Fluorescent DOC characteristics are related to streamflow and pasture cover in streams of a mixed landscape

Abstract

A growing body of research suggests that different land use activities may alter both the quantity and quality of dissolved organic carbon (DOC) exported from terrestrial landscapes. However, little is known about DOC from mixed-use landscapes where hydrology varies seasonally. This study examined how DOC and chemical properties of stream water were related to land use, drainage area, and streamflow in mixed-use landscape in the Willamette River Basin, Oregon. Stream water samples were collected at roughly monthly intervals over three water years from 21 sites whose drainage areas ranged from 1 to > 11,000 km2 and included pasture, forest, and developed land uses. DOC properties were characterized using PARAFAC (parallel factor) analyses of 3-D excitation and emission matrices (EEMs). We used the Cory-McKnight (CM) model, and we also developed a model unique to our samples. Thirteen components were identified using the CM model, and these were highly correlated with, and similar with respect to excitation and emission to the three components identified in our site-specific model. Fluorescent components of DOC were related to discharge and land cover, but not to drainage area. In our model, one component (C1) appeared to be associated with terrestrial detritus and was greater in streamflow from sites with forest/developed cover than from sites with high pasture cover. A second component (C2) was greater in streamflow from high-pasture sites than from high-forest sites. C2 was strongly correlated with a protein component identified in the CM model, and therefore we attributed this second component to more microbially-processed DOC. C1 increased significantly with discharge for both land covers, suggesting that periods of high flow produce less microbially-processed DOC from more surficial flow. C2 decreased significantly with discharge for both land covers, suggesting that deeper hydrologic flow paths produce more soil-associated, microbially processed DOC. SUVA254, often used as an index of chemical quality and aromaticity of DOC, was related only to streamflow but not land use, suggesting that while the chemistry of DOC differed among land use types as detected by EEMs analysis, the aromaticity of the DOC did not differ. The CM model and our site-specific model provided consistent results: those components that were highest in high pasture sites were also highest at low flow and appeared to be more microbially processed; components that were highest in low pasture sites were also greatest at high flow and appeared to reflect less microbially processed DOC. Taken together, these findings are consistent with a conceptual model of how differences in DOC chemistry among landscape types reflect differences in plant litter input chemistry, hydrologic connectivity, and degree of microbial processing. Such changes in DOC quality with land-use change can affect foodweb dynamics in receiving waters and change the balance between carbon storage and carbon flux to the atmosphere, and they imply that even non-intensive agricultural land use can have significant effects on terrestrial-aquatic carbon processes.