North Atlantic Oscillation modulates long-term ANPP dynamics via precipitation or temperature, depending on soil nutrient levels

Abstract

The North Atlantic Oscillation (NAO) is a dominant mode of atmospheric variability in the North Atlantic region, influencing winter and spring precipitation and temperature across Europe. These seasonal variations strongly control interannual dynamics in Aboveground Net Primary Production (ANPP). However, long-term fertilization (>100 years) may interact with these drivers and alter their influence on ecosystem functioning. To investigate this, we analysed a long-term dataset statistically to: (1) describe how hierarchically structured factors—global NAO and local precipitation and temperature—affect interannual variation in ANPP, and (2) assess how fertil­ization modifies the pathways through which climate influences ANPP. Data were taken from plots with ‘high’ and ‘low’ nutrient addition at the world’s longest ecological experiment; the ‘Park Grass’ experiment (Rothamsted Research, England, UK; analysis period: 1950− 2018). We used a structural equation model to evaluate relationships between ANPP and interannual climate variations, testing how monthly NAO changes affected precipitation and temperature, and how these, in turn, influenced ANPP. We detected a clear NAO signal on ANPP, but its influence varied by nutrient level. Under low nutrient addition, NAO effects were consistently mediated through precipitation (R² = 0.38, p < 0.001). In contrast, under high nutrient addition, the NAO had a weaker effect (R² = 0.17, p < 0.001), acting through April precipitation and May temperature. These results clarify how global atmospheric patterns influence local ecosystem functioning, revealing a shift from a precipitation-driven to temperature responses under ‘high’ nutrient conditions. These contrasting patterns sug­gest there is no simple way to explain the mechanisms by which global atmospheric patterns influence ecosystem functioning. They also reveal that long-term fertilization alters community composition and stability in ways that may differ from those observed in shorter-term experiments, with concomitant implications for interpreting ecosystem responses to human drivers.