Premium
Influence of Sample Size on Measurement of Soil Denitrification
Author(s) -
Parkin T. B.,
Starr J. L.,
Meisinger J. J.
Publication year - 1987
Publication title -
soil science society of america journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.836
H-Index - 168
eISSN - 1435-0661
pISSN - 0361-5995
DOI - 10.2136/sssaj1987.03615995005100060017x
Subject(s) - denitrification , environmental science , denitrifying bacteria , soil science , spatial variability , log normal distribution , hydrology (agriculture) , environmental chemistry , chemistry , nitrogen , mathematics , geology , statistics , geotechnical engineering , organic chemistry
The influence of sample size on the magnitude and variability of soil denitrification was studied by collecting soil cores, ranging in size from 1.7 to 5.4 cm in diameter, from no‐till and conventionaltill corn plots. Estimates of natural denitrification rates were obtained by incubating intact soil cores with C 2 H 2 and monitoring gaseous N 2 O production. In addition, maximum denitrification potential was determined by monitoring N 2 O production in anaerobic slurries amended with glucose, NO ‐ 3 and C 2 H 2 . Natural rate estimates were highly skewed and approximated lognormal distributions. The spatial variability of denitrification was characterized by large variation at small distances of <10 cm and only weak spatial dependence at distances of 10 to 100 cm. Studies of the effect of sample size on denitrification suggest that soil cores >4.2 cm in diameter yielded the most reliable estimates of natural denitrification rates. Using a computerized random resampling technique, we estimated that approximately 10 to 15 kg of soil was necessary to obtain a representative soil mass for estimating natural denitrification rates. The results of this study are consistent with the hypothesis that the source of variability associated with the natural denitrification rates is the patchy distribution of denitrifying “hot spots” in soil. Some implications associated with the application of classical statistical methods to lognormal data are also discussed.