Coregionalization models of geophysical data for site effect assessment at Anse-à-Veau, Haiti

The multivariate generalization of the kriging, in geostatistical analysis, called cokriging makes it possible to take advantage of the relationships between several variables. The information carried by a secondary variable may improve the precision of the estimation of the main variable. The multivariate structural model between the variables is then built from the spatial joint analysis of the data which can be realized by coregionalization linear models.  

The present study aims at assessing the site effects from the results of multivariate analysis of seismic data collected at Anse-à-Veau, a municipality in the Nippes Department of Haiti, characterized by a relatively high seismic activity – during the measurement campaign in August 2021 this region had been hit by the 2021 Nippes Earthquake. The surveys carried out include ambient noise and seismological recordings, seismic tests as well as electrical resistivity measurements along profiles. The two first were processed, respectively, in terms of Horizontal to Vertical Spectral Ratios (HVSR) and Standard Spectral Ratios (SSR), the seismic tests both as Seismic Refraction Tomography (SRT) and by Multichannel Analysis of Surface Waves (MASW) and the last measurements as Electrical Resistivity Tomography (ERT). In total, more than 100 HVSR recordings, 7 115m-long seismic profiles and 8 ERT profiles have been completed. All related results were then compiled within one multi-data (including also geological, geomorphic and geomechanical information) 3D geomodel and submitted to a spatial analysis. The coregionalization modelling applied within this analysis is expected to take advantage of the relationships between the different type of seismic data in order to better estimate the potential site effects in the study area.

Keywords: coregionalization model, site effects, multi-geophysics, geomodelling, Haiti.