Dynamic nitrogen resorption improves predictions of nutrient cycling responses to global change in a next generation ecosystem model

Nutrient resorption from senescing leaves is a critical process of plant nutrient cycling that can significantly affect plant nutrient status and growth, making it essential for land surface models (LSMs) in order to predict long-term primary productivity. Most models assume leaf resorption to be a fixed value of 50% for nitrogen (N) partially because we lack the knowledge of what drives this process, making it unclear what its implications are when simulating nutrient cycling. Based on our own analysis of global patterns of nutrient resorption from trait data (Sophia et al., 2024), we developed a dynamic scheme of N resorption driven by leaf structure and environmental limitation and implemented it in the QUINCY model. This scheme assumes that all metabolic N is fully mobilizable and available for resorption, representing the maximum resorption capacity for each plant functional type based on the leaf construction costs. Environmental limitations then downregulate the remaining mobilizable nutrients considering soil N availability relative to plant demand, adjusting their internal recycling in face of N stress and leaf C:N ratio. The model performance was validated by comparing the model's predicted values of N resorption against observational data analyzed in Sophia et al., 2024, using spatial-scale measurements of resorption efficiency across diverse plant types and climate zones, as well as gross primary productivity (GPP) observational data from plumber sites (Ukkola, et al., 2022) used for model application. We present the implications of this novel scheme for ecosystem functioning and show that we can improve the plant internal N available to growth with cascading implications for ecosystem nutrient pools and fluxes, better predicting plant and soil nutrient dynamics at steady state and crucially, under elevated CO2 conditions. For the first time, we show the importance of an ecologically realistic representation of nutrient resorption in an LSM and its implication for predicting the future of the terrestrial biosphere.