Louiselva, situated in Svalbard, is an alpine periglacial catchment underlain by shale bedrock. Svalbard’s hydrology is rapidly changing with freshwater discharge expected to double by 2100 under current climate models. Increasing solute transport, notably nitrate which is a limiting nutrient in Arctic marine environments, might therefore have profound impacts on downstream ecosystems. Although glacierized catchments have received most attention in the context of land-to-sea nitrate export, we show that their periglacial counterparts may be equally important, especially when underlain by shale bedrock; an appreciable source of geogenic nitrate. In this study, we aim to establish how changes in the periglacial landscape, such as the deepening of the active layer influence nitrogen transport by runoff. We use a source to sink approach combining monitoring records with weekly surface water geochemical sampling in the catchment through the 2023 and 2024 melt seasons. Sources of nitrate are differentiated using δ 15 N-NO 3 - and δ 18 O-NO 3 - to understand spatial and temporal trends in the system. Early season NO 3 - displays a strong meteoric signature. Decreasing δ 18 O-NO 3 - and high solute loads follow the snowmelt discharge peak and show that flowpaths increasingly acquire more nitrate after travelling through aerobic mineral soil layers that promote net nitrification. These flowpaths pass through a progressively deepening active layer, leading to a release of previously frozen NO 3 - derived largely from the geogenic and pedogenic sources (i.e. nitrification of nitrogen from remineralization in soils and erosion from the bedrock). Peak nitrate yield (max. 0.1 kg ha −1 day −1 ) therefore occurs following late summer rain events (mid-August onwards), when the active layer is deepest. Total NO 3 - delivery to the fjord for the 2023 and 2024 melt season was ∼1300 kg and ∼700 kg respectively. Our study shows that the fate of nitrogen in periglacial landscapes is highly dependent on late season dynamics, namely active layer conditions and precipitation patterns. This study highlights the importance of small Arctic catchments in regional nitrogen cycling and demonstrates that their hydrological and biogeochemical responses to climate change can diverge significantly from those of large Arctic river systems. Schematic representation of late season δ 15 N-NO 3 - and δ 18 O-NO 3 - dynamics in the Louiselva catchment, as well as in the nearby polygonal terrain at the Revneset raised beach (wetland). Letters A-D indicate potential nitrate sources with their measured range: A. ornithogenic (measured at the nearby Alkepynten bird cliff), B. geogenic (values derived from Dixon (2019)), C. pedogenic (values derived from Faucherre et al., 2018) and D. meteoric (i.e. snow). Arrows show the δ 15 N-NO 3 - and δ 18 O-NO 3 - range of exported nitrate from Louiselva and the nearby polygonal surface runoff in August-September. Note: δ 18 O-NO 3 - range for B and C marked with an * are calculated based on the δ 18 O range of H 2 O in the catchment assuming 5 oxygens from H 2 O and 1 from O 2 . Illustration of the catchment was made by Ingrid Valstad. • Peak NO 3 - yield (0.1 kg ha −1 day −1 ) occurs following late summer rain events from mid-August onwards, when the thaw layer is deepest. • Following snowmelt, NO 3 - in the catchment is derived largely from nitrified geogenic and pedogenic (including permafrost) nitrogen sources. • The key factors for mobilizing NO 3 - at Louiselva are the following: i) high hydraulic gradient and low biotic demand of inorganic N in the catchment, ii) relatively low ice content in the permafrost resulting in minimal impact from thermokarst processes, iii) continuous permafrost which restricts flowpaths to supra-permafrost runoff, iv) N-rich bedrock and v) deepening thaw layer with time (seasonally and inter-annually) that enable flowpaths in mineral horizons. • Unlike at the nearby Revneset raised beach, NO 3 - uptake by plant and denitrification are limited in the Louiselva catchment.
Verret et al. (2026) studied this question.