Search

Stable isotope analysis (SIA) is widely used in ecology, but samples are usually analysed in specialist facilities. Like any complex process, SIA is susceptible to human error. Despite the best efforts of the SIA laboratory, errors may occasionally remain in results released to clients.We used our experiences as ecologists and laboratory analysts to identify human error scenarios to which SIA performed anywhere might be susceptible. These scenarios ran from sample selection and submission by the ecologist, to the final results report from the SIA laboratory. Several scenarios involved errors hard to find by visual examination, by either ecologist or analyst.The ecologist can prevent pre‐submission errors by formalising their sample submission process, and can best perform a final results quality check by submitting their own quality control material with a matrix similar to that of the samples. Prevention of laboratory errors is best accomplished by a comprehensive risk‐based quality management system and a continual improvement, no‐blame laboratory culture.We recommend that the ecologist views the entire sample selection, submission and laboratory analysis as an integrated whole, establishing their own quality management processes to run alongside those of the SIA facility. This recommendation applies to any ecological methodology where samples are sent to a specialist facility for analysis. [ABSTRACT FROM AUTHOR]

Stable isotope analysis ( SIA) is widely used in ecology but is usually performed by specialist facilities. Consequently, the reliability of SIA data depends completely on the facility's quality systems unless ecologists take steps themselves. Although widely recommended, practical advice on how to do this is lacking., Here, we outline preparation of 150 vials of an in-house quality control material ( QCM) based on air-dried soil as an exemplar and describe methods of data analysis applicable to any QCM matrix. Homogeneity testing of the candidate material involves duplicate analysis of 10 randomly chosen vials, a Cochran test for outliers and 1-way anova. Given that many ecologists find SIA expensive, we propose that an ecologist preparing a soil QCM pretest for homogeneity using a simple colorimetric analysis of a soil quantity (bicarbonate-extractable P) sensitive to inhomogeneity., Once pretest homogeneity has been established, homogeneity under SIA can be tested. We obtained a 1-way anova P value of 0·824 in our laboratory pretest, then 0·782 ( δ13C) and 0·125 ( δ15N). QCM stability should also be confirmed., Submission of each batch of samples for SIA should be accompanied by a vial of the QCM, labelled as a normal sample. Control charting of the results then allows detection of within-batch issues encompassing any component of the analytical process (transport to the laboratory, sample preparation and subsampling, sample analysis, results calculation and reporting) and assessment of long-term analytical control. Any issue thus identified can form the basis of a conversation with the SIA laboratory about the affected batch of samples. [ABSTRACT FROM AUTHOR]

Marine nitrogen (N) has been traced into terrestrial and freshwater food chains at petrel (Aves: Procellariiformes) colonies on the South Island of New Zealand, but the effects of N or other marine-derived nutrients on the productivity of mainland forests are unknown. Mean ring widths in matai (Prumnopitys taxifolia; n = 3) and rimu (Dacrydium cupressinum; n = 2) trees growing in a Westland petrel (Procellaria westlandica) colony were 0.9-1.4 mm yr-1, similar to ring widths reported elsewhere. However, series of much wider rings in cores showed that trees in a petrel colony can experience periods of accelerated growth. δ15N values of whole wood were, at 3.2-11.1‰, significantly enriched in comparison to other temperate forest trees. The highest δ15N values were at the cambial and pith ends of different cores, so isotopic fractionation during N remobilisation is unlikely to have been significant. The radial extent of 15N enrichment suggests that petrels have bred at the site since at least the mid 18th century. [ABSTRACT FROM AUTHOR]