Improved Analysis of EGMS Data for Displacement Monitoring:

The Case Study of Regina Montis Regalis Basilica in Vicoforte, Italy

Davide Lodigiani

, Marica Franzini

and Vittorio Casella

Department of Civil Engineering and Architecture, University of Pavia, Via Ferrata 3, Pavia, Italy

Keywords: Cultural Heritage Conservation, SAR, Persistent Scatterers, Deformation Monitoring, Time Series.

Abstract: The Basilica of Vicoforte has always interested geotechnical engineers due to its location in a geologically

complex area. One part of the Basilica is built on marl, while the other is built on clay. These two types of

soil have different mechanical properties, which have caused the Basilica to experience various foundation

failures over time. Monitoring is necessary to evaluate structural evolution and prevent further damage. Radar

images are one of the geomatics techniques that can be used to perform these types of analyses; however,

SAR data processing is challenging and requires specialized skills and software to monitor deformations using

the PSInSAR approach. The European Ground Motion Service (EGMS) is useful for users and researchers,

but analyzing specific buildings or monuments requires a more refined grid. The paper proposes a package of

codes implemented using MATLAB release 2023b to manage grid spacing flexibly, customize it according

to structure dimensions, and manage potential blunders. After a thorough data analysis, it was concluded that

the monument exhibits no signs of subsidence trends. Instead, the analysis revealed that it undergoes seasonal

fluctuations closely associated with temperature changes. The proposed approach enhances data accuracy and

reliability, resulting in valuable insights and informed decisions.

1 INTRODUCTION

Italy is a country that boasts a diverse and extensive

cultural heritage, which has made it a popular tourist

destination for decades. The country's cultural

treasures are spread across its diverse regions and

span several millennia of human history. Artistically,

Italy is home to some of the world's most famous

artworks, from the masterpieces of the Renaissance to

contemporary art. The country's archaeological

heritage is equally impressive, with ancient ruins and

monuments scattered throughout its countryside.

According to UNESCO estimates, Italy is home

to a significant portion of Europe's cultural heritage,

including artistic, archaeological, architectural, and

environmental treasures (Benedikter, 2004). Since its

establishment, UNESCO has organized and

facilitated several conferences and conventions that

focus on conserving and restoring cultural artifacts,

historical sites, and traditions significant to

humanity's shared history and knowledge (Accardo et

https://orcid.org/0009-0009-5082-2022

https://orcid.org/0000-0002-3921-5178

https://orcid.org/0000-0003-2086-7931

al., 2003). The Italian Government has actively

promoted and founded various projects to provide

tools for preserving cultural heritage and mitigating

risks. Some of these projects are focused on raising

awareness about the importance of damage

prevention and reducing or eliminating its causes.

Geological hazards such as soil movements,

landslides, flooding, earthquakes, vulcanism,

sinkholes, cavities, and subsidence are among these

causes and pose significant risks to preserving Italian

cultural heritage. The Geo-Risks Assessment and

Mitigation for the Protection of Cultural Heritage

(GIANO) project is an Italian initiative that focuses

on assessing and mitigating geohazards that risk the

country's cultural heritage. The project has two main

objectives. Firstly, it aims to develop a robust

protocol for accurately quantifying the risks posed by

geohazards to Italy's cultural heritage. Secondly, it

aims to identify the most effective and least invasive

geotechnical mitigation techniques that can be used

Lodigiani, D., Franzini, M. and Casella, V.

Improved Analysis of EGMS Data for Displacement Monitoring: The Case Study of Regina Montis Regalis Basilica in Vicoforte, Italy.

DOI: 10.5220/0012730900003696

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 10th International Conference on Geographical Information Systems Theory, Applications and Management (GISTAM 2024), pages 71-82

ISBN: 978-989-758-694-1; ISSN: 2184-500X

for different monumental constructions to address

these hazards.

The GIANO project examines 14 noteworthy

combinations of historical landmarks and geo-

hazards. The list includes famous archaeological sites

like the Etruscan necropolis of Tarquinia and Pompeii,

churches such as Santa Maria Maggiore in Maratea

and San Paolo Cathedral in Rome, and castles like

Calatabiano and San Felice in Panaro. Among these

sites, and the subject of this paper, is the Basilica of

Regina Montis Regalis in Vicoforte. In particular, this

church is facing a geological hazard due to uneven

foundation settling caused by non-uniform deposits.

Unfortunately, the involved structures often lack

the monitoring systems that are helpful for the

GIANO project, resulting in insufficient data to

assess the structural conditions and any deformations

over time. This also makes it challenging to determine

the history of past deformations that the structure has

undergone. Consequently, conducting reliable

numerical and risk scenario modeling becomes

impossible, reducing the ability to protect and prevent

these monuments from the effects of natural hazards.

Despite having some structural information, the

Basilica of Vicoforte lacks a historical terrain

deformation dataset.

Geomatics has several effective non-invasive

monitoring systems for identifying and tracking

geohazards associated with ground movement or

slope instability affecting cultural heritage structures.

Deformation monitoring can be accomplished using

techniques like high-precision GNSS, spirit leveling,

high-precision topographic surveying, and satellite-

based SAR (Synthetic Aperture Radar)

interferometry. Differential Interferometric Synthetic

Aperture Radar (DInSAR) is a sophisticated type of

radar technology that employs multiple Synthetic

Aperture Radar (SAR) images to detect and measure

even the slightest changes in the Earth's surface over

time(Blanco-Sánchez et al., 2008). The technique

compares SAR images acquired at different times and

uses the data to generate precise ground deformation

maps. This remote sensing technique is especially

valuable for detecting and monitoring surface

deformation in areas with geological hazards, such as

landslides (Cascini et al., 2010), volcanic activities

(Hooper et al., 2004), and earthquakes (Wright et al.,

2004). By allowing scientists to detect and monitor

subtle ground movements, DInSAR plays a crucial

role in understanding and predicting geological

phenomena and ensuring the safety of people and

infrastructure in vulnerable areas (Crosetto et al.,

2016).

Persistent Scatterers InSAR (PSInSAR) falls

under the differential interferometric Synthetic

Aperture Radar (SAR) group. This technique uses

stable reflective points with strong radar signal

reflectivity and maintains temporal stability over time

(Ferretti et al., 2001). Exploiting the permanent

properties of scatterers, atmospheric effects can be

filtered out, and geometrical decorrelation can be

eliminated. Numerous missions, such as the C-band

ERS mission and the more recent X-band SAR

missions, COSMO-SkyMed and TerraSAR-X, have

been developed over the years. These missions have

enabled the systematic acquisition of thousands of

interferometric data stacks worldwide, allowing for

the precise measurement of surface deformations at a

high spatial resolution and with millimetric-level

precision. Best of all, it is a cost-effective alternative

to traditional geodetic observations and does not

require any ground-based instruments to be installed.

The historical archive of more than 30 years allows

the reconstruction of past deformations and the

analysis of their evolution over time.

The Copernicus programme (Jutz & Milagro-

Pérez, 2020; Showstack, 2014) is a crucial aspect of

this framework and deserves special attention. The

project involves two radar sensors, the Sentinel-1A

and 1B (S1A and S1B), designed to collect

interferometric C-band Synthetic Aperture Radar

(SAR) image stacks on a global scale. These sensors

are equipped to monitor ground deformation over

Earth's surface accurately. The temporal resolution of

the mission is impressive, with SAR image stacks

collected every 6 days, allowing for precise

monitoring of any changes that occur over time.

Furthermore, the spatial resolution is equally

impressive, with a resolution of about 10 meters. This

level of accuracy enables us to obtain a detailed

understanding of ground deformation processes

occurring globally (Torres et al., 2020; Torres &

Davidson, 2019).

SAR data is significant, leading to the

development of a dedicated service within the

Copernicus framework: the European Ground Motion

Service (EGMS) (EMGS, 2024). The European

Commission created this valuable tool to monitor and

detect ground deformations using InSAR technology.

The service processes and analyses all Sentinel-1

acquisitions over the Copernicus Participating States,

providing a comprehensive and reliable platform for

SAR data analysis and facilitating informed decision-

making in various domains (Costantini et al., 2022).

The paper will describe the characteristics of the

Basilica, focusing on monitoring the structures due to

geotechnical issues. Specifically, displacement trends

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

could impact further stability, so it is important to

conduct dedicated investigations. The EGMS will be

utilized to achieve this, and its main features are

described in Section 2. However, the online platform

has an insufficient spatial resolution to perform

reliable analysis, as discussed in Sections 2.1 and 2.2.

To address this issue, Matlab codes and routines have

been developed to manage grid spacing flexibly,

customize it according to structure dimensions, and

manage potential errors, as reported in Section 3.

Finally, the results will be presented in Section 4,

demonstrating that the Basilica does not suffer from

any displacement trend but shows significant

seasonal fluctuations closely associated with

temperature changes.

1.1 The Regina Montis Regalis Basilica

The Regina Montis Regalis Basilica is a massive

stone structure in Vicoforte, in northern Italy (Figure

1). This sanctuary is renowned for its extraordinary

oval-shaped dome, which measures around 37x25

meters internally, making it the fourth-largest dome

in the world (Chiorino et al., 2008).

Figure 1: View of the Regina Montis Regalis Basilica in

Vicoforte.

Duke Charles Emmanuel I of Savoy originally

envisioned the Basilica as the dynasty's mausoleum.

The construction of this architecture commenced in

1596, with Ascanio Vittozzi (1583-1615) serving as

the lead architect. Due to an unfortunate site

selection, the building suffered from significant

differential settlements in its foundations since the

early stages of construction (Chiorino et al., 2008).

The settlements were present due to the non-

uniformity of the deposit. The site is characterized by

a layer of compressible clayey silt that increases in

depth from east to west. This layer rests on top of a

marlstone layer that is exposed on the east side of the

Basilica. Upon investigation, five of the Basilica's

eight pillars were found on the compressible soft

layer. At the same time, the remaining three were

placed directly on the marlstone (Bandera et al.,

2023) due to this geological issue. However, concerns

over the drum-dome system's structural safety

emerged shortly after an extended system of

asymmetric wide cracks opened (Figure 2).

Figure 2: Pattern of the cracks observed by Eng. Garro on

the Basilica of Vicoforte (Bandera et al., 2023).

The structural integrity of the Basilica has been

monitored through permanent static (since 2004) and

dynamic (since 2015) multi-sensor monitoring

systems. The GIANO project aims to recreate

settlements and study how the subsoil's stress-strain

properties and behavior relate to the structure's

behavior. An essential aspect of geotechnical

numerical analyses is the geotechnical

characterization of the site, which includes analyzing

large-scale terrain deformations. This necessity forms

the basis of using the PSInSAR technique for

deformation monitoring.

2 THE EGMS SERVICE

EGMS is part of the Copernicus program, which

involves 30 European states that process and analyze

the Sentinel-1 dataset using a PSInSAR approach.

The project has a monitoring timeframe from 2015 to

2021, with an annual update (European

EnvironmentAgency (EEA), 2020).

The OpeRatIonal Ground motion INsar Alliance

(ORIGINAL) consortium, spearheaded by e-GEOS

from Italy, is responsible for data processing. The

consortium also comprises TRE Altamira from Italy,

NORCE from Norway, GAF AG from Germany, and

the support of five other companies. The processing

considers 750 S1 SAR SLC images, and each stack

consists of 260 SAR scenes for the baseline product.

The data utilized for processing requires a large

Improved Analysis of EGMS Data for Displacement Monitoring: The Case Study of Regina Montis Regalis Basilica in Vicoforte, Italy

amount of disk space, approximately 1.5 PB, which is

expected to increase by 350 TB annually (Crosetto et

al., 2020, 2022).

As previously mentioned, this service is based on

the Sentinel-1 constellation, which consists of two

satellites: Sentinel-1A and Sentinel-1B. Sentinel-1A

has operated since April 3, 2014, while its twin,

Sentinel-1B, was launched into orbit on April 25,

2016. Both satellites have a C-band SAR sensor and

use the Terrain Observation Progressive Scan SAR

(TOPSAR) mode to acquire data (De Zan &

Guarnieri, 2006). The Interferometric Wide Swath

(IWS) satellite imaging mode is a feature that enables

the sensor to capture high-resolution satellite images

with exceptional detail and coverage. With this mode,

the images can be obtained with a remarkable spatial

resolution of 5 meters in the direction parallel to the

satellite's line of flight (azimuth) and 20 meters along

the range. Moreover, the acquired area is extensive,

covering 200 kilometers along the azimuth and 250

kilometers along the range. This wide coverage

provides a more comprehensive view of the area,

making it ideal for large-scale mapping, land-use

monitoring, and environmental studies. Additionally,

this mode provides a revisit time of 12 days for a

single satellite, which means that the same area can

be imaged every 12 days. When two satellites are in

orbit, the revisit time is reduced to only six days,

allowing for more frequent and timely monitoring of

the area (Torres et al., 2012). EGMS is a well-

documented service reported in Costantini et al.,

2021; Crosetto et al., 2020, 2021; Del Soldato et al.,

2021.

The system offers a web-based interface that is

available to all users. This interface allows users to

search for a specific location and view a map of all

scatterers detected in the selected area. By selecting a

scatterer, the system generates a graph that displays

its displacements over time. Noticeable, the platform

is not limited to displaying time series; it allows

registered users to download the raw data in the form

of tables in three different ways (Crosetto et al.,

2021):

1. Basic (Level L2A): basic units of measurement,

which contain the displacement data along the

satellite’s line of sight as provided in the original

radar geometry grid, with geolocation and quality

indicators for each measurement point. These

products are generated for each relative orbit of

satellite image stacks;

2. Calibrated (Level L2B): these products are

derived from Level 2A products and are aligned

to a standard reference frame using external

information such as GNSS data. The Level 2B

products are also integrated and mosaicked using

best geodetic practices, and they are connected to

the EUREF datum for consistency;

3. Ortho (Level 3): East-West and Up-Down

deformation rates produced by combining Level

2B data from ascending and descending orbits,

allowing the East-West and Up-Down

components of the deformation rates to be

estimated. The Level 3 products are resampled on

a grid with 100 m x 100 m.

2.1 Data Visualization

On the EGMS platform, you can view points at

different levels - L2A, L2B, or L3. The L2A and L2B

levels only show movement along the line of sight,

while the L3 level allows you to separate horizontal

(East-West) and vertical (Up-Down) displacements.

However, it is worth noting that the PSs are divided

into 100x100 meter cells at the L3 level, which might

not be appropriate for all applications. Besides, the

grid is set by the platform and cannot be adapted to

the different needs of the users.

This limitation becomes even more apparent

when analyzing a structure's deformation or

displacement, such as the Basilica of Vicoforte. In

this scenario, the platform identifies the grid points

near the church rather than directly on it. As a result,

the displayed displacements are the outcome of

averaging results originating from permanent

scatterers located on the cathedral and its neighboring

elements (Figure 3).

Figure 3: The 100x100-meter grid over the Basilica of

Vicoforte available on the EGMS platform. The

background is the VHR Image Mosaic 2018 (Copernicus,

European Environment Agency).

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

Then, after selecting a location of interest, such as

the red circle shown in Figure 3, a time series graph

corresponding to that location is opened. The chart

displays the date of each acquisition on the x-axis and

the corresponding displacements in millimeters on

the y-axis. The default setting shows the vertical

displacements for the L3 level, but using the Data

Selection Window, it is also possible to display the

other information.

Figure 4 displays a visualization example,

precisely the gridded point highlighted in Figure 3.

This visualization illustrates horizontal and vertical

displacement and includes a synthetic summary

showing the annual mean velocity and the standard

deviation. However, the 100 x 100-meter grid

averaging process makes identifying any clear trend

or seasonality difficult.

Figure 4: Example EGMS chart showing both horizontal

and vertical displacement for a gridded point (L3 level).

2.2 Anomalous Data Management

It is crucial to remove any outliers from a dataset as

they deviate significantly from the behavior of the

rest of the data. In SAR data, outliers may arise from

measurement errors, atmospheric errors, or surface

changes. When considering displacement time series,

removing anomalous or erroneous values is

particularly important because they could

significantly affect the interpretation of the outcomes.

Rejecting blunders makes it possible to obtain a

homogeneous and reliable time series.

Unfortunately, the EGMS platform does not allow

the user to control the removal of outliers from the

displacement time series. The user cannot select the

most suitable method for rejecting data in their case

study, nor can they assess the effect of outliers on the

results. This inflexibility limits the user's capacity to

interpret the data accurately and draw sound

conclusions.

Figure 4 depicts an example of this issue with

some isolated points. Without proper knowledge and

control of data processing, it is challenging to

differentiate between blunders and actual values.

3 METHODOLOGY

Processing SAR datasets can be challenging,

requiring specific skills and dedicated software

packages. The PSInSAR approach is necessary to

monitor deformations. Therefore, massive data must

be processed to create the time series.

The EGMS service is a valuable tool for

conducting this analysis, especially for newbies. The

EGMS consortium extracts persistent scatterers (PS)

for ascending and descending orbits and provides

them on the web platform. This information is then

combined at the L3 level to generate vertical and

horizontal displacements in a gridded format.

However, as mentioned in the conclusion of Section

2, the data provided may not be sufficient for

analyzing a particular building or monument. In such

instances, a more refined grid is necessary as a grid

spacing of 100 meters is too coarse.

Within the GIANO project, the dome's behavior

at the Regina Montis Regalis Basilica in Vicoforte

requires careful analysis and a comparison to

surrounding buildings; in this case, the EGMS

platform's simple consultation is insufficient.

To address these limitations, a package of codes

was implemented using MATLAB release 2023b.

This approach has two main advantages: managing

grid spacing flexibly and customized according to

structure dimension and managing the presence of

potential blunders. Implementing this approach

makes enhancing the data's accuracy, consistency,

and reliability feasible, resulting in more valuable

insights and informed decisions.

3.1 EGMS Tables Management

The first step involves downloading the tables that

contain L2B data for both ascending and descending

geometry for a selected area. However, even if a

region of interest (ROI) is given, the platform

manages the data according to the burst of reference

S1 image. Hence, the persistent scatters downloaded

refer to a larger area. This results in a significant

amount of data, which, in our case, takes up around

1.5 GB per table.

Improved Analysis of EGMS Data for Displacement Monitoring: The Case Study of Regina Montis Regalis Basilica in Vicoforte, Italy

A specific code enables the selection of persistent

scatterers available inside the ROI, disregarding all

others and speeding up the following steps.

3.2 Grid Generation and PSs Indexing

Once both tables have been reduced to the scatterers

present in the ROI, they can be organized according

to the analysis requirements.

The ROI usually has a square or rectangular shape,

so it can be divided into a grid with regular spacing

determined using a parametric approach. Each cell

inside the ROI can be identified by an ID consisting

of two values representing its row and column

position. The program then checks the position of

each PS concerning the generated grid and assigns the

corresponding ID.

Users can then set the cell size according to

specific requirements, such as structure dimensions.

However, it is important to note that the number of

PSs available significantly impacts component

decomposition. At least one ascending and

descending PS is required for this operation, with

having more than one PS improving robustness.

Therefore, the cell size should be calibrated while

considering this requirement.

3.3 Time Series Generation

Displacement time series can then be generated after

harmonizing the two tables and indexing the PSs

based on the selected grid. The tool is made up of

several functions that allow you to perform the

following tasks:

 Removing cells that contain PSs having

discrepancies in their velocity values;

 Identifying outliers in the displacement time

series and recalculating them;

 Remove cells containing only one PS.

 Calculate the average displacement data for

ascending and descending geometries and place

them at the cell's centroid.

 Standardizing displacement dates between two

acquisition geometries by generating time-related

average data (weekly, monthly, quarterly, yearly).

 Generating a series of plots useful for data

interpretation and analysis.

The code tests the congruency of velocities between

PSs in each cell; the Copernicus consortium

determines velocities and available within the

downloaded datasets. Simply removing the PSs that

present such anomalies might not be sufficient to

address the issue of inconsistent velocities. Indeed,

several scatterers in such cells are often found to be

significantly different from one another. Therefore,

removing the entire cell altogether was decided,

providing greater accuracy. To understand which

areas have adequate cell numerosity for subsequent

steps, the code generates two graphs (one for each

acquisition geometry), in which valid cells are

highlighted in green and rejected cells in red.

Besides, to perform a slant displacement

decomposition in horizontal and vertical components,

at least two PSs are necessary - one for ascendant and

one for descendant geometry. If a cell contains only

one PS, the operation cannot be performed, and also,

in this case, the cell will be removed from the next

step.

The displacement time series of every PS is then

checked for possible outliers. Sometimes, the datasets

that contain the slant displacements may have

anomalous values due to atmospheric errors or

surface changes. If these values are not managed

correctly, they can negatively impact the time series

analysis. Instead of removing these values and

creating a gap in the dataset, the code replaces them

with the mean value between the previous and

following displacement. Also, in this case, the code

generates a graphical output for better interpretation.

Once the outliers have been replaced with average

values, the code calculates the mean and standard

deviation of the slant displacement for the two

acquisition geometries in every cell. This allows the

measurement to be regularized into a gridded format;

the calculated values are associated with each cell's

centroid coordinates. A graphic plot is also available

for this step.

Finally, the observations for the two geometries,

ascending and descending, are not taken

simultaneously but with a few days' difference

between them. To overcome this issue, the code

performs a time average between the dates, according

to the user's chosen criteria: weekly, monthly,

quarterly, or yearly. This creates a consistent

timetable for both acquisition geometries, which

helps decompose the displacement into vertical and

horizontal components.

Figure 5 illustrates the workflow described, with

blue shapes representing functions and red ones

indicating graphical outputs.

3.4 Slant Displacement Decomposition

By considering the slant displacements of the

ascending and descending geometry for each cell, it

is possible to calculate both vertical (up-down) and

horizontal (east-west) displacement components

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

Figure 5: Workflow implemented in MATLAB

environment.

(Fuhrmann & Garthwaite, 2019).

In particular, the two elements can be obtained:

























 (1)

Matrix A has the following expression:

A

sin







cos







cos







sin







cos







cos







 (2)

Variables have the following meanings:

 d



: vertical displacement (Up-Down);

 d



: horizontal displacement (East-West);

 d





: displacement along the satellite's line of

sight in ascending geometry;

 d





: displacement along the satellite's line of

sight in descending geometry;

 θ



: angle of incidence of ascending geometry;

 θ



: angle of incidence of descending geometry;

 α



: heading angle of ascending geometry;

 α



: heading angle of descending geometry.

Understanding the displacement signs is crucial to

comprehend the phenomenon in the selected area of

interest. The vertical displacement is negative when it

moves downwards away from the sensor, while it is

positive when moving upwards towards the sensor.

On the other hand, the horizontal displacement is

positive when the point moves in the East direction,

while it is negative when moving in the West

direction. For a better understanding, please refer to

Figure 6, which shows the legend of the signs for

vertical (Up-Down) and horizontal (East-West)

displacements.

Figure 6: Legend vertical (Up) and planimetric (East-West)

displacement.

4 RESULTS

The following is a selection of the results obtained for

the area of Regina Montis Regalis Basilica in

Vicoforte. The data used for the study spans five

years, from February 2016 to December 2021. The

section focuses on the immediate vicinity of the

Basilica, which covers an area of 0.15 square

kilometers, and focuses on a comprehensive analysis

of the dome's behavior.

Figure 7 presents an orthophoto of the area with

the persistent scatterers' availability displayed for

both the ascending and descending geometries. These

PSs are obtained from the L2B data collected from

the EGMS platform and are used to determine vertical

and horizontal displacements. The figure shows that

some areas have scatterers derived from only one

geometry, ascending or descending. The Basilica's

dome has this behavior due to its structure

configuration. Therefore, a careful choice of the grid

size is necessary in this case.

Based on the previous sections, a smaller grid size

is utilized to enhance the analysis of results

obtainable using the standard L3 level. In the EGMS

platform, a cell size of 100 m is used, whereas,

concerning the dome dimension and PSs location, a

cell size of 25 m is used here, which is a quarter of

that. A total of 238 cells were created and distributed

on 17 rows and 14 columns.

In order to gain a clearer understanding of the

distribution of cells within the area of interest, a plot

containing the grid and cell IDs is created. Figure 8

shows the generated grid, represented in black,

superimposed onto the orthophoto of the area. Each

cell containing at least one PS reports its index,

represented by the number of rows and columns;

empty cells correspond to areas without PS.

The code generates a figure displaying the grid

with cell ID in black, the number of scatterers in

Improved Analysis of EGMS Data for Displacement Monitoring: The Case Study of Regina Montis Regalis Basilica in Vicoforte, Italy

ascending geometry in red, and the number of

scatterers in descending geometry in blue (Figure 9).

Figure 9 shows that many cells contain a quantity

of PSs that are insufficient for performing the

decomposition process. Therefore, cells with only

ascending or descending scatterers and cells with

velocity incongruent values must be discarded. Table

1 reports some statistical values on valid and rejected

cells.

Figure 7: PSs positions for ascending acquisition geometry

are in red, and the descending one is in blue. The

background is the WMS service of the Piedmont region for

the orthophoto AGEA 2018.

Figure 8: The grid created for the ROI has a cell size of 25

m. Each cell containing the PSs reports its index,

represented by the number of rows and columns.

The EGMS platform partially bypasses the first

issue, choosing to discretize the ground into cells of

100 m in size since there is more possibility of having

ascending and descending PSs.

Fortunately, for GIANO project purposes, the

cells containing the Basilica’s Dome are sufficient in

number and quality for the next steps. The cells

involved are the (8,6) and (9,6).

Figure 9: Each cell displays the ID in black, the number of

ascending acquisition geometry scatterers in red, and the

number of descending geometry scatterers in blue.

Table 1: Total number of valid and rejected cells.

Acquisition

geometry

Total

cells

Valid

cells

Rejected

cells

Ascending 108 71 (66%) 37 (34%)

Descending 107 73 (68%) 34 (32%)

4.1 Dome Displacements Analysis

Out of the two cells covering the dome, only the cell

(8,6) will be considered as it has more PSs. This cell

is located in the South-Western part of the Basilica's

dome. Specifically, it consists of seven scatterers with

ascending geometry and six with descending

geometry.

The data was analyzed for outliers before

performing slant displacement decomposition. Two

figures were created to display the time series for

ascending and descending geometry - Figure 10 and

Figure 11, respectively. All present PSs (seven in the

first figure and six in the second) were considered.

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

The code highlights any blunders found with a red

point. Some values found blunders are relatively high

for some PSs, especially for ascending geometry.

Figure 10: The outliers identified for the ascending

geometry are indicated in red.

Figure 11: The outliers identified for the descending

geometry are indicated in red.

After replacing the outliers with the mean

displacement between the previous and following

data points, the displacement values for all the PSs in

the same cell are averaged and grouped by geometry.

The plot in Figure 12 shows the time series for the

slant displacement of the considered cell and the two

geometries. The results for the ascending orbit are

displayed in red, while those for the descending orbit

are in blue. The vertical bars in a graph show the

confidence interval of the mean, which is derived

from the standard deviation. This allows us to

visualize the spread of the data that contributed to the

mean. The former geometry generally shows low

standard deviation values, except for a few cases with

high values.

Upon analyzing the time series data, it becomes

immediately apparent that a recurring seasonal

pattern is present over time. Instead, when analyzing

the L3-level data present on the EGMS portal, this

phenomenon is not readily discernible. This is

because the L3-level data are aggregated at a coarser

level, masking the underlying seasonal pattern in the

raw data.

Decomposing the displacement into its vertical

and horizontal components is now possible. Figure 13

displays the displacement trend along with the

associated confidence interval and confidence bounds.

Both displacements exhibit a negligible annual

displacement rate of -0.1 mm/year. Besides, the

previously observed cyclic phenomenon depends

mainly on vertical displacement. Focusing on 2020

and 2021, Figure 14 shows that the positive peaks

occur in July-August, whereas the negative peaks

occur in January-February. This seasonal variation is

probably due to the material used to construct the

dome of the Basilica; indeed, it is made of copper,

which tends to expand and contract based on the

temperature, causing a change in its size. The

horizontal displacement, on the other hand, seems to

remain mostly stable.

Figure 12: Slant displacement time series for the selected

cell. Outcomes for ascending acquisition geometry are in

red, whereas those for descending one are in blue.

Figure 13: Vertical and horizontal displacement time series

for the selected cell.

Improved Analysis of EGMS Data for Displacement Monitoring: The Case Study of Regina Montis Regalis Basilica in Vicoforte, Italy

Figure 14: At the top is the vertical displacement time

series, and at the bottom is the horizontal one. The time

interval has been reduced to the last two years to show how

the maximum peaks occur in the summer period (July-

August) and the minimum peaks in the winter period

(February-March).

The vertical and horizontal displacement signals

were decomposed to understand the displacements

better. This process allows the time series to be

broken down into its components, which include the

trend, seasonality, and residual. Figure 15 displays

the vertical component of the displacement, where the

trend indicates a drop in the structure's position by

around 1.5 mm between 2015 and 2021. The

seasonality is evident with a fluctuation of ±2 mm,

and the residual falls within the same range.

Figure 15: Decomposition of the vertical displacement time

series for cell (8,6). From top to bottom, vertical

displacement seasonality, remainder, and the time series

trend.

Similarly, Figure 16 illustrates the horizontal

displacement, which also shows a decreasing trend,

although the displacement is only 0.6 mm over the

considered period. From 2017 onwards, the

contribution of seasonality appears to be much less

pronounced than for the vertical component, with

variations of ±1 mm. Like the vertical shift, the

residual seems to be within a range of ±2 mm.

Figure 16: Decomposition of the horizontal displacement

time series for cell (8,6). From top to bottom, vertical

displacement seasonality, remainder, and the time series

trend.

Based on the time series of displacement data,

there are no significant deformations in the structure

or the surrounding area. The deformations are

negligible. No localized deformations were observed

in the structure, but a seasonal pattern was identified

in the vertical component. Specifically, a periodic

signal was observed in the vertical displacement, with

maximum peaks occurring in August and minimum

peaks in January. Additionally, there is a slight

seasonal component in the horizontal component,

although it is not as significant as that of the vertical

displacement.

Figure 17: Decomposition of the vertical displacement time

series for cell (9,6). From top to bottom, vertical

displacement seasonality, remainder, and the time series

trend.

It is worth noting that while cell (8,6) has the most

significant number of PSs, it is possible to obtain

similar information for other cells as well. For

instance, Figure 17 shows the vertical displacement

of another cell, cell (9,6), located in the Basilica

dome. The two cells display similar trends regarding

average speed of movement and periodicity. This

similar behavior demonstrates that using tuned

parameters for PS analysis can improve the

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

interpretation of the EGMS datasets. However,

further investigation is needed for 2016, as there are

some conflicting values in Figures 15, 16, and 17.

5 CONCLUSION

The Basilica of Vicoforte is a magnificent religious

structure in a geologically complex area. It is even

more fascinating because it is built on two distinct

geological formations - marl and clay. These two

types of soil have different mechanical properties,

which have caused the Basilica to experience various

foundation failures over time. These failures have

resulted in cracks, detachments, and deformations in

the dome and other structure parts. Due to these issues,

the Basilica has required numerous restoration and

consolidation works since its construction. As a result,

geo-technicians continually monitor the Basilica to

detect new subsidence and intervene promptly to

ensure the structure's integrity.

Geomatics offers a range of noninvasive systems

to monitor structural deformation and displacement.

The Basilica has some monitoring systems, primarily

consisting of accelerometers, which measure the

frequency of structural movements. However, a

means to measure deformation concerning an

external stable framework was not available in the

short or long term. Therefore, a radar approach was

employed.

SAR data processing is a complex and

challenging task that requires specialized skills and

software. A more refined grid is necessary to monitor

deformations using the PSInSAR approach,

especially when analyzing specific buildings or

monuments. While the EGMS service can be helpful

for general purposes, it may not provide the necessary

detail for analyzing specific structures. Fortunately, a

package of codes implemented using MATLAB

release 2023b can flexibly manage grid spacing,

customized according to the dimensions of the

structure under consideration. This approach offers

greater control over the data, making it possible to

identify and manage potential blunders. It also

enhances data accuracy and reliability, providing

valuable insights to inform decisions.

The analysis with a customized MATLAB code

package helped refine the size of the interpolation

grid, manage the presence of any outliers, and analyze

the vertical and horizontal movements in terms of

seasonality and trends. Due to the aggregation level

present on the EGMS platform, this was not directly

possible with the original L3 level. The analysis

highlighted a seasonal phenomenon on the dome of

the Basilica, presumably due to the expansion of the

metal linked to temperatures, having an extent of

around two millimeters. However, no subsidence

trend is evident, recording movements of only 0.1

mm per year.

As no other displacement monitoring systems are

available, comparing the obtained results with other

measurements is not feasible. Nonetheless, the

paper's primary aim was to enhance the analysis of

data extracted from the EGMS database, and this

objective was achieved successfully. As part of the

GIANO project scope, other monitored monuments

will be tested with the same analysis approach,

selecting those with additional monitoring sensors.

ACKNOWLEDGMENTS

The authors thank the Italian Ministry of University

and Research (MUR) for funding the project through

the PRIN 2020 scheme.

REFERENCES

Accardo, G., Giani, E., & Giovagnoli, A. (2003). The risk

map of Italian cultural heritage. Journal of

Architectural Conservation, 9(2). https://doi.org/

10.1080/13556207.2003.10785342

Bandera, S., Gioffrè, D., & Lai, C. G. (2023). Geotechnical

Characterization of the Subsoil at the “Regina Montis

Regalis” Basilica in Vicoforte, Italy. Springer Series in

Geomechanics and Geoengineering. https://doi.org/

10.1007/978-3-031-34761-0_59

Benedikter, R. (2004). Privatisation of Italian cultural

heritage. International Journal of Heritage Studies,

10(4). https://doi.org/10.1080/ 1352725042000257393

Blanco-Sánchez, P., Mallorquí, J. J., Duque, S., & Monells,

D. (2008). The coherent pixels technique (CPT): An

advanced DInSAR technique for nonlinear deformation

monitoring. Pure and Applied Geophysics, 165(6).

https://doi.org/10.1007/s00024-008-0352-6

Cascini, L., Fornaro, G., & Peduto, D. (2010). Advanced

low- and full-resolution DInSAR map generation for

slow-moving landslide analysis at different scales.

Engineering Geology, 112(1–4). https://doi.org/

10.1016/j.enggeo.2010.01.003

Chiorino, M. A., Spadafora, A., Calderini, C., &

Lagomarsino, S. (2008). Modeling strategies for the

world’s largest elliptical dome at Vicoforte.

International Journal of Architectural Heritage, 2(3).

https://doi.org/10.1080/15583050802063618

Costantini, M., Minati, F., Trillo, F., Ferretti, A., Novali, F.,

Passera, E., Dehls, J., Larsen, Y., Marinkovic, P.,

Eineder, M., Brcic, R., Siegmund, R., Kotzerke, P.,

Probeck, M., Kenyeres, A., Proietti, S., Solari, L., &

Andersen, H. S. (2021). EUROPEAN GROUND

Improved Analysis of EGMS Data for Displacement Monitoring: The Case Study of Regina Montis Regalis Basilica in Vicoforte, Italy

MOTION SERVICE (EGMS). International

Geoscience and Remote Sensing Symposium (IGARSS),

2021-July. https://doi.org/10.1109/

IGARSS47720.2021.9553562

Costantini, M., Minati, F., Trillo, F., Ferretti, A., Passera,

E., Rucci, A., Dehls, J., Larsen, Y., Marinkovic, P.,

Eineder, M., Brcic, R., Siegmund, R., Kotzerke, P.,

Kenyeres, A., Proietti, S., Solari, L., & Andersen, H.

(2022). EGMS: a New Copernicus Service for Ground

Motion Mapping and Monitoring. EGU General

Assembly 2022, 23-27 May 2022.

Crosetto, M., Monserrat, O., Cuevas-González, M.,

Devanthéry, N., & Crippa, B. (2016). Persistent

Scatterer Interferometry: A review. In ISPRS Journal of

Photogrammetry and Remote Sensing (Vol. 115).

https://doi.org/10.1016/j.isprsjprs.2015.10.011

Crosetto, M., Solari, L., Balasis-Levinsen, J., Bateson, L.,

Casagli, N., Frei, M., Oyen, A., Moldestad, D. A., &

Mróz, M. (2021). Deformation monitoring at european

scale: The copernicus ground motion service.

International Archives of the Photogrammetry, Remote

Sensing and Spatial Information Sciences - ISPRS

Archives, 43(B3-2021). https://doi.org/10.5194/isprs-

archives-XLIII-B3-2021-141-2021

Crosetto, M., Solari, L., Barra, A., Monserrat, O., Cuevas-

González, M., Palamà, R., Wassie, Y., Shahbazi, S.,

Mirmazloumi, S. M., Crippa, B., & Mróz, M. (2022).

ANALYSIS OF THE PRODUCTS OF THE

COPERNICUS GROUND MOTION SERVICE.

International Archives of the Photogrammetry, Remote

Sensing and Spatial Information Sciences - ISPRS

Archives, 43(B3-2022). https://doi.org/10.5194/isprs-

archives-XLIII-B3-2022-257-2022

Crosetto, M., Solari, L., Mróz, M., Balasis-Levinsen, J.,

Casagli, N., Frei, M., Oyen, A., Moldestad, D. A.,

Bateson, L., Guerrieri, L., Comerci, V., & Andersen, H.

S. (2020). The evolution of wide-area DInSAR: From

regional and national services to the European ground

motion service. In Remote Sensing (Vol. 12, Issue 12).

https://doi.org/10.3390/RS12122043

De Zan, F., & Guarnieri, A. M. (2006). TOPSAR: Terrain

observation by progressive scans. IEEE Transactions

on Geoscience and Remote Sensing, 44(9). https://

doi.org/10.1109/TGRS.2006.873853

Del Soldato, M., Confuorto, P., Bianchini, S., Sbarra, P., &

Casagli, N. (2021). Review of works combining GNSS

and insar in Europe. In Remote Sensing (Vol. 13, Issue

9). https://doi.org/10.3390/rs13091684

EMGS. (2024). EGMS. https://egms.land.copernicus.eu/

European Environmnet Agency (EEA). (2020). Copernicus

Land Monitoring Service 2020. European Environment

Agency (EEA).

Ferretti, A., Prati, C., & Rocca, F. (2001). Permanent

scatterers in SAR interferometry. IEEE Transactions on

Geoscience and Remote Sensing, 39(1). https://

doi.org/10.1109/36.898661

Fuhrmann, T., & Garthwaite, M. C. (2019). Resolving

three-dimensional surface motion with InSAR:

Constraints from multi-geometry data fusion. Remote

Sensing, 11(3). https://doi.org/10.3390/rs11030241

Hooper, A., Zebker, H., Segall, P., & Kampes, B. (2004). A

new method for measuring deformation on volcanoes

and other natural terrains using InSAR persistent

scatterers. Geophysical Research Letters, 31(23).

https://doi.org/10.1029/2004GL021737

Jutz, S., & Milagro-Pérez, M. P. (2020). Copernicus: The

European earth observation programme. Revista de

Teledeteccion, 2020(56). https://doi.org/10.4995/

raet.2020.14346

Showstack, R. (2014). Sentinel Satellites Initiate New Era

in Earth Observation. Eos, Transactions American

Geophysical Union, 95(26). https://doi.org/10.1002/

2014eo260003

Torres, R., & Davidson, M. (2019). Overview of

Copernicus SAR Space Component and its Evolution.

International Geoscience and Remote Sensing

Symposium (IGARSS). https://doi.org/10.1109/

IGARSS.2019.8899134

Torres, R., Davidson, M., & Geudtner, D. (2020).

Copernicus Sentinel Mission at C- and L-Band: Current

Status and Future Perspectives. International

Geoscience and Remote Sensing Symposium (IGARSS).

https://doi.org/10.1109/IGARSS39084.2020.9323149

Torres, R., Snoeij, P., Davidson, M., Bibby, D., & Lokas,

S. (2012). The Sentinel-1 mission and its application

capabilities. International Geoscience and Remote

Sensing Symposium (IGARSS). https://doi.org/10.1109/

IGARSS.2012.6351196

Wright, T. J., Parsons, B. E., & Lu, Z. (2004). Toward

mapping surface deformation in three dimensions using

InSAR. Geophysical Research Letters, 31(1).

https://doi.org/10.1029/2003GL018827

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management