A stochastic parameterization for deep convection using cellular automata

A stochastic parameterization for deep convection using cellular automata
Technical memorandum
Date Published
Secondary Title
ECMWF Technical Memoranda
L. Bengtsson
M. Steinheimer
J.-F. Geleyn
Abstract A cellular automaton (CA) is introduced to the deep convection parameterization of two NWP models; the global ECMWF IFS, and the high resolution limited area model ALARO. The selforganizational characteristics of the CA allows for lateral communication between adjacent NWP model grid-boxes, and adds additional memory to the deep convection scheme(s). The CA acts in two dimensions, with finer grid-spacing than that of the NWP model. It is randomly seeded in regions where CAPE, CIN or already active convection exceeds a threshold value. Both deterministic and probabilistic rules are explored to evolve the CA in time, and the resulting CA field can be advected with the mid-tropospheric wind on the sub-grid level. Case-studies indicate that the scheme has potential to organize cells along convective squall-lines, and enhance advective effects. An ensemble of forecasts using the present CA scheme demonstrated an ensemble spread in the resolved wind-field, in regions where deep convection is large. Such a spread represents the uncertainty due to sub-grid variability of deep convection, and could be an interesting addition to an ensemble prediction system.
URL https://www.ecmwf.int/en/elibrary/73666-stochastic-parameterization-deep-convection-using-cellular-automata