Drought

The research on the changes in drought events was conducted at the Met Office Hadley Centre by Dr Eleanor Burke. This compared the occurrence of drought events in the pre-industrial climate to those that would be expected to occur in a world with a global average temperature +4 °C (7 °F) higher that the pre-industrial climate.

How drought is measured

Drought was defined by considering the amount of soil moisture available for plants to utilise. This type of soil moisture drought is a good representation of agricultural drought.

It is assumed that on average a drought event occurs 10% of the time (or one in 10 years) at every location in the period 1860-1891. This can then be used to define a threshold for drought in that period. The proportion of time the soil moisture goes below that threshold in the future period can be used to define the frequency of occurrence of drought in the future. Only regions which remain unfrozen for the whole year at least one year of the period 1860-1891 and are between 60S and 60N, were used in this study.

This value of soil moisture was analysed using data from the Met Office Hadley Centre HadCM3 model. This was run multiple times with slightly different assumptions and for two different green house gas emission scenarios, A1B and A1FI. The A1B scenario is referred to as the medium emissions scenario and the A1FI scenario is the high emissions scenario. The result was an ensemble of model runs that give a range of results. The average was taken of those simulations that showed a rise in global average temperature of +4 °C (7°F) or higher by the end of the century.

Fig 1. Proportion of time in drought in a climate with a global average temperature +4 ºC (7 ºF) higher than the pre-industrial climate, using models following the A1B emissions scenario. Fig 1. Proportion of time in drought in a climate with a global average temperature +4 ºC (7 ºF) higher than the pre-industrial climate, using models following the A1B emissions scenario.   Fig 2. Proportion of time in drought in a climate with a global average temperature +4 ºC (7 ºF) higher than the pre-industrial climate, using models following the A1B emissions scenario. Fig 2. Proportion of time in drought in a climate with a global average temperature +4 ºC (7 ºF) higher than the pre-industrial climate, using models following the A1B emissions scenario.

Findings

When explicit regions are selected, these were extracted as 'Giorgi regions'.

Drought does not increase everywhere

  • Between 60S and 60N 62% of land surface has more drought in the future, and 38% has less.

  • Similarly, between 23.5N and 23.5S 62% of land surface has more drought in the future, and 38% has less.

There is an overall increase in drought

  • Between 60S and 60N land surface is in drought 14.5% of the time compared with 10% in period 1860-1891.

  • Between 60S and 60N land surface is in drought on average 17% of the time compared with 10% in period 1860-1891.

Regions with increase in drought

  • Mediterranean basin from 10% to 20% (i.e. one in ten to one in five)

  • Amazon basin from 10% to 25%

  • Southern Africa from 10% to 19%

  • Southeast Asia from 10% to 18%

Regions with decrease in drought

  • Eastern Africa from 10% to 7%

  • South and Central Asia from 10 to 8%

References

Burke, E.J., S.J. Brown and N. Christidis, 2006: Modelling the recent evolution of global drought and projections for the twenty-first century with the Hadley centre climate model. Journal of Hydrometeorology, 7(5), 1113-1125.

Burke, E.J. and S.J. Brown, 2008: Evaluating uncertainties in the projection of future drought. Journal of Hydrometeorology, 9(2), 292-299.

Last updated: 3 December 2013