Geochemical correlation, variation, and petrogenesis, of ignimbrites in Central America and associated Caribbean Sea tephra layers

The Central American Tertiary Ignimbrite Province covers an area of 170,000 km2. A total of 112 marine tephra and 79 terrestrial samples from Nicaragua and Honduras were analyzed during this study, using an electron microprobe for major oxides and laser ICP-MS for trace elements. Cluster analysis resulted in the division of the samples into 6 geochemical groups. Visual inspection of these groups resulted in the reclassification of these 6 into 14 distinct geochemical groups. The geochemical correlation is strengthened using factor analysis in which the REE values for each sample were reduced to 2 factors. The Lesser Antilles can be ruled out as contributing to the Western Caribbean tephra due to the lack of any large-volume ignimbrite sheets within the arc, the atmospheric circulation patterns that transport tephra to the east, as well as the distinctly different REE trends of the magmas. The Sierra Madre Igneous Province in Mexico is also ruled out as a potential source, due mainly to a significant age difference between Sierra Madre ignimbrites and the Caribbean Sea tephra layers. There is evidence for sediment-derived fluid contamination, which includes a positive Eu-anomaly in two of the ignimbrite geochemical groups as well as the plotting of the ignimbrite magmas within mixing fields between NMORB and oceanic sediments. A plot of Ba/La along the inferred paleoarc shows a strong enrichment in Nicaragua. Evidence of the influence of the continental crust in the formation of the Central American ignimbrites include: (1) The very large volume of the rhyolite ignimbrites in comparison to occurrences of mafic magmas. (2) The majority of the ignimbrites fall within the classification fields of S-type granites. (3) All of the ignimbrites analyzed for delta18O and 87Sr/ 86 Sr isotopes have values within the range of continental crustal rocks. (4) REE and other trace element trends are similar to those of average lower continental crust. (5) Fractional crystallization models fail to explain the major and trace element compositions of the ignimbrites. (6) AFC and continental crustal melt mixing models produce evolutionary trends that are consistent with the compositions of the ignimbrites.