Abstract
Leaf optical properties (LOPs), including leaf reflectance and transmittance, are fundamental parameters in canopy radiative transfer processes. Recent studies indicate significant discrepancies between prescribed LOP values in current land surface models (LSMs) and measured data, particularly in the near-infrared (NIR) band. This study evaluates the impact of these discrepancies by running the Community Land Model (CLM5) and the Common Land Model (CoLM2014) with both default and measured LOP values. Results indicate that using measured LOPs significantly increases surface albedo in the NIR band by over 10% in regions like boreal forests, Southeast China, and the eastern United States during summer. This adjustment helps correct the negative albedo bias historically observed in boreal forests. Furthermore, incorporating measured LOPs reduces simulated net radiation and sensible heat flux by up to 11 W/m², emphasizing the importance of accurate LOP parameters for surface radiation budget simulations.
Introduction
Land Surface Models (LSMs) are the primary tools for simulating the exchange of energy, water, and carbon between the Earth's surface and the atmosphere. A critical component of these models is the representation of Leaf Optical Properties (LOPs), which dictate how solar radiation is reflected, transmitted, or absorbed by vegetation. Despite their importance, many modern LSMs—including CLM and JULES—continue to rely on prescribed LOP values derived from the Simple Biosphere (SiB) model developed in the late 1980s. These legacy "look-up tables" are increasingly seen as outdated. Recent field measurements suggest that default values significantly underestimate reflectance and transmittance in the near-infrared (NIR) spectrum, potentially leading to substantial errors in climate simulations and energy budget calculations.
Methods
The researchers conducted a comparative study using two widely adopted models: the Community Land Model (CLM5.0) and the Common Land Model (CoLM2014). Simulations were performed in "offline" mode for the period 2000–2009, driven by the CRUNCEPv7 historical atmospheric dataset. The core experiment involved running each model twice: once with its default prescribed LOPs and once with newly measured LOPs synthesized from recent spectral databases (Majasalmi and Bright, 2019). To ensure comparability, forest classifications in CoLM were mapped to match the geoclimatic zones (tropical, temperate, boreal) used in CLM5.0. The analysis focused on four primary plant functional types (PFTs): needleleaf evergreen trees (NET), needleleaf deciduous trees (NDT), broadleaf evergreen trees (BET), and broadleaf deciduous trees (BDT).
Results
The primary finding was a substantial discrepancy in NIR properties between the models' default values and measured data. While visible (VIS) band properties were relatively consistent with observations, measured leaves were found to be significantly more reflective and transmissive in the NIR band than previously assumed.
This discrepancy has a direct impact on surface albedo. In the boreal forest zone, Southeast China, and the eastern United States, the NIR surface albedo increased by 3% to 10% during the summer months (JJA) when using measured LOPs. This correction effectively addresses the long-standing "negative albedo bias" in boreal forests, where models previously predicted surfaces to be darker and more absorbent than they are in reality.
Furthermore, the shift in LOPs altered energy partitioning. Because leaves were found to be more transparent in the NIR than previously modeled, more radiation reaches the ground, shifting sensible heat flux from the vegetation canopy to the soil. Overall, incorporating measured LOPs led to a reduction in simulated net radiation and total sensible heat flux by as much as 11 W/m². Interestingly, these changes had minimal impact on Gross Primary Productivity (GPP) and land surface temperatures in these offline runs, suggesting that biological productivity—driven primarily by visible light—remains relatively stable despite the energy budget shift.
Discussion and Conclusion
The results of this study emphasize that LOP parameters have a considerable effect on the surface radiation budget and sensible heat flux simulations, necessitating their inclusion in future land surface model development. However, the study's reliance on offline simulations means that potential feedback loops—where changes in albedo affect air temperature, cloud formation, and precipitation—were not captured. Furthermore, the authors note that the "big-leaf" structure used in many models may mask more complex 3D canopy interactions. Future research should focus on incorporating these measured LOPs into coupled land-atmosphere models and developing dynamic LOP databases that account for seasonal variations in leaf health and color.
Summary updated