Santamaria Cerrutti ME, Moreau M. 2022. Comparison of ammonia-nitrogen concentrations between unpreserved and acid-sulfuric preserved groundwater samples. Lower Hutt (NZ): GNS Science. 15 p. (GNS Science report; 2021/27). doi:10.21420/3RJM-G633.
Abstract
In 2016, the National Groundwater Monitoring Programme (NGMP) sampling kit was modified by adding a 100 mL sulfuric-acidified sampling bottle for the analysis of ammonia-nitrogen (NH3-N). This change was prompted by a review applying the most recent version of Standard Methods (Rice et al. 2012), which is the basis for the sampling protocol adopted by the programme since 2006. Acidification was recommended to inhibit bacterial activity and prevent salt formation with organic bases. This report aims to inform on (i) the impact of this operational change on measured NH3-N concentrations at NGMP sites and (ii) the implication of this change on the long-term NGMP dataset. NGMP samples, including the new bottle, were collected by regional council staff as part of quarterly monitoring. Collected samples were then analysed at the New Zealand Geothermal Analytical laboratory for NH3-N, one of the monitored parameters using acid-preserved bottles, and for other parameters, including major ions and trace metals. Duplicate analysis of NH3-N (flow-injection analysis APHA 4500-NH3-H method) was performed on a subset of the unpreserved bottles for cost-effectiveness. The subset was selected to encompass a range of concentrations (<0.003 to 13.7 mg/L) and hydrogeological settings, while ensuring a dataset size to enable statistical testing. The dataset consisted of 251 matched pairs of NH3-N concentrations (unpreserved and acid-preserved groundwater samples). Overall, the distribution of concentrations was close to a 1:1 relationship. Identical pairs were observed in 30% of the dataset, and associated percentiles were very consistent between unpreserved and acid-preserved samples. Focusing on the matched pairs with differences, concentrations of acid-preserved samples were smaller than those of the unpreserved samples (Wilcoxon p-value = 0,001; confidence level of 95%). However, the magnitude of this difference fell within the uncertainty of measurement (estimated ±0.003 mg/L), and the calculated effect size was small (Wilcoxon effect size r = 0.29). These suggest that the acid preservation does not significantly impact NH3-N measured concentrations at NGMP sites and that it is unlikely that the long-term dataset is impacted by the change of sample collection from 2016. Some of the samples recorded a laboratory arrival temperature and sample holding time higher than the threshold listed in Standard Methods. However, when sub-setting the matched-pair dataset based on these more stringent criteria, similar results were obtained for statistical testing. Recommendations of this report are to: Accompany any change in operations with an analysis of potential impacts on the monitoring programme resources. For cost-effectiveness, this study supports reverting the NGMP sampling kit from acid-preserved to unpreserved samples for NH3-N analysis. Accompany any operational change that may occur within a programme with data collection in order to assess the impact of this change on the long-term dataset. This study suggests that the NGMP sample collection modification had a negligible impact on the integrity of the long-term dataset and therefore no correction factor is needed. Conduct similar comparisons to assess the impact of arrival temperature and sample holding time on laboratory analysis in NGMP samples. (The authors)
