FAIL (the browser should render some flash content, not this).

Modeling the West African Monsoon

Reference
Roehrig, R., Bouniol, D., Guichard, F., Hourdin, F. and Redelsperger, J.-L. 2013. The present and future of the West African Monsoon: A process-oriented assessment of CMIP5 simulations along the AMMA transect. Journal of Climate 26: 6471-6505.
The only way to know how well a specific climate model can predict the future of some particular region is to determine how well it can predict that region's past, where its predictions can be compared with real-world archived data that accurately describe the past. And, therefore, there have been a host of studies conducted to test various climate models in situations where the past climates of the regions involved are well documented.

Acting within this logical framework, Roehrig et al. (2013) analyzed the degree to which several models of the Coupled Model Intercomparison Project (CMIP) phase 5 (CMIP5) are able to reconstruct the past behavior of the West African Monsoon. And by doing so, the five French researchers report that (1) "CMIP5 climate change projections in surface air temperature and precipitation are found to be very similar to those of CMIP3," that (2) "as in CMIP3, the spread of model projections remains very large for both temperature and precipitation," that (3) "the precipitation response tends to be lower than the observed decadal variability in the second half of the twentieth century," that (4) "temperature changes also remain very uncertain," that (5) "CMIP5 coupled models still suffer major sea surface temperature biases in the equatorial Atlantic," that (6) "the averaged Sahel rainfall exhibits a large spread (±50%)," that (7) "the wrong phasing of the diurnal cycle of precipitation remains an issue," that (8) "most CMIP5 models ... have not reached yet a degree of maturity that directly makes them trustable to anticipate climate changes and their impacts," that (9) "many systematic and robust biases of the coupled and atmospheric models have not improved from CMIP3 to CMIP5," and that (10) "large surface radiative biases in arid and semiarid regions are a major issue in current simulations [as] they lead to departure from the observed radiative balance."

In the words of Roehrig et al., as they sum up their findings, the CMIP5 models "demand further work to achieve a reasonable realism."

Archived 27 November 2013