Premium
Water and Sediment Sampler for Plot and Field Studies
Author(s) -
Dressing S. A.,
Spooner J.,
Kreglow J. M.,
Beasley E. O.,
Westerman P. W.
Publication year - 1987
Publication title -
journal of environmental quality
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.888
H-Index - 171
eISSN - 1537-2537
pISSN - 0047-2425
DOI - 10.2134/jeq1987.00472425001600010012x
Subject(s) - sediment , silt , surface runoff , hydrology (agriculture) , sampling (signal processing) , environmental science , correlation coefficient , inflow , soil science , geology , mathematics , statistics , geotechnical engineering , geomorphology , ecology , oceanography , filter (signal processing) , computer science , computer vision , biology
The design and performance characteristics of a flush‐type sampling device for plot and field studies are described. The sampler is weld‐constructed and requires excavation and water conveyance for installation. It operates with no external power supply and collects consistently a known fraction of water and sediment passing through it. In laboratory tests, the sampler collected 2.65% (number of data points [ n ] = 54, standard deviation [ s ] = 0.0040) of all water passing through it at average flow rates ranging from 18 to 196 L min −1 . Sample volumes ranged from 0.75 to 18.7 L. Correlation analysis showed that sampling percentage was independent of flow rate ( n = 40, correlation coefficient = r = −0.04) over the range tested, in other laboratory tests, 30 sampling runs with inflow rates and total sediment concentrations ranging from 35 to 182 L min −1 and 252 to 1410 mg L −1 , respectively, showed that the ratios of waste to sample sediment concentrations were approximately one for total sediment (1.001), and for the sand (1.097), silt (1.008), and clay (1.020) fractions. Sand and clay ratios were shown to be statistically independent of total sediment concentration, but silt ( r = 0.30, n = 30) and total sediment ( r = 0.44, n = 30) ratios increased slightly with increasing total concentration. Monte Carlo simulation was performed to illustrate the suitability of the flush‐sampler for field and plot runoff studies. Simulation results indicated that for runoff estimates measurement error would exceed 10% with 33% probability for triplicate plots, but with only 16% probability in five plot studies. Additional simulation considering only measurement error associated with the sampler shows that the minimum number of paired samples required to detect sediment loss reductions of 50, 25, and 10% is 3, 4, and 16, respectively.