EarthRef.org Reference Database (ERR) -- Alexandrescu et al. 1997

Declination (D) and inclination (I) data from the Paris and London areas have been critically compiled by Alexandrescu et al. <1996a> and Malin and Bullard <1981>, respectively. Declination data begin in the mid-sixteenth century, and inclination data begin in the late sixteenth century. The data are remarkably well correlated between the two sites and can be used to construct two more complete and accurate synthetic series, in which many gaps or times with scattered or uncertain data can be amended. These synthetic series of D and I, believed to be valid for much of western Europe, are next filtered for external variations by applying an 11-year smoothing. Secular variation (the first time derivative) is then computed. Secular variation of annual means of D shows the well-known jerks of the twentieth century (1901, 1925, 1969, and 1978). Detailed analysis of monthly mean values confirms these events and further suggests jerks in 1913 and 1991, though it is still too early to be sure of the latter. Interestingly, the data confirm the existence of a large jerk in 1870. Analysis of the annual means from three other observatories supports the significance and extent of this event. From 1680 to 1870 there is no evidence of major jerks. Though more noisy and much shorter, the inclination data seem to confirm this general pattern and correlate rather well with those of D. In conclusion, pending discovery of other archives, we suggest that the two centuries from 1680 to 1870 may have been a period of smooth, parabolic secular variation devoid of jerks (in Western Europe). This quiet period was interrupted in 1870, ushering in a century of repeated jerks with a rather characteristic sawtooth pattern.¿ 1997 American Geophysical Union Declination (D) and inclination (I) data from the Paris and London areas have been critically compiled by Alexandrescu et al. <1996a> and Malin and Bullard <1981>, respectively. Declination data begin in the mid-sixteenth century, and inclination data begin in the late sixteenth century. The data are remarkably well correlated between the two sites and can be used to construct two more complete and accurate synthetic series, in which many gaps or times with scattered or uncertain data can be amended. These synthetic series of D and I, believed to be valid for much of western Europe, are next filtered for external variations by applying an 11-year smoothing. Secular variation (the first time derivative) is then computed. Secular variation of annual means of D shows the well-known jerks of the twentieth century (1901, 1925, 1969, and 1978). Detailed analysis of monthly mean values confirms these events and further suggests jerks in 1913 and 1991, though it is still too early to be sure of the latter. Interestingly, the data confirm the existence of a large jerk in 1870. Analysis of the annual means from three other observatories