Implementation considerations of a conceptual precipitation model

Some aspects of water management, such as flash flood analysis or sewerage management require spatially distributed rainfall estimates and forecasts o ver surface areas ranging from a few square kilometers to a few hundred squ are kilometers. Typically, these requirements cannot be satisfied by operat ional numerical weather prediction models. Faced with these constraints, an alternative solution consists of designing modeling tools consistent with observations routinely available for the survey of catchments, including gr ound meteorological data, voluminal radar data, satellite data, and operati onal numerical model output fields. This research headline is inspired by G eorgakakos and Bras [1984a] who proposed a simplified dynamical approach co nsidering an atmospheric column as a reservoir of liquid water to describe rainfall evolution. Initially, based on ground meteorological data, this ap proach was later adapted to voluminal radar data to model the evolution of vertically integrated rainwater content (VIL) in the atmospheric column. In the present work, the forecast lead time is extended through a proposed so lution consisting of implementing a simplified precipitation model explicit ly accounting for the cloud water content state. This paper demonstrates th e potential interest of taking into account the cloud water state through a feasibility study. The first part of the paper presents the model formulat ion introducing a reservoir representing the cloud water state. The second part of the paper evaluates the potential improvement gained by introducing this component, and the influence of cloud and rainwater uncertainties on model performance. This evaluation utilizes rainwater content, cloud water content, and related meteorological variables produced by a meteorological microphysical monodimensional model. Results of modeling and forecast exper iments are included to demonstrate the value of introducing the cloud water state. The experiments show improved forecast performance using the model accounting for cloud water compared with the simple extrapolation method an d a related precipitation model dealing only with the rainwater content sta te.