Bottom-waters in Smithtown Bay (Long Island Sound, NY) are subject to hypoxic conditions every summer despite limited nutrient inputs from waste-water and riverine sources, while modeling estimates of groundwater inputs are thought to be insignificant. Terrestrial and marine fluxes of submarine groundwater discharge (SGD) were quantified to Smithtown Bay using mass balances of ²²²Rn, ²²⁴Ra, ²²⁶Ra and ²²⁸Ra during the spring and summer of 2014/2015, in order to track this seasonal transition period. Intertidal pore waters from a coastal bluff (terrestrial SGD) and from a barrier beach (marine SGD) displayed substantial differences in N concentrations and sources, traced using a multi-isotope approach (²²²Rn, Ra, δ¹⁵N-NO₃⁻, δ¹⁸O-NO₃⁻). NO₃⁻ in terrestrial SGD did not display any seasonality and was derived from residential septic systems and fertilizer. Marine SGD N concentrations varied month-to-month because of mixing between oxic seawater and hypoxic saline pore waters; N concentrations were greatest during the summer, when NO₃⁻ was derived from the remineralization of organic matter. Short-lived ²²²Rn and ²²⁴Ra SGD fluxes were used to determine remineralized N loads along tidal recirculation flow paths, while long-lived ²²⁸Ra was used to trace inputs of anthropogenic N in terrestrial SGD. ²²⁸Ra-derived terrestrial N load estimates were between 20 and 55% lower than ²²⁴Ra-derived estimates (excluding spring 2014); ²²⁸Ra may be a more appropriate tracer of terrestrial SGD N loads. Terrestrial SGD NO₃⁻ (derived from ²²⁸Ra) to Smithtown Bay varied from (1.40–12.8) ∗ 10⁶ mol N y⁻¹, with comparable marine SGD NO₃⁻ fluxes of (1.70–6.79) ∗ 10⁶ mol N y⁻¹ derived from ²²²Rn and ²²⁴Ra. Remineralized N loads were greater during the summer compared with spring, and these may be an important driver toward the onset of seasonal hypoxic conditions in Smithtown Bay and western Long Island Sound. Seawater recirculation through the coastal aquifer can rival the N load from terrestrial SGD from a heavily polluted aquifer.