Marine phytoplankton stoichiometry links nutrient supply to marine carbon export. Deviations of phytoplankton stoichiometry from Redfield proportions (106C : 1P) could therefore have a significant impact on carbon cycling, and understanding which environmental factors drive these deviations may reveal new mechanisms regulating the carbon cycle. To explore the links between environmental conditions, stoichiometry, and carbon cycling, we compared four different models of phytoplankton C: P: a fixed Redfield model, a model with C: P given as a function of surface phosphorus concentration (P), a model with C: P given as a function of temperature, and a new multi-environmental model that predicts C: P as a function of light, temperature, and P. These stoichiometric models were embedded into a fivebox ocean circulation model, which resolves the three major ocean biomes (high-latitude, subtropical gyres, and tropical upwelling regions). Contrary to the expectation of a monotonic relationship between surface nutrient drawdown and carbon export, we found that lateral nutrient transport from lower C: P tropical waters to high C: P subtropical waters could cause carbon export to decrease with increased tropical nutrient utilization. It has been hypothesized that a positive feedback between temperature and pCO2,atm will play an important role in anthropogenic climate change, with changes in the biological pump playing at most a secondary role. Here we show that environmentally driven shifts in stoichiometry make the biological pump more influential, and may reverse the expected positive relationship between temperature and pCO2,atm. In the temperature-only model, changes in tropical temperature have more impact on the Δ pCO2,atm (~41 ppm) compared to subtropical temperature changes (~4.5 ppm). Our multi-environmental model predicted a decline in pCO2,atm of ~46 ppm when temperature spanned a change of 10 °C. Thus, we find that variation in marine phytoplankton stoichiometry and its environmental controlling factors can lead to nonlinear controls on pCO2,atm, suggesting the need for further studies of ocean C: P and the impact on ocean carbon cycling. [ABSTRACT FROM AUTHOR]