- Small mountainous watersheds are disproportionate sources of land-derived particulate organic matter (POM) to long-term sinks like lake bottoms and the ocean. As such, these ecosystems are an essential component of the global carbon cycle. The burial of POM in lacustrine and marine sediments contributes to the drawdown of atmospheric CO2 and therefore provides a mechanism of negative feedback for climate change.
POM export in mountainous landscapes is thought to be driven by a natural disturbance regime that, in the Pacific Northwest, historically have included both tectonic (i.e., earthquakes) and climatic (e.g., fires, storms, and flooding) forcings. After World War II, a dramatic increase in timber harvesting and road building resulted in the intensification of an additional anthropogenic forcing with unknown biogeochemical consequences.
In this dissertation, we studied patterns of POM mobilization, transport, and deposition to understand better how fundamental biogeochemical processes in forested mountainous watersheds respond to disturbances triggered by natural and human forcings. We approached our analysis from the paradigm of biogeochemical signals.
A biogeochemical signal is a relative change in POM chemistry correlated with watershed ecosystem processes involving nutrient exports. Here, we examined changes in elemental composition (%C, %N N:C ratios), stable isotopic composition (δ13C and δ15N), and terrestrial biomarkers (e.g., lignin-derived compounds) to characterize different pools of particles moving from headwaters to lake bottoms.
We used categories like carbon-rich/carbon-depleted particles (H.J. Andrews study); colloids, aggregates, fine particulate organic matter, and vascular plant detritus (Alsea study). In the Loon Lake study, we used a Generalized Least Squares regression model to test for significant correlations between specific particle size classes and the bulk chemical characteristics of deposited sediment layers. Based on the regression model, we introduced the concept of pseudo-property-property plots to delineate a general correspondence between particle size classes like colloids, aggregates, clays, and sands with elemental and isotopic composition of Loon Lake sediments.
Our results suggest that the superposition of disturbance forcings (both natural and anthropogenic) modulates the quantity and quality of POM exported from headwater streams and that changes in channel structure could be inferred from changes in variability of the probability distribution of discharge and sediment yields. We used elemental composition data to illuminate how sediment routing and other stream channel processes could affect the source/sink behavior of headwater ecosystems in the context of POM exports.
Along the steep channel of the Alsea River, characterized by high attrition rates and relatively simple geomorphology, we were able to obtain evidence of the role of physical instream processing of organic matter in the export of POM from a small mountainous river system in the Oregon Coast Range. We provide evidence that extensive physical fragmentation of the coarse organic matter transported in the suspended load was a significant source of fine particulate organic matter in the river during this storm. Our data suggest that physical processing during floods, associated with particle delivery and transport must be a driving force behind the imprinting of biogeochemical signatures in sedimentary records.
Lastly, we calculated spectral entropies and the information content of particle size spectra from a high resolution sedimentary archive from a lacustrine sink spanning more than 1500-years of natural history. These metrics, derived from Shannon’s information theory, constitute a distribution-free approach to analyze changes in variability. We found that particle size variability not only contains information about specific depositional processes, but also it has the potential to record climatic changes.
An overarching conclusion from this dissertation is the extraordinary heterogeneity that characterizes the biogeochemical signals of watershed’s response to disturbances which is in startling contrast with the availability of conceptual approaches to understanding such a heterogeneity. Our results suggest that generalized metrics of signal variability are central for translating between historical structural changes in natural patterns, and their timing and relevance in watershed history. Therefore, a more robust reconstruction of watershed history and a better identification of drivers for pressing environmental issues seems possible under the framework of Shannon’s information theory.