Understanding soil respiration behaviour in different environments is one of the
most crucial research questions currently in environmental sciences, since it is a
major component of the carbon cycle. It can be divided into many source
components, them being litter decomposition, soil organic matter, root respiration
and respiration in the rhizosphere. Many biotic and abiotic factors control soil
respiration through complicated relationship networks. Strong controlling factors
being soil temperature, soil moisture, substrate supply and quality, soil nitrogen
content, soil acidity and soil texture. As these relationships are biome-specific, they
must be understood in order to produce more accurate assessments worldwide.
In this study annual soil respiration rates and its controlling factors were
investigated globally in unmanaged and natural mature forest biomes. Observed
values were extracted from Soil respiration database (SRDB) v.5, and it was
complemented with spatially and temporally linked data from remotely sensed and
modelled databases to produce variables for forest productivity, meteorological
conditions and soil properties. Furthermore, empirical soil respiration models and
machine learning algorithms, as well as previous estimates, were compared to
each other. Locally, monthly manual soil respiration measurements from boreal
forest site in Hyytiälä, Finland from the years 2010-2011, with environmental, soil
temperature and soil water conditions were investigated to identify seasonal
differences in controlling factors of soil respiration rate.
Soil respiration controls were found to differ between biomes. Furthermore, the
Artificial Neural Network algorithm used was observed to outperform empirical
models and previous estimates, when biome specific modelling was implemented
with the continental division. Artificial neural networks and other algorithms could
produce more accurate estimates globally. Locally soil respiration rates were
observed to differ seasonally, with soil temperature controls being stronger during
the growing season and when snow depth exceeded 30 cm, soil water conditions,
controlled soil respiration strongly.
