African Easterly Wave variability in present and future climate simulations

The West African Monsoon (WAM) is a coupled atmosphere-ocean-land system characterised by summer rainfall and winter drought. It is of crucial importance to the people of sub-Saharan Africa as extremes in West Africa substantially impact water and food resources, as well as infrastructure. On an intraseasonal timescale however, the evolution of convection is less well-understood than other key features of the WAM even though its sum gives the seasonal mean, with large convective systems bringing about 70% of the total annual rainfall to the West African region. In response to this special anniversary session of the first observation of African Easterly Waves, we present first results from the new CMIP5 high resolution simulations in which we evaluate the ability of the GCMs participating in CMIP5 to simulate various aspects of the West African Monsoon (WAM). We explore the mechanisms that govern the intraseasonal variability of convection (internal forcing and feedback loops versus remote forcings) and of the impact of its variability over other regions of Africa with special emphasis given to the variability of precipitation, African Easterly Waves (AEW) and the African Easterly Jet (AEJ). The motivation for this study has been two-fold. Firstly, there is a need to better understand the key atmosphere, land and ocean interactions that govern the WAM evolution, as well as the interactions between dynamics and convection at intraseasonal and shorter timescales. Shorter timescales feedback on longer time- and larger space-scales in a continuum of process-related interactions. Therefore assessing short time scale variability associated with the WAM in models is an integral part of evaluating the CMIP5 simulations. Ultimately, understanding the origin of model biases in weather will inform model development and improve the reliability of their projections for this vulnerable region. Secondly, prediction, which informs adaptation for the African people, requires information about the statistics of the weather systems of the WAM region, such as the variability of the number and intensity of AEW. This is also important for downstream cyclogenesis of hurricanes. It is thus important to be able to quantify this uncertainty in current and future projections of these systems. Here, we present results focussing on two specific science questions: (1) How realistic are AEW spatial distributions, variability and structures and what are their sensitivity to model resolution and convective parameterizations? (2) Does resolving AEWs make a difference to climatology? Results include an evaluation of AEW climatology and variability compared with 4 high resolution reanalyses and an analysis of their structure, period and wavelength, using both the limited 6 hourly pressure level data as well as the multi-level 3 hourly model level data and the pioneering tracking technique developed in Hodges et al. (2003). A new system centered means of identifying regions associated with the waves has been used to extract data at various levels and for various relevant fields (eg. temperature, winds, humidity) to prepare composites. This has allowed the life cycles and full 3D structure of the waves to be studied in detail and the relationship of the AEWs with rainfall to be evaluated relative to the model convection scheme (eg. Strauss and Kiladis, 2009), similar to the approach used for extra-tropical cyclones (Catto et al, 2010). The model evaluation of the WAM and the generation of AEW statistics included the following hindcasts: AMIP, pre-industrial run and historical run. The AEW variability will also be assessed in the near-term decadel forecasts and latterly in different climate scenarios if these are made available in time. RCP4.5 and RCP8.5 forecasts have been compared to understand the impact of medium and high emission scenarios on the WAM and AEW variability.