Quantifying Spatial Variability in Runoff Quality in Semi-arid Urban Catchments

Urbanization alters ecosystem function and subsequently impacts quantity and quality of stormwater runoff. In arid and semi-arid urban ecosystems, solutes may accumulate in upland environments for several months between rainfall events, which flush these potential pollutants to stream channels. Although this decline in water quality is well documented with urbanization, it is unclear how different urbanization intensities affect catchment hydrologic responses and N dynamics in semi-arid regions. We expected that N would decline with increasing impervious cover. In heavily urbanized watersheds, N delivery would be controlled primarily by hydrologic transport and would exhibit a conservative flushing response. In watersheds with a lower density of development, N delivery would be related to complex source-sink relationships and would not exhibit a conservative flushing response. To address these hypotheses, we collected rainfall and stormwater runoff samples from 22 catchments in Tucson AZ that vary in percent impervious cover (IC, 2.92% to 90.7%) and catchment area (0.33km2 to 28.48km2.) We used a combination of 22 single-stage siphon samplers; one per catchment, paired with six multiple-stage automatic water collectors in a subset of catchments, to efficiently and inexpensively collect runoff samples across a wide these gradients of urbanization. There is a strong (r2 = 0.79) and significant (p < 0.05) correlation between the auto sampler and siphon sampler runoff quality. We found that pH and EC do not vary with catchment area, suggesting that the entire catchment may not be contributing surface runoff and solutes to streamflow. However, EC significantly increases with IC, while pH significantly decreases with IC. Data suggest that at high IC, pH of runoff begins to approximate the pH of rainfall (6.75). Given that urban catchments are designed to efficiently route runoff into waterways, it is plausible the increased IC results in more efficient mobilization of solutes, in a shorter time range, which may explain why pH begins to approximate rainfall at higher IC. We observe a significant decrease in EC over the monsoon (July to Sept) while at the same time observing a significant increase in pH, which may be the result of seasonal solute flushing. Preliminary data show that NH4-N does not vary with IC and catchment area, while PO4-P appears to decrease with catchment size and increase with IC, suggesting that some runoff solutes more directly impacted by land cover and biogeochemical catchment processes. Findings from our study suggest that imperviousness and seasonal dynamics impart a greater control on runoff quality than catchment size.