Watershed urbanization has been associated with declines in biodiversity and decreases in the proportion of pollution intolerant organisms. Some of these changes in community composition may be due to increased stormwater runoff from urbanized watersheds. In order to better understand how streamflow patterns change along a putative stormwater impact gradient and which abiotic factors may be driving shifts in macroinvertebrate community composition, we studied the hydrology, chemistry, and biology in six small streams in the Rocky River Watershed, northeastern Ohio. According to Ohio EPA classification, three of these streams were primary headwater streams, draining 0.41 to 1.22 km2; three headwater streams drained 3.47 to 7.25 km2. We measured flow at each stream using an Isco 2150 area-velocity flow meter from January 17 to December 9, 2010. We sampled each stream for macroinvertebrates and water and sediment chemistry three times: June 21 - 26, August 9 - 11, and September 30 - October 10, 2010.
Stream flashiness increased with watershed urbanization. The seasonality of high flow events was also altered in the more urbanized streams relative to the least urbanized streams, with least urbanized streams having the highest and urban streams having the smallest percentage of high flow days in the winter. The most urban primary headwater stream had larger average event peak flows than the least urbanized primary headwater
stream during the growing season, but not the dormant season. Peak flows did not differ among the larger headwater streams. Macroinvertebrate biodiversity, as measured by the Shannon Index, declined along the urbanization gradient. Chemical, hydrologic, and physical habitat variables were significantly correlated with macroinvertebrate community characteristics; however, the variable most strongly correlated with each macroinvertebrate community metric was a hydrologic variable. Potential multiple linear regression models to describe aspects of macroinvertebrate community composition were created from combinations of abiotic factors. The five models with the lowest Schwarz Bayesian Information Criterion and Mallow’s Cp statistics for each macroinvertebrate metric were examined further, and those that met the model assumptions were retained. In one multiple regression model, skewness of streamflow and total phosphorous concentration in stream water were highly predictive of Shannon’s Index, while in the another regression model skewness of streamflow, flow during sampling, and days since the ninety-fifth percentile of streamflow was exceeded were significant predictors of Shannon’s Index. The relative abundance of intolerant macroinvertebrate taxa was best described by skewness of streamflow. The increased flashiness of urban streams relative to least urbanized streams, the larger average flood peaks of urban primary headwater streams in the growing season, and the shifts in seasonality of high flow events provide evidence of hydrologic impairment of the more urban streams in this study. This hydrologic impairment appears to be one of the mechanisms by which urbanization influences stream macroinvertebrate communities. Therefore, increased use of low
impact design and more rigorous stormwater detention technologies within the watersheds may be needed to restore these streams.