Realistic measurement uncertainties for marine macronutrient measurements conducted using gas segmented flow and Lab-on-Chip techniques

Abstract

Accurate and precise measurements of marine macronutrient concentrations are fundamental to our understanding of biogeochemical cycles in the ocean. Quantifying the measurement uncertainty associated with macronutrient measurements remains a challenge. Large systematic biases (up to 10%) have been identified between datasets, restricting the ability of marine biogeochemists to distinguish between the effects of environmental processes and analytical uncertainty. In this study we combine the routine analyses of certified reference materials (CRMs) with the application of a simple statistical technique to quantify the combined (random + systematic) measurement uncertainty associated with marine macronutrient measurements using gas segmented flow techniques. We demonstrate that it is realistic to achieve combined uncertainties of~1-4% for nitrate + nitrite (ΣNOx), phosphate (PO 4 3-) and silicic acid (Si(OH) 4) measurements. This approach requires only the routine analyses of CRMs (i.e. it does not require inter-comparison exercises). As CRMs for marine macronutrients are now commercially available, it is advocated that this simple approach can improve the comparability of marine macronutrient datasets and therefore should be adopted as 'best practice'. Novel autonomous Lab-on-Chip (LoC) technology is currently maturing to a point where it will soon become part of the marine chemist's standard analytical toolkit used to determine marine macronutrient concentrations. Therefore, it is critical that a complete understanding of the measurement uncertainty of data produced by LoC analysers is achieved. In this study we analysed CRMs using 7 different LoC ΣNOx analysers to estimate a combined measurement uncertainty of < 5%. This demonstrates that with high quality manufacturing and laboratory practices, LoC analysers routinely produce high quality measurements of marine macronutrient concentrations.

Exploring Ocean Biogeochemistry Using a Lab-on-Chip Phosphate Analyser on an Underwater Glider

“…Seawater samples were analysed using a gas segmented flow manifold, Seal AA3 (Norderstedt, Germany), combined with a standard spectrophotometric technique (molybdenum blue) (Hydes et al, 2010). The chemistry is almost identical to that used in the LoC analyser, and previous studies have demonstrated that data generated via gas segmented flow analysis is directly comparable to that generated via LoC analysers (Clinton-Bailey et al, 2017;Grand et al, 2017;Birchill et al, 2019a). The limit of detection for gas segmented flow analysis is 0.01 µM in seawater (SEAL-Analytical, 2019).…”

“…The oceanographic research community is experiencing a dramatic increase in capability as autonomous technologies Grand et al, 2017). Error bars are laboratory estimates of combined (random + systematic) relative uncertainty (Birchill et al, 2019a). (B) Statistically significant inverse linear correlation between temperature and PO 4 3− concentration from dives 691, 692, and 693…”

“…The analytical figures of merit (accuracy, precision, limit of detection, and quantification) allow for the determination of PO 4 3− concentrations in all but the most oligotrophic regions of the ocean, and in situ results have been validated using water samplers and benchtop spectrophotometric reference methods (Grand et al, 2017). The combined (random and systematic uncertainty) measurement uncertainty associated with LoC PO 4 3− measurements from multiple LoC analysers was estimated to be 6.1% (u c ) (Birchill et al, 2019a).…”

Toward a Harmonization for Using in situ Nutrient Sensors in the Marine Environment

“…In this way, the quality control of field measurements will become as stringent as those for laboratory analysis and this for all the analytical approaches. One approach to determine uncertainty of nutrient measurements from LoC sensors is also presented by Birchill et al (2019). Quality control charts for the analysis of QC samples used by wet chemical analyzers should be recorded and checked to make sure that the analytical performance is continuously maintained and in accordance with the requirements.…”

“…Over the last decade, another important milestone in the quality control of nutrient measurements has been the development and application of CRMs for nutrients. A range of CRMs are now available, which, when used and applied correctly, are starting to improve the quality of nutrient analysis in laboratories and allows for the determination of analytical measurement uncertainty (Aoyama et al, 2018;Birchill et al, 2019).…”