Seasonal and vegetation driven shifts in Arctic dissolved organic carbon composition using metabolomics and fluorescence characterization

Background/Question/Methods

The production and assimilation of dissolved organic matter (DOM) are critical steps in the global carbon (C) cycle.  The large pool of organic C in Arctic tundra soils is vulnerable to increased decomposition rates due to the warming climate resulting in increased production of DOM. DOM is the portion of OM that can be directly assimilated by microbes, and in soils originates from the enzymatic depolymerization of detritus or soil organic matter, or directly from plant leachates and exudates. We examined the molecular composition of DOM at Toolik Field Station in Alaska (68° 38’ N, 149° 36’ W) to determine which components varied seasonally in soil pore water among three vegetation types (wet sedge: Carex aquatilis and Eriophorum angustifolium, moist acidic tussock: E. vaginatum, and shrub tundra: Betula nana and Salix sp.). These sites were sampled during winter/summer and summer/winter transitions in 2010. We expected the chemical composition of DOM in pore water to be distinct among plant communities due to differences in root exudates, litter chemistry and associated microbial communities; and vary seasonally due to shifting temperature and water availability and their impacts on decomposition of DOM. Soil pore water was isolated through centrifugation, characterized for metabolomics with ultra high performance liquid chromatography (UPLC) in line with a quadrupole time of flight mass spectrometer (QTOF-MS) and analyzed with principle components analysis. Excitation emission matrices (EEMs) generated by fluorescence spectroscopy were evaluated with parallel factors analysis and the Cory-McKnight model.

Results/Conclusions

We detected over 6000 mass features in pore water metabolite profiles. PCA analysis of the mass features revealed that sampling date was a better descriptor of DOM characteristics than vegetation type, with the summer to winter transition showing the largest shift in DOM composition. However, fluorescence EEMs indicated the opposite pattern with DOM composition more consistent by vegetation type than season. The Cory-McKnight model for DOM components exhibited high residuals so we developed a new model for the description of component variability in our data set.  The combined analyses suggest the metabolomics analysis may better characterize microbially processed components of the DOM pool while fluorescence EEMs may be more representative of plant-derived material. Furthering the understanding of DOM chemical characteristics and seasonal shifts in its composition can identify the material most sensitive to climatically driven changes, and thus inform future C decomposition dynamics in a warming Arctic and their impact on the global C cycle.