Challenging a Global Land Surface Model in a Local Socio-Environmental System

Area of Science:

• Earth System Science
• Terrestrial Ecosystem Modeling
• Hydrology and Biogeochemistry

Background:

• Land surface models (LSMs) are crucial for simulating Earth's climate system.
• Accurate representation of land surface processes is essential for predicting carbon, water, and energy fluxes.

Purpose of the Study:

• To evaluate the performance of the Community Land Model (CLM) against observational data at a single grid cell scale.
• To identify key areas for improving LSMs in human-dominated watersheds.

Main Methods:

• Comparison of CLM outputs with satellite and ground-based observational data for temperature, precipitation, land cover, leaf area index (LAI), greenness, productivity, soil moisture, and albedo.
• Analysis of model inputs and process simulations within a human-dominated watershed in Michigan, USA.

Main Results:

• CLM shows moderate to strong correlations for temperature and precipitation but significant mismatches in land cover distribution and phenology (LAI).
• Peak LAI in CLM was nearly double satellite estimates, indicating issues with seasonal vegetation dynamics.
• Simulated greenness and productivity showed better agreement, while soil moisture timing matched but not magnitude.
• Albedo correlations were significant in winter but not in summer.

Conclusions:

• CLM requires substantial improvements in land cover representation, phenology algorithms, and summertime radiative transfer modeling.
• Addressing plant stress responses is also critical for enhancing LSM accuracy in complex environments.
• Model enhancements are vital for reliable Earth system modeling and climate change projections.