Content uploaded by Massimo Rossi
Author content
All content in this area was uploaded by Massimo Rossi on Jun 21, 2015
Content may be subject to copyright.
Journal of Mediterranean Earth Sciences Special Issue (2013), 103-109
Multi-scale characterization of the Pleistocene-Holocene Tiber delta deposits as
a depositional analogue for hydrocarbon reservoirs
Mattia Marini1, Salvatore Milli1, Massimo Rossi2, Vittorio De Tomasi2, Marco Meda2, Nicola Lisi2
1Dipartimento di Scienze della Terra, SAPIENZA Università di Roma, Roma, Italy
2Eni E&P Division, San Donato Milanese, Milano, Italy
INTRODUCTION
The study of well exposed or shallow buried
depositional analogues of clastic coastal systems such as
estuaries and deltas is very attractive as it is cost-effective
and virtually provides any scale of detail. In the field,
stratigraphic-sedimentologic sections can be measured
and outcrop face surveyed using laser scanning
techniques that allow to reveal depositional geometries
and spatial distribution of facies with a resolution down to
few centimetres. Where these analogues are in the shallow
subsurface of urban or development areas, the availability
of a large number of well-spaced, public domain boreholes
and geophysical surveys, can allow a nearly three-
dimensional reconstruction of the depositional and
stratigraphic architecture at scales down to few tens of
metres. From these 3D volumes, significative cross-
sections can be obtained which permit to visualize
lithofacies distribution vs. depositional geometries and
understand large-scale permeability patterns.
Additionally, when the petrophysical properties are
available as input, two-dimensional seismic modeling can
be undertaken to obtain simulated seismic sections to be
used in order to image the most likely seismic features of
the internal reservoir geometry.
The paralic successions deposited under the influence
and supply of Tiber river during the Pleistocene and
Holocene, crop out extensively in the southwestern sector
of Rome thanks to the presence of many quarries or are
buried in the subsurface of the present-day Tiber delta
plain; they represent a suitable depositional analogue for
undertaking such an approach. The aim of the study
detailed here was to achieve a multi-scale characterization
of selected part of these paralic successions which may
serve as depositional analogues for hydrocarbon
reservoirs.
To this purpose, we conducted: i) the integration of
unpublished data from Milli (2013) with novel outcrop
and borehole data with the aim of better defining the
depositional architecture of study objects; ii) the two-
dimensional seismic modeling of selected cross sections
using as input seismic velocities deduced integrating
geophysical, geotechnical and petrophysical data from the
subsurface of the Tiber delta and the literature; iii) the
modeling of geometries and facies architecture of a gravel
beach conditioned to 3D photo-textured models obtained
through LIDAR and ortho photogrammetry techniques.
GEOLOGIC AND STRATIGRAPHIC OVERVIEW
The Pleistocene-Holocene deposits forming the subject
of this research represent part of the sedimentary fill of
the Roman Basin, which is located along the eastern
margin of the Tyrrhenian Basin, the youngest back-arc
basin of the Mediterranean area forming since the late
Miocene in response to back-arc extension in the
Apennines foreland basin system.
Along the Latium Tyrrhenian margin, the extensional
tectonics gave rise to NNW-SSE/NW-SE half-graben
basins, which were filled with syn- and post-rift clastic
sediments during the Plio-Pleistocene. Among these
extensional basins, the Roman Basin straddles the
present-day Tiber delta and extends for about 135 km in
a N-S direction. Its development started in the Late
Pliocene and was accompanied by tectonic uplift and
intense volcanic activity reaching their climax in the
Middle-Upper Pleistocene, when the Roman Magmatic
Province started up (Fig. 1).
Therefore, the stratigraphy of the Roman Basin resulted
from the close interaction of tectonic uplift, volcanic
activity and glacio-eustatic sea-level fluctuations driven
by the Quaternary climate changes (Milli, 1997). The
architecture of the basin is characterized by several units
forming high-frequency depositional sequences with
average duration spanning from 30,000 yr to 120,000 yr
(4th-order), stacked to form two composite 3rd-order
sequences, namely the Monte Mario Sequence (MMS;
Lower
Pleistocene) and the Ponte Galeria Sequence (PGS
hereafter; Late Lower Pleistocene-Holocene) Milli et al.,
2011, 2013). In the study area, the MMS is very poorly
exposed and its knowledge is mostly based on data from
relatively deep wells. Its deposits are represented by coastal
and transition-shelf depositional systems of the Late
Lowstand and Transgressive Systems Tracts. Oppositely,
the PGS is particularly well-exposed and shows a number
of adjacent depositional settings. It contains fluvial, fluvio-
lacustrine, barrier island-lagoon, and transition-shelf
J M E S
Journal of Mediterranean Earth Sciences
MARINI_ARGENTI 30/07/13 14:08 Pagina 103
depositional systems organised to constitute the lowstand
(LST), transgressive (TST) and highstand (HST) systems
tracts. PGS sediments are also interfingered with the
volcaniclastics deposits of the Albani and Sabatini
volcanic complexes.
THE PONTE GALERIA SEQUENCE (PGS)
In the studied area, the PGS is 10 to 110 m thick and lies
above the shelfal mud sediments of the MMS through a
polygenic erosional surface which is related to the sea-
level fall occurred between the MIS 31 and 27. It
represents a composite sequence consisting of twelve 4th-
order sequences (with thickness ranging from 5 to 80 m)
whose boundaries are sharp erosional surfaces recording
basin and downward facies shifts, subaerial exposure and
paleosols in the interfluves. From the sequence
stratigraphic point of view the sequences from PG01 to
lower PG1 stack to form the late LST, which developed
during a period of stillstand and slow relative sea-level rise
that produced a series of prograding and aggrading
wedges-shaped units. Sequences from PG1 to lower PG8
are interpreted as part of the TST, while the PG9 sequence,
namely the Tiber Depositional Sequence (TDS hereafter)
developed entirely during the HST.
The architecture of the PGS is characterized by a seaward
stacking pattern of its component 4th-order sequences
that contrasts with the evidence of a glacio-eustatic sea-
level rise which alone would have produced landward
migration of the equilibrium point of progressively
younger sequences. This counter-intuitive internal
stacking pattern of PGS is thought to be controlled by
regional tectonic uplift which forced the erosion and
incision of inland areas and the seaward migration of the
fourth-order sequences.
RESULTS
Modeling the small-scale depositional architecture of
the PG2 gravel beaches
Following the acquisition of a large LIDAR dataset from
two quarries (ESI and Tiberi quarries, see figure 2)
excavated in the lower PGS, the best exposed quarry faces
were chosen as input dataset for modeling in 3D the
depositional architecture of the PG2 gravel beaches.
The object to be modelled is a few m-thick gravel beach
depositional body representing a forced-regressive
deposit, which is exposed along two orthogonal faces of
the Tiberi quarry over about 350 m along strike (i.e., along
the palaeo-shoreline, namely in a NS direction) and 100 m
along dip (perpendicular to the palaeo-shoreline, that is
along the dip direction of the beach face clinoforms). In
the package to be modelled, the beach body profile can be
followed from the foreshore down to the upper shoreface
passing through the beachface which clearly shows
clinoforms with average dip angle of approximately 10°.
The photorealistic model reconstructed though LIDAR
and ortho photogrammetric techniques was interpreted
using Geco (ENI proprietary software for 3D visualization
Mattia Marini et al.
104 Journal of Mediterranean Earth Sciences Special Issue (2013), 103-109
Fig. 1 - Geological sketch of the Tyrrhenian margin, central Italy. 1) Messinian to Holocene sedimentary deposits; 2) Plio-Pleistocene
lavas and volcaniclastic deposits; 3) Meso-Cenozoic sedimentary deposits; 4) main buried faults; 5) strike-slip faults; 6) normal faults;
7) major thrusts. The black square indicates the study area.
MARINI_ARGENTI 30/07/13 14:08 Pagina 104
Multi-scale characterization of the Pleistocene-Holocene Tiber... 105
Journal of Mediterranean Earth Sciences Special Issue (2013), 103-109
Fig. 2 - Correlation panel showing the stratigraphic relationship between the high-frequency sequences cropping out in the Tiberi and ESI quarries. The correlation was worked out with the aid of
logs measured in the field and boreholes (from Milli, 2013, unpublished data).
MARINI_ARGENTI 30/07/13 14:08 Pagina 105
and interpretation of outcrop data) picking first the main
bounding surfaces and then discretizing the outcrop face
into grain size classes (sensu Blott and Pye, 2012). A grain
size-approach was preferred in place of a facies approach
as gravel beaches are generally constituted by packages of
very thin beds of rather homogeneous texture which
represent the actual building block of the beach itself and
are indicative of precise hydrodynamic condition along
the depositional profile. Sandy silts are only present at the
top of the modelled depositional body where their
deposition marks the maximum flooding surface
separating the two superimposed beach sub-units
constituting the PG2 in this area.
Planar parallel laminated gravelly sands dipping
seaward at very low angle (0° to 5°) represent the main
texture in the foreshore which along dip passes to sandy
and gravelly clinoforms internally made of dm-thick beds.
Occasionally, coarser bodies of sandy gravel are found in
the lower beach face where they would testify a seaward
transport of coarse sediment related either to distributary
channel mouths or to rip currents. The upper shoreface is
characterized by dm-thick alternations of sands and
gravely sands where the latter would record episodic
gravitative events ignited by either wave of storm-induced
shocks on the beachface. Because of the highly dynamic
nature of both foreshore and beachface, individual beds
prevalently show erosional bases across which they are
welded to form composite depositional bodies with
homogenous grain-size and highly complex geometry. In
the studied example, these bodies are commonly
represented by m-thick welded clinoforms of sandy
gravels showing lateral continuity of few 10’s m along
either strike and dip (Fig. 3)
The overall architecture of the lower PGS beach bodies
has been unravelled correlating the two quarries with the
aid of stratigraphic logs and boreholes (Fig. 2). The
continuity of the individual beach packages in strike
section can be estimated to at least 10 km in a N-S
direction whilst along dip it would not exceed few
hundred m before the beach body fades into lower
Mattia Marini et al.
106 Journal of Mediterranean Earth Sciences Special Issue (2013), 103-109
Fig. 3 - Dip-section photo panel of the Tiberi quarry (a) exposing the PG1 and PG2 sequences and Geco interpretation window (b)
showing the distribution of grain-size classes and depositional environments.
MARINI_ARGENTI 30/07/13 14:08 Pagina 106
shoreface facies association. Nonetheless, size and internal
architecture of these packages may be highly complex as
their deposition is intimately related to the interplay of
high frequency relative sea level changes and other
allocyclic-autocyclic factors, such as longshore sediment
drifting and compensation of local topography. Individual
beach packages are likely to represent single sandy-
gravelly flow units with primary permeability in range of
10-4-10-3 m/s containing even more permeable pockets
of either coarser or better sorted sediments (Fig. 4). The
two beach packages of PG2 are separated by a m-thick bed
of silty sands, a seal (Fig. 3) which landward is likely to be
eroded by the upper beach package.
The bottom and top of the package to be modelled were
reconstructed through a step-wise process which required
to linearly interpolate picked horizons with isolines in
plan view and then model surfaces with the convergence
interpolation algorithm. The output from Geco, which
consisted of a cloud of points coded according to grain
size classes, was imported in Petrel 2011 by Schlumberger
and up-scaled to a grid conformable to clinoforms with
2.5 m ·2.5 m ·0.1 m (I ·J ·K) cells. Grain size classes
were than modelled in either as a continuous or discrete
property using a Sequential Gaussian Simulation or
Sequential Indicator Simulator, respectively. Further
models were run to account for interpretation bias and
trends in grains size possibly related to changes in beach
dynamics and sediment supply.
Seismic modeling
Geometrical representations of the TDS and the
uplifted Pleistocene deposits to be modeled were digitized
from stratigraphic cross-sections which in turn were
reconstructed correlating a large number of boreholes and
sparse outcrops. For each deposition unit, parameters
such as effective and total porosity, fluid content and
mineral components percentages (quartz, feldspar, calcite,
illite, smectite, etc.) were defined basing on results of
petrophysical analysis conducted on samples from the
well ‘Pesce Luna’ (Milli et al., 2013). Density and elastic
moduli of the mineral components were derived from the
literature (Ahrens, 1995; Morcote at al., 2010; Wang et al.,
2001) in order to estimate seismic velocities and density
for each unit applying rock-physics equations (Hashin-
Shtrickman bounds, critical porosity and friable sand
models, Gassmann’s law, Mavko et al., 2009). Predictions
for Vp, Vs and density were computed for each model at
different burial depth and gas saturation. The seismic
response at different resolution was then computed using
a convolutional model and a Ricker wavelet in order to
highlight the relationships between stratigraphy and
seismic imaging and describe the possibility to seismically
resolve various aspects of a complex stratigraphy under
different conditions. Three different wavelet frequencies
were used: 30, 60 and 100 Hz.
The starting geomodels define the lateral and vertical
distribution of “impedance facies” (Fig. 5), obtained
through the computation of impedance values for each
facies in the cross sections. Although buonding surfaces
are often supposed to represent impedance boundaries
generating laterally-persistent reflections, our results
demonstrate that in heterogeneous reservoirs impedance
and polarity changes may be common along the same
bounding surface. This is more true for subaerial
sequence boundaries or regressive surfaces of marine
erosion, while some marine flooding surfaces tend to
show more laterally continuos impedance boundaries.
These conditions also dramatically change in relation to
the thickness changes of the facies above and below each
bounding surface and therefore show very different
imaging when varying wavelet frequency.
This implies that under different resolution potential,
seismic imaging defines an hypotetical reservoir
architecture that is changing not merely in terms of
resolved thickness and of an higher or lower degree of
detail (that could be possibly corrected through standard
up-scaling or down-scaling procedures), but more
problematically generating different shape of the
identified geobodies.
Multi-scale characterization of the Pleistocene-Holocene Tiber... 107
Journal of Mediterranean Earth Sciences Special Issue (2013), 103-109
Fig. 4 - Block diagram showing the scale and geometry of depositional bodies making up the gravel beach of the PG2.
MARINI_ARGENTI 30/07/13 14:08 Pagina 107
Mattia Marini et al.
108 Journal of Mediterranean Earth Sciences Special Issue (2013), 103-109
Fig. 5 - Example of seismic modeling of the Tiber sequence incised valley fill showing, from top to bottom: a) stratigraphic cross-section of the Tiber Depositional Sequence;.b) strike impedance
facies model; c) strike synthetic seismic profile at 100 Hz bandwidth.
MARINI_ARGENTI 30/07/13 14:08 Pagina 108
IMPLICATIONS OF RESULTS
FOR HYDROCARBON E&P
The preliminary results of this research are useful for
hydrocarbon reservoir characterization from both the
exploration and production perspectives as they prove
that:
i) the architecture of 4th-order sequences, despite a
relatively short duration (100 Kyr) and a small thickness
(10 to 70 m for the PG9) can be very complex internally
and result in a variety of possible connectivity scenarios,
which require closely and well distributed borehole data to
be explored and a detailed stratigraphic reconstruction.
From this standing point, the PG9 is a scholarly
depositional analogue comprising a number of peculiar
features of both estuary and deltaic systems;
ii) in shallow buried reservoirs, seismic can capture fine
scale architectural elements such as those described from
the PG9. Hence, the architecture of studied depositional
analogue detailed in this study represents a useful
reference model for interpreting seismic from reservoir
hosted in alike depositional sequences;
iii) in deeper conditions, imaging different “impedance
facies” becomes a crucial issue in heterogeneous
reservoirs. Although bounding surfaces are often
supposed to generate laterally-persistent reflectors, our
results demonstrate that impedance and polarity changes
may be common along the same bounding surface. This
implies that apparent changes in reservoir architecture are
not merely associated to the obvious decay in resolution
but may affect the correct identification of geobodies
shape and connectivity;
iv) the complexity of gravel beaches require a multi-
scale characterization to capture both their overall 3D
shape and connectivity of reservoir facies;
v) a successful facies modeling of outcrop data can allow
reconstructing semi-quantitatively the fine-scale
permeability structure. Results obtained in this study
using Petrel 2011 by Schlumberger can be exported to
analogue reservoir contributing to formation evaluation
and implementation of production strategy.
ACKNOWLEDGEMENTS - The authors would to like to
thank ENI E&P Division for having financially supported the
research project and for allowing the publication of data and
ISPRA, The Geological Survey of Italy for the sampling of Pesce
Luna well.
REFERENCES
Ahrens T.J. 1995. Rock Physics & Phase Relations - A Handbook
of Physicals Constants, American Geophysical Union, ISBN
1080-305X.
Blott S.J., Pye K. 2012. Particle size scales and classification of
sediment types based on particle size distributions: Review
and recommended procedures. Sedimentology: 59, 2071-
2096. doi: 10.1111/j.1365-3091.2012.01335
Mavko G., Mukerji T., Dvorkin J. 2009. The Rock Physics
Handbook, Cambridge university Press, ISBN 978-0-521-
86136-6.
Milli S. 1997. Depositional setting and high-frequency sequence
stratigraphy of the Middle-Upper Pleistocene to Holocene
deposits of the Roman basin. Geologica Romana: 33, 99-136.
Milli S., Palombo M.R., Anzidei A.P. 2011. I depositi
pleistocenici di Ponte Galeria e la Polledrara di Cecanibbio.
Guida all’escursione post-congresso AIQUA 2011 “Il
quaternario Italiano - Conoscenze e prospettive, Roma 26
Febbraio 2011, p. 37.
Milli S., D’Ambrogi C., Bellotti P., Calderoni G., Carboni M.G.,
Celant A., Di Bella L., Di Rita F., Frezza V., Magri D.,
Pichezzi R.M., Ricci V. 2013. The transition from wave-
dominated estuary to wave-dominated delta: the Late
Quaternary stratigraphic architecture of Tiber deltaic
succession (Italy). Sedimentary Geology: 284-285, 159-180.
http://dx.doi.org/10.1016/j.sedgeo.2012.12.003
Morcote A., Mavko G., Prasad M. 2010. Dynamic elastic
properties of coal. Geophysics: 75,227-234.
Wang Z., Wang H., Cates M.E. 2001. Effective elastic properties
of solid clays. Geophysics: 66, 428-440.
Multi-scale characterization of the Pleistocene-Holocene Tiber... 109
Journal of Mediterranean Earth Sciences Special Issue (2013), 103-109
MARINI_ARGENTI 30/07/13 14:08 Pagina 109
