Parameterization of Deciduous Vegetation Phenology for the Dynamic Land Model (DLM)

Abstract:

We developed a parameterization of deciduous vegetation phenology for the Dynamic Land surface Model (DLM), which integrated the Ecosystem-Atmosphere Simulation Scheme (EASS) model and the version 4 of Community Land Model (CLM) and has been coupled to the Community Earth System Model (CESM). In the phenological parameterization, we introduced the growing season index (GSI) model which combined the air temperature, day length and vapor pressure deficit simultaneously into an index to quantify the greenness of vegetation throughout the year and established the relationships between the GSI index and the carbon phenology fluxes. We tested this phenology scheme against observations focusing on two vegetation function types (deciduous forest and grass land). Preliminary results show that DLM has a better performance for simulating phenological events than CLM across the global eddy-covariance sites, and describes the deciduous vegetation seasonal changes more reasonably. As the same time, simulations for the other two important variables related with phenology closely, i.e., leaf area index (LAI) and gross primary productivity (GPP), by DLM are also in agreement with the observed and better than those by CLM. And, DLM can track the vegetation reversible nature of the spring recovery which is the shortcoming for CLM.