I. Hunstad,
1
A. Pepe, 2 S. Atzori, 1 C. Tolomei, 1 S. Salvi, 1 and R. Lanari, 2
1
Istituto Nazionale di Geofisica e Vulcanologia, Rome, Italy
2
Istituto per il Rilevamento Elettromagnetico dell’Ambiente (IREA), National Research Council of Italy
(CNR), via Diocleziano 328, Napoli, Italy
Accepted 2009 June 4 .
Received 2009 June 4; in original form 2008 October 29
Surface deformation in Central Italy
Hunstad Ingrid
Istituto Nazionale di Geofisica e Vulcanologia
Centro Nazionale Terremoti
Via di Vigna Murata, 605
00143 Rome Italy
Ph ++39 0651860423 fax ++39 0651860507 ingrid.hunstad@ingv.it
1

SUMMARY
We investigate the 1992-2000 surface deformation affecting the Abruzzi region, Central Italy, by exploiting two set of SAR data, acquired by the ERS-1/2 SAR sensors from ascending and descending orbits, respectively. We apply the Small BAseline Subset (SBAS) Differential Synthetic Aperture Radar
Interferometry (DInSAR) technique to the SAR data-set to estimate the displacement time series and the corresponding velocity maps.
Our DInSAR analysis allows us to detect a previously unreported major change in crustal deformation moving from the highest elevation of the Central Apennines to the Adriatic sector, that we infer as evidence of a complex tectonic boundary between two Adriatic microplates involving a wide interaction zone. An extension effect, at a rate of about 0.10·10-6 yr -1 , is also confirmed by our investigation. Our findings may have important implications in the seismic hazard assessment within the Adriatic region in
Central Italy.
Keywords: Satellite geodesy; Radar interferometry; Plate motions; Continental tectonics: extensional.
INTRODUCTION
The Italian peninsula is part of the broad deforming zone between Eurasia and Africa (M c Kenzie , 1972).
Within the overall convergence of this zone, Anderson & Jackson (1987) proposed the existence of an
Adriatic block as an independent microplate (Fig. 1), rotating counterclockwise with respect to Eurasia around a pole located at 45.8ºN/10.2ºE. Slightly different locations of the rotation pole have been proposed and debated ( Ward , 1994), as well as the location of microplate southern boundary (Calais et al.,
2002; Serpelloni et al., 2005; D’Agostino et al.
2008) .
Earthquake focal mechanisms and GPS measurements have contributed to define this relatively simple model that explains well the first order deformation patterns, and still represents a common reference model for tectonic interpretation. In this view, the shortening observed in the eastern part of the Southern Alps, in the Dinaric belt and Albania, the active NE-SW extension along the Apennines, and the different rates of deformation in the Northern and
2

Southern Apennines are kinematically in agreement with Anderson and Jackson’s observations. However, this model could be in contrast with compressive focal mechanisms computed for the Northern Apennines and, further north, with CGPS measured shortening (Serpelloni et al., 2005), although the signal-to-noise ratio for these observations is low and the debate is still open. CGPS sites east of the Central Apennines are still too sparse to contribute in defining areas of active extension and shortening. Major open questions are: i) whether the observed shortening is limited to the Northern Apennines; ii) is such a shortening process a first order tectonic feature? If so, this implies that the simple kinematic view introduced by
Anderson & Jackson (1987) does not fully explain the available observations. Accessibility of this plate boundary zone makes the Abruzzi region, central Italy, a suitable area for geodetic investigations on how the present day crustal deformation is distributed.
In this work, we analyze a Synthetic Aperture Radar (SAR) data-set, relevant to the Abruzzi region, acquired by ERS-1/2 satellites during the 1992-2000 time period. A multi-temporal Differential
Interferometric SAR (DInSAR) processing is carried out on these data in order to detect and measure crustal deformation in the Abruzzi region (Fig. 1), which extends from the Central Apennines toward the
Adriatic coast, i.e., at the boundary of the Adriatic block. The reference point for the DInSAR results is located in correspondence to the AQUI station (Fig. 1) of the Italian Space Agency CGPS network.
We discuss the implication of our results on the kinematics of the Adriatic region.
DInSAR DATA PROCESSING
Our analysis is based on applying the SBAS-DInSAR algorithm (Berardino et al., 2002) to a SAR data-set composed of 54 ERS-1/2 SAR data acquired from descending (track 308, frame 2754) and 42 from ascending (track 129, frame 843) orbits from 1992 to 2000 (see auxiliary material). The approach we considered relies on a proper combination of multilook interferograms, selected by imposing small temporal (< 1500 days) and spatial (< 200 m) baseline separations between the orbits in order to limit the noise effects affecting the interferograms. We compute, by using the 3-arcsec SRTM DEM
( http://www2.jpl.nasa.gov/srtm/ ), 165 and 128 differential interferograms (with pixel size of ~ 80 x 80 m)
3

from descending and ascending data, respectively, which are subsequently inverted by applying the SBAS method. Accordingly, the radar Line of Sight (LOS) time series and the associated velocity maps are retrieved, with typical accuracies of ~ 5 mm and 1 mm yr -1 , respectively (Casu et al., 2006).
We remark that high relief and temperate-humid climatic conditions of the Central Apennines induce significant atmospheric phase artifacts, which are minimized by the atmospheric filtering operation described in Berardino et al.
(2002). The original LOS velocity maps, retaining only pixels with temporal coherence (Pepe et al.
, 2006) values greater than 0.7 ( a typical threshold for SBAS-DInSAR analysis), are also subsequently low-pass filtered to further reduce the noise, ending up with a pixel size of ~ 400 m
(Fig. 2a, b).
From the Ascending and Descending data, the East and the Vertical deformation components are estimated by inverting, for each pixel, the following linear system
Ascending(
Descending i,
(i, j) j)


East(i,
East(i, j)
 j) coeff_E
A
 coeff_E
D
 
 

North(i,

North(i, j)
 j) coeff_N
A

  coeff_N
D i ,

 
Up(i,
 j)
Up(i,
 j) coeff_U
A

  coeff_U
D
  (1) wherein ( i,j ) identifies the discrete pixel coordinates, East(.), North(.) and Up(.) are the unknown components of the velocity field, Ascending(.) and Descending(.) are values retrieved through the SBAS-
DInSAR algorithm and coeff_E
A/D
, (.), coeff_N
A/D
(.), coeff_U
A/D
(.) are the coefficients (subscripts A and D designate Ascending and Descending ) defining the two LOS geometries ( Price & Sandwell , 1998) computed from precise satellite orbits (Parsons et al.
2006).
To solve the system (1) (three unknowns, two equations), it is worth observing that SAR measurements have little sensitivity to the north-south component of the displacement, with average absolute values for coeff_U , coeff_E and coeff_N being 0.93, 0.38, 0.08, respectively. Thus, even if assuming for the north velocity (relative to the reference point AQUI) a value of 1 cm yr -1 , exceeding what is reported in the available literature (Hunstad et al.
2003; D’Agostino et al.
2008), this contribution to the LOS velocity
4

would be significantly lower than the data uncertainty (1 mm yr -1 ). Therefore, we legitimately solve the system (1) by neglecting the north component.
RESULTS AND DISCUSSION
Fig. 2 shows the retrieved LOS mean velocity field for the ascending and descending orbits expressed in millimeters per year. To better display the tectonic meaning of our data we exploit, as discussed above, the ascending and descending velocity maps to estimate, through (1) the East and Up components reported in
Fig. 2c and 2d, respectively. We also show, along the profiles A-A' and B-B' highlighted in Fig 2c and 2d, the Up (see Fig. 2e and 2f) and East (see Fig. 2g and 2h) components of the velocity, interpolated with an inverse distance weighting with a radius of 2 km. For selected sites, we present the DInSAR time-series relevant to both the ascending and descending data-sets; these plots (Fig. 3) clearly show the rather continuous accumulation of the revealed displacements. By comparing Fig. 3b and 3d, it is evident that the revealed trend is still visible even in the absence of any atmospheric filtering operation.
A least squares regression of the Up velocity essentially yields a null slope on both profiles (Fig. 2e and
2f). Then, the prominent feature is a negligible vertical deformation rate over the Northern Abruzzi.
The analysis of geodetic levelling data led D’Anastasio et al. (2006) to the conclusion that extensional tectonics does not show an important signature on vertical motion in the Central Apennines except for the positive gradient south of the permanent AQUI GPS station, not yet explained and that we do not retrieve.
Estimated regional long-term uplift rates in Central Apennines is about 1 mm·yr -1 during the last 1 Ma
(D’Anastasio et al., 2006). Such a low rate is within the uncertainties of our DInSAR analysis and hence not clearly detectable. The evidence of negligible vertical displacement in the area is clearly shown in Fig.
3a and 3b having nearly opposite ascending and descending trends; this result also implies the presence of displacements in the east direction (Manzo et al., 2006). This is in contrast to the quasi-vertical displacement revealed in correspondence to the subsidence area within Pescara town, for which the retrieved displacement time series have nearly the same trends (Fig. 3c).
5

Least squares regressions of the East component of velocity on profiles A-A’ and B-B’ (Fig. 2g and 2h) represent the ∂v
1
∕∂x
1
component of the velocity gradient tensor L ij,
, where v i
and x i
are the velocity and position vectors measured in a reference frame where x
1
points eastward and x
2
points northward. The horizontal strain rate tensor
  ij is obtained by extracting the symmetric portion of the velocity gradient tensor. We retrieve the L
11
component from DInSAR analysis and,
 
11 being equal to L
11
, we compare our results with the estimate of
  known from previous studies. The least squares regression of the East
11 component on the westernmost side of the profile B-B’ yields a positive L
11
equal to 0.12 10 -6 yr -1 , concentrated along the axis of the Apennines in a belt 40 km wide (Fig. 2h), corresponding to the location of large earthquakes and elevated topography. This value is consistent with previous estimates of extension rate in the same area obtained through the analysis of both triangulation and GPS data (Hunstad
& England, 1999; Hunstad et al.
, 2003) and GPS data alone (Serpelloni et al., 2005; D’Agostino et al.,
2001).
Moreover, at about 60 km from the western end of the B-B’ profile, the velocity returns to zero (Fig 2h, suggesting that the increase in the East velocity towards the east, which means a positive L
11
, is concentrated on the mountain belt. A change in trend to a negative value of L
11
, moving from west to east, is also visible on the profile A-A’ of Fig. 2g.
The decrease of the east ground velocity from the eastern side of the belt to zero toward the Adriatic coast
(Fig. 2g and 2h) amounts to about 5 mm year -1 and is concentrated in a few tens of kilometers east to
Central Apennines. In terms of strain field, this finding means that the tectonic deformation of the Adriatic foreland is different from the Apennines, which are characterized by an extensional strain in the NE-SW direction, corresponding to a positive value of L
11
. This newly detected deformation, not revealed before due to the absence of dense networks in the zone, is distributed in an area that shows no remarkable seismic deformation (Selvaggi, 1998). The instrumental seismiticy map (Castello et al., 2007) shows that earthquakes never exceeded M = 4. Except for the two moderate magnitude historical earthquakes, that
6

occurred in this area (Fig. 1) in 1881 (M e
= 5.6) and 1882 (M e
= 5.4), the historical Seismic Catalog
(Stucchi et al .
, 2007) does not show any large earthquake in the past thousand years. Accordingly, further studies, based on future GPS measurements, are needed to define all the components of the strain rate tensor, thus allowing for resolving between shortening and extensional tectonics. In any case, despite the lack of available GPS data, the two stations from the national CGPS network (Selvaggi et al ., 2006),
INGP and RSTO, confirm our observation. Indeed, the computed ITRF05 east component of the velocity for these stations is 23.71±0.18 mm·yr -1 and 22.97±0.16 mm·yr -1 , respectively, showing a negligible strain in the east direction between INGP (close to DInSAR reference point location) and the Adriatic coast.
CONCLUSIONS
We have shown that it is possible to resolve the interseismic plate boundary deformation, through the
SBAS-DInSAR algorithm, in a region of low deformation rate. The known NE-SW extension across the
Apennines is consistent with our estimated value of L
11
. Moreover, our results show that this extension is localized on the mountain belt and that the Adriatic region is deforming where a rigid behaviour of the
Adria microplate is expected, at least from Anderson & Jackson’s (1987) model. This deformation is clearly different from what we see across the Apennines suggesting a major change moving from west to east. These observations are consistent with the recent paper of D’Agostino et al. (2008) that places the southern boundary of the Adria plate near the Gargano promontory, a few tens of km from the Abruzzi region. They interpret GPS measurements as evidence of the existence of two microplates in the Adriatic region. The northern one, Adria (Ad), extends from the Po Valley to the Central Adriatic (Gargano
Promontory); the southern newly defined one, Apulia, includes the Apulian Promontory, the Ionian Sea and the Hyblean Plateau. They suggest that the relative block motion between Adria and Apulia controls the deformation of the Central Adriatic, though a detailed description of deformation distribution is not presented because of the lack of GPS sites north of the Gargano promontory. We interpret the newly detected strain accumulation across the Adriatic foreland as the evidence that the boundary between Adria and Apulia is distributed over a wide zone rather than being concentrated on a single tectonic line.
7

Moreover, since the stored strain is a first order tectonic feature, it sets a new constraint on the simple kinematic model proposed by Anderson and Jackson for the Adriatic region.
ACKNOWLEDGEMENTS
We thank the European Space Agency who provided SAR ERS-1/2 data within Category-1 Project no.
3769 . This work was supported by the Italian FIRB program and the European Commission, Provision
3.15, Project of Regional Center of Competence ''Analysis and Monitoring of the Environmental Risk''
(CRdC-AMRA). The DEM was obtained from NASA SRTM archives. Precise ERS 1/2 satellite orbit state vectors are courtesy of Technical University of Delft, The Netherlands. We also thank Dr. G. Milano for his helpful remarks.
REFERENCES
Anderson, H. & J. Jackson, A., 1987. Active tectonics of the Adriatic region, Geophys. J. R. Astron. Soc.
,
16 , 504-522.
Berardino, P., Fornaro, G., Lanari, R. & Sansosti, E., 2002. A new algorithm for surface deformation monitoring based on small baseline differential SAR interferograms, IEEE Trans. Geosci. Remote Sens.
,
40 , 2375–2383.
Casu, F., Manzo, M. & Lanari, R., 2006. A quantitative assessment of the SBAS algorithm performance for surface deformation retrieval from DInSAR data, Remote Sens. Environ., 102 , no. 3/4, 195–210.
Calais, E., Nocquet, J.M., Jouanne, F. & Tardy, M., 2002. Current strain regime in the Western Alps from continuous Global Positioning System measurements, 1996-2001, Geology , 7 , 651- 654.
Castello, B., Olivieri, M. & Selvaggi, G., 2007. Local and duration magnitude determination for the Italian
Earthquake Catalog, 1981-2002, Bull. Seism. Soc. Am., 97 (1B), 128-139, doi 10.1785/0120050258.
8

D’Agostino, N., Giuliani, R., Mattone, M. & Bonci, L., 2001. Active crustal extension in the Central
Apennines (Italy) inferred from GPS measurements in the interval 1994 – 1999, Geophys. Res. Lett ., 28 ,
2121–2124.
D’Agostino, N., Avallone, A., Cheloni, D., D’Anastasio, E., Mantenuto, S. & Selvaggi, G., 2008.Active tectonics of the Adriatic region from GPS and earthquake slip vectors, J. Geophys. Res.
, 113 , B12413, doi:10.1029/2008JB005860.
D'Anastasio, E., De Martini, P.M., Selvaggi, G., Pantosti, D., Marchioni, A. & Maseroli, R., 2006. Shortterm vertical velocity field in the Apennines (Italy) revealed by geodetic levelling data, Tectonophysics ,
418 , 219–234.
Hunstad, I., & England, P., 1999. An upper bound on the rate of strain in the CentralApennines, Italy, from triangulation measurements between 1869 and 1963, Earth Planet. Sci. Lett., 169 , 261-267.
Hunstad, I., Selvaggi, G., D’Agostino, N., England, P., Clarke, P. & Pierozzi, M., 2003. Geodetic strain in peninsular Italy between 1875 and 2001, Geophys. Res. Lett ., 30 (4), 1181, doi:10.1029/2002GL016447.
M c Kenzie, D., 1972. Active tectonics of the Mediterranean region, Geophys. J. R. Astron. Soc.
, 30 , 109-
185.
Manzo, M., Ricciardi, G. P., Casu, F., Ventura, G., Zeni, G., Borgström, S.,Berardino, P., Del Gaudio, C.,
Lanari, R., 2006. Surface deformation analysis in the Ischia island (Italy) based on spaceborne radar interferometry, Journal of Volcanology and Geothermal Research , 151 , pp 399-416, doi:
10.1016/j.jvolgeores.2005.09.010.
Parsons, B., Wright, T. J., Rowe, P., Andrews, J., Jackson, J., Walker, R., Khatib, M., Talebian, M.,
Bergman, E., & Engdahl, E. R., 2006. The 1994 Sefidabeh (eastern Iran) earthquakes revisited: new evidence from satellite radar interferometry and carbonate dating about the growth of an active fold above a blind thrust fault, Geophys. J. Int ., 164 , 202-217, doi:10.1111/j.1365-246X.2005.02655.x.
Pepe, A. & Lanari, R., 2006. On the extension of the minimum cost flow algorithm for phase unwrapping of multitemporal differential SAR interferograms, IEEE Trans. Geosci. Remote Sens.
, 44 , no. 9, 2374–
2383.
9

Price, E. J. & Sandwell, D. T., 1998. Small-scale deformations associated with the 1992 Landers,
California, earthquake mapped by synthetic aperture radar interferometry phase gradients, J. Geophys.
Res.
, 103 , B27001-27016.
Selvaggi, G., 1998. Spatial distribution of horizontal seismic strain in the Apennines from historical earthquakes, Annali di Geofisica, 41 , 241–251.
Selvaggi, G. et al., 2006.
La “Rete Integrata Nazionale GPS” (RING) dell’INGV: una infrastruttura aperta per la ricerca scientifica, Atti 10a Conferenza Nazionale ASITA , 1749-1754 (in
Italian)
Serpelloni, E., Anzidei, M., Baldi, P., Casula, G. & Galvani, G., 2005. Crustal velocity and strain-rate fields in Italy and surrounding regions: new results from the analysis of permanent and non-permanent
GPS networks, Geophys. J. Int.
, 161 , 861-880.
Stucchi, M., et al., 2007. DBMI04, il database delle osservazioni macrosismiche dei terremoti italiani utilizzate per la compilazione del catalogo parametrico CPTI04. http://emidius.mi.ingv.it/DBMI04/ Quaderni di Geofisica, 49 , pp.38 (in Italian).
Ward, S.N., 1994. Constraints on the seismotectonics of the Central Mediterranean from Very Long
Baselines Interferometry, Geophys. J. Int.
, 117 , 441-452.
FIGURE LEGENDS
Figure 1
Geographic map of Abruzzi region. Circles correspond to instrumental earthquakes from Catalogue of
Instrumental Seismicity (Castello et al.
, 2007); black stars highlight historical earthquakes from CFTI04 catalogue (Stucchi et al.
, 2007). Triangles represent CGPS stations. PV, Po Valley; Ad, Adria microplate;
Ap, Apulian Promontory; GP Gargano Promontory; Hy, Hyblean Plateau; Io, Ionian Sea. Gray region in
10

overview map delineates Anderson & Jackson’s (1987) Adria microplate; dashed line represents boundaries of the two microplates, Adria (Ad) and Apulia (Ap+Hy+Io) (D’Agostino et al.
2008).
Figure 2
Mean LOS velocity field for ascending (a) and descending (b) orbits; black arrows show satellite track. (c) and (d) show estimated East and Up components of velocity field, respectively. Lines A-A´ and B-B´ correspond to profiles shown in the following panels from (e) to (h). (c) Triangles represent CGPS stations, squares 1 and 2 indicate pixels relevant to Fig. 3a and 3b respectively. (d) Square 3 indicates pixel relevant to Fig. 3c, L’Aquila and Pescara towns are shown. (e)-(f) Up velocity along profiles A-A´ and B-B´. Gray lines represent least square regression of observed data. (g)-(h) same as for (e)-(f) but for
East velocity.
Figure 3
(a), (b), (c) Comparison between DInSAR LOS deformation time series relevant to ascending data (red triangles) and descending ones (blue triangles) for pixels labeled 1, 2 and 3 in Fig. 2c and 2d. (d) Same as for (b) but with time series computed without any atmospheric filtering operation.
11

Figure 1
12

Figure 2
13

Figure 3
14