Generating artificial vertical soundings over complex terrain from the aLMo model output to drive a high resolution snow-drift model

Abstract

To obtain a high-resolution wind field and to model snow drift over complex topography, the Swiss Federal Institute for Snow and Avalanche Research (SLF) uses the atmospheric model ARPS (Advanced Regional Prediction System, University of Oklahoma). ARPS accesses horizontally homogeneous initial fields. Until recently, this model was driven by atmospheric soundings recorded at sites far away from the actual model domain. In order to optimise the initial conditions of ARPS and to provide the option of producing real forecasts, a new downscaling method has been developed, which can also be applied to other uses. Based on the vertical structure of the atmosphere produced by the operational mesoscale forecast model aLMo (Alpine Model, MeteoSwiss), an artificial sounding is created for a specified location within the ARPS grid. Some of the techniques applied include model bias adjustment (based on aLMo verification studies), the introduction of horizontal filters, the assimilation of surface observation data and the vertical displacement of the boundary layer, motivated by the distinct topographies of the ARPS and the aLMo model. The combination of the methods differs from parameter to parameter. This paper describes details of the downscaling technique and gives an example of the application of the method.