The content of the ARM website is available to any browser, but for the best experience we highly recommend you upgrade to a standards-compliant browser such as Firefox, Opera or Safari.
VIEW CART
primary link menu HOME SITE INDEX PEOPLE
skip to main content ABOUT ARMABOUT ACRFSCIENCESITESINSTRUMENTSMEASUREMENTSDATAPUBLICATIONSEDUCATIONFORMS

Cover image

How Tightly Does the Large-scale State Control the Fate of Convection?

Robert Pincus NOAA-CIRES Climate Diagnostics Center
Robert J. Hofmann University of Colorado and NOAA/ESRL/PSD

Category: Modeling

The parameterizations of convection in large-scale models are determinisitic: given a profile of temperature and humidity, the parameterizations compute the fate of convection (typically in terms of a mass flux profile) and its effects on the large-scale environment. In particular, each convective scheme has no "memory," so the convection at each time depends only on the instantaneous thermodynamic profile. These assumptions are thought to be one reason why the diurnal cycle of oceanic convection is so poorly reproduced in large-scalemodels. Here we explore the degree to which transport by convection may vary under similar environmental conditions. We simulate tropical convection at the ARM Climate Research Facility (ACRF) Tropical Western Pacific (TWP)sites using a cloud-resolving model. We make an ensemble of simulations, eachmember of which is forced by slightly different analyses produced by one of four assimilation systems. Each simulation encompasses an area roughly comensurate with a grid column within a large-scale model. At each snapshot in time, we identify the convective columns within the domain based on their vertical velocity, then calculate the amount of convective mass transport over each 20-minute time interval (corresponding to the time step of a global model). We stratify these estimates according to the variables often used to enclose ssumptions for convection schemes, including the convectively available potential energy and the convective inhibition. The results may suggest ways in which stochastic noise may be introduced to onvection parameterizations to break the exitising lock-step relationships.

This poster will be displayed at ARM Science Team Meeting.

POSTER in PDF: for proper viewing, it should be viewed with Adobe Acrobat Reader. Download the latest version from the Adobe Reader website.