Resolving Sahelian Thunderstorms Improves Midlatitude Weather Forecasts

The West African monsoon is the driving element of weather and climate during summer in the Sahel region. The summertime Sahel has the worldwide highest degree of thunderstorm organization into long-lived, several hundred-kilometer elongated, fast propagating mesoscale convective systems (MCSs) that contribute 90% to the annual rainfall. Practically all current global weather prediction and climate models represent thunderstorms using convection parametrization schemes. Shortcomings of the parametrizations, particularly the organization of convection on the mesoscale, deteriorate the modelled distribution of rainfall from individual storms and the entire West African monsoon circulation. Overall, this has negative impacts on forecasts over West Africa itself but may also affect remote regions, as waves originating from convective heating are badly represented.

Here we investigate those local and remote forecast impacts based on daily initialized 10-day forecasts for July and August 2016 and 2017 using the ICON model. One set of simulations employs the operational setup of the global model with a horizontal grid spacing of 13 km. It is compared with simulations using the two-way nesting capability of ICON. A second model domain over West Africa (the nest) with 6.5 km grid spacing is sufficient to explicitly resolve MCSs in this region. In the two-way nested simulations, the prognostic variables of the global model are influenced by the results of the nest through relaxation.

The nest with explicit convection is able to reproduce single MCSs much more realistically compared to the stand-alone global simulation with parametrized convection. Comparisons with radiosondes from the sparse observational network in West Africa indicate how this computationally inexpensive increase of model resolution over West Africa yields a better vertical distribution of moisture within the monsoon system. These improvements ultimately enhance 5–8-day tropical and midlatitude weather forecasts, most likely due to a better representation of African easterly waves and Rossby waves.

We advocate an operational use of a modelling strategy similar to the one presented here for a cost-effective improvement of global weather prediction and potentially even (sub-) seasonal and climate simulations.