Chaotic dynamics and fractals in magmatic interaction processes: a different approach to the interpretation of mafic microgranular enclaves

Abstract

Fractals have been used for describing a wide range of natural processes. However, applications of fractal geometry to the study of magma mingling/mixing processes is almost completely lacking, in spite of the fact that such an approach could give new insights into the nature of magmatic interactions. Fractal analyses of two geochemically linked groups of mafic microgranular enclaves (MME) from the Sithonia Plutonic Complex (Northern Greece) have been performed. Because of the difficulties involved in using contours of enclaves to estimate their fractal dimensions, a different approach is used in this paper. Fractal dimension is estimated using distribution maps of chemical elements inside enclaves. The results obtained indicate that a clear correlation between fractal dimensions of enclaves and geochemical evidences of interaction exists and hence the two groups of enclaves reflect different degrees of interaction. From a dynamical point of view, stretching and folding processes between two magmas with different rheologies are considered representative of magmatic interaction. This dynamic system can easily generate chaotic patterns recognized by the presence of fractal structures. The contemporaneous occurrence of well-mixed regions together with poorly-mixed regions in the same system is thought to be similar to the host/enclave system found in many granitoid plutons. In particular, MME are believed to represent islands of partially-mixed mafic fluid subjected to different degrees of interaction with the host granitoid, whereas host rocks are believed to represent mixed regions where the mixing process acted more efficiently producing almost completely hybrids. It is shown that fractal geometry and chaotic dynamics can be considered suitable techniques in studying complex petrological phenomena and they can represent helpful methods in the development of petrogenetic models.