The first archaeointensity records from New Zealand: evidence for a fifteenth century AD archaeomagnetic "spike" in the SW Pacific region?

All specimen-level palaeointensity data, quality parameters, checks and site averages. Columns give specimen ID (specimens marked with a * are those for which a high blocking temperature component was subtracted before the palaeointensity was estimated); palaeointensity estimate; uncertainty <i>σ</i>_B); T_min and T_max, the blocking temperature range over which the palaeointensity was calculated; <i>N</i>, the number of temperature steps in this range; <i>β</i> is the fractional uncertainty – the ratio of <i>σ</i>_B to the palaeointensity; <i>f</i> is the fraction of the NRM, (or low blocking temperature component) used in the calculation; <i>g</i> is the gap factor and <i>q</i> a quality factor (= <i>fg/β</i>) (both from Coe <i>et al</i>., 1978); <i>w</i> (=<i>q</i>/√<i>N</i>−2) is the weighting factor of Prevot <i>et al</i>. (1985); <i>δ</i>_CK and <i>δ</i>_TR are measures of the maximum differences found in pTRM checks and tail checks respectively; MAD<i>_free</i> is the maximum angular deviation of the principal component analysis fit to the remanence data over the selected temperature interval calculated without anchoring the fit to the origin of the vector component plot; TT* class is the overall quality classification of Leonhardt <i>et al</i>., 2004, modified by Paterson <i>et al</i>. (2014). The criteria for TT*A and TT*B classes are given at the end of the appendix. For each site a simple average and standard deviation is given, an inverse variance weighted average, and an average and standard deviation of data weighted according to the parameter <i>w</i> of Prevot <i>et al</i>., (1985). There is insignificant difference between the results of these different methods of averaging, and the <i>w</i>-weighted average was adopted.