Forest management and harvest offer a promising means of combating climate change by removing CO₂ from the atmosphere. However, most forest carbon (C) is held in soils. Thus, by disturbing soils and altering hydrology, forest management and harvest potentially displace large amounts of C from forest soils into aquatic ecosystems. My dissertation seeks to understand the fate of this forest C as dissolved organic matter (DOM) into various aquatic endpoints by tracing its molecular composition along the ephemeral water film that begins in upland soils and ends in streams. The fate and function of DOM in aquatic ecosystems are strongly affected by its chemical properties. Thus, recent progress in the molecular characterization of DOM has opened a new line of inquiry into harvest impacts on aquatic ecosystem functioning. My thesis advances this line of inquiry by applying high-resolution mass spectrometry to study the effect of forest disturbance on DOM in soils and connected streams.

Chapter 1 of this thesis gives a general overview of the molecular nature of terrestrial DOM sources, how these sources may be altered by harvest, and the subsequent transfer, fate, and properties of DOM once in streams. It also outlines the specific objectives I address with a combination of field experiments and synoptic surveys in the Batchewana watershed in Ontario, Canada. In Chapter 2, I tracked DOM along soil depths, and hillslope positions in four replicate forest headwater catchments of the Canadian hardwood forest. I related DOM composition to soil microbiomes and physical chemistry to establish baseline conditions before a harvest experiment. I found that DOM changed similarly along soil-aquatic gradients, irrespective of differences in environmental conditions. My results implicated continuous microbial reworking that shifts DOM towards a shared pool of compounds in soils. Such general degradation patterns can inform the management of soil-to-stream carbon losses by predicting DOM composition and its downstream reactivity along environmental gradients. In Chapter 3, I quantified and characterised the effects of logging on DOM composition over three years using the four experimental catchments from Chapter 2. Two catchments were experimentally logged, while the remaining two were left as controls. I found that DOM concentration in stream water from logged catchments increased in a pulse during the first year, but only the changes in the quality of DOM persisted. Using ultrahigh-resolution mass spectrometry, I showed that DOM released from deforested catchments was energy-rich and more chemically diverse, likely because of higher hydrological connectivity with intermediate and deep soil layers. I estimated that while logging increased the overall annual flux of dissolved organic carbon by approximately 8.5% of the extracted wood carbon, the exposure of deeper soil through logging released previously stable soil organic carbon to streams. The resulting changes to the molecular composition of DOM within headwater streams persisted for at least two years after logging, potentially disrupting aquatic ecosystems and making streams more likely to release terrestrial C into the atmosphere. In Chapter 4, I examined the chemical properties of DOM in stream water from over 200 Canadian headwater streams in an area with historical forest harvest. I demonstrated that using the fluorescence properties of streams, the effect of harvest, although detectable on a large spatial scale, is relatively minor compared to the effects of forest types and wetness gradients. These results have implications for land-water linkages under a changing climate that shift terrestrial sources of DOM. Finally, in Chapter 5, I discuss the implications of my findings for understanding the coupling between terrestrial and aquatic ecosystems and propose avenues for future research.