Multivariate analyses of crater parameters and the classification of craters |
| |
| Authors: | Barry S Siegal John C Griffiths |
| |
| Institution: | 1. Dept. of Geosciences, Pennsylvania State University, University Park, Pa., USA
|
| |
| Abstract: | Multivariate analyses were performed on certain linear dimensions of six genetic types of craters. A total of 320 craters, consisting of laboratory fluidization craters, craters formed by chemical and nuclear explosives, terrestrial maars and other volcanic craters, and terrestrial meteorite impact craters, authenticated and probable, were analyzed in the first data set in terms of their mean rim crest diameter (D r), mean interior relief (R i), rim height (R e), and mean exterior rim width (W e ). The second data set contained an additional 91 terrestrial craters of which 19 were of experimental percussive impact and 28 of volcanic collapse origin, and which was analyzed in terms ofD r,R i, andR e. Principal component analyses were performed on the six genetic types of craters; 90% of the variation in the variables can be accounted for by two components. 99% of the variation in the craters formed by chemical and nuclear explosives is explained by the first component alone. Classification using the CLUS procedure (Rubin and Friedman, 1967) indicates an optimum number of two groups; the main difference between the groups was the presence or absence of rims. The rimmed group of craters could, if desired, be subdivided on the basis of the dimension of the rim crest diameter. Several small rimless craters were classified as rimmed craters, this is believed to have resulted from the small difference in the dimension of their rim crest diameter to rim height, as contrasted with the larger rimless craters; thus, this distinction may also be size dependent. No evidence was found to support an exogenic-endogenic classification scheme of craters, with respect to the variables measured. |
| |
| Keywords: | |
| 本文献已被 SpringerLink 等数据库收录！ |
|
