The initiation of the squall line was found to be a result of complicated multi-scale interaction. It occurred in an environment of a cold vortex, which is a typical synoptic weather system in north China featured of a quasi-stationary upper level cold-core low pressure center. Locating in front of a mid-level short wave embedded in the cold vortex and near a surface mesoscale convergence line, the CI location experienced a destabilized environment through synoptic-scale ascending, sufficient low-level moisture supply, apparent upper-level cooling and surface heating. The timing and location of the CI were found to be mainly determined by the mesoscale lifting associated with outflow boundaries of a pre-existing MCS during a collision process between such outflow boundaries and the surface convergence line. Operational practice shows that quite often severe convective storm occurs around the CI location of this event when there is an MCS at the edge of the plateau.
Detailed CI process was examined through three-dimensional radar and surface observation analyses. A total of 11 gust fronts were identified to propagate southward from the pre-existing MCS toward the CI region. The CI occurred within half an hour after the boundary collision near two vertices of the scalloped gust fronts (GFs) where there are usually maximum horizontal convergence and maximum upward motion. At these two vertices, two convective cells formed, then merged with each other forming a convective line in parallel to the convergence line with a well-defined three-dimensional structure of a squall line.
Numerical simulations using WRF-ARW with a convection-permitting grid spacing of 4.5 km were performed to further investigate the three dimensional kinematic and thermodynamic CI features at different scales. By digesting mesoscale information provided by intensive surface and rawinsonde observations at the initial time, both the synoptic and mesoscale environments, the CI process as well as the subsequent evolution of the squall line were successfully captured by the simulation, including the boundary collision process between the surface convergence line and GFs. Numerical results revealed that the pre-existing surface convergence line lifted warm and moist air to some altitude but not high enough to pass the LFC, resulting in an accumulation of warm and moist air slightly under the LFC. The intrusion of GFs pushed the convergence line southward at the surface and enhanced the lifting under the LFC, which finally lifted the accumulated warm and moist air passing the LFC and thus resulted in the CI. The main features of this simulated boundary collision process were quite consistent with those found in the observed CI process.