Cello et al. (1997)
They apply remote sensing, aerial photo interpretation, and field survey 
methodologies to characterise the tectonic setting of the central Apennine; they 
identify a fault system, namely the Central Apennine Fault System (CAFS), that 
is comprised of major N-S trending, left-lateral strike-slip faults, which cut 
the entire crust, and several associated minor dip-slip normal faults, which are 
interpreted as being the evidence of a negative flower structure. The Colfiorito 
basin is bounded to the NE by a set of these latter faults. Recent activity of 
the N-S major faults is testified by lateral offset of stream incisions, while 
that of the minor faults is testified by displacement of Late Pleistocene 
deposits.

Amato et al. (1998)
They analyse the distribution of earthquake of the Umbria-Marche seismic 
sequence between September 4 and November 13, 1997. Using data from permanent 
and temporary networks they locate 2550 events among which 5 with Ml>5, and more 
than 20 with Ml>4. On this basis they conclude that the three largest shocks 
have ruptured distinct, adjacent, sub-parallel normal fault segments with a 
transversal offset between one to another. They also notice that the extent of 
the aftershock zone is longer than the cumulative length of the mainshock fault-
planes and that the dip of the aftershock distribution is in accordance with the 
dip of the mainshock fault-planes. In a more general perspective they relate the 
Umbria-Marche earthquake sequence to the active process of extension in the 
region and prompt the hypothesis that the faults ruptured along pre-existing 
thrust planes.

Basili et al. (1998)
These investigators carry out a geological field survey throughout the 
epicentral area of the 1997-98 seismic sequence. They made a distinction between 
the coseismic surface breaks visible in the northern area (Colfiorito and Costa-
Cesi basins, epicentral area of the two first mainshocks) that incidentally 
occurred in coincidence of pre-existing bedrock fault-scarps and those occurring 
in the southern area (Sellano, epicentral area of the third mainshock). A 
summary of the observations presented by the authors follows. The "ruptures" 
exposed in the northern area only occurred on very steep slopes with movement 
components always in accordance with maximum slope direction. The displacements 
appeared to be the result of scree or soil sliding along planes which mostly 
coincided with pre-existing bedrock fault planes. The ruptures never affected 
the nearby carbonate bedrock where bedrock fault planes were still visible. The 
total length of displacement was a very small portion (3%) of the total length 
of the exposed faults. On this basis, the authors interpret these displacements 
as the direct result of gravitational phenomena triggered by seismic shaking. 
Therefore, they claim that it is not possible to infer the source seismic 
behaviour on their basis. In contrast, the surface ruptures occurred the 
southern area did not seem to be influenced by either morphological or 
lithological factors. Indeed, they were mapped straight across successive 
valleys and hills and cutting both the carbonate bedrock and the unconsolidated 
debris covers. These surface breaks were thus interpreted as being coseismic 
tectonic features.

Cello et al. (1998)
They carry out a geological field survey in the epicentral area of the two first 
mainshocks of the 1997-98 seismic sequence. The ruptures found at Costa were 
correlated to the 26/09/97; 00:33 main shock. All other ruptures were not 
discernible between the two events (00:33 and 9:40) of the same day. The area of 
the 14/10/97 event is not analysed by these authors. They mapped displacements 
observed along pre-existing fault-scarps and calculated the total length of 
surface ruptures by summing the length of the observed displacements, that of 
minor surface breaks and fissures, and that of the trace in between. The three 
faults Costa-Corgneto, Tolagna and Colfiorito being respectively 1, 1.8 and 8 km 
long. All ruptures were interpreted as surface faulting on the basis of the 
constant offset along tens to hundreds metres in contrast to the variable 
thickness (centimetres to tens of metres) of the intervening slope deposits; the 
different geomorphic features of the slopes; the drained and stable conditions 
of deposits. Authors question the usefulness of discriminating components of 
gravity and tectonics in similar cases and suggest that their results can be 
used to characterise the local seismic source.

Ekstrom et al. (1998)
Using long-period seismograms from the Mediterranean seismographic network and 
additional data from the global seismographic network, they analyse focal 
mechanisms and moment release of the larger earthquakes (14 events) occurred 
during the Umbria-Marche seismic sequence of September-October 1997. Thirteen 
focal mechanisms indicated normal faulting on NW-SE striking fault planes with 
tension axes oriented in the range 40-60. Only one focal mechanism indicated a 
strike-slip mechanism indicating either right-lateral faulting on an E-W 
striking plane or left-lateral faulting on a N-S striking plane. The authors 
suggest the latter to be in agreement to the geological structures reported by 
Cello et al. (1997). The moment release of the main event (26/09/97; 9:40 UT) is 
1.20 x 10EXP18 Nm, that is nearly equal to the sum of the rest of the sequence 
(1.12 x 10EXP18 Nm). They also suggest that the distribution of the moment 
release in several events of moderate size rather than the concentration of it 
in a single largeevent may be related to the structural complexity of the 
region.

Cinti et al. (1999)
These workers report the results of field surveys performed throughout the 
epicentral area of the 1997-98 seismic sequence. The recognised and measured 
coseismic deformation features are comprised of fractures affecting paved and 
unpaved roads, buildings, soils and hard rocks and include displacements at the 
base of pre-existing bedrock fault-planes. On the basis of a statistical 
treatment, the data-set enabled the identification of four narrow bands of 
deformations which were then compared to the location of rupture planes inferred 
from the available seismological observations. Although most ruptures showed a 
NW-orientation and seem to be in agreement with the stress field inferred from 
CMT solutions, the authors conclude that the observed surface deformation was 
not the direct result of primary rupture of fault at depth associated with the 
three mainshocks. This interpretation was mainly driven by the assumption that 
primary surface faulting would have required larger rupture planes and that no 
fractures were detected where rupture planes project. The surface deformation 
was explained with the presence of high-angle normal faults which accommodate 
the deformation induced by deep displacement on the seismogenic structure and 
that are not necessarily connected with it. In addition, the authors suggest 
that the rupture planes of the first two mainshocks coincide to thrust planes, 
revealed by previous studies of seismic-reflection profiles.

Hunstad et al. (1999)
They propose a dislocation model (elastic, homogeneous, isotropic half-space) 
for the first two mainshocks of the Umbria-Marche seismic sequence of 1997 based 
on coseismic vertical and horizontal displacements measured by GPS at 5 
monuments of the Istituto Geografico Militare Italiano. The forward modelling 
procedure enable the authors to constrain the top of the fault associated to the 
9:40 event to be 0.6 km (imposing a non-uniform slip with a maximum of 65 cm in 
the first third of the fault) while the top of the fault associated to the 00:33 
event to be deeper than 1.7 km (imposing a uniform slip of 27 cm).

Meghraoui et al. (1999)
This paper utilises the 1997-98 earthquake sequence to introduce the concept of 
the fault-fragment that is described as: "the surface fault of the smallest 
earthquake for which the whole seismogenic layer is ruptured". The three 
mainshocks of the sequence are correlated to three distinct faults that are 
regarded to be fault fragments. Such fault-fragment is thus being interpreted to 
be the source for moderate earthquakes as opposed to the fault-segment which is 
the source for large ones.

Stramondo et al. (1999)
By generating and analysing an ERS-SAR differential interferogram they detect 9 
fringes in the area of the Colfiorito basin which correspond 25 cm of coseismic 
surface displacement. In addition, GPS measurements has shown 14+/-1.8 cm of 
maximum horizontal displacement and 24+/-3 cm of maximum subsidence. Numerical 
modelling of these data suggests that the fault rupture for the 26/09/97; 09:40 
event stopped very close to the surface but did not produce surface faulting.

Olivieri and Ekstrom (1999)
They use broadband teleseismic data to analyse the three mainshocks of the 1997-
98 sequence in Umbria-Marche. Their results show that the sources are quite 
shallow, having depths of the centre of slip to be in the range 1.0 to 4.0 km 
for the first two mainshocks and 1.0 to 5.0 km for the third. The authors 
suggest that the rupture of the second shock has an upward directivity with a 
horizontal component oriented toward the north and that the third shock has also 
an upward directivity. However they exclude a major southward horizontal 
component for this last event in contrast to what is suggested by felt reports. 
They also analyse the 26/03/98 Mw 5.2 earthquake located in the mantle. Depth is 
45 km and focal mechanism is oblique-normal with a NE-SW tension axis. In  
addition, the authors re-examine the available digital recordings for the 
19/09/1979 Norcia (30 km to the south of Colfiorito) earthquake which shows a 
depth range of moment release between 1.0 and 5.0 km (see source 15 for details 
on the Norcia faults).

Pino et al. (1999)
These authors derive the relative moment rate function for the three mainshocks 
of the 1997 sequence using broadband (MedNet) data. They obtain source time 
functions indicating south-eastward rupture propagation for the two smallest 
mainshocks and north-eastward rupture propagation for the largest one. They also 
determine for the three faults length (10.0 km; 14.0 km; 10 km), rupture 
propagation velocity (2.0 km/s; 2.6 km/s; 2.6 km/s) and direction (nearly 
horizontal at 152; slightly down-dip at 304; slightly up-dip at 107), and 
slip distribution (0.52 m at mid-length; two picks of 0.65 m and 0.32 m; 0.20 m 
at mid-length).

Zollo et al. (1999)
They model the strong motion records for the two first mainshocks of the 1997 
seismic sequence. Results show two normal fault with a minor dextral component, 
same strike at about 150, and same dip about 37. The first event rupture is 6 
km long, propagated almost up-dip and bilaterally at a velocity of 3 km/s, and 
has a maximum slip of 38 cm. The second is 12 km long and propagated up-dip and 
unilaterally to NW at about 2.6 km/s, and has a maximum slip of 55 cm. However, 
they have like average slip. The two fault planes are laterally offset by 5-7 
km. This offset is interpreted to be a barrier which controlled the local stress 
release between the two shocks.

Amato and Cocco (eds) (2000)
Amato and Cocco edited a compilation of papers on the 1997 Umbria-Marche 
earthquakes of 1997-1998. The special volume of J. Seismol. contains several 
geological, seismological and geodetic contributions that add detail to papers 
that were already published in the previous two years.

Barba and Basili (2000)
They analyse geological and seismological data about the 1997-98 earthquakes. 
They located about 1000 events using a local velocity model based on geophysical 
exploration data. On the basis of earthquake locations, fault plane solutions 
and geological mapping they point out that the two mainshocks of 26/09/97 and 
the minor events of 03/09/97, 22:07 UT and 26/09/97, 09:47 UT originated on the 
same structure at depth. Taking into accounts the possible source of 
uncertainty, the two main ruptures have to be blind with the top edge confined 
in a depth range between 0.5-4 km. Therefore the offsets occurred along pre-
existing fault scarps (see Basili et al., 1998) were interpreted as 
gravitational sliding triggered by the seismic shaking. The third mainshock 
(14/10/97) originated on another fault at a much shallower depth. In this case 
the fault top edge may be in the range of 0-2 km depth. Therefore, in spite of 
its lower magnitude, this earthquake could have produced tectonic ruptures at 
the ground surface. Based on several observations, among which the strong-motion 
data from Servizio Sismico Nazionale, these authors also suggest that the 
14/10/97 fault rupture may had propagated bilaterally. On the basis of 
paleostress reconstruction they suggest that the strain field in this region may 
be coaxial over a long time interval (ca. 1Ma).

Boncio and Lavecchia (2000)
These authors combine geological, geophysical and seismological data to derive a 
structural model for earthquake faulting in the Umbria-Marche Apennines. A low-
angle east-dipping normal fault, namely Altotiberina Fault (AF), controls 
distribution and character of seismicity in this region. Earthquakes originate 
on west-dipping normal faults detaching on the AF. The eastward deepening of the 
detachment seems to control the depth of earthquake foci as appears from other 
recent events in the region, e.g. 1984 Gubbio earthquake (M 5.2, depth 6-7 km) 
and 1979 Norcia (M 5.9, depth 11-12 km).

Calamita et al. (2000)
They collect several source of information, such as geologic, geomorphic and 
seismologic data and put them together to develop a seismotectonic model for the 
Umbria-Marche Apennines. In this model they interpret the 1997 earthquake 
causative faults as inverse reactivation at depth of pre-existing thrust faults. 
Thereby the low-angle thrust fault control the seismogenic zoning in the area, 
while the high-angle normal faults are linked to them and deform the uppermost 
layer of the crust.

Vittori et al. (2000)
They re-examine the effects at the ground surface related with the two 
mainshocks of 26/09/97 in the Colfiorito area. In addition to a description of 
secondary effects, such as landslides, hydrological variations and gas 
emissions, the authors propose a mechanical explanation to support the idea that 
the offsets observed in the Colfiorito and Costa-Cesi basins (see Basili et al., 
1998; Cello et al., 1998; Cinti et al., 1999) represent primary surface faulting 
related to the two 26/09/97 mainshocks. Also, they suggest that the offsets 
observed in the Renaro-Mevale zone, after the 14/10/97 mainshock, represent a 
"sackung-like movement".

