Monthly and annual mean seawater temperature, salinity and density from 26 tide gauge sites during 1855-1993 (NCEI Accession 0000817)

Tidal observers at primary tide gauges of the United States Coast and Geodetic Survey (now the NOAA National Ocean Service) routinely measured seawater temperature and density. Many records extend back to the early part of the last century, the longest record extant being San Francisco (Carvalho and Maul, 1997). For the most part, these records have not been digitized nor analyzed for climatic signatures to our knowledge. Unfortunately, many of the 81 or so USA tidal station temperature records have missing data and/or are too short for analysis of linear trend. Modern tide gauges typically have seawater temperature sensors reporting in their data-stream, but the traditional analog systems relied on the skill of the observer to measure, record, summarize, and report temperature and density. The missing records are probably not the fault of field observers, and may well yet be discovered in misfiled reports.

Herein, records that span most of the last 100 years are entered into spreadsheet files, and are made available. Most of the digitizing is the product of work-study students at Florida Tech. Quality control and quality assurance utilize standard methods including plots and visual outlier detection. All data were converted from Fahrenheit to Celsius.

Data were provided as photocopies (and more recently digitized records) of the station sheets filed with each tide station folder. In most cases the data included monthly and annual mean temperatures, monthly and annual maximum and minimum temperatures, monthly and annual mean seawater density and maximum and minimum density. Almost all data in the photocopies are hand entries; monthly and annual means were mechanically calculated. Temperatures and densities from the tidal station sheets were placed into spreadsheets manually. Plots of monthly data were made to determine errors or outliers. Finally the digitized tables were compared to the photocopies to ensure accurate copying. As with meteorological stations, tide gauge stations are moved from time to time. The usual emphasis in sea level continuity is through benchmark checks by differential leveling. No such practice seems to have been emphasized with the seawater temperature or density observations, but considering the tradition of excellence within the U.S. Coast and Geodetic Survey since its founding by President Thomas Jefferson in 1807, it is reasonable to assume the historical information is reliable. Quantifying the standard errors associated with these monthly or annual means is problematic. The observer traditionally visited the tide gauge on a daily basis, and recorded the data by hand on summary sheets and on the marigram itself (USC&GS, 1929). Temperature and density typically are from bucket samples and thus represent true sea surface observations. Well-designed and calibrated thermometers are usually read at least to the nearest one-degree F or to the nearest 0.5 degree C, and estimating the number of observations per month to be 30, it seems reasonable to assume the monthly mean standard error is less than +/- 0.1 deg C. Observer error then is probably negligible for the purposes of determining trends.

The bounding latitude/longitudes are the extremes from the 26 fixed positions for the tide gauge stations.

In this accession, NCEI has archived multiple versions of these data. The latest (and best) version of these data has the largest version number.