Confronting a multi-plume scheme with observations of continental shallow convection