Abstract Search

• Search again

Quantifying and evaluating uncertainty in carbon dioxide (CO2) fluxes from Alaskan North Slope tundra ecosystems

Luke Schiferl,  Lamont-Doherty Earth Observatory,  schiferl@ldeo.columbia.edu (Presenter)
Kyle Arndt,  San Diego State University,  karndt-w@sdsu.edu
Sebastien C. Biraud,  Lawrence Berkeley National Laboratory,  scbiraud@lbl.gov
Eugenie Euskirchen,  University of Alaska, Fairbanks,  seeuskirchen@alaska.edu
John Henderson,  AER,  jhenders@aer.com
Erik Larson,  Harvard University,  erik_larson@fas.harvard.edu
Kathryn McKain,  NOAA ESRL,  kathryn.mckain@noaa.gov
J. William Munger,  Harvard University,  jwmunger@seas.harvard.edu
Colm Sweeney,  NOAA Earth System Research Laboratory,  colm.sweeney@noaa.gov
Yonghong Yi,  JPL,  yonghong.yi@jpl.nasa.gov
Donatella Zona,  San Diego State University (USA),  dzona@mail.sdsu.edu
Róisín Commane,  Columbia University,  rc3195@columbia.edu

The Arctic is warming at twice the rate of the global average. Thawing of Arctic permafrost has the potential to release vast stores of carbon-containing gases into the atmosphere, thereby accelerating warming. We must first confidently simulate the current state of carbon dioxide (CO2) fluxes in order to understand how tundra ecosystems will respond to climate change on annual and decadal timescales. This study uses an improved ecosystem-resolved CO2 flux model (PVPRM-SIF) driven by functional relationships, which has been parameterized using eddy flux tower observations of CO2. These relationships have been generalized regionally to produce a spatially and temporally varying flux product. Our modeling framework allows for easy substitution of input parameters and configuration fields, such as different vegetation maps and solar-induced fluorescence (SIF) products. We find that changing the configuration assumptions can have a significant impact on the magnitude and sign of the resulting regional net ecosystem exchange (NEE), with selection of vegetation type distribution having the largest impact. We evaluate our model configurations using the calculated surface influence on airborne (e.g., ACME-V, ArctiCAP) and tower (e.g., Barrow) measurements of CO2 concentration from a Lagrangian atmospheric transport model (WRF-STILT). Our analysis identifies the spatial areas that are less accurately represented by the models and may indicate less confidence in the behavior of certain source vegetation types. The range of simulated fluxes does a good job of representing the observed growing-season seasonal cycle, including the transition into fall respiration-dominated NEE in September. However, the model underestimates the elevated respiration observed from October to December as the soil freezes. Attempts to use alternative respiration formulations, including Q10 relationships, fail to improve this underestimation. This points to an inaccurate representation of the soil temperature freeze-up which is likely driving the missing late fall to early winter respiration.

Presentation: iPoster

Associated Project(s): 

• McKain (TE 2016)
• Munger (CARBON 2016)

Presentation Type: Poster

Science Theme: Carbon Dynamics