Storm-event patterns and sources of dissolved organic matter (DOM) for stream runoff in a forested, mid-Atlantic watershed

DOM exports from catchments change dramatically and are typically highest during storm events. Thus, determining how DOM concentrations and quality change during events is critical for developing reliable budgets and for furthering our mechanistic understanding of DOM dynamics. We investigated the concentrations of dissolved organic carbon (DOC), nitrogen (DON), and DOM quality in stream water for multiple storm events over a two-year period. The study was conducted in a 12 ha forested watershed located in the Piedmont region of the mid-Atlantic USA. DOM quality was characterized using ultra-violet (UV) absorbance and fluorescence metrics such as - absorption coefficient at 254nm (a254), specific-UV absorbance (SUVA254), humification index (HIX), fluorescence index (FI), protein-like moieties and other indices derived from PARAFAC modeling of fluorescence excitation-emission matrices (EEMs). To identify the sources of stream runoff during events, an endmember mixing analysis (EMMA) was implemented. Tracers used in EMMA included sodium, magnesium, calcium, silica, DOC and a254. EMMA model was verified using 18O and other independent solutes. DOC and DON concentrations increased during storms and reached a peak concentration at or after the peak in streamflow discharge. Aromatic and humic constituents of DOM characterized by SUVA, a254, and HIX also increased during the event but peaked at different times during the event. In contrast, FI, protein-like, and “microbial” constituents of DOM followed a dilution trajectory with a minimum at or after the peak in streamflow discharge. Based on EMMA, we attributed the increase in aromatic and humic DOM constituents in stream runoff to contributions from humic-rich surficial sources such as litter-leachate and/or soil water. The dilution in % proteins and “microbial” constituents during the event was attributed to the relative drop in contributions from groundwater sources which had the highest % content of protein-like DOM. The magnitude and temporal patterns of the various DOM constituents also varied with event size, seasonal timing of the storms and the antecedent moisture conditions in the catchment. This study highlights that various DOM constituents may follow different patterns during storms and understanding these differences is critical for a comprehensive assessment of DOM dynamics in watersheds.