NRSA is part of EPA’s National Aquatic Resource Surveys which samples different surface waters on a 5-year cycle during the growing season. For logistical reasons, each NRSA cycle occurs over 2 years. Since nonwadeable systems were not sampled during the initial 2000–04 survey, we restricted our analyses to systems that have watershed areas <1,000 km2 across all 3 studied surveys for consistency in the sampled stream population over time. All of the national surveys described in this paper are overseen by the USEPA’s Office of Water http://www.epa.gov/nheerl/arm/designpages/monitdesign/survey_overview.htm) with the goal of creating unbiased assessments of aquatic resources across the 48 conterminous states.

Data on stream TN, NO3, ammonium (NH4), and TON concentrations originated from these summer NRSA surveys when one water sample was taken at each survey site. About 10% of sites were sampled twice during each survey for quality assurance and validation purpose—for these sites, we used only the first sample in this analysis. NRSA field and lab protocols are described in USEPA documents (67, 68). The detection limits for TN, NO3, and NH4 concentrations are respectively 0.05, 0.05, and 0.005 mg N L−1. Concentrations lower than the detection limits were censored and substituted with the detection limits. NH4 concentrations were generally quite low and near or below detection limits, so we do not examine the NH4 trend. TON concentrations were not directly measured during the survey and hence were calculated as the difference between TN and inorganic N (NO3 plus NH4). Negative TON values were found at 40 sites and were treated as values below the detection limit. We also included DOC concentrations and stream DOC/TON ratios from the surveys in the analysis to further understand changes in stream organic matter over time.