There are at present three distinct methods for the determination of tritium in water samples: direct measurement of tritium abundance by accelerator mass spectrometry (AMS), radioactive counting of tritium decay rate, and the daughter product (3He) ingrowth method. The first method (AMS) has not been used for the measurement of environmental tritium levels but is better suited to measuring high tritium concentrations in small samples, largely for biomedical tracer research (Brown et al., 2005; Chiarappa-Zucca et al., 2002; Glagola et al., 1984; Roberts et al., 2000). The second method usually involves isotopic enrichment of hydrogen from the water samples, either by electrolysis (e.g., Östlund and Werner, 1962) or thermal diffusion (Israel, 1962) followed by low-level radioactive counting, either by liquid scintillation (Momoshima et al., 1983) or gas proportional counting (Bainbridge et al., 1961). Measurements are made relative to prepared standards (Unterweger et al., 1980), and accuracy appears to be limited by the reproducibility of the enrichment process to 3 %–10 % (Cameron, 1967).

The third method, 3He ingrowth, is a three-step method. First, it involves degassing of a quantity (∼1 to 1000 mL) of water to remove all dissolved helium. Second, the degassed water is stored in a helium leak-tight container (usually low-He-permeability aluminosilicate glass or metal) for a period of several weeks to a year or more. Experience indicates that it is necessary to shelter the stored samples from cosmic rays since there is a latitude-dependent cosmogenic 3He production rate that masquerades as “tritium signal” (Lott and Jenkins, 1998). Finally, the ingrown 3He is extracted from the water sample and mass spectrometrically analyzed (Clarke et al., 1976; Ludin et al., 1997). The last method, although it involves a potentially lengthy incubation period, is chemically simpler and does not involve isotopic enrichment steps. As such, it offers intrinsically greater accuracy (limited by standardization of the mass spectrometer, typically better than 1 %) and a lower ultimate detection limit (Jenkins et al., 1983; Lott and Jenkins, 1998).

Water samples are usually drawn from Niskin bottles into a helium leak-tight container either for shipboard (Lott and Jenkins, 1998; Roether et al., 2013) or shore-based gas extraction. The latter involves either clamped (Weiss, 1968) or crimped copper tubing (Young and Lupton, 1983). The extracted gases are subsequently purified and concentrated, usually cryogenically (Lott, 2001; Lott and Jenkins, 1984; Ludin et al., 1997), and expanded into a mass spectrometer for isotopic analysis. While time-of-flight mass spectrometry has been used (Mamyrin, 2001; Mamyrin et al., 1970), most oceanic helium isotope measurements have been made using multi-collector magnetic sector instruments (whereby ions are electrostatically accelerated and deflected by a magnetic field, e.g., Bayer et al., 1989; Clarke et al., 1969; Lott and Jenkins, 1998, 1984; Ludin et al., 1997). Measurements are typically standardized to marine air and corrected for any sample-size-dependent ratio effects determined by measurement of different-sized air aliquots. Depending on the amount of tritium in the water and the length of time a water sample is stored prior to gas extraction, it is generally necessary to correct the 3He/4He results for decay of tritium during storage. For very deep and old samples, e.g., at 3000 m in the North Pacific, where tritium concentrations are very low, this correction may be inconsequential.

To make full use of the helium isotopic ratio measurements, one needs to know at least the concentration of helium in the samples as well. This allows investigators to calculate the actual concentration of both 3He and 4He in samples and, with some assumptions, to estimate the amount of non-atmospheric 3He – whether it be from tritium decay or hydrothermal input – and 4He. The estimate can be further improved with knowledge of neon concentrations. Since neon is similar to helium in terms of its solubility in seawater and there are no known significant non-atmospheric sources of neon in the ocean, its concentration is a direct tracer of processes like air bubble injection at the sea surface. As an example, we show in Fig. 1 the relationship between the helium and neon saturation anomalies

The saturation anomaly of a gas is the percent deviation of the measured concentration relative to the expected concentration of the sample in equilibrium with air at one atmosphere, for example, as defined by ΔHe=100(C/C∗-1), where C∗ is the solubility equilibrium concentration at the temperature and salinity of the sample.

Helium and neon concentration measurements are typically made by mass spectrometric peak height manometry, that is, by comparison of major isotope ion currents (4He+ and 20Ne+) between the unknown to helium and neon derived from an aliquot of marine air. The air aliquot size is determined from a knowledge of the barometric pressure, relative humidity, and temperature at which the previously evacuated air standard reservoir was filled and the volume of the aliquot. Generally the ion current is assumed to be a linear function of the sample size (number of atoms) over some narrow range but also can be corrected by construction of a standard curve using different sized aliquots. Some measurements (notably the GEOSECS expedition) were made by splitting the gas extracted from the water sample and measuring the helium and neon contents separately using isotope dilution (with 3He and 22Ne spikes).

We have compiled a comprehensive dataset consisting of helium isotope and tritium measurements in oceanic waters made by numerous laboratories over the past 6 decades. The dataset includes ∼60000 tritium and ∼63000 helium isotope measurements, ∼57000 dissolved helium concentrations, and ∼34000 dissolved neon concentrations in ocean water taken from 1952 to 2015 (for tritium) and from 1967 to 2015 for helium. In addition to “spot sampling”, there are ∼380 cruises, with sampling from >5400 locations for tritium and ∼5600 locations for helium. The helium data are from 8 different laboratories and the tritium data from 15 laboratories worldwide. In addition to including measurement uncertainties, a data quality flag, and data source, each data point is accompanied by location (latitude, longitude, depth) and time (decimal year) of sampling. When available, water temperature, salinity, and dissolved oxygen measurements are included.

A number of the earliest measurements were obtained from publications. In those cases, the publication source is given. If the data were transcribed from tables, the table number and page is also given. In the event that the data were only available graphically, a computer program to digitize the data from plots was used, and in the rare cases in which graph quality was sufficiently poor to degrade the precision of the data, the uncertainties were commensurately increased to reflect it. Where data had been assigned a Digital Object Identifier (DOI), this is also included.

The dataset (provided in several digital formats described below) consists of three tables. The REFERENCES table is a list of the data sources keyed by the text variable “Reference_Code” found in the main data table. This provides attribution or more information regarding the data origin. The METHODS table provides a more complete description of the fields “Tritium_Method” and “Helium_Method” in the main data table. This is intended to provide useful interpretive information regarding how the sampling and/or measurements were accomplished. The main data table fields are described in Table 1. Most data fields have an associated quality flag field whose meaning is summarized in Table 2. Following WOCE ocean data convention, normal acceptable data are associated with a quality flag of 2, whereas questionable data have a flag of 3. Results obtained by averaging two or more replicates are signified with a flag of 6. When fields are missing for a given record, the data are entered as -999 and the corresponding quality flag is 9. The tritium, helium, helium isotope, and neon data also have an associated uncertainty field (e.g., “Tritium_Error”), which is the estimated uncertainty in the data points. This is either provided by the data measurer or an estimate based on described procedures and can vary greatly between methods and laboratories, so the user is advised to be aware of this value.

In the spirit of the WOCE, CLIVAR, and GO-SHIP

WOCE is the World Ocean Circulation Experiment (e.g., see https://www.nodc.noaa.gov/woce/, last access: 1 April 2019), CLIVAR is the Climate and Ocean – Variability, Predictability, and Change (e.g., see http://www.clivar.org/about, last access: 1 April 2019), and GO-SHIP is the Global Ocean Ship-Based Hydrographic Investigations Program (see http://www.go-ship.org/, last access: 1 April 2019).

The dataset's DOI is 10.25921/c1sn-9631 (Jenkins et al., 2018b). The data are available for download from the U.S. National Oceanic and Atmospheric Administration's National Centers for Environmental Information website at https://www.nodc.noaa.gov/ocads/data/0176626.xml or alternately http://odv.awi.de/data/ocean/jenkins-tritium-helium-data-compilation/ in a number of formats. For maximum flexibility, we suggest one of the following three database formats: Microsoft Access^®, PostgreSQL, or ODV (Ocean Data View). In addition, the three tables are available as four files (the main data table is split into two to avoid spreadsheet row-number limitations) in Microsoft Excel^® or as comma-separated plain text files. The data are also available in NetCDF format. Finally, the data table is available for download as a MATLAB^® binary data file.

Although not the first to measure 3He/4He in the environment (that was done by Aldrich and Nier, 1948), Clarke made the first reported helium isotope measurements in seawater (Clarke et al., 1969). He made his first measurements using a modified single stage magnetic sector, single collector mass spectrometer to a precision of about 2 %. Clarke developed the first compact all-metal branch tube mass spectrometer specifically designed to make 3He/4He measurements ultimately to a precision of 0.1 % to 0.2 %. At the time, conventional wisdom dictated that such measurements (let alone precision measurements) were not possible with a single stage magnetic sector instrument for such high (106) abundance ratios, but Clarke forged ahead anyway. He initially constructed two instruments in the early 1970s, using one at McMaster University in Hamilton, Ontario, Canada, and setting the other machine up at the Scripps Institute of Oceanography in La Jolla, CA, USA, for Harmon Craig and John Lupton. These instruments were used in support of the GEOSECS program and subsequently for a wide variety of other research projects. One of his students (William J. Jenkins) moved to the Woods Hole Oceanographic Institution in Woods Hole, MA, USA, where he extended Clarke's design to construct three other instruments. A post-doctoral investigator from his laboratory (Zafer Top) moved to RSMAS at the University of Miami and constructed a similar machine. Clarke also collaborated with a UK mass spectrometer company to make a commercially produced mass spectrometer available to the global scientific community.

In his early career, Clarke contributed to the study of meteorites and nuclear physics using isotope mass spectrometry. Beginning in 1969 Clarke published a series of ground-breaking papers on 3He/4He measurements in seawater and lakes (Clarke et al., 1969, 1970; Clarke and Kugler, 1973; Craig and Clarke, 1970; Craig et al., 1975; Jenkins and Clarke, 1976; Top and Clarke, 1983; Torgersen and Clarke, 1985). He contributed research to geology, hydrology, limnology, nuclear physics, medicine, and other disciplines and was active until 2002 publishing on 3He-related evidence for and against cold fusion (Clarke, 2001; Clarke and Oliver, 2003; Clarke et al., 2001).

In addition to developing a mass spectrometer capable of measuring 3He/4He ratios to order 0.1 % on sub-nanomolar gas samples, Clarke also created a method to measure environmental levels of tritium (3H) in water samples by the 3He regrowth technique (Clarke et al., 1976), which has become the de facto state of the art in low-level tritium measurements. In addition to being intrinsically simpler than the traditional low-level method (which involved electrolytic enrichment combined with low-level proportional gas counting), this method has proved to be more precise (by more than a factor of 4) and extended the detection limit to lower levels (by as much as an order of magnitude) (Bayer et al., 1989; Jenkins et al., 1983; Lott and Jenkins, 1998).

Clarke was inventive and ingenious in the laboratory with a remarkable ability to recognize opportunities where no one else could and to pursue them to ultimate success. He was an accomplished glass blower and constructed vacuum lines and research apparatuses from scratch using many different kinds of glass (see Fig. 10). Clarke kept unusual work hours when not teaching, generally arriving at the laboratory after lunchtime and toiling into the night. Working with him was a delight due to his good nature and whimsical sense of Irish humor but sometimes challenging because he was an inveterate prankster.

Östlund was a pioneer in US ocean tracer measurements, establishing a world-class, low-level counting laboratory dedicated to the measurement of tritium and radiocarbon at the University of Miami. He made distinguished contributions to ocean, atmosphere, and groundwater sciences, in particular to the understanding of the timescales of the transport of fluids through these systems. Östlund had a life-long devotion to the high-quality measurement of radioactive species. He received a BS in chemistry in 1949 and a PhD in chemistry in 1958. Both were from the University of Stockholm. Between 1959 and 1961 Östlund developed the electrolytic enrichment of tritium and deuterium for low-level environmental tritium measurements by gas proportional counting at the Radioactive Dating Laboratory of the Swedish Geological Survey. In the early 1960s, he came to the Rosenstiel School of the University of Miami. At the Rosenstiel School, he built a world-class tritium and radiocarbon counting laboratory that set new standards for low-level counting. His laboratory also processed the samples rapidly, and he generously shared his data with colleagues. As a result, relatively routine collection and analysis of large quantities of samples in a timely manner enabled oceanographers to use the tracer data to gain new insights into the timescales of oceanographic processes. His work paved the way for the acceptance of the next generation of tracer oceanographers, those measuring tritium and helium-3 by mass spectrometry and those measuring the chlorofluorocarbons.

He played a key role in the creation and execution of early global ocean survey programs. He developed electrolytic enrichment techniques for low-level environmental tritium measurements by gas proportional counting (see Fig. 11).

Östlund was a member of the scientific steering committee for the Geochemical Ocean Sections Study (GEOSECS), which was the first global-scale survey of chemical, isotopic, and radiochemical tracers in the ocean (1972–1978). He produced the first large-scale, high-quality mapping of the distribution of tritium in the oceans, which opened oceanographers' eyes to the dynamic and rapid penetration of bomb-produced tracers into the deep ocean.

Östlund published over 100 papers in peer-reviewed journals on a wide range of subjects. Although his early interests focused on many areas including atmospheric transport, they quickly extended to the hydrologic cycle and the ocean for which he applied radioactive tracers to a spectrum of scientific problems. As a student, he participation in the discovery of the anaesthetic Xylocain. Soon after coming to Miami, he used tritium to show that evaporation from the ocean is the major energy source for hurricanes. To collect samples he even flew into the eye of a hurricane. He and a colleague were the first to use tritium data to show that vertical mixing in the upper layers of the open ocean was an order of magnitude smaller than predicted by mass balance and theory. This was corroborated 20 years later by other investigators using new techniques. A major interest of his was the Arctic Ocean. There he quantified the contributions from ice melt, runoff, and precipitation to the freshwater budget. This budget plays a critical role in the global overturning circulation, which is the leading candidate for modulating decadal to centennial climate.

Östlund was involved in the planning and implementation and served on the scientific steering committees of early global change programs: Geochemical Ocean Sections (GEOSECS), which was the first global-scale survey of chemical, isotopic, and radiochemical tracers in the ocean (1972–1978), followed in the 1980s by Transient Tracers in the Oceans (TTO). The data his laboratory produced and collected under the auspices of these programs have furthered our understanding of the timescales of ocean processes. For example, the data have been used to estimate the flux of anthropogenic carbon dioxide into the ocean, the rate of exchange between the atmosphere and ocean, and rates of deep-water formation. He produced the first large-scale, high-quality mapping of the distribution of tritium in the ocean, which opened a new vista on the dynamic and rapid penetration of bomb-produced tracers into the deep western North Atlantic Ocean. Östlund's leadership as a member and coordinator of the scientific advisory committee for GEOSECS and TTO and his vision and credibility in seeing that an accelerator mass spectrometry facility for 14C analysis was established in the USA – had a large influence on ocean science. The big oceanographic programs of the past 50 years (GEOSECS through WOCE, CLIVAR, and GO-SHIP) have provided platforms for obtaining large quantities of high-quality tracer data.

Östlund was soft-spoken and gentle in demeanor and generous with his time, advice, and data. He set an example and benchmark for subsequent generations of tracer geochemists for responsibility, honesty, and fairness.