Strong radiative effect induced by clouds and smoke on forest net ecosystem productivity in central Siberia

• Published in: Agricultural and Forest Meteorology
• Main Authors: Park, Sung-Bin, Knohl, Alexander, Lucas-Moffat, Antje Maria, Migliavacca, Mirco, Gerbig, Christop, Vesala, Timo, Peltola, Olli, Mammarella, Ivan, Kolle, Olaf, Lavric, Jost Valentin, Prokushkin, Anatoly, Heimann, Martin
• Format: Article in Journal/Newspaper
• Language: English
• Published: 2018
• Subjects: Text; article; ddc:630; Eddy covariance > Net ecosystem productivity > Smoke > Photosynthetically active radiation > Artificial neural networks > Central Siberia; taiga; Siberia
• Online Access: https://doi.org/10.1016/j.agrformet.2017.09.009; https://www.openagrar.de/receive/openagrar_mods_00043554; https://www.openagrar.de/servlets/MCRFileNodeServlet/openagrar_derivate_00017212/dn059790.pdf

Aerosols produced by wildfires are a common phenomenon in boreal regions. For the Siberian taiga, it is still an open question if the effects of aerosols on atmospheric conditions increase net CO2 uptake or photosynthesis. We investigated the factors controlling forest net ecosystem productivity (NEP) and explored how clouds and smoke modulate radiation as a major factor controlling NEP during fire events in the years 2012 and 2013. To characterize the underlying mechanisms of the NEP response to environmental drivers, Artificial Neural Networks (ANNs) were trained by eddy covariance flux measurements nearby the Zotino Tall Tower Observatory (ZOTTO). Total photosynthetically active radiation, vapour pressure deficit, and diffuse fraction explain at about 54–58% of NEP variability. NEP shows a strong negative sensitivity to VPD, and a small positive to fdif. A strong diffuse radiation fertilization effect does not exist at ZOTTO forest due to the combined effects of low light intensity, sparse canopy and low leaf area index. Results suggests that light intensity and canopy structure are important factors of the overall diffuse radiation fertilization effect.