Skip to main content
Dryad

Code for: Global carbon investment in terrestrial biological nitrogen fixation

Menge, Duncan1

Published Jul 27, 2026 on Dryad. https://doi.org/10.5061/dryad.tqjq2bwbn

Data files

Jul 27, 2026 version files 366.44 KB

Click names to download individual files

Abstract

Code that reads in a variety of data sources and analyzes them for the article with the following abstract. Biological nitrogen (N) fixation (BNF) provides the N needed to produce proteins and other biological building blocks, helping feed humanity and mitigate climate change. Due to its high energetic cost compared to other forms of N acquisition, investment in BNF indicates N limitation. Globally gridded BNF flux data provide an opportunity to determine the energetic investment in BNF across ecosystems, which would help reconcile conflicting indicators of N limitation. Here, we use a new BNF synthesis to quantify the relative importance of BNF in different N-fixing niches, in different biomes, and across the globe by calculating the fraction of net primary productivity (NPP) invested in BNF and the fraction of plant N acquisition provided by BNF. Larger fractions of non-agricultural NPP were invested in BNF in less-productive, higher-latitude biomes. This pattern was driven by biocrusts and mosses. Similarly, non-agricultural symbiotic N-fixing plants invested relatively more of their own NPP in BNF in less-productive, higher-latitude biomes. This symbiotic plant pattern was driven by shrubs and herbs, overriding the opposite pattern in trees. Symbiotic plants also acquired a higher fraction of their N from BNF at higher latitudes and in less productive biomes. Investments in symbiotic BNF were 10x higher in agricultural (2.9% of NPP) vs. natural (0.29%) biomes, providing 26% vs. 3.1% of ecosystem-scale plant N acquisition. These results support the paradigm of strong N limitation at higher latitudes, help understand the rarity of N-fixing trees at higher latitudes, underscore the dominance of human activity, and inform models.