Surface Layer Energy And Carbon Fluxes Are Vulnerable To The Representation Of Canopy Structural And Functional Profiles

The terrestrial carbon uptake currently accounts for about one-third of the annual global carbon sink in the atmosphere; however, future projections of the terrestrial carbon cycle is uncertain due to the complex, nonlinear interactions in the earth’s climate system. Studies have shown that carbon sequestration strength varies with different ecosystems, and its magnitude is greatly regulated by the amount of photosynthetic leaves inside plant canopies.

Although various ecophysiological treatments have been applied in different land surface models, the amount of plant leaves is unanimously represented by a dimensionless quantity called Leaf Area Index (LAI), a ratio between total green leaf area and ground surface area. Previous works on Amazon’s deforestation highlighted the impacts from LAI changes on ecosystem responses through shifting the energy partition from available energy into sensible and latent heat fluxes and thus affecting atmospheric boundary layer development and local and regional circulation patterns. As a result, a more realistic high-resolution surface vegetation LAI dataset, such as those available from satellite observations, is expected to improve global terrestrial carbon simulation.

The simulation results presented in Chang et al. (2018) confirmed this hypothesis. In their study, the accuracy of energy and carbon fluxes simulated with the use of different LAI datasets was evaluated by six years of field measurements encompassing deciduous forest, evergreen forest, and grassland. Their results indicated that the root mean square errors for the simulated water vapor and carbon fluxes are reduced by 10% and 15%, respectively, simply by improving the reliability of the model-driven LAI dataset. Their study demonstrated that the uncertainty in vegetation dataset would not only increase errors in terrestrial simulation but also lead numerical models to misinterpret land surface processes even with correct model physics.

Besides the reliability of the seasonal dynamics of LAI, Chang et al. (2018) highlighted the importance of canopy structure representation, i.e. how LAI is distributed inside vegetation canopies. They found that the inclusion of realistic canopy architecture profile improves scalar flux simulations, which advocates the use of multiple canopy layer representation in land surface simulation. Their results also suggested that the use of more advanced ecophysiological and turbulence transfer schemes can generally reduce the errors shown in energy and carbon fluxes simulation. Therefore, the use of a land surface model that reasonably represents ecosystem structural and functional responses to microclimate conditions driven by a realistic LAI dataset can thus properly represent surface layer exchange as driven by current and future climate drivers.

The simulation results presented in Chang et al. (2018) were conducted by a diabatic third order closure land surface model called ACASA developed at the University of California, Davis. The standard version of ACASA has 20 vertical canopy layers to represent the realistic turbulent fluxes of momentum, heat, moisture, and carbon dioxide above and within the simulated canopy at half-hourly to hourly time steps; the number of layers can be increased or decreased (Figure 1).

Figure 1. Schematic diagram for the ACASA model. The necessary inputs can be obtained from observations or atmospheric models. The multiple canopy layer feature in ACASA enables it to realistically capture local and non-local turbulence transport fluxes from the surface layer and heat and water fluxes from the soil layer. Image credit: Kuang-Yu Chang

These findings are described in the article entitled Canopy profile sensitivity on surface layer simulations evaluated by a multiple canopy layer higher order closure land surface model, recently published in the journal Agricultural and Forest Meteorology. This work was conducted by Kuang-Yu Chang, Kyaw Tha Paw U, and Shu-Hua Chen from the University of California, Davis.