This is the first version of a hydrographic climatology for the region 47 to 65 °N, 15 °W to 15 °E (Citation: Manfred Bersch, Viktor Gouretski, Remon Sadikni, and Iris Hinrichs (2013): Hydrographic climatology of the North Sea and surrounding regions, Centre for Earth System Research and Sustainability (CEN), University of Hamburg). It contains monthly and yearly mean temperature and salinity data at 179 depth levels on a 0.25° x 0.5° latitude-longitude grid for the period 1890 to 2011. For the calculation of the means all available temperature and salinity data from water sample, CTD, and float measurements in this period were selected from the World Ocean Database 2009 (Boyer et al., 2009), the ICES (International Council for the Exploration of the Sea) data base, and the BSH (Federal Maritime and Hydrographic Agency of Germany) data base, rejecting duplicate stations (profiles). The derived data base consisted of about 735,000 stations (profiles).
The original profiles of temperature and salinity were interpolated on 179 depth levels:
0 to 10 m every 2 m,
10 to 200 m every 5 m,
200 to 500 m every 10 m,
500 to 1000 m every 20 m,
1000 to 5000 m every 50 m.
There was no interpolation if the gap between two measurements in the original profile was too large. This yielded about 13.3 and 12.8 Mio. temperature and salinity data, respectively, at the selected depth levels. The data were then sorted in 0.25° x 0.5° latitude-longitude boxes.
For these boxes an extensive data quality control was carried out, including various statistical checks for different time intervals and detection of erroneous data by visual methods. About 2.4 and 7.9 percent of the temperature and salinity data, respectively, were rejected. Though the quality of the data base was significantly improved, some undetected erroneous data may have remained and could have influenced the monthly and yearly mean values of individual boxes. Thus, the climatology has to be used with care, especially on small spatial scales.
Before calculating the monthly and yearly means of each box, the seasonal and intra-monthly variations of the data in the upper 200 m were eliminated by the subtraction of an appropriate polynomial fit. Some residuals of the seasonal and intra-monthly variations may have remained and thus contribute to the uncertainty of individual means. For the yearly means the uncertainty is increased if there were only data from few months in a box. The calculated monthly and yearly means were assigned to the centers of the boxes:
latitude: 47.125, 47.375, …, 64.625, 64.875 °N
longitude: -14.75, -14.25, …, 14.25, 14.75 °E.
Finally, some outliers were corrected in the horizontal distributions and the latter were slightly smoothed.
The derived data fields contain monthly and yearly means of temperature and salinity and the number of data used for the calculation of the means at 72 x 60 latitude-longitude grid points on 179 depth levels for 122 years.
The new climatology was compared coarsely with existing climatologies, e.g. HadISST1 (Rayner et al., 2003), showing no obvious erroneous deviations.
Further details can be found in the Documentation "Processing of the hydrographic data".
The meteorological part of this Climatology can be found here.
Due to a format error a part of the original salinity measurements (about 10 percent) were assigned to wrong positions and thus led to erroneous monthly and yearly means for a lot of boxes. So the salinities of version 1.0 should not be used further. The temperature values of version 1.0 were not affected by this error.
In version 1.1 this error of the salinities is corrected. All salinty data were again quality controlled as before. 6.2 percent of the salinity data were flagged as bad. Mean values were calculated in the same way as in version 1.0. The temperature values of version 1.0 were not changed and were included in version 1.1.
Version 2.0 of the North Sea climatology for temperature and salinity (Bersch et al., 2016) was extended to the period 1873 to 2013 and based on 1.23 Mio. hydrographic stations. Besides additional CTD, water sample, and drifter measurements XBT and MBT data were added from several sources as World Ocean Database 2013, ARGO, ICES, BSH, NIOZ, BODC, IFREMER, Danish Centre for Environment and Energy, Marine Scotland Science. The MBT and XBT data were corrected for depth errors after Gouretski and Reseghetti (2010). Only ARGO data of the delayed mode were used.
As in the version before station duplicates and erroneous measurements were rejected. The quality control of the data enclosed statistical methods based on the standard deviation in each geographic box and its surrounding at each depth level and in different time periods ( from total to monthly). Additionally, the profiles of the single boxes were intensively manually controlled. Profiles with more than 50 percent of erroneous data were omitted.
Before yearly and monthly means were calculated the annual and monthly variations, respectively, in the upper 300 m were eliminated by the subtraction of an appropriate polynomial fit. The calculated means were then despiked and smoothed weakly. The monthly means were calculated in the upper 600 m only. The error of the yearly and monthly means in the upper 100 m of a 2° x 2° area were estimated at about 0.06 °C and 0.027 psu for temperature and salinity, respectively.