Insights into the Nitrogen Budget of Earth from investigation of the mantle, moon, core, and meteorites.

Abstract:

We present a new N budget for the Earth based on extensive literature review and compilation. The Bulk Silicate Earth (BSE) contains 11.6±6.7×10¹⁸ kg N, which is 2.9±1.7 times the present atmospheric mass of N (PAN). The continental crust (2.3±0.1×10¹⁸ kg N, 0.5 PAN) and the mantle (9.3±6.3×10¹⁸ kg N, 2.3 PAN) are significant N reservoirs.

In addition, the transition zone and lower mantle may sequester a much greater mass of N, due to the presence of Fe-metal, but the absence of samples precludes any robust estimates. We use the N content of lunar basalts to estimate the N content of the BSE after core segregation to be between 25±1×10¹⁸ to 118±4×10¹⁸ kg. Based on experimental results and comparison with iron meteorites, we suggest that the core itself is likely a major reservoir of N. Estimates for the N content of the core are 240±40×10¹⁸ to 420±40×10¹⁸ kg N. Our work supports the notion that N retention in the core could help explain the "missing" N dilemma. Assuming a chondritic source for the accretion of Earth and 10% N retention during accretion gives a closed, self-consistent budget.

Determining the current distribution of N on Earth is important, since many aspects of its long-term cycling between the surface and deeper Earth remain unresolved.

Surface N has clearly been subducted and cycled through the mantle, as revealed by correlation between N₂ and ⁴⁰Ar in mid-ocean ridge basalts. In contrast, N₂ does not correlate with primordial ³⁶Ar, strengthening the case for a surficial origin. The timing, rate, and magnitude of N movement to the deeper Earth is somewhat constrained for modern subduction zones, but fluxes in the past and the amount of N retained in the mantle in are more enigmatic.

N sequestration into the solid Earth has a direct impact on atmospheric evolution. It is possible that the atmosphere during the Archean contained up to 2-3x the present mass of N, though other evidence suggests the N content has been constant since that time. Any attempt to verify these states with a model would be greatly enhanced if a starting N content of the Earth could be obtained, which we have attempted to undertake herein.