Moho relief modelling by gravity data using the Maximum Difference Reduction method: the case of central Zagros, Iran

We have conducted a gravity study using ground gravity data, via an automated iterative forward modelling method, to determine the Moho depth map in parts of the Zagros Mountains. The method used, the Maximum Difference Reduction (MDR) method, consists in a modification of the well-known Bott method for solving nonlinear inverse problems. The main characteristics of the MDR algorithm are: 1) the subsurface is divided into rectangular blocks to determine the gravity effect; 2) the density contrast is defined as a priori information, and is considered constant in the whole body. Therefore, we have no non-uniqueness; 3) depth correction in any station stops when the sign of a residual gravity anomaly differs from that of an observed one. As a result, the algorithm moves in one direction, and convergence occurs; 4) the inversion process stops when the residual anomalies fall within a predefined error range, which is controlled by the chi-square criteria. The method has been examined through a synthetic model and, then, applied to determine the Moho depth under the Zagros Mountains. The maximum Moho depth in the Zagros fold-thrust belt, Sanandaj-Sirjan zone, Urumieh-Dokhtar magmatic arc and central Iran, are found respectively at 45, 68, 52, and 42 km.