The dynamics of the marine nitrogen cycle across the last deglaciation

[1] We use a geochemical box model to investigate the changes in marine N-fixation and denitrification required to match the observed sedimentary ı15N changes between 30 kyr B.P. and the late Holocene. This is achieved by optimizing a set of seven parameters that describe the strengths of three ocean-internal N feedbacks and the response of the oceanic N cycle to four external forcings. Scenarios that best match the ı15N constraints indicate a strong transient decrease in N-fixation in the early deglacial in response to the decrease in iron input by dust. Around 15 kyr B.P., N-fixation rebounds primarily in response to an abrupt increase in water column denitrification caused by an expansion of anoxia. Benthic denitrification is not well constrained by our model but tends to increase in sync with water column denitrification. As a result of the transient imbalance between N-fixation and denitrification, we infer a glacial-to-interglacial decrease in the marine N inventory of between 15 and 50%. The model diagnoses this reduction in order to simultaneously fit the data from all ocean basins, requiring it to reduce the degree by which water column denitrification in the oxygen minimum zones is influencing the ı15N of nitrate of the whole ocean (dilution effect). Our optimal solution suggests a glacial N cycle that operated at nearly the same rates as that in pre-industrial times, but sensitivity cases with substantially lower rates fit the data only marginally worse. An important caveat of our study is the assumption of an unchanging ocean circulation. An initial sensitivity experiment shows that this affects primarily the magnitude of the change in the N inventory, while the diagnosed deglacial dynamics with global marine N-fixation taking a dip before the onset of denitrification remains a robust result.