Findings show that results from sites that were chosen randomly indicated no widespread change in chloride at a national scale, but that results from sites chosen to focus on larger, more developed watersheds indicated significant increases in chloride at almost half of the sites.

Many water-quality monitoring organizations across the United States use two different sampling approaches to track water-quality conditions in streams.  The first approach, in which sampling sites are randomly chosen, is called a probabilistic design and provides a representative picture of water quality across a larger region, such as a state or the country as a whole. The second approach, in which sampling sites are chosen for a specific reason, is called a targeted approach, and can answer site-specific questions like “How has urban development in this watershed affected water quality?”

The USGS and USEPA study used changes in chloride concentrations to illustrate the complementary nature of probabilistic results from the National Rivers and Streams Assessment (NRSA) and targeted results from the USGS National Water-Quality Assessment (NAWQA) Project and other federal, state, tribal, and local monitoring programs. The probabilistic assessment, where more than 900 randomly distributed sites were sampled once every 5 years during summer baseflow conditions, provides a spatially unbiased view of chloride concentrations in the nation’s streams and rivers, but only for a particular season and type of flow. The targeted assessment, where sites were sampled more frequently during a range of seasons and stream flows, provides a temporally unbiased view, but mostly in larger and more developed watersheds.

The results from the probabilistic assessment indicate little change in chloride concentrations at the national scale from between the early 2000’s and the early 2010’s. The results from the targeted assessment, however, reveal that chloride concentrations increased during that same period at 132 sites—44% of those sampled. Integration of the two approaches suggests that chloride is not responding to a widespread, common driver across the country, but rather that chloride is responding more to factors that could be managed regionally or locally.

These combined results capture the behavior of the population of U.S. rivers and streams during summer baseflow conditions and supplement that with more detailed site-specific information on continuous long-term changes in rivers and streams within and across years at a more limited number of sites.  The most complete and robust monitoring approach to support future management of water quality at multiple scales would require changes to stream monitoring in the United States in a way that combines the benefits of both probabilistic and targeted monitoring—that is, sample collection at a large number of randomly distributed river and stream sites throughout the year on a continued long-term basis.