In wetlands, denitrification is an important ecosystem process that permanently removes reactive nitrogen from systems. Much research on denitrification has occurred in non-urban or highly managed urban wetlands. However, in urban landscapes nitrogen-rich water is often discharged into areas not designed or managed to reduce nitrogen loads. “Accidental” wetlands resulting from these discharges may have the capacity to remove nitrate, but can contain novel soils and vegetation, and are subject to unique hydrologic conditions that could create unexpected spatial and temporal patterns of denitrification. Our objectives of this study were to (1) examine spatial and temporal patterns of denitrification, and (2) identify predictors of denitrification in accidental urban wetlands.
We measured denitrification potential (DNP) on soils from nine wetlands forming at storm drain outfalls in Phoenix, AZ. Wetland sites were categorized into three hydroperiods: perennially flooded, intermittently flooded (~9 months/year), and ephemerally flooded (2-6 weeks/year). To assess spatial variation of DNP, we collected samples from 3-4 dominant vegetation patch types within each wetland. To assess temporal variation of DNP, samples were also collected during three seasons differing in rainfall pattern. To determine predictors of DNP, we measured plant traits hypothesized to affect denitrification, and physical and chemical properties of soil.
We found small- and large-scale spatiotemporal patterns in DNP that have important implications for management of urban wetlands for stormwater quality. Plant patches had higher DNP than unvegetated patches at intermittent and perennial wetlands, but unexpectedly not at ephemeral wetlands. Further, DNP was significantly higher in patches of Ludwigia peploides, indicating that plant species type may mediate within-wetland variations in NO3- removal capacity. We found a range of responses in DNP among wetlands to seasonal monsoon floods, which interacted with wetland hydroperiod. In perennially flooded wetlands, monsoon floods equalized DNP among plant patches; however, at ephemerally flooded wetlands overall DNP unexpectedly decreased after monsoon floods. Significant predictors of DNP were soil moisture, soil organic matter, and rooting depth; however, the relative importance of each varied among seasons. The plant traits we examined were not the most important significant predictor of DNP, suggesting we need a better understanding of how vegetation affects the soil environment and ultimately denitrification. Together, these findings offer novel insights into the complex interactions in accidental urban wetlands among plant patches, monsoon floods, and wetland hydroperiod.