NEURAL-NET AIDED DETECTION OF ASTRONOMICAL PERIODICITIES IN GEOLOGIC RECORDS

Astronomically controlled variations in the Earth's climate induce cyc lic trends in the sedimentary process and record (Milankovitch periodi city). One of the main difficulties to be solved in order to choose am ong the registered periodicities is the conversion from the spatial (i .e. recurrent variations along the stratal sequences) to the temporal domains of the astronomically induced frequencies present in the rock record. We discuss here how this problem can be circumvented by teachi ng a neural net how to recognize periodicities in the signal. The appl ication to two sequences of shallow water carbonate deposits from the Cretaceous of Southern Italy has shown this approach to be particularl y effective, confirming the existence of Milankovitch-type periodiciti es in the records examined, where climate, sediments and biota concomi tantly react to the variation in the solar constant induced by secular perturbations of the Earth's orbital elements.