date: Fri Aug 1 09:56:46 2003 from: Tim Osborn subject: Mann series to: simon.tett@metoffice.com Hi Simon, an extra bit of information to go with the Mann record is that the "lowf" standard error column actually seems more applicable to the record when it is smoothed with a 50-year filter rather than a 40-year filter. But I've got hold of the calibration residuals from Mike now, and can check the answer by using an alternative method - and if it works, I can quantify errors for any smoothing filter we want. Cheers Tim