Dynamics of gap winds in the Great Rift Valley, Ethiopia: emphasis on strong winds at Lake Abaya

Abstract. Lake Abaya, located in the Great Rift Valley (GRV) in Ethiopia, is affected by regularly occurring strong winds that cause water waves, which in turn affect the lake's ecology and food web. The driving forces for these winds, however, are yet unexplained. Hence, the main goal of this study is to provide a physical explanation for the formation of these strong winds in the GRV and especially at Lake Abaya. To this aim, two case studies were performed based on measurements, ERA5 reanalysis data and mesoscale numerical simulations conducted with the Weather Research and Forecasting (WRF) model. The simulations revealed that in both cases a gap flow downstream of the narrowest and highest part of the GRV (i.e. the pass) led to high wind speeds of up to 25 m s−1. Two types of gap flow were identified: a north-eastern gap flow and a south-western gap flow. The wind directions are in line with the orientation of the valley axis and depend on the air mass distribution north and south of the valley and the resulting along-valley pressure gradient. The air mass distribution was determined by the position of the Intertropical Convergence Zone relative to the GRV. The colder air mass was upstream of the GRV in both case studies. During the day, the convective boundary layer in the warmer air mass on the downstream side heated up more strongly and quickly than in the colder air mass. The most suitable variable describing the timing of the gap flow was found to be the pressure gradient at pass height, which corresponds roughly to the 800 hPa pressure level. In both cases the gap flow exhibited a strong daily cycle, which illustrates the importance of the thermal forcing due to differential heating over complex terrain in addition to the large-scale forcing due to air mass differences. The start, strength, and the duration of the gap winds within the valley depended on location. For both cases, the strongest winds occurred after sunset and in the ongoing night downstream of the gap and on the corresponding lee slope. The ERA5 reanalysis captures both events qualitatively well but with weaker wind speeds than in the mesoscale numerical simulations. Hence, ERA5 is suitable for a future climatological analysis of these gap flows.