Ammonia oxidation and nitrate reduction marker genes are key indicators of nitrogen losses in temperate forest catchments.

Chronic nitrogen inputs can alleviate N limitation and potentially impose N losses in forests, indicated by soil enrichment in 15 N over 14 N. However, the complexity of the nitrogen cycle hinders accurate quantification of N fluxes. Simultaneously, soil ecologists are striving to find meaningful indicators to characterise the "openness" of the nitrogen cycle. We integrate soil δ 15 N with constrained ecosystem N losses and the functional gene potential of the soil microbiome in 14 temperate forest catchments. We show that N losses are associated with soil δ 15 N and that δ 15 N scales with the abundance of soil bacteria. The abundance of the archaeal amoA gene, representing the first step in nitrification (ammonia oxidation to nitrite), followed by the abundance of narG and napA genes, associated with the first step in denitrification (nitrate reduction to nitrite), explains most of the variability in soil δ 15 N. These genes are more informative than the denitrification genes nirS and nirK, which are directly linked to N 2 O production. Nitrite formation thus appears to be the critical step associated with N losses. Furthermore, we show that the genetic potential for ammonia oxidation and nitrate reduction is representative of forest soil 15 N enrichment and thus indicative of ecosystem N losses.