Disturbance alters relationships between soil carbon pools and
aboveground vegetation attributes in an anthropogenic peatland in
Patagonia
Abstract
1 Vegetation attributes derived from species and plant functional types
(PFTs) directly or indirectly drive the carbon (C) cycle in peatlands.
However, anthropogenic-based disturbances may alter petland soil-plant
interactions and their ability to sequester carbon. Likewise, it is
unclear how the soil-plant linkages among different soil C
decomposition-based pools and plant attributes vary under disturbance
conditions. 2 We aimed to assess how anthropogenic disturbances affect
the relationships between aboveground vegetation attributes and
belowground C pools in a peatland located in Northern Patagonia, Chile.
We further evaluated if attributes derived from PFTs are better suited
to predict soil C pools than attributes derived from species. We used
structural equation modeling and regression analyses to explore these
differences. 3 We found that undisturbed peatland has more soil-plant
significant relationships between soil C pools and vegetation
attributes, yielding higher predictive accuracies than disturbed areas.
The species-based attributes yielded consistently better results
predicting soil C pools than PFT-based attributes. However, PFT-based
information showed significant interactions with the highly-decomposed C
pools in the undisturbed peatland. Likewise, plant height and diversity
were only significant with C pools in the undisturbed peatland. 4 We
observed that water-logged plant communities have different soil-plant
interactions than dryer communities. These differences were observed in
both areas but were higher in the disturbed peatland, making it
impossible to find meaningful soil-plant relationships across vegetation
types and taxa. 5 Our results highlight the importance of accounting for
disturbance or management when linking vegetation attributes to soil C
pools in peatlands. This implies that up-to-date extensive monitoring of
peatland disturbances is needed to accurately monitor soil C attributes
at the regional level using vegetation as proxies. We also need to
aggregate species into specific plant functional types that hold these
soil-plant interactions across landscapes, regions, and disturbances to
generalize the soil-plant relationships accurately.