The ELS was conducted in 1984 by the U.S. Environmental Protection Agency to describe the chemical status of lakes in the eastern United States, particularly with regard to acidification. Therefore, the presence of low-alkalinity, low-pH systems is expected to be overrepresented in this data set compared to lakes in the United States generally [Linthurst et al., 1986]. Nevertheless, the size and scope of this database provides a frame of reference against which to compare the results in the Shingobee headwater watershed. A total of 1797 lacustrine ecosystems were sampled in the northeastern, southeastern, and upper midwestern United States as part of this study. The ELS excluded from analysis water bodies meeting any of the following conditions: surface area <0.04 km2, intense anthropogenic disturbance, flowing water (stream), high conductance (>1500 μS cm−1), bay or estuary, or too shallow to obtain a sample free of debris at 0.5 m depth. Therefore, we did not include Little Shingobee Lake when comparing our results to the ELS results because it has a surface area of 0.03 km2 (Table 1). For the purposes of our study, we also excluded several other types of water bodies from the ELS data set: those determined to be swamps, reservoirs, or for which lake type was undefined, and those for which lake water residence time (RT) was not defined. All told 1532 lakes from the ELS data set were included in our analysis. Most lakes were sampled once just after fall turnover in the autumn of 1984. A water sample was collected with a Van Dorn bottle from 0.5 or 1.5 m depth, depending upon the depth of the lake, at the deepest part of the lake. Samples were analyzed for major anions and cations, pH, DOC, and other water quality parameters. The presence of surface water inlets and outlets were determined by visual analysis of maps, whereas lake area (AL) and watershed area were determined by planimetry [Linthurst et al., 1986]. Furthermore, an algorithm was developed that allowed estimation of lake water residence time (RT) [Linthurst et al., 1986]. For the purposes of our study, we used lake hydrologic setting and RT as a means of comparing our study site to other freshwater lakes in the eastern continental United States. We also calculated DOCX for the ELS lakes using the following equation:
where DOC is the measured DOC concentration (g m−3) of each lake sampled as part of the ELS, VL is the calculated lake volume (m3), AL is expressed in m2, and ZAVG is average lake depth (m), and RT is expressed in units of years. Linthurst et al.  describe the calculations and assumptions used to determine the ZAVG and RT of each lake in the ELS. Our chief assumption was DOC concentration was relatively constant in these lakes. This assumption is valid in the Shingobee Headwaters Watershed [Stets et al., 2009], but in other areas, there can be very large intraannual and interannual trends in DOC [Pace and Cole, 2002]. Therefore, these calculations provide a rough estimate of lake hydrologic characteristics, but taken in aggregate, they can convey a sense of DOCX in freshwater lakes. In our analyses of the ELS database, we calculated descriptive statistics for all lakes and for lakes categorized by hydrologic setting (open- or closed-basin depending upon the presence of a surface-water outlet).