A time-varying subsidence parameterization for the atmospheric boundary layer

Abstract

This study examines the effect of a time-varying parameterization for subsidence in the atmospheric boundary layer (ABL) on a one-dimensional coupled land-atmosphere model. Measurements of large-scale divergence in the ABL are scarce and often marred by error, providing the motivation to model this important physical process and estimate its values from indirect but related observations. Constant parameterizations of-large- scale divergence and/or subsidence velocity are adequate for periods within a characteristic synoptic time scale, but longer studies require a parameterization that yields to local atmospheric change. After confirming the potential significance of subsidence in the ABL, this experiment investigates two key areas: (1) the ability to model subsidence change as a response to estimated time-varying model error and (2) the net improvement and potential benefits of this enhancement. This study indicates a consistent reduction of root-mean-square error scores for the time-varying subsidence (divergence) parameter scheme versus a constant parameterization for the 2 m specific humidity measurement, with negligible change to the 2 m temperature measurement.

(cont.) Model error does not improve explicitly, in spite of the presumed improvement to model physics. However, the unknown nature of the model error precludes an accurate diagnose of change, thus leaving the root-mean-square-error scores as the principal tool of evaluation and hence the justifying the potential usefulness of the time-varying parameterization.