Embedding a novel phenology model into the Common Land Model for improving the modeling of land-atmosphere fluxes

Land surface models (LSMs) are essential tools for simulating the energy and momentum exchange between the land surface and the atmosphere. The vegetation phenology module, a pivotal component of LSMs, reflects the feedback of vegetation to climate variations. However, the vegetation phenology module in several LSMs, such as the Common Land Model (CoLM), face challenges concerning its rationality. Currently, the original satellite phenology mode in CoLM, which uses satellite-based leaf area index (LAI) data directly and statically, constrains the model's ability to accurately represent complex and dynamic ecosystem processes. In this study, we embedded a dynamic vegetation phenology model into CoLM, serving as a novel phenology module. We conducted an extensive comparison of the effects of various phenology modules on CoLM, focusing on their performance in the modeling of gross primary productivity (GPP) and evapotranspiration (ET) at both site and global scales. Results demonstrate an improvement in CoLM for modeling GPP and ET following the integration of the novel phenology module. Specifically, the Pearson's correlation coefficient for GPP increased by 1.29 % to 33.74 %, and for ET by 0.32 % to 7.47 % on a site scale. Furthermore, on a global scale, when compared with global reference datasets, the CoLM embedded with the novel phenology module successfully captures the spatiotemporal distribution of GPP and ET, and its performance is closer to the reference data than that of the CoLM with the original satellite phenology mode. This study highlights the importance of vegetation phenology in LSMs and its potential to improve the modeling of dynamic interactions between vegetation processes and environmental changes within LSMs.