The influence of storm characteristics and catchment structure on particulate organic matter transport in an arid city

Background/Question/Methods

In urban landscapes, flood infrastructure and altered surface permeability influence the movement of stormwater.  Stormwater is an important carrier of biotic nutrients and sediments.  Thus, factors influencing stormwater dynamics influence materials transport and retention within urban catchments.  Our goal is to understand how the structure of urban catchments controls material transport in an arid, urban landscape. Understanding these controls will contribute to more thorough understanding of urban ecohydrological processes and ultimately to better designs for stormwater infrastructure and management. We focused on particulate organic matter (POM) dynamics.  We collected stormwater samples and measured water flux during storm events from October 5 – December 29, 2010 at four sites in the Indian Bend Wash watershed of Scottsdale, Arizona.  These sites represented various catchment sizes with varying types of stormwater infrastructure.  Water samples and basic hydrologic data (water depth, flow, and rainfall) were collected with automatic water sampling equipment (ISCO autosamplers).  Water samples were filtered (GF-F filters) and the filters were dried, then ashed to quantify POM content. We calculated POM flux using concentration and water flow data.  Statistical models were used to extrapolate discrete measurements of POM concentration across the entire storm hydrograph to determine event-based POM export from the catchments.

Results/Conclusions

Export of POM from catchments showed that POM export is highly variable within and between catchments.  For a catchment with wash infrastructure, POM export was highest for a summer, monsoonal event (4.18 kg/ha) and lowest for a winter frontal event (0.07 kg/ha).  The hydrology of these events varied greatly; the summer event experienced high rainfall intensities and discharge, providing increased capacity for POM transport.  For this same summer event, POM export varied across catchments with values ranging from 1.10 kg/ha to 4.18 kg/ha. Room for improvement exists in the statistical models used to calculate POM flux.  Initial exploration into model improvement has yielded success.  POM concentrations were generally higher at the beginning of each event (indicating flushing of POM with the onset of runoff); we created an index of discharge per unit time elapsed to account for this ‘flushing effect’.  This index and rainfall data were used to create a multiple regression model to predict POM concentrations.  This approach generated a more accurate representation of POM transport, predicting 62% (R²=0.619, P<0.0001) of the variation in POM concentration for the aforementioned wash catchment during a summer, monsoonal event.  Predictive power may increase as analysis continues to determine the best predictor variables.