The IPSI 1.4 database contains information collected up until December 2019 concerning horizontal stress orientations inferred from different stress indicators2. Data are relative to borehole breakouts from deep wells, crustal earthquake focal mechanisms and fault data. All data refer to a depth interval down to 40 km, that corresponds to an average maximum crustal thickness in Italy. The depth ranges of the different stress indicators are almost complementary, as faults yield stress information about the surface, breakouts in the first kilometers and earthquakes mostly in the deepest ranges (Fig. 2), providing a complete picture of crustal present-day stress field.
In particular, IPSI 1.4 includes more than 10% data with respect to the first version of the database1,26, relatively to focal mechanisms, formal inversions and faults mainly located in central Italy. In addition to the bibliographic references associated with the single data, it contains all the information on the available wells (stratigraphy and geophysical logs), and since 2018, we have no longer considered Centroid Moment Tensor earthquake solutions but those obtained from high-quality data30.
IPSI was mainly conceived for users interested in studying the stress field of the Italian crust in a more accurate and complete way. Most of the data are also part of the WSM (http://www.world-stress-map.org/), as evidenced by many links that redirect to this database, and all these data are regularly delivered to the WSM project. The IPSI website (http://ipsi.rm.ingv.it/) provides access to data in a standard map viewer where data can be selected for plotting (category and/or quality) and downloaded in common file formats (Fig. 1). The legend on the left shows the basic information of the different data, the tectonic regime assignment and quality ranking method with pop-up windows and linked files. The main information of each element (type, quality, orientation) can be viewed by hovering over the related symbol on the map and more details appear by clicking the selected element (Fig. 3).
In this section, we describe the methods used to analyse the different indicators for defining stress orientations. According to the WSM ranking scheme31, a quality category (A to E, from best to worst) is assigned to each stress data orientation record. Quality A–C indicates that the stress orientation is within ± 15°, ± 20° and ± 25°, respectively, D-quality means that stress orientation is questionable (within ± 40°) and E-quality denotes unreliable or insufficient information.
The stress regime expresses the relative magnitude of the three principal stress axes (S1, S2 and S3). Stress magnitudes are defined using standard geologic/geophysical notation with positive compressive stress: S1 as the maximum, S2 as the intermediate and S3 as the minimum principal stress axis. Assuming that one of the principal stresses is vertical (Sv), SHmax and Shmin are the maximum and minimum principal components of the stress tensor on the horizontal plane, respectively.
As related to fault kinematics, the main categories of a tectonic regime are thrust, normal and strike-slip faults32. Only when faults are optimally oriented with respect to the stress field is the stress regime coincident with the tectonic regime33,34,35. In normal faults, S1 is vertical, Shmin corresponds to S3 and SHmax to S2. In thrust faults, S3 is vertical, Shmin corresponds to S2 and SHmax to S1. In strike-slip faults, S2 is vertical, Shmin corresponds to S3 and SHmax to S1 (Fig. 4).
On the IPSI website, we report results in terms of the minimum horizontal stress orientation on the map (corresponding to either S2 in thrust regime or S3 in a normal or strike-slip regime), and both Shmin and SHmax in the data records. We classify the data, except for borehole breakout data, into five tectonic regime categories: normal faults (NF); thrust faults (TF); strike-slip faults (SS); normal faults with a strike-slip component (transtension, NS) and thrust faults with strike-slip component (transpression, TS) according to the WSM categorisation6.
The database also includes a single overcoring datum that has been taken from WSM36. Usually overcoring data are referred to a very shallow depth and then comparable to the fault data. For a more complete description of overcoring technique please refer also to ref. 37.
Borehole breakout data
Borehole breakouts are stress-induced “enlargements” of a wellbore cross-section that occur discontinuously when a well is drilled in rocks within an anisotropic stress field38,39. The “enlargements” develop on opposite sides of the borehole wall along the Shmin direction. Following the main criteria reported in refs. 40,41,42, we determine breakouts using records from a four-arm caliper tool in deep wells (approximately 0.450–7 km depth). We take into account wells with deviation not more than 15° and not less than 0.5° from the vertical, usually the deviation is much less than 10°. We apply the circular statistics of ref. 43 to compute the mean breakout orientation and standard deviation (95% confidence) for each borehole weighed by the breakout zone length. Up to now we have not been able to use image logs that would reduce the uncertainty in breakout orientations, as shown by recent studies (e.g. ref. 44).The WSM quality ranking system31 (A to E) used to classify the breakout orientation of each well accounts for the number of breakout zones, total length of the breakouts and standard deviation of the orientation (Table 1). We assign E-quality to well data without reliable breakouts (standard deviation > 40°) or evidencing no breakouts along the borehole, therefore stress orientations and related information are not provided in the database. We always analyse in detail the possible reasons of such results, especially in order to verify if some stress orientations with large standard deviation could be related to stress rotations nearby major fault zones.
Borehole breakout orientations alone do not allow evaluation of tectonic regime. They can be combined with other borehole data, for instance along with rock strength or together with leak off tests to estimate or constrain stress magnitude and infer faulting regime (e.g. refs. 45,46,47).
Once selecting an element on the IPSI website, a pop-up window provides the related identification code and summary information from the data record. The upper part includes the type, quality, stress orientation, tectonic regime, geographic coordinates and date of the first online publication and its reference location. The lower part includes additional information such as the well name where the breakout was inferred, drilling year, vertical well depth, top and bottom of the breakout data and code of the well in the National Mining Office database for hydrocarbon and geothermal energy (UNMIG) of the Italian Ministry of Economic Development. The latter includes a link to the Videpi archive (https://www.videpi.com) where well stratigraphic log can be viewed and downloaded (Fig. 3a). Usually, associated to the stratigraphy there are some geophysical logs (e.g. resistivity and sonic) that are useful for a better characterization and interpretation of the breakout data.
Earthquake focal mechanism data
This category contains crustal earthquakes with M ≥ 4, usually moment magnitude (Mw), and maximum depth of 40 km. Concerning the magnitude cut off, it is related to most of the used dataset that include focal mechanisms of M ≥ 4 earthquakes. Since even smaller events can help constraining the stress tensor48, in the next IPSI releases, using higher quality data, we could insert earthquakes with lower magnitude, particularly useful in areas with few data. Relatively to the depth, most of crustal earthquakes occur within the upper 20 km and we do not consider seismicity related to the still active subduction zones.
For the oldest 23 seismic events that occurred between 1908 and 1975, we consider results from the polarity solutions of earthquakes computed by a range of previous studies. For seismic events from 1976 to 2017, we account for Centroid Moment Tensor (CMT) solutions of earthquakes selected from the European-Mediterranean RCMT catalogue49 and the Italian CMT dataset50,51. For earthquakes that occurred from 2018 to present, we use the Time Domain Moment Tensor (TDMT) catalogue52. The focal mechanism solutions30 are obtained from the high-quality data of the Italian broadband and Mediterranean seismographic networks using the long-period full waveform inversion code originally proposed by ref. 53. Taking into account the systematic error of CMT-like solutions ( ± 14° according to ref. 54), range of stress orientations that would be consistent with each focal mechanism, and that the orientation of P (compression), B (null) and T (extension) axes may slightly deviate from the principal stress orientations55, the WSM31,56 assigns C-quality to these data (see also ref. 25 for a detailed explanation). Following the WSM criteria, we assign C-quality to all focal mechanism data.
To identify the Shmin azimuth and tectonic regime, we use the plunge of the P, T and B axes by applying the criteria of ref. 6, modified for Shmin (Table 2). We discard all of the focal solutions with P, T and B axes that do not define a clear tectonic regime. Although focal plane solution principal axes may not be indicative of stress axes, the possible differences between Shmin derived from P, T and B axes and Shmin from slip vectors lie within the error of the attributed quality category, as shown in ref. 16. Moreover, regional compilations show that the average orientations of P, B, and T axes determined from a number of earthquakes yield a good indication of the stress orientation throughout a region (e.g. ref. 57). Thus, the orientation of the kinematic axes (P, B, and T) is assumed to coincide with those of the dynamic axes (S1, S2 and S3).
When selecting an element on the IPSI website, a pop-up window provides the related identification code and summary information from the data record. The upper part provides the type, quality, stress orientation, tectonic regime, geographic coordinates, date of the first online publication and a reference to the focal mechanism data source. The lower part provides the earthquake date, magnitude, depth, strike, dip and rake of the two nodal planes (Fig. 3b).
Formal inversion of earthquake focal mechanism data
This category includes stress orientations determined from the inversion of P, B, and T axes of diffuse seismicity. The data are located in close geographic proximity where a homogeneous stress field can be hypothesised and where formal inversions of more than 8 well-constrained single events are present with a standard deviation or misfit angle less than 20°. We account for literature data that satisfy the above criteria and assign them as mostly B-quality.
We use the same method as that for earthquake focal mechanisms to define the tectonic regime and corresponding Shmin for each inversion (Table 2). When selecting an element on the IPSI website, a pop-up window provides the related identification code and summary information from the data record. The upper part provides the type, quality, stress orientation, tectonic regime, geographic coordinates, date of the first online publication and the reference of the inversion of focal mechanisms. The lower part provides the code name, region, number of earthquakes used in the inversion procedure, year range of the events, magnitude range, depth range, misfit, azimuth and dip of S1, S2 and S3 (Fig. 3c).
Fault slip data
This stress indicator category includes data from single faults with known attitude and primary sense of slip. As described in ref. 16, we do not include fault data related to earthquakes whose focal mechanisms are available, the stress information can be found in the focal mechanism category. The only exception concerns the latest surface faulting related to the 2016 central Italy seismic sequence. Starting from that date, with the aim to provide more complete information, we decided to include surface faulting data even if focal mechanism solutions are available. We are aware that this choice requires more attention when using the entire dataset to avoid duplicates, but also allows the use of complete fault category data only if needed.
Each fault, strike, dip, slip and kinematics provided in the original studies are used to define the Shmin orientation and tectonic regime in the same way as the focal mechanism data. When the slip is unknown, the Shmin orientation is assumed to be perpendicular to the fault strike for normal faults. We assign C-quality to all fault data, as suggested by the WSM guidelines.
When selecting an element on the IPSI website, a pop-up window provides the related identification code and summary information from the data record. The upper part provides the type, quality, stress orientation, tectonic regime, geographic coordinates, date of the first online publication and a reference to the field studies of the fault. The lower part provides the region and name (Fig. 3d).