New integrated geodetic monitoring system at Stromboli volcano (Italy)

www.elsevier.com/locate/enggeo

Engineering Geology 7

New integrated geodetic monitoring system

at Stromboli volcano (Italy)

G. PuglisiT, A. Bonaccorso, M. Mattia, M. Aloisi, A. Bonforte, O. Campisi,

M. Cantarero, G. Falzone, B. Puglisi, M. Rossi

Istituto Nazionale di Geofisica e Vulcanologia—Sezione di Catania, Piazza Roma, 2 Catania, Italy

Received 15 February 2004; received in revised form 8 August 2004; accepted 1 October 2004

Available online 25 April 2005

Abstract

Stromboli is a composite volcano, forming the northernmost island of the Aeolian Archipelago volcanic arc (Tyrrhenian

Sea). On December 28th 2002, lava flows outpoured from the northern wall of the NE crater and descended into the Sciara del

Fuoco (SdF), a deep depression marking the NW flank of the volcano edifice. On December 30th 2002, two landslides occurred

on the northern part of the SdF; they moved a mass in the order of tens of millions of cubic meters both above and below sea

level. The landslide produced a tsunami causing significant damage to the eastern coast of the island, reaching the other Aeolian

Islands and the Sicilian and southern Italian coasts. This event led to the upgrading of the ground deformation monitoring

system, already existing on the island; the new requirement was the real-time detection of the deformations related to potential

slope failures of the SdF. To this end, a remotely controlled monitoring system, based both on GPS and geodetic techniques,

was planned and set up in a few months. The new monitoring system allowed to continuously measure the ground deformations

occurring on the SdF, by integrating both terrestrial and satellite geodetic techniques. Although this system was severely

damaged during the 7-month long eruption, it allowed to monitor important eruptive phases. For the first time, an accurate data

set on the actual mass movements of the SdF and the crater area was available. It provided data that significantly supported the

Civil Protection Authorities in making decisions and constrains theories related to landslide movements and volcanic activity.

After the end of the eruption, the system was overhauled in order to optimize the instruments and set up a monitoring system

aimed at measuring deformations forecasting other flank collapses.

D 2005 Published by Elsevier B.V.

Keywords: Stromboli; Ground deformation; Real-time monitoring; Eruption forecasting; Landslide forecasting

0013-7952/$ - see front matter D 2005 Published by Elsevier B.V.

doi:10.1016/j.enggeo.2004.10.013

T Corresponding author. Fax: +39 95 435801.

E-mail address: [email protected] (G. Puglisi).

1. Introduction

Stromboli is an active volcano, about 2500 m high

above the sea floor. Roughly only the last kilometer of

this volcano emerges from the sea, forming an island

whose diameter ranges from 2.4 to 5 km. It belongs to

9 (2005) 13–31

G. Puglisi et al. / Engineering Geology 79 (2005) 13–3114

the Aeolian Islands and represents the most active

volcano of this archipelago (Fig. 1). Its conic shape is

clearly characterized by a big depression that marks

the northwestern flank of the edifice: the Sciara del

Fuoco (SdF); the uppermost southeastern rim of the

SdF (Pizzo Sopra la Fossa—PSF) is the highest point

on Stromboli island (924 m a.s.l.). SdF continues

below the sea, down to 700–750 m b.s.l.; this scar was

produced by a huge sector collapse, in the order of 1

km3 that very likely originated in the Holocene. Three

active craters are aligned along a SW–NE trend, at

about 750 m a.s.l., in the central part of a flat area

(named Crater Terrace, Bertagnini et al., 1999) that

characterizes the upper southeastern part of the SdF,

just below PSF. Usually, the craters are called 1 to 3

Fig. 1. Ground deformation monitoring system before the 2002–2003 eru

during the first two days of the 2002 eruption: the fracture opened on the n

(black area) and the slid area (gray area). In the inset, the geographical se

from the northeast to southwest (Bertagnini et al.,

1999). The volcanic activity of Stromboli has been

well known since the Greek epoch and several

chroniclers report the almost regular and continuous

explosive activity occurring at its craters. Thanks to

this peculiar constant activity, volcanologists have

termed the intermittent explosive activity often

observed on active volcanoes as dstrombolianT. Some-

times, flank eruptions start from vents opening on the

outer side of the Crater Terrace, producing lava flows

that reach the sea, running along the steep flanks of

the SdF. Barberi et al. (1993), by analysing historical

data of the Stromboli eruptive activity, define a

medium interval between lava eruptions of 3.7 years,

with a maximum interval of 10 years.

ption onset. The map also shows the main volcanic features formed

orthern flank of the NE crater (white line), the outpoured lava flows

tting of Stromboli volcano.

G. Puglisi et al. / Engineering Geology 79 (2005) 13–31 15

The volcanic hazard of Stromboli is mainly

characterized by explosive activity, and in particular

by paroxysmal episodes that, for example, killed 10

people and injured dozens in 1919 and 1930. In the

last century, ballistic ejecta emitted during these

paroxysms destroyed many buildings and fires

induced by scoria caused severe damage to crops.

During some paroxystic explosions, tsunamis have

Fig. 2. Ground deformation monitoring system in the SdF; triangles mark t

used as reference frame; stars mark the control points (C.P.) installed to mo

networks are distinguished with different colors.

been signaled (6 times since 1879; Barberi et al.,

1993). Although chronicles of these events do not

report any specific origin of these phenomena, Barberi

et al. (1993) relate it to hot avalanches or pyroclastic

flows entering the sea during the paroxysm.

Since 1993, Stromboli ground deformations have

been monitored by GPS and clinometric techniques.

Before the 2002–2003 eruption, the monitoring

he pillars where measurements were carried out and the benchmarks

nitor movements. The control points belonging to the different sub-


system on the island was aimed at measuring the

effects of the plumbing system dynamic at medium-

shallow depths. To this end, four GPS and three

tiltmetric permanent stations were set up between

1993 and 1997 at a lower altitude (50–150 m) around

the island’s perimeter (Fig. 2). This system was able to

detect magmatic intrusions as, for instance, in 1995

(Bonaccorso, 1998). The Stromboli 2003 volcanic

emergency prompted setting up a powerful and

complex system for monitoring ground deformations

both in the SdF area and over island as a whole, based

on the integration of different surveying techniques:

terrestrial (first EDM and then Automatic Total

Station) and satellite (GPS).

2. 2002–2003 eruption

On December 28th 2002, after about 6 months of

increasing explosive activity at the summit craters,

accompanied by some lava flows from the NE rim

Crater, a NE-trending eruptive fissure broke the

Crater 1 flank and the upper eastern part of the SdF

(Fig. 1). The lava outpoured from this fissure, flowed

through the SdF and reached the sea during the

evening; the emission lasted about half a day and on

the morning of December 29th the lava flows were no

longer being fed from this fissure. In the afternoon,

December 29th, two vents opened at about 500–600

m altitude, within the SdF, forming new lava flows.

On December 30th, at about 13:15 local time, two

landslides broke off from the northern part of the SdF,

both on the emerged and submarine parts of the scar,

and generated a tsunami that caused severe destruc-

tion on the island, produced some damage to the

Aeolian archipelago and the northern coast of Sicily,

and was observed in several coastal places in the

south-Tyrrhenian sea. The total volume of landslides

was estimated at about 17 million cubic meters from

photogrammetric and bathymetric data (reported in

Tinti et al., 2003). After these events, the eruption

continued by showing only effusive activity on vents

opened at about 500 m altitude, more or less in the

same position as those formed on December 29th. On

February 16th, after a gradual decrease in effusive

rate, new vents opened at about 600 m altitude, along

the fissure formed during the first day of the eruption,

feeding a new lava flow. Up to the end of March

2003, no explosive activity was observed on the

Summit Craters, although the seismic network

recorded seismic signals related to strombolian

degassing processes. On April 5th, at 9:13 local time,

a violent paroxysmal explosion occurred at the

Summit Craters (1 and 3), producing pyroclastic

flows within the SdF and scoria fall out that mainly

reached the southwestern part of the island. Bombs,

some cubic meters large, reached the Ginostra village,

located on the eastern coast of Stromboli, destroying

walls and houses. This explosion may be compared

with an analogous one occurring in 1930, when 4

people were killed and for many years the island was

abandoned for the fear of another similar episode.

After this event, the explosive activity gradually

resumed at the Summit Craters (now only 1 and 3).

The effusive vents located at 600 m altitude produced

a compound lava flow field that during the following

months entirely covered the upper part of the land-

slide niche, located just below the SdF’s rim, named

Bastimento, and formed a fan that reached the sea. At

the end of the effusion, on July 22, the thickness of

lava field in the area below the Bastimento was in the

order of 30–50 m.

3. Monitoring system description

The events of December 2002 (landslides and

tsunami) compelled geophysicists and volcanologists

to rapidly confront new phenomena having a serious

impact and producing a broad risk scenario, ranging

from simple lava flow emissions to a sector collapse

of the NW flank of the volcano. After the December

30th 2002 landslide, it became evident that the most

urgent requirement of the monitoring system was to

detect the deformations related to potential slope

failures of the SdF. For this purpose, a complex

monitoring system, based on the integration of

different measurement techniques, was planned; the

set up of this new monitoring system took several

months and was developed in three steps. At the same

time, the breinforcingQ of the previous networks (Tilt

and GPS) was carried out.

On January 16th 2003, a first monitoring network,

based on EDM measurements, was set up on the

upper flank of the SdF, around the craters area. Before

the end of January 2003, this network was enlarged,


including all the northern sector of the Sciara, along

the path of the 30th December 2002 landslide. At the

beginning of February 2003, there were 22 monitoring

control points within the Sciara and two pillars were

installed for setting up the EDM instruments.

The next step was the installation of a permanent

GPS network within the SdF, specifically designed for

the real-time monitoring of movement in the core of

the potential failure zone and close to the summit

craters. By February 12th 2003, the stations were built

and the network was up and running.

The last step was the installation of a robotized

Total Station to automate the terrestrial geodetic

measurement along the northern sector of the SdF.

This system required more than 4 months to be

routinely introduced in the monitoring system, both

for technical and volcanological reasons (e.g. the

April 5th 2003 paroxysmal explosion).

In this paper we describe these three steps and

provide technical details and scientific implications.

3.1. Terrestrial geodetic networks

3.1.1. EDM monitoring network

The EDM was the first technique adopted to

evaluate the hazard evolution in the SdF during the

initial days of the volcanic crisis. The first sub-

network installed on January 16th, on the summit area

of the volcano was in the Fossa area, the part of Crater

Terrace included between the PSF and the Craters.

This sub-network (Fig. 2) consisted of five reflectors

(named FOS1 to FOS5) that were installed around the

southeastern flanks of the summit craters, along a

profile crossing several fractures and fumaroles on the

ground. These control points were surveyed from the

iron pillar set up at PSF (STR in Fig. 2). At the same

time, four reflectors (BAST, SLF, FORT and VANC)

were installed on the upper part of the volcano outside

and around the upper part of the SdF area, on sites

considered stable (Fig. 2). This second sub-network

allowed to check the actual stability of these areas and

verify the repeatability of the distance measurements

carried out from the STR pillar, in order to detect

systematic errors due, for example, to atmospheric

effects. A third monitoring sub-network was installed

below the Bastimento. This sub-network consisted of

six control points (from PST1 to PST6), placed in the

area of the niche of the December 30th landslide (Fig.

2). These benchmarks were measured from the STR

pillar and two of them (PST4 and PST6) were

equipped with two reflectors, in order to be measured

also from another pillar, located on the lower flank of

the volcano, at Punta Labronzo (PLB in Fig. 2). This

pillar was set up with the aim of enabling monitoring

on the lower part of the SdF area, not visible from the

STR pillar. In fact, the suitability of the station of

Punta Labronzo allowed the installation of another six

reflectors along the SdF northern slope, from the

niche area down to the sea (SDF1 to SDF6, Fig. 2).

Finally, two other benchmarks were installed in the

central part of the SdF (SDF7 and SDF8) and a third

pillar at Punta del Corvo was set up (CRV), with the

view to enlarging the network and controlling the

western sector of the Sciara in the future (Fig. 2).

Throughout the EDM monitoring activities, angu-

lar measurements were carried out from the pillars by

using a theodolite (DKM3), in order to evaluate the

three-dimensional movements of the benchmarks. The

long time required to perform accurate angular

measurements and the severe operative conditions

did not allow using the theodolite as frequently as the

EDM, so that the data set of this first stage of the

monitoring activity in the SdF was based mainly on

distance variations. In any case, the 3D vectors,

through integration of EDM and theodolite measure-

ments, provided helpful information on the dynamics

(direction and displacement velocity) of the landslide.

3.1.2. Automatic EDM measurements (THEODOROS)

The experience of the first days of monitoring

activities highlighted the need to both automate the

geodetic measurement and provide 3D displacement

vectors of all the measured control points in the SdF.

To achieve these objectives a robotic total station has

been installed near the Labronzo pillar (Fig. 3).

The set up of this new system was a real challenge,

given the logistic and operative conditions existing on

Stromboli during the volcanic crisis. The chosen

instrument was the Leica TCA 2003 Total Station

equipped with GeoMos software that allows remote

control of the sensor. The acronym of this new system

was THEODOROS (THEOdolite and Distancemeter

Robot Observatory of Stromboli). To guarantee a

continuous stream of data from the instrument, it

requires a continuous power supply and a continuous

link with the PC that controls the Total Station’s

Fig. 3. Robotized Total Stations belonging to THEODOROS

system.


activities, installed on the S. Vincenzo Observatory

(SVIN), where the National Department of Civil

Protection (DPC) control room is located. These

technical solutions were achieved by exploiting two

facilities already installed by the DPC and the Istituto

Nazionale di Geofisica e Vulcanologia (INGV) during

the volcanic crisis: the electric line for supplying the

DPC satellite video-camera and the optical line

installed by INGV to collect all geophysical data

from the Punta Labronzo area to the S. Vincenzo

Observatory. Because the head of the optical line was

near the Labronzo lighthouse, we have implemented a

WiFi link to connect the Total Stations with the

optical line. Although the two-way link system was

quite complex (optical+WiFi), it was efficient enough

to guarantee the real-time control of the Total Stations

from the Observatory.

A reference system of five reflectors (400, CURV,

BORD, SEMF, SPLB2) was installed around the Total

Station, on sites we considered stable; it aimed at

guaranteeing the same orientation in the space (both

horizontally and vertically) through the measurements

of the Total Station (Fig. 4). At the very beginning

THEODOROS carried out the measurements only on

four reflectors installed on the SdF (SDF5, SDF6,

SDF7 and SDF8), which were the only remaining

ones after the destruction due to the lava flows, and on

three reflectors installed around the site belonging to

the reference system (SEMF, 400, SPLB2). The fine-

tuning of the THEODOROS system required several

months because it was necessary to identify and

overcome all possible error sources or phenomena

perturbing the data acquisition. These adjustments

intended to achieve the highest accuracy. Among the

various interventions required to improve the system,

we mention the building of a cabin to reduce solar

perturbations on the different parts of the system

(concrete basement, pillar, sensor, etc.; see Figs. 3 and

4), the improvement of the configuration of the

reference system, the improvement of the of the PC

controlling the system, the assessment of the best

strategy to carry out the measurements, etc. At the end

of July, when the eruption finished, THEODOROS

was already running properly. The eruption end

allowed the implementation of the network by setting

up other reflectors in the SdF and redefining the

reference network. The actual system aims to follow

the post-eruptive movements of the recent lava flow

field. At present, THEODOROS counts 5 reflectors

for the reference system (400, BORD, SEMF, SPLB2

plus ELIS), 10 reflectors for monitoring movements in

the SdF (the four previously installed plus SDF9,

SDF10, SDF11, SDF12, SDF14, SDF16) and 2 more

reflectors to check the stability of the measurement

both on short and very long distance measurements

(CURV and CRV) (Fig. 4). The current measurement

strategy assumes that the horizontal geometry and the

relative distance of the reference system are fixed (by

fixing the coordinates of the reference system bench-

marks), while the Total Station adjusts its own vertical

positions by using only the internal vertical compen-

sators. The system performs 48 cycles of measure-

ments per day. At the beginning of each cycle, the

Total Station first carries out a series of measurements

on the reference system to define its orientation and

Fig. 4. THEODOROS system: end of July 2003 configuration.


the atmospheric effect; afterwards, it measures all 17

reflectors, by carrying out three series of left/right

measurements. All these choices result from the 3-

month-long testing of the THEODOROS system.

3.2. GPS permanent network (SciaraDat)

After the events of December 30th, the INGV

section of Catania, in agreement with the National

Department of Civil Protection, decided to set up a

GPS network, aimed at the real-time monitoring of

displacements in the highest part of the SdF, under the

realistic hypothesis that any future flank collapse

should be preceded by initial deformations in these

areas. To achieve this objective, two challenging

technical obstacles had to be overcome: the first

concerns the capability to install and remotely control

autonomous GPS stations within the SdF, and the

second is to produce positioning data of these stations

with a compatible time scale to the expected precursor

phenomena. The overcoming of these drawbacks led

to setting up the SciaraDat GPS permanent network

in less than 1 month. SciaraDat counted four stations

[SCRA, SDIC, SSBA, SVIN; pink circles in Fig. 2]

and a radio bridge near the PSF.

Adequately designed stations were placed on the

upper SdF slope and around the summit craters by

using helicopters, thus reducing the risk. The National

Department for Civil Protection provided two heli-

copters with high loading capacity that were used for

the two different types of remote stations, specifically

designed and realized: the first blightQ one (Fig. 5) wasused in SCRA, SDIC and for the radio-bridge near the

PSF, while the bheavyQ one (4000 kg; (Fig. 6)) was

used for the SSBA (Sotto Bastimento). All stations

had a prismatic shape, with triangular cross section;

one of the sides is the basement of the stations while a

second one supports the solar panels. Instruments and

batteries are contained within the prism.

Two identical iron frames formed each blightQstation. A steel pillar, ballasted with concrete (approx-

imately 400 kg) and buried, acted as the benchmarks

for the GPS antenna.

The set up of the heavy station was required in the

area below Bastimento because here it was impossible

to install any pillar and in order to work safely for

enough time to connect cables, arrange the iron

structure, etc. For these reasons, we planned to build

a station capable of being the benchmark for the GPS

antenna itself and able to be left on the field by only

Fig. 6. Photos and designing of SSBA station.

Fig. 5. Photos of SCRA and SDIC stations. We show the design of a steel pillar ballasted with concrete and buried that was used as the

benchmarks for the GPS antenna.



one helicopter flight, without the work of specialized

technicians on the ground (Fig. 6). A special high

density concrete with melted iron was used for a fast

construction (48 h), without the use of traditional

reinforced concrete. The GPS antenna was placed on

the roof of the station. The traditional approach based

on the post processing of 24 h/day of GPS measure-

ments, sampled every 30 s, was not able to give fast

enough responses for an early warning to the

population of large regions of Southern Italy, poten-

tially affected by a large-scale tsunami (Tinti et al.,

1999). To provide data at suitable frequency, we

adopted the CRNet software (Leica and Geodetics

Inc., 2002) which applies the LAMBDA (Least

squares AMBiguity DecorrelAtion technique; Teunis-

sen, 1993) method for the fast resolution of the integer

ambiguities of the beating phase and is able to resolve

this problem in the so-called bEpoch-by-EpochQ(Bock et al., 2000) algorithm. Each time a single

epoch of GPS data is collected, this algorithm

produces a 3D position of the remote stations, with

respect to a bmasterQ whose coordinates are assumed

fixed. The levels of accuracy in positioning are in the

order of 1–2 cm in planimetry and 2–6 cm in altimetry

as demonstrated with internal tests performed on a

selected baseline on Mt. Etna during the month of

August 2002 (Mattia and Rossi, 2004). Other param-

eters used in the current operations of the data

processing procedure were the 108 cutoff angle and

the use of elevation dependency in variance/cova-

riance matrix, while we did not calculate the tropo-

spheric delay because the calculated baselines were

very short (1.5–3 km). In this way, we are potentially

able to retrieve 20 Hz frequency 3D positions data or

more, depending only on the GPS receiver character-

istics, and each epoch is independent of the other. For

optimizing and speeding-up the network setup, we

used Leica SR530 receivers, equipped with the Leica

AT 502 antennas, for the three remote stations

installed on the slopes of the volcano, and Leica AT

504 choke ring for the master station in San Vincenzo

Observatory. The total power consumption of the

remote stations amounted to 9 W, so four solar panels

(maximum power 65 W each one) and two deep cycle

110 Ah batteries have been utilized. We adopted 1 Hz

as sample frequency at SciaraDat.

The choice of the characteristics of the data

transmission system was strictly related to the adopted

software. The main requirements were to have enough

band amplitude and bi-directionality of the connec-

tions between the remote stations and the PC where

the CRNet software runs. In Fig. 7 the whole system

is shown. The system is divided in two parts: the first

connects the GPS receivers in the SdF to a radio-

bridge and the second one connects the radio-bridge

to the San Vincenzo Observatory, in the village of

Stromboli. The UHF-band radio-modem connects the

serial ports of the Leica SR530 remote receivers to the

bridge located near PSF, where a Serial-to-Ethernet

converter routes the signal (transformed in TCP/IP

protocol) over a Wi-Fi board (2.4 GHz, 802.11b

standard). A high-gain (24 dB) antenna on the San

Vincenzo Observatory roof receives the signals from

the bridge and then a Local Area Network (LAN)

connection brings the received data stream to the PC

where the software manages the raw GPS files. To

guarantee the continuity of the data stream, specific

software was implemented on the PC, simulating the

flux when the signal from one or more station was

absent; this might happen, for instance, when tempo-

rary failure occurs at the receiver or the radio-modem.

On the PC located at the master station of San

Vincenzo where CRNet runs, the data collected from

the remote stations are archived both in RINEX and

raw Leica LB2 format. The software processes the

data with 1-s latency and displays different levels of

positioning results on different windows on the PC

screen. Usually three kinds of result were suitable for

each remote station: the 1 Hz position, the median of

10 min of data positions and the median of 1 day of

data positions. One of the aims of this network was to

produce an alarm when significant deformation

phenomena are observed on the GPS data. To achieve

this, some useful statistics were introduced. CRNet

multi-epoch statistics are based on two robust

estimators, the median and the interquartile range

(IQR), which are less sensitive to data outliers than the

traditional mean and standard deviation. When a data

sample is drawn from a normal distribution, its mean

very nearly equals its median and its standard

deviation equals about three quarters of the IQR.

For intervals longer than a single epoch, the position

coordinates are determined as the median value of all

single-epoch positions within the interval, excluding

position coordinate outliers which are defined as

appropriate multiple of the IQR. Although the 1 Hz

Fig. 7. Sketch of SciaraDat data flow.


data are very precise, multipath effects and local

transient radio signals can potentially degrade the

level of precision. For this reason the alarm has been

set up on the median of 10-min data set; for a fist test

we set the threshold alarm at 0.1 m on planimetry and

0.2 m in altimetry.

4. Data discussion

Data collected from this complex monitoring

system allowed to effectively monitor the deforma-

tions related both to the downslide movements of the

SdF and to the evolution of the volcanic activity. Here

we briefly discuss some significant volcanological

events observed in the data produced from the

monitoring system. For simplicity we can group the

data in three intervals, based on the main events

recorded during the volcanic eruption.

4.1. From January up to 16 February 2003

The first measurements collected from the STR

pillar confirmed that the reference network installed

around the summit area of the Stromboli did not

record any significant deformation (Fig. 8); this was

the first significant result of the monitoring that

excluded the possibility of a sector collapse involving

the summit craters area that might produce a large

phreatomagmatic explosion and a giant tsunami.

The measurements carried out on the benchmarks

located within the Fossa area allowed us to make some

hypotheses on the dynamic of the craters. We observed

a stepwise increase in the distances between the STR

pillar and the benchmarks of Fossa. This peculiar

deformation pattern could be linked to the movements

of the magmatic column within the craters; probably,

when the top of the magma drops down, the distance

increases between the crater flanks and the STR.

Fig. 8. EDM data collected from STR pillar, related to the bReference SystemQ and bFossaQ sub-networks.


The monitoring system gave us the opportunity to

observe the effects of the migration of effusive vents

from 550 to 600 m altitude. This phenomenon

produced a deformation field, whose evolution was

monitored both by GPS and EDM systems. Since

February 15th, several distances measured both from

the STR and PLB pillars showed significant varia-

tions with respect to the previous days, on bench-

marks located below the Bastimento area (Fig. 9).

Similarly, the SSBA GPS station, just installed,

measured a significant NW displacement. The climax

of this phenomenon was on February 16th. The

integration between the 3D movement detected by

the GPS station and the dense distribution of the

reflectors installed below the Bastimento enabled us

to understand this phenomenon. In fact, the point

PST4 recorded a very high ground deformation

(about 80 cm between 15 and 16 February) while

the other points close to PST4 recorded one order of

magnitude smaller movements (Fig. 10). The SSBA

data showed that the source of the movement was

upslope from the station. The abrupt decay of the

Fig. 9. EDM data collected from STR pillar, related to the sub-network, installed below the Bastimento. The data are shown in the top and

bottom panels at different scale.


deformation and the stability of the reference points

on the summit area, as well as the absence of

movements at the GPS permanent network at the

foot of the volcano, confirmed that this phenomenon

was very local and did not affect either the entire

volcano or the whole unstable flank. This sustained

the hypothesis of a very shallow migration of

magma, towards the new active vents that opened

on February 16th, at 600 m altitude. All these data

have been provided in breal timeQ to the scientists

involved in the monitoring system. This capability,

the appropriate dense network configurations and the

integration between information provided by differ-

ent techniques, allowed us to give proper and timely

information to the DPC, avoiding raising a false

alarm of the possible large failure of SdF flank (as

Fig. 10. 3D vectors deduced by EDM and theodolite at PST4, from February 13th to February 19th. To show the abrupt decay of the

deformation, we reported a table with the modules and the vertical variations of many points located near the PST4.


hypothesized in the very first hours of this ground

deformation event).

4.2. From 16 February to April 5th 2003

After the opening of the vents at 600 m we

observed a slowing down of the ground deformation

pattern along the Sciara flank (Figs. 8, 9 and 11). The

activity of the vents at 600 m destroyed SSBA station

and many reflectors installed below the Bastimento

and on the SdF slope; furthermore, the emplacement

of lava flows along the line of sight from the Punta

Labronzo pillar hid the upper reflectors in the SdF.

The points on the lower sector of the SdF, near the

sea, (SDF5 and SDF6) recorded movements probably

related to the lithostatic pressure of the lava front (Fig.

12). Moreover, the reflectors at the center of the

Sciara did not record significant ground deformation,

demonstrating the stability of this part of the slope

(Fig. 11).

The sudden paroxysmal explosion occurring on

April 5th at about 7:13 GMT destroyed the entire

permanent monitoring system of the summit area. The

SciaraDat network had recorded data until the very

first phase of the explosions. These data did not show

large variations, except in the last seconds, when

variations in the order of decimeters were observed.

This is the first time that ground deformations

produced by a paroxysmal volcanic explosion was

recorded with such detail both in space and time, so

the volcanological interpretation of these data is a

very difficult task and certainly beyond the scope of

this paper. However, the main insight from this

experience is that no particular precursors in terms

Fig. 11. EDM data collected from PLB pillar. As described in the text, PST4 recorded a very high ground deformation between 15 and 16

February, so it goes off the plot early in the time-series.


of ground deformation have been observed in a

minimal span of time for a reasonable alert for this

kind of volcanic phenomena. In particular, from the

available data set we can deduce that the source of the

ground deformations associated with the paroxysmal

events is located in the body of the volcano edifice,

very near the crater area, and this might be the same as

Fig. 12. 3D vectors deduced by EDM and theodolite at SDF5 and

SDF6, from February 13th to March 21st.

the normal strombolian activity (Chouet et al., 2003),

as also hypothesized by Falsaperla and Spampinato

(2003).

4.3. After April 5th 2003

After the explosion of April 5th, the geodetic

networks were seriously damaged; only four bench-

marks (SDF5, SDF6, SDF7 and SDF8) out of more

than 20 remained. Unfortunately, these benchmarks

are not in ideal positions for detecting movements

related to eventual sliding on the SdF by using EDM

only. Due to their position, the principal component of

this kind of movement is almost orthogonal to the line

of the sight, so that the number of measurements

carried out from Labronzo pillar diminished and the

principal efforts were focused at setting up THEO-

DOROS. The comparison among the few distance

measurements carried out from the pillar and the first

measurements carried out by the THEODOROS

system–during its test phase–confirm that the ground

deformation pattern of four reflectors is quite stable up

to the end of July 2003 (Fig. 13).

The following improvement in the network con-

figuration allowed monitoring the post-eruptive

Fig. 13. EDM data collected from PLB pillar integrated with the data recorded by THEODOROS.


assessment of the SdF (Figs. 14–16). Looking at the

summary of the measurement data set from the

beginning of August up to the end of December

2003, we can confirm that the reference system

around the Total Station is stable, both considering

the near (CURV) and far (CRV) benchmarks. More-

over, the monitoring of the SdF enabled us to

distinguish different ground deformation patterns.

First of all, the movements are detected only in the

sector of the SdF where the December 30th 2003

landslides occurred, while on its central part (SDF7

and SDF8 benchmarks) the flank of the volcano is

stable (Fig. 15). Within the areas showing movements,

different deformations rates are observed between the

lower and the upper part of the slope. On the lower

part (SDF5 and SDF6), the rates are relatively small

(in the order of mm/day) and seem to increase on the

lowest benchmarks (Fig. 15). On the upper part of the

SdF, at the benchmarks located on the recent lava flow

field, we measured both horizontal and vertical

movements up to 1 cm/day. However, we can

distinguish different areas with homogeneous rates.

The fastest area is located along the northern

boundary of SdF (SDF9, SDF10 and SDF12) (Fig.

16). SDF14 and SDF11 benchmarks seem located

over another area where the rate of movement is in the

order of 0.5 cm/day (Fig. 16). A further block, where

SDF16 is located, has practically stopped its move-

ment today (Fig. 16).

5. Conclusions

The twofold aspect (eruption and landslide) of the

Stromboli volcanic crisis emphasizes the role of

geodetic monitoring activities on active volcanoes.

Flank failures have occurred worldwide on several

volcanoes during their history, so the experience

gained on Stromboli might be usefully exploited in

other similar cases.

The strategy adopted at Stromboli was to imple-

ment as many different types of survey techniques as

possible (taking economic and logistic constraints into

account), by choosing those able to provide real-time

or near real-time data, deployed in the field by

appropriate configuration, in such a way that if one

technique should fail another would ensure continu-

ous information on the current dynamic.

This experience underlines that the current devel-

opment of surveying techniques allow performing a

detailed breal-timeQ monitoring of ground deforma-

tions, in a very wide spectrum of operational

400 BenchmarkSPLB2 BenchmarkSEMAFORO BenchmarkELISUR BenchmarkCURVONE BenchmarkSCORVO Benchmark

gon

gon

Met

erAzimuth Angle Variations - Reference Benchmarks

Vertical Angle Variations - Reference Benchmarks

Slope Distances Variations - Reference Benchmarks

0.020

0.015

0.010

0.005

0.000

-0.005

-0.010

-0.015

-0.020

0.020

0.015

0.010

0.005

0.000

-0.005

-0.010

-0.015

-0.020

0.020

0.015

0.010

0.005

0.000

-0.005

-0.010

-0.015

-0.020

01-Aug-03 31-Aug-03 30-Sep-03 30-Oct-03 29-Nov-03 29-Dec-03 28-Jan-04

Fig. 14. Data collected by THEODOROS system, regarding the benchmarks of the reference system. Angles are measured in bgonQ(400 gon=3608).


SDF5 BenchmarkSDF6 BenchmarkSDF7 BenchmarkSDF8 Benchmark

gon

gon

Met

er

Azimuth Angle Variations - Benchmarks of Low S. d. F.

Vertical Angle Variations - Benchmarks of Low S. d. F.

Slope Distances Variations - Benchmarks of Low S. d. F.

0.025

0.020

0.015

0.010

0.005

0.000

-0.005

-0.010

-0.015

0.025

0.020

0.015

0.010

0.005

0.000

-0.005

-0.010

-0.015

0.025

0.020

0.015

0.010

0.005

0.000

-0.005

-0.010

-0.015


Fig. 15. Data collected by THEODOROS system, regarding the benchmarks of the low SdF.


SDF9 BenchmarkSDF10 BenchmarkSDF11 BenchmarkSDF12 BenchmarkSDF14 BenchmarkSDF16 Benchmark

gon

gon

Met

erAzimuth Angle Variations - Benchmarks of High S. d. F.

Vertical Angle Variations - Benchmarks of High S. d. F.

Slope Distances Variations - Benchmarks of High S. d. F.

0.045

0.040

0.035

0.030

0.025

0.020

0.015

0.010

0.005

0.060

0.050

0.040

0.030

0.020

0.010

-0.200

-0.400

-0.600

-0.800

-1.000

-1.200

-1.400


Fig. 16. Data collected by THEODOROS system, regarding the benchmarks of the high SdF.



conditions. The achievable accuracies in 3D position-

ing are suitable for scientific and early warning

purposes. The successfully applied strategy at Strom-

boli allowed the Civil Protection Authorities to

appropriately manage the volcanic crisis.

From the volcanological point of view, the geodetic

monitoring of the Stromboli volcano enables collect-

ing important data on the dynamic of the eruption,

both on the eruptive vents (episode of February 16th)

and craters (explosion of April 5th); this information

certainly represents a valued data set for future

researches to understand how the plumbing system

of this volcano works, mainly during eruptive events.

Furthermore, the monitoring system now provides

very useful data relevant to the post-eruptive dynamic

of the SdF; these positive results suggest this system is

a useful tool to identify movements for forecasting

large future landslides that could occur on the SdF.

Acknowledgments

The authors are indebted to G. Bertolaso, Director

of the National Department of Civil Protection (DPC),

and the DPC team working at the Advanced Operative

Centre at S. Vincenzo Observatory, led by B. de

Bernardinis, for their support in installing and

managing the complex monitoring system during the

eruption. Special thanks are also due to M. Zaia and

the other brave and generous Volcanological Guides,

the bGuardia di FinanzaQ staff and the helicopter pilots

and technicians who aided the difficult fieldwork

during the eruption.

The authors are grateful to E. Boschi, President of

INGV, who encouraged this work. We also acknow-

ledge M. Rosi for the thought-provoking discussions

and for his support in installing the EDM monitoring

system. Last but not least, sincere thanks are due to

P.G. Scalato and the many other INGV colleagues and

researchers of other institutions involved in the

Stromboli emergency for the stimulating professional

discussions and personal support during the tasking

days of the eruption.

References

Barberi, F., Rosi, M., Sodi, A., 1993. Volcanic hazard assessment at

Stromboli based on a review of historical data. Acta Vulcanol. 3,

173–187.

Bertagnini, A., Coltelli, M., Landi, P., Pompilio, M., Rosi, M., 1999.

Violent explosions yield new insights into dynamics of

Stromboli volcano. EOS Trans. Am. Geophys. Union 80,

633–636.

Bock, Y., Nokolaidis, R.M., de Jong, P.J., Bevis, M., 2000.

Instantaneous geodetic positioning at medium distances with

the Global Positioning System. J. Geophys. Res. 105 (B12),

28223–28253.

Bonaccorso, A., 1998. Evidence of a dyke-sheet intrusion at

Stromboli volcano inferred through continuous tilt. Geophys.

Res. Lett. 25 (22), 4225.

Chouet, B., Dawson, P., Ohminato, T., Martini, M., Saccorotti, G.,

Giudicepietro, F., De Luca, G., Milana, G., Scarpa, R., 2003.

Source mechanisms of explosions at Stromboli determined from

moment tensor inversion of very long period data. J. Geophys.

Res. 108, 2019.

Falsaperla, S., Spampinato, S., 2003. Seismic insights into explosive

paroxysm at Stromboli volcano, Italy. J. Volcanol. Geotherm.

Res. 125, 137.

Leica and Geodetics Inc., 2002. Software Crnet User Manual.

Mattia, M., Rossi, M., 2004. Sperimentazione di un sistema di

monitoraggio vulcanico per mezzo di tecniche GPS in tempo

reale. INGV CT internal report.

Teunissen, P.J.G., 1993. Least-Squares Estimation of the Integer

GPS Ambiguities. LGR Series, vol. 6. Delft Geodetic Comput-

ing Centre.

Tinti, S., Bortolucci, E., Romagnoli, C., 1999. Computer

simulations of tsunamis due to flank collapse at Stromboli,

Italy. J. Volcanol. Geotherm. Res. 96, 103–128.

Tinti, S., Pagnoni, G., Zaniboni, F., Armigliato, A., 2003.

Simulations of the 30 December 2002 Stromboli tsunami, Italy.

Geophys. Res. Abstr. 5, 10804 (EGS Assembly, 6–11 April

2003, Nizza).

New integrated geodetic monitoring system at Stromboli volcano (Italy)

Documents

Transcript of New integrated geodetic monitoring system at Stromboli volcano (Italy)