Information about HadSST2: the ocean and sea component of the global surface temperature dataset HadCRUT3.
HadSST2 is a dataset of maps of anomalies - the temperature relative to the average during a reference period - in the temperature of the water near the surface of the oceans and seas for each month since January 1850. It is updated monthly with new data. This page covers the following topics:
Sea surface temperature (SST) observations - a term used for measurements of the temperature of the water near the surface of both the oceans and seas - come from about 1,200 drifting buoys deployed across the world's oceans and around 4,000 ships in the Voluntary Observing Ship programme. There are also numerous moored buoys in the tropics and in coastal regions, principally around the USA. Together they take around 1.5 million observations each month. These are checked by computer and any obviously inaccurate readings are excluded.
The monthly updates are combined with archives of historical observations that have been gathered over the past 160 years. The historical data are adjusted to minimise the effects of changes in the way measurements were made.
The observations are compared to the long-term average for the locations where they were recorded. The differences from the long-term average, known as anomalies, are averaged onto a grid that divides the earth's surface into 'squares' that are equally spaced in latitude and longitude at 5º intervals. The period used to calculate the long-term average is 1961 to 1990. There are often large areas from which we receive few, or no, observations.
The short answer is no. Before drifting and moored buoys were introduced, the vast majority of sea-surface temperature observations were made by voluntary observing ships. The observations were made for a variety of different purposes - for example to map ocean currents such as the Gulf Stream, for early weather forecasts - but they were not made to satisfy the later needs of global climate records. Consequently, the observations were made using a variety of methods, often dictated by expedience rather than accuracy.
An example of this is measurements made using a bucket to haul a sample of water to the ship's deck early in the data record. Many of these buckets were essentially canvas sacks and as they were hauled to deck evaporation from the sides of the buckets tended to cool the water sample, particularly on cold and windy days. Computer models of these buckets have been made and tested in wind tunnel experiments. These allow us to estimate the errors caused by using buckets to make SST measurements and to adjust the data accordingly.
As well as adjusting the data to reduce systematic measurement problems, the uncertainty in each temperature anomaly value is calculated. This includes an estimate of the uncertainty that arises from having to use observations made at discrete locations to estimate the average temperature anomaly over an area.
The image at the top right shows a map of anomalies in SST in the most recent month of the HadSST2 dataset. The anomalies shown are the differences between the actual temperatures and the averages in the same month during a reference period of 1961 to 1990, so the map is showing where the oceans and seas are currently cooler or warmer near the surface of the water than they were during this reference period. See our climate bulletins for information about what has happened in the climate this month.
Time series calculated from the HadSST2 dataset Enlarge The data for each month of the HadSST2 dataset can be averaged together to give an estimate of the global average SST anomaly for each year. The results are shown in the time series plot on the right. The plot shows our best estimates of the annual average temperature anomaly for each year (red bars) and the 95% confidence range for each (black). This confidence range represents how much certainty we have in the values. In 19 out of 20 cases we expect the true value of the annual average temperature anomaly to be within this range.
Also shown are the values after smoothing using a 21 point binomial filter to reveal long period fluctuations in the temperatures (thick blue lines; dashed where being close to the end of the series has affected the filtering) and the 95% confidence range in those values (thin blue lines).
Our page about global surface temperature has discussion about features that can be seen in time series of SST and also shows how changes in the temperature of the sea surface compares to the land.
A more detailed description of the methods used to generate the dataset can be found in:
Rayner, N.A., P. Brohan, D.E. Parker, C.K. Folland, J.J. Kennedy, M. Vanicek, T. Ansell and S.F.B. Tett 2006: Improved analyses of changes and uncertainties in sea surface temperature measured in situ since the mid-nineteenth century: the HadSST2 data set. Journal of Climate.19 (3) pp. 446-469
A version of this is available from the link above.