The formation of large-scale upwellings with lateral extents of several hundreds of kilometers, reaching up to $10,000 km or more, still remains a hotly debated topic. Some seismic imaging studies based on high-resolution data suggest that the main superplumes underneath Africa and South-central Pacific are clusters, composed of several individual plumes rather than being a single large mantle upwelling. The iron spin transition in the lower mantle minerals may present a new idea on the origin and the formation of such superplumes, notably sourcing such features in the midmantle. Stagnation of both cold sinking slabs and hot rising plumes can be caused by density and viscosity variation due to the spin transition in iron in ferropericlase (Fp) and a possible spin-dependent bulk modulus hardening in bridgmanite silicate perovskite (Pv). This process produces intermittent downward spin transition-induced midmantle avalanches (SIMMA) of the cold sinking flow as well as upward spin transition-induced midmantle superplume avalanches (SIMMSA) of the rising hot plumes, triggered at the spin transitioninduced thermal boundary layer at around 1600 km depth. Our high-resolution axi-symmetric models reveal that the hot upwellings, trapped below $1600 km depth, can suddenly penetrate into the upper levels in the mantle and spread laterally for hundreds of kilometres. Owing to the upward penetration of the midmantle-rooted superplumes, as broad as $1500 km across, a large amount of heat can be delivered to the upper mantle and base of the lithosphere with implications for large volcanic episodes.