Premium
Mass‐spectrometer bias in stable isotope ecology
Author(s) -
Mill Aileen C.,
Sweeting Christopher J.,
Barnes Carolyn,
AlHabsi Saoud H.,
MacNeil M. Aaron
Publication year - 2008
Publication title -
limnology and oceanography: methods
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.898
H-Index - 72
ISSN - 1541-5856
DOI - 10.4319/lom.2008.6.34
Subject(s) - isotope ratio mass spectrometry , sample (material) , statistics , variance (accounting) , variation (astronomy) , isotope , ecology , environmental science , mass spectrometry , chemistry , mathematics , accounting , biology , physics , chromatography , quantum mechanics , astrophysics , business
Stable isotope analysis (SIA) is recognized as a powerful analytical tool with numerous ecological applications. This has been highlighted by the increase in popularity of the isotope ratio mass spectrometry (IRMS) technique and the large number of studies reporting isotopic data. Comparisons of new isotopic data with previously published results and the use of large volumes of isotopic ratios in meta‐analyses to explain isotopic variance are commonplace. Such data often originate from different IRMS instruments and are assumed to be readily comparable as all instruments are calibrated to International Atomic Energy Agency (IAEA) standards. To test the validity of this assumption, we analyzed a single ecological sample (homogenized cod muscle, Gadus morhua ) on eight anonymous IRMS instruments and found significant variation in both δ 15 N and δ 13 C. We used a one‐way analysis of variance (ANOVA) with random effects to estimate the average variability of laboratory results within and among instruments. Overall, 74% of variation in δ 15 N and 35% of variation in δ 13 C of a single ecological sample was explained by differences in the IRMS instrument used. In light of these findings, researchers are encouraged to submit their own sample reference to provide an independent check on variation between runs and between instruments; consistent discrepancies between instruments should be corrected through linear regression. Comparisons of data obtained from multiple instruments should acknowledge inter‐instrument variation as a potential source of error.