Multiridge Method for Studying Ground-Deformation Sources: Application to Volcanic Environments

Abstract

Volcanic phenomena are currently monitored by the detection of physical and chemical observations. Generally, the ground deformation field is the most relevant shallow expression of the geometric and physical parameters variations in the magmatic reservoir. In this study, we propose a novel method for the direct estimation of the geometric parameters of sources responsible for volcanic ground deformation detected via the DInSAR technique. Starting with the biharmonic properties of the deformation field, we define an approach based on the Multiridge and ScalFun methods to achieve relevant information about both the positions and shapes of active sources, such as the Mogi source. Our methodology is definitely different from the methods currently used for modeling ground-deformation sources, mainly based on forward or inverse techniques. In fact, (i) it does not require any assumptions about the source type, and (ii) it is not influenced by the distribution of medium elastic parameters or (iii) the presence of high-frequency noise in the dataset. For synthetic cases, we accurately estimate the depth to the source within a 3% error. Finally, we study the real case of the Okmok volcano ground-deformation field and achieve results compatible with those in previous works.

Multiscale Analysis of DInSAR Measurements for Multi-Source Investigation at Uturuncu Volcano (Bolivia)

“…Indeed, the proposed methodology has already been used to investigate the inflation source responsible for the 2003-2004 Okmok volcano (Alaska-USA) LOS ground deformation pattern, verifying its validity: its results do not depend on the physical property parameters, such as elastic modula or source pressure changes. Moreover, the achieved results are stable with respect to the high-frequency noise in the dataset [37].…”

“…We specify that each ridge is determined by a best-fit linear regression within a 95% confidence interval; in particular, we calculate the determination coefficient R 2 , which represents a statistical measure of how the selected data (edges) are close to the fitted regression line (ridges). Moreover, we evaluate the solution uncertainties (intersection at the ridges) by considering the error on the best-fit linear regression coefficients (intercept and slope constants) [37].…”

“…Regarding the deformation case, Castaldo et al [37] have exploited the physical properties of deformation fields, showing how to apply the Multiridge method to detect Mogi's kind of source [49] by DInSAR measurements. Indeed, the proposed methodology has already been used to investigate the inflation source responsible for the 2003-2004 Okmok volcano (Alaska-USA) LOS ground deformation pattern, verifying its validity: its results do not depend on the physical property parameters, such as elastic modula or source pressure changes.…”

“…zone of APVC during the 2005-2008 time interval (Figure 4a) and the vertical deformation component at Uturuncu volcano from August 2006 to February 2007 (Figure 5a). To apply the Multiridge method, we perform for each dataset the constant reduction level procedure [37], whose level, in this case, is equal to 6 km a.s.l.. Appendix A describes this procedure and the vertical gradient transformation (Figures A1 and A2).…”

“…We apply two multiscale methods to investigate possible active multi-source localized at different depths. In particular, focusing on the vertical component relevant to the 2005-2008 time interval, we use the Cross-correlation analysis [34,35] and the Multiridge method [36,37] to emphasize the spatial and temporal multiscale properties of the field during the considered period.…”