Estimating land surface variables and sensitivity analysis for CLM and VIC simulations using remote sensing products

Area of Science:

• Earth and Environmental Sciences
• Atmospheric Sciences
• Hydrology

Background:

• Accurate land surface models (LSMs) are crucial for understanding hydrological and biophysical processes.
• Heterogeneous terrain and climate regimes present challenges for LSM parameterization and validation.

Purpose of the Study:

• To assess and compare the performance of the Community Land Model (CLM 4.0) and the Variable Infiltration Capacity (VIC-3L) model.
• To evaluate energy flux estimations (net radiation, sensible heat, latent heat, ground heat) at diverse geographical locations.

Main Methods:

• Utilized CLM 4.0 and VIC-3L to model land-atmosphere interactions.
• Analyzed energy fluxes at two contrasting sites: Freeman Ranch-2 (lowland Texas) and Providence 301 (mountainous Sierra Nevada).
• Compared model outputs with MOD16 satellite data and in situ measurements from 2003 to 2013.

Main Results:

• CLM underestimated net radiation at the high-elevation P301 site due to its snow scheme.
• VIC-3L showed a positive bias in latent heat flux during summer precipitation at P301, while CLM had a negative bias at FR2.
• Ground heat flux (G) performance was weaker in CLM (residual term) compared to VIC-3L at P301.
• MOD16 data showed lower correlation with in situ measurements at P301 compared to LSMs, especially in summer.
• Sensitivity analysis revealed negative trends between soil moisture and turbulent fluxes, particularly latent heat at P301, influenced by topography and snow cover.

Conclusions:

• VIC-3L demonstrated better performance in estimating energy fluxes compared to CLM 4.0 in heterogeneous terrain.
• Model discrepancies highlight the need for improved parameterization, especially for snow hydrology and ground heat flux in complex environments.
• Findings support enhancing LSMs and satellite-based estimations of water and energy fluxes in challenging high-elevation regions.