An Examination of Trends and Patterns of Ecologically important streamflow variables in association with climate change predictions

Streamflow is a key factor in the biodiversity of freshwater ecosystems. In this paper we examine how ecologically important streamflow variables vary across the continental US. We also examine their trends and investigate how changes are associated with or driven by climate change. We selected 1512 reference quality stream gages from the USGS GAGES dataset [Falcone et al., 2010] and at each station used daily streamflow to compute 16 streamflow variables selected to quantify ecologically important components of the streamflow regime. We used rotated principal components of these variables with K-means clustering to classify the streams across the contiguous US. We also developed a statistical model to predict the streamflow regime classes for ungaged sites using watershed and climate attributes. The spatial variability of the 16 variables had a pattern consistent with climate and physiography drivers, while trends indicated changes that may contribute to changes in biodiversity in these streams. We observed daily mean flow increases in the Midwest and Northeast US. Regions with increases in daily mean flows correspond to regions with increase in higher values of precipitation. Principal Component Analysis with varimax rotation on the 16 normalized variables reduced the dimensionality to 4 components and six classes were mapped using K means classification. The model for prediction of stream class based on climate and watershed attributes provides information on how future climate conditions may affect streamflow regime important for biodiversity.