Precipitation, Evaporation and Runoff in the ECMWF coupled atmosphere/land system

Anton Beljaars,  Head of the Physical Aspects Section at European Centre for Medium-Range Weather Forecasts (ECMWF)

Special Seminar on Friday, February 23, 2018 at 3 pm in Harvill 318 

After a general introduction on progress in numerical weather prediction (NWP), this seminar will focus  on the coupling between atmosphere and land surface through precipitation and evaporation in the ECMWF system. A overview will be given of two major changes to the parametrization of convection. These changes brought the level of variability in the system closer to observations, they improved tropical cyclone track forecasts and delayed the diurnal cycle over land. The convection modifications improved tropical cyclone track forecasts even more than a resolution increase that was introduced before.

Furthermore, summer precipitation/evaporation coupling over land will be discussed. Published research suggests that at the convection scale, convective events are more likely over dry areas than over surrounding wet areas, i.e. on a local scale there is a negative feedback. However, the ECMWF model does not have the resolution to resolve small scale convective events. It turns out that in summer at larger scales, evaporation is the main moisture source for precipitation. This implies a positive feedback, e.g. more evaporation leads to more precipitation which makes more water available for evaporation. Numerical experimentation for the 1993 Mississippi flood case exemplifies this. Additional sensitivity experiments indicate that precipitation over large scale continental areas follows evaporation rather closely. The crucial element for seasonal soil drying is the residual between precipitation and evaporation. This makes summer evaporation and precipitation very sensitive to the representation of runoff. Experimentation where runoff is artificially increased does indeed indicate such a sensitivity. To improve the simulation of water cycling between land surface and atmosphere, it is tentatively concluded that progress is needed on the representation of runoff which requires a close collaboration between meteorologist and hydrologist.