as shown in Figure 6.8c, much increased frequency and intensity of such events as the twenty-first century progresses.

Of even more impact are changes in extremes connected with the hydrological cycle. In the last section it was explained that in a warmer world with increased greenhouse gases, average precipitation increases and the hydrologi-cal cycle becomes more intense.24 Consider what might occur in regions of increased rainfall. Under the more intense hydrological cycle the larger amounts of rainfall will come from increased convective activity: more really heavy showers and more intense thunderstorms. This is well illustrated by Figure 6.9 which shows how, on doubling the carbon dioxide concentration, the number of days with large rainfall amounts (greater than 25 mm day?-1) doubled. Although from a climate model of some years ago, it illustrates a robust result from all climate models that more intense precipitation events and more dry days are to be expected during the twenty-first century as global warming increases. The substantial degree of model agreement is illustrated in Figure 6.10 which shows an analysis of results from nine different models, for precipitation intensity. Similar information for other indices related to extremes, for instance heatwaves, frost days, dry days, etc. is provided in the article from which Figure 6.10 is taken.25

More hot weather More record hot weather

Increase in mean and variance

Cold

Average

Figure 6.8 Schematic diagrams showing the effects on extreme temperatures when (a) the mean increases leading to more record hot weather, (b) the variance increases and (c) when both the mean and the variance increase, leading to much more record hot weather.

Cold

Average

Much more hot weather

J More .jr record hot weather

Figure 6.8 Schematic diagrams showing the effects on extreme temperatures when (a) the mean increases leading to more record hot weather, (b) the variance increases and (c) when both the mean and the variance increase, leading to much more record hot weather.

160

140

120

N?

100 -

>,

u c

80

<D

3

CJ-<U

60

C

40

<D

C

tti

20

U

Australian land points

Over 25.6

Daily rainfall class (mm day-1)

Figure 6.9 Changes in the frequency of occurrence of different daily rainfall amounts with doubled carbon dioxide as estimated by a CSIRO model in Australia.

1880 1920 1960 2000 2040 2080 I I I II I I I I standard deviation

Year -1.25 -1 -0.75-0.5-0.25 0 0.25 0.5 0.75 1 1.25

Figure 6.10 Changes in extremes based on multi-model simulations from nine global coupled climate models, adapted from Tebaldi et al. (2002). (a) Globally averaged changes in precipitation intensity (defined as the annual total precipitation divided by the number of wet days) for low (SRES B1), middle (SRES A1B) and high (SRES A2) scenarios. (b) Changes of spatial pattern of precipitation intensity based on simulations between two 20-year means (2080-99 minus 1980-99) for the A1B scenario. Solid lines in (a) are the 10-year smoothed multi-model ensemble means, the envelope indicates the ensemble mean standard deviation. Stippling in (b) denotes areas where at least five of the nine models concur in determining that the change is statistically significant. Extreme indices are calculated only over land and are calculated following Frich et al. 2002. Because the study focused on analysing the direction and significance of the changes and the degree of inter-model agreement, the indices plotted are shown in units of standard deviation rather than absolute magnitude. Each model's time series was centred around its 1980-99 average and normalised (rescaled) by its standard deviation computed (after detrending) over the period 1960-2099, then the models were aggregated into an ensemble average, both at the global average and the grid-box level.

1880 1920 1960 2000 2040 2080 I I I II I I I I standard deviation

Year -1.25 -1 -0.75-0.5-0.25 0 0.25 0.5 0.75 1 1.25

Figure 6.10 Changes in extremes based on multi-model simulations from nine global coupled climate models, adapted from Tebaldi et al. (2002). (a) Globally averaged changes in precipitation intensity (defined as the annual total precipitation divided by the number of wet days) for low (SRES B1), middle (SRES A1B) and high (SRES A2) scenarios. (b) Changes of spatial pattern of precipitation intensity based on simulations between two 20-year means (2080-99 minus 1980-99) for the A1B scenario. Solid lines in (a) are the 10-year smoothed multi-model ensemble means, the envelope indicates the ensemble mean standard deviation. Stippling in (b) denotes areas where at least five of the nine models concur in determining that the change is statistically significant. Extreme indices are calculated only over land and are calculated following Frich et al. 2002. Because the study focused on analysing the direction and significance of the changes and the degree of inter-model agreement, the indices plotted are shown in units of standard deviation rather than absolute magnitude. Each model's time series was centred around its 1980-99 average and normalised (rescaled) by its standard deviation computed (after detrending) over the period 1960-2099, then the models were aggregated into an ensemble average, both at the global average and the grid-box level.

Drought.

What does this mean in terms of floods and droughts? More intense precipitation means more likelihood of floods. Illustrated in Figure 6.11 is a modelling study showing that if atmospheric carbon dioxide concentration is doubled from its pre-industrial value, the probability of extreme seasonal precipitation in winter is likely to increase substantially over large areas of central and northern Europe. The increase in parts of central Europe is such that the return period of extreme rainfall events would decrease by about a factor of five (e.g. from 50 years to 10 years). Similar results have been obtained in a study of major river basins around the world.26

Note also from Figure 6.9 that the number of days with lighter rainfall events (less than 6 mm day-1) is expected to decrease in the globally warmed world. This is because, with the more intense hydrological cycle, a greater proportion of the rainfall will fall in the more intense events and, furthermore, in regions of convection the areas of downdraught become drier as the areas of updraught become more moist and intense. In many areas with relatively low rainfall, the rainfall will tend to become less. Take, for instance, the

1 2 3 5 7

Figure 6.11 The changing probability of extreme seasonal precipitation in Europe in winter as estimated from an ensemble of 19 runs with a climate model starting from slightly different initial conditions. The figure shows the ratio of probabilities of extreme precipitation events in the years 61 to 80 of 80-year runs that assumed an increase of carbon dioxide concentration of 1% per year (hence doubling in about 70 years) compared with control runs with no change in carbon dioxide.

2020 2040 2060 2080

Year

Figure 6.12 Proportion of the world's land surface in drought (extreme, severe and moderate) each month as projected for the twenty-first century by the Hadley Centre climate model. In each case results from three simulations with the A2 emissions scenario are shown.

2020 2040 2060 2080

Year

Figure 6.12 Proportion of the world's land surface in drought (extreme, severe and moderate) each month as projected for the twenty-first century by the Hadley Centre climate model. In each case results from three simulations with the A2 emissions scenario are shown.

situation in regions where the average summer rainfall falls substantially -as is likely to occur, for instance, in southern Europe (Figure 6.7). The likely result of such a drop in rainfall is not that the number of rainy days will remain the same, with less rain falling each time; it is more likely that there

The sea surface temperature in August 2005. Orange and red depict regions where the conditions are suitable for hurricanes to form (at 28°C or higher). Hurricane winds are sustained by the heat energy of the ocean and Hurricane Katrina in 2005 caused catastrophic damage along the Gulf coast from Florida to Texas, and in particular in New Orleans, Louisiana

The sea surface temperature in August 2005. Orange and red depict regions where the conditions are suitable for hurricanes to form (at 28°C or higher). Hurricane winds are sustained by the heat energy of the ocean and Hurricane Katrina in 2005 caused catastrophic damage along the Gulf coast from Florida to Texas, and in particular in New Orleans, Louisiana will be substantially fewer rainy days and considerably more chance of prolonged periods of no rainfall at all. Further, the higher temperatures will lead to increased evaporation reducing the amount of moisture available at the surface - thus adding to the drought conditions. The proportional increase in the likelihood of drought is much greater than the proportional decrease in average rainfall.

A recent study of the incidence of drought27 has employed the Hadley Centre climate model to simulate droughts over all continents, first during the second half of the twentieth century so that confidence in the model could be established by comparing simulations with observed droughts. Droughts are divided into three categories: extreme, severe and moderate.28 Averaged over the period 1952-98, the percentages of the world's land area at any one time under extreme, severe and moderate drought were 1%, 5% and 20%. By the beginning of the twenty-first century these proportions had risen to 3%, 10% and 28%. Projections for the twenty-first century under the SRES A2 scenario (Figure 6.12) show the proportions of land area under extreme drought rising to over 10% by 2050 and 30% by 2100, the increases occuring not because droughts are much more frequent but much longer in duration. Their study indicates the areas most vulnerable to drought, broadly in agreement with the areas of reduced rainfall indicated in Figure 6.7.

In the warmer world of increased greenhouse gases, therefore, different places will experience more frequent droughts and floods - we noted in Chapter 1 that these are the climate extremes which cause the greatest impacts and will be considered in more detail in the next chapter.

What about other climate extremes, intense storms, for instance? How about hurricanes and typhoons, the violent rotating cyclones that are found over the tropical oceans and which cause such devastation when they hit land? The energy for such storms largely comes from the latent heat of the water which has been evaporated from the warm ocean surface and which condenses in the clouds within the storm, releasing energy. It might be expected that warmer sea temperatures would mean more energy release, leading to more frequent and intense storms. However, ocean temperature is not the only parameter controlling the genesis of tropical storms; the nature of the overall atmospheric flow is also important. Further, although based on limited data, observed variations in the intensity and frequency of tropical cyclones show no clear trends in the last half of the twentieth century. AOGCMs can take all the relevant factors into account but, because of the relatively large size of their grid, they are unable to simulate reliably the detail of relatively small disturbances such as tropical cyclones. From projections with these models there is no consistent evidence of changes in the frequency of tropical cyclones or their areas of formation. However, during the last few years a number of studies with regional models and more adequate resolution with large-scale variables taken from AOGCMs (see next section on regional modelling) project some consistent increases in peak wind intensities and mean and peak precipitation intensities. An indication of the size of the increases is provided from one study that projected an increase in 6% in peak surface wind intensities and 20% increase in precipitation.29

Regarding storms at mid latitudes, the various factors that control their incidence are complex. Two factors tend to an increased intensity of storms. The first, as with tropical storms, is that higher temperatures, especially of the ocean surface, tend to lead to more energy being available. The second factor is that the larger temperature contrast between land and sea, especially in the northern hemisphere, tends to generate steeper temperature gradients, which in turn generate stronger flow and greater likelihood of instability. The region around the Atlantic seaboard of Europe is one area where such increased stormi-ness might be expected, a result indicated by some model projections. However, such a picture may be too simple; other models suggest changes in storm tracks that bring very different changes in some regions and there is little overall consistency between model projections.

For other extremes such as very small-scale phenomena (e.g. tornadoes, thunderstorms, hail and lightning) that cannot be simulated in global models, although they may have important impacts, there is currently insufficient information to assess recent trends, and understanding is inadequate to make firm projections.

Table 6.2 summarises the state of knowledge regarding the likely future incidence of extreme events. Although many of the trends are clear, many more studies are required that provide quantitative assessments on regional scales of likely changes in the frequency or intensity of extreme events or in climate variability.

0 0

Post a comment