Séminaire de Christian Jakob

In recent years it has become apparent in numerous studies that the amount of rainfall in tropical regions is strongly non-linearly related to the vertically integrated water vapor content of the atmosphere. This relationship has prompted the development of a number of simple models of the interaction of water vapor and tropical convection. More recently it has been hypothesized that while low-level processes likely control the occurrence of convection, the amount of rainfall, and by association the strength of convection is controlled by the amount of water vapor in the mid-troposphere.

This raises the question which processes are primarily responsible for moistening the middle troposphere and hence providing the environment for strong convection. Likely candidates are horizontal or vertical advection, and non- or weakly precipitating convection of limited vertical extent. In this study we use a long-term data set of the large-scale and subgrid-scale state of the atmosphere derived using the standard variational analysis technique developed in the ARM program at the Darwin TWP-site. Using this long-term data set we first show that this location is representative for the earlier findings on the water-vapor rainfall relationship. We then demonstrate the main source of water vapor for the mid-troposphere prior to heavy rainfall is vertical advection by the mean flow, and not as is often postulated, moistening by non-precipitating convection. Our findings are consistent with the notion that heavy rainfall (and hence strong convection) requires significant dynamical “forcing” to occur.