Seasonal leaf dynamics for tropical evergreen forests in a process based global ecosystem model

M. De Weirdt, Hans Verbeeck, Fabienne Maignan, Philippe Peylin, Benjamin Poulter, Damien Bonal, Philippe Ciais, Kathy Steppe2012 · DOI

TL;DR

This paper introduces a new model for simulating seasonal leaf dynamics in tropical evergreen forests, addressing a significant simplification in current global ecosystem models. By incorporating a more realistic leaf turnover mechanism driven by Net Primary Production (NPP), the model better captures observed carbon fluxes, particularly Gross Primary Productivity (GPP), in these critical ecosystems.

What Problem Does This Paper Address?

Global vegetation models have historically oversimplified the seasonal behavior of tropical evergreen forests, often assuming constant leaf area properties. This is problematic because tropical forests play a huge role in the global carbon cycle and climate regulation. Recent studies suggest that seasonal phenology (changes in leaf cover and photosynthetic capacity) in these forests significantly impacts carbon and water fluxes, and thus climate dynamics. The authors aim to bridge this gap by developing a more process-based representation of seasonal leaf dynamics within the ORCHIDEE global ecosystem model, leading to more accurate simulations of these vital ecosystems.

Methods

The core of the paper is the development of a new "NPP-based Leaf Turnover" (NLT) scheme for the ORCHIDEE model. This scheme replaces the previous age-dependent leaf litterfall model with one where daily leaf litterfall is directly linked to daily leaf Net Primary Production (NPP). The rationale is that in a dense canopy, older, less efficient leaves are shed to make way for new, more productive ones, a process driven by resource availability and light. This new scheme also involves adjusting two key parameters: increasing the fraction of NPP allocated to leaves and decreasing fine root maintenance respiration. The model was then evaluated against in-situ measurements from a flux tower site in French Guiana.

Datasets

The study primarily uses data from the Guyaflux tower site in French Guiana, including:

  • Eddy covariance measurements of carbon fluxes (GPP, NEE) from 2004-2009.
  • Leaf litterfall measurements collected over several years.
  • Meteorological data (down-welling short-wave radiation, rainfall, temperature, etc.).

Additionally, leaf litterfall data from the Tapajós km 67 site in Brazil was used for further validation.

Key Insights and Ideas

  • Leaf Litterfall as a Dynamic Process: The study highlights that leaf litterfall in tropical evergreen forests isn't just a passive process but is tightly coupled to productivity (NPP). This is a crucial shift from previous static assumptions.
  • Optimality in Canopy Dynamics: The new model is grounded in the idea that canopies optimize resource use. Shedding older leaves to invest in new ones is a strategy to maximize carbon gain under light-limiting conditions.
  • Improved GPP Simulation: By better representing seasonal leaf dynamics, the model significantly improves its ability to simulate Gross Primary Productivity (GPP), especially capturing the pronounced dry season increase observed in field data.
  • Interconnectedness of Carbon Pools: The modifications to leaf allocation and root respiration demonstrate how changes in one part of the ecosystem model can have cascading effects on overall carbon cycling and productivity.

Key Results

  • The ORCHIDEE-NLT model, incorporating the new leaf turnover scheme, shows a more realistic seasonal pattern of leaf litterfall that aligns better with field inventory data at both Guyaflux and Tapajós sites.
  • The fit to GPP flux data was substantially improved, with RMSE decreasing and correlation increasing compared to the standard ORCHIDEE model.
  • The model better captures the seasonal variations in GPP, particularly the higher GPP observed during the dry season, by simulating a more pronounced dry season increase (27% in the modified version vs. 23% in the standard version).
  • The mean modeled LAI decreased slightly in the ORCHIDEE-NLT version, remaining more constant throughout the year due to balanced leaf allocation and litterfall.
  • The canopy integrated photosynthetic capacity (Vc,max) increased in the modified model, contributing to the improved GPP simulations.

Limitations and Future Work

  • Applicability to Disturbed Forests: The authors acknowledge that the "self-thinning" approach, based on optimality, might be less applicable to secondary or disturbed forests that haven't reached their maximum canopy biomass or LAI.
  • Spatial Variability: The current study focuses on single sites. Future work should extend the analysis to other tropical forest types (e.g., secondary, flooded) and investigate the impact of spatial variations in soil fertility and nutrient content on Vc,max and GPP.
  • Linear vs. Exponential Vc,max Profile: The study suggests testing a linear decrease in Vc,max with canopy depth, as opposed to the current exponential implementation, which might lead to higher simulated photosynthetic capacity.
  • Drought Stress Representation: While the model improved dry season GPP, the authors note that processes related to drought stress might still need further refinement.

Summary updated

Abstract

The influence of seasonal phenology in tropical humid forests on canopy photosynthesis remains poorly understood and its representation in global vegetation models highly simplified, typically with no seasonal variability of canopy leaf area properties taken into account. However, recent flux tower and remote sensing studies suggest that seasonal phenology in tropical rainforests exerts a large influence over carbon and water fluxes, with feedbacks that can significantly influence climate dynamics. A more realistic description of the underlying mechanisms that drive seasonal tropical forest photosynthesis and phenology could improve the correspondence of global vegetation model outputs with the wet-dry season biogeochemical patterns measured at flux tower sites. Here, we introduce a leaf Net Primary Production (NPP) based canopy dynamics scheme for evergreen tropical forests in the global terrestrial ecosystem model ORCHIDEE and validated the new scheme against in-situ carbon flux measurements. Modelled Gross Primary Productivity (GPP) patterns are analyzed in details for a flux tower site in French Guiana, in a forest where the dry season is short and where the vegetation is considered to have developed adaptive mechanisms against drought stress. By including leaf litterfall seasonality and a coincident light driven leaf flush and seasonal change in photosynthetic capacity in ORCHIDEE, modelled carbon and water fluxes more accurately represent the observations. The fit to GPP flux data was substantially improved and the results confirmed that by modifying canopy dynamics to benefit from increased light conditions, a better representation of the seasonal carbon flux patterns was made.
BibTeX
@misc{weirdt2012seasonal,
  author = {M. De Weirdt and Hans Verbeeck and Fabienne Maignan and Philippe Peylin and Benjamin Poulter and Damien Bonal and Philippe Ciais and Kathy Steppe},
  title = {{Seasonal leaf dynamics for tropical evergreen forests in a process based global ecosystem model}},
  year = {2012},
  doi = {10.5194/gmdd-5-639-2012},
}

Loading related content...

Figure