A dynamic-flow carbon-cycle box model and high-latitude sensitivity

Most of the hypotheses put forward to explain glacial-interglacial cycles in atmospheric pCO₂ are centred on Southern-Ocean-based mechanisms. This is in large part because: (1) timing constraints rule out changes in the North Atlantic as the trigger; (2) the concept of "high-latitude sensitivity" eliminates changes in the non-polar oceans as likely contenders. Many of the Southern-Ocean-based mechanisms for changing atmospheric pCO₂ on glacial-interglacial time-scales are based on results from highly simplified box models with prescribed flow fields and fixed particulate flux. It has been argued that box models are significantly more "high-latitude sensitive" than General Circulation Models. In light of this, it is important to understand whether this high-latitude sensitivity is a feature common to all box models, and whether the apparent degree of sensitivity changes for different tracers and parameters. We introduce a new metric for assessing how "high-latitude sensitive" a particular solution is to perturbations. With this metric, we demonstrate that a given model may be high-latitude sensitive to certain parameters but not to others. We find that the incorporation of a dynamic-based flow field and a Michaelis-Menten type nutrient feedback can have a significant impact on the apparent sensitivity of the model to perturbations. The implications of this for current box-model-based estimates of atmospheric pCO₂ drawdown are discussed.