- The spatial morphology of the mei-yu front and the location and nature of the precipitation evolve simultaneously in a coherent way. At the beginning (stage 1), the mei-yu front appeared as a quasi-zonal belt, and banded rainfall occurred on the front. Subsequently in stage 2, weakening and bending of the mei-yu front leads to heavy rainfall in the warm sector south of the mei-yu front. With the warm sector rainfall, the mei-yu front eventually resumed the original belt shape and the rainband re-emerged in stage 3.
- In the stages (stage 1 and 3) of frontal rainfall, water vapor convergence occurred due to the wind field difference between the north and south of the mei-yu front. Warm, humid air in the lower troposphere rises along moist isentropes in the symmetric instability region near the front, and is further accelerated after entering convectively unstable regions in the midtroposphere. In the warm sector rainfall stage (stage 2), the mei-yu front weakened and deformed with the water vapor convergence zone moving to the exit of the LLJ and the precipitation echoes developing mainly in the warm sector of deep convective instability south of the curved mei-yu front.
- During the frontal precipitation stages, the mei-yu frontogenesis is dominated by FG1a (diabatic heating) and the frontolysis by FG4 (tilting). The overall effect is frontolytical and is responsible for the weakening and bending of the mei-yu front. In the warm sector rainfall stage, FG1b (moisture depletion) and FG4 are frontolytical in the mei-yu front area. However, diabatic heating and tilting are strongly frontogenetic in the precipitation south of the mei-yu front, where a belt-shaped front was restored and a rainband reappeared.
The observations presented here suggest a symbiotic relationship between a mei-yu front and its associated rainfall. When the front is strong, steady, and in a banded shape, its dynamical (horizontal wind shear) and thermodynamic (large temperature/moisture gradient) structures create zones that are symmetrically unstable, driving the sloping ascent of warm humid air. This updraft is further intensified after it enters midtropospheric zones of convective instability, ultimately generating a rainband nearly parallel to the mei-yu front. The banded rainfall weakens and bends the front through (direct) diabatic and (indirect) tilting effects, and the rainfall subsequently moves to the warm sector south of the front in a zone of pronounced convective instability. Diabatic and tilting effects are both strongly frontogenetic in the region of warm sector rainfall, later creating a banded front and thus banded rainfall again. This completes a cycle characterizing the coupling between the mei-yu front and its associated rainfall. By going through all the persistent mei-yu events in the same area during the period from 2010 to 2020, six new cases exhibiting the same cyclic behaviors are identified. Similar to the case in 2020, the morphology of the mei-yu front and the location and shape of the precipitation in these cases coevolved in a symbiotic way.
The rainfall–front cyclic behavior outlined here suggests potential, synoptic-scale periodic oscillations in mei-yu rainfall. The intrinsic connection between such oscillations and the mei-yu front indicates the necessity of accurate model representations of the frontal morphology and its evolution in any effort that aims to improve the short-term forecast skill of mei-yu rainfall. The importance of cloud microphysics cannot be overemphasized due to its essential role in determining the diabatic heating that drives frontogenesis in both direct and indirect ways.

