Abstract. During 11-12 June 2012, quasistationary linear mesoscale convective systems (MCSs) developed near northern Taiwan and produced extreme rainfall up to 510 mm and severe flooding in Taipei. In the midst of background forcing of low-level convergence, the back-building (BB) process in these MCSs contributed to the extreme rainfall and thus is investigated using a cloud-resolving model in the case study here. Specifically, as the cold pool mechanism is not responsible for the triggering of new BB cells in this subtropical event during the meiyu season, we seek answers to the question why the location about 15-30 km upstream from the old cell is still often more favorable for new cell initiation than other places in the MCS.With a horizontal grid size of 1.5 km, the linear MCS and the BB process in this case are successfully reproduced, and the latter is found to be influenced more by the thermodynamic and less by dynamic effects based on a detailed analysis of convective-scale pressure perturbations. During initiation in a background with convective instability and near-surface convergence, new cells are associated with positive (negative) buoyancy below (above) due to latent heating (adiabatic cooling), which represents a gradual destabilization. At the beginning, the new development is close to the old convection, which provides stronger warming below and additional cooling at mid-levels from evaporation of condensates in the downdraft at the rear flank, thus yielding a more rapid destabilization. This enhanced upward decrease in buoyancy at low levels eventually creates an upward perturbation pressure gradient force to drive further development along with the positive buoyancy itself. After the new cell has gained sufficient strength, the old cell's rear-flank downdraft also acts to separate the new cell to about 20 km upstream. Therefore, the advantages of the location in the BB process can be explained even without the lifting at the leading edge of the cold outflow.

Abstract. During 11–12 June 2012, quasi-stationary linear mesoscale convective systems (MCSs) developed near northern Taiwan and produced extreme rainfall up to 510 mm and severe flooding in Taipei. Evident back-building (BB) process in these MCSs contributed to the extreme rainfall, and thus is investigated using a cloud-resolving model. Specifically, we seek answers to the question why the location about 15–30 km upstream from the old cell is often more favorable for new cell initiation without the cold pool mechanism in this subtropical event during the mei-yu season. With a horizontal grid size of 1.5 km, the model successfully reproduced the linear MCS and the BB process, which is found to be influenced by both dynamical and thermodynamical effects. During initiation in a background with convective instability, new cells are associated with positive (negative) buoyancy below (above) due to latent heating (adiabatic cooling), which represent a gradual destabilization. At the beginning, the new development is close to the old convection, which provides stronger warming below and additional cooling at mid-levels from evaporation of condensates, thus yielding a more rapid destabilization. This enhanced upward decrease in buoyancy at a lower height eventually creates an upward perturbation pressure gradient force to drive further development along with the buoyancy itself. After the new cell has gain sufficient strength, a descending branch at the old cell's rear flank acts to separate the new cell to about 20 km upstream. Therefore, the advantages of the spot in the BB process can be explained.