Bonan et al. (2025) Beyond surface fluxes: Observational and computational needs of multilayer canopy models – A walnut orchard test case

Identification

• Journal: Agricultural and Forest Meteorology
• Year: 2025
• Date: 2025-12-02
• Authors: Gordon B. Bonan, Sean P. Burns, Edward G. Patton
• DOI: 10.1016/j.agrformet.2025.110960

Research Groups

• NSF National Center for Atmospheric Research (NCAR), Boulder, CO, USA
• Department of Geography, University of Colorado, Boulder, CO, USA

Short Summary

This study evaluates the Community Land Model's multilayer canopy model (CLM-ml v2) against comprehensive, multi-level observations from a walnut orchard, demonstrating its strong performance for most atmospheric conditions but highlighting limitations in simulating within-canopy mixing during strongly stable regimes.

Objective

• To compare the Community Land Model's multilayer canopy model (CLM-ml v2) with high-resolution observations of air temperature, specific humidity, wind speed, and turbulent fluxes at multiple heights within and above a walnut orchard.
• To identify the strengths and weaknesses of CLM-ml v2 across various atmospheric stability regimes.
• To provide guidance on observational needs and parameter estimation for advancing multilayer canopy models.

Study Configuration

• Spatial Scale: A walnut orchard (Juglans regia ‘Chandler’) in Dixon, California, with an average canopy height of approximately 10 meters. Measurements were taken from a 30-meter tower with 13 levels within and above the canopy.
• Temporal Scale: Observations from the Canopy Horizontal Array Turbulence Study (CHATS) field campaign, specifically focusing on May 2007 (when leaves were fully emerged). Data were processed into 30-minute average statistics from 5-minute average data.

Methodology and Data

• Models used:
  • Community Land Model (CLM) multilayer canopy model (CLM-ml v2) for primary simulations.
  • CLM5 for generating initial soil moisture and temperature profiles.
• Data sources:
  • Canopy Horizontal Array Turbulence Study (CHATS) field campaign (May 2007, Dixon, California).
  • 30-meter tower measurements: air temperature, specific humidity, horizontal wind speed, net radiation, sensible heat flux, latent heat flux, and momentum flux at multiple heights.
  • Ancillary observations: incoming and outgoing shortwave and longwave radiation, precipitation, barometric pressure, soil moisture (at 5 cm depth), and soil heat flux (at 5 cm depth).
  • NSF NCAR Earth Observing Laboratory (EOL) high-rate CHATS data (version 2) for 5-minute statistics.
  • Global Soil Wetness Project Phase 3 for CLM5 spin-up meteorology.
  • California Irrigation Management Information System (CIMIS) for precipitation gap-filling.
  • Literature review for walnut physiology parameters (e.g., Vcmax25, marginal water-use efficiency, leaf water potential, hydraulic conductance, vegetation optics).

Main Results

• CLM-ml v2 accurately simulates above-canopy net radiation, sensible heat flux, latent heat flux, and friction velocity across a range of atmospheric stability regimes (strongly unstable to strongly stable).
• Vertical profiles of wind speed within and above the canopy closely match observations under all stability regimes.
• Vertical profiles of air temperature and specific humidity are well simulated, except during strongly stable conditions.
• Under strongly stable conditions, the model's first-order turbulence closure cannot represent non-local vertical mixing within the canopy, leading to a significant warm bias in near-ground air temperature and large gradients in temperature and specific humidity.
• The model successfully replicates the observed displacement height and its flow-dependent variation.
• Heat storage in the canopy air space and leaves is well reproduced by the model.
• Numerical improvements in CLM-ml v2 (fourth-order Runge–Kutta method with 5-minute meteorological forcing interpolation) significantly smooth simulated fluxes and reduce the need for ad-hoc numerical fixers present in CLM5/6.
• Model sensitivity analysis indicates that stomatal conductance parameters (maximum carboxylation rate, marginal water-use efficiency, leaf water potential at 50% conductance) and stem hydraulic conductance are critical for accurate daytime flux and temperature simulations. Revised vegetation optical properties also improve model performance.

Contributions

• Provides a rigorous, multi-level evaluation of a multilayer canopy model (CLM-ml v2) against comprehensive field observations, establishing a benchmark for future model development.
• Highlights the critical need for detailed within-canopy micrometeorological measurements (air temperature, humidity, wind speed, fluxes) and leaf-level gas exchange data to constrain and improve multilayer models.
• Demonstrates the capability of multilayer models to simulate the surface air space, extending the scientific scope of land surface models beyond simple flux boundary conditions to explicitly represent forest microclimates.
• Shows that CLM-ml v2 offers a more physically consistent and theoretically sound approach to canopy physics, reducing reliance on ad-hoc parameterizations found in traditional big-leaf models like CLM5/6.
• Identifies a key deficiency in current first-order turbulence closure schemes for multilayer models, specifically their inability to capture non-local mixing under strongly stable atmospheric conditions, pointing to a crucial area for future research and model advancement.

Funding

• NSF National Center for Atmospheric Research (NCAR), United States (Cooperative Agreement No. 1852977).

Citation

@article{Bonan2025Beyond,
  author = {Bonan, Gordon B. and Burns, Sean P. and Patton, Edward G.},
  title = {Beyond surface fluxes: Observational and computational needs of multilayer canopy models – A walnut orchard test case},
  journal = {Agricultural and Forest Meteorology},
  year = {2025},
  doi = {10.1016/j.agrformet.2025.110960},
  url = {https://doi.org/10.1016/j.agrformet.2025.110960}
}