The 6k data are based on two types of data: 1) alkenones, a geochemical proxy, and 2) transfer function estimates of sea-surface temperature based on faunal assemblages of planktic foraminifers. All alkenone data have previously been compiled within the GHOST project (Kim et al., 2004a,b) (http://www.pangaea.de/Projects/GHOST/). For the foraminiferal data, raw census data were composed and two different transfer function techniques (Maximum Likelihood (ML) and the Modern Analog Technique (MAT)) were used to calculate the sea-surface temperatures to illustrate, at least to some extent, the magnitude of the differences in estimated SST that are related to the choice of calculation method. Hence, in the data set you will find both the two separate calculations for the foraminifers as well as the average 6k value combining data from the two different techniques. When it comes to errors of different SST estimation techniques, it is fair to say that most methods operate with a statistical error close to +- 1 degree C. However, due to the geographical autocorrelation these types of data possess, the assumption of data independence is violated and the errors are most likely underestimated.
There are 17 alkenone records and 12 foraminiferal records in the data set. From a few of the cores present in the data set have both alkenone and foraminiferal data have been derived.
The data set is presented on three work sheets in an Excel workbook. The fist page contains average values for 6k, as well as modern atlas data for the selected sites, whereas the second and third sheets contain the actual Holocene time series of alkenone and planktic foraminiferal data, respectively.
In order to download the data you will need a password. If you have not received a PMIP/MOTIF email with this information, you can obtain the password by contacting Carin Andersson
If you wish to use these data as a diagnostic, please cite Kim et al., (2004a,b) , and this website (http://www.bridge.bris.ac.uk/resources/databases.htm) as the source.
Technical queries about the data sets should be addressed to Sandy Harrison or Mark Creighton.