High-frequency monitoring reveals seasonal and event-scale water quality variation in a temporally frozen river

Abstract

Potential influences of climate change on water quality, riverine suspended sediments, nitrogen and organic matter loads in temporally frozen rivers, which have ice-covered flow and snow-affected basins, are poorly understood. However, before being able to understand potential future changes, the impact of ice and snow needs to be investigated more thoroughly for years which were hydrologically different. We investigated seasonal and event scale concentration-discharge (C-Q) dynamics of total suspended solids/turbidity, nitrate-N (NO3-N) and chemical oxygen demand (COD), which is indicative of the amount of organic matter in river water. In particular, the influence of ice cover, contrasting spring thaw, and soil frost conditions on intra-annual fluxes and the C-Q response of the three solutes are detected based on over four years of hourly data. Seasonally flow-weighted suspended solids and NO3-N concentrations were at their highest in either the autumn or spring thaw, but COD concentrations were the highest each year in autumn. NO3-N and COD levels typically decreased during winter. The ice-covered river water was less turbid compared to open-channel water at an equivalent river discharge likely due to in-stream factors. Storms during the freshet period introduced flushing of organic matter and suspended solids. The ratio of organic matter yield to water yield was similar each freshet and was independent of the amount of precipitation as snow or soil frost status. The freshet NO3-N yield per water yield was higher during the years with a thick snowpack and the consequent thawed soil compared to a year with soil frost and minor snowpack. 91 storm events studied revealed differences and similarities in storm dynamics in between the three variables. Anti-clockwise hysteresis was most common for the variables, with turbidity peaking faster than the COD or NO3-N concentration in most of the storms. Snowmelt storms showed highly variable C-Q responses inbetween the variables. However, spring thaw-related COD concentration peaks abated more slowly compared to turbidity or NO3-N. NO3-N showed a strong dilution pattern during several autumn storms during an extremely wet year, indicating limited N sources for flushing from the catchment. As a result, the flow-weighted mean NO3-N concentration was not the largest during the year of largest water yield instead which was true for suspended solids and COD. We found no evidence that warmer winters with precipitation as rain instead of snow would increase suspended sediment, organic matter and NO3-N load at entire winter-spring season or annual timescales.