AbstractTraditionally, it is assumed that contributions of crust‐derived CO2 emissions at arc settings are minor, but this requires well‐constrained field studies to determine the extent of magma‐carbonate reaction and concomitant C released. The Jurassic Bonanza arc on Vancouver Island (Canada) was built on a Triassic limestone platform and makes for an ideal setting to examine arc magma‐carbonate interactions beneath an island arc. We examine how km‐scale magma bodies might react with carbonate from a well‐constrained study of a gabbro pluton that intrudes limestone. The pluton shows muted to nil carbonate interaction preserving primary igneous 87Sr/86Sr values (∼0.703) except for a thin (<2 m wide) marginal gabbro (∼0.706) in contact with a decarbonized metamorphic aureole. A lack of 87Sr/86Sr or δ18Ocpx enrichment in gabbro ∼10 to >1,000 m from the contact suggests that any reaction with limestone wallrock is limited to at most the outer ∼10 m of the pluton. More enhanced magma‐carbonate interaction and CO2 production occurs via a network of shallow dikes and sills (<0.2 GPa) compared to deeper plutons, consistent with experimental data showing increased carbonate assimilation at shallower depths (≤0.5 GPa). Plutons are an important heat source to release CO2 from carbonate wallrock by contact metamorphism, where >89% of crust‐derived CO2 is liberated by wallrock decarbonation and <11% is liberated by magma that assimilated limestone. Nonetheless, we show that neither magmatic nor metamorphic crust‐derived CO2 contributions compare to mantle‐derived CO2. Our study places realistic and quantitative limits on arc‐derived CO2 from upper crustal limestone sources.