Skip to main content
ARS Home » Midwest Area » Columbus, Ohio » Soil Drainage Research » Research » Publications at this Location » Publication #153101


item King, Kevin
item Harmel, Daren

Submitted to: Transactions of the ASAE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/12/2004
Publication Date: 10/30/2004
Citation: King, K.W., Harmel, R.D. 2004. Comparison of Time-Based Sampling Strategies to Determine Nitrogen Loading in Plot Scale Runoff. Transactions of the ASAE. 47(5):1457-1463.

Interpretive Summary: What is the best sampling procedure to use to accurately indicate the amount of pollutants being transported in runoff water? This study compared pollutant loads determined from samples taken at various time intervals to the pollutant load determined by capturing the entire runoff volume during a two-hour runoff duration. Sampling at a one minute interval was required to accurately indicate the total amount of pollutants. The results from this study will facilitate the selection of time-based sampling strategies for small plot studies. This result is important for designing sampling schemes for research studies and for water quality monitoring applications.

Technical Abstract: Estimates of water quality loadings are often presented without any knowledge of the relation of the estimates to the true loads. A laboratory runoff study was designed and conducted to compare true loads with several time-based sampling schemes. True loads were measured for three constituents (NO3+NO2-N, NH4-N, and PO4-P) on three 2.2-m2 plots topped with bermudagrass (Cynodon dactylon L. Pers.) sod and set at a 5% slope. Water samples were manually collected on 1-min intervals for a runoff duration of 2-hr. The true loads were compared to an array of time-discrete and time-composite sampling schemes. The schemes included time-discrete sampling at 1, 2, 3, 4, 5, 10, 15, and 30-min and composite sampling that included 2, 3, 4, and 5 sample composites based on the same time-discrete intervals as well as a single composite sample collected at 1-min intervals. Time-discrete sampling at 1-min intervals was the only discrete sampling scheme that produced load calculations not statistically different (alpha = 0.05) from the true loads. A single composite sample based on 1-min interval sampling and a 5 sample composite based on 1-min intervals were the only tested composite schemes that produced loads not significantly different from the true loads. To preserve the true load from plot-scale studies, more intensive sampling is required. The results from this study will facilitate the selection of time-based sampling strategies for small plot studies.