Dear Theo,
even though there are some data gaps in the CM SAF SIS data set, the gaps in your data cannot be explained by the gaps in the original data set. For me it seems like you are missing the data from Meteosat 5, which has been operational from 1994 until 1997. This data is available from the CM SAF, maybe there was a problem with the data transfer etc. As the data comes in packages per satellite, maybe this one has not been prepared or so. I suggest you repeat your request focussing on the gap in 1994 to 1997.
Another issue I noticed when looking at your time series plot is that there are some days that have a SIS value of zero. This is obviously not realistic (even in the UK!) and has recently been updated in the CM SAF data base. These files have been replaced by files that contain missing_data instead of zero values, which should allow the users to automatically remove these data from the analysis. Alternatively you might want to set all zero-values to NA.
In addition we found a couple of days with suspicious data (weird stripes across the data) that also have been set to NA recently in the data base. These days include 15.08.1992, 14.12.1992, 13.06.2000. Please make sure to remove these data from you analysis (or update your data).
Sorry for an inconvenience the suspicious data has caused, obviously we try to minimaze these problems, but things like these cannot always been ruled out.
Hope this help, kind regards,
Jörg
P.S. You should be able to merge the two times series, once you have extracted the specified location.