Seawater temperature records for the UK Shelf - 01 - Cefas Coastal Temperature Network dataset https://portal.medin.org.uk/portal/start.php?tpc=009_CEFAS9d5852de-2498-4726-840e-cacb63161b07 The Coastal Temperature Network consists of Cefas (and predecessor) originated data and data from external suppliers, who have agreed their data can be published as part of the network (Jones, 1981). The earliest data are from 1875 (Owers Light vessel) and have been supplied by the Met Office. The longest continuous record provided here is from Eastbourne (1892–2014). Sampling is from piers and breakwaters 50-200m from the shore where possible (Jones, 1981). The present network covers the temperature condition of coastal waters around the coast of England and Wales and was operationally combined with the salinity and temperature conditions across the Southern Bight of the North Sea. Individuals on behalf of Cefas, councils, companies and other organisations have obtained records of coastal sea surface temperature, for some stations, of more than 100-year duration. Approximately half of the stations started recording coastal temperatures in the mid–1960s. There are 41 stations in England and Wales where 20 out of 41 are still in operation. Cefas observers record coastal sea surface temperature using calibrated thermometers approximately 6 – 14 times per month, usually close to the time of high water. Other organisations record sea surface temperature ranging from daily values to monthly means. Since 2012, the data from Dover Council is recorded every minute. Data are published as monthly means (Joyce, 2006); the extracted data are the measurements used to calculate the means. The Cefas instruments are calibrated at Lowestoft to an accuracy of ±0.1°C. The accuracy of other instruments is not known, but is thought to be at least to an accuracy of ±0.2°C. The ferry route observers record offshore sea surface temperature from the ships main seawater pipe using a calibrated thermometer 4 times a month. The temperatures are recorded to at least an accuracy of ±0.2°C. The seawater samples are taken from the sea water main pipe to the harbour pump about 1.5 metres inboard. Quality assurance checks are applied to the data for each station by comparing the current dataset with either a 5 or 10 year running mean for each month. The data is first tested to see whether it is normally distributed i.e. whether all the data are close to average. The standard deviation is calculated to see how tightly the data are clustered around the mean; three standard deviations are then calculated to account for 99% of the data. If the data are outside this range (3 std dev) then the value is flagged and removed from subsequent analysis. See Joyce (2006) for details of the duration and history of individual datasets. Inevitably, there are changes in the number and location of monitoring stations over such a long period. At its peak the network reported on about 100 locations. This has reduced to around 30 in the late 20th century. Jones & Jeffs (1991) show the locations of early coastal stations. In addition, operating sites are moved and data recording upgraded, e.g. Eastbourne from a manual coastal site (see Joyce, 2006) to, in 2013, an electronic logging system mounted on an offshore buoy. These changes are reflected in the positions associated with the extracted data. The longest time-series include those from Eastbourne (1892 - present), Dover (1926 - present) and Port Erin, Isle of Man (1903 - present) although most time series began in the 1960s or 1970s. 20160310 20150108