Macroscale water color patterns in large US rivers

Abstract:

Large river basins are ideal for studying macrosystems ecology. The patterns we observe are the result of the interaction of local and distal processes integrated over millions of square kilometers of land and water. However, there are few examples of spatial ecological measurements in rivers for understanding macroscale phenomena. Optical imagery from satellites, such as Landsat, can measure the color of water as perceived by the human eye. Water color is an intuitive measure of water quality that encodes information about the amount of suspended sediment, chlorophyll-a, and colored dissolved organic matter. Using Google Earth Engine and the Landsat 5,7,8 surface reflectance product from 1984-2018, we approximated river color as the dominant wavelength (400-700 nm) in 32,245 river and lake reaches (i.e. connected lakes only) totaling 105,000 river kilometers and 15.9 million color observations. Our goal was to provide an initial description of broad trends and patterns. We asked 1) What is the most common color of rivers? 2) How does river color vary over time and space? We found 1) rivers were dominantly yellow (56% of observations) or green (38%), and greenness was associated with connected lakes; 2) river color had three distinct seasonal patterns that were synchronous with regional flow regimes; 3) one-third of large US rivers had significant color shifts over the last 35 years. Our observations may help disentangle regional and local drivers of trends in water quality, pinpoint hotspots of environmental change, and contribute to the study of macrosystems ecology in rivers.