Sudeep Kanungo, Gosia Skowron, Eiichi Setoyama and Sabita Silwal
EGI, University of Utah
Corresponding author: Sudeep Kanungo; firstname.lastname@example.org
The graphic correlation method using the composite standard database is a technologically advanced bio-chronostratigraphic technique that provides definitive, absolute-age-based correlations of rock units defined by lithologic, log, and/or seismic data. Starting with well-by-well chronostratigraphic interpretations, in conjunction with lithologic and seismic correlations, the technique produces high-resolution and synchronous correlations that are reproducible on a local or regional scale, thereby enabling an exploration team to reconstruct basin history in thin time-slice intervals and evaluate structural and stratigraphic prospects.
The foundation of this method is the former Amoco Composite Standard Database, which was granted to the University of Utah (EGI) in 1999. Since then, EGI has successfully built upon and enhanced the database through chronostratigraphic interpretations of new sections, latest time scale calibrations and expansion of taxonomy, while preserving the core methodology developed by the former Amoco Corporation geologists. Representing nearly five decades of paleontological research, this multi-million dollar legacy database provides global capabilities in chronostratigraphy and paleoenvironmental applications unparalleled in either industry or academia.
2. The need for high-resolution stratigraphic correlation
Exploration programs and methods in the industry lean heavily towards finding and drilling seismic structures. Seismic and geologic structure maps based on correlation of lithologic units therefore represent the main tools for exploring hydrocarbons. However, the lack of integrated and multidisciplinary paleontological dating methods to supplement lithologic correlations, even where rock units are regionally continuous, has led to significant miscorrelation of formational units and incorrect seismic interpretations. This includes the commonly misplaced theory that formation boundaries are synchronous, which they are not, unless deposited along strike. In all other directions, the formation boundary is time-transgressive. Accurate time control is therefore necessary to reconstruct the correct stratigraphic framework of facies and paleogeography, in order to fully understand the geological history of any area. This cannot be learned from seismic data or from the correlation of lithologic units alone.
Traditional methods of paleontological correlation are based on mono-disciplinary zonation schemes and biomarkers that often conflict with each other, even within the same facies. They produce thick biostratigraphic zones, especially in older Mesozoic sediments, which frequently parallel rock units and are not always synchronous from place to place. The graphic correlation technique uses the entire suite of paleontological data, irrespective of biostratigraphic zones, and is a quantitative stratigraphic method that provides the required level of stratigraphic refinement on a basin to prospect scale.
3. Concepts and technique
a. Composite Standard Database:
A composite standard database, or simply a composite standard, is a paleontological database that contains stratigraphic ranges of multidisciplinary fossil species (taxa) calibrated to absolute age in mega-annum (Ma). The composite standard, as the name implies, is made up of composited stratigraphic ranges of species defined from several localities (wells and outcrops) within a geographic area to synthesize the complete stratigraphic range of those species (Carney and Pierce, 1995). Any one section is unlikely to record the complete stratigraphic range of a species; so, by compositing the lower and upper end of the species range from multiple localities, eventually the complete stratigraphic range from the first to last appearance can be constructed. It is important to note that a mere listing of species from a geographic area does not make a composite standard; it is only when those taxa are calibrated to an absolute time scale (e.g., GTS 2012), that a composite standard is established with calibrated datums (tops and bases) in it. A composite standard progresses towards maturity as more and more sections are interpreted and composited into it, so that species ranges are at, or close to, their true, maximum extent. The database uses two kinds of datums for age interpretations: tops (last stratigraphic occurrence/last appearance datum/first downhole occurrence) and bases (first stratigraphic occurrence/first appearance datum/last downhole occurrence) of a species. The distinct advantage is that datums from a number of different fossil groups can be combined with magnetic and isotopic events into an integrated scheme that can be tied to an absolute time scale. An example of a Cenozoic composite standard from the Nile Delta is illustrated in Figure 1.
There are at least twenty fossil groups represented in the EGI composite standard database, including micro- and macrofossils. The microfossil groups that have the greatest application in age and paleoenvironmental interpretations encompass calcareous microfossils (planktonic/ benthic foraminifera, nannofossils, and ostracodes) and the organic-walled/palynology group (dinoflagellate cysts and pollen & spore). Siliceous microfossils such as radiolaria, diatoms, and silicoflagellates are also available. Paleozoic restricted palynological groups such as acritarchs and chitinozoans, and conodonts are represented in some of the Paleozoic composite standards. Rare microfossil groups include tintinnids, ebridians, and calcispheres. Macrofossils exist in the database, even though limited, and include a variety of disciplines such as molluscs (bivalves, gastropods, cephalopods), brachiopods, vertebrates, arthropods, worms, etc., and some very rare Paleozoic groups such as conulariids and hyolithids.
b. Geological time scale:
EGI’s composite standards span across the entire Phanerozoic time interval, which includes the Paleozoic, Mesozoic, and Cenozoic eras. The Mesozoic and Cenozoic composite standards bear stronger calibration of paleontological datums compared to the Paleozoic standards, as more projects have been commissioned on the Mesozoic–Cenozoic interval at EGI in the last decade.
The Atlantic Composite Standard, one of our most comprehensive and mature standards, is the first standard to be upgraded to the GTS 2012 time scale. Other standards can similarly be migrated to GTS 2012, depending on project needs. Previous time scales reflected in many of the composite standards include Berggren et al., (1995) for the Cenozoic interval, and de Graciansky et al., (1999) for the Mesozoic. The GTS 2004 time scale has been applied to some North African standards, e.g., the Libya Standard. However, the Paleozoic time scale developed by Amoco was an internal product and has not been upgraded at EGI.
c. Graphic correlation:
The methodology of graphic correlation (Figure 2a), originally developed by Shaw (1964), provides the means for compositing stratigraphic ranges of species from multiple localities. The Y-axis represents a well or outcrop (in meters or feet) that is being interpreted in relation to a relevant composite standard (e.g., Atlantic CS) shown on the X-axis (scale = absolute age in mega annum, Ma), with time getting younger on the right side of the X-axis. The datums that plot are those in common between the particular locality on the Y-axis and the composite standard on the X-axis. The biostratigrapher constructs a line of correlation (LOC), which in effect, is a time-depth curve that characterizes the age and depositional history of the section on the Y-axis. The LOC has only two components: off-vertical segments (slopes) that represent periods of rock accumulation; and horizontal offsets (terraces) that denote temporal hiatuses, a period of time for which there is no surviving section in the locality, and/or condensed sections (Figure 2a). As more localities are interpreted via graphic correlation, the composite standard improves and evolves into a database with complete stratigraphic ranges of taxa in it. In summary, graphic correlation is a process; the composite standard is the database that the geologist creates using that process.
The graphic correlation technique can provide higher stratigraphic resolution because it integrates data from multiple fossil disciplines, e.g., foraminifera, nannofossils, palynomorphs, siliceous microfossils, macrofossils, etc., rather than isolating disciplines with the traditional biozonation approach. Graphic correlation, therefore, allows for integrated or unified age determination for correlation purposes, as opposed to an independent age from a particular paleontologic discipline. This is schematically illustrated in Figure 2b.
4. Discrete outputs from graphic correlation
a. Line of Correlation (LOC):
The standard format of a graphic correlation chart (GCC) used by EGI is shown in Figure 3 with all the outputs from an interpreted well. The LOC on a graphic correlation plot is drawn based upon the graphical representation of datums in common between the section to be analyzed and the composite standard. All datums, however, are not of equal value. Certain microfossil taxa have more intrinsic chronostratigraphic value (e.g., zonal markers) than others that are long-ranging and facies-dependent. A datum may represent reworking, or in ditch cuttings samples, downhole contamination. Stratigraphic ranges may be truncated, not because of a missing section, but due to paleoenvironmental change. For these reasons, the LOC cannot be generated by mathematical regression (based upon equal consideration of all plotted datums), but only through the interpretation by an experienced paleontologist.
The LOC describes the age of the interpreted section in absolute time (Ma), derived from the relevant composite standard time scale. It provides one-to-one correlation between any depth in the well and absolute age. The age of the section is interpreted via discrete datums, rather than biostratigraphic zones, which are evaluated only through quantitative or semi-quantitative biostratigraphic distribution data (as opposed to a ‘tops’ listing). Because the datums that define the LOC are multidisciplinary in nature, they are annotated by different colors representing the respective fossil groups in EGI GCCs.
As the composite standard technique scales sections in absolute age or time, correlation of sections in multiple wells is also achieved via absolute time. This gives independence from the traditional biostratigraphic zonal schemes in which successful correlation is dependent, to a large degree, upon the number of zonal markers in common between the sections analyzed. It also gives some independence from facies, allowing correlation of sections composed of disparate facies, e.g., carbonate platform to deep-water basins, or continental to marine sections, in which few, if any zonal markers are in common.
b. Rate of rock accumulation:
The angle of the sloping parts of the LOC is directly correlated with the rate of rock accumulation; steeper slopes represent more rapid accumulation. Slope determination therefore permits calculation of rock accumulation rates, assuming that the measured intervals represent true stratigraphic thickness. Rock accumulation rates can be converted to sedimentation rates when compaction is taken into account.
c. Hiatus and hiatal intervals:
Lines of correlation of thick rock sections usually exhibit a stair-step pattern. The horizontal offset (step or terrace) represents a period of geologic time not represented by any detectable rock in a stratigraphic section. The length of the terrace is equal to the duration of the hiatus in absolute time that can be read from the time scale on the X-axis. The missing rock (or sediment) may have been removed by erosion or never deposited (sediment bypass). Alternatively, it may have been removed by normal faulting. The distinction between erosion and non-deposition cannot be made by graphic correlation alone, as additional data pertaining to basin evolution is needed to confirm the nature of unconformity.
In some cases, a series of stacked, high-order hiatuses (e.g., ravinements, condensed intervals) of short duration occur within a narrow stratigraphic interval. Each event is beyond the resolution of graphic correlation, but the sum of these events can be captured as a single terrace. The rock interval that contains these stacked terraces is termed a hiatal interval. Hiatal intervals are rarely recognized in subsurface stratigraphic analyses.
d. Chronostratigraphic summary:
The duration of stratigraphic events (deposition vs. hiatus) can be read from the time scale at the bottom of the graph via the chronostratigraphic summary bar. This gives a visual estimate of how much section is actually preserved vs. periods of time for which there is no surviving section in a locality.
5. Paleoenvironmental analysis
a. Interpretation using benthic microfossils
Using the same paleontological data employed in the chronostratigraphic analysis, changes in biofacies, and hence paleoenvironment, can be recognized through time for an interpreted well (Figure 3). Biofacies are defined based on paleontological parameters, such as species composition, diversity, equitability, and abundance, and provide information about depositional environment, including paleobathymetry, productivity, oxygenation, and displaced sediments. Biofacies often reflect paleoenvironmental changes that may not be obvious in lithology, e.g., a change in bottom-water oxygenation from anoxic to dysoxic, and can supplement lithofacies interpretation. Paleoenvironmental interpretation is mainly derived from benthic microfossils, especially calcareous benthic and agglutinated foraminifera, because of their good recovery potential from ditch cuttings and cores, and their ubiquity and bottom dwelling habitat that reflects bottom water environmental changes from the littoral zone to the deepest part of the oceans.
Paleobathymetry: Benthic foraminiferal assemblage compositions are regarded as a good indicator of the depth of deposition and commonly used to estimate paleobathymetry, i.e. paleo-water depth. Studies on modern benthic foraminifera revealed that while their bathymetric distribution is not governed directly by water depth via hydraulic pressure, it is still controlled by multiple depth-related physicochemical parameters, such as light intensity, salinity, temperature, nutrient availability, and oxygen concentration. Such studies also show that morphology of foraminiferal tests, or shells, can be related to their ecological preferences, and thus the shell morphology could be used to infer paleoenvironment and paleobathymetric preferences of extinct taxa. Moreover, even for many of those benthic foraminiferal taxa which are extinct and do not have living relatives, paleobathymetric limits have been approximated by a series of studies on their paleo-depth ranges, notably in the Atlantic passive margins, which enables determination of relative changes in paleobathymetry for older sediment (inner shelf, upper slope, etc.).
Recognition of displaced sediments: Paleobathymetric interpretation based on paleontological data can also be applied for detection of displaced (penecontemporaneously reworked) sediment facies. Foraminifera are biogenous sediment particles, and a sediment flow picks up and transports foraminifera as well as other fossil debris from different paleoenvironmental settings as it moves from the shelf into the basin. Therefore, allochthonous facies can contain a mixture of fossils from two or more paleoenvironmental settings (e.g., mixed shelf and lower bathyal fauna), and such displaced sediments can be distinguished from autochthonous facies, which contain only in situ fossils of a single paleoenvironment. Recognition of different paleoecological assemblages in allochthonous sediments may, additionally, shed light on the flow pathway. An assemblage containing benthic foraminifera from the inner shelf, middle shelf, and lower slope, for example, indicates mixing of sediments from those settings, and in turn, suggests that a sediment flow started from the inner shelf (provenance), bypassed the outer shelf and upper slope, and finally the sediments were deposited on the lower slope.
Productivity and oxygenation: The flux of organic matter to the sea floor, especially in deep-sea conditions, is considered a main environmental parameter affecting the benthic foraminiferal assemblage composition, along with bottom-water oxygen concentration, which is a limiting factor when it is low. The dynamic correspondence between organic matter flux (productivity) and oxygen level can be interesting. For example, dysoxic biofacies, in general, show low species richness and equitability, dominated by a few low-oxygen tolerant taxa. If the low oxygen level is related to high productivity, and it is moderately dysoxic, benthic foraminifera can be very abundant. On the other hand, if it is anoxic, then benthic foraminifera are usually absent. In oligotrophic (low productivity), oxic environments, benthic foraminifera are usually diversified and small-sized, but rare, except for an area with seasonal phytoplankton bloom where opportunistic taxa occur in high abundance.
Low oxygen conditions are the result of an imbalance between microbial oxygen consumption and oxygen renewal in an area. Thus, while suboxic conditions are often related to a high organic matter flux to the sea floor from productive surface waters, leading to consumption of oxygen by degradation of organic matter, oxygen depletion can be achieved even under low productivity surface waters through poor bottom-water circulation in restricted basins. Antithetically, oxygenated bottom water currents can maintain oxic conditions under highly productive surface waters. The understanding of such a complex dynamic directly impacts source rock intervals in hydrocarbon exploration. Quantitative analysis of benthic foraminiferal assemblages enables the discrimination between such subtly different oxygenation regimes through varying paleontological parameters, such as abundance, diversity, equitability, and species composition. Figure 3 illustrates the full spectrum of paleoenvironmental interpretations that can be made during the interpretation of a well, which is calibrated to the LOC, giving precise time control for each of the paleoenvironmental events.
b. Wheeler diagrams
Paleoenvironmental interpretation scaled in absolute time permits the construction of a “Wheeler” diagram. A Wheeler diagram (or Wheeler) aids the detection of regional unconformities and/or condensed sections, and visually summarizes vertical changes in depositional environments; it finds applications in facies integration on a regional scale. In the example shown in Figure 4, the paleoenvironmental interpretation includes lithofacies, paleobathymetry, organic productivity, and intervals of penecontemporaneous (displaced) sediment movement. Any or all of these components have been calibrated to absolute time via the LOC, giving the duration of the paleoenvironment in question. As this is done for multiple wells, the Wheeler diagram can be overlain with all kinds of facies and sedimentological information calibrated from the individual wells. Any horizontal line on a Wheeler diagram of this type is by definition a time line that can permit a time-slice paleoenvironmental reconstruction, e.g., Eocene and Maastrichtian time lines illustrated in Figure 4.
6. Applications in hydrocarbon exploration:
a. Basin modeling
Absolute age constraints on depositional events and hiatuses are critically important in basin modeling. Graphic correlation is the most reliable dating technique for resolving the uncertainty around absolute ages. Numerical data on rock accumulation (or sedimentation) rates and absolute age of hiatuses are therefore two key inputs required to develop a good basin model. Confidence on the position of slopes and terraces in the chronostratigraphic interpretation increases if they are tied to specific datums, which provides higher confidence to the basin model. The position of the end points of the terrace or slope can sometimes be uncertain, creating questionable slopes and terraces, especially when biostratigraphic data is sparse or of poor quality.
b. Paleoenvironmental reconstructions/maps
Biofacies identified for multiple wells can be scaled in absolute time through the Wheeler diagram, and, given enough geographic coverage of data, a regional paleoenvironmental/facies map can be generated to visualize the lateral distribution of facies at any given point in time. Such maps assist the prediction of sediment transport direction and geographic distribution of source and reservoir rocks for selected time slices (e.g., 70 Ma, Early Maastrichtian), by showing the provenance of sediment flows and the geographic extent of different depositional environments, which are important components of play fairway analysis (Figure 5). High temporal resolution on such paleoenvironmental reconstructions is a distinct advantage, with attainable durations often in the order of 0.5–2 Ma, depending on number of well points and the quality of data. Such reconstructions have a high degree of reliability that truly coeval facies are being integrated.
c. Log calibration
Wireline logs (e.g., Gamma Ray, Neutron, SP, Resistivity, Sonic, etc.) can be directly calibrated to absolute age via the LOC, in as much as they are scaled properly to the Y-axis. When lithofacies discontinuities are recognized in the log signature they can be compared to the chronostratigraphy to determine whether or not they correlate to a temporal discontinuity (i.e., hiatus), or if they occur within a continuously deposited interval. Hiatuses which are not flagged by lithofacies discontinuities can be noted as shown in the example (Figure 6).
c. Seismic calibration and sequence stratigraphy
The application of the composite standard technique is not limited to logs, but can be extended to seismic data to strengthen sequence stratigraphic interpretations. As a time-depth curve, the LOC can be exported as a pseudo-log (X-Y coordinates of time and depth) and imported into integrative software programs such as Landmark(R) (Figure 7). In this way, depositional packages can be directly integrated with the overall stratigraphic architecture, and hiatuses can be correlated to major seismic reflectors. When this is done for several wells, regional sequence boundaries can be identified and the median ages for these surfaces determined.
As shown in Figure 8, an up-dip and down-dip well on the continental margin of Brazil were graphically interpreted. Inflection points of the LOC for each well were imported as pseudo-log tracks and overlain on the seismic section. Although a sequence stratigraphic model had already been constructed for this margin, correlating between the up-dip and down-dip sections on this structurally complex margin based upon seismic alone proved to be difficult without the chronostratigraphic interpretation provided by graphic correlation for the two wells. With this temporal data, successful correlation became easier to achieve.
It can be concluded that the composite standard methodology provides several competitive advantages for high-resolution stratigraphic correlation in an exploration or production setting, related to conventional or low-permeability reservoirs, as described below:
- Chronostratigraphic interpretation based upon the entire suite of paleontologic data set and assemblage (not just zonal markers or ‘tops’) resulting in a single, integrated and multidisciplinary age interpretation. However, more emphasis is placed on planktonic groups (nannofossils, planktonic foraminifera, dinoflagellate cysts and diatoms, etc.) for age determination, whereas benthic foraminifera, supported by planktonic groups provide for paleoenvironmental delineation.
- Ability to integrate pre-existing data from multiple biostratigraphic reports (from multiple vendors/vintage) with disparate age conclusions into a consistent graphic age interpretation.
- Correlations of multiple well sections are done via absolute age and are therefore independent of zonal markers or facies.
- Important numerical data for basin modeling, such as rock accumulation rate and hiatus duration, can be easily measured from the graphic interpretation of any well.
- Paleoenvironmental and facies interpretations can be directly calibrated to absolute time via the LOC for multiple wells in a region. Once accomplished, high-resolution maps of the order of 0.5–2 Ma can be constructed.
- Any geological data that is overlain for the well on the Y-axis (sedimentology, log, or seismic), can be directly calibrated to absolute time via the LOC. Pseudo-log tracks (points of inflection on the LOC) can be overlain on seismic sections for direct integration of chronostratigraphic interpretation with the overall stratigraphic architecture. The ease of geologic data integration is perhaps the greatest advantage provided by this technique, and could be of immense value to an exploration team. Sequence stratigraphy interpretations, such as identification of maximum flooding surfaces/condensed sections, transgressive/regressive packages, etc., can be readily accomplished.
- Since the acquisition of the composite standard database over a decade ago, research on at least 88 chronostratigraphy-based projects has been conducted at EGI for CA member companies. These projects, among other things, have contributed to the tremendous growth of the database via development of new composite standards, localities, and datums. The table below provides a synopsis of the database, as it stands today.
|COMPOSITE STANDARD DATABASE SPECIFICS AT EGI|
|# of Composite Standards||141|
|# of Localities (wells and outcrops)||4031|
|# of Paleontologic Datums (tops and bases)||52,530|
|# of Fossil Groups||20|
The author thanks Dr. Paul Sikora, former EGI Head of Chronostratigraphy, for his valuable inputs on the article.
Berggren, W.A., Kent, D.V., Swisher III, C.C., Aubry, M.-P., 1995. A revised Cenozoic geochronology and chronostratigraphy. In: Berggren, W.A., Kent, D.V., Hardenbol, J. (Eds.), Geochronology, Time Scales and Global Stratigraphic Correlations, SEPM Special Publication No. 54, pp. 129–212.
Carney, J.L., and Pierce, R.W., 1995. Graphic correlation and composite standard databases as tools for the exploration biostratigrapher. In: Mann, K.O., and Lane, H.R. (Eds.), Graphic Correlation, SEPM Special Publication No. 53, pp. 23–45.
de Graciansky, P.-C., Hardenbol, J., Jacquin, T., and Vail, P.R., (Eds.), 1998. Mesozoic-Cenozoic Sequence Stratigraphy of European Basins, SEPM Special Publication No. 60, 786 pp.
Gradstein, F.M., Ogg, J.G., and Smith, A.G., (Eds.), 2004. A Geologic Time Scale 2004, Cambridge University Press, 589 pp.
Gradstein, F.M., Ogg, J.G., Schmitz, M.D., and Ogg, G.M., (Eds.), 2012. The Geologic Time Scale 2012, Elsevier, vol. 2, 1144 pp.
Shaw, A.B., 1964. Time in Stratigraphy: New York, McGraw-Hill, 365pp.
List of figures:
Figure 1: Example of a Cenozoic composite standard from the Nile Delta; composited stratigraphic ranges of species calibrated against the composite standard time scale and arranged by last stratigraphic occurrences (‘tops’); different colors represent different microfossil groups, e.g., blue = benthic foraminifera, red = planktonic foraminifera, green = nannofossils, etc.
Figure 2a: Methodology of graphic correlation that provides a means for compositing stratigraphic ranges (from first occurrence/base to last occurrence/top) of species from multiple localities (wells/outcrops) within a region. The process also allows for the calculation of rock accumulation rate and hiatus delineation using the absolute-age composite standard time scale.
Figure 2b: Schematic illustration of multidisciplinary fossil integration via graphic correlation to give a unified, absolute-age interpretation of a fossiliferous rock section.
Figure 3: Standard format of a graphic correlation chart (GCC) showing discrete outputs generated from an interpretation, such as; line of correlation (LOC) representing the rock accumulation rate, absolute age of hiatus in mega-annum and the chronostratigraphic summary. In the chronostratigraphic summary at the bottom, colored blocks indicate deposition, blank blocks are hiatuses placed at a given depth. The paleoenvironmental interpretation on the right hand side demonstrates the ease of calibrating paleoenvironments to absolute time via LOC for an interpreted well.
Figure 4: Example of a multi-well Wheeler diagram constructed via graphic correlation, with biofacies overlain on chronostratigraphic interpretations for five well sections. Two time lines (Eocene and Maastrichtian) are illustrated on the Wheeler diagram. Biofacies are represented by colored blocks. Different colors represent different paleoenvironments (legend on the top right). Intervening segments between biofacies denote hiatuses, the number refers to the depth at which the individual hiatus was placed. The time scale is on the vertical axis, so any horizontal line becomes a time line.
Figure 5: High-resolution paleoenvironmental reconstruction for the Maastrichtian (~70 Ma) time line in the Equatorial Atlantic region, based upon absolute age calibration of facies from graphic correlations of several wells.
Figure 6: A line of correlation tied to a gamma ray (GR) log suite, illustrating how unconformities can be tied to major lithofacies discontinuities (or lack thereof). Hiatuses/maximum flooding surfaces can be more reliably interpreted by integrating log data with paleontological data.
Figure 7: Example of integrating seismic line with chronostratigraphic data for a continental margin well. The LOC (red) for a well (green) can be overlain on seismic data, with correlation of temporal hiatuses and regional seismic reflectors.
Figure 8: Correlation of a composite seismic section is made easier with chronostratigraphic overlays from an up-dip and a down-dip well along the continental margin of Brazil. The two LOC’s are in red; the wells are in green. The LOC’s assign an absolute age (Middle Campanian, 78 Ma) to the purple seismic reflector, facilitating regional seismic and sequence stratigraphic interpretation.