Aridity threshold for alpine soil nitrogen isotope signature and ecosystem nitrogen cycling.

Determination of tipping points in nitrogen (N) isotope (δ 15 N) natural abundance, especially soil δ 15 N, with increasing aridity, is critical for estimating N-cycling dynamics and N limitation in terrestrial ecosystems. However, whether there are linear or nonlinear responses of soil δ 15 N to increases in aridity and if these responses correspond well with soil N cycling remains largely unknown. In this study, we investigated soil δ 15 N and soil N-cycling characteristics in both topsoil and subsoil layers along a drought gradient across a 3000-km transect of drylands on the Qinghai-Tibetan Plateau. We found that the effect of increasing aridity on soil δ 15 N values shifted from negative to positive with thresholds at aridity index (AI) = 0.27 and 0.29 for the topsoil and subsoil, respectively, although soil N pools and N transformation rates linearly decreased with increasing aridity in both soil layers. Furthermore, we identified markedly different correlations between soil δ 15 N and soil N-cycling traits above and below the AI thresholds (0.27 and 0.29 for topsoil and subsoil, respectively). Specifically, in wetter regions, soil δ 15 N positively correlated with most soil N-cycling traits, suggesting that high soil δ 15 N may result from the "openness" of soil N cycling. Conversely, in drier regions, soil δ 15 N showed insignificant relationships with soil N-cycling traits and correlated well with factors, such as soil-available phosphorus and foliage δ 15 N, demonstrating that pathways other than typical soil N cycling may dominate soil δ 15 N under drier conditions. Overall, these results highlight that different ecosystem N-cycling processes may drive soil δ 15 N along the aridity gradient, broadening our understanding of N cycling as indicated by soil δ 15 N under changing drought regimes. The aridity threshold of soil δ 15 N should be considered in terrestrial N-cycling models when incorporating 15 N isotope signals to predict N cycling and availability under climatic dryness.