Use of a relational database for the classification of fluvial sedimentary systems and the interpretation and prediction of fluvial architecture

1
Use of a relational database for the
classification of fluvial sedimentary systems and
the interpretation and prediction of fluvial
architecture

• Luca Colombera, Nigel P. Mountney, William D.
  McCaffrey

Fluvial Eolian Research Group University of
Leeds
2
Fluvial architecture
Interpretations and subsurface predictions of
fluvial architecture rely on classification
schemes, facies models and depositional
models qualitative approaches based on limited
number of examples
Orton Reading (1993)
Shanley McCabe (1994)
3
Overview
Creation of a relational database for the
digitization of fluvial sedimentary architecture
the Fluvial Architecture Knowledge Transfer
System (FAKTS)

• Quantitative characterization of fluvial
  architecture applicable to
• determination of importance of controlling
  factors
• develop quantitative synthetic depositional
  models
• derive constraints on subsurface predictions
• identify modern and ancient reservoir analogues

4
Approach to DB design
The sedimentary and geomorphic architecture of
preserved ancient successions and modern rivers
is translated into the database schema by
subdividing it into geological objects common
to the stratigraphic and geomorphic realms
which belong to different scales of observation
nested in a hierarchical fashion.
FAKTS conceptual and logical schemes
5
Implementation
Each object type is assigned to a table and each
individual object is given a unique identifier to
implement the nested containment relationships.
The same numerical indices are also used for
re-creating neighbouring relationships between
objects belonging to the same scale.
6
Genetic units classifications
Implementation
DEPOSITIONAL ELEMENTS

ARCHITECTURAL ELEMENTS

FACIES UNITS

2 classes Channel-complex Floodplain
14 classes partly based on Mialls (1996)
scheme enhanced geomorphic expression
24 textural structural classes partly based on
Mialls (1996) scheme
Dataset/subset classifications
METADATA

INTERNAL PARAMETERS

EXTERNAL CONTROLS

• Authors/reference
• Basin
• Lithostratigraphic unit
• River
• Age
• Methods/data type
• Data Quality Index
• etc

• Basin gradient
• Discharge regime
• River pattern
• Drainage pattern
• Aggradation rates
• Load-type dominance
• Relative distality
• etc

• Subsidence rates/types
• Basin/catchment climate
• Basin/catchment vegetation
• Relative eustatic change
• Catchment lithologies
• Catchment uplift rates
• Catchment geomorphic processes
• etc

7
Data Entry
Cain (2009)
Amorosi et al. (2008)
Cain (2009)
Robinson McCabe(1997)
North (1996) at present, much is being
published in the format of multiple vertical
profiles, photomontages and line drawings
because we still do not really know how to handle
all the available facts.
8
Database Output unit proportions
North (1996) at present, much is being
published in the format of multiple vertical
profiles, photomontages and line drawings
because we still do not really know how to handle
all the available facts.
9
Database Output unit DIMENSIONS
Miall Jones (2003) the database on
large-scale fluvial architecture, especially
sandbody width and length, remains extremely
small
10
Database Output unit Transitions
Transition count matrices
COUNT (Z) Sh Sl Sm Sp Sr Ss St
Sh 555 116 218 145 211 59 169
Sl 122 283 151 89 25 33 121
Sm 215 142 561 119 51 25 103
Sp 143 87 106 350 56 22 155
Sr 152 19 50 37 121 4 76
Ss 68 55 16 20 7 58 57
St 208 145 124 137 103 42 698

Facies transition within 4th order channel-fills
N 1024
11
Database Output FILTERING ON ARCHITECTURAL
PROPERTIES
Facies overlying 4th order BS
Facies overlying 5th order BS
Possibility to filter on linked architectural
properties dimensions, type of genetic units,
bounding surfaces, etc.
N 432
N 260
Right-hand strike lateral transitions from AEs
left-hand neighbouring CH elements
N 515
Right lateral AE
Left lateral AE
12
Spatial and temporal evolution
KAYENTA FM. Jurassic SE Utah
Quantitative investigation of spatial and
temporal sedimentary trends
ORGAN ROCK FM. Permian SE Utah (data from Cain
2009)
13
Synthetic depositional models
FILTERS
MODEL
Facies proportions
CH channel-fill characterization
All systems
NO FILTERS
Sandy deposits
41 case studies 28 basins 19 Formations 11
rivers 1,408 Depositional El.s (1,192
classified ) 1,344 DE transitions 2,591
Architectural El.s (2,274 classified) 4,885
AE transitions 11,908 Facies units (11,100
classified) 13,581 FU transitions
Architectural element proportions
N 2274
Brierley (1996) By definition, individual
models must synthesize information from a range
of examples otherwise, each case
study could be considered a model itself.
14
Synthetic depositional models
FILTERS
MODEL
CH channel-fill characterization
Facies proportions
All systems
NO FILTERS
Sandy deposits
Braided systems
River pattern BRAIDED
Architectural element proportions
23 case studies 11 Basins 8 Formations 6 River
s 396 Depositional El.s 1163 Architectural
El.s 4,948 Facies units
N 964
Brierley (1996) By definition, individual
models must synthesize information from a range
of examples otherwise, each case
study could be considered a model itself.
15
Synthetic depositional models
FILTERS
MODEL
CH channel-fill characterization
Facies proportions
All systems
NO FILTERS
Sandy deposits
Braided systems
River pattern BRAIDED
Architectural element proportions
Basin climate SEMIARID
Braided semiarid systems
8 case studies 2,704 genetic units
N 438
Brierley (1996) By definition, individual
models must synthesize information from a range
of examples otherwise, each case
study could be considered a model itself.
16
Synthetic depositional models
FILTERS
MODEL
Facies proportions
CH channel-fill characterization
All systems
NO FILTERS
Sandy deposits
Braided systems
River pattern BRAIDED
Architectural element proportions
Basin climate SEMIARID
Braided semiarid systems
Braided semiarid ephemeral systems
Discharge regime EPHEMERAL
N 86
Brierley (1996) By definition, individual
models must synthesize information from a range
of examples otherwise, each case
study could be considered a model itself.
17
Subsurface applications
North Prosser (1993) Are the results from
outcrop and modern environment studies being
translated into predictive tools
suitable for modelling subsurface geology?
de Marsily et al. (2005) future work should be
focused on improving the facies models A
world-wide catalog of facies geometry and
properties, which could combine site genesis and
description with methods used to assess the
system, would be of great value for practical
applications.

• QUANTITATIVE INFORMATION FROM
• identified modern and ancient reservoir
  analogues
• synthetic depositional models used as synthetic
  analogues
• TO BE USED FOR
• guiding subsurface correlations
• deriving constraints for stochastic reservoir
  modelling
• genetic/material unit proportions, absolute and
  relative dimensional parameters,
• Indicator auto- and cross-variograms, transition
  probabilities/rates

18
Facies modelling applications
RELATIVE DIMENSIONAL PARAMETERS COMPUTATION
INDICATOR VARIOGRAM COMPUTATION
19
Facies modelling applications
Possibility to tailor the models filtering on
genetic units
FLUVSIM (Deutsch Tran 2002) simulations
and on boundary conditions.
SISIM (Deutsch Journel 1998) simulations
20
Conclusions

• FAKTS database
• Quantitative characterization of fluvial
  architecture applicable to
• determine the importance of controlling factors
• develop quantitative depositional models
• derive constraints on borehole correlations
• derive constraints on stochastic simulations of
  fluvial architecture
• identify modern and ancient reservoir analogues
• compare the geomorphic organization of modern
  rivers with preserved stratigraphic architecture
• assess the influence of 1D data sampling density
  on observations and interpretations