Image-D5.6: Mt-Inversion Techniques With External Constraints

• Main Authors: Hersir, Gylfi Páll, Manzella, Adele
• Format: Text
• Language: English
• Published: Zenodo 2017
• Subjects: geothermal energy; exploration; MT-Inversion; data modelling; iceland; italy; geophysics; Greenfield; Hengill; Iceland
• Online Access: https://dx.doi.org/10.5281/zenodo.1288708; https://zenodo.org/record/1288708

This report describes MT-inversion techniques with external constraints, Deliverable D5.6 within the IMAGE project. Report analysing the results of MT-inversion techniques with external contraints, describing optimal MT layout and data modelling for obtaining maximum information from MT surveys, including recommendations for opitmised MT site spacing for different a-priori constraints. The first part of the report, which was performed by ISOR, discusses the origin and nature of the static shifts and some tested methods for static shift correction, i.e. joint 1D inversion of co-located TEM and MT and spatial filtering and statistical assumptions about the shifts. A software is introduced which inverts the two datasets for both the resistivity model and the shift of the MT. Besides determining the static shift, joint inversion is an important quality check of the TEM and MT data sets, i.e. weather they are compatible. In EM surveying, software should be used make a preliminary joint inversion of TEM and MT at base camp as a quality control to make sure the field mission is not terminated until god quality data have been collected. Finally, the claim that 3D inversion of MT can deal with the static shifts, i.e. introduce shallow resistivity structures (not resolved by the data) to account for the shifts is tested. The second part was done by ISOR. On one hand the depth-location of a low-resistivity anomaly, as observed from borehole data, is build into the starting model, giving the program a headstart into gaining information on the resistivity in the survey area and on the other hand information on the ductile-brittle bounday location is used to infer the location of a deep low-resistivity anomaly, which is put into the starting model of the inversion. Several tests have been done on the site spacing in MT for 3D inversion.n This was done for MT data from the Hengill area in SW-Iceland, starting with a measurment and model grid of 1 km by 1 km (see, Árnason et al., 2010). Later a measurement and model grid seize of 500 m by 500 m was tested which improved the resolution significantly for the uppermost 1 to 2 km (Árnason, pers. comm., June 2017). Finally, a measure-ment and model grid seize 250 m by 250 m was tested (Benediktsdóttir, pers. comm., June 2017). It only improved the resolution moderately. In the third part performed by CNR, an integrated approach is proposed that greatly improved the knowledge on the deep structures of the system on the basis of the critical review of deep well data, geological and geophysical data and the analysis of new and previously acquired MT data in the Larderello-Travale field. Resistivity models by 2D deterministic inversion was achieved, focusing on the understanding of the reliability of the a-priori model for the inversion procedure. Three sets of starting models were implemented and tested: A homogeneous (without external constraints), a geological model (from the integrated model) and finally an interpolation of 1D models. The resulting models constrained with detailed and accurate geological information as well as using 1D models showed higher resolution than those unconstrained (i.e. from homogeneous half-space). It is demonstrated how the a-priori information from the analysis of MT data, i.e. by PSO optimization in this case, greatly improved the inversion results even those geologically constrained, which is not a trivial issue for the exploration of geothermal greenfield, due to lack of underground data. : FP7