Abstract:

The aim of this project was to collect samples of stalagmites from Northern and Central America in order to produce records of the palaeosecular variation of the earth's magnetic field. Two stalagmites were sampled from Western Canada and ten from Mexico and Guatemala which could be compared with contemporaneous stalagmite records from these areas (Latham, 1981; Latham et al, 1982; 1986; 1987; 1989).
The stalagmites were generally weakly magnetised but remanence directions were stable upon stepwise thermal and alternating-field demagnetisation. Consistency in directions recorded between central and corresponding lateral sub-samples within two stalagmites (MSC2 from Canada and CP1 from Guatemala) inferred that any depositional errors caused by surface effects were less than the measurement errors. Grain size analysis showed the presence of a fine-grained magnetic fraction (0.01 - 0.1?:m) sourced from the cave drip-waters (either by direct deposition or by chemical precipitation) and a coarser magnetic fraction (0.01 - >10?:m) sourced from the flood-borne detritus. The latter source was dominant in stalagmites which were regularly inundated with water. The type of magnetic mineral present was determined by the geology of the catchment area; magnetite dominated in the Vancouver Island stalagmites, titanomagnetite in the Mexican stalagmites and haematite in the Guatemalan stalagmite.
Uranium-series dating of samples was hindered by the young ages of many of the samples, by low uranium concentrations and by the presence of allogenic thorium. If significant amounts of allogenic thorium were present, a sample age could be calculated based on an estimate of the initial thorium ratio ([230Th/232Th]0). Analysis of samples from Sumidero Recuerdo in Mexico, however, suggested that this ratio is not constant with time and may vary by a factor of two over approximately 1700 years. Due to these imprecisions many dates were out of stratigraphic sequence and age estimates were made assuming constant growth rates, except where growth had ceased for a finite length of time.
Records of sequential change of palaeomagnetic direction were obtained from the Mexican stalagmite SSJ3 and the Canadian stalagmite MSC2. The reliability of the latter record was confirmed by comparison with another Canadian stalagmite record (Latham et al, 1987) and contemporaneous lacustrine records. Other records were disappointing due to poor temporal resolution; each sub-sample represented a period of approximately 1000 years in Mexican stalagmites SSJ2 and SSJ4. Such slow growth rates are insufficient for the resolution of secular variation features with periods of less than 2000 years and are only suitable to gain information about the nature of long-term secular variations, for example the far-sided virtual geomagnetic poles and low inclinations predominant throughout the Holocene in Southern Mexico.
The existence of matching contemporaneous stalagmite records of secular variation together with the demonstrated lack of depositional inclination errors is encouraging, despite the sometimes "hit or miss" aspects of sample selection. Nevertheless it has been proved that speleothem records have the potential to complement the existing archaeomagnetic, lava and lacustrine data.