the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Explainable machine learning for modeling of net ecosystem exchange in boreal forests
Ekaterina Ezhova
Anna Lintunen
Pasi Kolari
Tuomo Nieminen
Ivan Mammarella
Keijo Heljanko
Markku Kulmala
Related authors
Boreal rivers are significant sources of carbon dioxide (CO2) and methane (CH4) to the atmosphere but the controls of these emissions are uncertain. We measured four months of CO2 and CH4 exchange between a regulated boreal river and the atmosphere with eddy covariance. We found statistical relationships between the gas exchange and several environmental variables, the most important of which were dissolved CO2 partial pressure in water, wind speed, and water temperature.
We present a novel version of an aerosol number size distribution instrument, showcasing its capability to measure particle number concentration and particle number size distribution between 1 and 12 nm. Our results show that the instrument agrees well with existing instrumentation and allows for both the accurate measurement of the smallest particles and overlap with more conventional aerosol number size distribution instruments.
Related subject area
Drained peatlands cause high CO2 emissions. We assessed the effectiveness of subsurface water infiltration systems (WISs) in reducing CO2 emissions related to increases in water table depth (WTD) on 12 sites for up to 4 years. Results show WISs markedly reduced emissions by 2.1 t CO2-C ha-1 yr-1. The relationship between the amount of carbon above the WTD and CO2 emission was stronger than the relationship between WTD and emission. Long-term monitoring is crucial for accurate emission estimates.