Nathan D. Rebuck1,2, David
W. Townsend1 and Maura A. Thomas1
2-NOAA NEFSC Narragansett
page updated 23 Feb 2012, data updated 23 Feb 2012
Figure 1 Data in the greater Gulf of Maine region.
Figure 2 Data in the Gulf of Maine
proper, color coded by year of collection.
Table 1 Data sources listed in text, along
with number of samples and years
represented in database.
Figure 3 Histogram of observations shown
in Figure 1, binned by month of collection.
This database is an archive of nutrient and hydrographic data collected in the greater Gulf of Maine region. The spatial distribution is shown in Figure 1 for the broader scope, and Figure 2 for the Gulf of Maine proper. It is intended to be a 'nutrient' database, however we have used temperature, salinity, and nitrate (reported as either NO3 or NO3+NO2) as variables necessary for inclusion, which may or may not include corresponding silicate or phosphate. Phosphate and silicate data (without a nitrate measurement) may exist but have not been included here. Because we have simplified the dataset for ease of examination, the metadata have been excluded here but raw data, including cruise numbers, program designations, and funding agencies when available are retained on file with the UMaine group.
The majority of data come from two public sources: the World Ocean Database maintained at the National Oceanographic Data Center (WOD, Boyer et al. 2006), and data assembled at the Marine Environmental Science Division at the Bedford Institute of Oceanography and the Marine Environmental Data Service in Ottawa, both a part of the Department of Fisheries and Oceans Canada (updated from the published Petrie et al. 1999). The two public sources were supplemented with additional data that have remained in provisional or unreleased status. Much of the new and unreleased data come from samples collected and processed by the DW Townsend Lab at the University of Maine (DWT). In addition, observations were provided by the Atlantic Zone Monitoring Program (AZMP), the University of New Hampshire Coastal Ocean Observing Center (UNHCOOA), data collected on Alexandrium and other survey cruises by D. McGillicuddy at the Woods Hole Oceanographic Institute (WHOI), the Massachusetts Water Resource Authority (MWRA), and the Marine Monitoring and Assessment Program at the Northeast Fisheries Science Center of NOAA (MARMAP).
The number of contributions from each source to the archive is displayed in Table 1. While much of the data are publicly available, some is not, and many individuals and institutions have expended time and effort (and money!) to collect, compile, and maintain these data. We are trying to keep this dataset as updated, comprehensive, and simplified as possible, without compromising the ability to verify any previously published results. Questions, comments, additions, subtractions (errors missed by any of the QA/QC methods described below), requests for raw data including data flags and metadata, and general inquiries should be directed to Nathan Rebuck, or to David Townsend or Maura Thomas.
The WOD database spans the years 1930‐2006 and here contains the Dec 2007 update of the WOD05. As described above, criteria for data inclusion are temperature, salinity, and nitrate. Historical samples were analyzed by a variety of analytical and manual methods, as can be expected with an archive reaching back to the 1930s. Comparison analyses show general agreement between analytical and manual wet chemistry methods, with results from historical and present methods falling within the experimental standard deviations at all but the lowest concentrations (Berberian and Barcelona 1979, Airey and Sandars 1987).
One source of observational error within the historical WOD05 and Petrie datasets is the reporting of NO3 data as a combination of nitrate (NO3) and nitrite (NO2). Some early data were reported as NO3 but conceivably measured as the sum total of both nitrate and nitrate, however metadata for the analytical methods of much of the historical data is unavailable (Johnson et al. 2005). Historical data from Petrie et al. (1999) are reported as NO3, however it is possible that the data represented a combination of both NO3 and NO3+NO2 measurements. For most marine environments, including the Gulf of Maine, the quantity of NO2 (typically <.2μM) is negligible relative to NO3 and introduced errors should be on the order of 1% in all but the most depleted or rapidly nutrient cycling environments. The conventions "NO3" and "nitrate" are used here to describe the measured oxidized inorganic nitrogen species (NO3 and NO2), with the understanding that some data may in fact be solely nitrate but that the error introduced by inclusion or exclusion is not significant at most spatial and temporal scales.
The publicly available data from WOD and Petrie have undergone quantitative methods designed to eliminate duplicate, erroneous, and mislabeled data prior to public release (Conkright et al. 2002, Johnson et al. 2005, Petrie et al. 1999). Similar methods were applied to all other preliminary and pre‐processed data after incorporation into the larger assemblage, such as removing data >3 standard deviations (SD) from the mean at standard depth intervals, where applicable. While these methods do not ensure all data are accurate, they do remove obvious outliers that would otherwise bias analyses. In addition to screening based on the probability of natural variability, all data were scanned for errors based on the following criteria unlikely to be encountered in the GoM: salinity>37, temperature>35C, nitrate>50μM, silicate>100μM (some estuarine data may exceed this threshold). Other data were scanned using a more subjective, yet logical criteria. Data were eliminated that were at a depth >150m and <5μM nitrate as these are unlikely to occur naturally, although they do not fall outside the 3 SD exclusion criteria. We have also eliminated zero values, as metadata suggests zero was often incorrectly used to represent a lack of data. Other quality control techniques applied to the WOD involve spatial interpolation and identification of bulls eyes in estimated fields, as well as illogical vertical profiles, such as density inversions. As the Gulf of Maine is a coastal shelf sea with a great seasonal and bathymetric influence, it is impractical and possibly misleading to attempt to apply all of these methods to the other preliminary data in such a geographically confined space. Figure 3 shows a histogram of the months of collection. As sea-going conditions are generally more pleasant and most nutrient studies center on the spring bloom and summer conditions, the data are more common in spring and summer months. Many of the fall data are provided by the MWRA in the vicinity of the diffuser in and around Massachusetts Bay, otherwise data are quite sparse in Sept, Oct, and Nov. More comprehensive discussion on the exclusionary criteria and methodology can be found in Rebuck (2011).
***February 2012 Update: Inclusion of 2010 field sampling from the GOMTOX program.***
If you have trouble loading, reading, or other general inquiries please email Nathan Rebuck (note that the data as a whole are too large for Excel).
Data columns are as follows, missing values indicated with NaN:
month(1); day(2); year(3); decimal longitude(4); decimal latitude(5); bottom depth(6); sample depth(7); temperature(8); salinity(9); NO3+NO2(10); Si(OH)4(11); PO4(12); extracted chlorophyll(13); PO4 quality flag(14); Si(OH) quality flag (15); NO3+NO2 quality flag (16) .
QA/QC flags: 0-passed all criteria; 1-outside UMaine range check; 2-WOD questionable cruise results; 3-WOD questionable cast results; 4-fail extreme criteria listed in text; 9-NaN, on land, zeros for all nutrients.
Airey D. and Sandars G.A. 1987. Automated Analysis of Nutrients in Seawater. CSIRO Mar. Lab. Report 166, 95 pps.
Berberian G.A. and Barcelona M. 1979. Comparison of Manual and Automated Methods of Inorganic Micro-Nutrient Analysis. NOAA Tech. Mem. ERL AOML-40, 26 pps.
Boyer, T. P., J. I. Antonov, H. Garcia, D. R. Johnson, R. A. Locarnini, A. V. Mishonov, M. T. Pitcher, O. K. Baranova, and I. Smolyar, 2006: World Ocean Database 2005, Chapter 1: Introduction, NOAA Atlas NESDIS 60, Ed. S. Levitus, U.S. Government Printing Office, Washington, D.C., 182 pp., DVD.
Conkright M.E., Locarnini R. A., Garcia H. E., O' Brien T., Boyer T. P., Stephens C., and Antonov J.I. 2002. World Ocean Atlas 2001: Objective Analyses, Data Statistics, and Figures CD-ROM Documentation, NODC Internal Report 17, Silver Spring MD, 17 pp.
Johnson, D.R.,. Boyer T.P, Garcia H.E., Locarnini R.A., Mishonov A.V., Pitcher M.T., Baranova O.K., Antonov J.I., and Smolyar I.V., 2005. World Ocean Database 2005 Documentation. Ed. Sydney Levitus. NODC Internal Report 18, U.S. Government Printing Office, Washington, D.C., 163 pp,
Petrie, B., Yeats P., and Strain P.. 1999. Nitrate, silicate and phosphate atlas for the Scotian Shelf and Gulf of Maine. Can. Tech. Rep. Hydrog. Ocean Sci. 203, vii, 96 pp.
Rebuck, N.D.. 2011. Nutrient distributions in the Gulf of Maine: An analysis of spatial and temporal patterns of dissolved inorganic nitrate and silicate. PhD Dissertation, University of Maine, Orono.