METHODS FOR DETERMINTION OF PALEOSTRESS IN ROCKS

Abstract

Fault slip data inversion for the getting the paleo-stress (original stress tensor) that may have generated that fault slip lineations, is a linear to non-linear technique depending on the assumptions one consider. There are different methods that are available for stress inversion ranges from linear to non-linear mode. As these methods compiles different techniques as well as assumptions, then there must be some limitation to their use depending on different tectonic setting. That’s why these methods must be compared as to have a clear understanding of their correct use while discarding the others. So, to compare the methods, we have taken 4 methods under consideration (1) Direct Inversion by Angelier(1990), (2) Linear Inversion by Hansen(2013), (3) Iterative Method by Etchecopar(1981) and (4) Genetic Algorithm by Thakur(2016). Firstly synthetic data for fault slip was generated for homogeneous stress while taking consideration of Wallace and Bott hypothesis. And thus the fault slip data being generated as per the Byerlee’s friction law which validates the slipping on the fault. After all the data been inverted through all the above described methods, the error calculation is done. For comparison purpose, we have done the Root Mean Squared Error (RMSE) for all the methods as it gives a closer look of what method is better than the other. So after testing all the data and the error calculation, it is shown that the Linear Inversion by Hansen (2013) produces the most accurate and consistent result in comparison with other methods. Though there is one thing to do before coming to final result is to solve the unflipped cases in the Linear Inversion as it gives two result at the same time and the probability of that happening of completely random. So if the flipping case of Hansen’s method is taken under consideration, his method for stress inversion is by far the best while Direct Inversion by Angelier (1990) is the worst of all. And all the other two methods are good and fairly consistent.