Forward and Inverse modelling of gravity anomaly for geothermal applications: Case study of Los Humeros (Mexico).

Abstract

Forward modelling of gravity anomaly may be used to validate structural models of the subsurface, particularly when the deeper subsurface is poorly constrained. When the modelled gravity anomaly correlates well to the measured anomalies, the structural model is validated. Otherwise, an inverse modelling with data assimilation will allow to reduce the uncertainties of the density anomaly present in the structural model. Basin3D software developed by TNO, is a gravity anomaly modelling software for both forward and inverse modelling. The methodology of the software is based on previous work in the field by Caratori Tontini et al. (2009) and Cooley and Tukey (1965), and is implemented in Java. This study presents the results of benchmark tests performed on the forward modelling module of Basin3D, versus Geomodeller software (Geophysics, 2015), which uses highly comparable methodology. Identical structural model of Los Humeros geothermal complex (Evanno, 2017), and equivalent modelling parameters, have been used in both the forward modelling module of Basin3D and forward modelling module of Geomodeller. Results presented in this study have been assessed using Mean Absolute Error (MAE) for error calculation, and Root Mean Square Error (RMSE) to MAE ratio for evaluation of the error distribution, also Pearson’s Correlation Coefficient (PCC) has been used for equivalence assessment. We found that the gravity anomaly modelled in Basin3D correlates very well with the gravity anomaly modelled in Geomodeller, reaching 0.999977 of correlation. Errors are also minimal, lower than 0.1 mGal, and the error distribution is satisfactory at a ratio equivalent of 1.252. Thus, we find that Basin3D performs well in the benchmark test, and it is a reliable tool for forward modelling of gravity anomaly. Next, we use the same structural model of Los Humeros geothermal area and the results of the forward gravity modelling, together with the digitised map of gravity anomaly for the area, to evaluate the validity of the structural model by Evanno (2017). We find that the correlation between the modelled and observed gravity is very low, at only 0.24370, and the error is high at around 10 mGal. Thus, the modelled distribution of the anomalies and their amplitudes do not match the observed gravity anomalies. Therefore, to achieve a better match, we inversely model the gravity anomaly and corresponding densities with use of data assimilation of the measured values (digitised observed gravity map). The results presented in this study allow for identification of three main areas of density anomaly within the Los Humeros area, with the most significant low density anomaly located just below the Los Humeros caldera. Low density anomaly below the caldera may indicate the presence of hotter material, possibly a magma chamber. There is no current consensus on the location, structure and extend of the magma chamber. Therefore, any constrains on its depth and location below the Los Humeros caldera are of great importance for the geothermal prospecting of the Los Humeros geothermal complex. Thus, based on the location of the density anomalies and the direction of change, we suggest two new structural models. First, we suggest a model with limestone layer of irregular thickness, thicker in the south-west of the model and thinner in the north-east, to compensate for the low and high gravity anomalies respectively. The second model is proposed with an additional low density layer, possibly hot material, reaching the shallowest depths just below the Los Humeros caldera. We suggest forward modelling of gravity anomaly will be performed in the future studies for the two proposed models, and then compared to the measured values of gravity anomaly. Thus, either validating or rejecting the hypothesis that the observed low gravity anomaly below the caldera is caused by a less dense material, possibly a magma chamber. Finally, we conclude that, both the forward and inverse modelling of the gravity anomaly for the Los Humeros case study produced results which will aid future research in the field.