Dataset: Mercury, methylmercury, nutrients, and physicochemical measurements from water column samples collected in the Gulf of Maine and Penobscot Estuary during several cruises in 2023 and 2024

Description:
Total mercury, methylmercury, and ancillary variable data were collected in the Gulf of Maine along the Gulf of Maine North Atlantic Time Series (GNATS) transect and on a transect in the Penobscot Estuary and River. Sampling took place over four cruises on the R/V Endeavor (April 2023, August 2023, November 2023, April 2024), and two separate trips in a small boat to sample further upstream in the Penobscot River (September 2023, May 2024). 21 stations were sampled in total, with multiple depths (up to 6) extending throughout the water column at all stations except for the shallow Penobscot River stations.

This dataset includes bulk and particulate total mercury and monomethylmercury, dissolved organic carbon, particulate organic carbon and nitrogen, chlorophyll-a, phaeopigments, nitrate, phosphate, silicate, fluorescence indices (including the freshness index, BIX, and HIX), and components identified during Parallel Factor (PARAFAC) analysis. The percentage of methylated mercury (bulk methylmercury/bulk total mercury), percentage of particulate methylated mercury (particulate methylmercury/particulate total mercury), and percentage of each PARAFAC component are also included. The dataset also includes salinity, temperature, and dissolved oxygen data from the sensors on the CTD, as well as the calculated apparent oxygen utilization. Nutrient data in August in the Gulf of Maine was replaced with data supplied by DFO from a transect similar to the GNATS transect, which occurred in September 2023.

Methods & Sampling:
During the four cruises, samples were collected using a trace metal clean rosette with Go-Flo bottles. During the August 2023 cruise, the trace metal clean rosette and winch system suffered a technical issue that was unable to be resolved, thus Stations 3, 4, 6, 7, and 10-15 were collected with the same Go-Flo bottles attached to a non-trace metal clean, Seabird rosette. Water samples were collected over the side of a small watercraft for the Penobscot River sites, taken either by hand or via Go-Flo bottles. Bulk water samples were collected in 2-liter (L) acid-cleaned Teflon bottles and then brought into a trace metal clean van (or laminar flow hood for the Penobscot River sites) for typically immediate, but if not same-day, filtering and subsampling.

Bulk Methylmercury samples were preserved with sulfuric acid (0.5%) and kept cold until analysis. They were analyzed on a Tekran 2700 Automated Methylmercury Analysis System following standard techniques (Hammerschmidt and Fitzgerald, 2006). Briefly, bulk MeHg samples were digested overnight with sulfuric acid, buffered with acetate buffer, followed by the addition of ascorbic acid, neutralized with potassium hydroxide, and ethylated using sodium tetraethylborate. Once on the Tekran 2700, samples were purged with argon, and the Hg species were trapped on a Tenax trap. Upon heating the trap, the Hg species were released and underwent gas chromatographic separation prior to decomposition of the Hg species to elemental Hg for cold vapor atomic fluorescence detection. Samples were run within 6 months of collection at the University of Connecticut Avery Point.

Particulate methylmercury was filtered through a pre-weighed and pre-combusted (400 degrees Celsius (°C), 15-18 hours), 0.45-micrometer (μm) pore-size, quartz fiber filter using a vacuum pump and frozen until analysis. Filters were freeze-dried for 7 days, weighed, and then digested in 4.5 N nitric acid overnight in a 60°C water bath, buffered with acetate buffer, neutralized with potassium hydroxide, and ethylated using sodium tetraethylborate, where the sample then underwent the same process on the Tekran 2700 as described above. The remaining digest - 3 milliliters (mL) out of 7 mL total - was frozen and preserved within 1 week for THg analysis by diluting to 10 mL with DI water, adding bromine monochloride, and storing in the dark at room temperature. Filters were run within 1 year of collection at the University of Connecticut Avery Point.

Bulk total mercury samples were filled with no headspace and kept cold until analysis on a Tekran 2600 following EPA method 1631, refined by Hammerschmidt and Fitzgerald (2006). Briefly, both particulate and bulk samples were digested overnight with bromine chloride, reduced with hydroxylamine hydrochloride to remove excess reductant, oxidized with stannous chloride to form elemental Hg, and run using cold vapor atomic fluorescence spectrometry. Samples were run within 6 months of collection at the University of Connecticut Avery Point.

Particulate total mercury was analyzed using the same filter as for particulate methylmercury and run using the method described above. Filters were run within 1 year of collection at the University of Connecticut Avery Point.

Chlorophyll-a and phaeopigments were collected by filtering 100 mL of sample through a 25-millimeter (mm), 0.7 μm pore-size glass fiber filter using a vacuum pump and immediately covering with tin foil and freezing. The filters were pre-combusted (450°C, 2 hours) for the April 2024 cruise and 2024 Penobscot river samples only. Samples were analyzed using EPA method 445. Briefly, the filters were digested overnight in 90% acetone and run in dimmed light with and without hydrochloric acid acidification using a Trilogy Fluorometer. Samples were only collected for the shallowest 3 to 4 depths at each station for all cruises except April 2023 because the chlorophyll-a was found to be below the limit of detection (LOD) at depth during the April 2023 cruise. The bottom depths for chlorophyll-a were replaced with ½ the LOD, based on the CTD fluorescence scans, which confirmed the lack of chlorophyll-a at depth. Phaeopigment data at depth were left blank due to inability to confirm or deny the lack of phaeopigments at depth. Chlorophyll and phaeopigment concentrations were calculated using equations based on the EPA method 445. Some phaeopigment values came back negative; these samples were assumed to contain no phaeopigments. Samples were run within 1 month of collection, except for in April 2023 where they were run within 3 months of collection at the University of Connecticut Avery Point.

Particulate organic carbon and nitrogen samples were collected by filtering 200 mL (300 mL in November 2023) of sample through a 25 mm, 0.7 μm pore-size, pre-combusted (450°C, 2 hours) glass fiber filter using a vacuum pump and immediately covering with tin foil and freezing. POC/N filters were saturated with hydrochloric acid for 4 to 12 hours, allowed to dry overnight in the hood, and dried at 45°C for at least 3 days in the oven. Filters were weighed and re-dried for a couple of hours to ensure the drying process was complete and kept in a sealed desiccator until analysis (within 1 year of collection). Samples were run on a Fisons NA 1500 series 2 elemental analyzer by Costech Analytical Technologies using procedures based on the EPA method 440. Several samples did not have any peaks associated with nitrogen; these samples were replaced with ½ the LOD. Samples were processed within 14 months of collection at the University of Connecticut Avery Point.

Dissolved organic carbon samples were filtered through a 25 mm, 0.7 μm pore-size, pre-combusted (450°C, 2 hours) glass fiber filter and were frozen until analysis. Samples were acidified to 0.5% with hydrochloric acid and analyzed with a Shimadzu TOC analyzer within seven months of collection at the University of Connecticut Avery Point.

Nitrate, phosphate, and silicate samples were filtered through a 25 mm, 0.7 μm pore-size glass fiber filter and were frozen until analysis. Filters were pre-combusted (450°C, 2 hours) for the April 2024 cruise and Penobscot river trip only. Samples were analyzed via standard colorimetric methods using a Shimazdu UV-1900i Spectrophotometer (Hansen & Koroleff, 1999). Samples were analyzed within approximately one year of collection at Dartmouth College. Results from April 2023 and August were not included due to concerns about data quality. August data was replaced with data from a cruise by DFO from September 27th to the 29th, 2023, which followed a transect close to the GNATS transect (Figure S1). Data was substituted if it was collected from within 15 meters of this study’s depths.

Dissolved organic matter samples were filtered through a 25 mm, 0.7 μm pore-size, pre-combusted (450°C, 2 hours) glass fiber filter and were frozen until analysis (they were defrosted for short periods to obtain a subsample for DOC analysis). Samples were run using a Cary Eclipse Fluorescence Spectrophotometer (excitation: 220-550 nanometer (nm), emission: 300-700 nm, slit width: 5 nm, increment by 5 nm, PMT voltage: 800, scan speed: 900 nm/minute) within 14 months of collection at the University of Connecticut Avery Point. Absorbance scans were run using a Shimazdu UV-1900i Spectrophotometer from a wavelength of 200 to 700 nm within 18 months of collection at Dartmouth College. Only samples from the upper 30 meters of the water column were analyzed. Data was processed by Urban Wünsch (Technical University of Denmark) using Parallel Factor (PARAFAC) analysis to obtain fluorescence indices and PARAFAC components.

Salinity, temperature, and dissolved oxygen were collected using either the CTD or a hand-held Sonde (for the Penobscot River sites).

Apparent Oxygen Utilization was calculated using the equation and constants from Garcia & Gordon (1992).

Hg data was blank and spike corrected and ancillary variables were blank-subtracted, with the exception of the nutrient data. THg and MeHg bulk samples were run in triplicate. For the other variables, triplicate samples were collected at varying stations and depths throughout the cruises. Any data below the limit of detection (LOD) was replaced with the LOD. The LOD was defined as 3 times the standard deviation of the blanks. A table with the average QA/QC parameters is included (Table 1).

Table 1. Average QA/QC data for all variables.

	LOD	Field Blank	Method Blank	% Recovery of Standard
MeHg (pmol/L)	0.06	0.07 ± 0.04	0.05 ± 0.02	102 ± 12%
pMeHg (pmol/L)A	0.005	0.014 ± 0.004	0.003 ± 0.002	103 ± 10%
THg (pmol/L)	0.4	0.4 ± 0.4	0.1 ± 0.1	101 ± 7%
pTHg (pmol/L)A	0.04	0.13 ± 0.08	0.12 ± 0.06	97 ± 9%
Chlorophyll (µg/L)	0.02	0.006 ± 0.005	0.001 ± 0.002	B
Phaeopigments (µg/L)	0.02	0.01 ± 0.01	0.012 ± 0.008	NA
DOC (µM)	7.06	11 ± 11	2 ± 2	97 ± 9%
POC (µg)C	26.1	14 ± 9	15 ± 4	100 ± 6%
PON (µg)C	2.16	0.2 ± 0.7	0.5 ± 2	98 ± 5%
NOx (µM)	0.62	0.3 ± 0.2	NA	108 ± 10%D
PO43- (µM)	0.016	0.043 ± 0.005	NA	105 ± 5%D
SiO42- (µM)	0.08	0.01 ± 0.03	NA	110 ± 27%D

^A pMeHg and pTHg values are lower than the bulk counterparts because the particulate values were processed the same as the samples (accounted for the ratio of the total volume of digest versus the volume of digest analyzed) to facilitate comparison with the processed particulate data.

^B For chlorophyll, a solid standard was used to ensure fluorometer stability prior to each run. The fluorescence of the solid standard was within the acceptable range prior to use, with the acceptable range based on a correlation between a liquid chlorophyll standard and the Trilogy fluorometer.

^C POC/N field blank values were converted to µg/L prior to being subtracted from sample data.

^D The %recovery of standard for nitrate (NOx), phosphate (PO4), and silicate (SiO4) refers to the recovery of spiked samples rather than recovery of a standard.