We experimentally investigated the response of phytoplankton and understory macroalgae to canopies of giant kelp, Macrocystis pyrifera, by following changes in their biomass and net primary production over 17 months in 600-m 2 plots where giant kelp was continually removed or left intact and allowed to vary naturally. Production by phytoplankton was two times greater and understory algae five times greater where Macrocystis was removed relative to the intact forest. Understory biomass, but not phytoplankton biomass, was suppressed inside the forest, leading to a higher magnitude of effect on net primary production (NPP) by understory relative to phytoplankton. Following a natural decline of the Macrocystis canopy by winter storms, understory macroalgae and phytoplankton increased production in the Macrocystis control plot. This response was delayed for both groups, with phytoplankton production increasing in spring and understory increasing later during summer. The longer delay for understory macroalgae was likely due to restrictions in the timing of macroalgal recruitment and their slower growth rates compared with phytoplankton and with increased competition for light resulting from greater light absorption by the spring phytoplankton bloom. Surprisingly, total ecosystem production that included NPP by Macrocystis, phytoplankton, and understory algae did not differ between the Macrocystis control and removal plots for much of the study. NPP by understory algae, which comprised the bulk of the ecosystem NPP in the Macrocystis removal plot, can compensate to varying degrees for the loss of Macrocystis production following its removal by winter storms.