Assessment of ECMWF model bias in the AMMA region with observations from the ARM Mobile Facility at Niamey

 
Poster PDF

Authors

Anton Beljaars — European Centre for Medium-Range Weather Forecasts
Maike Ahlgrimm — Deutscher Wetterdienst
Anna Agusti-Panareda — European Centre for Medium-Range Weather Forecasts

Category

Modeling

Description

The ARM Mobile Facility was deployed to Niamey, Niger, in 2006 in the wider context of the African Monsoon Multidisciplinary Analysis (AMMA). The short-term forecast of the ECMWF model quickly develops biases, intensifying the Saharan heat low, placing the ITCZ further south than observed, and underestimating the frequency and intensity of the intermittent deep convection and associated precipitation in the Sahel region. Observations from the ARM Mobile Facility at Niamey provide an opportunity to assess how physical processes in the model contribute to the developing bias. The incoming shortwave radiation is overestimated in the model. A combination of lacking aerosol optical depth and cloud cover contribute to this overestimation. The net radiation absorbed by the surface is also overestimated. The model develops a warm and dry bias at the surface, which is adjusted in the analysis increments by increasing soil moisture and thus enhancing evaporation. As a result, the model’s surface latent heat flux is unrealistically high during the dry months of the year. The model develops a deep, well-mixed boundary layer. Analysis increments adjust the temperature profiles to reduce the near-surface warm bias. Since radiosonde locations are sparse in the Sahel, this localized cooling induces a secondary circulation resulting in subsidence and further suppression of convection.