EGU2020-15930
https://doi.org/10.5194/egusphere-egu2020-15930
EGU General Assembly 2020
© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Carbon dioxide fluxes on agricultural grasslands – Uncertainty associated with gap-filling

Henriikka Vekuri1, Juha-Pekka Tuovinen1, Mika Korkiakoski1, Laura Heimsch1, Liisa Kulmala1,2, Juuso Rainne1, Timo Mäkelä1, Juha Hatakka1, Jari Liski1, Tuomas Laurila1, and Annalea Lohila1,2
Henriikka Vekuri et al.
  • 1Finnish Meteorological Institute, Helsinki, Finland (henriikka.vekuri@fmi.fi)
  • 2Institute for Atmospheric and Earth System Research, University of Helsinki, Helsinki, Finland

Mitigation of climate change requires – besides reductions in greenhouse gas emissions – actions to increase carbon sinks and storages in terrestrial ecosystems. Agricultural lands have a high potential for increased carbon sequestration through climate-smart land management and agricultural practices. However, in order to make climate-smart farming an accredited solution for climate policy, carbon markets and product footprints, reliable verification of carbon sequestration is needed. Direct measurement of the changes in soil carbon stock is slow, laborious and expensive and has significant uncertainties due to large background stocks and high spatial variability. An alternative is to infer the soil carbon stock change from measurements of the gaseous carbon fluxes between ecosystems and the atmosphere using the micrometeorological eddy covariance (EC) method.

Eddy covariance measures net ecosystem exchange (NEE), which is a small difference between two large components: carbon uptake by photosynthesis and losses due to plant and soil respiration. Therefore, small changes in either of them results in a large change in NEE. This sensitivity is also reflected in uncertainty estimates, which are critical for defining confidence intervals for annual carbon budget estimates and for making statistically valid comparisons of different management practices.  In addition, there are inevitable gaps in the data due to instrument failure, power shortages and non-ideal flow conditions. Therefore, in order to calculate daily and annual sums, the collected data must be temporally upscaled or gap-filled, which constitutes a major additional source of uncertainty. This study compares two different gap-filling methods for CO₂ fluxes: (1) an artificial neural network and (2) non-linear regression, which uses temperature and radiation as drivers. Uncertainties associated with both methods are estimated and discussed. The analysis is based on EC flux measurements conducted at two agricultural grassland sites in Finland.

How to cite: Vekuri, H., Tuovinen, J.-P., Korkiakoski, M., Heimsch, L., Kulmala, L., Rainne, J., Mäkelä, T., Hatakka, J., Liski, J., Laurila, T., and Lohila, A.: Carbon dioxide fluxes on agricultural grasslands – Uncertainty associated with gap-filling, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-15930, https://doi.org/10.5194/egusphere-egu2020-15930, 2020