Small head water contributions in a heterogenous watershed of the Connecticut River Valley

Headwater streams in valleys comprise land surfaces with deep flow paths as well as impervious surfaces where urban centers have developed strategically near water. The impaired Mill River Watershed discharges into the Connecticut River of the Northeastern US after flowing through heterogenous land-uses in a valley system. In this study, we young water fractions across from watershed sources contributing to the CR as well as to Lake Warner. We compare theses dynamics and the inuence of urban, agricultural, and rural land use from May 2021 to November 2024, on the young water fraction (Fyw ) and mean transit times (MTT) using stable water isotopes (δ 2H, δ 17O, and δ 18O) across 13 sites (collected monthly) within the Lake Warner-Mill River watershed (LWMRW) of the Connecticut River Valley. In addition, we monitored 3 storms in Spring, Summer, and Fall to determine proportions of stream pre-event flow that coincides with pollutants such as dissolved organic carbon (DOC). Finally, we performed synoptic sampling at 33 sites between and just after storm to further compare the spatial expression on flow across the network. The local meteoric water line for the 3.25 years of collected data was δ H = 7.5 δ O+ 9.5 (n=254), with high variability of the precipitation isotopes mean stable water isotope composition of -7.3 per mille (oxygen-18, δ O) and standard deviation (SD) of 3.7 per mille.  For the monthly surface water collected data, mean results were similar.  Forested sites had mean of -7.92 per mille (SD = 0.53), urban sites -7.69 per mille (SD = 0.88), and agricultural sites -7.57 per mille (SD = 0.74). Correspondingly, Fyw was 0.17 (rural), 0.29 (urban), and 0.24 (agricultural), with rural sites having the longest MTT (327 days) and urban sites having the shortest MTT (189 days). Mean transit times for forested sites declined in value as mean O18 of the streams increased, however there was no trend for urban sites. During the three storm events, a high proportion of pre-event flow was determined to contribute to the hydrograph. In September 2024, this transported over 3 mg/L DOC at its peak. Interestingly, the synoptic showed the most enriched samples across the 33 sites in late fall rather than mid-summer sampling. This work contributes to efforts at better understanding the hydrological dynamics of the watershed network and its heterogeneous contributions to Lake Warner and the Connecticut River.