3d‐dynamic modelling of cretaceous sandstones at the outcrop scale (galve sub‐basin, iberian...
TRANSCRIPT
3D‐dynamic modelling of Cretaceous sandstones
at the outcrop scale (Galve Sub‐basin, Iberian Basin).
Application to the studies of CO2 injection.
Ph.D. thesis December 2015
Fernanda de Mesquita L. Veloso
Universidad de Zaragoza Departamento de Ciencias de la Tierra ‐ Area de Estratigrafía
3D‐dynamic modelling of Cretaceous sandstones at the
outcrop scale (Galve Sub‐basin, Iberian Basin). Application
to the studies of CO2 injection.
Ph.D. thesis, December 2015
Fernanda de Mesquita L. Veloso
Supervisors: Ana Rosa Soria de Miguel (Universidad de Zaragoza) Maria Nieves Melendez Hevia (Universidad Complutense de Madrid)
Cover Picture:
CO2 gas saturation in the study case TsunV2 (constant rate regime) at the
beginning of the injection simulation and three years later.
https://es.linkedin.com/pub/fernanda‐de‐mesquita‐lobo‐veloso/9b/727/5a4
To access my LinkedIn profile:
This thesis is dedicated to my parents and my grandmother.
Essa tese é dedicada aos meus pais e a minha avó.
Esa tesis es dedicada a mis padres y mi abuela.
“Agir, eis a inteligência verdadeira. Serei o que quiser. Mas
tenho que querer o que for. O êxito está em ter êxito, e não em
ter condições de êxito. Condições de palácio tem qualquer terra
larga, mas onde estará o palácio se o não ficarem ali.”
Fernando Pessoa (1888‐1935)
Livro do desassossego (1982)
ACKNOWLEDGEMENTS
This Ph.D. project was constructed and achieved with the assistance
and advice of wonderful and competent persons, as well as with the support
of University of Zaragoza, Ciencias de la Tierra department, laboratories as
such the “Servicio de Preparación de Rocas y Materiales Duros” and members
of Stratigraphy section.
I offer my sincerest gratitude to my supervisors, Ana Rosa and Nieves
Melendez who believed in this project and made me turn it in a Ph.D. They set
excellent examples as successful and accomplished geologist scientist women
(and mothers). I also thank Peter Frykman (from GEUS) who has supported me
when it was still a project, then throughout its 4‐years’ development, and until
now.
I would like to thank the CNPq (Conselho Nacional de Pesquisa for the
fellowship that I was awarded to conduct this Ph.D., and the Spain government
(proyectos I+D, CGL2011‐23717) for the financial support to realize this
research. Many thanks to Carlos Liesa, Antonio Casas and Luis Arlegui from the
Geotransfer Research Group, and Enrique Arranz from petrography section,
for their technical support to collect, prepare and analyze samples. Special
thanks to Lope, Rocio and Roy of the stratigraphy section for the nice
moments together (when I was there...).
The international collaborations contributed immensely to my personal
and professional competences. Thanks to Laboratoire de Fluides Complexes et
ses Réservoirs of the UPPA (Université de Pau et de Pays de l’Adour), specially
to Patrice Creux, Charles Aubourg and Jean Paul Caillot for the outstanding
scientific stay and enlightening professional discussions. Thanks to GEUS
(Geological Survey of Denmark), in particular to Carsten, Lars, Neils and
everyone from the Reservoir department, for the short but intense and
efficient stay during the final steps of the Ph.D. Also I would like to thank
Schlumberger for making the modelling and simulation steps of this thesis
possible using one of the best available commercial software (Petrel and
Eclipse). I acknowledge the IEAGHG for the opportunity to join the 6th
international interdisciplinary CCS (CO2 Capture and Storage) summer school
which permitted me to create and expand my network in the CCS, as well as to
understand the CCS chain and deployment.
I take this opportunity to express my gratitude to my climber friends
(from Sagunto, Benidorm, Zaragoza and France) who helped me maintain a
clear brain. Above all, I also thank my parents for the unceasing
encouragement, support and attention. I am also grateful to Jean (my
husband), who endured me through this venture and stayed on my side in all
happy and painful moments. Thanks also to Françoise and Fabio for their
understanding when reading my “English” version.
Obrigada!
Gracias!
Merci!
Thanks!
Tak!
CONTENTS
ABSTRACT ........................................................................................................................... 3
RESUMEN ............................................................................................................................ 5
RESUMO .............................................................................................................................. 7
1. INTRODUCTION ........................................................................................................ 11
1.1. THE STUDY AREA AND THE SELECTED OUTCROP: THE ALIAGA OUTCROP .................................... 13
1.2. GOALS ........................................................................................................................ 14
1.2.1. Framework .......................................................................................................... 14
1.3. STATE OF THE ART ......................................................................................................... 16
1.3.1. CO2 geological storage ........................................................................................ 16
1.3.2. Reservoir and outcrop heterogeneities ............................................................... 19
1.3.3. The numerical simulator: Eclipse300 .................................................................. 20
2. ALIAGA OUTCROP: SEDIMENTOLOGY AND PETROPHYSICS ........................................ 23
2.1. GEOLOGY AT THE BASIN SCALE ......................................................................................... 23
2.2. GEOLOGY AT THE OUTCROP SCALE .................................................................................... 29
2.2.1. Materials and methods ....................................................................................... 34
2.2.2. Sandstone deposits at the macroscale ................................................................ 37 2.2.2.1. Tsunami deposit .......................................................................................... 37 2.2.2.2. Barrier island ‐ tidal inlet deposit ................................................................ 39
2.2.3. Sandstone deposits at the microscale ................................................................. 43 2.2.3.1. Microfacies of the tsunami deposit ............................................................ 43 2.2.3.2. Microfacies of the barrier island ‐ tidal inlet deposit .................................. 50
2.3. PETROPHYSICS ............................................................................................................. 55
2.3.1. Petrophysics of the tsunami deposit ................................................................... 55
2.3.2. Petrophysics of the barrier island ‐ tidal inlet samples ....................................... 59
2.4. DISCUSSION ................................................................................................................. 63
2.5. CONCLUSION ............................................................................................................... 70
3. GEOLOGICAL AND PETROPHYSICAL MODELLING OF THE ALIAGA OUTCROP ............... 75
3.1. MATERIALS AND METHODS ............................................................................................. 76
3.1.1. Geostatistical analyses ........................................................................................ 76 3.1.1.1. Spatial data description .............................................................................. 77 3.1.1.2. Stochastic simulation algorithms ................................................................ 79
3.2. GRID CONSTRUCTION .................................................................................................... 81
3.2.1. Input surface/horizon construction ..................................................................... 83
3.2.2. Layering .............................................................................................................. 86
3.3. FACIES MODELLING ....................................................................................................... 87
3.3.1. Facies modeling of the tsunami deposit ............................................................. 87 3.3.1.1. Results of facies modelling of the tsunami deposit .................................... 92
3.3.2. Facies modelling of the barrier island‐tidal inlet deposit .................................... 96
2
3.3.2.1. Results of facies modelling of the barrier island‐tidal inlet deposit ............ 99 3.3.3. Discussion on facies modelling .......................................................................... 103
3.3.3.1. Discussion on tsunami facies modelling .................................................... 104 3.3.3.2. Discussion on barrier island – tidal inlet deposit ...................................... 110
3.4. PETROPHYSICAL MODELLING ........................................................................................ 114
3.4.1. Petrophysical modelling of the tsunami deposit ............................................... 114 3.4.1.1. Porosity modelling of the tsunami deposit ............................................... 114
3.4.1.1.1. Results of porosity modelling of the tsunami deposit ...................... 117 3.4.1.2. Permeability modelling of the tsunami deposit ........................................ 121
3.4.1.2.1. Results of permeability modelling of the tsunami deposit ............... 122 3.4.2. Petrophysical modelling of the barrier island – tidal inlet deposit .................... 123
3.4.2.1. Porosity modelling of the barrier island – tidal inlet deposit .................... 123 3.4.2.1.1. Results of porosity modelling ........................................................... 126
3.4.2.2. Permeability modelling of the barrier island – tidal inlet deposit ............. 131 3.4.2.2.1. Results of permeability modelling .................................................... 131
3.4.3. Discussion on Petrophysical Modelling ............................................................. 134
3.5. MODELLING CONCLUSION ............................................................................................ 136
4. CO2 INJECTION IN THE TSUNAMI AND BARRIER ISLAND – TIDAL INLET RESERVOIRS AT
THE OUTCROP SCALE ........................................................................................................141
4.1. RESERVOIR MODEL ..................................................................................................... 141
4.2. FLUID MODEL ............................................................................................................ 142
4.3. FLUID PROPERTIES ....................................................................................................... 145
4.3.1. Density .............................................................................................................. 145
4.3.2. Viscosity ............................................................................................................ 146
4.3.3. pH calculation ................................................................................................... 146
4.3.4. Saturation functions .......................................................................................... 146
4.3.5. Diffusion ............................................................................................................ 147
4.4. RESULTS OF RESERVOIR SIMULATION ............................................................................... 148
4.4.1. Simulation on the Tsunami reservoir ................................................................ 148
4.4.2. Simulation on the barrier island – tidal inlet deposit reservoir ......................... 152
4.4.3. Flow at the boundary grid blocks of reservoirs ................................................. 157
4.5. DISCUSSION ............................................................................................................... 158
4.6. CONCLUSION ............................................................................................................. 162
5. CONCLUSIONS AND PERSPECTIVE ............................................................................165
CONCLUSIONES Y PERSPECTIVA ........................................................................................173
REFERENCES .....................................................................................................................183
3
ABSTRACT
Geological and reservoir modelling are mandatory in studies that regard
the geological storage of CO2. The aim of this study was to investigate the
intra‐unit heterogeneity of the two sandstone deposits observed in the Aliaga
outcrop at metre scales, and to examine how heterogeneity can impact the
behaviour of CO2 in the zone close to the injector well during injection and
post‐injection processes in a deep saline aquifer scenario.
The Aliaga outcrop, is an 8400‐m² 2D vertical face in the upper part of
the Camarillas Fm. (Early Cretaceous, Galve sub‐basin). The studied sandstone
deposits correspond to a tsunami and a barrier island – tidal inlet, which were
described in macroscale (centimetre to metric) and microscale (micron);
further cores were drilled along outcrop to collect samples for porosity and
permeability measurements. The two sandstone deposits were generated by
distinct sedimentary processes under the same sedimentary system, and
showed distinct petrophysical characteristics.
The modelling process was different for each deposit, and honoured
the petrophysical characteristics that were used to build the reservoir model
post‐hoc. The petrophysics models reflected the sandy variability, which was
represented by the facies distribution. The facies were defined as a function of
the sand sorting at the petrographic scale. The tsunami facies and porosity
distributions are homogeneous, whereas the barrier island‐tidal inlet facies
and porosity distributions are heterogeneous. Porosity and permeability are
strongly correlated in both deposits; thus, the permeability modelling was
carried out as a function of the porosity model by applying a regression
equation.
Although the permeability is usually low (tens of mD), the two deposits
behaved as a reservoir. At a short‐time scale (7 years), both reservoirs stored
at least of 60% injected CO2, with the 20‐40% dissolved in the brine. At the
sub‐metric scale, under the same reservoir conditions and fluid model
parameters, the thickness of reservoir has the major impact in the amount of
CO2 dissolution rather than the permeability contrast.
4
5
RESUMEN
La modelización geológica y de reservorios es obligatoria en los
estudios sobre el almacenamiento geológico de CO2. El objetivo de este
estudio ha sido investigar las heterogeneidades intra‐capa en dos cuerpos
arenosos a escala de afloramiento (escala métrica) observados en el
afloramiento de Aliaga, y examinar cómo estas heterogeneidades pueden
afectar al comportamiento del CO2 en la zona próxima al pozo inyector,
durante los procesos de inyección y post‐inyección, en un escenario
correspondiente a un acuífero salino profundo.
El afloramiento de Aliaga, situado en la parte superior de la Fm.
Camarillas (Cretácico Inferior, Subcuenca de Galve), tiene una superficie de
afloramiento 2D de 8.400 m². Los dos cuerpos arenosos estudiados
corresponden al depósito de un episodio de tsunami y de un complejo isla de
barrera/inlet. Estos depósitos fueron cartografiados a escala de afloramiento y
descritos a todas las escalas (desde la macroescala a la microescala) y además,
en ellos, se extrajeron núcleos de sondeos sobre los que se realizaron las
medidas de porosidad y permeabilidad. Ambos depósitos se generaron por
procesos sedimentarios distintos dentro de un mismo sistema sedimentario
(ambientes de back barrier) y muestran características petrofísicas distintas.
El proceso de modelización fue diferente para cada depósito, y respetó
las características petrofísicas que se emplearon para construir
posteriormente el modelo de yacimiento. Los modelos petrofísicos reflejan
principalmente la variabilidad de las areniscas, representada por la
distribución de las distintas facies arenosas definidas. Estas facies arenosas
han sido definidas a partir de la selección de las areniscas a escala
petrográfica. La distribución de las facies del depósito de tsunami es
homogénea, así como la distribución de su porosidad, mientras que en el
depósito de isla de barrera/inlet, tanto la distribución de las facies como la de
la porosidad son heterogéneas. La porosidad y permeabilidad están
fuertemente correlacionados en ambos depósitos; por lo tanto, el modelo de
permeabilidad se llevó a cabo como una función del modelo de porosidad
mediante la aplicación de una ecuación de regresión.
6
A pesar de que la permeabilidad es generalmente baja en ambos
depósitos (decenas de mD), ambos se comportaron como reservorios. A una
escala de tiempo corto (7 años), ambos reservorios almacenan al menos un
60% de todo el CO2 inyectado, con un 20‐40% de él disuelto en la salmuera. A
escala submétrica, para las mismas condiciones de yacimiento y los mismos
parámetros del modelo de fluido, la potencia del reservorio tiene un mayor
impacto en la cantidad de disolución del CO2 que el contrate de permeabilidad
del yacimiento.
7
RESUMO
A modelagem geológica torna‐se necessária em estudos para
caracterização de um reservatório para armazenamento geológico de CO2.
Com este propósito, o presente estudo teve como objetivo investigar as
heterogeneidades intra‐layer (intracamadas) em dois corpos aflorantes de
rochas areníticas, em escala métrica, do afloramento denominado Aliaga e
analisar como essas heterogeneidades podem afetar o comportamento do CO2
na área próxima ao poço de injeção durante os processos de injeção e pós‐
injeção em um cenário de armazenamento em aqüífero salino profundo.
O afloramento de Aliaga situa‐se na parte superior da Formação
Camarillas, pertencente à sub‐bacia de Galve, considerada de idade do
Cretáceo Inferior. Em superfície aflora abrangendo uma área de 8.400 m²
(afloramento em 2D). Os dois corpos de arenito estudados para serem
utilizados com reservatórios geológicos de CO2 correspondem a um episódio
sedimentar relacionado a eventos proporcionados por tsunami e ilha barreira /
canal inlet. Ambas as unidades foram gerados por diferentes processos
deposicionais dentro do mesmo sistema sedimentar (parte continental da ilha
barriera) e mostram diferentes características petrofísicas. Estas duas
unidades foram mapeadas na escala de afloramento e descritas em escalas
microscópica a macroscópica e, ainda, foram efetuadas sondagens e os
testemunhos foram utilizados para obtenção das medições de porosidade e
permeabilidade.
O processo de modelagem aplicado foi distinto para cada unidade
estudada, e assim, respeitadas suas características petrofísicas que foram
utilizadas posteriormente para construir os modelos geológico dos
reservatórios. Modelos petrofisicos refletem, principalmente, a variabilidade
dos arenitos, representada pela distribuição dos vários fácies arenosos
definidos. Estes fácies arenosos foram classificados a partir da seleção de grãos
descritos em escala petrográfica. Os fácies do depósito de tsunami apresentam
uma distribuição homogênea, assim como a porosidade, enquanto que os
facies do depósito de ilha barreira/ canal inlet tanto a distribuição de fácies e a
porosidade são heterogêneos. Porosidade e permeabilidade estão fortemente
correlacionados em ambos os depósitos, por conseguinte, o modelo de
8
permeabilidade foi realizado como uma função do modelo de porosidade por
meio de uma equação de regressão .
Embora a permeabilidade seja geralmente baixa em ambas as unidades
(dezenas de mD), as mesmas podem se comportar como reservatórios. Numa
escala de tempo curto, de aproximadamente 7 anos, ambos os reservatórios
apresentam capacidade de armazenamento de, pelo menos, 60 % do CO2
injectado e entre 20 e 40 % desse CO2 de se dissolver na solução salina. Na
escala submétrica, para as mesmas condições de reservatório e os mesmos
parâmetros do modelo de fluido, a espessura do reservatório tem um impacto
maior sobre a quantidade de CO2 dissolvido do que o contraste de
permeabilidade do reservatório.
9
CHAPTER 1
INTRODUCTION
1.1 The study area and the selected outcrop: The Aliaga
outcrop_________________________________________13
1.2 Goals______________________________________14
1.3 State of the art______________________________16
10
11
1. Introduction
The ‘greenhouse effect’ refers to processes that trap heat in the
atmosphere and thus prevent heat loss to space. The effect is primarily the
result of enhanced concentrations of carbon dioxide (CO2) and other gases in
the atmosphere, that absorb and re‐radiate energy. Human activity related to
the burning of fossil fuels is largely responsible for recent changes in the
natural CO2 balance of the Earth–atmosphere system, and is thus responsible
for the recent intensification of the greenhouse effect on Earth (IEAGHG,
2013). In Europe alone, the total emission of CO2 from the 28 member states
(EU28) caused by human activity (excluding land use activities, land‐use
changes and forestry; LULUCF) was 2.99 billion tonnes in 2012. The main
source of CO2 emissions was public electricity and heat production (PEHP),
which correspond to 27% of the total CO2 emission. Spain was responsible for
7.5% of the European CO2 emission, occupying 7th place in terms of emissions
after Poland (EEA, 2015). In 2013, Spain produced only 28% of its PEHP,
showing a strong dependency on imported energy (SEE, 2013), despite
estimated carbon reserves in Spain which, in 1992, were ~3463.4 Mt (IGME,
2012); however, Spain’s carbon energy production in 2012 was only 21 Mt
(IGME, 2012), or 1% of its estimated carbon reserves.
Carbon dioxide capture and geological storage (CCS) is a bridging
technology that will contribute to the mitigation of climate change, as it
reduces the amount of carbon in the atmosphere and provides a sustainable
supply of raw materials. The CCS strategy consists of capturing CO2 from
industrial emissions and transporting it to storage sites for injection into
suitable underground geological formations for permanent storage. However,
the characterization and selection of appropriate CCS sites is a lengthy and
costly process, and so must begin early in the project planning process.
Geological storage of CO2 requires satisfactory characterization of reservoir
and caprock geology at both local and regional scales, to elucidate CO2
migration patterns and overall storage potential. The characterization and
assessment of potential storage sites is based on dynamic modelling
comprising a variety of time‐step simulations of CO2 injection into the storage
site, using three‐dimensional static geological Earth models and a complex
computerized storage simulator (EU, 2009).
12
Geological models for the dynamic study of CO2 behaviour in deep
saline aquifers should represent the properties of rock heterogeneity, as
heterogeneity maximizes the long‐term storage of CO2 in cases where
buoyancy forces drive CO2 movement by density differences (Frykman, 2009).
The complexity of geological models is a function of the purpose(s) that the
model addresses; also, geological models need to be easily updated with
monitoring data, so as to predict safe storage and reduce storage uncertainties
(Norden and Frykman, 2013). Rock heterogeneity depends on the scale of
observations, and on the phenomenon being investigated (Cushman, 1997;
Bachu et al., 2007; Frykman, 2009). Heterogeneity at reservoir scales
(kilometres) has been studied in an attempt to understand how it impacts fluid
flow (Hornung and Aigner, 1999; Eaton and Bradbury, 2003; Felletti, 2004;
Eaton, 2006; Issautier et al., 2013; Norden and Frykman, 2013; Asharf, 2014).
Studies of heterogeneity at outcrop scales (metres to hundreds of metres)
have shown the impact of sedimentary heterogeneity on fluid flow and on
aquifer groundwater flow (Robinson and McCabe, 1997; Bersezio et al., 1999;
Klingbeil et al., 1999; Dalrymple, 2001; Heinz et al., 2003; Tye, 2004; Wood,
2004; Huysmans et al., 2008; Frykman et al., 2013). Studies of heterogeneity at
microscopic scales (microns to millimetres) have demonstrated the impact and
sensitivity of capillary pressure and relative permeability on CO2 trapping
mechanisms as a function of CO2 saturation (Juanes et al., 2006; Spiteri and
Juanes, 2006; Plug and Bruining, 2007; Pini et al., 2012; Boxiao et al., 2013;
Frykman et al., 2013).
The challenge in building a geological model is the integration at
different scales of heterogeneity and relevant petrophysical characteristics
that impact fluid flow in reservoirs (Corbett and Potter, 2004). The
incorporation of high‐resolution sedimentary heterogeneity into reservoir and
groundwater flow models improves the accuracy of predictions regarding fluid
flow behaviour. Reservoir models are often constructed at field scales (tens to
hundreds of square kilometres), and practical limits on the size of reservoir
simulation models are often imposed (AAPG, 2015). The grid block size in
reservoir models is a function of reservoir heterogeneity, well distances, and
computational expense or capabilities; the grid block size or cell dimension is
usually 40–150 m on horizontal scales and 1–15 m at vertical scales (Asharf,
2014; Issautier et al., 2014; AAPG, 2015). The outcrop scale is a bridge
13
between seismic and core scales, as the outcrop represents the scale of
individual bedforms (metres to hundreds of metres) and laminae (millimetres
to metres) (Yoshida et al., 2001).
1.1. The study area and the selected outcrop: The Aliaga outcrop
Site selection for geological storage of CO2 is a complex issue involving a
variety of geological and non‐geological variables. The quality of the reservoir
rocks and the seal system are of particular importance in site selection, as is
the proximity of the site to CO2 emission sources. The Lower Cretaceous
Camarillas Fm. is a potentially good candidate for CO2 storage because of its
sedimentological characteristics and its geographical location. In this study, we
examined the Camarillas Fm. in the province of Cuencas Mineras, which has
been one of the most important regions in Spain for the supply of raw
materials and for the generation of electricity from coal‐fired power stations
through the centuries. Today, the Andorra power plant, located 60 km from
the study area, is one of the most important heat and electricity generation
facilities in the Spain. However, a large‐scale study of the potential of the
Camarillas Fm. for CO2 storage would be costly and prolonged; consequently,
this study does not examine or appraise the storage capacity of the Camarillas
Fm. per se, but rather presents a low‐cost approach for investigating the
dynamic behaviours of two sandstone units within the Camarillas Fm.,
determined at outcrop scales.
Previous stratigraphic and sedimentological studies of the Iberian Basin
by Soria (1997), Navarrete et al. (2013, 2014) and Navarrete (2015) identified a
fault‐bounded sub‐basin, the Galve Sub‐basin, within the Maestrazgo basin.
The Camarillas Fm., which was deposited during the Barremian synrift phase, is
one of the most important sedimentary units in the Galve Sub‐basin (Soria,
1997); the formation consists of red clays and sandstones, and reaches a
thickness of up to 800 m (Navarrete, 2015). The Aliaga outcrop, located in the
upper part of the Camarillas Fm., exposes two sandstone bodies that were
generated by distinctive sedimentary process; one of the sandstones is a
tsunami deposit, and the other is a barrier island ‐ tidal inlet deposit. Despite
their limited thicknesses (1–7 m in each case), both are recognized over vast
areas (35 km²). The Aliaga outcrop reveals important information about
14
variations in the sandstone, in terms of the size, sorting and nature of
component sedimentary grains, over a distance of 200 m, allowing for a study
of the distribution of sand in the deposits at different scales of observation.
1.2. Goals
The aim of this study was to investigate the intra‐unit heterogeneity of
the two sandstone deposits observed in the Aliaga outcrop at metre scales,
and to examine how heterogeneity can impact the behaviour of CO2 in the
zone close to the injector well during injection and post‐injection processes in
a deep saline aquifer scenario. Dynamic 3D models were constructed from
outcrop data obtained at the Aliaga outcrop; the outcrop provides direct
access to geological information and sedimentary samples for sedimentary
and petrophysical analyses.
1.2.1. Framework
Constructing a geological model for flow simulation requires a step by
step validation approach. The framework of this study, presented in Fig. 1.1, is
organized into three connected parts, referred to as panels, corresponding to
Chapters 2, 3 and 4 of the text. Although the parts (panels) are inter‐
dependent, each part has an independent methodology used to accomplish its
objective.
Panel 1 is discussed in Chapter 2, Sedimentology and Petrophysics of
Sandstone Deposits (Fig. 1.1). The objective of this study was to establish a
correlation between sandy facies and petrophysical characteristics at outcrop
scales. The stratigraphic and sedimentological studies of Navarrete et al. (2013
and 2014), conducted at basin‐wide scales, identified the Camarillas Fm. as a
good candidate for geological CO2 storage purposes. The Camarillas Fm. crops
out for many kilometres in the study area, and is composed of sandstones
interbedded with shales and marls. The Aliaga outcrop exposes the upper part
of the Camarillas Fm., which consists of a transitional sedimentary interval
from sandy‐dominant to carbonate‐dominant deposits. Tsunami and barrier
island ‐ tidal inlet deposits, recognized at basin‐wide scales, were described in
macroscale (centimetre to metre) and microscale (microns) over the 200‐m
length of the outcrop. In addition, cores were drilled along the outcrop to
15
Fig. 1.1: Framework of undertaken actions in this study.
collect samples for porosity and permeability measurements. Then, a facies
coding scheme was developed, taking into consideration the sedimentary
characteristics that are most relevant to, and most highly correlated with,
hydrodynamic parameters.
Panel 2 is discussed in Chapter 3, Geological and Petrophysical
Modelling of the Aliaga Outcrop (Fig. 1.1). This section describes the
construction of 3D models of tsunami and barrier island/inlet deposits for
reservoir simulation studies. The modelling process uses a combination of
geostatistical methods based on the original data distribution, and is
conditioned by the quality and quantity of input data. The vertical and
horizontal resolution of the 3D grid was constructed as a function of
sedimentary heterogeneity. The facies code defined in Panel 2 was used in the
16
facies modelling process, and the facies model was conditioned by the
petrophysical modelling.
Panel 3 is discussed in Chapter 4, CO2 injection in the tsunami and
barrier island – tidal inlet reservoirs at the outcrop scale (Fig. 1.1). The panel
describes the dynamic analyses of the two sandstone deposits during and after
CO2 injection. Each deposit was investigated as an individual reservoir at
conditions that allowed CO2 to behave as a critical fluid. The CO2, injected into
the reservoir as a dry gas, interacted with brine from the first days of injection,
until the end of simulation, after the injection was stopped. The physical and
chemical processes of injected CO2 into the reservoir are complex, with the
calculation time depending on: number of grid cells, complexity of the model
and complexity of the fluid behaviour. Because the fluid model is simple, the
major complexity is related to model resolution (size and number of grid cells)
and the distribution of petrophysical characteristics. The sensibility of the
injection regime and the injector well location were tested for four study cases
in each deposit.
Finally, the chapter 5 presents the general conclusions and perspectives
of this Ph.D. thesis.
1.3. State of the art
1.3.1. CO2 geological storage
Geological CO2 storage is achieved through a combination of physical
and chemical trapping mechanisms that are effective over different
timeframes and spatial scales (IPCC, 2005; Bachu et al., 2007). Key geological
characteristics used to evaluate the practicability of geological CO2 storage
include: reservoir depth, reservoir thickness, porosity, permeability, seal
integrity and aquifer salinity (Chadwick et al., 2008). Thus, the geological
storage capacity of CO2 depends on properties of the reservoir rock and
boundary rocks/faults, as these play a role in physical and chemical trapping
mechanisms.
Four main trapping mechanisms (Fig. 1.2) are required for permanent
and safe CO2 storage in reservoir rocks (IPCC, 2005; Chadwick et al., 2008). (1)
Structural and stratigraphic trapping: CO2 is physically trapped by low‐
17
permeability and low‐diffusivity top‐seal rocks or faults; structural traps
include those formed by folded or fractured rocks. (2) Residual saturation
trapping: capillary forces and adsorption onto surfaces of mineral grains in the
rock matrix immobilise a proportion of the injected CO2 as residual CO2 phase.
(3) Dissolution trapping: dissolution and trapping of injected CO2 within
reservoir brine. (4) Geochemical trapping: reaction of dissolved CO2 with
native pore fluids and/or minerals constituting the rock matrix or reservoir.
Geological CO2 storage can be undertaken in a variety of geological
settings in sedimentary basins: oil fields, depleted gas fields, deep coal seams
and saline formations are all possible storage formations (IPCC, 2005). Some
studies have shown that storage in saline formations has the greatest
potential, with an estimated storage capacity of 1,000–10,000 Gt (IPCC, 2005;
IEAGHG, 2013).
Fig. 1.2: Evolution of CO2 storage mechanisms through time. The horizontal axis shows the time since the start of injection; the right vertical axis shows the trapping contribution percentage of the four main storage mechanisms; the left vertical axis shows the qualitative evolution of CO2 storage mechanisms. Modified from IPCC (2005).
18
Predicting the sequestration potential and long‐term behaviour of CO2
in deep saline reservoirs requires calculations of the pressure (P), temperature
(T) and composition (X) of CO2–H2O mixtures at depths where temperatures
are <100°C, at pressures of up to several hundred bars. The P–T diagram of
pure CO2 phases is presented in Fig. 1.3. At the critical point, CO2 behaves as a
gas (IPCC, 2005), and the amount of H2O in the CO2‐rich phase is small, such
that CO2 properties can be approximated by those of pure CO2 (Spycher et al.,
2003).
The dissolution of CO2 in water (brine or saline formation water)
involves a number of chemical reactions between gaseous and dissolved CO2,
carbonic acid (H2CO3), bicarbonate ions (HCO3−) and carbonate ions (CO3
2–),
which can be represented as (IPCC, 2005):
CO2 (gas) ↔ CO2 (aqueous)
CO2 (aqueous) + H2O ↔ H2CO3 (aqueous)
H2CO3 (aqueous) ↔ H+ (aqueous) + HCO3‐ (aqueous)
HCO3– (aqueous) ↔ H+ (aq) + CO3
2− (aqueous).
Fig. 1.3: Pressure–temperature phase diagram of pure CO2. The arrow indicates the average initial pressure and temperature conditions of the two reservoirs examined in this study. Modified from IPCC, 2005 (from ChemicaLogic Corporation, 1999).
19
1.3.2. Reservoir and outcrop heterogeneities
Sedimentary heterogeneity in reservoir models is usually expressed by
the distribution of low‐permeability structural or diagenetic features, such as
faults, breccia or deformation bands (Eaton, 2006), or by the distribution of
low‐permeability facies, such as mud drapes or shale layers (Ashraf, 2014;
Issautier et al., 2014). Detailed outcrop models have shown the impact of sand
heterogeneity (such as heterogeneities in grain size, sorting indices, net to
gross (ratio of sand and clay content) and rock texture) on the local spatial
distribution of petrophysical properties (such as porosity, permeability and
capillarity entry pressure) (Hornung and Aigner, 1999; Klingbeil et al., 1999;
Heinz et al., 2003; Sun et al., 2007; Ambrose et al., 2008; Huysmans et al.,
2008; Frykman et al., 2013). Studies of groundwater flow or contaminant
movement in aquifers have also demonstrated that sand heterogeneity
determines local groundwater flow patterns and plume dispersion in aquifers
(Koltermann and Gorelick, 1996; Zheng and Gorelick, 2003).
Sedimentary heterogeneity at outcrop scales can be directly observed
and sampled, from the fine‐scale to large‐scale features. The geomodel built
from outcrop data incorporates reservoir and top‐seal heterogeneity and
architecture to investigate the dynamic influence of intra‐body heterogeneities
on reservoir flow simulations (Robinson and McCabe, 1997; Dalrymple, 2001;
Tye, 2004; Wood, 2004; Ekeland et al., 2008). Outcrop models are useful when
the main focus of geomodels is to study injectivity and estimate the capacity of
structural and dissolution trapping as dominant mechanisms, as the location of
the injection well can induce significant changes in migration and trapping
efficiency (Le Gallo et al., 2010). The major impacts of injection occur close to
the wellbore region, and a detailed modelling approach in these regions
contributes to better estimates of the injectivity (Le Gallo, 2009). Therefore,
analogous outcrop studies can supply detailed geological information which
can elucidate geological gaps in local zones on the reservoir model.
Some significant limitations of the study exist with regard to outcrop‐
based studies. First, the data are typically 2D (Lantuéjoul et al., 2005) and thus
present observational biases. Consolidated rock is often preserved in outcrops
while weaker rocks (e.g., shales) are often eroded; the dominance of weaker
rocks can prevent the formation of outcrops altogether. Also, weathering and
20
unloading of rocks may change the nature of outcrop exposures and obscure
features that are relevant in the in situ state (Pyrcz and Deutsch, 2014).
1.3.3. The numerical simulator: Eclipse300
The chemical reactivity of CO2 supercritical fluid with brine in reservoirs
is an important determinant of its flow behaviour in reservoirs. The simulator
Eclipse300 (E300) is a commercial (Schlumberger Company) compositional
simulator based on a cubic equation of state and a pressure‐dependent
permeability value. Technical descriptions reported here are found in the
Eclipse Technical Descriptions, Version 2013.1 (Schlumberger, 2013a).
Several E300 functions are available depending on the site and
operational conditions that need to be modelled. Four equations of state are
available, implemented through Martin's generalized equation (e.g., Martin,
1979): Redlich–Kwong, Soave–Redlich–Kwong, Peng–Robinson and
Zudkevitch–Joffe. The program is written in FORTRAN and operates on any
computer with an ANSI‐standard FORTRAN90 compiler and with sufficient
memory.
To model geological conditions in saline storage aquifers, the
CO2STORE option of E300 offers the possibility of modelling three additional
phases: a CO2‐rich phase (labelled ‘gas’), an H2O‐rich phase (labelled liquid)
and a solid phase. This option gives accurate mutual solubilities of CO2 in
water, and water in the CO2‐rich phase. Solids (salts) can also be included and
described as components of the liquid and/or solid phase.
21
CHAPTER 2
ALIAGA OUTCROP:
SEDIMENTOLOGY AND PETROPHYSICS
2.1 Geology at the basin scale_____________________23
2.2 Geology at the outcrop scale___________________29
2.3 Petrophysics________________________________54
2.4 Discussion__________________________________63
2.5 Conclusion__________________________________70
22
23
2. Aliaga Outcrop: sedimentology and petrophysics
This section attempts to investigate the correlation between the
characteristics of sandy facies and their petrophysical parameters such as
porosity and permeability at the outcrop scale. The Camarillas Fm. is a
relatively thick unit (100‐800 m) consisting of interbedded sandstones, shales,
and marls. The Aliaga outcrop is an 8400‐m² 2D vertical face in the upper part
of the Camarillas Fm.; this outcrop was included in the sedimentological and
stratigraphic studies of Navarrete et al. (2013, 2014) and Navarrete (2015).
The description of the Aliaga outcrop provided here consists of lithological
descriptions of two sandstone deposits at both macroscopic (outcrop) and
microscopic scales. Regionally, the two sandstone deposits are recognized over
vast areas at the basin scale (<7 km), although their individual thickness is
limited to 1–7 m. Macroscale descriptions of the sandstone deposits were
based on the lithofacies descriptions of Navarrete et al. (2013, 2014).
Macrofacies were also sampled for petrographic and petrophysical analyses to
establish a new sandy facies code. This new code describes the variability of
the sandy facies, taking into account the petrophysical parameters that may
have the greatest effect on fluid flow.
2.1. Geology at the basin scale
The study area is located in the Cretaceous Galve sub‐basin in the
Iberian Chain of central–eastern Iberia (Fig. 2.1A). The Galve sub‐basin (40 km
long and 20 km wide, elongate NNW–SSE) developed during Late Jurassic–
Early Cretaceous rifting (e.g., Salas and Casas, 1993; Capote et al., 2002) at an
expansion centre (RRR triple junction type) in the western Tethys (Antolín‐
Tomas et al., 2007). The sub‐basin represents a western marginal
sedimentation area of the coastal Maestrazgo Basin that formed during the
Early Cretaceous. The activity of two main fault sets, one trending NNW–SSE
(e.g. the Alpeñés, Ababuj, Cañada Vellida, and Miravete faults) and the other
trending ENE–WSW (the Campos, Santa Bárbara, Aliaga, Camarillas and
Remenderuelas faults) (Fig. 2.1B and C) determined the Early Cretaceous
extensional structure of the Galve sub‐basin (Soria 1997; Liesa et al., 2000;
Soria et al., 2001; Navarrete et al., 2013, 2014). The Tertiary structure of
24
Fig. 2.1: Geological setting of the study area in the Galve sub‐basin, modified from Navarrete et al. (2013). (A) Location of the Maestrazgo Basin and the Galve sub‐basin; the area shown in Fig. 2.2 is highlighted by the red square (modified from Capote et al., 2002). (B) Block diagram showing the tectonic setting of the Galve sub‐basin during deposition of the El Castellar, Camarillas and Artoles formations. (C) Close‐up of the area highlighted in B, and location of the outcrop between the Aliaga–Miravete Anticline and the Camarillas–Jorcas Syncline; point “8” shows the location of the transitional interval on the far side of the Camarillas–Jorcas Syncline (Fig. 3.3B in Chapter 3). (D) Chronostratigraphic diagram and sedimentary record of the Galve sub‐basin and the studied interval (s.u., synrift unconformity; RT, rift transition). (B) and (D) are modified from Liesa et al. (2006) and Rodríguez‐López et al. (2009).
25
the sub‐basin shows the superimposition of two orthogonal fold‐and‐thrust
structural trends, one striking NNW–SSE and the other WSW–ENE (Guimerà,
1988) (Fig. 2.1B). Both structural trends represent the rejuvenation and
inversion of normal faults, basically inherited from Mesozoic extensional
and/or post‐Variscan fracturing (Guimerà et al., 1996; Soria, 1997; Liesa et al.,
2004). Present‐day morphotectonics are the result of extensional deformation
that began on the eastern margin of the Iberian Peninsula during the mid‐
Miocene (Simón, 1982, 1989), related to rifting in the Valencia Trough (Álvaro
et al., 1979; Simón, 1982).
Outcrops in the study area exposing Palaeogene–Neogene fold
structures provide excellent material for observations and sampling.
Moreover, geological mapping of the steeply dipping (>75°) western limb of
the NNW–SSE‐trending Aliaga–Miravete anticline (Fig. 2.1C) provides a cross‐
sectional view of Cretaceous Galve sub‐basin infill across WSW–ENE‐striking
listric normal faults (Fig. 2.2).
Synrift sedimentation in the Galve sub‐basin spans the late Hauterivian
to the early Albian (Soria, 1997; Soria et al., 2000; Salas et al., 2001; Liesa et
al., 2004, 2006; Peropadre, 2012), and comprises the following units (Fig.
2.1D): (1) an alluvial and lacustrine series (El Castellar Fm.; Soria, 1997) that
records the transition from initial rifting to rift climax (Liesa et al., 2006;
Meléndez et al., 2009); (2) red clays and sandstones (Camarillas Fm.)
previously interpreted as a low‐sinuosity fluvial system with broad flood plains
(Salas, 1987; Soria, 1997); (3) marls and limestones (Artoles Fm) rich in
calcareous algae, planktic foraminifera and molluscs, interpreted as a shallow
marine to transitional carbonate system (Salas, 1987; Soria, 1997) that evolved
northwards (towards the Las Parras sub‐basin) to a coastal lacustrine system
(Soria, 1997); (4) a series of siliciclastic and/or carbonate marine platforms
(Morella, Chert, Forcall, Villarroya de los Pinares and Benasal Fm.)
characteristic of Aptian sedimentation (e.g., Vennin and Aurell, 2001;
Peropadre et al., 2008; Peropadre, 2012); and (5) a late Aptian–early Albian
transitional siliciclastic series with coal beds (Escucha Fm.; Rodríguez‐López et
al., 2009).
The studied outcrop is stratigraphically located in the upper part of the
Camarillas Fm. in a synrift sequence of the Galve sub‐basin (Fig. 2.1D). The
26
27
Camarillas Fm. constitutes one of the most important sedimentary units to be
deposited in the Galve sub‐basin during the synrift phase of the Barremian
(Soria, 1997). The unit exhibits large thickness variations (150–800 m) that are
related to extensional faulting (Soria, 1997; Navarrete et al., 2013) that
occurred at the climax of Cretaceous rifting (Liesa et al., 2004, 2006; Navarrete
et al., 2013). A recent sedimentological study by Navarrete (2015) indicated
that the Camarillas Fm. is a transitional continental‐to‐marine sedimentary
system composed of tidal mud flat with tidal channel deposits at the base,
changing upward into barrier island and lagoon deposits.
The contact between Camarillas Fm. and Artoles Fm. is a transitional
interval characterized by two stages of back‐barrier sedimentation,
represented by mudstone and sandstone beds that have 45 m thick, at least 5
km long, and 7 km wide (Navarrete et al., 2013) (Fig. 2.1D and 2.2). Stage 1
(Fig. 2.3A) is characterised by deposits in extensive back‐barrier mud flats with
tidal creeks and minor washover fans, interbedded with lagoonal carbonates
and influenced by local syn‐sedimentary tectonics and showing thickness
variations, rotated blocks and angular unconformities (Fig. 2.3A). Back‐barrier
stage 2 (Fig. 2.3A) comprises washover fan deposits interbedded with lagoonal
carbonates, the complete absence of back‐barrier tidal mud flats and
associated channels, and well‐developed ebb‐ and flood‐tidal deposits. Local
tectonics did not affect the stratigraphic architecture, thus resulting in flat‐
lying units. Below the back‐barrier system, an exceptional tsunami deposit (up
to 3 m thick) was identified by Navarrete et al. (2014); a remarkable dinosaur
megatracksite of track casts is preserved at the base of the deposit.
‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐
Fig. 2.2 (previous page): The Aliaga–Miravete anticline and the Galve sub‐basin sedimentary infill. (A) High‐resolution satellite image at a scale of 1:5000 (available on the SITAR web page of the Aragón Government). (B) Geological map of Galve sub‐basin units across WSW–ENE‐striking listric faults; the transitional interval and the outcrop locations are marked, as are the locations of stratigraphic profiles. Modified from Navarrete et al. (2013).
28
29
The development of extensive back‐barrier tidal mud flats with tidal
channels during stage 1, and their absence during stage 2, probably indicates
that stage 1 developed under a mesotidal regime while stage 2 developed
under a microtidal regime. This change in the tidal regime coincided with the
retreat of the barrier island system to the east, likely associated with a change
in basin configuration, ultimately controlled by the development of an
extensional basin (Navarrete et al., 2013).
2.2. Geology at the outcrop scale
The Aliaga outcrop is located on the steeply dipping (>75°) western limb
of the NNW–SSE‐trending Aliaga–Miravete anticline, between the ENE–WSW‐
striking Remenderuelas and Camarillas listric faults (Figs. 2.1B, 2.1C and 2.2), in
the transitional interval at the top of Camarillas Fm. (Fig. 2.2). The sedimentary
history recorded at the outcrop corresponds to the stage 1 of back‐barrier
system including the sedimentary datum (Fig. 2.3B), as well as the tsunami
deposit at the base of the interval (Fig. 2.4). The outcrop is included in the
sedimentological and stratigraphic studies of Navarrete et al. (2013, 2014).
Five sedimentary sections were logged in the outcrop (sections 2–8, except 5
and 6 in Fig. 2.3B) and facies associations were described through an
integrated study of sedimentology, stratigraphy and ichnology.
The Aliaga outcrop is a 2D vertical face, 210 m wide and 40 m high (Fig.
2.4), located on road TEV 8008, 5 km from Aliaga village. The road divides the
outcrop into a North sector (on the north side of the road) and a South sector
(on the south side of the road) (Fig. 2.4). Sedimentary sections 2–4, 7 and 8 of
Fig. 2.3B were renamed sections 1–5 (Fig. 2.4) in this study. The Aliaga outcrop
‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐
Fig. 2.3 (previous page): (A) Sedimentary correlation panel of the transitional interval between the Camarillas Fm. and Artoles Fm. (Fig. 2.2) with descriptions of facies associations. The sedimentary datum is the barrier island and tidal inlet deposit facies association (FA4) with the basal wave ravinement surface (wRs). (B) Detail of sedimentary sections representing outcrop face; mfs is the maximum flooding surface. Source: Navarrete et al. (2013).
30
Fig. 2.4: Aerial photograph of Aliaga outcrop, supplied by the Aragon Visor 2D (http://sitar.aragon.es/). The Aliaga outcrop is outlined
by blue rectangle; the sedim
entary sections 1‐5 correspond to sedim
entary sections 2‐8, except 4 and 5 of Fig. 2.3.
31
is delimited at the base by the bottom surface of the tsunami deposit and at
the top by top surface of the barrier island/tidal inlet deposit. These deposits
(Fig. 2.4) were examined in detail to investigate the correlation between sandy
facies and petrophysical characteristics of the relatively homogenous
sandstone deposits at the scale of the sedimentary formation.
Schematic sedimentary section of the studied interval is presented in
Fig. 2.5. The facies descriptions, taken from Navarrete et al. (2013, 2014), take
into account large (km)‐scale sedimentary characteristics; however, many of
the sedimentary structures described can be observed at the Aliaga outcrop.
Brief descriptions of the facies associations observed at the Aliaga outcrop
section are given below.
(A) Tsunami deposit (Fig. 2.4 and 2.5). This facies consists of a
sandstone body of 1 to 5 m thick; the sediments were deposited on a clay
layer bearing dinosaur trackways (including a 7‐km‐long dinosaur
megatracksite). At least five sedimentary facies successions can be recognized
in the sandstone deposit, as related to inflows and back flows of the tsunami
wave train (Fig. 2.6). Each sedimentary succession is composed of three
lithofacies (from base to top): a conglomerate facies (LF1), a ripple and
megaripple facies (LF2) and an oscillation ripple facies (LF3).
(B) Tidal mud flat (FA1 in Fig. 2.5). This facies association is composed of
red and green mudstones, red fine‐grained sandstones and decimetre‐thick
tabular sandstones. The mudstones are generally massive and arranged in
metre‐thick tabular strata, occasionally showing slickensides and mottling. The
red fine‐grained sandstones are centimetre‐ to decimetre‐thick lenticular
layers, occasionally with irregular bioturbated bases and vertical burrows. The
sandstones pinch out towards the north and south; they exhibit cross‐
lamination, asymmetric ripple laminations, bivalve fragments and Arenicolites
and Skolithos ichnofossils. This facies association is interpreted as a tidal flat
environment in which the background sedimentation of fine‐grained particles
settled from suspension under low‐energy conditions. The centimetre‐ to
decimetre‐thick sandstones containing Arenicolites are interpreted as distal
washover fan deposits interbedded with mudflat mudstones.
32
Fig. 2.5: Schematic sedimentary section of the studied outcrop; the facies associatios, which are the object of this study,are red colored (modified from Navarrete et al., 2013).
(C) Carbonate lagoon (FA2 in Fig. 2.5). This facies association is
composed of carbonates and interbedded ochre‐ and white‐coloured
sandstones. The carbonates are limestones and marls containing bivalves,
gastropods and charophytes, which are interpreted as representing
sedimentation in a carbonate lagoon. The sandstones, which are fine‐grained
centimetre‐ to decimetre‐thick tabular strata thinning northwards and
southwards, pinch out in lagoonal sediments; they exhibit sub‐parallel and
undulating lamination, cross‐lamination, asymmetric ripple lamination and
local deformation as slump and flame structures. The sandstones are
33
interpreted as distal washover fan deposits that interrupted the background
low‐energy conditions conducive to carbonate sedimentation in the lagoon.
(D) Flood‐tidal delta (FA3 in Fig. 2.5). This facies association appears as
a fining‐upwards sandy body 6 m thick and 4 km long. In the study area, the
facies pinches out in the southern sector of the profile within the carbonate
lagoon facies. The sand body shows a generally lenticular geometry with a
sharp and flat horizontal base (with tool casts and locally horizontal
bioturbation) and a convex top. The facies association is interpreted as a flood‐
tidal delta comprising facies of flood ramp and flood delta origin. The facies
FA2 and FA3 are vertically related, suggesting a flood‐tidal delta encased in
mixed carbonate lagoonal deposits.
(E) Barrier island–tidal inlet (b.i./inlet) (FA4 in Fig. 2.4 and 2.5). This
facies association consists of a 6‐m‐thick lenticular body of very coarse‐ to fine‐
grained sandstone containing scattered oysters shells, fish teeth, clay and
quartzite pebble lags, vertebrate bones and metre‐long tree trunks at the
base. Internally, the sand body is divided by large‐scale planar to slightly
concave‐up lateral accretion surfaces, inclined to the northeast and with
aligned basal lags (clay, quartzite and minor bioclasts). The deposit is
interpreted as a multistory body representing a barrier island system;
furthermore, the occurrence of a channel body showing large‐scale accretion
surfaces is interpreted as a tidal inlet encased in a barrier spit.
The tsunami and b.i./inlet deposits are relatively homogeneous at the
scale of the sedimentary formation; they were recognized by Navarrete (2015)
in a 35‐km² area on both sides of the Camarillas–Jorcas syncline (Fig. 2.1C),
despite their limited thicknesses (1–7 m). The sandy homogeneity into both
deposits is dependent on the scale of observation. At the basin scale, the
deposits are homogenous sands with horizontal continuity, whereas at the
metre and even micron scales, the deposits show variability within the sandy
facies in terms of grain size distribution, lithic and fossil content, sedimentary
structures, etc. The lithologies of both deposits are here described at macro
(metre) and micro (micron) scales. In addition, hydrodynamic measurements
of porosity and permeability on the outcrop samples were performed to
complement the lithological descriptions.
34
Fig. 2.6: Vertical sedimentary architecture of Tsunami deposit (Navarrete et al. 2014).
Other facies associations between the tsunami and b.i./inlet deposits
(Fig. 2.5) consist of clay, marl and micritic carbonate lithofacies; these facies
associations were not included in this study, which focused on high‐resolution
characterization of heterogeneities in the sandstone, and the relationship of
these heterogeneities to porosity and permeability structures. The sandstone
deposits of the FA1 (Tidal mud flat with channels) and FA3 (Flood‐tidal delta)
facies associations were also excluded from this study because the facies are
laterally discontinuous and pinch out in the study area at the Aliaga outcrop.
2.2.1. Materials and methods
The logged sedimentary sections (see the outcrop panel in Fig. 2.3B)
were plotted on aerial photograph of Villarluengo 543‐12 (scale, 1:5000) (Fig.
2.4). Using the sedimentary sections and the aerial photo as geographic
references, a low aerial photograph of the outcrop, obtained by a camera fixed
to a drone (unmanned aerial vehicle), was georeferenced to the ED‐50 UTM‐30
geographic coordinate projection. Based on sedimentary interpretations of the
drone photographs and the sedimentary descriptions of Navarrete et al. (2013
35
and 2014), a sampling schema was developed to systematically collect samples
for petrographic analyses and petrophysical measurements according to
lithological variations, and to provide upcoming geostatistically analysis.
The Aliaga outcrop was divided into two sectors, the North sector and
the South sector, separated by the Aliaga–Miravete road (Fig. 2.4). From the
North to the South sectors, fifty‐three samples were collected at regularly
spaced intervals along the tsunami and b.i./inlet deposits using a portable rock
core drill (Pomeroy D026‐GT10 Gear‐Reduced Core Drill with a 48 cm3 Stihl
motor) and a 4” (10 cm) outer‐diameter drill bit (BS‐4Pro). In areas where the
thicknesses of deposits were >1 m, samples were collected at the base, middle
and top of each bed. Orientations of the extracted cores were measured; their
dimensions varied from 10 to 20 cm in length and 8 to 9.5 cm in diameter. The
cores were obtained parallel to the dip direction, and were taken on the
outcrop face due to constraints imposed by the orientation of the outcrop;
therefore core lengths correspond to bed width (Veloso et al., 2013). Samples
were taken from each core for petrographic and petrophysical analyses.
Lithofacies described in the field, based on the studies of Navarrete et
al. (2013 and 2014), were refined by petrographic descriptions so as to analyse
the grain size distribution, sorting, shape and nature of grains, matrix
components, sedimentary structures and the presence of fossils or cement.
Many samples could not be dissolved or disaggregated by conventional
methods used for sedigraph grain size distribution analyses due to the high
concentrations of carbonate clasts and fractured quartz grains; therefore,
grain size distributions, estimates of sand sorting and sand nature were
obtained by quick petrographic observations according to a reference chart
(U.S. GeoSupply Inc., 2015). In both sandstones, sixty‐five thin sections for
petrographic study were made from samples collected from the cores and
hand specimens. The thin sections were oriented perpendicular to bedding,
and were chemically stained to identify carbonate.
The petrophysical measurements included estimations of sample
porosity and permeability by direct measurements on plugs. Fifty‐six plugs
were taken from the cores for measurements of porosity (Phi) and horizontal
permeability (Kh); a further 23 plugs were used for measurements of vertical
permeability (Kv). The petrophysical measurements were conducted at the
36
Petrophysics Institute Foundation (IPF), Madrid, Spain, on plugs 60 mm long
and 40 mm in diameter. The plugs were cut from the cores as shown in Fig.
2.7, where the vertical plugs were taken perpendicular to the strike direction
or in the North direction, and the horizontal plugs were taken parallel to the
strike direction or the North direction. Porosity was estimated by a helium
pycnometer at atmospheric conditions and ambient temperatures through gas
displacement inside a known cell volume; the pore volume was obtained
according to Boley’s law (IPF, 2012). The horizontal (Kh) and vertical (Kv)
permeabilities were estimated using a gas permeameter at steady‐state
conditions; the gas permeability was calculated according to Darcy’s law and
was then corrected to the equivalent liquid permeability using the Klikenberg
correction factor (IPF, 2012).
Fig. 2.7: Schema for obtaining plugs for petrophysical measurements. The core length is equal to the width of deposit; “V” plugs were used to measure the vertical permeability and “H” plugs were used to measure the horizontal permeability; “T” and “B” correspont to the top and base of bedding; respectively.
37
2.2.2. Sandstone deposits at the macroscale
Macroscale (decimetres to tens of meters) descriptions of the geometry
of internal lithofacies and their distributions were obtained in 2D perspectives:
horizontal in N‐S direction and vertical (thickness). The lithofacies descriptions
of Navarrete et al. (2013, 2014) were used for macroscale facies mapping; the
mapping was based on interpretations of drone photographs and field
observations.
2.2.2.1. Tsunami deposit
The tsunami deposit, located at the base of the Aliaga outcrop (Fig. 2.4
and 2.5), consists of a multiple‐bed deposit of sandstone and conglomerate of
variable thickness (30–50 cm to 3–4 m). The base is eroded in many areas, and
the deposit rests on an important dinosaur trackway site. The tsunami deposit
represents a catastrophic sedimentary event with at least five episodes of
inflow and backflow of wave trains (Fig. 2.6), related to sediment deposition
and reworking, respectively (Navarrete et al., 2014). The lithofacies of the
tsunami deposit is described in Navarrete et al. (2014). The sedimentology and
architecture of the track‐bearing sandstone are based mainly on field
observations made from 20 stratigraphic sections logged in detail, and
interpretations from drone photographs. Sedimentological features, such as
grain size and sedimentary structures, allowed the discrimination of four
lithofacies (Navarrete et al., 2014), with three being present at the Aliaga
outcrop (Fig. 2.6): a conglomerate facies (LF1), a ripple and megaripple facies
(LF2) and an oscillation ripple facies (LF3).
Lithofacies LF1 is a conglomeratic facies with matrix‐supported clasts of
greenish carbonate pebbles, and plant and dinosaur bone fragments in a
whitish coarse‐grained sandy matrix of feldspar and quartz grains along with
minor zircon, rutile, tourmaline, glauconite and apatite grains (Fig. 2.8A, B and
C). The LF1 lithofacies is recognized mainly at the base of deposit, as dinosaur
track cast infill (Fig. 2.8A), where tracks are present. In other zones of the
North sector, this facies is locally preserved in the middle of the bed at the
base of an intermediate sedimentary succession, as in Fig. 2.8B, Fig. 2.9.
38
Fig. 2.8: Field view of Tsunami lithofacies. (A) The lithofacies LF1 occurs as dinosaur cast infill (Section 2, Fig. 2.3B). (B) Field view of the sedimentary succession in section 2 (Fig. 2.3B) showing the lithofacies arrangements in the sedimentary successions. Palaeocurrent indicators in lithofacies LF2 indicate flow towards the SSE. (C) Lithofacies LF1, showing cross‐stratification and a palaeocurrent direction towards the NNE (Section 2, Fig. 2.3B). (D) Detail of Fig. 2.9D, showing lithofacies LF1 and LF2. (E) Detail of lithofacies LF2 in Section 2 (Fig. 2.3B), showing a fining upward sequence with centimetre‐scale planar cross‐stratification; the palaeocurrent direction is towards the SSE.
39
The LF1 lithofacies is generally massive, although locally it shows faint cross‐
laminations, horizontal laminations and fining upward (Fig. 2.8C). Cross‐
stratification planes measured in sedimentary sections 3 and 4 (Fig. 2.4)
indicate a N–NNE palaeo‐flow direction (Fig. 2.8C). The geometry of the LF1
lithofacies is usually lenticular, with a length of 1–10 m and thickness of 2–15
cm (Fig. 2.8B and D and Fig. 2.9). The drone photographs show that the
lithofacies occurs in zones where the thickness of the lithofacies is >5 cm and
the length is >1–2 m.
The ripple and megaripple lithofacies (LF2) is the most abundant facies
in the both sectors of outcrop (Fig. 2.6, 2.8 and 2.9). It is composed of fine to
coarse feldspar and quartz sand with fining‐upward grain‐size sequences and
centimetre‐ to decimetre‐scale sets of cross‐bedding and planar cross‐bedding
(Fig. 2.8E); its geometry is wedge‐shaped, with a thickness of 15–160 cm (Fig.
2.8D and 2.9).
Lithofacies LF3 is scarce and is composed of fine to medium feldspar
and quartz sand with an upward decrease in grain size; its thickness at the
Aliaga outcrop is 5–28 cm and its contact with LF2 is gradual (Fig. 2.6 and
2.8B). This facies shows asymmetrical climbing ripples and cross‐laminations
and parallel laminations.
The tsunami sandstone bed has the geometry of pseudo‐sand sheet
deposit, with an irregular basal surface. A preliminary map of the macro facies
was drawn on a drone photograph to delimit the geometry of the lithofacies
deposit. Macrofacies LF2 and LF3 are similar in colour and were difficult to
distinguish on the drone photographs (Fig. 2.9).
2.2.2.2. Barrier island ‐ tidal inlet deposit
The barrier island ‐ tidal inlet (b.i./inlet) deposit is located at the top of
outcrop profile (Figs. 2.3, 2.4 and 2.5). This deposit comprises a 6‐m‐thick
lenticular body of ochre‐coloured very coarse‐ to fine‐grained sandstone with
scattered oysters shells, fish teeth, pebbles, and clay and quartzite pebble lags.
In the study area, Navarrete et al. (2013) interpreted the deposit as a tidal inlet
sandstone encased in a barrier spit that recorded changes in flow energy and
40
Fig. 2.9: Tsunami deposit in a drone photograph and a map of the lithofacies as identified in the sedimentary sections and described in Navarrete et al. (2014). Drillcores are indicated by a star. (A) Excerpt of the South sector. (B) Excerpt of the North sector; the square labelled D indicates the area shown in Fig.2.8D.
41
interruptions in water flow; these conditions give the deposit different scales
of heterogeneity that can be observed by comparing the North and South
sectors of the outcrop.
In the South sector, this deposit is characterised by a 3 to 7 m‐thick
lenticular body of lihofacies LF4 (Fig. 2.10A and B); this lithofacies is an ochre‐
coloured medium‐ to fine‐grained sandstone with decimetre‐thick trough
cross‐bedding sets (maximum thickness, 35 cm) and centimetre‐thick planar
cross‐bedding sets (maximum thickness, 8 cm). The basal surface is slightly
concave and includes metre‐long tree trunks and vertebrate bone fragments;
the top surface is flat (Fig. 2.11A).
In the North sector, two main lithofacies (1 to 3 m thick) are
distinguished (Fig. 2.11B). Lithofacies LF4 is an ochre‐coloured coarse‐ to fine‐
grained sandstone with decimetre‐thick trough cross‐bedding sets and clay
and quartzite pebble lags (Fig. 2.11B and 2.10). Lihofacies LF5 is a grey very‐
fine‐grained cemented sandstone with centimetre‐thick trough cross‐
lamination sets, drapes, locally interfering ripple crests and asymmetric wave,
ripples on top of the cross‐bed sets (Fig. 2.10C), scattered oysters, fishes teeth
and centimetre‐thick accumulations of bioclast or carbonaceous plant
fragments (Fig. 2.10D).
Navarrete et al. (2013) measured the palaeocurrent directions
throughout the Miravete anticline, and reported unidirectional flow
representing a flood into the barrier island system. The authors interpreted
major bedding surfaces inclined towards the NE as tidal inlet accretion
surfaces developed by primary longshore drift currents.
At the outcrop scale, the b.i./inlet deposit shows heterogeneous
lithofacies distributions and deposit geometries between the North and South
sectors. In the South sector (Fig. 2.11A), the deposit is thicker (3–7 m) and
lithofacies LF4 is dominant. This lithofacies was generated by the migration of
minor megaripples that were moved by flood and ebb water fluxes under low‐
energy flow regimes in the shoreface zone of a tidal inlet/barrier spit
(Navarrete et al., 2013). In the North sector (Fig. 2.11B), the thickness of the
42
Fig. 2.10: Field view of the barrier island ‐ inlet facies. (A) and (B) Lithofacies LF4 in the South sector showing decimetre‐thick trough cross‐bedding and planar cross‐bedding sets. (C) Lithofacies LF4 and LF5 in the North sector showing centimetre‐thick trough cross‐bedding laminations and asymmetric wave structures (arrow). (D) Plant fragments in sample 59 (see in Fig.2.12B).
43
deposit is 1–3 m and two facies are dominant: lithofacies LF4 with bioclasts
and pebbles, and lithofacies LF5, a grey cemented sandstone with drapes and
accumulations of bioclasts or plant fragments, and asymmetric wave and
interference ripples structures. These structures in LF5 indicate variations in
the flow regime, probably associated with tidal flows from brackish to marine
conditions (Navarrete et al., 2013).
2.2.3. Sandstone deposits at the microscale
The petrological characterization of the lithofacies was based on
analyses of 65 thin sections of samples obtained from cores and hand
specimens. The classification used for the mineralogical analyses was based on
that of Pettijohn et al. (1973). The estimates of grain size distribution and sand
sorting, as well as of bulk composition, were made by quick petrographic
observations, according to the reference chart of U.S. GeoSupply Inc. (2015).
A preliminary analysis of samples of tsunami and b.i./inlet deposits
showed that nearly every sample was an arkosic or subarkosic sandstone, with
an average composition of 10% clay matrix, 60% quartz (monocrystalline and a
few polycrystalline samples), 20% feldspar (plagioclase and K‐feldspar) and
10% lithic fragments. This mineralogical homogeneity excluded a mineralogical
classification of the lithofacies, and thus the lithofacies were classified as a
function of grain size distribution, sand sorting, and cement content.
2.2.3.1. Microfacies of the tsunami deposit
The petrological characterization of the tsunami lithofacies was based
on analyses of 36 thin sections (Table 2.1). The locations of these samples
along the outcrop are shown in Fig. 2.12. All samples are subarkosic–arkosic
sandstones, containing 10% clay matrix, 50%–70% quartz, 10%–20% feldspar
and 5%–15% lithic fragments with mud clasts and micas as the main lithic
grains. Kaolin and calcite are the most common cements (Caja, 2004); their
44
Fig. 2.11: D
rone photographs of the b.i./inlet deposit. A: South sector. B: N
orth sector.
45
contents vary from trace amounts to 10% (Fig. 2.13A); a few samples contain
trace amounts of dolomite cement (Fig. 2.13D). The main accessory minerals
are tourmaline and trace amounts of phosphate and organic matter.
Sedimentary structures, where present, consist of trough cross bedding with
an apparent direction to North. The upward decrease in grain size is apparent
at the microscale (Fig. 2.13A). Fractured quartz and feldspar grains are
abundant in many samples (Fig. 2.13A and B). The red colour of the sandstone
in the southernmost part of the South sector profile (Fig. 2.9A and 2.12) is
probably related to the relatively high K‐feldspar content (sample 83, Fig 2.9A
and Fig. 2.13B).
The samples were first classified as a function of sorting (Table 2.2 and
Fig. 2.12); four sorting facies were defined: (0) poorly sorted sandstone (PSs
facies), (1) moderately sorted sandstone (MSs facies), (2) cemented sandstone
(CSs facies) and (3) well‐sorted sandstone (WSs facies). Samples with bimodal
grain‐size distribution were classified as MSs facies, and those with a cement
content > 10% as CSs facies. Table 2.2 summarizes the number of samples by
sorting class and the equivalent lithofacies at the macroscale. Half of the
samples were classified as WSs facies; in the South sector, almost every
sample was WSs facies, whereas all facies were present in the North sector
(Fig. 2.12). The cemented samples were found in both sectors at the base,
middle and top of the profile, and comprised the WSs and PSs facies (Fig. 2.12,
Table 2.1). The dinosaur footprints were sampled infrequently, and are
generally filled with PSs facies.
Samples were also classified as a function of grain size distribution
following Folk (1980). They were grouped into five grain‐size facies classes:
coarse grained, medium–coarse grained, medium grained, fine–medium
grained and fine grained (Table 2.3). The equivalent grain‐size facies classes
and macroscale lithofacies are listed in Table 2.3. Half of the samples were
classified as fine‐ and fine‐ to medium‐grained sand, three samples were
classified as coarse‐grained sand, and the others were classified as fine‐ to
coarse‐grained sand. Generally, the fine‐grained sands are well (WSs) or
moderately sorted (MSs) sands, and the coarse‐grained sands are poorly
sorted (Pss; Table 2.3); coarser samples are located in the North sector (Fig.
2.12).
46
Fig. 2.12: D
rone photographs showing the spatial distribution of collection localities of tsunami samples, classified according degree of sorting
(see legend).The photographs extend from the South (upper photograph) to the North (lower photograph) sector.
47
The tsunami facies are relatively homogeneous when classified in terms
of sorting, with 70% of the facies classified as WSs and MSs facies. In terms of
grain size, almost half the tsunami samples are classified as fine‐ to medium‐
grained, and the other half are classified as coarse‐grained, medium‐ to
coarse‐grained or medium‐grained. Generally, the lithofacies at the
macroscale are more correlated with grain size facies classes than sorting
facies classes; however, the sorting facies classes allow identification of
heterogeneities between sector: in the South sector, the WSs facies is
dominant (Table 2.1), while in the North sector, all facies are present (Fig.
2.12).
Table 2.1: Tsunami samples and their petrographic and petrophysical characteristics. Table headings: Phi is measured porosity; Kh and Kv is measured horizontal and vertical permeabilities respectively; %Clast is the sum of quartz, feldspar and lithics grains; %Cement is the amount of cement content which can be calcite, kaolin and dolomite; Sorting is the sand sorting at microscopic scale; Grain Size is the average sand size at the microscopic scale; Facies is the code to distinguish sorting facies class; and Sector is the location of sample in the outcrop face.
Sample
Phi
(m³/m³)
Kh (mD)
Kv (m
D)
% Clast
(Qtz+F+L)
% cement
Sorting
Grain Size
Facies
Sector
1 80 well medium 3
North
2 80 3 Moderate (bi‐modal)
fine‐coarse 3
3 80 3 well fine‐medium 3
5 50 50 moderate medium‐coarse 2
41 0.197 13.9 14 65 7 well fine‐medium 3
42 0.0663 0.17 65 7 moderate medium‐coarse 1
43 0.076 2.1 70 5 moderate medium‐coarse 1
44 0.1676 7.91 85 well medium 3
45 0.1859 10.3 85 well fine‐medium 3
46 0.1272 85 3 poor coarse‐very
coarse 0
47 0.1795 7.84 85 3 well medium 3
48 0.1685 10.45 70 7 poor fine‐medium 0
49 0.184 14.24 10.89 85 3 moderate fine‐medium 1
50 0.171 5.15 4.57 70 10 well fine‐medium 2
51 0.164 4.83 80 3 moderate medium‐coarse 1
48
Sample
Phi
(m³/m³)
Kh (mD)
Kv (m
D)
% Clast
(Qtz+F+L)
% cement
Sorting
Grain Size
Facies
Sector
52 0.176 6.55 80 3 well fine‐medium 3
53 0.173 7.77 7.73 65 10 poor fine‐coarse 2
54 0.143 5.17 65 10 poor medium‐coarse 2
55 0.189 8.09 7.7 70 7 moderate medium‐coarse 1
56 0.184 7.47 75 3 well fine‐medium 3
57 0.179 6.5 75 3 poor fine‐coarse 0
23 70 10 moderate medium‐coarse 2
South
27 80 5 well fine 3
28 0.111 1.57 1.96 65 10 well fine 2
83 0.187 14.96 80 5 well medium 3
84 0.204 26.65 20.89 70 5 moderate medium 1
85 0.166 5.91 7.02 85 5 well medium 3
86 0.213 28.76 24.09 75 7 well medium‐coarse 3
87 0.196 11.04 75 10 well fine‐medium 3
88 0.187 10.19 11.14 75 7 well fine‐medium 3
89 0.163 8.29 21.21 70 10 well fine‐medium 2
90 0.171 9.25 9.06 70 10 well fine‐medium 2
91 0.156 5.48 80 7 well fine‐medium 3
92 0.168 8.14 75 10 well fine 3
93 0.176 6.99 75 7 well fine 3
94 0.143 2.99 75 7 well fine‐medium 3
‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐
Fig. 2.13: Thin sections of tsunami samples (next page). (A) Sample no. 2: WSs facies showing upward‐decreasing sand size from medium to fine sand. Cross polarized light. (B) Sample no. 83: WSs facies showing a medium sand size, with the K‐feldspar content increasing towards the top of the sample. Plane polarized light. (C) Sample no. 28: Css facies showing a fine sand size and good sorting. Plane polarized light. (D) Sample no. 46: Pss facies showing traces of dolomite cement and fractured quartz. Plane polarized light.Location of samples is given in Fig. 2.12.
49
50
Table 2.2: Tsunami facies classes based on degree of sorting and equivalent lithofacies. Table variables: Name is the abbreviation of facies classification by sorting; Lithofacies macroscale is the equivalent facies described by Navarrete et al. (2014); North and South are the quantity of samples in each sector; and Total is the sum of samples by facies.
Tsunami
Code NameLithofacies macroscale
North South Total
0 PSs LF1 3 0 3
1 MSs LF2 6 2 8
2 CSs LF1 or LF2 4 3 7
3 WSs LF2 or LF3 8 10 18
total 21 15 36
Table 2.3: Tsunami classification based on grain size and equivalent sorting facies class and macroscale lithofacies. ‘Sorting class’ for a given class are listed in order of abundance.
Number of samples
Class North South Total Sorting class Lithofacies macroscale
Coarse 4 0 4 PSs LF1
Medium‐coarse 6 2 8 MSs,WSs, PSs LF2
Medium 3 3 6 Wss, MSs LF2
Fine‐medium 8 6 14 WSs,MSs,PSs LF2
Fine 0 4 4 WSs LF3
Total 21 15 36
2.2.3.2. Microfacies of the barrier island ‐ tidal inlet deposit
The petrological characterization of the barrier island ‐ tidal inlet
(b.i./inlet) lithofacies was based on observations of 30 thin sections (Table 2.4)
obtained from cores and hand specimens. The locations of the samples in the
outcrop are shown in Fig. 2.11. All samples are subarkosic–arkosic sandstones
consisting of 5%–15% clay matrix, 50%–70% quartz, 10%–20% feldspars, and
10%–25% lithic fragments dominated by mud clasts and bioclasts. Tourmaline
is the most abundant accessory mineral, followed by opaque minerals and
phosphate. Sedimentary structures, when present, are trough cross bedding
with apparent N–S orientations. Some samples have a bimodal grain‐size
51
distribution (Fig. 2.14A). Fractured quartz grains are less numerous than in the
tsunami deposit, and authigenic kaolinite is locally present Fig. 2.14B). The
cement is principally dolomite and locally calcite or kaolin Fig. 2.14C), and it
generally makes up less than 10% of the whole rock (Fig. 2.14), except in some
samples in the North sector from the LF5 lithofacies, where the matrix has
been entirely replaced by dolomite cement (Fig. 2.14D).
Samples were classified by the degree of sorting Table 2.5), based on
the classification used in Section 2.2.3.1, into the PSs, MSs, WSs and CSs facies.
Samples with bimodal grain‐size distribution were classified as MSs facies, and
cement samples were considered as those with a cement content that made
up >10% of the total rock. The spatial distribution of samples classified by
sorting facies class is heterogeneous in the deposit (Fig. 2.11 and Table 2.5);
nearly all samples classified as WSs and MSs are located in the South sector of
the profile, whereas most samples in the North sector were classified as PSs
sorting facies class (Table 2.5 and Fig. 2.11). The CSs facies is only present in
the North sector (Table 2.4, Fig. 2.11); CSs facies samples are sub‐classified as
MSs and PSs facies.
Samples were also classified by grain size facies class based on the
classification outlined in Section 2.2.3.1. The spatial distribution of samples
classified in different grain size classes shows that samples of the fine–medium
facies class occur in the South sector (Table 2.6), whereas those from other
grain size facies classes are regularly distributed throughout the deposit (Table
2.6). Generally, samples of fine–medium and medium grain size facies classes
are also classified as WSs and MSs sorting facies classes, and samples of coarse
or medium–coarse grain size classes are also classified as PSs sorting facies
(Table 2.6).
The heterogeneity of the sandy facies distribution is more apparent
when samples are classified by sorting facies class rather than grain size facies
class. A clear difference between the sectors is observed in the abundances of
the WSs and MSs facies in the South sector, whereas the PSs and CSs facies are
abundant in the North sector. Samples classified by grain size facies class show
a homogeneous distribution in the deposit, except those of the fine–medium
grain size class, which are more abundant in the South sector.
52
Table 2.4: B.i./inlet samples and their petrographic and petrophysical characteristics. Table headings: Phi is measured porosity; Kh and Kv is measured horizontal and vertical permeability respectively; %Clast is the sum of quartz, feldspar and lithics grains; %Cement is the amount of cement content which can be calcite, kaolin and dolomite; Sorting is the sand sorting at microscopic scale; Grain Size is the average sand size at microscopic scale; Facies is the code to distinguish sorting facies class; and Profile sector is the location of sample in the outcrop face.
Samples
Phi
(m³/m³)
Kh (mD)
Kv (m
D)
% Clast
(Qtz+F+L)
% cement
Sorting
Grain Size
Facies
sector
18 70 5 moderate Fine‐medium 0
North
19 70 10 moderate fine 2
20 70 10 moderate fine_medium 2
39 0.115 0.9 0.5 60 5 poor medium 0
60 0.1545 8.24 9.58 70 7 poor coarse 0
61 0.165 14.42 75 3 poor medium‐coarse 0
62 0.0236 0.01 50 40 moderate medium 2
63 0.1489 4.28 2.85 65 7 poor medium‐coarse 0
64 0.1492 3.53 75 3 poor medium‐coarse 0
64.5
0.0239 0.02 50 30 poor fine‐medium 2
65 0.064 0.09 70 7 poor medium‐coarse 0
66 0.1127 0.7 75 3 moderate medium‐coarse 1
36 0.21 16 16.5 90 3 well fine_medium 3 South
37 0.14 2 4.2 70 5 poor fine‐medium 0
67 0.1947 8.87 85 3 moderate medium‐coarse 1
68 0.2028 11.23 70 5 moderate medium‐coarse 1
69 0.1804 9.6 75 5 moderate medium‐coarse 1
70 0.1516 6.79 85 moderate fine‐medium 1
71 0.1945 15.33 12.14 80 7 well medium 3
72 0.1803 9.25 14.2 80 5 well medium 3
73 0.1793 6.79 85 5 poor fine‐medium 0
74 0.1871 6.64 7.38 80 5 poor fine‐medium 0
75 0.2086 11.9 90 3 moderate medium‐coarse 1
76 0.1842 7.59 75 3 well fine‐medium 3
77 0.1886 8.25 80 5 moderate coarse 1
78 0.1864 9.08 8.3 75 7 moderate medium‐coarse 1
79 0.1778 8.1 8.51 75 5 well fine‐medium 3
53
Samples
Phi
(m³/m³)
Kh (mD)
Kv (m
D)
% Clast
(Qtz+F+L)
% cement
Sorting
Grain Size
Facies
sector
80 0.1135 2.03 70 7 moderate medium‐coarse 1
81 0.2175 12.89 14.63 80 3 well fine‐medium 3
82 0.1279 5.72 70 7 moderate medium‐coarse 1
Table 2.5: B.i./inlet sorting facies classes and equivalent lithofacies
Code Lithofacies macro‐scale
Name North South Total
0 LF4 or LF5 PSs 6 3 9
1 LF4 or LF5 MSs 2 9 11
2 LF4 CSs 4 0 4
3 LF4 WSs 0 6 6
total 12 18 30
Table 2.6: B.i./Inlet grain size facies classes and equivalent sorting facies classes and macroscale lithofacies.
Number of samples
Class North South Total Sorting class Lithofacies macroscale
Coarse 1 1 2 PSs, MSs LF4
Medium‐coarse 5 7 12 MSs, PSs LF4
Medium 2 2 4 WSs, MSs, PSs LF4 or LF5
Fine‐medium 3 8 9 WSs, PSs, MSs LF4 or LF5
Fine 1 0 0 MSs LF5
Total 9 18 27
54
55
2.3. Petrophysics
Fifty‐six plugs for porosity (Phi) and horizontal permeability (Kh)
measurements, and 23 plugs for vertical permeability (Kv) measurements were
taken from drilled cores. All samples were classified by sorting facies class and
grain size facies class (Table 2.1 and Table 2.4). For all samples (from both
deposits and sectors), the range of porosity (Phi) values was 2%–22% and the
permeability values were 0.01–28 mD. A plot of Phi versus Kh for all samples
(Fig. 2.15) shows a strong correlation (correlation coefficient, r2, of 0.8 and 0.9
for the tsunami and b.i./inlet regressions, respectively). In both deposits, Kv is
strongly correlated with Kh (r2 = 0.85) (Fig. 2.16); therefore, the permeability is
considered to be isotropic, and no further analyses of Kv were performed in
this study. The Phi and Kh values of the tsunami lithofacies are usually >14%
and 3 mD, respectively, whereas the Phi and Kh values of the b.i./inlet deposit
vary widely.
2.3.1. Petrophysics of the tsunami deposit
The porosity (Phi) and horizontal permeability (Kh) of tsunami samples
were measured on 30 plugs; the vertical permeability (Kv) was measured on 12
of these plugs. Samples were classified as a function of sorting facies class and
grain size facies class (Table 2.1). The Phi versus Kh plot (Fig. 2.17) of tsunami
samples shows porosity varying from 14% to 22% (as outlined by the ellipse in
Fig. 2.17). Samples were classified by their sorting facies class, and cemented
samples were also sub‐classified according to their sand sorting.
‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐
Fig. 2.14 (previous page): Photomicrographs of b.I./inlet samples. (A) Sample no. 75 (South sector): MSs facies showing apparent bi‐directional palaeocurrent directions. Plane polarized light. (B) Sample no. 79 (South sector): WSs facies showing fine sand and authigenic kaolinite. Cross polarized light. (C) Sample no. 68 (South sector) from the MSs facies showing clay matrix replaced by dolomite. Plane polarized light. (D) Sample no. 62 (North sector) from the CSs facies with matrix entirely replaced by dolomite. Plane polarized light. (E) Sample no. 66 (North sector) from the PSs facies showing very coarse grain sizes. Cross polarized light.
56
Fig. 2.15: Phi versus Kh for all samples. Red squares, tsunami samples; blue circles, B.I./inlet samples.
Fig. 2.16: Horizontal permeability (Kh) versus vertical permeability (Kv) for all samples.
57
Despite cementation in some samples, all Ph vs. Kh values lie within a
narrow range. A comparison of facies in the North and South sectors (Fig.
2.18A and B) shows that samples from the South sector are comprised of the
WSs facies class whereas those of the North sector belong to the Wss and Mss
facies classes, with some PSs facies samples. The values of Phi and Kh in North
sector samples are lower for the MSs facies than for those in the South sector;
values of Phi and Kh for samples of the PSs facies class lie within the ellipse
shown in Fig. 2.17.
Fig. 2.17: Phi versus Kh plot for the tsunami samples, classified by sorting facies. Except for three samples, the porosity values lie in the range 0.14–0.22 (blue ellipse). The cemented samples (CSs class) are shown by pale red symbols and sub‐classified by their sand sorting.
Samples on the Phi vs. Kh plot (Fig. 2.19) are classified according to
their grain size facies class. Two samples that fall outside of the ellipse in Fig.
2.19 belong to the medium–coarse grain size class; one sample belonging to
the fine grain size class is also classified as the CSs facies. In the South sector
(Fig. 2.20A), samples are principally classified in the fine‐medium grain size
class. In the North sector (Fig. 2.20B), half of the samples are classified in the
medium‐coarse grain size class. The sorting and grain size classes are
correlated; generally, finer grain size classes correspond to the WSs and MSs
sorting facies classes, and the coarser grain size classes correspond to the PSs
58
and MSs facies classes (Table 2.3). Both classifications show the heterogeneity
of the sandy facies class distribution between the North and South sectors.
Fig. 2.18: Phi versus Kh plot of tsunami samples, classified by sorting facies. (A) Samples from the South sector. (B) Samples from the North sector. The cemented samples are shown by pale red symbols according to the CSs class and sub‐classified by their sand sorting.
59
Fig. 2.19: Phi versus Kh plot for tsunami lithofacies, classified by grain size facies class.
2.3.2. Petrophysics of the barrier island ‐ tidal inlet samples
The porosity (Phi) and horizontal permeability (Kh) of barrier island ‐
tidal inlet (b.i./inlet) samples were measured on 26 plugs; the vertical
permeability (Kv) was measured on 11 of these plugs (Table 2.4). The Phi
versus Kh plot of samples classified by sorting facies (Fig. 2.21) shows two
categories of values, representing low‐porosity (<17%) and high‐porosity
(>17%) samples. The high Phi values are mainly classified as WSs and MSs
sorting facies, while the low porosity values are mainly classified as PSs, MSs
and CSs sorting facies; the CSs samples show the lowest Phi and Kh values. The
porosity of Wss facies is >17%, and the porosity of MSs facies falls into two
groups: half of the samples show values of 10%–15%, while the other half
show values of >17% (Fig. 2.21).
Samples are plotted separately for each sector profile in Fig. 2.22. In the
South sector (Fig. 2.22A), permeability values are >1 mD and porosity values
are >10%, with MSs and WSs being the dominant sorting facies classes. In the
North sector (Fig. 2.22B), the Pss facies is dominant and porosity values are
<17%; permeability values are <10 mD, and half of the values are <1 mD.
60
Fig. 2.20: Phi versus Kh plot for the tsunami lithofacies, classified by grain size facies class. (A) South sector. (B) North sector.
61
Fig. 2.21: Phi versus Kh plot for the b.i./inlet lithofacies, classified by sorting facies class. The cemented samples are shown by pale red symbols according to the CSs class and sub‐classified by their sand sorting.
Classification of the samples by grain size facies class (Fig. 2.23) shows
two main grain size classes, medium–coarse and fine–medium facies class,
representing samples with high Phi values (<17%). Plotting the samples for
each sector (Fig. 2.24) shows that in the South sector (Fig. 2.24A) Phi values
are >10% and Kh values are >1 mD, and the dominant facies are fine–medium
and medium–coarse facies classes. In the North sector (Fig. 2.24B), Phi is <17%
and the dominant facies is the medium–coarse facies class.
A comparison of the Phi vs. Kh relationship for samples classified by
sorting facies class (Fig. 2.21) and those classified by grain size facies class (Fig.
2.23) shows that the correlation is stronger for the sorting facies class, and
that Phi values fall into at least three ranges that define the respective sorting
facies classes: WSs, PSs and CSs. The porosity of the WSs facies is 17‐22%; the
porosity of the PSs facies, which is mainly located in the North sector (Fig.
2.24B), is 6‐16%, while porosity values are lowest for the CSs facies (< 5%). The
porosity values for the MSs facies are 11%–21%.
62
Fig. 2.22: Phi versus Kh plot for the b.i./inlet lithofacies classified by sorting facies class. (A) South sector. (B) North sector.
63
Fig. 2.23: Phi versus Kh plot for the b.i./inlet lithofacies, classified by grain size facies class.
2.4. Discussion
The Aliaga outcrop is an 8400‐m² 2D vertical face in a transitional
interval in the upper part of the Camarillas Formation (Fig. 2.1D). The present
study examined two sandstone deposits over a distance of 200m‐long and
attempted to establish a correlation between sedimentary heterogeneity and
the petrophysical properties of the sandstones. The principal sedimentary
heterogeneities within and between the deposits were variability in grain size,
sorting of sand, and the presence of cement. The tsunami and b.i./inlet
deposits were described at macro (decimetre–metre) and micro (micron–
millimetre) scales, and the sandy facies were analyzed as a function of their
porosity and permeability, as measured on plugs obtained from drilled cores.
Porosity and permeability are strongly correlated in both sandstone deposits
(Fig. 2.15).
At the macroscale, the tsunami deposit exhibits three main lithofacies
that are homogeneously distributed within the deposit (Fig. 2.8). In drone
photographs, two of the lithofacies (LF2 and LF3) are indistinguishable from
each other, and the other (LF1) is recognized wherever its thickness exceeds 5
cm and its length is >1–2 m (Fig. 2.9). Lithofacies LF1 usually fills dinosaur
tracks at the base of the North sector profile (Fig. 2.9B); this lithofacies is also
64
preserved in an intermediate sedimentary succession at specific zones (Fig.
2.8B and D) without lateral continuity. Lithofacies LF2 is (Fig. 2.8B, D and E)
dominant in both sector. At the microscale, the tsunami deposit is composed
of subarkosic–arkosic sandstones, with 70% of the samples belonging to the
WSs and MSs sorting facies classes (Table 2.2); almost half the samples are
Fig. 2.24: Phi versus Kh plot for the b.i./inlet lithofacies, classified by grain size facies class. (A) South sector profile. (B) North sector profile.
65
classified in the fine–medium grain size class and the other half vary between
medium, medium–coarse and coarse grain size classes (Table 2.3). Generally,
the lithofacies at the macroscale are better correlated with grain size class
than with sorting facies class; however, the sorting facies class allows the
identification of clear heterogeneities in facies between sectors. That is, in the
South sector WSs is the dominant facies (Fig. 2.12A) whereas in the North
sector the samples are little heterogeneous and every facies is present,
principally the WSs and MSs facies (Fig. 2.12B).
The distribution of sand in onshore tsunami deposits is complex, as
tsunami run‐up processes and the preservation of sediment depend on factors
such as seabed and nearshore topography, and sediment source areas
(Nanayama and Shigeno, 2006; Dawson and Stewart, 2007; Sugawara et al.,
2008; Martini et al., 2010; Apotsos et al., 2012; Phantuwongraj and Choowong,
2012; Sugawara et al., 2014). At geological time scales, tsunami deposits occur
instantaneously, as single sedimentary events (Fujiwara and Kamataki, 2007).
According to the sedimentological and architectural features described by
Navarrete et al. (2014), the studied tsunami deposit is an onshore multiple‐
bed tsunami sandstone (Fujiwara, 2008), in which the five succession of fining‐
upwards sequences corresponds to successive incursions and outflows of the
tsunami wave train. The onshore tsunami sediments accumulated on back‐
barrier mudflats, along with lagoonal sediments, and filled and covered
exceptionally well‐preserved dinosaur footprints. During sediment deposition,
the Aliaga outcrop area was located approximately 5–8 km from the coastline.
The first tsunami wave inflow caused little erosion, although the wave train
removed large volumes of sand from the barrier island. Approximately 70% of
the palaeocurrents measured from cross‐stratification planes are oriented to
the SSE, and represent backflows of the tsunami wave train.
At the Aliaga outcrop, the upper layers of the tsunami deposit are
relatively fine‐grained and thick; in addition, the higher beds are more widely
distributed than the lower beds. These characteristics are similar to those
identified by Nanayama and Shigeno (2006) in onshore tsunami deposits
caused by a magnitude 7.8 earthquake along the Japan coast in 1993.
Nanayama and Shigeno identified two lithofacies, a gravel lobe facies (GLF)
and a sand‐sheet facies (SSF), organized into two fining‐upwards sequences,
66
and whose upper sequence is more widely distributed and thicker than the
lower sequence. The upper sequence is coarser grained than the lower
sequence, as the second run‐up was larger than the first.
Matsumoto et al. (2010) described the sedimentary characteristics of
deposits of the 2004 Indian Ocean tsunami in and around Periya Kalapuwa
Lagoon, Sri Lanka. The Kalapuwa tsunami deposits consist of well‐sorted fine–
coarse sand. The mean grain size of the Kalapuwa tsunami deposits (as
determined from core samples) is 0.64–2.35 φ (average, 1.06 φ) and mud
contents are generally low (average, 0.61 vol. %). Generally, thick‐bedded (>10
cm) tsunami deposits showed single or multiply graded bedding structures.
The thickness of deposits decreases landward (westward) from 66 to 5 cm
over a distance of 165 m. The landward extent of the tsunami deposits is
greater along two transects near two inlets (~1000 m long in each case),
suggesting that tsunami deposition was controlled by runup flow through the
inlets. Moreover, the sediment source was mainly scoured from sand dunes
near the two inlets, as the grain‐size distribution of the sand dunes is similar to
that of the tsunami deposits. Some characteristics described in Kalapuwa
tsunami deposits are similar to the Aliaga tsunami, such as the sediment
composition which is mainly composed of well‐sorted sand, independently of
the sand size, and the sediment source, which was from erosion of the sand
dunes.
In the Aliaga outcrop, samples from studied tsunami deposit are
relatively homogeneous in terms of the sorting facies class, with 70% of the
samples classified as WSs and MSs, and in terms of porosity values (14%–22%);
this homogeneous texture of the sand facies in the deposit is independent of
grain size (Matsumoto et al., 2010; Sun et al., 2007). The textural homogeneity
is probably related to:
‐ The homogeneous nature and great sediment supply from barrier island;
‐ Flatling deposition surface composed of mud and carbonate lagoon
sediments;
‐ Large distance from coast plain;
‐ Limited erosion of the deposited sediments by tsunami wave;
67
‐ Single sedimentary event at the geological time scale, principally deposited
by backflows currents of turbidity flow. Deposit was not reworked or eroded
by overlaid sediments.
The studied b.i./inlet deposit is an inlet deposit in a back barrier system
(Navarrete et al., 2013). At the macroscale, two main lithofacies are
recognized in the studied deposit, and lithofacies heterogeneity is clear. In the
South sector (Fig. 2.11A), LF4 is the dominant lithofacies (an ochre‐coloured
medium‐ to fine‐grained sandstone with decimetre‐thick trough cross‐bedding
and centimetre‐thick planar cross‐bedding sets). In the North sector (Fig.
2.11B), two lithofacies are present, LF4 and LF5; LF5 is a grey cemented
sandstone with local drapes and accumulations of bioclasts or plant fragments.
At the microscale lithofacies are composed of subarkosic–arkosic sandstones,
and differences between the North and South sectors are found in the
distribution of sorting facies classes (Fig. 2.11); the WSs and MSs facies are
abundant in the South sector Fig. 2.11A) whereas the PSs and CSs facies are
abundant in the North sector (Fig. 2.11B). When the facies are classified by
grain size class, the samples exhibit a homogeneous distribution of fine‐
medium to medium‐coarse sand size, except that those in the fine–medium
grain size class are more abundant in the South than in the North sector.
In the South sector, lithofacies LF4 (Fig. 2.10A and B) was generated by
the migration of minor megaripples moved by flood and ebb water fluxes
(Navarrete et al., 2013) that promoted the development of straight sinuous
crest ripples from the sandy bedload (Collison, 1996; Miall, 1996; Ghazi and
Mountney, 2009) under a low‐energy flow regime in the shoreface zone of the
tidal inlet/barrier spit (Navarrete et al., 2013). These low‐energy processes
deposited and preserved a homogeneous and usually well or moderately
sorted sand (Table 2.5). In the North sector profile, the existence of drapes
formed by carbonaceous plant fragments and asymmetric wave and
interference ripple structures (Fig. 2.10C and D) indicates variations in the flow
regime, probably associated with tidal flows from brackish to marine
conditions (Navarrete et al., 2013). The variations in flow regime and the
mixing conditions and properties of water in the system resulted in a
heterogeneous sorting of sand; the sand is generally poorly sorted, with locally
important cementation (Table 2.5).
68
Inlet deposits associated with barrier island systems are typically
represented by complex fill patterns marked by multiple episodes of erosion,
lateral filling and migration (Nishikawa and Ito, 2000; Mallinson et al., 2010).
Heterogeneity in the deposit is determined first by processes dominant in
barrier systems, and secondarily by the architectural elements of such systems
(Davis and Barnard, 2003; Simms et al., 2006; Hodgkinson et al., 2008;
Mallinson et al., 2010). The inlet deposits change dynamically and complexly
through time, and their evolution depends on the dominant process (wave‐
dominant or tidal‐dominant) and the interactions between the processes
(Davis and Gibeaut, 1990; Davis and Barnard, 2000; Nishikawa and Ito, 2000;
Simms et al., 2006; Navarrete et al., 2013). Inlet flow conditions and changes
in the dominant process drive intra‐deposit heterogeneity and the spatial
distribution of sorting facies. Jackson and Rawn‐Schatzinger (1993) reported
large variations in the facies associated with laterally migrating tidal inlets in
barrier island systems. Tidal inlet fill and tidal delta facies contain the coarsest
and least well‐sorted (i.e., with the greatest standard deviation in grain size)
sands in barrier island systems, while middle shoreface, tidal creek and tidal
channel facies consist of the finest sands with the best sorting (i.e., with the
smallest standard deviation). The differences in grain size and sorting between
sand facies are related to the tidal energy flux, sand sources and channel
configuration.
A comparison of the characteristics of the studied b.i./inlet deposit with
the Upper Cretaceous outcrop of inlet deposits associated to barrier island
described by Jackson and Rawn‐Schatzinger (1993) shows that in the North
sector, the studied deposit is composed of facies related to inlet fill of coarser‐
grained and more poorly sorted sand, whereas in the South sector the facies
represents a tidal channel composed of finer‐grained and moderately to well‐
sorted sand. For the b.i./inlet deposit, the sorting facies class has a stronger
relationship with porosity than does the grain size class (Fig. 2.22).
Some zones in the tsunami and b.i./inlet deposits are cemented. The
cement is principally kaolin and calcite; the amount of cement in the both
sectors is small but is filling the pore space as observed by Caja (2004) and
Bauluz et al. (2014). The kaolinite crystallizes from a fluid rich in Si and Al
cations, which is a result of silicate alterations, such as feldspars and even
69
quartz minerals. The process to cristalize kaolinite might have started prior to
the diagenesis but it continued and reached a maximum of crystallization
during diagenesis (Bauluz et al., 2014). The carbonate cementation in both
deposits can be related to burial processes (Curtis, 1978) and/or early
diagenesis (Machent et al., 2007; Henares et al., 2014) where early non‐
ferroan calcite precipitates first with solutes derived from marine water and
sulphate reduction, principally in transgressive shallow marine–coastal plain
deposits during relative sea‐level rise and flooding (Machent et al., 2007). Early
diagenesis also probably induced early compaction and consolidation of
barrier sequences (deVries Klein, 1974).
Compaction of tsunami deposits causes fracturing of quartz and
deformation of micas (Fig. 2.13) at micron scales. Kaolin and calcite cements
may be linked to compaction and diagenesis, as observed by Hammer et al.
(2010). Hammer et al. (2010) described cementation in the Upper Triassic–
Lower Jurassic Åre Formation, and demonstrated a link between compaction,
diagenesis and cementation; these findings are applicable to the cementation
of tsunami deposits in the Aliaga outcrop. The Åre Formation is a deep
reservoir in the offshore Heidrun oil field of central Norway, and consists of
fluvial channel (FCH) sandstones, floodplain fines (FF), and sandy and muddy
bay‐fill sediments (SBF and MBF, respectively). In the FCH sandstones,
authigenic kaolinite is identified as the only pore‐filling clay mineral, and in
some places its volume exceeds 10% of the total rock volume. Carbonate
cement, which is more common than authigenic kaolinite, occurs as cemented
lamina and small‐scale patches. The kaolinite is interpreted to have formed
both by replacement through leaching of feldspar and mica, and by eogenetic
precipitation from pore fluids.
In inlet deposits associated with barrier islands, Jackson and Rawn‐
Schatzinger (1993) observed large amounts of calcite cement in oyster‐rich
beds. The cement of the studied b.i./inlet deposit is principally dolomite, along
with some calcite in the North sector. The cementation is particularly affecting
the macro‐facies LF5, which contains scattered oysters, fish teeth and
centimetre‐thick accumulations of bioclasts or carbonaceous plant fragments.
Cementation may also be related to fluid migration through fractures and
faults. Syn‐sedimentary activity related to extensional structures in the
70
Cretaceous Galve sub‐basin has been highlighted in previous studies (Soria,
1997; Liesa et al., 2004, 2006; Navarrete et al., 2013a, 2013b), and such
structures could have acted as preferred pathways for fluid remobilization
during and after compaction. Martín‐Martín et al. (2012) examined dolomite
bodies occurring in association with basement faults in the southeast
Maestrazgo Basin. The dolomites appear to be intercalated with very low‐
porosity mud‐dominated facies and/or early cemented grain‐dominated facies.
Most of the dolostone volume (60%–70%) consists of replaced dolomite
showing characteristic fabric‐retentive textures and low porosities. The
dolomitization was controlled primarily by tectonic structures in the area, and
secondarily by depositional fabrics and early diagenetic processes. The fluid
flow pattern through strata and along faults could be part of a larger scale
convective system that may have been active in the Maeztrat basin during the
Late Cretaceous post‐rift episode (Martín‐Martín et al., 2015).
The cemented facies (CSs) in the tsunami deposit of the present study is
sparsely distributed, and may reduced porosity and permeability (Henares et
al. 2014) to levels that are within the tsunami porosity range (14%–22%), but
which would be higher in the absence of cementation. However, the
cementation in the CSs facies of the studied b.i./inlet deposit strongly reduces
its porosity and permeability (Fig. 2.21).
2.5. Conclusion
Porosity and permeability are strongly correlated in both sandstone
deposits (Fig. 2.15). Samples from the tsunami deposit are relatively
homogeneous in terms of the sorting facies class (70% composed of WSs and
MSs facies) and porosity distribution (14‐22%) whereas b.i./inlet samples have
heterogeneous distribution of sorting facies between profile sectors. For the
b.i./inlet deposit, the sorting facies class has a strong relationship with porosity
and is distributed as following: the porosity of the Wss facies is >17%; the
porosity of the PSs facies, which is mainly located in the North sector (Fig.
2.24B), is <17%, while porosity values are lowest for the CSs facies (< 5%). The
porosity values for the MSs facies are 10%–22%.
The sedimentary processes of deposition and preservation control the intra‐
and inter‐sandstone heterogeneity, and the sand heterogeneity controls the
71
distribution of petrophysical characteristics. The tsunami deposit is a single
sedimentary event deposited principally bay backflow currents under good
sedimentary conditions for its preservation. The b.i./inlet deposit is a complex
combination of sediment erosion, reworking and preservation under distinct
and dynamic flow conditions. Despite the small size of the outcrop, the two
studied sandstone deposits show distinct sedimentary processes under the
same palaeogeographical context, as well as distinct petrophysical
characteristics (Fig. 2.15). The sorting facies class seems to be more
appropriate for describing sand heterogeneity, and thus influencing
petrophysical parameters in both deposits.
The heterogeneity of the studied deposits is primarily a function of the
observation scale and secondarily of sedimentary processes. Both sand
deposits are homogeneous sandstones at the scale of the sedimentary
formation, and both were deposited under the same sedimentary system.
However, the deposits are heterogeneous at the metre scale (Fig. 2.8 and
2.10) and are very heterogeneous at the micron scale (Fig. 2.13 and 2.14).
With regard to the sedimentary formations, the intra‐bed heterogeneity is
directly related to depositional processes and flow conditions. Therefore, the
first step in understanding how heterogeneity affects the sandstone deposits is
to study the genesis and conditions of sediment deposition that generated the
distribution of heterogeneity, at both the macroscale and the microscale.
The sedimentary heterogeneity of the sandstone deposits described
here can be hierarchically ordered. First‐order heterogeneity is related to the
observation scale, and can be distinguished through detailed sedimentological
studies at the basin scale. Second‐order heterogeneity is related to genesis
and the conditions of sediment deposition and preservation; these
heterogeneities can be recognized at the meter scale, and can be important at
the microscale. Third‐order heterogeneity is related to external elements to
sedimentary system such as palaeo‐relief and/or the presence of syn‐
sedimentary faults, which can locally change flow conditions (Pochat and Van
Den Driessche, 2007) and/or accommodation space (García‐García et al., 2006;
Foix et al., 2013; Navarrete et al., 2013). Moreover syn‐sedimentary faults play
an important role in fluid remobilization during compaction and diagenesis, as
well as in deeper conditions as conduct or barrier to fluid flow.
72
In reservoir studies, the first‐order heterogeneity is assessed through
sedimentological studies of seismic data and core and well logs; second‐ and
third‐order heterogeneities require more detailed studies in terms of facies
distributions and their relationship to petrophysical characteristics. Analogous
outcrop studies should be useful restating the purpose principally in clastic
systems where petrophysical parameters (porosity and/or permeability) are
commonly correlated with specific sandy lithofacies (Hornung and Aigner,
1999; Heinz et al., 2003; Huysmans et al., 2008; Norden et al., 2010, Pyrcz and
Deutsch, 2014). The use of analogous outcrops bridges the inter‐well‐scale gap
that is generally used to define the spatial variability of reservoir properties
(White et al., 2004) as sandy or cementation distribution.
73
CHAPTER 3
GEOLOGICAL AND
PETROPHYSICAL MODELLING OF THE ALIAGA OUTCROP
3.1 Materials and methods______________________76
3.2 Grid construction___________________________81
3.3 Facies modelling___________________________87
3.4 Petrophysical Modelling____________________114
3.5 Modelling Conclusion______________________136
74
75
3. Geological and Petrophysical Modelling of the Aliaga
Outcrop
Modelling of CO2 injection and storage in geological formations is a key
component of the implementation of carbon storage in geological materials,
and is important for understanding the various phases of the CO2 injection and
storage process, such as site geology heterogeneity, drilling, well testing, risk
assessment, storage operations, site monitoring, and site closure (IEAGHG,
2009). Large‐scale geological models are usually based on limited inputs of
data and a simplified modelling approach; however, such models must still
incorporate all relevant geological information related to controls on CO2
behaviour in the reservoir. The complexity of geological models often requires
a prolonged and time‐consuming process to calculate and update a reservoir
model with monitoring data. The requirements of small‐scale geological
models, on the other hand, are different to those of large‐scale models, and
are thus useful for investigating reservoir behaviour in specific zones, such as
around boreholes during the injection of CO2 to monitor fluid behaviour or
pressure build‐ups over short time scales (Norden and Frykman, 2013).
Furthermore, small‐scale models allow us to improve or update large‐scale
models with relevant geological information. In particular, outcrop modelling
has the advantage of providing direct access to geological information at
different scales of observation.
In Chapter 2, the Aliaga outcrop was studied at different scales to
establish a correlation between sandy facies and petrophysical characteristics.
Tsunami and b.i./inlet deposits were described in terms of the sandy facies
distribution at macro (metric) and micro (micron) scales; in addition, 56
samples collected from drillcores recovered along the outcrop were analysed
for porosity and permeability values determined from measurements on plugs.
The macro‐lithofacies defined in the field were refined by petrographic and
petrophysical studies, and the samples were then grouped into four sandy
facies as a function of the relationship of sand sorting to petrophysical
characteristics (see Section 2.3 in Chapter 2). The sandy facies groupings were
used to populate the facies model, and the facies model was used to condition
the petrophysical modelling. In the next step, the petrophysical models are
converted to a reservoir model to investigate, at outcrop scales, how porosity
76
and permeability contrasts observed at sub‐metric scales can impact CO2
behaviour around an injector well during a short injection period and during
the post‐injection period.
3.1. Materials and methods
To construct 3D static models of the deposits, a grid resolution was
defined as a function of model purpose, sedimentary heterogeneity, the
spatial distance between available data and quantity of input data. The
modelling process consisted of mapping distributed properties and discrete
and continuous variables, such as facies, porosity and permeability, onto a 3D
grid. Modelling processes were performed using Petrel software (courtesy of
Schlumberger‐SIS).
The modelling process was based on an interpolation of values or
properties, such as facies, porosity, or permeability, between available data; a
stochastic method was used for the interpolations. The original data
distribution was honoured during the interpolation process, and geostatistical
analyses helped to identify trends or heterogeneities in the data distribution in
a given spatial dimension. Properties of the sample data and sedimentary
sections were up‐scaled to a 3D grid before geostatistical analyses were
applied; the up‐scaling of properties was based on an assignment of sample
values to the grid cells that intersected the data.
3.1.1. Geostatistical analyses
The stochastic method is similar to the Kriging method, except that the
stochastic method uses a variogram and the input data are distributed by
adding additional values between available data (Schlumberger, 2013b). The
Kriging method, also called linear estimation, uses a weighted average of the
sample values, with closer samples having more weight (Matheron, 1971;
Clark, 2001). The stochastic method, on the other hand, generates multiple
equiprobable images of the variable using a random seed, in addition to the
input data (Bohling, 2007); thus, while consecutive runs with the same input
data give similar results, the details of the results are different (Schlumberger,
2013b).
77
Before starting the modelling process, the data were mathematically
described in space to identify trends and/or continuities in the data
distribution. Stochastic algorithms are based on the distribution of input data,
the descriptions of spatial data and probability curves to calculate the
probability occurrences of variables at a particular location (Deutsch and
Journel, 1998; Pyrcz and Deutsch, 2014). In addition, some algorithms require
that the data conform to a standard normal distribution, in which case the
continuous variables were transformed according to needs of the original data
distribution.
3.1.1.1. Spatial data description
The spatial distribution of data in the reservoir or inside an area can be
mathematically described by the similarity and continuity of the data. The
similarity between variable values can be calculated from the cumulative
frequency histogram, standard deviation, variance and coefficient of variation
(Isaaks and Srivastava, 1989). The variance (σ²) measures the spread of the
data distribution, and is given by the average squared difference of the
observed values from their arithmetic mean:
∑ Eq. 3.1
where is the number of data values, represents the data values, and is
the arithmetic mean. The standard deviation is the square root of the variance
and its units are the same as those of the variable being described. The
coefficient of variation (CV) is used to describe the symmetry of the data
distribution histogram, and is calculated as follows (Isaaks and Srivastava,
1989):
Eq. 3.2
where, and σ are the original mean and standard deviation of the data
variable, respectively.
The spatial continuity of data can be described by a variogram (Isaaks
and Srivastava, 1989), which can be used to understand the spacing of data
and the expected length scales of geological features (Pyrcz and Deutsch,
78
2014). Supposing that the value difference between two samples depends on
the distance ( ), then the mean ( ) and the variance (σ²) can be described as
a function of (Clark, 2001) as follows:
∗ ∑ Eq. 3.3
∗ ∑ Eq. 3.4
where represents the value of the variable, is the position of one sample in
the pair, is the position of the other sample in the pair, and is the
number of pairs or correlated values for . The function ∗ is the
calculated average mean difference or expected mean difference between two
samples, and ∗ is the calculated average variance difference between
two samples. The term ∗ , which is also called a variogram or an
experimental semi‐variogram, as the function varies with distance (and
direction), is usually represented as , and is generally referred to as the
lag distance. Clark (2001) and Bohling (2007) summarize the elements of an
experimental semi‐variogram as follows (Fig. 3.1):
Sill: the semi‐variance value at which the semi‐variogram levels off,
where data become uncorrelated; also used as the “amplitude” of a
certain variable value.
Range: the lag distance at which the semi‐variogram reaches the sill
value. The range is the distance at which samples become independent
of one another; presumably, autocorrelation is essentially zero beyond
the range.
Nugget: represents the variability at distances smaller than the sample
spacing, including measurement error.
79
Fig. 3.1: Experimental semi‐variogram and its elements. Modified from Clark (2001).
In the case of = 0, two samples taken from the same location will
show no difference in ; therefore, should pass through the origin of the
graph (0). For values of smaller than the sampling spacing, the nugget effect
estimates this range of the variability in the data. The shape of the semi‐
variogram can be described by semi‐variogram models (Fig. 3.2); the most
common models are the spherical, Gaussian, exponential, potential and linear
models. The nested semi‐variogram model is a linear combination of a short‐
range spherical and a longer‐range spherical semi‐variogram model (Petrowiki‐
SPE, 2015). Nested variograms can be useful for examining the correlations of
samples within outcrop sectors and between outcrop sectors.
3.1.1.2. Stochastic simulation algorithms
The algorithms for stochastic simulations are mathematical functions
that interpolate between available data. The interpolation algorithm is defined
as a function of the variable and its distribution; multiple algorithms are
available for modelling process. In this study, two algorithms were used in the
simulation: the sequential indicator simulation (SIS) and the sequential
Gaussian simulation (SGS).
Sill
Distance, h (lag distance)
Nugget
Semi‐variance Estimation
g(h)
Range
80
Fig. 3.2: Comparison of the exponential and spherical semi‐variogram models with the same range and sill values.
The SIS, which was used to conduct the facies modelling, generates a
stochastic distribution of the facies using a pre‐defined histogram and
directional settings, such as a variogram and extensional trends. The SIS
algorithm runs in three steps: first, it assigns a data value to the closest grid
node; second, it establishes a random path through all grid nodes (using the
variogram); and third, the algorithm checks each grid node to find nearby data
and previously simulated grid nodes, to build the conditional distribution by
Kriging. The conditional distribution is constrained by the global proportion of
each facies type, local data or the relationship with other variables, such as the
seismic attribute (Deutsch and Journel, 1998).
The SGS was used to simulate continuous variables, such as porosity
and permeability. The SGS algorithm assigns data to the nearest grid node, and
then determines a random path through all the grid nodes by finding nearby
data and previously simulated grid nodes and building the conditional
distribution by Kriging. During the simulation, local highs and lows are
generated between input data locations, and the positions of these highs and
lows are determined by a random number. The conditional distribution
honours the global histogram, local data and secondary data; its shape is
Gaussian, with the mean and variance given by Kriging; variogram
reproduction can be poor, principally if the variogram range is larger than the
Sill (C)
Distance (h)
spherical model
g(h)
Range of influence (a)
81
grid node spacing (Deutsch and Journel, 1998). The SGS algorithm requires
that the data are stationary and conform to a standard normal distribution.
The normal distribution of the data means that most of the samples in the
dataset are close to the mean value, while relatively few samples tend to one
extreme or the other (Schlumberger, 2013b). The traditional method of
standardizing data (normal transformation) is to calculate the standard
residual, Y = (Z – m)/σ, where m and σ are the mean and standard deviation of
the original variable Z, respectively; the mean and standard deviation of the
new variable Y are 0 and 1, respectively (Pyrcz and Deutsch, 2014).
3.2. Grid construction
The gridding process defines the size of the geological model, the
resolution of the grid cell and the reservoir thickness; in fact, the geometry of
modelled properties is dependent on the geometry of the existing grid, and
the grid size must be appropriate to the data spacing and to the scale of
geological heterogeneities. A stochastic algorithm propagates property values
along the grid layers, interpolating between data points. In the first step of the
gridding process, input surfaces or horizons marking the boundary surfaces of
the deposit are created to define the geometry and horizontal resolution of
the model; a layering process is then applied to define the vertical resolution
of the model.
The aim of geological and petrophysical modelling is to a build reservoir
model that simulates CO2 injection into the reservoir; 3D dynamic models are
more appropriate than 2D models for obtaining reliable simulation results (EU,
2009). To investigate the CO2 behaviour in zones close to the injection well, a
third direction (40 m long) was created in a westerly direction at the Aliaga
outcrop face. The Aliaga outcrop (Fig. 3.3A) is a 2D face inside the transitional
interval (Fig. 2.2 in Chapter 2) which was mapped by Navarrete et al. (2013
and 2014) on both sides of the Camarillas–Jorcas syncline (Fig. 3.3B and Fig.
2.1C in Chapter 2). Navarrete et al. (2013 and 2014) noted that the thickness
of the profile decreases to the west for 7 km; however, a constant reservoir
thickness was assumed over the 40‐m length of the western outcrop (Fig.
3.3B).
82
Fig. 3.3: (A) Correlation of
logged
sedim
entary sections
along the
outcrop profile
(Fig.
2.3B in Chapter
2;
modified from N
avarrete et
al., 2013).
(B)
Correlation of
outcrop
profile
in
the
western
direction,
located on the
opposite
limb
of
the
Camarillas–Jorcas
syncline
(point
8 on Fig.
2.1C of
Chapter 2).
83
The outcrop data were georeferenced to the ED‐50 UTM‐30 geographic
coordinate projection in the X (Easting UTM) and Y (Northing UTM) directions,
where the Y direction represents the length of the outcrop and the X direction
represents the outcrop thickness (as the sedimentary bedding is sub‐vertical).
The X coordinate was converted to a Z coordinate (representing thickness) in
AutoCAD; the new coordinate axes of outcrop panel were now in the Y
(Northing UTM) and Z (thickness) directions projected onto X plan 693630
UTM (Fig. 3.4).
3.2.1. Input surface/horizon construction
The 2D contour map of the sandstone deposits (Figs 2.9 and 2.11 in
Chapter 2) (Z/Y axes projected on the X plan 693630 UTM) (Fig. 3.4) was
imported into Petrel as lines in the ED 50‐30UTM coordinate geographic
system, at a reference depth of –1000 m for the highest zone (top surface) of
the b.i./inlet deposit. These lines were duplicated in the X direction, spaced at
a distance of 10 m and 20 m, based on the original X coordinate data (693630
UTM), thus maintaining their original Z/Y values (Fig. 3.4 and Fig. 3.5). Four
horizons corresponding to the basal and top surfaces of both deposits (Fig. 3.5)
were created for the 40‐m‐long transect (in the X direction), using an
algorithm of convergent interpolation that relays the original and duplicated
lines of the deposit boundary surfaces. The created horizons have a width of
40 m, a length of 210 m, and variable thicknesses, according to the geometry
of the original drafted lines (Fig. 3.4).
The dimensions of the Aliaga model are 40‐m in the X direction, 210‐m
in the Y direction, and 34.31‐m in the Z direction (Table 3.1). The grid cell size
in the X (E‐W) and Y (N‐S) directions is a function of sample locations and
heterogeneities in a horizontal section; therefore, the Y cell size is 0.5 m and
the X cell size is 1 m. The grid dimensions of each deposit are given in Table
3.2. The zone between the deposits remains unchanged (without horizons and
layering).
84
Fig. 3.4: Base and top contours of the sandstone deposits at X = 693630 UTM
. The sedim
entary sections are oriented S–N
(SedLog5,
SedLog4, SedLog3, SedLog2 and SedLog1). The vertical scale is exaggerated by 4x.
85
Fig. 3.5: H
orizons from the original and duplicated contour lines of the sandstone deposits. The vertical scale is exaggerated by 4x.
86
Table 3.1: Geographical coordinates of Aliaga model grid box
Axis Min (UTM) Max (UTM) Delta (meters)
X 693600 693640 40
Y 4500850 4501060 210
Elevation depth [m] ‐1032.28 ‐997.97 34.31
Table 3.2: Geographical coordinates of sandstone deposits in the Aliaga model
B.i./inlet deposit
Axis Min (UTM) Max (UTM) Delta (meters)
X 693600 693640 40
Y 4500855 4501060 205
Z 0.11 10.53 10.42
Tsunami deposit
Axis Min (UTM) Max (UTM) Delta (meters)
X 693600 693640 40
Y 4500850 4501060 210
Z 0 3.02 3.02
3.2.2. Layering
The layer step defines the cell thickness or the Z cell size. The layering is
a function of the thickness of the deposit and internal sedimentary
heterogeneity in a vertical section. The zone between the top and base
surfaces of each deposit (Fig. 3.5) was divided into proportional layers of the
same thickness, given an average layer size of 0.21 m, as illustrated in Fig. 3.6.
Thus, the b.i./inlet deposit was divided into 30 layers, and the tsunami deposit
was divided into 10 layers.
The final average 3D cell size in the X, Y and Z directions was 1 m, 0.5 m
and 0.2 m, respectively. The layer numbering in the b.i./inlet deposit (from 1
to 30) indicates layer indexing from the top to the bottom layer, given a total
of 504,000 3D cells. The layer numbering in the tsunami deposit (from 33 to
42) indicates layer indexing from the top to the bottom layer, given a total of
168,000 3D cells.
87
Fig. 3.6: Schematic proportionality of layering in Petrel. Layers 1–5 have the same thickness, and were constructed by proportional division of the zone between the top and base surface (modified from Schlumberger, 2013b).
3.3. Facies modelling
The objective of the facies modelling was to distribute facies between
available data, so as to realistically preserve reservoir heterogeneity and
interpolate between existing data (Schlumberger, 2013b). The facies modelling
process was performed using the SIS algorithm. The facies model conditions
the petrophysical model, which is used to simulate CO2 injection. In Chapter 2,
a facies classification (Fig. 3.7) was defined as a function of sand sorting; the
facies classification was well correlated with sample porosity and permeability.
Fig. 3.7: Colour shading for the sorting facies classification used here.
3.3.1. Facies modeling of the tsunami deposit
Before proceeding with the facies modelling, a geostatistical analysis of
the data was performed to identify trends and/or spatial continuity in the data
distribution. Vertical probability curves were constructed from the vertical
distribution of samples after upscaling to the 3D grid. The input data were
0
12
3
Poor sorted sandstone ‐ PSS
Moderate sorted sandstone ‐MSs
Cemented sandstone ‐ CSs
Well sorted sandstone ‐WSs
88
composed of 36 samples classed by sorting facies (see Table 2.2 in Chapter 2),
collected from drillcores and hand specimens, and giving a distribution
principally on the Y axis and also on the Z axis (depth) (Fig. 3.8, and Figs 2.11
and 2.12 in Chapter 2), and from five sedimentary sections (Fig. 3.3 and Fig.
3.8) logged from outcrop and giving a distribution principally on the Z axis
(depth) and also on the Y axis. A vertical probability curve (Fig. 3.9) calculated
for every facies was used in the algorithm that distributed facies in the vertical
(Z) direction. Figure 3.9 shows the original data for both samples and
sedimentary sections, and calculated probability curves; the calculated curves
coupled the original data distributions of samples and sedimentary sections.
Sorting facies of the tsunami samples are composed of 50% facies WSs,
20% facies MSs and 30% facies PSs and CSs (Table 2.2, Chapter 2). The South
sector is more homogeneous than the North sector in terms of the sorting
facies distribution (see Fig. 2.12 in Chapter 2), with facies WSs being dominant.
The distribution of sorting facies in the horizontal direction, principally in the
N–S direction, can be analysed using variograms. In spite of the limited
quantity of data, variograms can be adjusted manually to add sedimentary
variations observed in the field into the model.
The variogram was conditioned by the sample spacing, which varied
from 4 to 10 m, with more proximate samples (spacing, <2 m) obtained in
zones of thicker deposits (thickness, >1 m). Before proceeding with the
variogram, a cursory analysis of the distance between samples of the same
facies was undertaken (see Table 3.3); the minor distance is the distance
between more proximate samples and the major distance is the distance
between more distal samples. The distance is indicative, and helps to
understand the micro‐scale facies geometry within macro‐scale patterns.
Samples with a large major distance (>100 m) are likely present in both sectors
of the profile, as in the MSs, CSs and WSs facies.
89
Fig. 3.8: Data from the Aliaga outcrop on aerial photo 543‐12: Villarluengo, at a scale of 1:5000 (available from the SITA
R w
ebpage of the
Aragon Government) showing the locations of sedim
entary sections and collected
samples.
90
Fig. 3.9: Proportions of Tsunami facies in the up‐scaled samples and sedimentary sections. The right graph shows the calculated probability curve as a function of the original facies proportions. Colour codes for the facies are given in Fig. 3.7.
Table 3.3: Horizontal distance (meters) between samples from the same facies of Tsunami deposit
Facies Minor Major
PSs 10 10
MSs 3 190
CSs 1.8 150
WSs 0.5 200
A nested spherical semi‐variogram was constructed for each facies from
the sample data, in an attempt to describe sample correlations within and
between sectors; in facies with few samples, as in the PSs facies, the
experimental semi‐variogram was adjusted to observations obtained in the
field (Fig 2.9 in Section 2) and the distances given in Table 3.3. Table 3.4
summarizes the experimental semi‐variogram calculated for the tsunami
facies; the X and Y data ranges in Table 3.4 show the ranges of data in the
semi‐variogram (Fig. 3.1). Structure 1 describes sample similarity at short
ranges, and Structure 2 describes similarity at longer ranges. The vertical range
calculated for Structure 2 is larger than the deposit thickness, to avoid adding
heterogeneity. The Y range for Structure 2 of less than 50 m, as in facies PSs
91
Table 3.4: Experim
ental variograms elem
ents of the tsunami facies
92
and CSs, indicates facies discontinuity along both sectors of the profile. The X
range of both Structures 1 and 2 is adjusted to accentuate (or not) the
anisotropy of facies, given an elongated or circular shape of the semi‐
variogram.
3.3.1.1. Results of facies modelling of the tsunami deposit
The stochastic simulation produces multiple results and each result is
unique. The original sorting facies fraction constrains the simulation algorithm,
and the variograms and probability curves add facies variability to the 3D grid.
A comparison of originals, samples and sedimentary sections, and modelled
frequency histograms (Fig. 3.10) shows similar facies fraction distributions
principally within the sample histogram. The sedimentary sections histogram
overestimates PSs facies and underestimates WSs facies amounts, because the
PSs facies consists principally of filled dinosaur footprints, and the sections
were logged at footprint zones.
Fig. 3.10: Frequency histogram of the Tsunami facies fraction for sample data, sedimentary sections and results of modelling. See facies code in Fig. 3.7.
93
The selected Tsunami facies model, illustrated in Fig. 3.11, has a vertical
exaggeration of 4x. The view from the East (Fig. 3.11A) and West (Fig. 3.11B) of
tsunami facies model shows the following:
an elongate and discontinuous geometry of the PSs and CSs facies in a
N–S direction, with both facies being abundant at the base of the
deposit;
the MSs facies geometry elongate in the Y direction, but which is more
continuous than the PSs and CSs facies, and is present in both sectors
of profile; and
the WSs facies is the most abundant facies in the model, but
principally in the upper part of the deposit.
Fig. 3.12 shows the vertical evolution of the facies model, layer by layer.
The facies distribution and geometry evolve upwards into:
an increasing in proportion of WSs and MSs facies;
a higher proportion of WSs facies in the South sector; and
a decreasing area and frequency of CSs and PSs facies patches.
Table 3.5 summarizes the 3D geometry of the modelled facies; the
geometry is given by the experimental semi‐variogram ranges.
Table 3.5: Model geometry of the Tsunami facies
Facies Shape Length (m) Width (m) Thickness (m)
PSs lenticular 10‐40 5‐10 0.2‐0.4
MSs Wedge‐shaped 20‐120 5‐10 0.2‐0.8
CSs lenticular 10‐40 1‐10 0.2
WSs Sand sheet 20‐200 1‐30 0.2‐1.2
94
Fig. 3.11: Facies m
odel of the Tsunami deposit. View from the East (A) and W
est (B) of the deposit. The arrow indicates the North direction and
the upward or downward perspective of the model is indicated by the green
colour and red
colour, respectively. V
ertical exxageratiion of 4x.
95
Fig. 3.12: View of the Tsunami facies model, layer by layer from the base (K42) to the upper layers (K38‐37). The arrows indicate the heterogeneity in the X direction, from east to west. The vertical scale is exaggerated by 4x. The facies code colours are as in Fig. 3.7.
96
3.3.2. Facies modelling of the barrier island‐tidal inlet deposit
The barrier island‐tidal inlet (b.i./inlet) input data are composed of 30
samples classified by sorting facies (Fig. 3.7), and collected from core and hand
specimens (Table 2.4 in Chapter 2 and Fig. 2.11 in Chapter 2) and five
sedimentary sections (Fig. 3.3 and Fig. 3.8), giving a spatial distribution of data
similar to that of the Tsunami deposit. A vertical probability curve was
calculated for each facies, as illustrated in Fig. 3.13, which shows the original
and calculated probability curves and a modelled histogram similar to both of
the original histograms, for both samples and sedimentary sections. The PSs
facies proportion decreases towards the upper layers, the WSs facies
proportion increases towards the upper layers and the CSs samples are located
in the lower layers.
A clear heterogeneity between North and South sectors was noticed in
the facies distribution of b.i./inlet samples: in the North sector, samples were
classified principally as PSs and MSs facies, whereas in the South sector they
were classified mainly as MSs and WSs facies (Table 2.5 in Chapter 2). The
Fig. 3.13: Original up‐scaled facies proportions of samples and sedimentary sections from the b.i./inlet. The right graph shows the calculated probability as a function of the original facies proportions. The facies code colours are as in Fig. 3.7.
97
variograms attempted to capture the spatial variability of sorting facies
samples, and the data were used to construct experimental variograms after
completion of an initial analysis of sample distances.
The horizontal distance in the N–S (Y) direction of samples from the
same facies class is given in Table 3.6, where the “minor distance” shows more
proximate samples and the “major distance” shows more distal samples; the
sample spacing is 4–10 m; the more proximate samples (<2 m) were obtained
from the thicker deposit zones (thickness, >1 m). Samples with a high major
distance (>100 m) are likely present in both sectors of the profile, as in the PSs
and MSs facies.
Table 3.6: Horizontal distance (meters) between samples from same facies of b.i./inlet deposit
Facies Minor Major
PSs 5 160
MSs 0.2 160
CSs 12 12
WSs 0.53 49
Nested semi‐variograms were calculated for facies of the b.i./inlet
deposit, except for facies CSs (Table 3.7), as this facies is present only in the
North sector. The Y range of Structure 2 (Table 3.7) is <50 m for all facies, and
reflects the heterogeneity and discontinuity of facies in both sectors. The X
range of both Structures 1 and 2 (Table 3.7) is adjusted to accentuate (or not)
the anisotropy of facies, given an elongate or circular shape of the semi‐
variogram, as in the WSs and PSs facies, respectively.
98
Table 3.7: Experim
ental sem
ivariogram elements of b.i./inlet facies
99
3.3.2.1. Results of facies modelling of the barrier island‐tidal inlet
deposit
The chosen b.i./inlet facies model for upcoming reservoir studies shows
a global facies fraction similar to the original facies fraction (Fig. 3.14) and
facies geometries according to descriptions of b.i./inlet macro‐ and micro‐scale
patterns, as described in Chapter 2 and in Navarrete et al. (2013).
The frequency histogram of the facies fractions in the b.i./inlet deposit
is given in Fig. 3.14, which compares the facies fractions of samples,
sedimentary sections and modelled facies. The modelled facies fraction is
similar to the observed facies fraction for sample data (±3%) and also for the
sedimentary sections fraction.
Fig. 3.14: Frequency histogram of the b.i./inlet sorting facies fraction for sample data, sedimentary sections and results of modelling. See facies code in Fig. 3.7.
100
The facies model selected for the b.i./inlet deposit for the reservoir
study is illustrated in Fig. 3.15, with a vertical exaggeration of 4x. The facies
model (Fig. 3.15) shows the following:
an increasing of WSs ad MSs facies towards the upper layers;
a lateral discontinuity of facies, except of the MSs facies; and
local patches of the CSs facies, mainly in the North sector.
The vertical evolution, modelled layer by layer (Fig. 3.16) from the
bottom layer (K30) to the top layer (K15, K09, etc.), shows the following:
the proportion of the MSS and the PSs facies decreases;
the PSs facies patches are amalgamated within the MSs facies and
become more discontinuous and distant from one another;
the CSs facies patches decrease in area and are located mainly in the
North sector;
the proportion of the WSs facies increases by forming small elongate
E–W oriented patches (Fig.3.16, small blue arrows in K27 and K22),
which become increasingly abundant until they amalgamate within the
MSs facies in the upper layers.
Table 3.8 summarizes the 3D geometry of the modelled facies; the
geometry is related to the experimental semi‐variogram ranges.
Table 3.8: Modelled geometry of facies of the b.i./inlet deposit
Facies Shape Length (m)
Width (m)
Thickness (m)
PSs lenticular 2‐10 5‐20 0.2‐0.4
MSs Sand sheet 200 40 0.4‐1.2
CSs lenticular 5‐20 1 0.2‐0.4
WSs Amalgamated to MSs 3‐200 1‐30 0.2‐0.8
101
Fig. 3.15: Facies m
odel of the b.i./inlet deposit. View of the top and base of the deposit. The arrow indicates the North direction and green
and
red colours represent up and down views of the model, respectively.
102
103
3.3.3. Discussion on facies modelling
Realistic stochastic facies models depend on knowledge of the
sedimentary model and available input data, and the model should be
designed according to its specific purpose (Norden and Frykman, 2013). Facies
models developed in this study are high‐resolution models based on sub‐
metre‐scale outcrop data, and are intended to be used for petrophysical
modelling through direct facies–porosity correlations (where correlations can
be made). The facies were first defined on the basis of lithology and sand size
distribution, for the purpose of building a sedimentary model at the basin‐
wide scale. The facies were then described at micron scales, including porosity
and permeability measurements. The sorting facies classification used in this
study, which is defined in terms of sand sorting observed at micron scales,
correlates well with petrophysical data. In addition, a cemented facies was
defined independently of sorting criteria, given the influence of cementation
on petrophysical properties. The architecture of the facies model therefore
controls the architecture of the petrophysical model, and the facies model
must be validated before proceeding with the petrophysical modelling.
The geometry and arrangement of modelled facies represent a probabilistic
version of the micron‐scale facies distribution established through kriging
stochastic simulation. The simulation maintains the proportions and spatial
resolutions of available data, and variograms introduce horizontal variations in
the data in the N–S (Y) and E–W (X) directions, while vertical probability curves
were used to calculate variations in the vertical (Z) direction. The input data
were spatially distributed along the Y and Z axes of the X plan 693630 UTM; no
samples were collected in the E–W (X) direction. The Aliaga–Miravete road
separates the sectors of the deposit and represents a gap of 50 m (Fig. 3.8).
‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐
Fig. 3.16 (previous page): View of the b.i./inlet facies model by layer; from the base (K30) to the upper layers (K15–09). The large arrows indicate facies variations in the N–S direction, and small arrows indicate WSs patches elongate in an E–W direction.
104
Experimental nested variograms were calculated in all directions, with
available data distributions being convenient only in the N‐S direction. Nested
variograms allowed a definition of facies variability within and between model
sectors (at short and long ranges, respectively).
3.3.3.1. Discussion on tsunami facies modelling
The pseudo sand‐sheet geometry of the tsunami deposit extends over
an area of 35 km², and shows evidence for the transport of highly
concentrated suspended sediments driven by wave‐dominated sedimentary
processes (turbidity flow) over relatively short periods of time (hours or
minutes) and rapid accumulations of sediment under high‐energy conditions
(Dawson and Smith, 2000; Sugawara et al., 2014). A comparison of the sand
distribution in the studied tsunami deposit with sand distributions in other
tsunami deposits worldwide indicates that such sand‐sheet deposits are
relatively horizontally homogeneous (Clague et al., 1994; Nanayama and
Shigeno, 2006; Jankaew et al., 2008; Prendergast et al., 2012), and that vertical
heterogeneity is expressed by graded sequences of ‘fine‐ to coarse‐grained
sand (Cantalamessa and Di Celma, 2005; Dawson and Stewart, 2007; Jankaew
et al., 2008; Prendergast et al., 2012). The studied tsunami deposit is
composed of relatively homogeneous subarkosic–arkosic sandstones, with
70% of the samples belonging to the WSs and MSs sorting facies classes (see
Table 2.2 in Chapter 2), and with nearly half the samples being classified as
fine‐ to medium‐grained sand. The geometry of the facies in the model (Table
3.5) is given by variograms (Table 3.4) that were manually adjusted mainly in
the E–W (X) direction to reflect the facies heterogeneity described in section
2.2.3.1 in Chapter 2. Facies with poor (PSs) and moderate (MSs) sorting are
located principally in the North sector (Table 2.2 in Chapter 2), the WSs facies
is located in the South sector, and the cemented facies (CSs) occurs in specific
zones in both sectors. The variogram ranges of Structure 1 (Table 3.4) for a
given facies (Table 3.3) primarily reflect the distance between samples in a
sector. The ranges of Structure 2 (Table 3.4) reflect facies continuity; i.e.,
within a sector (distances of <50 m), such as in the PSs and CSs facies; between
sectors, such as in the MSs and WSs facies; or along the entire outcrop, such as
in facies WSs. Thus, the PSs and CSs facies are lenticular and elongate in the N–
S (Y) direction (Table 3.5), whereas the MSs and WSs facies are widely
105
distributed (Table 3.5), with the WSs facies being abundant in the South sector
and higher in the section (Fig. 3.11).
The distribution of sand in the modelled outcrop was well documented
in the N–S direction, along the western limb of the Aliaga–Miravete anticline,
by Navarrete et al. (2014) (see Fig. 2.2 in Chapter 2). Fig. 3.17A compares the
modelled geometry and distribution of facies in a N–S direction with the
descriptions of Navarrete et al. (2014). The coherency between the modelled
facies (Fig. 3.17A) and the described facies is good, and the similarities
between the two are listed below (Fig. 3.17B and C).
‐ The PSs facies and the LF1 lithofacies, located in the bottom layer,
principally fill dinosaur footprints, where present.
‐ The PSs facies is discontinuous in the N–S direction, and its geometry
depends on the presence of footprints.
‐ The proportion of well and moderated‐sorted facies increases
upwards and towards the southern area of the model.
‐ A stacked sedimentary succession is represented in the model by
interbedded facies (arrows in Fig. 3.17A and C).
The studied tsunami deposit is particularly thick (1–3 m) and covers a
vast area. Single‐bed tsunami deposits described in onshore regions are
usually 25–50 cm thick (e.g. Morton et al., 2008), and multiple‐bed tsunami
deposits reach over 1 m in thickness in some cases (Fujiwara and Kamataki,
2007; e.g. Chagué‐Goff et al., 2011). A comparison of the studied deposit with
other palaeo‐tsunami deposits is difficult on account of the exceptional
thickness of the studied tsunami record. However the modelled facies
geometry is comparable to facies heterogeneities described by Nanayama and
Shigeno (2006) and Fujiwara and Kamataki (2007). The facies described by
these authors are defined by their sand size distributions, while the modelled
facies is described as a function of sand sorting. However, in the modelled
facies the sorting facies classification is correlated with the grain size
distribution at the micron scale (see Table 2.3 in Chapter 2). Generally, the
fine‐grained sands are well sorted (WSs) or moderately sorted (MSs), and the
coarse‐grained sands are poorly sorted (Pss).
106
107
Nanayama and Shigeno (2006) described in detail the provenance and
grain size distribution of an onshore tsunami deposit up to 36 cm thick, caused
by a magnitude 7.8 earthquake along the coast of Japan in 1993. The
earthquake epicentre was located 200 km from the current study area. The
authors defined two lithofacies, a gravel lobe facies (GLF) and a sand sheet
facies (SSF), along two sedimentary sections 240 m long and spaced 20 m
apart. They are oriented perpendicular to the shoreline and located at 500 m
from it onshore. Two fining‐upwards sequences were observed in each of four
stratigraphic units. The GLF is composed mainly of matrix‐supported pebble
and cobble sands, with a graded or inverse graded structure and a weak gravel
fabric. The SSF lies above the GLF facies and is composed mainly of fine sand,
has a graded structure, and contains current ripples indicating a landward flow
direction. Over an area 40 m long and 20 m wide, the GLF lobe geometry is
discontinuous and its thickness is variable (0–20 cm), whereas the SSF
lithofacies is widely distributed. Although Nanayama and Shigeno (2006)
defined only two lithofacies, the GLF and SSF facies are comparable to the PSs
and WSs facies, showing the following similarities.
‐ The GLF and PSs facies are discontinuous in both horizontal directions,
N‐S (Y) and E‐W (X).
‐ The GLF and PSs facies are commonly located at the base of the
sedimentary succession (see Fig. 2.6 in Chapter 2) or at the base of
fining‐upwards sequences.
‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐
Fig. 3.17 (previous page): Comparison of the tsunami facies model with the sedimentary sections described by Navarrete et al. (2014). (A) Facies model (colour coding as in Fig. 3.7; vertical exaggeration, 4x). (B) Detail of the area indicated by the rectangle in A, showing sedimentary sections of the South sector and their locations on the drone photo. (C) Detail of the area indicated by the rectangle in A, showing sedimentary sections of the North sector and their locations on the drone photo (modified from Navarrete et al. 2014). Locations of core samples are indicated in (B) and (C). Arrows indicate similarities in the facies distribution between the model and the sedimentary sections: black arrows, PSs distribution; thin red arrows, WSs distribution.
108
‐ The lobed geometry of the GLF facies is probably related to its
proximity to the coastline (within 500 m). The lenticular geometry of
the PSs facies is probably related to its great distance from the
coastline (5–8 Km).
‐ The SSF and WSs facies have sand sheet geometries and are widely
distributed.
‐ The SSF and WSs facies are the most abundant facies in the deposit.
Fujiwara and Kamataki (2007) described Holocene outcrops of shallow
offshore tsunami deposits in eastern Japan. The authors identified three types
of lithology in the deposits: sandy, gravelly and conglomeratic. The deposits
show remarkable lateral facies changes at kilometre scales. The sandy deposits
consist mainly of laminated, packed or weakly cemented sand, and include
gravels and mollusc shells. They are several centimetres to 40 cm thick, and
show generally fining‐upward trends. Four units were identified, with each
deposit being composed of several sub‐layers and representing the results of
repeated sediment flows. Unit Tna is composed of fine‐ to coarse‐grained sand
and is up to 15 cm thick; Unit Tnb consists of laminated sand (coarser grained
than the sand in Unit Tna) and is up to 25 cm thick; Unit Tnc reaches a
thickness of 10 cm and consists of fine sand layers alternating sandy silt layers
covered by mud drapes with concentrated plant debris; and Unit Tnd is
composed of sandy silt or silt beds with a total thickness of up to 15 cm. The
four depositional units exhibit a distinctive stacking pattern and marked grain‐
size differences. Unit Tna, composed of relatively fine‐grained components, is
locally truncated by Unit Tnb; sub‐layers are truncated in their upper parts.
Unit Tnb is divided into more than four sub‐layers, each less than 10 cm thick,
and each showing amalgamated contacts; sub‐layers are composed of sorted
sand with some gravels and mollusc shells, and exhibit planar to gently
undulating laminations and wedge‐shaped cross‐stratification. The variability
in sand sorting and grain size of Units Tna and Tnb is comparable to the
variability observed in the modelled facies, as follows.
‐ Unit Tna would represent a single layer in the facies model; its
lithology is similar to that of the MSs facies. The MSs facies is
composed of moderately sorted fine–medium to medium–coarse sand.
109
‐ Although Unit Tnb is coarser than Unit Tna, Unit Tnb is well sorted,
like the WSs facies. The sub‐layers of Unit Tnb are thicker as those of
the WSs facies and contacts between the layers are amalgamated, as
the contact between the MSs and WSs facies. At the studied scale, the
WSs facies and the MSs facies are amalgamated.
‐ The upper units (Units Tnc and Tnd) are finer grained than the lower
units (Units Tna and Tnb). In the tsunami facies model, the WSs facies is
usually abundant in the upper zone of the deposit, which is composed
mainly of fine and fine–medium sand.
Studies of cementation in tsunami deposits are scarce. Only Fujiwara
and Kamataki (2007) reported weak cementation in tsunami sandstone, and
this may have been linked to compaction and diagenesis, and/or
remobilization of fluids through faults at deeper levels. The tsunami deposits
described by Nanayama and Shigeno (2006) and Fujiwara and Kamataki (2007)
were deposited recently, as are most of the tsunami deposits described in the
literature; thus, compaction by lithostatic pressure is weak. The cemented
facies (CSs) in the studied tsunami deposit is scarce, although the cementation
apparently reduces the porosity and permeability of the beds (see section 2.4
in Chapter 2). Nevertheless, in these cemented facies, porosity values do not
decrease significantly, and are within the tsunami porosity range (14%–22%).
Hammer et al. (2010) observed that Upper Triassic–Lower Jurassic fluvial
channel deposits with carbonate cements occur more commonly in cemented
laminae and small‐scale patches, as occurs in the modelled CSs facies. The
effects of carbonate cementation on reservoir quality are relatively limited, as
the cementation is restricted to millimetre‐scale laminae of mica and organic
debris, as the calcite cementation observed in studied tsunami deposit (see
Fig. 2.13 in Chapter 2).
Based on these observations, it can be concluded that the variograms
and probability curves used to construct the tsunami facies model represent
the geometries of the facies at micrometre scales, which is the same scale as
that used for petrophysical measurements. The sand distributions in the facies
model and those observed by different authors are remarkably similar, despite
the differences in scale. Thus, the selected tsunami facies model is consistent
110
with the scenario described by Navarrete et al. (2014) and the scenarios
proposed for other tsunami deposits.
3.3.3.2. Discussion on barrier island – tidal inlet deposit
The studied b.i./inlet deposit is composed of subarkosic–arkosic
sandstones. The sorting facies distribution of the sandstones is different in the
North and South sectors, and is linked to sedimentary processes. The WSs and
MSs facies are abundant in the South sector (see Fig. 2.11A in Chapter 2) while
the PSs and CSs facies are abundant in the North sector (see Fig. 2.11B in
Chapter 2). When the facies are classified by grain size class, the samples
exhibit homogeneous distributions of fine–medium to medium–coarse sand,
except that facies with fine–medium sand are more abundant in the South
than in the North sector. In the South sector, facies were generated by the
migration of minor megaripples moved by flood and ebb water fluxes
(Navarrete et al., 2013), deposited under low‐energy flow regimes. In the
North sector, sedimentary structures, such as drapes formed by carbonaceous
plant fragments and asymmetric waves and interference ripple structures,
indicate variations in the flow regime, probably associated with tidal flows
from brackish to marine conditions. Thus, heterogeneity in the b.i./inlet
deposits, as well as the geometry and spatial distribution of facies, are the
result of a complex fill pattern marked by multiple episodes of erosion, lateral
filling and migration of inlets within the barrier island system (Nishikawa and
Ito, 2000; Mallinson et al., 2010). Recent studies of inlet morphology and the
lithologies of barrier island systems describe a large range in the sizes and
shapes of inlets, and a strong dependency of the lithofacies distribution on
waves or tidal processes, as well as the source area of sediments (Davis and
Barnard, 2003; Simms et al., 2006; Hodgkinson et al., 2008).
The facies model basically consists of two facies with lenticular
geometries, and two facies with sand sheet geometries. The PSs facies is
lenticular shaped and oriented NE–SW and NW–SE (large arrows in Fig. 3.16);
the CSs facies is also lenticular shaped, and is oriented N–S. The WSs facies is
amalgamated with the MSs facies, forming a sand sheet geometry (small
arrows in Fig. 3.16). These directions of anisotropy in the lenticular facies
represent the palaeocurrent direction measured by Navarrete et al. (2013),
111
who suggested that the barrier island was oriented NW–SE with a tidal inlet
evolving in an ENE–WSW direction and an open marine environment to the S–
SW. The variogram ranges, which give the geometry and distribution of facies
in the model, attempt to reflect the observed heterogeneity (see section
2.2.3.2 in Chapter 2). All variograms (Table 3.7) exhibit a range smaller than
that of the profile sector (~60 m) (except for the CSs facies variogram) to
reproduce observed facies variability between the North and South sectors.
The CSs facies is located in the North sector and matrix is entirely replaced by
dolomite in some samples, and the variogram ranges for the CSs facies are
larger than the model dimension in the X and Y direction.
A comparison of the facies model and sedimentary sections described
by Navarrete et al. (2013) (see Fig. 3.18) yields the following results.
‐ The stack of layers in the thicker zones of the South sector is well
represented by the interbedding of sorting facies (circled zones in Fig.
3.18).
‐ The fraction of WSs and MSs facies increases towards the upper parts
of the deposit (circled zones in Fig. 3.18).
‐ Coarser lithofacies, as well as the PSs facies, are located mainly at the
bottom of the sections.
The facies model is compared with inlet deposits associated with
barrier islands as described by Jackson and Rawn‐Schatzinger (1993) and
Mallinson et al. (2010). Jackson and Rawn‐Schatzinger (1993) described, at
outcrop scales, an Upper Cretaceous inlet deposit of Rock Springs (Wyoming,
United States) composed of stacked laterally discontinuous sandstone layers
that show large variations in both the grain size and sorting of the sand
(medium–coarse sand and poor–moderate sorting). The authors observed
rapid lateral changes and complex geometries resulting from the truncation of
tidal inlets in shallow marine sands. The major sand bodies are associated with
the lateral migration of tidal inlets, and commonly contain oysters at the base
of the deposit. The tidal channel facies are similar to the sedimentary
characteristics of facies in the South sector (see section 2.4 in Chapter 2). The
tidal channel facies is composed of moderately to well‐sorted fine–medium
sand associated with ebb flow currents, with planar tabular bedding 1 m thick
112
Fig. 3.18: Barrier island/inlet facies m
odel and description of sedim
entary sections logged
by Navarrete et al. (2013). The red circles show zones
in the model that correspond to the sedim
entary sections. Vertical exaggeration is 4x.
113
and 6–16 m wide. The sedimentary characteristics of the tidal inlet fill facies
are similar to those of the North sector (see section 2.4 in Chapter 2). The tidal
inlet facies is composed of moderately to poorly sorted coarse sand, with an
erosive base containing laterally continuous shell lag and spit accretion beds.
The cross‐sections of outcrops studied by Jackson and Rawn‐Schatzinger
(1993) show lateral changes of facies from tidal inlet fill to tidal channel, in an
interval 162 m long; this distance is similar to that of facies changes observed
between the North and South sectors. Furthermore, the authors observed a
large amount of calcite cementation in the oyster‐rich bed, as observed in the
present study in the North sector. Mallinson et al. (2010) described pre‐
historic inlet deposits on the Outer Banks barrier island system of North
Carolina, USA, and some of the features that they describe are also found in
the present study; e.g. sandy lithofacies are heterogeneous and most
abundant, and sandy lithofacies are subdivided into heterogeneous sub‐
lithofacies according to composition, colour and texture. Mallinson et al.
(2010) study reveals a complex fill pattern within the palaeo‐inlet facies,
marked by multiple episodes of erosion, lateral filling and southward
migration; it also exhibits marked variations in thickness, increasing from 1 to
3 m over a distance of ~300 m.
In terms of sand heterogeneity, the modelled heterogeneity given by
the variogram is consistent with the facies descriptions in previous studies
(Jackson and Rawn‐Schatzinger, 1993; Mallinson et al., 2010; Navarrete et al.,
2013), which validates the facies model. Cementation in inlet deposits, which
has been described by Jackson and Rawn‐Schatzinger (1993), has a strong
influence on the petrophysical properties, by decreasing porosity and
permeability. In the Aliaga section, cementation is more pronounced in the
North sector, where it mainly affects the PSs facies (containing oyster and shell
fragments) (see section 2.2.3.2 in Chapter 2); this effect is consistent with the
observations of Jackson and Rawn‐Schatzinger (1993). The modelled CSs facies
(Fig. 3.16) is abundant at specific zones in the North sector close to the
Remenderuelas and other minor associated faults (see Fig. 2.2 in Chapter 2),
which suggests that the observed cementation is related to the compaction
and diagenesis of sedimentary facies, and probably also to fluid remobilization
in the fault zone (see section 2.4 in Chapter 2).
114
3.4. Petrophysical Modelling
Petrophysical modelling of porosity and permeability variables in the 3D
grid are based on stochastic simulations. The statistical algorithm used for
calculating porosity and permeability probabilities at a given location is the
sequential Gaussian algorithm, conditioned by spatial data locations, input
distributions, variograms and trends (Deutsch and Journel, 1998; Pyrcz and
Deutsch, 2014). The porosity and permeability of a sedimentary deposit are
represented by a range of values which depend on the rock texture (Corbett
and Potter, 2004; Sun et al., 2007); these properties are often simulated using
a facies model. The facies classification used in this study is correlated to
porosity and permeability measures of plugs extracted from drillcores
recovered from the outcrop (Figs 2.7, 2.17 and 2.21 in Chapter 2).
3.4.1. Petrophysical modelling of the tsunami deposit
3.4.1.1. Porosity modelling of the tsunami deposit
The porosity of the Tsunami deposit and its relationship to sorting
facies was studied in Section 2.3.1; the study of the petrophysics of the
tsunami deposit was based on 30 plugs (Table 2.4 in Chapter 2). The porosity
values are generally in the range of 14%–22% for all sorting facies (Fig. 2.17,
Chapter 2). The porosity model construction of tsunami deposit conditioned by
facies was discarded because no real relationship was observed between
porosity and sorting facies; thus, any analysis of the data according to facies
does not lead to distinguishable flow properties (Pyrcz and Deutsch, 2014).
Therefore, the tsunami deposit porosity was distributed in the grid using a
probability curve and semi‐variograms of measured porosity. The arithmetic
mean of 16.6% and the standard deviation of 3% (Table 3.9) show the low
range of variation of measured values, with some extreme low values denoted
by the variance coefficient. A probability curve was calculated from the
frequency histogram (Fig. 3.19), and this curve was then normalized (Fig. 3.19,
blue curve).
115
Table 3.9: Main statistics of porosity values for the tsunami deposit.
Tsunami
Number of values 30
Mean 0.16605
Standard deviation 0.0334702
coefficient of variance 0.201567
Min 0.02
Max 0.2175
The porosity distribution was analysed spatially using nested semi‐
variograms (Fig. 3.20); the semi‐variogram elements are given in Table 3.10.
The experimental semi‐variogram for the Tsunami deposit porosity (Table
3.11) attempts to represent the correlation of porosity values along an outcrop
sector through Structure 1, and between outcrop sectors through Structure 2.
The semi‐variogram in the X (E–W) direction for both structures 1 and 2 was
calculated using a large tolerance angle (Table 3.10), thus, the variogram range
in the X direction is similar to that in the Y (N–S) direction (Table 3.11).
Fig. 3.19: Histogram of porosity values (bars) for the Tsunami deposit. The blue curve is the porosity probability curve
116
Fig. 3.20: Semi‐variograms for the spatial distribution of porosity in three dimensions: vertical, the major direction Y and the minor direction X. The distance (in meters) is the lag distance listed in Table 3.10.
117
Table 3.10: Experimental semi‐variogram elements for measured porosity in the Tsunami deposit.
Direction
Azimuth
Dip
Number
lags
Lag
distance
Search
radius
Ban
d
width
Tolerance
angle
Lag
tolerance
Thickness
Vertical NA 90 20 0.2 4 1.8 90 50 NA
Major 0 0 18 5 90 17.8 55.8 50 0.001
Minor 270 0 20 5.9 118 123.3 90 50 0.001
Table 3.11: Modelled nested Gaussian semivariogram for measured Tsunami porosity
3.4.1.1.1. Results of porosity modelling of the tsunami deposit
The porosity model of the Tsunami deposit selected for permeability
modelling was obtained after multiple iterations, which were repeated until
the modelled frequency histogram achieved a coherent model design that
fitted the original data. A comparison of modelled and measured frequency
histograms (Fig. 3.21) shows similar shapes and ranges of values in both
histograms, with an overestimation of 14%–15% of values in interval 10
(porosity = 19‐20% of porosity) and an underestimation of 14%–16% of values
in interval 8 (porosity = 16‐17%) for the modelled distributions; however, the
main statistics of the distributions (Table 3.12) in the modelled and measured
histograms are similar. Improvements in the modelled distribution are
reflected by decreasing covariance values.
Structure 1 Structure 2
Type Gaussian Gaussian
Sill 1.7894 0.846
Major range (Y) 23.37 55.235
Minor range (X) 23.271 83.902
Vertical range (Z) 0.91 8
118
Fig. 3.21: Frequency histogram of modelled and measured porosity in the Tsunami deposit.
Table 3.12: Statistics of measured and modelled porosity values for the Tsunami deposit.
Porosity
m³/m³
Min:
Max:
Delta:
Number of
defined
values:
Mean
:
Std. d
ev.
Variance:
Covarian
ce
Measured 0.07 0.22 0.16 30.00 0.17 0.03 0 0.20
Modelled 0.07 0.21 0.15 92715 0.18 0.03 0 0.16
The Tsunami porosity model from the East and bottom perspectives,
illustrated in Fig. 3.22A and B, respectively, shows higher (>20%) and lower
(<8%) porosity values in specific zones. The layer‐by‐layer evolution of porosity
in the model, presented in Fig. 3.23, shows lower values arranged as individual
patches with thicknesses of >60 cm located beside the road (black arrows in
Fig. 3.23), and higher values (>20%) located in specific zones with thicknesses
of <1 m (ellipses in Fig. 3.23). The porosity variation in the model is divided
into four main intervals, with more or less defined geometries, as enumerated
119
Fig. 3.22: Tsunami porosity m
odel: (A) perspective from the East, and (B) perspective from the bottom. V
ertical scale exaggerated is 4x.
120
Fig. 3.23: Layer‐by‐layer evolution of the porosity model of the Tsunami deposit from the bottom (K42) to the upper (K37) layers. See the legend in Fig. 3.22 for details. The arrows indicate patches of lower porosity values (< 8%). The ellipses indicate the zones of higher porosity values (>20%).
121
in Table 3.13. The porosity background of interval 3 (16%–19%) is dominant,
with intervals 1 (6%–11%) and 4 (19%–22%) evolving within it; interval 2 (11%–
16%) is usually associated with interval 1.
Table 3.13: Geometry of porosity patches for a given interval of values.
Interval Code
Porosity interval (%)
Shape Length (m)
Width (m)
Thickness (m)
1 6‐11 spherical 5‐10 5‐30 0.2 ‐ 1.4
2 11‐16 around patch 1 0.2 ‐ 1.6
3 16‐19 Background , everywhere
0.2 ‐ 3
4 19‐22 more or less elongated patches in X direction
10‐20 5‐40 0.2 ‐ 3
3.4.1.2. Permeability modelling of the tsunami deposit
The permeability modelling involved distributing both vertical (Kv) and
horizontal (Kh) permeabilities into the 3D grid. Two main characteristics
describe the distribution of measured permeability values in the 29 samples
(Table 2.1 in Chapter 2): (1) vertical and horizontal permeabilities are strongly
correlated (correlation coefficient of 0.85) (Fig. 2.16 in Chapter 2), and the
permeability behaviour is assumed to be isotropic; and (2) a strong correlation
between porosity and horizontal permeability (Fig. 2.15 in Chapter 2). The
permeability model was constructed as a function of the porosity model in the
X, Y and Z directions, each assuming the same model for Kh(x), kh(y) and Kv(z).
The correlation between measured permeability and porosity is 0.82; thus, it is
assumed that permeability depends on porosity, and that the relationship can
be described by an exponential regression curve (Equation 3.5), as illustrated
in Fig. 2.15 in Chapter 2, expressed by
. . ∗ (Eq. 3.5)
122
3.4.1.2.1. Results of permeability modelling of the tsunami deposit
When applying equation 3.5 to distribute Kh values in the tsunami
deposit, the measured permeability values were not up‐scaled to the 3D grid.
A comparison of measured and modelled permeability frequency histograms
(Fig. 3.24) shows a perfect exponential distribution of the modelled histogram
with an improvement of value intervals (bar width in Fig. 3.24). With regard to
the main statistics of the modelled and measured Kh values (Table 3.14), the
mean and standard deviation remain nearly equal, while the variance,
covariance and delta values are smaller in the modelled than in the original
distribution. The permeability model depends on the porosity model, and the
permeability distribution in the model evolves according to the porosity
distribution (Fig. 3.25); however, the permeability values show less variability
than the porosity values. Permeability values of <1 mD are found in the North
sector of the model, at the upper layers with a thickness of 0.4 m (arrows in
Fig. 3.25).
Fig. 3.24: Frequency histograms of measured and modelled Kh values showing the permeability distribution of the Tsunami deposit.
123
Table 3.14: Main statistics of measured and modelled permeability in the Tsunami deposit.
Kh (mD)
Min:
Max:
Delta:
Number
of
defined
values:
Mean
:
Std. d
ev.
Variance
: Covarian
ce
Measured 0.17 28.76 28.59 29 8.92 6.29 39.6 0.70
Modelled 0.5 21.81 21.31 92715 10.41 5.59 31.28 0.53
3.4.2. Petrophysical modelling of the barrier island – tidal inlet deposit
3.4.2.1. Porosity modelling of the barrier island – tidal inlet deposit
A good correlation between the measured porosity and the sorting
facies was established in Chapter 2, Section 2.3.2, based on the barrier island –
tidal inlet (b.i./inlet) petrophysical characteristics determined from 27 plugs
(Table 2.4 in Chapter 2). The porosity distribution can be divided into intervals
according to sorting facies (Fig. 2.22 in Chapter 2), as enumerated in Table
3.15; lower porosity values are classified as CSs facies and higher values as WSs
facies. Clear lithological heterogeneities are observed between profile sectors,
with the WSs facies located in the South sector and the CSs facies in the North
sector.
The porosity modelling was carried out according to the facies model
using the sequential Gaussian algorithm; porosity intervals (Table 3.15),
variograms and probability curves were defined for every facies. The
probability curves (Fig. 3.26) represent the frequency of the porosity interval
Table 3.15: Measured porosity intervals in the b.i./inlet deposit in both profile sectors, grouped by sorting facies class and frequency.
Sorting Facies class CSs PSs MSs WSs
Porosity interval (%) 0‐6 6‐17 11‐21 17‐22
Facies Fraction in North sector 100% 66% 10% 0%
Facies Fraction in South sector 0% 33% 90% 100%
124
Fig. 3.25: Layer‐by‐layer permeability model for the Tsunami deposit, from the bottom (K42) to the upper (K37) layers. The arrows indicate the zones of lower permeability values (<0.1 mD).
125
Fig. 3.26: Histograms of b.i./inlet deposit porosity by facies, with respective probability curves (blue lines): (A) PSs facies; (B) MSs facies; (C) CSs facies; and (D) WSs facies.
with respect to facies; the curves were smoothed and normalized by a normal
transformation, and the mean and standard deviation used in the modelling
process are specified (Table 3.16).
Table 3.16: Statistics of b.i./inlet porosity distribution by facies
Facies PSs MSs CSs WSs
Mean (m³/m³) 0.1485 0.166 0.023 0.194
Std (%) 0.0377 0.037 0.0002 0.016
The limited number of samples of each facies precludes an acceptable
statistics analysis to calculate experimental semi‐variograms; therefore, semi‐
variogram ranges (Table 3.17) were set with regard to the analyses in Section
2.3.2 of Chapter 2 and Fig 2.22 (see Chapter 2), and in some cases, larger
ranges than model dimension were set to avoid adding artefact heterogeneity.
126
Table 3.17: Semivariograms used in the modelling process for each facies.
SEMIVARIGRAM FOR FACIES
Facies Nugget Major Minor Vertical Sill
PSs 0.10 50.00 50.00 10.00 1
MSs 0.10 75.00 20.00 10.00 1
CSs 0.50 200.00 100.00 10.00 1
WSs 0.10 75.00 40.00 3.00 1
3.4.2.1.1. Results of porosity modelling
Multiple realizations were generated to match the porosity frequency
histogram of input data with that of the model, in addition to matching
coherent porosity variations related to facies geometry (Table 3.8, and
information in Section 2.3.2 in Chapter 2). A comparison of the modelled and
measured porosity histograms (Fig. 3.27) shows similar exponential shapes in
both histograms, with small variations in the range; the main statistics of
modelled and measured porosities are similar, with slight improvements in the
covariance of the modelled values (Table 3.18).
Table 3.18: Statistics measured and modelled porosities for the b.i./inlet deposit.
Porosity (%) Measured Modelled
Min: 0.02 0.02
Max: 0.22 0.22
Delta: 0.19 0.19
Number of defined values: 27 199822
Mean: 0.15 0.16
Std. dev. 0.05 0.04
Variance: 0 0
Covariance: 0.33 0.25
127
Fig. 3.27: Histograms of modelled and measure porosities for the b.i./inlet deposit.
Fig. 3.28 and Fig. 3.29 illustrate the porosity and facies models, from
the East and West perspectives, respectively. The porosity variations in the
MSs and PSs facies are important (Table 3.15), and create contrasting porosity
values in some zones (see the ellipses in Fig. 3.28 and Fig. 3.29). Generally,
within a given facies the porosity values transition from low in the North
sector to high in the South sector according to distribution of Fig. 2.22 (see in
Chapter 2). The vertical layer‐by‐layer evolution of porosity (Fig. 3.30) shows
an increase in porosity values towards the upper layers, with higher values in
the South sector. Porosity evolves in two main directions, as indicated by the
large arrows in Fig. 3.30: (1) a NNW–SSE direction, related to variations in
porosity of <16%; and (2) a N–S direction, as related to variations in porosity of
>16%; the lowest values (<6%), corresponding to CSs facies patches (Fig. 3.16),
are located mainly in the North sector.
128
Fig. 3.28: View from the East of b.i./inlet models. The sedimentary sections are represented as wells. The green/red
arrows indicate North, and
green
shading indicates the top of the model. The vertical exaggeration is 4x. The ellipses indicate the zones of high permeability within a given
facies. (A) Porosity m
odel. (B) Facies m
odel.
129
Fig. 3.29: View from the West of the b.i./inlet models. North is indicated by the green/red
arrows; the top of the model is indicated by green
shading. The vertical exaggeration is 4x. The ellipses indicate the zone of high permeability contrast within a given
facies. (A) b.i./inlet porosity
model. (B) B.i./inlet facies model.
130
Fig. 3.30: Vertical evolution of the b.i./inlet porosity model from the bottom layer (K30) to the upper layers (K15, K09). The large arrows indicate the apparent lateral trend in porosity; the small vertical arrows and the circled zones indicate patches of lower porosity.
131
3.4.2.2. Permeability modelling of the barrier island – tidal inlet
deposit
The methodology used for modelling the permeability of the Tsunami
deposit was also applied to modelling of the barrier island – tidal inlet
(b.i./inlet) deposit. The 27 permeabilities measured on plugs exhibit two main
characteristics: (1) a correlation coefficient of 0.85 between the vertical (Kv)
and horizontal (Kh) permeabilities (Fig. 2.16 in Chapter 2), indicating isotropic
permeability of the deposit; and (2) a correlation coefficient of 0.94 between
the porosity and the horizontal permeability. Thus, the permeability model
was constructed as a function of the porosity model for X, Y and Z permeability
directions, and using the same model for Kh(x), Kh(y) and Kv(z). The
dependence of permeability on porosity is described by a power regression of
the form (Equation 3.6):
∗ . (Eq. 3.6)
where is the permeability and is the porosity (see Fig. 2.15 in Chapter 2).
3.4.2.2.1. Results of permeability modelling
Equation 3.7 was applied to the distribution of permeability (Kh) in the
barrier island – tidal inlet (b.i./inlet) deposit; however, the measured
permeability values were not up‐scaled to the 3D grid. A comparison of the
frequency histograms of the measured and modelled permeabilities (Fig. 3.31)
shows that the shapes of the two histograms are similar, with some
differences in the data spread. The modelled values are distributed close to
the ideal power curve; thus, the mean and standard deviation of the modelled
and measured values are similar, but the variance is lower in the modelled
distribution (Table 3.19). The layer‐by‐layer permeability model is given in Fig.
3.32, from the bottom layer (K30) to the upper layers (K09, K15, etc.). Because
permeability is conditioned by porosity, the heterogeneity in permeability
132
values in the model (Fig. 3.32) is very similar to that in the porosity model (Fig.
3.30), with the orientation trends of permeability being similar to those of
porosity (large arrows in Fig. 3.32); however, the variations in the permeability
model are smoothed relative to those in the porosity model.
Fig. 3.31: Histogram of measured and modelled permeability for the b.i./inlet deposit.
Table 3.19: Statistics of permeability values for the b.i./inlet deposit.
Permeability Kh (mD) Measured Modelled
Min: 0.01 0.01
Max: 16 14.89
Delta: 15.99 14.88
Number of defined values: 27 199822
Mean: 7.05 6.85
Std. dev: 4.8 3.63
Variance: 23.06 13.17
Covariance 0.68 0.52
133
Fig. 3.32: Vertical evolution of the b.i./inlet permeability model from the bottom layer (K30) to the upper layers (K15, K09). The large arrows indicate the apparent trends of lateral permeability variations; the small arrows and circles indicate regions of lower permeability.
134
3.4.3. Discussion on Petrophysical Modelling
The petrophysical modelling process was unique for each deposit, and
accorded with the distribution of petrophysical characteristics and their
relationship to the sorting facies of each deposit. Permeability was modelled in
the tsunami and the b.i./inlet deposits as a function of porosity, on account of
the strong porosity–permeability correlation (see section 2.3 in Chapter 2). The
strong correlation allowed permeability to be modelled using an exponential
function and a power function for the tsunami and b.i./inlet deposits,
respectively. This method smoothed the permeability distribution in the
reservoir, as it was based on the most relevant petrophysical parameters
(Corbett and Potter, 2004), and thus improved the statistical power of the
parameters (Tables 3.14 and 3.19) and the small‐scale variations (Fig. 3.24 and
Fig. 3.31) of the permeability distribution for each deposit.
The porosity distribution in the tsunami deposit is relatively
homogeneous (14%–22%) and independent of the facies. The permeability
distribution is mainly in the range of 3–22 mD, and only one sample showed a
permeability of <1 mD. Despite the grain size variations in the deposit, the
sand is moderately to well sorted, and cementation is weak and sparse.
Therefore, the porosity was modelled as a function of the porosity
distribution, as the distribution is homogeneous (a standard deviation of
values of only 3%). The distribution of the modelled porosity improved the
original porosity distribution by fine‐tuning the main statistics, as well as
decreasing the covariance. The spatial location of lower and higher porosity
values is given by the kriging estimation, and the distribution of values is
ranged in intervals within a geometry (Table 3.13) given by the variogram.
Regions with lower porosity (6%–11%, Table 3.13) are mainly located close to
the road (Fig. 3.23) and to the Remenderurelas syn‐sedimentary fault on the
northern sector of the outcrop, whereas high porosity values (>16%) evolved
as a background or a sheeted geometry (Table 3.13 and Fig. 3.23). The
presence of low‐porosity patches near the road also could reflect the presence
of the fault in this zone.
In the literature, any data is available on the porosity and/or
permeability of tsunami deposits. This lack of data is probably because most
studies are focused on recent tsunami deposits and thickness of these deposits
135
is low (usually less than 1 m), then they are not considered as reservoirs.
However, here the distribution of porosity reflects the homogeneous facies
distribution which is related to the well sorted sand composition of facies.
The b.i./inlet porosity distribution is heterogeneous between sectors
and is correlated with facies (Table 3.15: low‐porosity values correspond to the
CSs and PSs facies and are located in the North sector, while high‐porosity
values correspond to the WSs and MSs facies and are located in the South
sector). The porosity and permeability models reflect the sedimentary
heterogeneity observed in each sector of the profile. The permeability values
in the South sector are >1 mD, whereas in the North sector, permeability
values of half the samples are <1 mD (see Fig. 2.22 in Chapter 2). While all
permeability values are <22 mD, the modelled petrophysical heterogeneity is
consistent with the permeability distribution observed by Ambrose et al.
(2008) for reservoirs in beach and barrier‐island systems in the West Ranch
field (Gulf Coast, USA). These authors reported that barrier island deposits are
typically well sorted, continuous, sandy, and internally homogeneous as a
result of their high‐energy, shallow‐marine origin. Such deposits consist of
elongate shore‐parallel sand bodies separated from the shoreline by muddy
lagoons, and are commonly crosscut by tidal‐inlet and distributary‐channel
deposits that introduce facies heterogeneity and permeability variations. Over
an area of 165 m² and thickness of 3.3 m, which is similar to the dimensions of
the model presented here, the authors described permeability variations in
the inlet‐fill facies in the West Ranch field of up to three orders of magnitude
(<500 to 1000 mD), whereas permeability values in the barrier‐core facies
display less variability, with the highest permeability occurring at the top of
the succession (>2000 mD).
In the b.i./inlet deposit of this study, the trend in porosity variations in a
north–south direction (large arrows in Fig. 3.30), principally in the bottom
layers (K30, K27, K24, and K22), is marked by lower values in the North sector
changing gradually to higher values in the South sector, and also changes to
higher values upwards in the section. The porosity model depends on the
facies model, and the variograms, probability curves and kriging estimations
give the porosity variation in the facies, as indicated in Table 3.15. Semi‐
variogram ranges were manually set to create the horizontal porosity and
136
permeability variations. The ranges represent the large permeability range in
the North sector (0.01–10 mD), which corresponds to the inlet infill facies, and
the more homogeneous permeability distribution in the South sector (10–22
mD), which corresponds to the channel inlet or barrier spit facies.
Furthermore, the lowest permeability values in the North sector correspond to
the CSs facies. Cementation occurs in the inlet infill facies, and in proximity to
the Remenderurelas Fault, as discussed in section 2.4 in Chapter 2.
In both deposits, the tsunami and b.i./inlet, the vertical permeability is
strongly correlated with the horizontal permeability (see Fig. 2.16 in Chapter
2), demonstrating the isotropic behaviour of permeability vectors. The low
anisotropy of permeability in plug samples has also been noted by Meyer and
Krause (2006) in samples without clay or carbonate laminae, at millimetre
scales; moreover, the permeability anisotropy is likely to be dependent on the
scale of the measurement. Here, the permeability anisotropy of the deposits is
represented by the high resolution of the grid, which gives an important
permeability contrast in both the vertical and horizontal directions, creating
permeability anisotropy at the scale of the model.
3.5. Modelling Conclusion
The porosity and permeability models of both deposits were developed
for use in a simulation study of CO2 injection. The small size of the grid blocks
(1 m in the X direction, 0.5 m in the Y direction and 0.2 m in the Z direction)
represents the dimensions of sedimentary heterogeneities, and is in accord
with the sample spacing in the Y and Z directions. Each sample was up‐scaled
to one grid block (grid cell), to avoid the need for averaging to obtain values
for each grid block.
The modelling process was different for each deposit, and honoured
the petrophysical characteristics that were used to build the reservoir model
post‐hoc. The tsunami deposit has a homogeneous porosity distribution in all
facies, despite the small differences in the sorting facies distribution between
the North and South sectors. Consequently, the tsunami porosity model was
constructed as a function of the porosity distribution, and the modelling
process produced a model that is consistent with field observations. The
b.i./inlet porosity modelling was conducted based on the facies models, and
137
performed well in reproducing the observed heterogeneities in the deposit, as
well as the observations of other authors regarding these deposits (Jackson
and Rawn‐Schatzinger, 1993; Ambrose et al., 2008).
Although some factors affect the reliability of the variograms (Oliver
and Webster, 2014), such as the absence of samples taken along the X axis and
the small number of samples for any given facies (Table 3.11), the
petrophysical heterogeneity created by the modelling process is a probable
result of scenarios selected between multiple stochastic iterations. The chosen
models represent heterogeneity in grid block at sub‐metre scales, over interval
200 m long. In fact, the models represent up‐scaled microscopic (micron–
centimetre) features within macroscopic (metre–hectometre) features, as the
facies and petrophysical characteristics were analysed at microscopic scales.
138
139
CHAPTER 4
CO2 INJECTION IN THE
TSUNAMI AND BARRIER ISLAND – TIDAL INLET RESERVOIRS AT THE
OUTCROP SCALE
4.1 Reservoir Model___________________________________141
4.2 Fluid Model_______________________________________142
4.3 Fluid properties____________________________________145
4.4 Results of reservoir simulation________________________148
4.5 Discussion________________________________________158
4.6 Conclusion________________________________________162
140
141
4. CO2 injection in the tsunami and barrier island – tidal inlet
reservoirs at the outcrop scale
Reservoir modelling is essential in studies of geological storage of the
CO2 to understand the behaviour of CO2 in the reservoir for a safely and
efficiently storage. The reservoir model is a petrophysical model reflecting the
rock heterogeneity which maximizes the CO2 storage in the reservoir through
different time scales. The petrophysical model and its architecture is of course
important, but also the fluid model and its physical and chemical properties
which lead the processes of CO2 trapping (see section 1.3.1 in Chapter 1) and
movement. Buoyancy forces are the most important mechanism driving the
CO2 movement in the reservoir by the difference of density (IPCC, 2005;
Frykman, 2009).
The petrophysical models of the tsunami and barrier island –tidal inlet
(b.i./inlet) deposits built in Chapter 3 were converted into reservoir models for
simulation study cases of CO2 injection. The size of reservoir models here
represent nearly one grid block (grid cell) of classic reservoir models. The
complexity of tsunami and b.i./inlet models is related to the grid resolution
(size and number of grid blocks) and the petrophysics distribution in the
reservoir. Four study cases of each reservoir, the tsunami and b.i./inlet, tested
the sensibility of the injector well location and the reservoir thickness in the
behaviour of injected CO2 during a short‐injection and pos‐injection period.
The studies of CO2 injection simulation were undertaken in Eclipse 300 (E300)
software using the CO2STORE option, a courtesy of Schlumberger.
4.1. Reservoir Model
The porosity (see Fig. 3.22, 3.28 and 3.29 in Chapter 3) and permeability
models (see Fig. 3.25 and 3.32 in Chapter 3) of each deposit of the Aliaga
outcrop, the tsunami at the bottom and the b.i./inlet at the top (see Fig. 3.3 in
Chapter 3), supplied two independent reservoir models from different
sedimentary processes with a 15 m‐thick of shales, marls and micrit carbonate
deposits separating them. The dimensions of Aliaga 3D grid and of each
deposit are given in Table 4.1 (see section 3.2 in Chapter 3); the tsunami
142
reservoir has 92,659 active grid blocks between layers K 33‐42; the b.i./inlet
reservoir has 199,818 active grid blocks between layers K 1‐30.
In order to avoid a fast pressure build‐up related to the small size of
reservoir models, the pore volume of the four end faces of each reservoir
model was multiplied by 1000 (Fig. 4.1).
Table 4.1: 3D grid dimensions of Aliaga model and each resevoir.
Aliaga model Horizontal i East‐West
Horizontal j North‐South
Vertical k thickness
Aliaga grid bocks 40 420 42
Tsunami grid blocks 40 420 10
B.i./inlet grid blocks 40 420 30
Average size of grid block (meters)
1 0.5 0.21
4.2. Fluid Model
The fluid model defines the fluid components and phases in the
reservoir at the initial conditions and during the simulation. The temperature
of fluid model was constant of 32⁰C and the initial equilibrium pressure was
103 bars for Tsunami model and 98 bars for b.i./inlet model, which correspond
to the hydrostatic pressure at 1003 and 998 m‐deep from the ground,
respectively. At this range of temperature (T) and pressure (P), the state of CO2
is very close to its critical point in P x T phase diagram (see Fig. 1.3 in Chapter
1); thus the CO2 is in a supercritical state where it behaves as a gas (IPCC,
2005).
A fluid model of 4 components was established with three phases, a
CO2‐rich gas phase, a H2O‐rich liquid phase and a solid phase. The components
at the intial conditions were H2O and CO2 in aqueous and gas phases
respectively, and NaCl and CaCl2 in aqueous phases. The mutual solubility and
partitioning of CO2 and H2O in aqueous and gas phases during simulation are
calculated to match the experimental data of Spycher and Pruess (2005). The
salt components (NaCl and CaCl2) can be presented in both aqueous and solid
143
Fig. 4.1: Top view of tsunami and b.i./inlet reservoir m
odel showing the four end faces of each reservoir m
odel which were multiplied
by 1000. V
ertical scale is exaggerated of 4x.
144
phase during the simulation; the salt equilibrium in the reservoir is calculated
according to reactions (Schlumberger, 2013a) as following:
H2O H+ + OH‐
CO2 + H2O HCO3‐ + H+
HCO3‐ CO3
‐ + H+
CaCl+ Ca2+ + Cl‐
NaCl(s) Na+ + Cl‐
CaCl2(s) CaCl+ + Cl‐
CaCO3(s) Ca2+ + CO32‐
The dissolved salts are essentially non volatile and formulation hardly
changes. In salinity ranging up to ionic strength of around 6 molal (below halite
saturation), the water activity equals its mole fraction on the basis of a fully
ionized salt. Thus, the water mole fraction in the CO2‐rich phase ( ,
Equation 4.1) and the CO2 mole fraction in the aqueous phase ( , Equation
4.2) are respectively expressed as (Spycher and Pruess, 2005):
(Eq.4.1)
.
(Eq.4.2)
K0 is the thermodynamic equilibrium constant for each component at temperature T and
reference pressure P0=1 bar, for respective reactions H2O (liquid) H2O(gas) and
CO2(aqueous) CO2(gas/liquid).
P is total pressure.
V is the average partial molar volume of each pure condensed phase over the pressure range
P0‐P.
F is the fugacity coefficient of each component in the CO2‐rich (compressed gas) phase.
R is the gas constant.
is the activity coefficient for aqueous CO2.
145
4.3. Fluid properties
The physical and chemical properties of fluid model components,
principally of the CO2, determine the behaviour and trapping mechanism of
injected CO2 and its evolution in the reservoir during and after the injection
operation.
4.3.1. Density
The density of supercritical CO2 (at depths below 800 m, see Fig.1.3)
range from 50 to 80% of the density of water, under these conditions,
buoyancy forces tend to drive CO2 upwards (IPCC, 2005). However, at the
interface between supercritical CO2 and brine, CO2 dissolves in the brine to
form a solution that is heavier than the underlying brine (Nomeli et al, 2014).
The simulator (E300) leads to fast density changing for the CO2 rich gaseous
phase, and the gas density is calculated by a cubic equation of state (EOS),
modified from Redlich‐Kwong EOS and adjusted to match experimental results
of Spycher and Pruess (2005). For the H2O rich phase, density is firstly
computed from the density of pure water following the thermo‐dynamical
properties detailed in “The International Association for the Properties of
Water and Steam” (IAPWS‐IF97), then the Ezrokhi’s method (Equations 4.3 and
4.4) is applied to calculate the effect of salt and CO2 (Prost, 2008; e.g. Zaytsev
and Aseyev, 1993).
∑ (Eq.4.3)
r is the density of water and r0 is the density of pure water.
ci is the mass fraction of each component.
Ai is the activity coefficient for each component
, , , (Eq.4.4)
, i are coefficients for a series of electrolytes and
stands for temperature in degree Celsius.
Solid density of NaCl and NaCl2 were respectively 2170 kg/m3 and 2150
kg/m3 at the reference pressure of 100 bars and temperature of 32⁰C
(Schlumberger, 2013a).
146
4.3.2. Viscosity
Supercritical CO2 is much less viscous than water (by an order of
magnitude or more) and migration is controlled by the contrast in mobility of
CO2 and the in situ formation fluids (IPCC, 2005; Celia et al., 2005; Nordbotten
et al., 2005). The viscosity of CO2 is calculated from the equations of Vesovic et
al. (1990) and Fenghour et al. (1998) which declare the viscosity as a function
of temperature and pressure. The viscosity of H2O rich phase is calculated as
the H2O density. Firstly, the H2O viscosity is calculated for pure water and then
the Ezrokhi’s method is applied (Equations 4.3 and 4.4).
4.3.3. pH calculation
The dissociation of CO2 leads to the transformation of dissolved CO2
into bicarbonate ions (ionic trapping) inducing a lowering of the pH in the
formation water (Chadwick et al., 2008). The pH is calculated as given in the
equation 4.5 (Schlumberger, 2013a):
(Eq.4.5)
is the activity coefficient of the H+ ion. is themolality of theH+ ion.
4.3.4. Saturation functions
Interactions between the CO2 and H2O rich‐phases are represented at
the grid‐block scale by the capillary pressure and relative permeability
functions (Doughty, 2007), which depend on the fluid saturation in the porous
media (Corey, 1954) or the grid block. Many studies have showed the impact
and the sensibility of the capillary pressure and relative permeability in the
CO2 trapping mechanism by increasing the residual trapping mechanism
and/or accelerating the CO2 dissolution in the brine (Juanes et al. 2006; Spiteri
and Juanes, 2006; Plug and Bruining, 2007; Pini et al., 2012; Boxiao et al., 2013;
Frykman et al., 2013). The water (WSF) and gas saturation (GSF) functions of
Table 4.2 were taken from the CO2STORE.data Schlumberger study case, as
well as the gas‐water drainage capillary pressure (Pcog). However, sensibility
tests of these parameters for the Camarillas Fm. are strong recommended,
147
because the capillary heterogeneity in the reservoir induces the capillary
trapping and reduces the migration of CO2 plume (Frykman et al., 2013).
Table 4.2: Water and Gas saturation function used in this study
WSF (water saturation function)
GSF(Gas saturation function)
Sw Krw Sg Krg Drain Pcog (bars)
0.3 0 0 0 0
0.38 0.000152 0.08 0 0.6
0.46 0.002439 0.16 0.000407 0.78
0.53 0.012346 0.23 0.005831 0.93
0.61 0.039018 0.31 0.024131 1.09
0.69 0.09526 0.39 0.064892 1.26
0.77 0.197531 0.47 0.140566 1.49
0.84 0.36595 0.54 0.269314 1.84
0.92 0.624295 0.62 0.484797 2.53
1 1 0.7 1 10
The solid saturation reduces the mobility of flow by the solid adsorption
on to the rock formation (Schlumberger, 2013a). However, at the time scale of
this study, the solid saturation is low.
4.3.5. Diffusion
Diffusion or mass transfer is the dominant mechanism to transport
dissolved CO2 into the low‐permeability regions from the high‐permeability
pathways (Chang et al., 2014). The molecular diffusion in E300 is calculated for
diffusive flows in terms of the liquid mole fractions and vapour mole fractions.
The diffusion coefficients ( ) of each component, entered by user, solve the
condition (equation 4.6) of Reid et al. (1987) to define the molar vapour flux
( ) of each component per unit of area:
(Eq.4.6)
is the molar concentration
is the molar concentration gradient of component i
148
Water diffusion coefficients are typically an order of magnitude lower
than gas diffusion coefficients (Schlumberger, 2013a) and were entered for the
four components of the fluid model. Although the diffusion mechanism is slow
and not relevant to time scale of this study, Schlumberger recommended the
utilization of diffusion coefficients, which are summarized in Table 4.3.
Table 4.3: Diffusion coefficients used in this study
Water diffusion coefficients Gas diffusion coefficients
H2O CO2 NaCl CaCl2 H2O CO2
0.0001 0.0001 0.0001 0.0001 0.001 0.001
4.4. Results of reservoir simulation
In order to investigate the impact of the permeability and porosity
contrast and the reservoir thickness in trapping mechanism of the CO2, for
each reservoir were simulated four study cases with two different injection
rate regimes at two distinct well locations. The CO2 was injected as a dry gas.
4.4.1. Simulation on the Tsunami reservoir
The Table 4.4 summarizes the study cases of the tsunami reservoir
model in terms of the injection rate, injection time and observation time after
injection. The study cases TsunV2 and TsunV3 had the injector well located in
the North sector of model, while in the study cases TsunV4 and TsunV5 the
injector well was located in the South sector. The cases TsunV2 and TsunV5
had a constant injection flow rate of 80 sm³/day and the cases TsunV3 and
TsunV4 had three different injection rates and each one is spaced from
another by ~ 1 year observation time or stopped injection time. In all study
cases, the injector well was horizontal in the ‘Y’ direction and located at the
bottom of model connecting 7 grid blocks. The well diameter at the
connection of grid block is 19 cm; this value was required to calculate the
connection transmissibility factor and well productivity/injectivity index.
149
Table 4.4: Injection rate, simulation time and observation time after injection of Tsunami study cases
Case Rate (sm³/day) Injection time
(years) Observation time (Years)
TsunV2 & TsunV5
80 3.56 3.97
TsunV3 & TsunV4
200 0.92 0.97
150 0.51 0.97
100 0.34 3.17
The CO2 is present in the reservoir as dissolved in the brine (saline
water), trapped in the gas phase and mobile gas phase. Table 4.5 summarizes
the CO2 amount in the reservoir for different CO2 states. Although the total
amount of CO2 injected in the cases of different rate regimes was 3.7% higher
than the cases of constant rate, the constant rate cases TsunV2 (northern
injector well) and TsunV5 (southern injector well) dissolved quite more CO2
than the cases TsunV3 and TsunV4 for the respective sector of model. This
difference in the amount of injected CO2 produces a slightly larger plume
extension of gas saturation (CO2) in the cases of different injection rates (see
arrows in Fig. 4.2B, D and Fig. 4.3A, C). The maximum saturation of CO2 gas
into the saturation plume at the end of simulation reached up 30% and 25% in
the cases with the injector well located in the North and South sectors,
respectively, independent on the rate regime (Fig. 4.2 and Fig. 4.3). During the
simulation, the gas saturation in the gas plume up to 45% in the cases of
constant rate regime (covert picture) and to 50% in the cases of different rate
regimes. Further, the southern cases (TsunV4 and TsunV5, Fig. 4.3) stored
quite more (1‐2%) CO2 as dissolved in the brine than northern cases (Table
4.5). The average reservoir pressure of all case studies remained ~122 bars at
the end of simulation, and up to 130 bars at the wellbore bottom during
injection.
150
Table 4.5: Summary of CO2 distribution into tsunami reservoir at the end of simulation
151
Fig. 4.2: Plume extension of the CO2 saturation at the end of simulation for Tsunami study cases with the injector well (‘W’) in the North sector. Vertical exaggeration of 4x. (A) Reservoir top view of the case TsunV2 with constant injection rate; the injector well is represented by the vertical line. (B) Reservoir top view of the case TsunV3 with different rate regimes; arrow indicates the CO2 saturation of around 24%. (C) North‐South cross section of CO2 saturation between points 1 and 1’ in A of TsunV2 case; the injector well is represented by vertical bars marked ‘W’. (D) North‐South cross section of CO2 saturation between points 2 and 2’ in B of TsunV3 case, the arrow indicates the northern border of the plume extension.
152
Fig. 4.3: Plume extension of CO2 saturation at the end of simulation for Tsunami study cases with the injector well (‘W’) in the South sector; arrows indicate the CO2 plume extension in North direction. Vertical scale exaggerated of 4x. (A) Reservoir top view of case TsunV4 with different rate regimes. (B) Reservoir top view of case TsunV5 with constant injection rate. (C) North‐South cross section of CO2 saturation between points 1 and 1’ in A (case TsunV4) at well location. (D) North‐South cross section of CO2 saturation between points 2 and 2’ in B (case TsunV5) at well location; injector well is represented by blue bar.
4.4.2. Simulation on the barrier island – tidal inlet deposit reservoir
Table 4.6 summarizes the barrier island – tidal inlet (b.i./inlet) study
cases in terms of the injection rate, injection time and observation time after
injection. The injector well was located in the North sector in the study cases
InletV3 and InletV4 and in the South sectors in cases InletV5 and InletV6. The
cases InletV4 and InletV5 had a constant injection rate of 260 sm³/day and the
cases InletV3 and InletV6 had three different injection rate regimes spaced of
~1.6 year observation time after injection period. In all cases, the injector well
was horizontal in the ‘Y’ direction and located at the bottom of model
153
connecting 7 grid blocks. The well diameter at grid block connection is 19 cm;
this value was required to calculate the connection transmissibility factor and
the well productivity/injectivity index.
Table 4.6: Rate and simulation time of b.i./inlet study cases
Injection rate (sm³/ day)
injection time (year)
Observation time (year)
IV3 and IV6
400 0.54 1.96
150 0.34 1.45
150 0.34 1.95
IV4 and IV5 260 1.22 5.34
Table 4.7 summarizes the amount of CO2 dissolved in the water, and
mobile and trapped in gas phase for all study cases: The comparison of cases
with the injector well located in the North (Fig. 4.4) shows a larger extent of
the CO2 saturation plume and a more homogeneous distribution of the CO2
saturation into the plume in the case of constant injection rate (InletV4) at the
end of simulation, despite of the quantity of injected CO2 (Table 4.7) in InletV4
and InletV5 was a little lower. Although the cases InletV5 and InletV6
(southern injector well) show similar CO2 plume extension (Fig. 4.5), nearly 3%
more CO2 is dissolved in the case of constant rate (InletV5) (Table 4.7). The
cases with the injector well in the North sector dissolved 20% more CO2 than
the cases of southern location of the injector well. The average reservoir
pressure at the end of simulation is similar between study cases (Table 4.7);
maximum bottom well pressure during injection was around 130 bars (± 3
bars) for study cases of different rate regimes (InletV3 and InletV6) and 128
bars and 118 bars for InletV4 and InletV5, respectively.
154
Table 4.7: Summary of CO2 distribution into b.i./ inlet reservoir at the end of simulation
155
Fig. 4.4: Plume extension of the CO2 saturation at the end of simulation for b.i./ Inlet study cases with the injector well (‘W’) in the North sector. The arrows indicate the CO2 plume extension in the North‐South direction. Vertical scale is exaggerated of 4x. Reservoir view from the top of (A) the case InletV3 with different rate regimes and (B) the case InletV4 with constant injection rate. (C) Cross section of CO2 saturation between points 1 and 1’ in A (case InletV3). (D) Cross section of CO2 saturation between points 2 and 2’ in B (case InletV4).
156
Fig. 4.5: Plume extension of the CO2 saturation at the end of simulation for b.i./inlet study cases with the injector well in the South sector (‘W’). Vertical scale is exaggerated of 4x. Reservoir view from the top of (A) the case InletV5 with constant injection rate and (B) the case InletV6 with different rate regimes. (C) Cross section of the CO2 saturation between points 1 and 1’ in A (case InletV5). (D) Cross section of CO2 saturation between points 2 and 2’ in B (case InletV6).
157
4.4.3. Flow at the boundary grid blocks of reservoirs
The pore volume of the four end faces of reservoir models were
multiplied by 1000 to avoid a fast increasing of pressure linked to the small
volume of reservoir grid (Fig. 4.1). The end faces grid blocks represents ~5% of
the total grid blocks with defined continuous values for both reservoir models
(Table 4.8). Their total pore volume correspond to ~98% of the total pore
volume in the reservoir, however the amount of CO2 (in all phases) into these
grid blocks at the end of simulation represented less than 1% of the total CO2
activity in the entire model (Table 4.8) for both reservoirs the tsunami and
b.i./inlet. The flow is inactive outside the reservoirs.
Table 4.8: Summary of defined values for pore volume and CO2 activity in the entire model (column ALL), boundary zone of pore volume multiplied by 1000 (MULTPV) and the main zone with the original pore volume (MUTPVNO).
ALL MULTPV MULTPVNO
Statistics for pore volume (rm³) at initial conditions
InletV3
Type of data: Continuous
Number of defined values:
199818 10110 189708
Sum of values: 175551 172080 3471
volume percentage: 100 98.0227968 1.9772
Statistics for CO2 at the end of simulation
InletV3
Number of defined values:
199818 10110 189708
Sum: 12565.493 108.3761 12446.9
volume percentage 100 0.86248984 99.0561
TsunV5
Number of defined values:
92659 4868 87791
Sum: 10417.1272 79.5426 10341.9
volume percentage: 100 0.7635752 99.278
158
4.5. Discussion
The CO2 injected at the end of the simulation was, in all study cases,
distributed in the reservoir as a gas phase and as a phase dissolved in brine.
The dissolution of CO2 in brine is an important mechanism for its safe long‐
term entrapment (IPCC, 2005). Moreover, as a mobile gas, the CO2 in the
reservoir moves upwards, driven by buoyancy, which increases the risk of
leakage (IEA‐GHG, 2009) by diffusion across caprock formations, mobilization
through natural faults and fractures, and through human‐made features such
as wellbores (Celia et al., 2005). Leakage types range from short‐term large
leakage to long‐term diffuse leakage (Chadwick et al., 2008). Although the CO2
trapped by residual saturation is effectively immobile, an upwards leakage
pathway may degas CO2 as saturated brine is depressurized (IPCC, 2005).
However, when CO2 is dissolved it no longer exists as a separate phase,
thereby eliminating the buoyant forces that drive it upwards (IPCC, 2005). In
addition, the dissolved CO2 may, over thousands of years, be converted to
stable carbonate minerals, thus trapping the CO2 as minerals in the reservoir
(see Fig. 1.2 in Chapter 1).
The partitioning of CO2 between aqueous and gas phases is strongly
dependent on the fluid model, temperature and pressure (IPCC, 2005).
Reservoir heterogeneity also influences the CO2 distribution in the reservoir.
Injection in homogeneous high‐permeability zones provides conduits for flow
circulation and the lateral distribution of CO2 as an aqueous phase (Akatu,
2008; Ambrose et al., 2008; Sifuentes et al., 2009), while heterogeneous
permeability zones have reduced injectivity (Hovorka et al. 2004; Deng et al.
2012), although such zones trap more residual CO2 (Ambrose et al., 2008,
Frykman et al. 2013).
The heterogeneity of reservoir models is expressed by the contrast of
porosity and permeability in a reservoir. The distribution of permeability and
porosity values is strongly correlated with the facies classified as a function of
the sand sorting (see Chapter 2); thus, their spatial distribution is controlled by
the geometry of the depositional facies and the degree of continuity of sand
bodies (Ambrose et al., 2008). The tsunami reservoir model has a
homogeneous porosity and permeability distribution (see Figs. 3.23 and 3.25
in Chapter 3), with lower values located in specific zones, principally in the
159
North sector. The cross‐sectional view of the tsunami porosity model at well
locations (Fig. 4.6A), which is strongly correlated with the permeability model
at these locations (Fig. 4.6B), shows high porosity values (>20%) close to the
wellbore in the South sector, whereas in the North sector the wellbore is close
to lower porosity values (11%–13%). The injector well located in the South
sector stores slightly more CO2 dissolved in brine and trapped in the gas phase
(Table 4.5); in addition, the maximum gas saturation is relatively low (Fig. 4.3).
With regard to the injection rate regime, a constant injection rate seems to
enhance CO2 dissolution in both sectors at the studied scale.
The b.i./inlet reservoir model is heterogeneous in terms of the porosity
and permeability distribution between the North sector (lower porosity and
permeability values) and the South sector (higher porosity and permeability
values). The reservoir thickness also varies; in the South sector, the thickness
reaches 7 m, as compared with a thickness of only 1–2 m in the North sector.
The cross‐sectional view of the b.i./inlet permeability model at the wellbore
location in both sectors (Fig. 4.7) shows the highest permeability values (>10
mD) close to the wellbore in the South sector (Fig. 4.7A and C) and the lowest
values (1–5 mD) close to the wellbore in the North sector (Fig. 4.7B and C).
Despite the large thickness (5–6 m) and the high permeability values in the
South sector, study cases in the North sector stored 20%–25% more CO2
dissolved in brine and 10%–15% less CO2 as a mobile gas phase (Table 4.7). A
comparison of cases in the South sector shows that a constant injection rate
increases the amount of dissolved CO2 by 3%. A comparison of cases in the
North sector shows that a constant injection rate influences the partitioning of
the CO2 in the reservoir and the CO2 saturation gas in the plume. Injection at a
constant rate trapped more CO2 dissolved in brine and as residual gas; in
addition, the gas plume showed a more homogenous saturation distribution,
with smaller patches of high saturation values displaced in a southeast
direction relative to the injection point (Fig. 4.4).
Despite the simplicity of the fluid model, at the studied scale the cases
in which the injector well is located in a thinner reservoir zone stored more
CO2 dissolved in brine, principally in the b.i./inlet reservoir, where
concentrations of dissolved CO2 were 20% higher. Furthermore, when CO2 is
160
Fig. 4.6: Tsunami models (vertical exaggeration, 4x) and locations of wellbores (represented by vertical bars m
arked
‘W’) (A) North–south
cross‐section of the porosity model at the well locations in the North and South sectors. (B) Overview of the permeability m
odel
161
injected in thicker zones (6–7 m), more CO2 is present as mobile gas. At the
reservoir scale, thin reservoirs and caprock are both susceptible to damage by
the build‐up of pressure during injection, and CO2 injectivity may be
diminished (Ambrose et al., 2008). The rate of injection in tsunami deposit
cases was lower than that in the b.i./inlet cases, to control for pressure build‐
up during injection at the bottom of the hole. A similar pressure build‐up was
observed in all tsunami cases (~30 bars above the initial reservoir pressure).
For the b.i./inlet reservoir, the pressure build‐up was only 18 bars in the case
of a constant injection rate, based on an injector well in the South sector
(InletV5), while in other cases the pressure build‐up was 30 bars.
Fig. 4.7: The b.i./inlet permeability model (vertical exaggeration, 4x) and locations of wellbores (represented by vertical bars marked ‘W’). (A) North–south cross‐sectional view of the permeability model at the location of wells in the South sector. (B) North–south cross‐sectional view of the permeability model at the location of the well in the North sector. (C) Overview of the permeability model.
162
4.6. Conclusion
Despite the relatively low permeability of tsunami and b.i./inlet
reservoirs (0.1–20 mD), the two deposits exhibit an important permeability
contrast in a restrained volume. In the tsunami reservoir, the location of the
injector well in a homogenous and higher‐permeability zone improves CO2
dissolution and decreases the gas concentration in the gas plume. In the
b.i./inlet reservoir, with the injector well located in thinner heterogeneous
zones (high permeability contrasts, as in the North sector), the constant rate
regime seems to improve CO2 dissolution and decrease the gas concentration
in the gas plume. The amount of dissolved CO2 in the northern b.i./inlet cases
(thinner zones) is similar to the amount of dissolved CO2 in tsunami cases.
However, the tsunami cases seem to store more CO2 as residual gas
At the metre scale, reservoir thickness has the greatest impact on the
amount of dissolved CO2 under the simulated conditions. The CO2 injected in a
thin reservoir (1–3 m thick) improves its dissolution in formation water, which
reaches a total of 40% injected CO2 in a short period (7 years). In addition,
constant rate regimes can control pressure build‐up and enhance CO2
distribution in the reservoir in its different phases. Although heterogeneity is
observed with higher permeability contrasts (0.01–20 mD) in the North sector
of the b.i./inlet reservoir, this zone seems to be the best for improving CO2
dissolution at higher injection rates, principally under a constant injection rate.
163
CHAPTER 5
CONCLUSIONS AND
PERSPECTIVE
CONCLUSIONES Y PERSPECTIVA
164
165
5. Conclusions and Perspective
Geological and reservoir modelling are mandatory in studies that regard
the geological storage of CO2. The scale of the model is a function of the
phenomena that is been investigated. Modelling performed at the outcrop
scale allows to quantify the impact and the variability of the geological
parameters, further, to improve reservoir models with the most relevant
characteristics which will affect the CO2 behaviour. In this doctoral thesis, the
geological and reservoir modelling of a tsunami and barrier island‐tidal inlet
(b.i./inlet) deposits from the Camarillas Fm. were achieved at the outcrop scale
(200 m‐length, 40 m‐width and maximum 7 m of individual thickness), with a
high resolution of the grid blocks at the sub‐metric scale (1 cm in i (X or E‐W),
0.5 cm in j (Y or N‐S), and 0.2 cm in z (thickness) directions). The adopted fluid
model and scenario of simulation were hypothetical and consider the same
parameters used in other study cases and laboratory studies (see Chapter 4).
The high resolution of constructed models allowed to investigate the
behaviour of the CO2 in zones closer to the injector well. The CO2 is injected as
a dry gas and behaves as a supercritical gas at the simulated reservoir
conditions, of which the risk of leakage is important, as the CO2 moves
upwards. Thus the CO2 behaviour in the wellbore region controls the well
efficiency in terms of CO2 distribution in the reservoir and its interaction with
formation water, especially at the beginning of the CO2 injection operation.
The conclusions of this study parallel the observed sedimentary
heterogeneity (see section 2.4 in Chapter 2) and also can be hierarchized in
three orders.
First‐order heterogeneity is related to the observation scale, and can
be distinguished through detailed sedimentological studies at the basin scale.
First‐order conclusion is related to the upscalling of some features in the
reservoir scale studies for the geological storage of CO2. The studied tsunami
and barrier island‐tidal inlet deposits are laterally continuous over an area of
35 km², and were classified as fine‐coarse grained sandstones by Navarrete et
al. (2013 and 2014). Detailed studies at the basin scale of Navarrete (2015)
have showed that the deposits were generated by distinct sedimentary
process within the same sedimentary system. The features which were
166
observed here can be up‐scaled into reservoir studies of clastic back barrier
sedimentary systems; as follows.
‐ The nature and distribution of sand barrier control the
characteristics of the tsunami, washover and storm deposits. The well‐
sorted sand grains, independently of their size, eroded from the barrier
island and accumulated on back‐barrier mudflats, along with lagoonal
sediments, are well preserved and show a good potential as reservoir for
geological storage purpose. The mud and carbonate sediments should
play a role as top seal rocks and/or as flow barriers into reservoir which
would block the upwards movement of the CO2.
‐ The petrophysics models of both studied deposits reflected the
sandy variability which was represented by the facies distribution. The
facies were defined as a function of the sand sorting at the petrographic
scale. The tsunami facies and porosity distributions are homogeneous,
whereas the barrier island‐tidal inlet facies and porosity distributions are
heterogeneous.
‐ A most likely and realist reservoir model depends on the
knowledge of sedimentary model and on the correlation between the
sedimentology and petrophysics. The good correlation between facies
and petrophysics allowed to performer the petrophyscical modelling as a
function of facies model, and to produce geologically realist spatial
distributions of porosity and permeability in the reservoir.
In addition, the performed CO2 flow modelling, which mainly
conclusions are summarized in second order conclusions, provides
considerations that are also interesting to the basin scale studies, such as:
‐ Although the permeability is usually low (tens of mD), the two
deposits behaved as a reservoir. At a short‐time scale (7 years), both
reservoirs stored at least of 60% injected CO2, within 20‐40% dissolved in
the brine.
‐ The CO2 injection in thinner reservoir zones (1‐3 m) dissolved at
least 20% more CO2 compared to the injection in thicker zones such as
the southern case of barrier island‐tidal inlet deposit (5‐6 m). At the
reservoir scale, the presence of impermeable layers (< 0.5 mD) into thick
167
sandstone reservoirs (> 10 m) could improve the CO2 dissolution by
restraining the upwards movement of CO2 and creating hydrodynamism
in the reservoir.
Second‐order heterogeneity is related to the genesis of sedimentary
process and to the conditions of sediment deposition and preservation; these
heterogeneities can be recognized at the meter scale, and can be important at
the micro‐scale. Thus, the conclusions of second‐order are related to the
characteristics of each sandstone deposit which control the petrophysics
distribution, modelling process and simulation input; they are:
‐ The tsunami deposit has an important variation of sand grain
size (coarse to fine), however the sand sorting is moderated to well, and
the porosity variation is low, between 14‐22%. The deposit is a single and
instantaneous event considering the geological time scale. The dominant
sedimentary process was a turbidity flow which deposited sand
sediments principally from backflow currents of the tsunami wave train;
further, deposited sediments were not reworked or eroded afterwards.
The homogeneity of sedimentary process and the excellent preservation
of deposit controlled the homogenous sorting of sand and hence the
porosity distribution, despite of the variability in sand grain size.
‐ The barrier island‐tidal inlet deposit is composed of inlet infill
facies in the North sector and tidal channel facies in the South sector. In
the North sector, the sand is poor‐sorted and the cementation is
important, whereas in the South sector, the sand is moderate‐sorted and
the cementation is low. In an extension of 200m in length, the deposit
showed significant variation in facies and petrophysics. The porosity in
the North sector varies from 3 to 16%, and in the South sector from 11
to 22%. Moreover, the porosity varies according to facies, the higher
porosity values correspond to the well‐sorted facies and the lower
porosity values correspond to the cemented facies. This great variation
of facies and porosity is mainly related to the sedimentary nature of the
deposit. The facies associated to the inlet infill showed a fill pattern
more complex than the facies of tidal channel. Inlet deposits are
characterized by multiple episodes of erosion, migration and lateral
filling of inlets within the barrier island system, therefore, the
168
heterogeneous distribution of porosity reflects the heterogeneous
nature of the deposition process.
‐ Petrography description and petrophysics measurements were
performed at the same scale (micron‐centimetres), and showed a good
correlation between the petrographic facies (sorting facies) and the
porosity distribution. Indeed, the facies and petrophyscial modelling
represented up‐scaled microscopic (micron–centimetre) features within
macroscopic (metre–hectometre) features.
‐ The sampling spacing constraints the geostatistic analyses and
should represent the facies variability observed on field. A regular
sampling spacing and adequate concerning the sedimentary variability
gives a better analysis of the spatial variability of facies through the
variograms for stochastic modelling.
‐ The sequential indicator algorithm used in facies modelling
reproduced possible scenarios according to the sedimentary model and
the input data distribution. Nested variograms showed to be useful in
modelling of heterogeneous deposit as the barrier island‐tidal inlet
sandstone. The knowledge of sedimentary model allowed to adjust the
experimental variogram ranges which did not match with the variogram
model and then to better determine the gemetry of facies.
‐ Porosity modelling of the tsunami and barrier island‐tidal inlet
deposits was achieved through a Gaussian algorithm. The porosity
modelling of the tsunami deposit honoured its porosity distribution due
to the low variation of the porosity values. The porosity modelling of the
b.i./inlet deposit was conditioned by the facies model once the defined
facies clearly controlled the porosity distribution.
‐ The porosity and permeability have a strong correlation in both
deposits. The porosity model of both deposits conditioned the
permeability model through application of regression functions.
Exponential and power regression functions defined the permeability
variation in the tsunami and barrier island‐tidal inlet deposits,
respectively.
169
‐ Although permeability anisotropy is low, the high resolution of
reservoir model with sub‐metric size of grid blocks created a contrast of
permeability at sub‐metric scale and thus, created permeability
anisotropy in the reservoir model.
‐ The total pore volume of both reservoirs is low and the injection
rate was limited to control the pressure build‐up. A constant injection
rate seems to improve the CO2 dissolution and to decrease the maximum
gas saturation into the gas plume in the zone close to the injector well.
‐ The results from simulation studies which regard the behaviour
of CO2 under storage conditions depend on the petrophysics model, fluid
model, reservoir conditions and simulator capacities. At the sub‐metric
scale, under the same reservoir conditions and fluid model parameters,
the thickness of deposit has the major impact in the amount of CO2
dissolution rather than the permeability contrast.
‐ In the tsunami reservoir, the location of the injector well in a
more homogenous zone with better permeability values improves the
CO2 dissolution and decreases the gas saturation in the gas plume. The
amount of dissolved CO2 in the northern b.i./inlet cases (thinner zones) is
similar to the amount of dissolved CO2 in tsunami cases; however the
injection rate of b.i./inlet case was higher.
Third‐order heterogeneity is related to external elements to
sedimentary system such as palaeo‐relief and/or the presence of syn‐
sedimentary faults, which could locally change flow conditions. Moreover syn‐
sedimentary faults play an important role in fluid remobilization during
compaction and diagenesis, as well as in deeper conditions as conduct or
barrier to fluid flow. Third‐order conclusions are related to the presence of
such external elements and their implication in the petrophysics of deposits at
the metric scale.
‐ At the sandstone bed scale, the intense cementation in the
North sector of b.i./inlet deposit seems to be related to the proximity
with the Remenderuelas and other minor associated syn‐sedimentary
normal faults, and to the inlet infill nature of facies in this sector. Faults
might remobilize the fluid during and after diagenesis, thus the
170
cementation could be located close to faults. On the order hand, the
cementation is related to facies which contain oysters, as observed by
Jackson and Rawn‐Schatzinger (1993) and here (see Chapter 2).
The perspective of this study concerns the Camarillas Fm and its
potential as reservoir for geological storage of CO2. Therefore, the presence of
syn‐sedimentary faults has also an impact in the unit scale. The lower
permeability (mD) observed in both deposits is related to compaction and
probably to cementation of remobilized fluid through faults. Cementation
might be affecting only specific zones of Camarillas Fm, thus the radius of fault
influence into sandstone beds could be restrained by certain distance (tens of
meters?) from fault. A structural and digenetic study of Camarillas Fm. could
elucidate the relationship between faults and cementation. Moreover, the
hydraulic behaviour of syn‐sedimentary faults conducts or blocks the flow of
injected CO2. The simulation studies at the sedimentary formation scale should
take into account the hydraulic behaviour of faults and their vertical extension
to quantify their radius of influence and risk of leakage.
Finally, a study to estimate the geological storage capacity of the
Camarillas Fm. is interesting, as the Camarillas Fm. is located close to a major
coal‐fired power plant and the transport of CO2 within the Carbon Capture and
Storage (CCS) technology, from the emission source to the storage site, is risky
and costly. On the other hand, the Camarillas Fm is composed of interbedded
sandstones with shale and marls which could be considered, for the storage
purpose, as a multi‐layer reservoir. The shale and marls play a role of flow
barrier to the upwards movement of CO2, while the thinner reservoir zones
enhance the CO2 dissolution. The injection in thin reservoirs needs an
exhaustive monitoring of the injection rate and pressure build‐up to avoid
reservoir and top seal damages, whereas the injection in thick reservoir zones
improves the injection rate with a lower pressure build‐up. However, the
injected CO2 in thin reservoir zones (< 3m) showed a higher rate of dissolution
in the brine over few years (7 years). Therefore, a compromise between the
needs of storing high amount of CO2 and increasing the CO2 dissolution in a
short‐time scale is needed. Moreover, the lower permeability zones located
close to normal syn‐sedimentary faults suggest that areas of better porosity
and permeability values should be located outside the zone of fault influence.
171
The possible presence of zones with better petrophysics values is a plus in the
interest of the Camarillas Fm. as reservoir for geological storage purposes.
172
173
Conclusiones y Perspectiva
La modelización geológica y de reservorios es obligatoria en los
estudios sobre almacenamiento geológico de CO2. La escala del modelo se
define en función de los fenómenos investigados. La modelización a escala de
afloramiento permite la cuantificación del impacto de dicho almacenamiento y
la variabilidad de los parámetros geológicos, con el fin de mejorar los modelos
de reservorios teniendo en cuenta las características más relevantes que
afectarán al comportamiento del CO2. En esta tesis doctoral se ha abordado la
modelización geológica y de reservorio de dos cuerpos arenosos de la Fm.
Camarillas correspondientes al depósito de un episodio de tsunami y de un
complejo isla barrera/inlet realizados a escala de afloramiento (200 m de
longitud, 40 m de ancho y un máximo de 7 m de espesor individual), con una
alta resolución de los bloques de cuadrícula en la escala sub‐métrica: 1 cm en
la dirección i (X o dirección E‐W), 0,5 cm en la j (Y o dirección N‐S), y 0,2 cm en
la dirección z (espesor). El modelo de fluido adoptado y el escenario de la
simulación utilizados fueron hipotéticos, pero están de acuerdo con los
parámetros que se encuentran en otros casos descritos en la literatura y en
estudios de laboratorio (ver Capítulo 4). La alta resolución de los modelos
elaborados permitió investigar el comportamiento del CO2 en las zonas
cercanas al pozo inyector. El CO2 se inyecta en forma de gas seco y se
comporta como gas supercrítico en las condiciones del yacimiento simulado.
En estas condiciones el riesgo de fuga es importante ya que el CO2, como fase
gaseosa, se mueve hacia arriba. Por lo tanto, el comportamiento del CO2 en el
área cercana al pozo controla la eficiencia, en términos de distribución del CO2
en el reservorio, y su interacción con el agua de formación, especialmente
durante el comienzo de la operación de inyección.
Las conclusiones de este estudio están estrechamente ligadas a la
heterogeneidad sedimentaria observada (ver sección 2.4 en el Capítulo 2) y
también pueden ser jerarquizadas en tres órdenes.
Las heterogeneidades de primer orden están relacionadas con la
escala de observación, y se pueden identificar a través de estudios
sedimentológicos detallados a escala de cuenca. Las conclusiones de primer
orden están relacionadas con la aplicación que los estudios de reservorio a
174
escala de afloramiento para el almacenamiento geológico de CO2 puedan
tener a mayor escala (escala de cuenca). Los depósitos de tsunami y de isla
barrera /inlet son lateralmente continuos en un área de 35 km², y se
clasificaron como areniscas de grano fino a grueso por Navarrete et al. (2013 y
2014). Los estudios detallados a escala de la cuenca realizados por Navarrete
(2015), han mostrado que estos depósitos se generaron por procesos
sedimentarios distintos dentro de un mismo sistema sedimentario (sistemas
de islas barrera). Las principales conclusiones de este estudio que se pueden
aplicar a mayor escala en estudios de reservorios de sistemas sedimentarios
clásticos de back barrier son:
‐ La naturaleza y distribución espacial de las barreras arenosas
controlan las características de los depósitos de tsunami, de washover y
de tormenta. Las areniscas que componen estos depósitos están bien
seleccionadas, y con independencia de su tamaño de grano, proceden de
la erosión de la isla barrera y se depositan en las llanuras fangosas de
backbarrier y en los sedimentos margosos del lagoon, por lo que tienen
una buena preservación y potencial como reservorio para fines de
almacenamiento. Los sedimentos lutíticos y carbonatados deben jugar
un papel como rocas sello y / o como barreras de flujo dentro del
reservorio que bloquearían el movimiento hacia arriba del CO2.
‐ Los modelos petrofísicos de ambos depósitos reflejan la
variabilidad de las areniscas, representada por la distribución de facies.
Las facies han sido definidas como una función de la selección de las
areniscas a escala petrográfica. La distribución de las facies y de la
porosidad del depósito de tsunami es homogénea, mientras que en el
depósito de isla barrera/inlet, tanto la distribución de las facies como la
distribución de porosidad son heterogéneas.
‐ Un modelo de reservorio más probable y realista depende del
conocimiento del modelo sedimentario y de la relación que se establece
entre la sedimentología y la petrofísica. Una buena correlación entre
facies y petrofísica permitió ejecutar la modelización petrofísica en
función del modelo de facies, y producir distribuciones espaciales,
geológicamente realistas, de la porosidad y la permeabilidad dentro del
reservorio.
175
Además, la modelización de flujo de CO2 realizada, cuyas conclusiones
se resumen en las conclusiones de segundo orden, permite hacer algunas
consideraciones que también pueden ser interesantes para los estudios a
escala de la cuenca, tales como:
‐ A pesar de que la permeabilidad es generalmente baja (decenas
de mD), ambos depósitos se comportaron como reservorios. A una
escala de tiempo corto (7 años), ambos modelos de reservorios
almacenan al menos un 60% de todo el CO2 inyectado, con entre un 20‐
40% de él disuelto en la salmuera.
‐ La inyección de CO2 en las zonas más delgadas de los reservorios
(1‐3 m) disuelve al menos un 20% más de CO2 que cuando la inyección se
realiza en las zonas más potentes (5‐6 m), como en el caso de la
inyección en el sector meridional del depósito de la isla barrera / inlet. A
escala del reservorio, la presencia de capas impermeables (<0,5 mD)
dentro de depósitos potentes arenosos (> 10 m) podría mejorar la
disolución de CO2 por disminución del movimiento ascendente del CO2 y
creación de hidrodinamismo en el depósito.
Las heterogeneidades de segundo orden se relacionan con la génesis
del proceso sedimentario y con las condiciones de depósito de los sedimentos
y su preservación. Estas heterogeneidades pueden ser reconocidas a escala
métrica, y pueden ser importantes a escala micro. Por lo tanto, las
conclusiones de segundo orden están relacionadas con las características de
cada depósito arenoso, que controlan tanto la distribución petrofísica
obtenida, como el proceso de modelización y de simulación. Éstas son:
‐ El depósito de tsunami tiene una importante variación de
tamaños de grano (desde arena gruesa a fina), sin embargo su selección
es de moderada a buena, y la variación de la porosidad es baja, entre el
14‐22%. Este depósito es un evento único y considerado instantáneo a
escala de tiempo geológico. El proceso sedimentario dominante fueron
flujos turbidíticos que depositaron sedimentos arenosos principalmente
desde corrientes de reflujo de trenes de olas del tsunami y cuyos
sedimentos depositados no fueron retrabajados o erosionados
posteriormente. Este proceso sedimentario y la excelente preservación
del depósito controlan la selección arenosa y, por tanto, la distribución
176
de la porosidad, a pesar de la variabilidad observada en el tamaño de
grano.
‐ En el depósito de isla barrera / inlet se han identificado facies de
relleno de inlet en el sector norte del afloramiento y facies de canal
mareal en el sector sur. En el sector norte, las areniscas están mal
seleccionadas y la cementación es importante, mientras que en el sector
sur, la arena está moderadamente seleccionada y la cementación es
baja. En una extensión de 200 m de longitud, este depósito muestra
variaciones significativas tanto en las facies como en las características
petrofísicas. La porosidad en el sector norte varía del 3 al 16%, y en el
sector Sur desde 11 a 22%. Por otra parte, la porosidad varía en función
de las facies, de tal manera que los valores de las porosidades más altas
corresponden a facies bien seleccionadas, mientras que los valores de
porosidad más bajos corresponden a las facies cementadas. Esta gran
variación de facies y de porosidades se relaciona principalmente con los
diferentes subambientes y procesos sedimentarios del depósito. Las
facies asociadas al inlet muestran un patrón de relleno mucho más
complejo que las facies de canal mareal. Los depósitos de inlet se
caracterizan por episodios múltiples de erosión, migración y de relleno
lateral, por lo tanto, la distribución heterogénea de la porosidad refleja
la heterogeneidad y complejidad de los procesos sedimentarios
involucrados.
‐ La descripción petrográfica y las mediciones petrofísicas se
realizaron a la misma escala (micras‐centímetros), y se ha demostrado
una buena correlación entre las facies petrográficas (selección de grano)
y la distribución de la porosidad. De hecho, la modelización de facies y la
modelización petrofísica representan una integración de la escala micro
(micras‐centímetros) dentro de las características de la macro‐escala
(metro‐hectómetro).
‐ El espaciado del muestreo limita el análisis geoestadístico y debe
representar la variabilidad de facies observadas en el campo. Un
muestreo con espaciado regular, que además sea adecuado a la
variabilidad sedimentaria, proporciona un mejor análisis de la
177
variabilidad espacial de facies, a través de los variogramas, para la
modelización estocástica.
‐ El indicador secuencial del algoritmo de simulación utilizado en
la modelización de facies reproduce posibles escenarios de acuerdo al
modelo sedimentario y a la distribución de los datos de entrada. Los
variogramas nested mostraron ser útiles en la modelización de depósitos
heterogéneos como las areniscas del depósito de isla barrera/inlet. El
conocimiento del modelo sedimentario permitió ajustar los rangos de
variogramas experimentales que no coincidían con el modelo del
variograma y estimar mejor la longitud de la variación de facies.
‐ La modelización de la porosidad en ambos depósitos se llevó a
cabo a través de un algoritmo Gausiano. La modelización de la porosidad
del depósito de tsunami se realizó a través de la distribución de
porosidad, debido a la baja variación de dichos valores. El modelo de la
porosidad del depósito de isla barrera/inlet está condicionado por el
modelo de facies, ya que las facies definidas controlan claramente la
distribución de la porosidad.
‐ La porosidad y permeabilidad tienen una buena correlación en
ambos depósitos. El modelo de porosidad de ambos depósitos
condicionó el modelo de permeabilidad a través de funciones de
regresión. Funciones exponenciales y de regresión potencial definen la
variación de la permeabilidad en los depósitos de tsunami y de isla
barrera/inlet, respectivamente.
‐ Aunque la anisotropía de la permeabilidad es muy baja, la alta
resolución del modelo de reservorio con un tamaño sub‐métrico de los
bloques de la cuadrícula genera un contraste de la permeabilidad a
escala sub‐métrica creando la anisotropía de la permeabilidad en el
modelo de reservorio.
‐ El volumen total poroso de ambos depósitos es bajo y la
velocidad de inyección se limitó a controlar la presión acumulada. Una
velocidad de inyección constante mejoró un poco la disolución del CO2 y
disminuyó la saturación máxima del CO2 en la pluma gaseosa en la zona
cercana al pozo inyector.
178
‐ Los resultados de los estudios de simulación que consideran el
comportamiento del CO2 en condiciones de almacenamiento dependen
del modelo petrofísico, del modelo de fluido, de las condiciones del
yacimiento y de las capacidades del simulador. A la escala submétrica, a
las mismas condiciones del yacimiento y con los mismos parámetros del
modelo de fluido, la potencia del reservorio mostró tener un mayor
impacto en la cantidad de disolución de CO2 que el contraste de
permeabilidad.
‐ En el depósito tsunami, la ubicación del pozo inyector en el
sector meridional, que es una zona algo más homogénea y de mayor
permeabilidad, mejora un poco la disolución del CO2 y disminuye su
concentración en la pluma de gas. La cantidad de CO2 disuelto en los
casos de inyección simulados en el sector septentrional del sistema isla
barrera / inlet, correspondiente a la zona menos potente, es similar a la
cantidad de CO2 disuelto en todos los casos estudiados del depósito de
tsunami; sin embargo, su tasa de inyección fue mayor.
Las heterogeneidades de tercer orden están relacionadas con
elementos externos al sistema sedimentario como la presencia de
paleorelieves y /o fallas sinsedimentarias, que pueden modificar localmente
las condiciones de flujo. Además, las fallas sinsedimentarias juegan un papel
importante en la removilización de los fluidos durante la compactación y
diagénesis, e incluso en condiciones más profundas, pudiendo actuar como
vías preferentes de circulación o como barreras al flujo. Las conclusiones de
tercer orden están relacionadas con la presencia de estos elementos externos
y su implicación en la petrofísica de los depósitos a escala métrica.
‐ A escala de las capas de areniscas, las zonas altamente
cementadas en el sector norte del depósito de isla barrera/inlet están
probablemente relacionadas con la proximidad a la falla normal
sinsedimentaria de Remenderuelas y otras fallas normales menores
asociadas, así como con las características de las facies de relleno de
inlet en este sector. Estas fallas podrían permitir la removilización de
fluidos durante la diagénesis o posteriormente a ella y cementar zonas
cercanas a éstas. Por otro lado, la cementación también podría estar
179
relacionada con facies de ostreídos, según lo observado por Jackson y
Rawn‐Schatzinger (1993) y en este estudio (ver Capítulo 2).
La perspectiva de este estudio se refiere al potencial que tiene la Fm.
Camarillas como almacén geológico de CO2, y en este sentido, la presencia de
fallas sinsedimentarias tienen un impacto a escala de yacimiento. A escala de
yacimiento, la baja permeabilidad (mD) observada en ambos depósitos podría
relacionarse con la compactación y con la cementación por removilización de
fluidos a través de fallas. Esta cementación estaría afectando solamente a
zonas específicas de la Fm. Camarillas, debido a que su radio de influencia
dentro de las capas de areniscas parece estar limitado a una cierta distancia
(decenas de metros) de las fallas. Un estudio más profundo tanto estructural
como diagenético de la Fm. Camarillas podría dilucidar la relación entre las
fallas sinsedimentarias y la cementación carbonatada. Además, el
comportamiento hidráulico de las fallas sinsedimentarias podría permitir o
bloquear el flujo de CO2 inyectado. Los estudios de simulación a escala de
formación sedimentaria deben tener en cuenta el comportamiento hidráulico
de las fallas y su extensión vertical para cuantificar su radio de influencia y el
potencial riesgo de fuga.
Por último, sería interesante estimar la capacidad de
almacenamiento de la Fm. Camarillas, sobre todo teniendo en cuenta que esta
unidad está relativamente cerca de una gran planta de energía térmica de
carbón y que el transporte de CO2, dentro de la tecnología de captura y
almacenamiento de carbono (CCS), desde la fuente de emisión a la zona de
almacenamiento, es arriesgado y costoso. Por otro lado, la Fm. Camarillas se
compone de areniscas con intercalaciones de arcillas y margas pudiendo
considerarse, para el almacenamiento propuesto, como un reservorio multi‐
capa. Las arcillas y margas juegan un papel de barrera para el movimiento
ascendente del CO2, mientras que las zonas de reservorios menos potentes, de
escala métrica, mejoran la disolución de CO2. La inyección en estos reservorios
poco potentes necesita una exhaustiva monitorización de la velocidad de
inyección y de las presiones acumuladas a fin de evitar daños en el reservorio y
en el sello, mientras que la inyección en zonas de reservorios potentes mejora
la velocidad de inyección con una presión acumulada más baja. Sin embargo,
la inyección de CO2 en reservorios delgados (< 3m) permite una mayor tasa de
180
disolución del CO2 en la salmuera en unos pocos años (7 años). Esto nos obliga
a encontrar un compromiso entre la necesidad de almacenar grandes
cantidades de CO2 y la necesidad de aumentar la disolución de CO2 en una
escala de tiempo corto (que haría el almacenamiento más seguro en el menor
tiempo posible). Además, las zonas de menor permeabilidad localizadas cerca
de fallas sinsedimentarias normales, sugieren que los valores de mayor
porosidad y permeabilidad se deben localizar fuera de su zona de influencia. La
posible presencia de zonas con mejores valores petrofísicos es un dato a tener
en cuenta para valorar el interés de la Fm. Camarillas como reservorio para
fines de almacenamiento geológico.
181
REFERENCES
182
183
REFERENCES
AAPG (American Association of Petroleum Geologist), 2015. AAPG Wiki.
ttp://wiki.aapg.org/Main_Page.
Akaku, K., 2008. Numerical Simulation of CO2 Storage in Aquifers without
Trapping Structures. International Petroleum Technology Conference, 3‐5
December, Kuala Lumpur, Malaysia. IPTC‐12304‐MS.
Alvaro, M., Capote, R., Vegas. R., 1979. Un modelo de evolución geotectónica
para la Cadena Celtibérica. Acta Geológica Hispanica 14, 172–177.
Ambrose, W.A., Lakshminarasimhan, S., Holtz, M.H., Núñez‐López, V., Hovorka,
S.D., Duncan, I., 2008. Geologic factors controlling CO2 storage capacity and
permanence: case studies based on experience with heterogeneity in oil and
gas reservoirs applied to CO2 storage. Environmental Geology 54, 1619–1633.
Antolín‐Tomás, B., Liesa, C.L., Casas, A.M., Gil‐Peña, I., 2007. Geometry of
fracturing linked to extension and basin formation in the Maestrazgo basin
(Eastern Iberian Chain, Spain). Revista de la Sociedad Geológica de España 20,
3‐4, 351‐365.
Apotsos, A., Gelfenbaum, G., Jaffe, B., 2012. Time‐dependent onshore tsunami
response. Coastal Engineering 64, 73–86.
Asharf, M., 2014. Geological storage of CO2: Heterogeneity impact on the
behaviorof pressure. International Journal of Greenhouse Gas Control 28, 356–
368.
Bachu, S., Bonijoly, D., Bradshaw, J., Burruss, R., Holloway, S., Christensen,
N.P., Mathiassen, O.M., 2007. CO2 storage capacity estimation: Methodology
and gaps. International journal of greenhouse gas control 1, 430–443.
Bauluz, B., Yuste, A., Mayayo, M. J., Canudo, J. I., 2014. Early kaolinization of
detrital Weald facies in the Galve Sub‐basin (Central Iberian Chain, north‐east
Spain) and its relationship to palaeoclimate. Cretaceous Research 50, 214‐227.
Bersezio, R., Bini, A., Giudici, M., 1999. Effects of sedimentary heterogeneity
on groundwater flow in a Quaternary pro‐glacial delta environment: joining
facies analysis and numerical modelling. Sedimentary Geology 129, 327–344.
184
Bohling, G., 2007. Introduction to Geostatistics. Courses notes of
Hydrogeophysics: Theory, Methods, and Modeling. Boise State University,
Idaho, USA. http://people.ku.edu/~gbohling/BoiseGeostat.
Boxiao, L., Tchelepi, H.A., Benson, S.M. 2013. The Influence of Capillary Entry‐
Pressure Representation on CO2 Solubility Trapping. Energy Procedia 37, 3808
– 3815.
Caja, M.A., 2004. Procedencia y diagénesis de los sedimentos del Jurásico
superior‐Cretácico inferior (facies Weald) en las subcuencas occidentales de la
Cuenca del Maestrazgo, Cordillera Ibérica Oriental. Tesis Doctoral, Univ.
Complutense de Madrid, 293 p.
Cantalamessa, G., Di Celma, C., 2005. Sedimentary features of tsunami
backwash deposits in a shallow marine Miocene setting, Mejillones Peninsula,
northern Chile. Sedimentary Geology 178, 259–273.
Capote, R., Muñoz, J.A., Simón, J.L., Liesa, C.L., Arlegui, L.E., 2002. Alpine
tectonics I: the Alpine system north of the Betic Cordillera. In: Gibbons, W.,
Moreno, T. (eds), Geology of Spain. The Geological Society, London, 367–400.
Celia, M.A., Bachu, S., Nordbotten, J.M., Gasda, S.E, Dahle, H.K., 2005.
Quantitative estimation of CO2 leakage from geological storage: analytical
models, numerical models, and data needs. Greenhouse Gas Control
Technologies 7, I, 663‐671.
Chadwick, A., Arts,R., Bernstone,C., May,F., Thibeau, S., Zweigel, P., 2008 ‐ Best
practice for the storage of CO2 in saline aquifers. British Geological Survey,
Occasional Publication 14, 277 p.
Chagué‐Goff, C., Schneider, J.L., Goff, J.R., Dominey‐Howes, D., Strotz, L., 2011.
Expanding the proxy toolkit to help identify past events—lessons from the
2004 Indian Ocean Tsunami and the 2009 South Pacific Tsunami. Earth‐Science
Reviews, 107 (1), 107‐122.
Chang, C., Zhou, Q., Guo, J., Yu, Q., 2014. Supercritical CO2 dissolution and
mass transfer in low‐permeability sandstone: Effect of concentration
difference in water‐flood experiments. International Journal of Greenhouse
Gas Control 28, 328–342.
185
Clague, J.J., Bobrowsky, P.T., Hamilton, T.S., 1994. A sand sheet deposited by
the 1964. Alaska tsunami at Port Alberni, British Columbia. Estuarine, Coastal
and Shelf Science 38 (4), 413–421.
Clark, I., 2001. Practical Geostatistics. Author home page:
http://www.kriging.com/PG1979/PG1979_pdf.html.
Collison, J.D., 1996. Alluvial sediments. In: Reading, H.G. (ed), Sedimentary
Environments and Facies. Blackwell Publishing, Oxford, 37–82.
Corbett, P.W.M., Potter, D.K., 2004. Petrotyping: a basemap and atlas for
navigating through permeability and porosity data for reservoir comparison
and permeability prediction. International Symposium of the Society of Core
Analysts, Abu Dhabi, UAE, 5‐9 October.
Corey, A.T., 1954. The interrelation between gas and oil relative
permeabilities. Producers Monthly 19 (November): 38–41.
Curtis, C.D., 1978. Possible links between sandstone diagenesis and depth‐
related geochemical reactions occurring in enclosing mudstones. Journal of the
Geological Society 135(1), 107‐117.
Cushman, J.H., 1997. The Physics of Fluids in Hierarchical Porous Media:
Angstroms to Miles. Springer, Heidelberg. 484 p.
Dalrymple, M., 2001. Fluvial reservoir architecture in the Statfjord Formation
(northern North Sea) augmented by outcrop analogue statistics. Petroleum
Geoscience 7, 115‐122.
Davis, R.A., Barnard, P.L., 2000. How anthropogenic factors in the back‐barrier
area influence tidal inlet stability: examples from the Gulf Coast of Florida,
U.S.A. In: Pye, K., Allen, J.R.L. (eds), Coastal and Estuarine Environments:
Sedimentology, Geomorphology and Geoarchaeology. Geol. Soc. London Spec.
Publ. 175, 293‐303.
Davis, R.A., Barnard, P.L., 2003. Morphodynamics of the barrier‐inlet system,
west‐central Florida. Marine Geology 200, 77‐101.
Davis, R.A., Gibeaut, J.C., 1990. Historical Morphodynamics of Inlets in Florida:
Models for Coastal Zone Planning. Tech Paper 55. Florida Sea Grant College,
Gainesville, FL, 81 p.
186
Dawson, A.G., Stewart, I., 2007. Tsunami deposits in the geological record.
Sedimentary Geology 200, 3, 166‐183.
Dawson, S., Smith, D.E., 2000. The sedimentology of Middle Holocene tsunami
facies in northern Sutherland, Scotland, UK. Marine Geology, 170, 1, 69‐79.
Deng, H., Stauffer, P.H., Dai, Z., Jiao, Z., Surdam, R.C., 2012. Simulation of
industrial‐scale CO 2 storage: Multi‐scale heterogeneity and its impacts on
storage capacity, injectivity and leakage. International Journal of Greenhouse
Gas Control 10, 397‐418.
Deutsch, C.V., Journel, A.G., 1998. GSLIB‐ Geostatistical Software Library and
User's Guide, Second Edition. New York Oxford. Oxford University Press, 376 p.
DeVries Klein, G., 1974. Estimating water depths from analysis of barrier island
and deltaic sedimentary sequences. Geology 2, 8, 409‐412.
Doughty, C., 2007. Modeling geologic storage of carbon dioxide: Comparison
of non‐hysteretic and hysteretic characteristic curves. Energy Conversion and
Management 48, 1768–1781.
Eaton, T.T., 2006. On the importance of geological heterogeneity for flow
simulation. Sedimentary Geology 184, 187–201.
Eaton, T.T., Bradbury, K.R., 2003. Hydraulic transience and the role of bedding
fractures in a bedrock aquitard, southeastern Wisconsin, USA. Geophysical
Research Letters 30, 18. doi:10.1029/2003GL017913.
EEA – (European Environment Agency), 2014. Annual European Union
greenhouse gas inventory 1990–2012 and inventory report 2014. Technical
report No 09/2014.
Ekeland, A., Pedersen, N., Howell, J., Nemec, W., Keogh, K., Viste, I., 2008.
Modeling of Intra‐Channel Belt Depositional Architecture in Fluvial Reservoir
Analogs from the Lourinha Formation, Portugal. Search and Discovery Article
number: 50146.
EU, 2009. European Parliament and of the Council Directive 2009/31/EC on the
geological storage of carbon dioxide and amending Council Directive
85/337/EEC, European Parliament and Council Directives 2000/60/EC,
187
2001/80/EC, 2004/35/EC, 2006/12/EC, 2008/1/EC and Regulation (EC) No
1013/2006.
Felletti, F., 2004. Statistical modelling and validation of correlation in
turbidites: an example from the Tertiary Piedmont Basin (Castagnola Fm.,
Northen Italy). Marine and Petroleum Geology 21, 23‐39.
Fenghour, A., Wakeham, W.A., and Vesovic, V., 1998. The Viscosity of Carbon
Dioxide. Journal of Physical and Chemical Reference Data 27, 1, 31‐44.
Foix, N., Paredes, J.M., Giacosa, R.E., 2013. Fluvial architecture variations
linked to changes in accommodation space: Río Chico Formation (Late
Paleocene), Golfo San Jorge basin, Argentina. Sedimentary Geology 294, 342–
355.
Folk, R.L., 1980. Petrology of Sedimentary Rocks. Hemphill Publish Company.
Austin, USA. 184 p.
Frykman, P., 2009. The importance of geological heterogeneities. CO2GeoNet
Open Forum March 18‐20, Venice, San Servolo Island, Italy.
Frykman, F., Nielsen, C.M., Bech, N., 2013. Trapping effects of small scale
sedimentary heterogeneities. Energy Procedia 37, 5352 – 5359.
Fujiwara, O., 2008. Bedforms and sedimentary structures characterizing
tsunami deposits. In: Shiki, T., Tsuji, Y., Yamazaki, T., Minoura, K. (eds),
Tsunamiites — Features and Implications. Elsevier, Amsterdam, Oxford, 51–62.
Fujiwara, O., Kamataki, T., 2007. Identification of tsunami deposits considering
the tsunami waveform: an example of subaqueous tsunami deposits in
Holocene shallow bay on southern Boso Peninsula, Central Japan. Sedimentary
Geology 200, 3, 295‐313.
García‐García, F., Ferñandez, J., Viseras, C., Soria, J.M., 2006. Architecture and
sedimentary facies evolution in a delta stack controlled by fault growth (Betic
Cordillera, southern Spain, late Tortonian). Sedimentary Geology 185, 79 – 92.
Ghazi, S., Mountney, N.P., 2009. Facies and architectural element analysis of a
meandering fluvial succession: the Permian Warchha Sandstone, Salt Range,
Pakistan. Sedimentary Geology 221, 99–126.
188
Guimerà, J.J,. 1988. Estudi estructural de l'enllaç entre la Serralada Ibèrica i la
Serralada Costanera Catalana. PhD Thesis, Universitat de Barcelona, 600p.
Guimerà, J., Salas, R., Vergés, J., Casas, A., 1996. Extensión mesozoica e
inversión compresiva terciaria en la Cadena Ibérica: aportaciones a partir del
análisis de un perfil gravimétrico. Geogaceta 20 , 7, 7697‐7694.
Hammer, E., Mørk, M.B.E., Næss, A., 2010. Facies controls on the distribution
of diagenesis and compaction in fluvial‐deltaic deposits. Marine and Petroleum
Geology, 27, 8, 1737‐1751.
Heinz, J., Kleineidam, S., Teutsch, G., Aigner, T., 2003. Heterogeneity patterns
of Quaternary glaciofluvial gravel bodies (SWGermany): application to
hydrogeology. Sedimentary Geology 158, 1–23.
Henares, S., Caracciolo, L., Cultrone, G., Fernández, J., Viseras, C., 2014. The
role of diagenesis and depositional facies on pore system evolution in a
Triassic outcrop analogue (SE Spain). Marine and Petroleum Geology 51, 136‐
151.
Hodgkinson, J., Cox, M.E, McLoughlin, S., Huftile, G.J., 2008. Lithological
heterogeneity in a back‐barrier sand island: Implications for modelling
hydrogeological frameworks. Sedimentary Geology 203, 64–86.
Hornung, J., Aigner,T., 1999. Reservoir and aquifer characterization of fluvial
architectural elements: Stubensandstein, Upper Triassic, southwest Germany.
Sedimentary Geology 129, 215–280.
Hovorka, S.D., Doughty, C., Benson, S.M., Pruess K., Knox P.R., 2004. The
impact of geological heterogeneity on CO2 storage in brine formations: a case
study from the Texas Gulf Coast. In: Baines S.J., Worden R.H. (eds), Geological
storage of carbon dioxide. Geological Society of London Special Publication
233, 147–163.
Huysmans, M., Peeters, L., Moermans, G., Dassargues, A., 2008. Relating small‐
scale sedimentary structures and permeability in a cross‐bedded aquifer.
Journal of Hydrology 361, 41– 51.
IAPWS (The International Association for the Properties of Water and Steam),
1997. Industrial Formulation 1997 for the Thermodynamic Properties of Water
and Steam.
189
IEAGHG (IEA Greenhouse Gas R&D Programme), 2009. Summary report of 1st
IEA‐GHG CO2 Geological Storage Modelling Network Meeting. BRGM, Orleans,
France. http://ieaghg.org/docs/Modelling/2009‐5_Summary.pdf.
IFP (Instituto Petrofísico de Madrid), 2012. Helium Porosity and Permeability
measurements. Technical Report.
IGME (Instituto Geológico y Minero de España), 2012. Panorama Minero.
http://www.igme.es/PanoramaMinero/actual/CARBON12.pdf
IPCC (Intergovernmental Panel on Climate Change), 2005. IPCC Special Report
on Carbon Dioxide Capture and Storage. Cambridge University Press, New
York. 431 p.
Isaaks, E.H., Srivastava, R. M., 1989. Applied geostatistics (Vol. 2). New York:
Oxford University Press, 561 p..
Issautier, B., Fillacier, S., Le‐Gallo, Y., Audigane, P., Chiaberger, C., Viseur, S.,
2013. Modelling of CO2 Injection in Fluvial Sedimentary Heterogeneous
Reservoirs. Energy Procedia 37, 5181 – 5190.
Issautier, B., Viseur, S., Audigane, P., le Nindre, Y.M., 2014. Impacts of fluvial
reservoir heterogeneity on connectivity: Implications in estimating geological
storage capacity for CO2. International Journal of Greenhouse Gas Control 20,
333–349.
Jackson, S.R., Rawn‐Schatzinger, V., 1993. Data from selected Almond
Formation outcrops‐‐Sweetwater County, Wyoming (No. NIPER‐‐724). National
Inst. for Petroleum and Energy Research, Bartlesville, OK (United States).
Jankaew, K., Atwater, B.F., Sawai, Y., Choowong, M., Charoentitirat, T., Martin,
M.E., Prendergast, A., 2008. Medieval forewarning of the 2004 Indian Ocean
Tsunami in Thailand. Nature 455, 1228–1231.
Juanes, R., Spiteri, E.J., Blunt, M.J., 2006. Impact of relative permeability
hysteresis on geological CO2 storage. Water Resources Research, 42, W12418,
doi:10.1029/2005WR004806.
King M.B., Mubarak A., Kim J.D., Bott T.R., 1992. The mutual solubilities of
water with supercritical and liquid carbon dioxide. Supercritical Fluids 5, 296–
302.
190
Klingbeil, R., Kleineidam, S., Asprion, U., Aigner, T., Teutsch, G., 1999. Relating
lithofacies to hydrofacies: outcrop‐based hydrogeological characterisation of
quaternary gravel deposits. Sedimentary Geology 129, 3‐4, 299–310.
Koltermann, C.E., Gorelick, S., 1996. Heterogeneity in sedimentary deposits: a
review of structure imitating, process‐imitation, and descriptive approaches.
Water Resource Research 32, 9, 2617–2658.
Lantuéjoul, C., Beucher, H., Chilès, J.P., Lajaunie, C., Wackernagel, H., Elion, P.,
2005. Estimating the trace length distribution of fractures from line sampling
data. In Leuangthong, O., Deutsch, C. V. (eds), Geostatistics Banff 2004.
Quantitative Geology and Geostatistics 14, 165–174. Springer Netherlands,
Dordrecht.
Le‐Gallo, Y., 2009 ‐ Injectivity. Session 2 Assessment Objectives for Modelling.
CO2 Geological Storage Modelling Workshop, 10‐12 February, Orleans, France.
Organised by IEA Greenhouse Gas R&D Programme.
Le Gallo, Y., Fillacier, S., Lecomte, A., Munier, G., Hanot, F., Quisel, N.,
Rampnoux, N., Thomas, S., 2010. Technical challenges in characterization of
future CO2 storage site in a deep saline aquifer in the Paris basin. Lessons
learned from practical application of site selection methodology. Proceedings
of the 10th International Conference on Greenhouse Gas Control Technologies
(GHGT‐10), 19‐23 September 2010, Amsterdam.
Liesa, C., Soria, A.R., Meléndez, A., 2000. Lacustrine evolution in a basin
controlled by extensional faults: the Galve subbasin, Teruel, Spain. In:
Gierlowski‐Kordesch, E.H., Kelts, K.R. (eds), Lake Basins through Space and
Time. American Association of Petroleum Geologists, Studies in Geology, 46,
295–302.
Liesa, C.L., Casas, A.M., Soria, A.R., Simón, J.L., Meléndez, A., 2004. Estructura
extensional cretácica e inversión terciaria en la región Aliaga‐Montalbán. In:
Colombo, F., Liesa, C.L., Meléndez, G., Pocoví, A., Sancho, C., Soria, A.R. (eds),
Itinerarios Geológicos por Aragón. Sociedad Geológica de España, Geo‐Guías 1,
151–180.
191
Liesa, C.L., Soria, A.R., Meléndez, N., Meléndez, A., 2006. Extensional fault
control on the sedimentation patterns in a continental rift basin: El Castellar
Formation, Galve sub‐basin, Spain. Geological Society of London 163, 487–498.
Machent, P.G., Taylor, K.G., Macquaker, J.H.,Marshall, J.D., 2007. Patterns of
early post‐depositional and burial cementation in distal shallow‐marine
sandstones: Upper Cretaceous Kenilworth Member, Book Cliffs, Utah, USA.
Sedimentary Geology 198, 1, 125‐145.
Mallinson, D.J., Smith, C.W., Culver, S.J., Riggs, S.R., Ames, D., 2010. Geological
characteristics and spatial distribution of paleo‐inlet channels beneath the
outer banks barrier islands, North Carolina, USA. Estuarine, Coastal and Shelf
Science 88, 175‐189.
Martin, J.J., 1979. Cubic Equations of State‐Which? Industrial and Engineering
Chemistry Fundamentals 18, 81‐97.
Martín‐Martín, J.D., Gómez Rivas, E., Travé i Herrero, A., Salas, R., Vergés, J.,
2012. Dolomías controladas por fracturas en carbonatos aptienses de la zona
de Benicàssim (SE Cuenca del Maestrat): distribución y características
petrográficas. Geogaceta 51, 19‐22.
Martín‐Martín, J. D., Travé, A., Gomez‐Rivas, E., Salas, R., Sizun, J. P., Vergés, J.,
Alfonso, P., 2015. Fault‐controlled and stratabound dolostones in the Late
Aptian–earliest Albian Benassal Formation (Maestrat Basin, E Spain): Petrology
and geochemistry constrains. Marine and Petroleum Geology 65, 83‐102.
Martini, P.M., Barbano, M.S., Smedile, A., Gerardi, F., Pantosti, D., Del Carlo,
P., Pirrotta, C., 2010. A unique 4000 year long geological record of multiple
tsunami inundations in the Augusta Bay (eastern Sicily, Italy). Marine Geology
276, 42–57.
Matheron, G., 1971. La théorie des variables régionalisées et ses applications.
Fasc. 5, École National Supériure des Mines, Paris.
Matsumoto, D., Shimamoto, T., Hirose, T., Gunatilake, J., Wickramasooriya, A.,
DeLile, J., Murayama, M., 2010. Thickness and grain‐size distribution of the
2004 Indian Ocean tsunami deposits in Periya Kalapuwa Lagoon, eastern Sri
Lanka. Sedimentary Geology 230, 3, 95‐104.
192
Meléndez, M.N., Liesa, C.L., Soria, A.R., Meléndez, A., 2009. Lacustrine system
evolution during early rifting: El Castellar Formation (Galve sub‐basin, Central
Iberian Chain). Sedimentary Geology 222, 64‐77.
Meyer, R., Krause, F.F., 2006. Permeability Anisotropy and Heterogeneity of a
Sandstone Reservoir Analogue: An Estuarine to Shoreface Depositional System
in the Virgelle Member, Milk River Formation, Writing‐on‐Stone Provincial
Park, Southern Alberta. Bulletin of Canadian Petroleum Geology 54, 4, 301‐
318.
Miall, A.D., 1996. The Geology of Fluvial Deposits, Sedimentary Facies, Basin
Analysis, and Petroleum Geology. Springer‐Verlag, New York, 582 p.
Morton, R.A., Goff, J.R., Nichol, S.L., 2008. Hydrodynamic implications of
textural trends in sand deposits of the 2004 tsunami in Sri Lanka. Sedimentary
Geology 207, 1, 56‐64.
Nanayama, F., Shigeno, K., 2006. Inflow and outflow facies from the 1993
tsunami in southwest Hokkaido. Sedimentary Geology 187, 139–158.
Navarrete, R., 2015. Controles alocíclicos de la sedimentación barremiense en
la Subcuenca de Galve (Fm. Camarillas, margen occidental de la Cuenca del
Maestrazgo). PhD Thesis, Univ. de Zaragoza. 444 p.
Navarrete, R., Rodríguez‐López, J.P., Liesa, C.L., Soria, A.R., Veloso, F.M., 2013.
Changing physiography of rift basins as a control on the evolution ofmixed
siliciclastic–carbonate back‐barrier systems (Barremian Iberian Basin, Spain).
Sedimentary Geology 289, 40–61.
Navarrete, R., Liesa C.L., Castanera D., Soria A.R, Rodríguez‐López, J.P.,
Canudo , J.L., 2014. A thick Tethyan multi‐bed tsunami deposit preserving a
dinosaur megatracksite within a coastal lagoon (Barremian, eastern Spain).
Sedimentary Geology 313, 105–127.
Nishikawa, T., Ito, M., 2000. Late Pleistocene barrier‐island development
reconstructed from genetic classication and timing of erosional surfaces,
paleo‐Tokyo Bay, Japan. Sedimentary Geology 137, 25‐42.
Nomeli, M.A., Tilton, N., Riaz, A., 2014. A new model for the density of
saturated solutions of CO2–H2O–NaClin saline aquifers. International Journal
of Greenhouse Gas Control 31, 192–204.
193
Nordbotten, J.M., Celia, M.A., Bachu, S., 2005. Injection and storage of CO2 in
deep saline aquifers: Analytical solution for CO2 plume evolution during
injection. Transport in Porous Media 58, 339–360.
Norden, B., Frykman, P., 2013. Geological modelling of the Triassic Stuttgart
Formation at the Ketzin CO2storage site, Germany. International Journal of
Greenhouse Gas Control 19, 756–774.
Norden, B., Förster, A., Vu‐Hoang, D., Marcelis, F., Springer, N., Le Nir, I., 2010.
Lithological and Petrophysical Core‐Log Interpretation in CO. SPE Reservoir
Evaluation and Engineering 13, 2, 179‐192.
Oliver, M.A., Webster, R., 2014. A tutorial guide to geostatistics: Computing
and modelling variograms and kriging. Catena 113, 56–69.
Peropadre, C., 2012. El Aptiense del margen occidental de la cuenca del
Maestrazgo: controles tectónico, eustático y climático en la sedimentación.
PhD Thesis, Universidad Complutense de Madrid, 649 pp.
Peropadre, C., Meléndez, N., Liesa, C.L., 2008. Variaciones del nivel mar
registradas como valles incisos en la Formación Villarroya de los Pinares en la
subcuenca de Galve (Teruel, Cordillera Ibérica). Geo‐temas 10, 167–170.
Petrowiki, 2015. Petrowiki webpage published by Society of Petroleum
Engineers (SPE). http://petrowiki.org/PetroWiki.
Pettijohn, F.J., Potter, P.E., Siever, R., 1973. Sand and sandstone. Springer‐
Verlag, New York‐Heidelberg‐Berlin, 618 p.
Phantuwongraj, S., Choowong, M., 2012. Tsunamis versus storm deposits from
Thailand. Natural Hazards 63, 31–50.
Pini, R., Krevor, S.C.M., Benson, S.M., 2012. Capillary pressure and
heterogeneity for the CO2/water system in sandstone rocks at reservoir
conditions. Advances in Water Resources 38, 48–59.
Plug, W.J., Bruining, J., 2007. Capillary pressure for the sand–CO2–water
system under various pressure conditions. Application to CO2 sequestration.
Advances in Water Resources 30, 2339–2353.
194
Pochat, S., Van Den Driessche, J., 2007. Impact of synsedimentary metre‐scale
normal fault scarps on sediment gravity flow dynamics: An example from the
Grès d’Annot Formation, SE France. Sedimentary Geology 202, 796–820.
Prendergast, A.L., Cupper, M.L., Jankaew, K., Sawa, Y., 2012. Indian Ocean
tsunami recurrence from optical dating of tsunami sand sheets in Thailand.
Marine Geology 295‐298, 20–27.
Prost, P., 2008. Numerical Simulations of CO2 Injection into Saline aquifers:
estimation of storage capacity and srrival times using multiple realizations of
heterogeneous permeability fields. PhD Thesis, University of Stuttgart,
Germany, 69 p.
Pyrcz, M. J., Deutsch, C.V. 2014. Geostatistical reservoir modeling. Oxford
university press, 448 p.
Reid, R.C., Prausnitz, J.M., Polling, B.E., 1987. The Properties of Gases and
Liquids. New York, USA, McGraw‐Hill.
Robinson, J.W., McCabe, P.J., 1997. Sandstone‐body and shale‐body
dimensions in a braided fluvial system: salt wash sandstone member (Morrison
Formation), Garfield County, Utah. AAPG Bulletin 81, 1267–1291.
Rodríguez‐López, J.P., Meléndez‐Hevia, N., Soria, A.R., De‐Boer, P.L., 2009.
Reinterpretación estratigráfica y sedimentológica de las formaciones Escucha y
Utrillas de la Cordillera Ibérica. Revista de la Sociedad Geológica de España 22,
3‐4, 163‐219.
Salas, R., 1987. El Malm y el Cretaci inferior entre el Massis de Garraf y la Serra
d'Espadà. PhD Thesis, Universidad de Barcelona, 345 pp.
Salas, R., Casas, A., 1993. Mesozoic extensional tectonics, stratigraphy and
crustal evolution during the Alpine cycle of the eastern Iberian basin.
Tectonophysics 228, 1, 33‐55.
Salas, R., Guimerà, J., Más, R., Martín‐Closas, C., Meléndez, A., Alonso, A.,
2001. Evolution of the Mesozoic Central Iberian Rift System and its Cenozoic
inversion (Iberian Chain). In: Ziegler, P.A., Cavazza, W., Robertson, A.F.H.,
Crasquin‐Soleau, S. (eds), Peri‐Tethys Memoir 6: PeriTethyan Rift/Wrench
Basins and Passive Margins. Mémoires du Muséum national d'Histoire
naturelle 186, 145–185.
195
Saripalli, P., McGrail, P., 2002. Semi‐analytical approaches to modeling deep
well injection of CO2 for geological sequestration. Energy Conversion and
Management 43, 2, 185–198.
Schlumberger, 2013a. Eclipse Technical Descriptions, Version 2013.1, 58 p.
Schlumberger, 2013b. Petrel Help Center Manual.
SEE (Secretaria de Estado de Energia), 2014. La Energia en España 2013.
Ministerio de Industria, Energia y Turismo, 333 p.
Sifuentes, W.F., Giddins, M.A., Blunt, J.B., 2009. Modeling CO2 Storage in
Aquifers: Assessing the key contributors to uncertainty. SPE 123582.
http://dx.doi.org/10.2118/123582‐MS.
Simms, A.R., Anderson, J.B., Blum, M., 2006. Barrier‐island aggradation via
inlet migration: Mustang Island, Texas. Sedimentary Geology 187, 105–125.
Simón, J.L., 1982. Compresión y distensión alpinas en la Cadena Ibérica
oriental. PhD Thesis, Universidad de Zaragoza. Publicaciones del Instituto de
Estudios Turolenses, 269 p.
Simón, J.L., 1989. Late Cenozoic stress field and fracturing in the Iberian Chain
and Ebro Basin (Spain). Journal of Structural Geology 11 (3), 285‐294.
Soria, A.R., 1997. La sedimentación en las cuencas marginales del surco ibérico
durante el Cretácico inferior y su control estructural. PhD Thesis. Universidad
de Zaragoza. 363 p.
Soria, A.R., Meléndez, M.N., Meléndez, A., Liesa, C.L., Aurell, M., Goméz‐
Fernández, J.C., 2000. The Early Cretaceous of the Iberian Basin (Northeastern
Spain). In: Gierlowski‐ Kordesch, E.H., Kelts, K.R. (eds), Lake basins through
space and time. American Association of Petroleum Geologists, Studies in
Geology 46, 285–294.
Soria, A.R., Liesa, C.L., Meléndez, A., Meléndez, N., 2001. Sedimentación
sintectónica de la Formación El Castellar (Cretácico Inferior) en la subcuenca
de Galve (Cuenca Ibérica). Geotemas 3, 2, 257‐260.
Spiteri, E., Juanes, R., 2006. Impact of relative permeability hysteresis on the
numerical simulation of WAG injection. Journal of Petroleum Science and
Engineering 50, 115– 139.
196
Spycher, N., Pruess, K., 2005. CO2‐H2O mixtures in the geological
sequestration of CO2 II. Partitioning in chloride brines at 12‐100 C and up to
600 bar. Geochimica et Cosmochimica Acta 69, 13, 3309‐3320.
Spycher, N., Pruess, K., Ennis‐King, J., 2003. CO2‐H2O mixtures in the
geological sequestration of CO2. I. Assessment and calculation of mutual
solubilities from 12 to 100°C and up to 600 bar. Geochimica et Cosmochimica
Acta 67, 16, 3015–3031.
Sugawara, D., Minoura, K., Imamura, F., 2008. Tsunamis and Tsunami
sedimentology. In: Shiki, T., Tsuji, Y., Yamazaki, T., Minoura, K. (eds),
Tsunamiites — Features and Implications. Elsevier, Amsterdam, Oxford, 9‐50.
Sugawara, D., Goto, K., Jaffe, B.E., 2014. Numerical models of tsunami
sediment transport — Current understanding and future directions. Marine
Geology 352, 295–320.
Sun, S., Shu, L., Zeng, J.C, Feng, Z., 2007. Porosity–Permeability and textural
heterogeneity of reservoir sandstones from the Lower Cretaceous Putaohua
Member of Yaojia Formation, Weixing Oilfield, Songliao Basin, Northeast
China. Marine and Petroleum Geology 24, 109–127.
Tye, R.S., 2004. Geomorphology: an approach to determining subsurface
reservoir dimensions. AAPG Bulletin 88: 1123–1147.
U.S. GeoSupply, Inc., 2015. Wellsite Geological Solutions. Website:
www.usgeosupply.com
Veloso, F.M.L., Navarrete, R., Soria, A.R., Meléndez Hevia, M.N., 2013. Estudio
de afloramiento de escala intermedia como almacén geológico (Aliaga,
provincia de Teruel). In: Garcia‐Hidalgo, J.F., Gil, Barroso, F., Diaz de Neira, J.A.
(eds), V Congreso Cretácico de España, 136‐139.
Vennin, E., Aurell, M., 2001. Stratigraphie sequentielle de l'Aptien du sous‐
basin de Galvé (Province de Teruel, NE de l'Espagne). Bulletin de la Société
Géologique de France 172, 397–410.
Vesovic, V., Wakeham, W. A., Olchowy, G.A., Sengers, J.V., Watson, J.T.R.,
Millat, J., 1990. The Transport Properties of Carbon Dioxide. Journal of Physical
and Chemical Reference Data 19, 3, 763.
197
White, C.D., Willis, B.J., Dutton, S.P., Bhattacharya, J. P., Narayanan, K., 2004.
Sedimentology, statistics, and flow behavior for a tide‐influenced deltaic
sandstone, Frontier Formation, Wyoming, United States. In: Grammer, G.M.,
Harris, P.M., Eberli, G.P (eds), Integration of outcrop and modern analogs in
reservoir modeling. AAPG Memoir 80, 129‐152.
Wood, L.J., 2004. Predicting tidal sand reservoir architecture using data from
modern and ancient depositional systems . In: Gramer, M., Harris, P.M.,
Eberli, G.P. (eds), Integration of Outcrop and Modern Analogs in Reservoir
Modeling. American Association of Petroleum Geologists Memoir 80, 45–66.
Yoshida, S., Jackson, M. D., Johnson, H. D., Muggeridge, A. H., Martinius, A. W.,
2001. Outcrop studies of tidal sandstones for reservoir characterization (Lower
Cretaceous Vectis Formation, Isle of Wight, southern England). In: Martinsen,
O.J., Dreyer, T. (eds), Sedimentary Environments Offshore Norway ‐ Palaeozoic
to Recent. NPF Special Publication 10, 233‐257. Elsevier Science B.V.,
Amsterdam.
Zaytsev, I.D., Aseyev, G.G., 1993. Properties of Aqueous Solutions of
Electrolytes. Boca Raton, Florida, USA CRC Press.
Zheng, C.M., Gorelick, S.M., 2003. Analysis of solute transport in flow fields
influenced by preferential flowpaths at the decimeter scale. Ground Water 41,
2, 142–155.