Sharing knowledge
© Crown Copyright
African street market
African street market

Output 2

Improved application of science and climate models for early warning systems and adaptation planning.

Led by  Adam Scaife.

Climate forecasts from months to a decade ahead need to account for the current phase of natural climate variability and increases in greenhouse gases as both play a role in near term climate. Tropical climates are also generally more predictable than climate in extra-tropical regions and forecast systems skilfully predict major tropical climate drivers such as El Niño and anthropogenic change months and years ahead. This leads to skill in forecasts in potentially vulnerable regions of Africa but the regions and quantities for which skill is present are still unknown and are not being utilised by African decision makers. By using monthly-to-decadal predictions we can better inform adaptation to climate change in Africa and there are many user-focused applications that could be developed in consultation with DFID.

While it is tempting in some quarters to blame every extreme weather event on climate change, inappropriate adaptation responses could follow if based on incorrectly assuming that every event is attributable to increased greenhouse gas concentrations in the atmosphere. While the evidence for human influence on global climate is clear, extreme weather events also occurred in pre-industrial climate, and other drivers in addition to greenhouse gases, such as aerosols and land use changes can have major effects on climate. By considering how external drivers of climate and modes of variability have changed the probability of particular weather events we can improve the evidence basis to better inform adaptation.

  1. Monthly-to-seasonal prediction
    Through the initial consultation process we will determine high priority climate variables for vulnerable regions. This will focus initial attention on specific areas and diagnostics to analyse in our existing seasonal forecast system. We will assess current skill, determine what climate drivers are responsible and focus our development work on improving forecast skill for these regions and variables. We expect this to include estimates of the likelihood of extreme events, estimates of exceeding thresholds and some monthly forecast information such as onset or duration of rains.
  2. Decadal predictions
    These longer range predictions will be more heavily influenced by climate change signals and bridge the gap between initialised seasonal forecasts and uninitialized centennial climate projections. They will provide experimental forecasts for adaptation to climate change on decadal timescales. Decadal predictions need to be "seamless" with monthly and seasonal forecasts from 2.1 to avoid providing conflicting information.
  3. Attribution
    We will develop the capability to provide updates in near-real time of the causes of extreme events. This will include an assessment of the probability of extreme events and how they have changed and will investigate the effects of external forcings and variability on regional climate changes and extremes. Examples include: links between aerosol forcing, shifts in the ITCZ (Inter-Tropical Convergence Zone) and Sahel drought, the effects of biomass burning on African climate, the effects of land use changes on extreme temperatures. Based on the understanding developed, and an analysis of the relevant driving processes in the prediction models, a near-real time attribution capability will be developed that will sit alongside the forecasting system. This will provide regularly updated appraisals of evolving climate conditions and extreme weather events in Africa, assessing the extent to which modes of variability and external forcings of climate have altered the probability of their occurrence.


Output 2 - highlight results 

Last Updated: 8 May 2012