Over tropical oceans precipitating clouds occur in ensembles in which phenomena ranging from shallow isolated cells to large mesoscale convective systems form, move, and dissipate in a bewildering variety of complex patterns. A fundamental part of understanding the behavior of oceanic tropical cloud ensembles is determining how the clouds regenerate themselves and propagate. Kwajalein Atoll in the Marshall Islands is a venue of concentrated study of precipitating tropical oceanic cloud ensembles because of its research quality S-band radar and the KWAJEX field campaign conducted there in 1999. In this study we are using high resolution numerical modeling to understand the initiation and propagation of convection in this environment.

Precipitating convective storms typically generate subcloud cold pools, and these pools often play a key role in storm maintenance and longevity by providing a lifting mechanism for warm, moist air. Although this ascent tends to be punctuated into individual convective cells, this mechanism has become known as continuous propagation, the continuity being provided by the cold pool's persistence. Yet, new cells are often seen to initiate ahead of established convection. When the new cells appear out ahead of the cold pool, a mechanism other than cold pool lifting must be operating. Some of these cells subsequently merge with the established convection while others remain entirely independent entities. In either case, the convection can be thought of as propagating in a discrete fashion.

The figure depicts a sequence of events from a rather typical simulation of tropical oceanic convection. The MM5 model was used at 4 km resolution with a domain centered on the Marshall Islands. The colored field is surface air temperature, the vertically integrated condensation field is contoured and the vectors represent model estimated 10 m winds. Only a small portion of the simulation domain is shown.

The first panel shows convection in the southwestern and eastern portions of the domain. These storms are moving towards the north and west, respectively. As highlighted on the second panel, some new clouds have appeared ahead of both systems during the next 30 min. These new clouds are far removed from any existing cold pool. The remaining two panels reveal that some of these cells die quickly while others continue to develop, some of those merging with the established convection and some remaining independent.

We propose to discuss how and why this kind of convective initiation and discrete propagation occurs over the tropical ocean. We will compare the behavior of the model convection with the radar observations obtained at Kwajalein to assess whether the model is faithfully simulating the fundamental characteristics of convection in that area. Preliminary examination of the radar data indicates that the radar echoes in this oceanic regime exhibit a lot of merging, suggesting that echoes form near to but isolated from other convective entities and subsequently incorporate with other echoes.