RESUMEN
Plants are arguably the most visible components of stormwater bioretention basins and play key roles in stabilizing soils and removing water through transpiration. In regions with cold winters, bioretention basins along roadways can receive considerable quantities of deicing salt, much of which migrates out of the systems prior to the onset of plant growth but the rest remains in the soil. The resulting effects on plants presumably vary with time (due to annual weather patterns), space (because stormwater exposure is location-dependent), and biology (because plant taxa differ in their salt tolerance). The goal of this study was to investigate the magnitude of deicing salt's effects on bioretention plants and how it varies with spatial, temporal, and biological factors. The study took place in a set of five bioretention basins in Philadelphia, USA that receive runoff from a major highway. Over a five-year period, the electrical conductivity (EC) of influent stormwater frequently exceeded 1 mS cm-1 in winter, and occasionally surpassed that of seawater (â¼50 mS cm-1). In both of the years when soil EC was measured as well, it remained elevated through all spring months, especially near basin inlets and centers. Mortality of nine plant taxa ranged widely after three years (0-90%), with rankings largely corresponding to salt tolerances. Moreover, leaf areas and/or crown volumes were strongly reduced in proportion to stormwater exposure in seven of these taxa. In the three taxa evaluated for tissue concentrations of 14 potentially toxic elements (Hemerocallis 'Happy Returns', Iris 'Caesar's Brother', and Cornus sericea 'Cardinal'), only sodium consistently exceeded the toxicity limit for salt intolerant plants (500 mg kg-1). However, exceedance of the sodium toxicity limit was associated with plants' topographic positions, with median concentrations greatest in the bottom of basins and least on basin rims. This study demonstrates that deicing salts can have detrimental effects on plants in bioretention basins, with the strongest effects likely to occur in years with the greatest snowfall (and therefore deicing salt use), in portions of basins with greatest stormwater exposure (typically around inlets and centers), and in plants with minimal salinity tolerance. Our results therefore underscore the value of installing salt-tolerant taxa in basins likely to experience any frequency of deicing salt exposure.