February 21, 2023

Print Friendly, PDF & Email

Modeling Carbon and Energy Exchange of an Amazonian Palm Swamp Peatland

Advancing biophysical functions and multiobjective parameter optimization boosts carbon cycle model capability for tropical peatlands.

Eddy covariance flux measurements at Sitio de Monitoreo Intensivo de Carbono, Quistococha Forest Reserve Amazonian AmeriFlux Site (PE-QFR).

[Courtesy Instituto de Investigaciones de la Amazonía Peruana.]

The Science

Tropical peatlands are an important global carbon sink and represent a major biophysical feedback factor in the climate system. Researchers use models with empirical data to represent carbon cycle processes for these complex ecosystems. Unfortunately, the lack of field observations for these ecosystems leads to a substantial knowledge gap when simulating real-world tropical forested peatlands. Incorporating field observations from a newly established peatland site in Iquitos, Peru, allowed researchers to boost a land surface model’s ability to simulate carbon dioxide (CO2) and methane (CH4) fluxes and energy balance for tropical forested peatlands by advancing tropical-specific biophysical functions and multiobjective parameter optimization.

The Impact

This study advanced three key tropical-specific biophysical functions to reduce model structure bias. Model bias from parametric estimates was further reduced using surrogate-assisted Bayesian optimization. This study lowered model uncertainties in simulating carbon cycle processes and budgets in tropical forest peatlands. It also improved understanding of how these ecosystems function and respond to future climate change. This improved representation will increase confidence in projecting biophysical feedbacks associated with tropical forested peatlands.


In this study, researchers evaluated and improved the performance of the Energy Exascale Earth System Model (E3SM) Land Model (ELM) in simulating CO2 and CH4 fluxes and energy balance of an Amazonian palm swamp peatland in Iquitos, Peru. Three algorithms were improved according to site-specific characteristics, and key parameters were optimized using an objective surrogate-assisted Bayesian approach. Modified algorithms included soil water retention curve, water coverage scalar function for CH4 processes, and seasonally varying leaf carbon-to-nitrogen ratio function. The revised tropics-specific model better simulated diel and seasonal patterns of carbon and energy fluxes of the tropical forested peatland. Global sensitivity analyses indicated that the strong controls on carbon and energy fluxes were mainly attributed to parameters associated with vegetation activities. Parameter relative importance depended on biogeochemical processes and shifted significantly between wet and dry seasons. This study advanced understanding of biotic controls on carbon and energy exchange in Amazonian palm swamp peatlands and highlighted knowledge gaps in simulating tropical peatland carbon cycling.

Principal Investigator

Timothy Griffis
University of Minnesota

Program Manager

Daniel Stover
U.S. Department of Energy, Biological and Environmental Research (SC-33)
Environmental System Science


This research is based upon work supported by the Biological and Environmental Research (BER) Program within the U.S. Department of Energy’s (DOE) Office of Science under award number DE-SC0020167. Support was always received from the Sustainable Wetlands Adaptation and Mitigation Program (SWAMP, Grant MTO-069018), which is funded by the U.S. Agency for International Development and implemented by the U.S. Forest Service and Center for International Forestry Research. The Global Comparative Study on REDD+ (Grant agreement #QZA-12/0882) by the Government of Norway and the Instituto de Investigaciones de la Amazonía Peruana (IIAP) also provided support.

Related Links


Yuan, F., et al. "Evaluation and Improvement of the E3SM Land Model for Simulating Energy and Carbon Fluxes in an Amazonian Peatland." Agricultural and Forest Meteorology 332 109364  (2023). https://doi.org/10.1016/j.agrformet.2023.109364.