Table of Contents

• Coverage
• Dataset Description
  • Methods & Sampling
  • Data Processing Description
  • BCO-DMO Processing Description
  • Problem Description
• Data Files
• Related Publications
• Parameters
• Instruments
• Deployments
• Project Information
• Program Information
• Funding

Sample Collection

Samples were collected on 21 Hawaii Ocean Time-series (HOT) cruises, onboard the R/V Kilo Moana throughout 2020-2023. All samples were collected at Station ALOHA (22.45°N, 158.0°W) following trace metal clean procedures.

For all cruises except for October and December 2021, trace metal clean profile samples were collected using a powder-coated aluminum 12-place 'trace metal' rosette (Seabird) with 8 L external spring sampling bottles (Ocean Test Equipment), deployed using metal-free line (Amsteel). For the October and December 2021 cruises, samples were collected with C-Free bottles (Ocean Test Equipment) that were attached directly to the metal-free line and triggered with Teflon-coated messengers. Prior to sampling, sampling bottles were subjected to a ca. 18 hr soak in seawater collected at a test station near Oahu.

Upon recovery, sampling bottles were taken to a positive-pressure clean van onboard. Dissolved trace metal sample samples were filtered through 0.2 um Acropak filters into acid-cleaned 1 L LDPE bottles. Total dissolvable samples were collected unfiltered into acid-cleaned 1 L LDPE bottles. Samples were double-bagged and transported to the lab at the end of the cruise. Prior to each cruise, 1 L LDPE bottles were soaked in Citranox for one day, soaked in 10% HCl for 7 days, and rinsed multiple times with 18.2 M-Ohm Milli-Q water (Millipore).

Sample Processing and Measurement

Samples were processed via batch extraction following Conway et al. (2013) and measured by isotope dilution using Inductively Coupled Plasma Mass Spectrometry (ICP-MS). Briefly, samples were acidified to pH 1.8 using Optima grade HCl and left to equilibrate for a minimum of 2 months. Most samples were acidified for >6 months before processing.

500 uL of a stable isotope spike (containing ⁴⁷Ti, ⁵⁷Fe, ⁵⁸Fe, ⁶¹Ni, ⁶²Ni, ⁶⁵Cu, ⁶⁴Zn, ⁶⁷Zn, ¹¹⁰Cd, ¹¹¹Cd, and ²⁰⁷Pb) was added to the samples and allowed to equilibrate at least overnight (up to 1 week). Sample pH was adjusted to a pH of 6 using an ammonium hydroxide acetic acid buffer. Nobias PA-1 resin (stored in 1M HNO₃) was rinsed in water and added to the sample, which was then shaken for at least 3 hours or overnight. Resin beads were collected on a 5 um polycarbonate filter, rinsed with water and transferred to a disposable low-pressure chromatography column (Bio-rad, stored in 1M HNO₃). Extracted metals were eluted from the resin using 10 mL 1M HNO₃ with a 10 ppb In internal standard added. The same process was repeated regularly with 1 L Milli-Q water as a processing blank.

Samples were measured using iCAP-TQ ICPMS (Thermo Scientific) with In used as an internal standard. For Ti, Fe, Ni, Cu, Zn, Cd, and Pb, concentrations in the resin eluent were calculated using isotope-dilution equations, then converted to sample concentrations using the mass of each sample. For other elements (Mn, Co), concentration was calculated based on comparison to a 10 ppb multi-element standard (Inorganic Ventures), assuming analogous recovery to Fe and Ni (calculated from isotope dilution equations). The samples were largely processed and measured as two sets. The medians of 72 Milli-Q blanks for set 1 (70 pM Fe, 54 pM Zn, 36 pM Ni, 13 pM Cu, 1.4 pM Pb, 1.1 pM Cd, 11 pM Mn, 2.2 pM Ti) and 44 Milli-Q blanks for set 2 (125 pM Fe, 41 pM Zn, 34 pM Ni, 26 pM Cu, 1.3 pM Pb, 1.0 pM Cd, 6 pM Mn, 14 pM Ti) were subtracted from each sample in the corresponding set.

Accuracy of dissolved metal samples was assessed by measurement of GEOTRACES community consensus standards GSC (n = 3) and GDI (n = 4), which is similar to the GD standard.  For both standards, agreement with reported values was found for Fe, Cd, Ni, Cu, Zn and Pb. Concentrations for Ti were not available on the GEOTRACES website.

We observed variable recovery on Mn in both standards, likely due to the known pH sensitivity of Mn binding to the Nobias PA-1 resin. As a result, we subjected Mn measurements to additional scrutiny during quality control.

Samples were not UV-oxidized, which is essential for complete determination of dissolved Co concentration. As a result, data are reported as labile dissolved Co and total dissolvable labile Co.

This study represents the longest high-resolution timeseries of metals in the ocean. To generate a consistent data product across several years of sample collection, we subjected our analyses to a rigorous quality control scheme to flag and remove contaminated samples. We periodically observed higher concentrations in filtered samples, compared to unfiltered samples, which implies release of particulate TMs from within the capsule filter. To objectively identify impacted samples, we flagged all samples where dTM > (tdTM * X), where X for each element is noted below. We also occasionally observed "fliers" for unfiltered samples from clear contamination. To systematically identify impacted samples, we flagged all tdTM samples were the sample was greater than (the profile sample below + Y nM) and greater than (the profile sample above + Y nM), where Y for each element is noted below.

X = 1.2 (Cd, Cu, Ni, Pb), 1.3 (Fe, Ti, Co), 1.5 (Mn, Zn)
Y = 0.5 nM (Fe, Zn), 5 pM (Cd), 0.25 nM (Cu, Ni), 0.3 nM (Mn), 2 pM (Co), 0.02 nM (Ti, Pb)

Clear Zn contamination was noted for two cruises, HOT333 and 334, when samples were collected using an alternate sampling regime. Dissolved Ti  was not used until HOT333 (tdTi) and 334 (dTi). Data before these cruises is not reported. 

We also observed a consistent elevation of dPb and tdPb at 25 m on almost all cruises. This may be related to Pb contamination associated with the research vessel.

For Mn, poor recovery during column chemistry was the primary concern rather than contamination. Thus, for some samples, dMn > tdMn due to poor Mn recovery in the total dissolvable sample. We are expecting to update these fields with additional quality control schemes at a later date.

After the QC scheme was applied, we manually identified a small number of oceanographically inconsistent values, as well as oceanographically consistent values that were rejected by our QC scheme.

The percentage of samples rejected for each elements are as follows:

Fe: 21% of dissolved, 8% of total dissolvable
Cd: 2% of dissolved, 1% of total dissolvable
Cu: <1% of dissolved, 1% of total dissolvable
Mn: 3% of dissolved, 14% of total dissolvable
Zn: 15% of dissolved, 22% of total dissolvable (including the full profiles of HOT333 and 334)
Ni: <1% of dissolved, <1% of total dissolvable
Co: 1% of dissolved, 4% of total dissolvable
Ti: 16% of dissolved, 8% of total dissolvable
Pb: 9% of dissolved, 11% of total dissolvable

[v1 drafted version 2025-05-29]
* Sheet 1 of submitted file "HOTdTM_QCflags_nM.xlsx" was imported into the BCO-DMO data system for this dataset.  Had appeared as Data File: 962986_v1_hot-watercolumn-dissolved-metals.csv (along with other download format options).

* Data within sheet 2 "Flag Key" appeared redundant with already entered metadata so was not added (see Problems/Issues section for quality flag info)

* ISO DateTime with timezone (UTC) column added in ISO 8601 format.

* Column names adjusted to conform to BCO-DMO naming conventions designed to support broad re-use by a variety of research tools and scripting languages. [Only numbers, letters, and underscores.  Can not start with a number]

* After correspondence with the data submitter, metal columns were rounded to three decimal places, except for dCo, dCd, tdCo, and tdCd which was rounded to 5 decimal places.

* After discussion with the data submitter, Cruise ID changed to HOT-### to match the project convention

[Revised v2 2025-08-29]
* Dataset was revised before public release (scheduled for release 2026-06-01).
* Metadata Field "Data Processing" was updated to reflect the changes in this file version which explains the revised strategy for calculations for columns dTi, dTi_qc, tdTi, and tdTi_qc.
* Sheet 1 of submitted file HOTdTM_QCflags_nM_v2Ti.xlsx was exported as csv and imported into the BCO-DMO data system. This data will appear in this dataset as 962986_v2_hot-watercolumn-dissolved-metals.csv (along with other download format options).
* The same processing pipeline was used for the revised data, however, steps to round columns was not performed since the resubmitted file already had rounding to desired precision.
* As before, the previx HOT- was added to cruise ids.
* Column names were not altered from the revised file since the met BCO-DMO naming conventions.

NSF Award Abstract:
Phytoplankton are the base of marine food webs but their ability to grow in the open ocean by photosynthesis is limited by the scarcity of key nutrients especially iron. To understand how phytoplankton respond to global environmental changes, it is essential to predict how the nutrient content of seawater will change as well. Iron is essential to the light-harvesting machinery of phytoplankton but is an extremely small fraction of seawater (1 part per billion) . Iron is much more abundant in soils and when dust storms blow these soils out to sea, the iron content of seawater increases. It is unknown how long the effects of these iron supply events last, which depends on how well the marine ecosystem can recover and reuse iron before it sinks to the seafloor. It is also unknown if human activities have added to the natural Fe supply. The proposed research will address these questions by conducting a 3 year time-series of iron measurements in the North Pacific Ocean. Here, dust supply from Asia occurs mainly during spring, allowing the loss of iron over the summer and fall months to be documented. Unique chemical signatures will be used to distinguish iron supply from the deposition of desert dust or from human sources. This record of the marine iron cycle will be important for validating ecosystem models that are used to predict how climate change will influence the growth of phytoplankton in the future. The research would make a scientific contribution to the Hawaii Ocean Time-Series, help improve biogeochemical iron models, student training at the graduate and undergraduate level, and support an early career scientist.

A 3 year time-series of iron (Fe) measurements is proposed to constrain the magnitude of external Fe input and Fe recycling in the open ocean. Near-monthly observations will be conducted in the North Pacific Subtropical Gyre onboard Hawaii Ocean Timeseries cruises, which receives regular dust input during springtime and is minimally influenced by deep mixing. Water column profiling of dissolved and particulate Fe concentrations – combined with the flux of Fe recorded in trace-metal-clean sediment traps – will define a residence time of Fe in the upper water column. Iron uptake rates will be quantified through short-term incubations using a novel stable isotope technique and will be used to derive a turnover time with respect to biological uptake. Finally, the isotopic composition of dissolved and particulate Fe in the mixed layer will be measured to evaluate the potential importance of anthropogenic and Hawaiian Fe sources, which are poorly constrained. Together, these measurements will define the tempo and variability of the open ocean Fe cycle and provide a means to validate models that simulate the biogeochemistry of this key micronutrient.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

See more information about the Hawaii Ocean Time Series (HOT) on the related project page:  https://www.bco-dmo.org/project/2101 

Program description text taken from Chapter 1: Introduction from the Global Intercomparability in a Changing Ocean: An International Time-Series Methods Workshop report published following the workshop held November 28-30, 2012 at the Bermuda Institute of Ocean Sciences. The full report is available from the workshop Web site hosted by US OCB: http://www.whoi.edu/website/TS-workshop/home

Decades of research have demonstrated that the ocean varies across a range of time scales, with anthropogenic forcing contributing an added layer of complexity. In a growing effort to distinguish between natural and human-induced earth system variability, sustained ocean time-series measurements have taken on a renewed importance. Shipboard biogeochemical time-series represent one of the most valuable tools scientists have to characterize and quantify ocean carbon fluxes and biogeochemical processes and their links to changing climate (Karl, 2010; Chavez et al., 2011; Church et al., 2013). They provide the oceanographic community with the long, temporally resolved datasets needed to characterize ocean climate, biogeochemistry, and ecosystem change.

The temporal scale of shifts in marine ecosystem variations in response to climate change are on the order of several decades.  The long-term, consistent and comprehensive monitoring programs conducted by time-series sites are essential to understand large-scale atmosphere-ocean interactions that occur on interannual to decadal time scales.  Ocean time-series represent one of the most valuable tools scientists have to characterize and quantify ocean carbon fluxes and biogeochemical processes and their links to changing climate.

Launched in the late 1980s, the US JGOFS (Joint Global Ocean Flux Study; http://usjgofs.whoi.edu) research program initiated two time-series measurement programs at Hawaii and Bermuda (HOT and BATS, respectively) to measure key oceanographic measurements in oligotrophic waters. Begun in 1995 as part of the US JGOFS Synthesis and Modeling Project, the CARIACO Ocean Time-Series (formerly known as the CArbon Retention In A Colored Ocean) Program has studied the relationship between surface primary production, physical forcing variables like the wind, and the settling flux of particulate carbon in the Cariaco Basin.

The objective of these time-series effort is to provide well-sampled seasonal resolution of biogeochemical variability at a limited number of ocean observatories, provide support and background measurements for process-oriented research, as well as test and validate observations for biogeochemical models. Since their creation, the BATS, CARIACO and HOT time-series site data have been available for use by a large community of researchers.
 
Data from those three US funded, ship-based, time-series sites can be accessed at each site directly or by selecting the site name from the Projects section below.