Implementation of a dynamic rooting depth and phenology into a land surface model: Evaluation of carbon, water, and energy fluxes in the high latitude ecosystems

Abstract

Recent studies and observations have shown that northern high latitude ecosystems (NHLE) are strongly responsive to environmental changes, particularly warming temperature. Ecosystem models are important tools that help us to understand and assess the impact of environmental changes in the NHLE. However, models lack processes that are essential for modeling ecosystem dynamics for the NHLEs. In this study, NHLE-specific dynamic phenology and dynamic rooting distribution and depth parameterizations was implemented in a land surface model, the Integrated Science Assessment Model (ISAM), to improve the estimated carbon, water, and energy fluxes in the NHLs. These parameterizations account for light, water, and nutrient stresses while allocating the assimilated carbon to leaf, stem, and root pools. The model parameters related to these processes were calibrated and evaluated using measured data from 16 sites (12 fluxnet sites and 4 non-flux net sites) representative of the dominant NHLEs. By including these dynamic processes, ISAM was able to capture the measured seasonal variability in leaf area index (LAI) and root distribution in the soil layers. The evaluation of the model results suggested that without including the dynamic processes, the modeled growing season length (GSL) in the NHLE was almost two times higher, as compared to measurements. To quantify the implication of these processes on the C, water, and energy fluxes, we compared the results of two different versions of ISAM, a dynamic version that includes dynamic processes (ISAMDYN) and a static version that does not include dynamic processes (ISAMBC), with measurements from 12 eddy covariance flux sites. The results showed that ISAMDYN, unlike ISAMBC, was more capable to capture the flux site-based seasonal variability in GPP, water, and energy fluxes. Regional analysis revealed that the growing season length increased on average by about 5 days in the NHLs in the 2000s compared to 1990s.