In a mesoscale convective system (MCS), regions of convection that redevelop over (i.e. back-build), and/or repeatedly pass (i.e. train) over a particular region for an extended period of time can contribute to extreme rainfall and flash flooding. Past studies have indicated that both mesoscale ascent and gradual lifting of the inflow layer by a cold pool or bore are important when this back-building/training convection is displaced from the leading line (sometimes called rearward off-boundary development or ROD). However, Plains Elevated Convection at Night (PECAN) field campaign observations suggest that stability of the nocturnal boundary layer is highly variable and some MCSs with ROD have little indication of a surface cold pool. The MCS simulated in an environment with strong low-level stability presented in this work suggests that it is possible for ROD to be supported by mechanisms other than those mentioned above. 

Simulations were initialized using an environmental sounding with a strong stable layer from a PECAN MCS case with both forward propagation and ROD. When a single warm bubble was used to initiate convection, the simulation produced an MCS remarkably like that observed despite the homogeneous initial conditions. Analysis of this simulation leads to a number of findings that challenge our current understanding of nocturnal MCSs, and in particular the interaction of downdrafts with a stable boundary layer; Notably, downdrafts can reach the surface, and different mechanisms support different regions of the MCS (even along the leading line), which influences the relevant inflow layer. Unlike previous studies of ROD, parcel lifting may be due to the downdraft air itself in the form of an intrusion (essentially an elevated cold pool), favorably amplified gravity waves, or some combination of the two.