|
Titel |
A pantropical analyses of soil organic carbon storage in tropical forests and savannas |
VerfasserIn |
Gustavo Saiz, Michael Bird, Tomas Domingues, Ted Feldpausch, Elmar Veenendaal, Jonathan Lloyd |
Konferenz |
EGU General Assembly 2011
|
Medientyp |
Artikel
|
Sprache |
Englisch
|
Digitales Dokument |
PDF |
Erschienen |
In: GRA - Volume 13 (2011) |
Datensatznummer |
250052482
|
|
|
|
Zusammenfassung |
Tropical forests and savannas contain ~28% of the world’s soil organic carbon (SOC), and
consequently have considerable potential to impact the global carbon cycle and hence global
climate. Despite their significance, our knowledge of the drivers of SOC stocks
and fluxes remains incomplete for these ecosystems. Within the framework of the
TROBIT (TROpical Biomes In Transition) project, 63 permanent plots have been
established covering tropical forest and savannas across three continents. Large
differences in SOC stocks are observed between different geographical areas. These
differences are mostly explained by climatic and soil characteristics of the sites, as
demonstrated by significant correlations between SOC stocks and variables such as
temperature, precipitation, and relative abundance of minerals characteristic of these
highly weathered tropical soils. It is generally the case that higher aboveground
biomass will be present in ‘mature’ tropical forest compared to ‘mature’ tropical
savanna under similar environmental conditions. However, this may not always
be the case for SOC. This is because SOC storage in tropical areas is not just a
function of environmental conditions and vegetation type, which directly influence
the inputs of organic matter to the soil, but is also strongly dependant on complex
interactions of the different soil organic compounds with minerals and aggregates in the
soil.
To further advance our understanding of tropical SOC dynamics, this project
focused on implementing a combined physical and chemical fractionation scheme
aimed at validating conceptual SOC pools predicted by the process-based RothC
model across the soils of the TROBIT plot network. It is expected that this will
enable better characterization of interactions between individual site conditions
and SOC dynamics as well as enhance confidence in the use of models to predict
SOC dynamics in these ecosystems. Invariably, most of the carbon is stored in the
physical protected fractions forming aggregates. The results revealed strong regression
coefficients (r2 >0.75, with slopes close to 1) between SOC stocks in measured soil
fractions and most equilibrium-modelled pools in Roth-C. However, the model
consistently failed to satisfactorily predict the highly resistant SOC measured fraction.
To address this issue, the elemental and stable carbon isotopic composition of the
bulk soil and the chemically resistant fraction (defined as that fraction surviving
dichromate oxidation at 60Ë for 72 hours) were used as a proxy for the role of fire in
producing chemically resistant compounds ‘pyrogenic carbon’ in the soils. Using this
proxy we found that, sites with higher soil δ13C values (indicative of C4-containing
ecosystems likely susceptible to fire) contained relatively larger amounts of resistant
carbon compared to those soils with low δ13C values characteristic of C3-only
environments.
However, pyrogenic carbon still failed to account for most of the resistant carbon
observed in forests with a presumed low (if any) incidence of fire, and in some savannas
established on rich basalt soil environments. This suggests that physical protection has to be
playing a major role in the occurrence of chemically resistant organic carbon compounds in
some of the soils from the TROBIT sites. An attempt was made to include both forests and
savannas in a single predictive function for resistant SOC based on total SOC and the
relative abundance of soil mineral components, with either quartz or iron contents
giving excellent results (r2 >0.79). These results indicate that failure to account for
stocks of pyrogenic carbon and the role of aggregates protecting SOM in these
fire-prone and highly weathered ecosystems may result in underestimation of a
significant C pool (~20% of TOC on average) with a long mean residence time, with
implications for accurate predictions of future SOC dynamics in tropical ecosystems. |
|
|
|
|
|