Watershed urbanization increases the concentrations of major ions in downstream freshwater ecosystems. Non-point source ions from human activities and the chemical weathering of infrastructure are efficiently transported by stormwater runoff through subsurface pipes directly into streams. While the increase in mean streamwater solute concentrations in urban watersheds because of non-point source loading is a commonly documented phenomenon, the temporal dynamics of urban streamwater solute chemistry and its relationship to development patterns is less well understood. We continuously monitored streamflow, temperature, and conductivity as well as bimonthly variation in the solute chemistry of 24 watershed outlets in the Raleigh-Durham-Chapel Hill metropolitan area in North Carolina, USA for two years. Watersheds were selected to have similar proportions of development (10–36% development) while spanning the full measured range in road and stormwater pipe density for urban watersheds in this metropolitan area. We found remarkable temporal variation in the chemical regimes of the urban streams draining this set of watersheds, despite their similar proportion of development. For multiple major ions (SO42−, Cl−, NO3−), bimonthly concentrations varied tenfold across the study streams and temporal variation in streamwater chemical composition increased across watershed road and stormwater pipe density gradients. Total temporal variation in baseflow and stormflow dissolved ions, as measured by the coefficient of variation of specific conductance, was highly correlated with road density after accounting for underlying geology (R2 > 0.60). Using structural equation modeling, we found that subsurface piping mediates the relationship between roads and stream chemistry. In watersheds with high stormwater pipe density, streamwater ionic strength was low at baseflow while high and highly variable during event flow. This ‘flashy’ chemical signal in those watersheds, mirrors their more responsive hydrographs. In contrast, in our urban watersheds with the lowest pipe density, baseflow ionic strengths were higher, and event flow ionic strengths were lower and less variable. These results suggest that the extent to which pavement is drained by pipes creates a tradeoff between routing urban salts directly to streams via pipes versus indirectly to streams via loading to groundwater. Results of this study demonstrate the importance of considering altered temporal chemical variability, not just elevated solute concentrations, as a key feature of the impact of urbanization on streams. Future urban development design which minimizes the extent of roads and stormwater pipes in watersheds through denser development may be an effective strategy to abate impacts on downstream water quality.
