Central Caspian Basins Vitor Abreu ExxonMobil Exploration, Houston, Texas, U.S.A. Dag Nummedal Colorado Energy Research Institute, Colorado School of Mines, Golden, Colorado, U.S.A. ABSTRACT T he oil industry has been active in Azerbaijan for centuries, and the Ap- sheron Peninsula, Apsheron sill, onshore, and the shelf margin of Azer- baijan are considered mature areas for exploration. However, large areas of the offshore Caspian, including the deep-water South Caspian, Turkmenistan shelf, and Central Caspian are still exploration frontiers. An understanding of the stratigraphy of reservoir rocks and seals in these areas could significantly reduce exploration risk. The interplay of the paleo-Volga, paleo-Amu Darya, and paleo- Kura deltas, since the late Miocene, provides the first-order controls on prospect distribution. A continuous trend of coastal onlap on the western margin exists for the South Caspian and in the Central Caspian basins from the upper Miocene to lower Pliocene, with onlap of these units over Miocene and Cretaceous rocks. These coastal onlap trends are associated with an overall rise in lake level from the lower to the upper productive series. The three delta systems exhibit sig- nificant differences in depositional style and timing, reacting in different ways to the rising lake level. Strong progradation of the paleo-Amu Darya delta occurred on the Turkmenistan shelf, on the eastern margin of the South Caspian Basin. This progradation is related to the Pliocene deposition of the Red series in Turkmenistan (equivalent to the Pereryva to Surakhany suites in Azerbaijan). Thus, the paleo-Amu Darya delta prograded during rising lake level, controlled primarily by sediment supply. During the deposition of the Pereryva suite, the paleo-Volga delta aggraded in the Apsheron region (northern margin of the South Caspian). A transgressive Chapter 11 Abreu, V., and D. Nummedal, 2007, Miocene to Quaternary sequence stratigraphy of the South and Central Caspian basins, in P. O. Yilmaz and G. H. Isaksen, editors, Oil and gas of the Greater Caspian area: AAPG Studies in Geology 55, p. 6586. 65 Copyright n2007 by The American Association of Petroleum Geologists. DOI:10.1306/1205845St553000 trend marks the Central Caspian Basin from the upper Balakhany to the Su- rakhany suites and may indicate backstepping of the paleo-Volga delta at that time. In the paleo-Kura system, on the southwest margin of the South Caspian Basin, a backstepping trend occurred during the deposition of the upper Ba- lakhany and Sabunchi suites (lower Pliocene). A downlap surface developed at the base of the paleo-Kura delta in the middle Surakhany suite. This downlap surface along the western Caspian margin correlates to the upper part of the progradational phase of the paleo-Amu Darya delta on the eastern margin of the basin. A paleo-Kura prograding wedge developed concurrent with the deposi- tion of the upper Surakhany and Akchagylian (upper Pliocene). The impact of sediment supply from the Alborz Mountains in Iran could not be evaluated be- cause of lack of data. Climatic fluctuations did exert a dominant control on the style of sedi- mentation in the South Caspian Basin through their direct impact both on lake levels and on sediment supply. The entire Productive Series reflects the Pliocene golden climate, when the Earth, overall, was much warmer than today. In addi- tion, on shorter time scales, the stratal pattern is controlled by high-frequency climatic cycles. Late lowstand deposits are dominated by aggradational braided streams and braid deltas. Transgressive and highstand deposits consist of ex- tensive lake shales interbedded with silts and sands. The transgressive shales can act as pervasive seals and permeability barriers and baffles in the reservoirs. Very little sand appears to have entered the lake during periods of falling lake level. INTRODUCTION Eight megasequences are recognized in the South Caspian Basin based on stratigraphic characteristics and relationship to periods of deformation: Triassic or Paleozoic, CretaceousJurassic, pre-Miocene Ter- tiary, PontianMessinian, early and middle Pliocene (Productive Series), late Pliocene (Akchagylian), early Pleistocene (Apsheronian), and late Pleistocene Holocene (Hickmanet al., 1999; Hickmanand Stuart, 2001). A significant increase in subsidence and sedi- mentationrates at the beginning of deposition of the Productive Series forms a natural division between the latest Cenozoic and all older strata (Hickman et al., 1999; Hickmanand Stuart, 2001). This increase in subsidence and sedimentation rate also coincides with the isolation of the Caspian from the Black Sea. The compressive deformation that started in the late Akchagylian or earliest Apsheronian further sub- divides the young succession into four composite se- quences (Hickman et al., 1999; Hickman and Stuart, 2001). This paper focuses on the stratigraphic evolution of the South and Central Caspian basins from the upper Miocene to the Holocene (Figure 1). The da- tabase for the study is regional two-dimensional (2-D) seismic surveys offshore Azerbaijan and Turk- menistan. Seismic-sequence-stratigraphic interpre- tations are integrated with well-log and outcrop data. The chronostratigraphic framework used here is derived from the integration of paleontological and radioisotopic dates with graphic correlation (J. Stein and R. Witmer, 1998, personal communica- tion) and tuning of events with the oxygen isotope record. Depositional infilling of the South Caspian Basin produced stratal geometries controlled by relative changes in lake level and local changes in sediment supply. Inthis case, lake-level changes are more likely controlled by changes in climate instead of eustasy, although there may be some indirect climatic link- age. Major tectonic unconformities are related to the closure of the connections between the Caspian and the Black Sea (the upper Miocene top of the Pontian Formation) and the onset of deformation along the Azeri coast and the Apsheron sill (top of Surakhany), the Apsheronian orogeny. CHRONOSTRATIGRAPHY ANDCLIMATE CYCLES The oxygen isotope curve from Deep Sea Drilling Program site 846 in the equatorial Pacific (from Shack- leton et al., 1995) is shown in Figure 1, together with 66 / Abreu and Nummedal the chronostratigraphy of the Azeri part of the South Caspian Basin. We infer that the very warm episode at the beginning of the Pliocene is associ- ated with the arrival of the sandy Pereryva delta and associated fluvial facies into the South Caspian Ba- sin (Figure 1). Overall, the isotope signature remains light throughout the deposition of the upper pro- ductive series. This persistent warm climate, and probable northward shift in the subtropical high- pressure zone, is consistent with the high sand-to- mud ratio for the Pereryva formation and con- trasts that unit with both the lower productive FIGURE 1. Chronostratigraphy of the late Miocene to Quaternary formations in the South and Central Caspian basins. The figure also shows the oxygen stable isotope record obtained by Shackleton et al. (1995) for the central Pacific Ocean. MioceneQuaternary Sequence Stratigraphy of the South and Central Caspian Basins / 67 series below and the rest of the upper productive series above. The base of the Balakhany formation may cor- respond to the first significant cold snap in the Plio- cene. This cold episode could be related to Caspian water-level rise, increasedpaleo-Volga discharge, wide- spread flooding, and the deposition of the basal Ba- lakhany shale member. Well logs through the Bala- khany interval in the Megastructure oil field show nine high-frequency sequences, each one consisting of basal fluvial sandstone followed by an overlying lacustrine mudstone interval. Overall, the Balakhany interval (Figure 1) probably represents a colder cli- mate than does the Pereryva. Climate remained relatively warm during deposi- tion of Sabunchi and Surakhany. In fact, the time interval from4.5 to3.5 Ma is commonly referred to as the golden age of climate, because of its equitable warmclimate. This time interval is also characterized by a rather monotonous lithologic succession, domi- nated by fine sand and siltstones inferred to have been deposited in a lake. The golden age termi- nated abruptly at about 3.3 Ma (early Piacenzian), and this may be associated with the onset of the Akchagylian flooding. After that, the climate began a cooling trend, leading into the major glaciations of the late Pliocene and Quaternary. WELL LOGS The Pliocene section of the paleo-Volga delta on- shore and offshore Azerbaijan appears to have been controlled by climatic cycles. Cores from the Mega- structure oil field show stacked, high-frequency se- quences (1015 m [3349 ft] thick), with sand at the base (lowstand systems tract), intercalated thin sand, and shales on top (transgressive systems tract), culminating with a shale bed with intense fractur- ing and/or bioturbation (maximum flooding sur- face), followed by a coarsening-upward succession of sand interbedded with mudstones (highstand sys- tems tract) with an erosional contact at the top of the highstand systems tract (sequence boundary), followed by fluvial sands, sometimes with mud clasts at the base (lowstand systems tract). These deposi- tional sequences are nested into lower frequency se- quence sets. The physical character of Pliocene sequences in the South Caspian Basin is here primarily addressed through the analysis of well logs and seismic re- flection profiles. The lowest frequency sequence ob- served in well-log patterns in this section extends from the base of Pereryva to the lower part of Bala- khany (represented in the two columns at the right- hand side of Figure 2). The two blocky sands observed in the Pereryva formation (from 2935 to 3050 m [9629 to 10,006 ft] in the well in Figure 2) corre- spond to a late lowstand systems tract. The trans- gressive systems tract corresponds to the thinner sands showing a fining-upward log motif from 2935 to 2740 m (9629 to 8989 ft). At the depth of 2740 m (8989 ft), a change intrends fromfining tocoarsening upward occurred, corresponding to the maximum flooding surface and marking the base of the high- stand systems tract, which extends until the depth of 2660 m(8727 ft) in the lower Balakhany. On average, this sequence is 415 m (1361 ft) thick, and consider- ing a rough estimate for the sediment accumulation rate in the Productive Series of 1 m/k.y. (3.3 ft/k.y.), this cycle could correspond to the 400-k.y. Milanko- vitch cycle. This low-frequency sequence can be subdivided in two higher frequency cycles, with a sequence boundary at the base of the blocky sand close to the base of Balakhany at 2860 m (9383 ft) (Figure 2). It is also possible to subdivide the apparent 400 k.y. in four higher frequency cycles, with two sequences in the Pereryva formation in the well in Figure 2 and two sequences in the lower part of the Balakhany formation. Thus, in the case of the Pereryva forma- tion in this well, each of the blocky sands would roughly correspond to the lowstand systems tract of a high-frequency sequence. The average thickness of these four high-frequency sequences is about 100 m (330 ft), which may correspond to the 100-k.y. Mi- lankovitch cycles (Figure 2). Deposition of the Pereryva and Balakhany suites seems to be basically controlled by orbital tuned climatic cycles, considering the strongly cyclic pat- tern and the different orders of cyclicity observed. The autocyclic component of the deposition would be restricted to shifting in delta fronts, bars, and channels in a same sequence. In one sequence, it is possible to observe varia- tion in the sand content in different systems tracts along strike and dip without a particularly predict- able behavior. This is probably related to autocyclic deposition. The dominant depositional systems for each sys- tems tract are inferred to be a late lowstand systems tract dominated by amalgamated fluvial sands deposited during a very early rise of the lake level. The transgressive systems tract would correspond to 68 / Abreu and Nummedal a time when lake level rose faster and appears be dominated by backstepping delta parasequences (wave dominated) interbedded with lacustrine sedi- ments. The highstand systems tract seems to be the least developed systems tract, at least in the South Caspian Basin, and is dominated by lacustrine mud- stones or locally by small coarsening-upward dis- tributary mouth-bar deposits. The maximum flooding surface in the middle of Pereryva is one of the best markers for regional cor- relation. Actually, the maximum flooding surface of the four high-frequency sequences is also excellent for regional correlation. INTERPRETING COASTAL ONLAP AND RELATIVE CHANGES IN BASE LEVEL Looking at large-scale basin fill, it has long been recognized that stratigraphic packages show cyclic- ity at several scales (e.g., transgressive-regressive cy- cles). One cyclic pattern of interest to seismic strati- graphers is coastal encroachment or coastal onlap. Coastal onlap is the progressive landward onlap of coastal (littoral or lacustrine) deposits ina givenstrati- graphic unit. Coastal aggradation (black arrows in Figure 3A) and coastal encroachment (gray arrows in Figure 3A) are the vertical and horizontal components FIGURE 2. Different orders of cyclicity observed in the Pereryva and lower Balakhany. MioceneQuaternary Sequence Stratigraphy of the South and Central Caspian Basins / 69 of coastal onlap, respectively. Coastal onlap indicates a relative rise of sea level (or lake level). A relative rise of sea level is an apparent sea level rise with respect to the underlying depositional profile. However, a basinward shift in coastal onlap indicates a relative fall in sea level (Figure 3B). Vail et al. (1977) proposed a methodology to de- termine the coastal onlap configuration of strata, re- lating the observed trends to relative changes in sea level. Detailed seismic-stratigraphic interpretationin- tegrated with biostratigraphic data allows the con- struction of chronostratigraphic charts (Figure 4). Measuring coastal aggradation, coastal encroachment, and downward shift in coastal onlap allows the con- struction of coastal onlap charts, which are graphic representations of relative changes insealevel through time (Figure 4). In this study, we used the methodology defined by Vail et al. (1977) and other classic seismic-stratigraphic methodologies (Mitchumet al., 1977) todefine changes in coastal onlap in the South and Central Caspian strata ranging in age from the uppermost Miocene to the Holocene and interpret the relative changes in lake level through time. Variations in coastal onlap through time were interpreted from the seismic data available. Vertical stacking of observed trends in coastal onlap is in- dicative of relative changes in lake level. Relative rise in lake level is inferred when coastal onlap of a younger sequence is located landward of the coastal onlap of the previous sequence. However, relative fall in lake level is inferred when coastal onlap of a younger sequence is located basinward of the coastal onlap of the previous sequence (downward shift in coastal onlap). These coastal onlap trends were correlated to the stacking patterns observed from wells in each of the three regions studied in this paper. FIGURE 3. Schematic diagrams showing coastal onlap (A), which indicates a relative rise in sea level, and a downward shift in coastal onlap (B), which indicates a relative fall in sea level (from Vail et al., 1977). 70 / Abreu and Nummedal Coastal onlap curves were built by plotting the position of the onlap against geologic time for each seismic horizon. Where the coastal onlap lies landward of the end of the seismic line, an estimate was made using stacking patterns observed in well logs, changes in offlap break, and sometimes, paleontological infor- mation. The relative lake-level curves were built using the position of sequence boundaries and maximum flooding surfaces as inflectionpoints for the falling and rising limbs of the lake-level curve, respectively. FIGURE 4. Schematic diagrams from Vail et al. (1977), showing a sequence-stratigraphic interpretation of five sequences (A). Diagram B shows a chronostratigraphic chart based on the interpretation shown in (A). The chronostratigraphic chart is an important step toward building a coastal onlap chart (C). MioceneQuaternary Sequence Stratigraphy of the South and Central Caspian Basins / 71 SEISMIC-STRATIGRAPHIC INTERPRETATION OF THE CASPIAN BASIN Western Margin of the South Caspian Basin: Interplay of the Paleo-Volga and Paleo-Kura Deltas Acontinuous landward shift of coastal onlap marks the upper Miocene and lower Pliocene sections (Pontian to Sabunchi) along the southwest margin of the South Caspian off the Kura delta of Azerbaijan (Figures 57). The paleo-Kura delta is interpreted to be representedinthe seismic profile showninFigure 5 as eight prograding depositional sequences during the Akchagylian. In map view, the offlap breaks of the paleo-Kura sequences are roughly parallel to the actual shoreline, with a straight outline typical of a wave-dominated delta. The top of the Akchagylian is associated with a major transgression along this basin margin. Figure 6 shows a Wheeler Diagram based on the seismic line shown in Figure 5. The lithology dis- tribution shown in the chart is an extrapolation of lithology trends observed in well logs in the region. Velocity estimates for time-depth conversion used stack velocities fromoriginal seismic data. Avery good visual correlation exists between the transgressive- regressive trends observed on seismic data and the li- thologyvariationtrends observedinwell logs (Figure 6). The landward shift in coastal onlap observed from Pontian to Sabunchi on seismic data (interpreted as transgression) is equivalent to a shaly section in the type well shown in Figure 6. The upper Surakhany and Akchagylian regressive phase of the paleo-Kura delta correlates to a sandy section in the type well. Coastal onlap and relative lake-level curves for the Azerbaijan margin were built on the trends observed in the offlap-break curve interpreted from seismic profiles (Figures 6, 7). Light gray color indicates a FIGURE 5. Line drawing of an interpreted seismic profile from the western margin of the South Caspian Basin, near the paleo-Kura delta. Offlap-break curve is shown in red (dashed line indicates the uncertainty in the location of the offlap break). 72 / Abreu and Nummedal coastal onlap not observed in the seismic profile because of lack of the data coverage. In these cases, the coastal onlap is tentatively interpreted from the offlap-break curve, stacking patterns observed in well logs, and paleontology data. Variation in cyclicity inferred from seismic and well information in the Central Caspian and Turk- menistan shelf in the Productive Series (400-k.y. cy- cles) and Quaternary (100-k.y. cycles) is not clearly observed along the Azerbaijan margin. There is a more even distribution of sequences through time and a less clear relationship between the number of sequences observed and possible predominant frequencies influencing the depositional sequences (Figures 57). Because the paleo-Volga delta was mostly back- stepping since the lower Pliocene whereas the paleo- Kura delta was prograding since the uppermost lower Pliocene, the sedimentation rate of the paleo-Volga delta was decreasing, and the sedimentation rate of the paleo-Kura delta was increasing through time. Consequently, the higher frequency cycles were sup- pressed in the seismic data at the distal margin of the paleo-Volga delta. Therefore, the strong interplay of the receding paleo-Volga and the prograding paleo-Kura along the FIGURE 6. Wheeler Diagram based on seismic profile shown in Figure 5. Vertical scale in the Wheeler Diagram is in geologic age (see Figure 1 for ages of the formations). The vertical scale for the type well shown in the right-hand column is in meters, with interpreted tops for the South Caspian formations in red, correlated to the interpreted tops in seismic data, shown in the Wheeler Diagram. The type well shows a sand-percentage curve interpreted from available well logs (gray color is mud, and yellow color is sand). Colors in the Wheeler Diagram represent interpreted relative changes in sand percentage associated with the different seismic units extrapolated from well information (increasing sand content from green [mud rich], to brown, light-yellow, and yellow [sand rich]). MioceneQuaternary Sequence Stratigraphy of the South and Central Caspian Basins / 73 Azerbaijan margin in the upper Pliocene and Qua- ternary probably inhibited the development of cyclic patterns similar to the ones observed elsewhere inthe South and Central Caspian basins. The interplay of the high-frequency lake-level fluctuations and the formation of the compressional structures during the upper Pliocene and Quaterna- ry controlled the sediment distribution along the FIGURE 7. Coastal-onlap, lake-level, and offlap-break curves for the western margin of the South Caspian Basin interpreted from the seismic profile shown in Figure 5 and Wheeler Diagram shown in Figure 6. In the coastal onlap column, dark gray color indicates a coastal onlap directly observed in the seismic profile, and light gray color indicates coastal onlap trends extrapolated from well information and the offlap-break curve. The lake level column shows two interpreted lake-level curves (high frequency = solid line; low frequency = dashed line). The high-frequency lake level curve was built using the coastal onlap curve, with sequence boundaries and maximum flooding surfaces used to define the inflection points of the falling and rising limbs of the lake-level curve. The low-frequency lake-level curve (dashed line) was built by connecting the points of high lake level from the high-frequency curve (solid line). The offlap-break column shows the offlap-break curve shown in Figure 5, converted from depth in two-way traveltime (seconds) to geologic age in millions of years. The type well shown in the right-hand column is the same as that shown in Figure 6, with vertical scale in meters. 74 / Abreu and Nummedal western margin of the South Caspian Basin. The growth of these compressional structures seems to have affected the lake floor at least since the upper Surakhany. Elongated minibasins formed between these structures, and the different stratigraphies in the minibasins and the ridge crests indicate periodic emergence of elongated islands during lowlake level. All systems tracts, including lowstand systems tract, are present within the minibasins, but only the maxi- mum flooding surface is present at the crest of the structures, indicating that the structures would be submerged only during periods of high lake level. During lowstands, the exposed structures would have formed broad and elongate islands, shedding sediments to localized lowstand packages in the minibasins. Central Caspian Basin: Influence of the Paleo-Volga and Samur Deltas Broad seismic coverage in the Central Caspian (Figure 8) shows a rising lake-level trend for the en- tire Productive Series. Strata related to the Pontian, lower productive series, and Pereryva and Balakhany suites onlap against the Kara-Bogaz high. Coastal onlap of depositional sequences associated withthese formations shows younger sequences continuously onlapping in a landward direction comparing to older ones, indicating a continuous rise inrelative lake level during this time. Basal strata associated with the Sa- bunchi suite onlap against the basinward parts of the Balakhany suite, indicating a major relative fall in lake level. Depositional sequences associated with the Surakhany suite bury the Kara-Bogaz high. This may indicate for the first time that the lake in the South and Central Caspian basins would be linked to or have flooded the North Caspian Basin in this region. In seismic profiles, the Surakhany and Akchagy- lian packages are mostly aggradational (Figure 8). Regression marks the base of the Apsheron, reflected by the prograding wedge of the paleo-Samur deltaic system originating at the northern Azerbaijan mar- gin. Thus, whereas the paleo-Volga probably con- tinuously backstepped during the upper Pliocene FIGURE 8. Line drawing of a seismic profile across the Central Caspian Basin demonstrating long-term onlap onto the Kara-Bogaz high. Offlap-break curve is shown in red (dashed line indicates uncertainty in the location of the offlap break). MioceneQuaternary Sequence Stratigraphy of the South and Central Caspian Basins / 75 and Quaternary, deltaic systems from the northern Azerbaijan margin (probably originated because of the uplift of the Caucasus) were prograding. The stratal pattern described above is consistent with a transgressive trend observed in well logs in the Central Caspian. Figure 9 shows a well drilled in the Central Caspian showing a fining-upward trend that correlates with the transgressive trend from the Pereryva to the top of Balakhany observed in the seismic data. Regressionduring the Sabunchi interval and renewed transgression in the Surakhany also seem to be represented in the well. An increase in sand toward the top of the Akchagylian seems to correlate withthe regressionof the paleo-Samur delta observed in seismic profiles. The Wheeler Diagram was built from the seismic line in Figure 8, using the top of the formations extrapolated from wells in the region. The Productive Series in the Central Caspian Basin is composed of eight seismically defined depositional sequences. Considering the duration of the Produc- tive Series (about 2.8 Ma), one sequence exists about every 350 k.y. These sequences might represent the 400-k.y. climate cycles because of orbital eccentric- ity, as defined in well logs along the Megastructure oil field. The Apsheron and younger strata are rep- resented by 12 depositional sequences, corresponding approximately to one sequence every 130 k.y. This apparent change in cyclicity could be related to a change in the predominant frequency that is regu- lating climatic cycles, from400 k.y. during the upper Miocene and lower Pliocene (Productive Series) to 100 k.y. during the Quaternary. Figure 10 shows a comparison between the coast- al onlap and relative lake-level curve with the offlap break curve obtained from a seismic line across the Central Caspian Basin. The influence of two distinct deltaic systems controlling the stacking patterns in the Central Caspian is remarkable on seismic lines. The transgressive trend observed during most of the Productive Series could indicate the backstepping of the paleo-Volga delta, and the regressive trend ob- served in the upper part of the section (Quaternary) is probably related to the prograding wedge of the paleo-Samur delta. Across the area that now constitutes the southern marginof the Central CaspianBasin(Apsheronsill and Apsheron Peninsula), the lake was probably shallow FIGURE 9. Wheeler Diagram for the Central Caspian Basin, obtained from the seismic profile shown in Figure 8. 76 / Abreu and Nummedal during the deposition of most of the Productive Series. The lake probably deepened gradually to the south- east and south, as suggested by pronounced clino- form reflectors associated to the upper productive series in the Amu Darya delta in Turkmenistan. Eastern Margin of the South Caspian Basin: Influence of the Paleo-Amu Darya Delta Two major regression phases mark the evolution of the paleo-Amu Darya delta since the uppermost Miocene. The first extends fromthe lower productive series to the mid-Surakhany, the second one fromthe Apsheron to Holocene (Figure 11). Similar to the paleo-Kura and paleo-Volga deltas, a major trans- gression characterized the paleo-Amu Darya delta during the Akchagylian. A pronounced basinward shift of the offlap break in the Productive Series is characterized by fast progradation from the lower productive series to middle Balakhany, with five seismically defined depositional sequences in this interval (Figure 11). From early Balakhany to late Surakhany, the offlap break shifted in a basinward FIGURE 10. Coastal-onlap, relative lake-level, and offlap-break curves for the Central Caspian compare with a sand percentage curve from a well in the region. MioceneQuaternary Sequence Stratigraphy of the South and Central Caspian Basins / 77 direction at a slower rate than at the base of this pro- grading wedge (Pereryva to early Balakhany), forming a progradational to aggradational stacking pattern (Figure 11). Pronounced extensional faulting characterizes the shelf break of the paleo-Amu Darya delta in the Pro- ductive Series, forming a detached extensional- compressional system in the southeast part of the South Caspian. The paleo-Amu Darya shelf break reached its most basinward position in the Produc- tive Series towards the end of the deposition of the Surakhany. Following the Akchagylian transgression, the paleo-Amu Darya again prograded continuously from the Pleistocene to the Holocene, forming a wedge with 18 seismically defined depositional se- quences (Figures 11, 12) representing about 1.8 m.y. Logs fromwells located on the Turkmenistan shelf show cycles similar to the ones inferred from seismic data (Figure 13). A continuous regressive trend exists from Sabunchi to the mid-Surakhany, with a contin- uous increase of sand up to the base of the Akchagy- lian that seems to be associated with the regressive trend observed inseismic profiles. The Surakhany and Akchagylian suites on the Turkmenistan shelf have the highest sand content in the South Caspian (as much as 60%). The Akchagylian and Apsheronian are marked by a single, long-term transgressive-regressive cycle with high sand content (Figure 13). A sharp de- crease in sand in the upper Pleistocene and Holocene is present. In this case, this fining-upward trend re- lates to muddy coastal sediments being deposited landward of the paleo-AmuDarya prograding wedge. FIGURE 11. Seismic profile from the eastern margin of the South Caspian Basin. 78 / Abreu and Nummedal Based on the sequence-stratigraphic interpreta- tion of the southeast Caspian seismic line, a Wheeler Diagramwas constructed and calibrated to the lithol- ogy estimates from well logs (Figure 12). The coastal onlap and relative lake-level curves for the Turkmeni- stan shelf (Figure 13) were built mostly based on the trends observed in the offlap-break curve interpreted from the seismic line (Figure 11). The change in dominant cycle durations observed in the Central Caspian from the Productive Series (400 k.y.) to the Quaternary (100 k.y.) is alsopresent onthe Turkmeni- stan shelf. The Productive Series is composed of nine depositional sequences, but of highly variable dura- tions. The four lower sequences in the upper produc- tive series are 100 k.y. each. Above these are longer duration cycles; some of those are about 400 k.y. in duration. Eighteen depositional sequences are ob- served in the Quaternary, which corresponds to ex- actly one sequence every 100 k.y. Thus, the 400- and 100-k.y. Milankovitch cycles seem to be domi- nant during the lower Pliocene and Quaternary, respectively, also on the Turkmenistan shelf. Figure 13 shows a comparison between coastal- onlap, relative lake-level, and offlap-break curves. Two major regressive phases of the paleo-Amu Darya delta occurred during the deposition of the Produc- tive Series and in the Quaternary. These regressive phases are represented in seismic lines (Figure 11) as two prograding wedges and are graphically shown in Figure 13 as relative falls in lake level and basin- ward shifts in the offlap-break curve. Although the predominant frequency of these two cycles is dif- ferent, it appears that the paleo-Amu Darya delta had a similar size and is a dominant factor in pro- gradation of the southern Turkmenistan margin both in the early Pliocene and the Quaternary. FIGURE 12. Wheeler Diagram constructed based on the seismic profile shown in Figure 11 from the eastern margin of the South Caspian Basin (Turkmenistan shelf). MioceneQuaternary Sequence Stratigraphy of the South and Central Caspian Basins / 79 TRANSGRESSIVE-REGRESSIVE CYCLES The greater percentage of sand in wells from the Central Caspian in the Pereryva to the Sabunchi suites (Figure 14) is caused by the proximity of these wells to the axis of the paleo-Volga delta. The paleo- Volga delta was located close to the Apsheron Pen- insula during much of the lower Pliocene. How- ever, the Balakhany thick observed in the Central Caspian suggests that the major paleo-Volga delta complex had moved north from the Apsheron Pen- insula at that time. Moreover, the transgressive pat- tern in the Productive Series observed in well logs in the Central Caspian Basin and consistent with the coastal onlap trend observed in seismic profiles in the Central Caspian (Figures 10, 14) seems to be related to the backstepping of the paleo-Volga delta. The paleo-Kura delta was present in the western margin of the Caspian Basin during the deposition of the Pereryva and Balakhany suites. Apparently, the FIGURE 13. Coastal-onlap, relative lake-level, and offlap-break curves for the eastern margin of the South Caspian Basin compared with a sand percentage curve from a well in the region. 80 / Abreu and Nummedal paleo-Kura sedimentation rate was in equilibrium withthe rising lake-level trendobservedinthe coastal onlap, resulting in an overall aggradational stacking. Backstep of the paleo-Kura delta occurred during the depositionof the uppermost Balakhany andSabunchi suites. Tworegressive trends are associatedtothe paleo- Kura delta in the Surakhany and Akchagylian, and Apsheron, separated by a major flooding event at the top of the Akchagylian. Surakhany and Akchagylian strataare thicker inthe Kuraregionthaninthe Central Caspian, related to the progradation of the paleo-Kura delta. Well logs from the eastern margin of the South Caspian Basin show a coarsening-upward trend (in- ferred to be a regressive pattern) from the base of the well to mid-Surakhany (Figure 14). From mid- Surakhany to the top of the Akchagylian, the well logs show a fining-upward trend (inferred to be a trans- gressive pattern). This pattern is consistent with the offlap-break curve for the eastern margin of the South Caspian (Figure 13). A maximum flooding surface marks the top of the Akchagylian in the Central and South Caspian basins (Figures 69, 11, 13, 14). Surakhany and Akchagylian strata are thicker and richer in sand on the Turkmenistan shelf and the western Caspian margin in the Kura region than in the Central Caspianregionbecause of regressionof the paleo-Amu Darya and paleo-Kura deltas (Figure 14). The whole Productive Series is thicker along the east- ernmargin of the South Caspian Basinthanin the rest of the region, indicating that the paleo-Amu Darya had greater sedimentation rate than the other deltaic FIGURE 14. Well-log correlation across the South and Central Caspian basins. Horizontal arrows at the base of the figure show the influence of the deltaic systems in the different regions. The Surakhany suite is shown in gray, as a reference. MioceneQuaternary Sequence Stratigraphy of the South and Central Caspian Basins / 81 systems. However, it is also apparent that the paleo- Volga system was richer in sand and with higher sedi- mentation rate during Pereryva and, perhaps, lower Balakhany. The stacking patterns on the eastern and western margins of the South Caspian are remarkably differ- ent in the upper productive series. On the western margin, a trend of aggradation to transgression exists fromthe lower productive series to the Sabunchi suite (Figures 14, 15), followed by regression in the Sura- khany and Akchagylian. On the eastern margin, the paleo-Amu Darya delta is mostly regressive until late Surakhany, with a major transgression during the Akchagylian (Figures 14, 15). The Central Caspian shows a trend similar to the western margin of the South Caspian Basin, with a continuous transgres- sion until the end of the Akchagylian (Figure 15). Thus, there is good evidence for rising lake level from Pereryva to Akchagylian in the South and Central Caspian basins. The regressive trend observed at the eastern mar- gin of the South Caspian throughout the Productive FIGURE 15. Summary of the transgressive and regressive trends in the South and Central Caspian basins. Solid gray lines represent the offlap-break curves for the west and east margins of the South Caspian and Central Caspian from Figures 7, 10, and 13. Green dashed lines represent transgressive trends, and red dashed lines represent regressive trends in the different localities. 82 / Abreu and Nummedal Series (Figure 15) is related to the high sedimenta- tion rates in the paleo-Amu Darya delta, probably associated with the uplift of the Pamirs and the Himalayas. High sedimentation rates of the paleo- Amu Darya overcame the rate of rising lake level, causing progradation on the Turkmenistan shelf. Overall maximum transgression for the Caspian Sea occurred at the end of the Akchagylian, right at the start of the Pleistocene (Figures 14, 15). This also appears to be the time when the Caspian was last linked to the world oceans (Nummedal et al., 2000). Regression was the dominant pattern for all these margins of the Caspian during the Pleistocene and the Holocene. The regression of the Caspian during the Quaternary is probably related to the continu- ous uplift of the Caucasus and the global ice-house climate (Figures 14, 15). No regression is observed in the paleo-Volga delta in the Central and South Caspian since the Akchagylian. The paleo-Volga del- ta was probably continuously backstepping at least since the Surakhany, as indicated by the coastal on- lap trends observed in the Central Caspian Basin (Figures 9, 10, 15). PALEOGEOGRAPHY The incised valley of the paleo-Volga is a con- spicuous seismic feature in the Central Caspian, lo- cated north of the Apsheron Peninsula and with northsouth orientation (Figure 16). To the south of this incised valley lies the early Pliocene paleo- Volga delta, forming a broadly arcuate geometry with a sandy depocenter at the eastern part of the Apshe- ron Peninsula (Figure 16). The sand percentage de- creases away from this depocenter in an asymmetric pattern skewed toward the east. Considering this broad distribution of fluvial and deltaic sediments in the Pereryva and the lack of observed prograding wedges associated with this system, it is inferred that the paleo-Volga delta was probably deposited on a ramp setting. Figure 16 shows maps of onlap terminations of the base of the Tertiary, base of Pereryva, top of lower Balakhany, and middle Balakhany (P. Ware, 1998, personal communication). The base Pereryva is re- stricted to the southwest part of the Central Caspian because of its onlap against the Kara-Bogaz high to the north. The incised valley of the paleo-Volga river is clearly shown in the map for the top of Balakhany, with a northsouth orientation in the western side of the Central Caspian Basin. The trans- gressive trend is continuous during the Balakhany, as illustrated by the landward shift of the coastal onlap from the tops of the lower and middle Ba- lakhany and the broadening of the paleo-Volga val- ley. A good correspondence exists between the po- sition of the paleo-Volga incised valley defined in seismic profiles in the Central Caspian Basin and the position of the paleo-Volga delta defined by sand package correlation in well logs (interpretation by A. Milton). The prograding wedge of the paleo-Amu Darya delta occupied the inner part of the Turkmenistan shelf duringthe depositionof the Pereryva (Figure 17). The deepest water depth in the lake appears to have been directly in front of this system. Reconstruc- tion of seismic lines across the Turkmenistan mar- gin indicates that the Caspian Sea was several hun- dreds of meters deep at this location during the deposition of the Pereryva (Hickmanet al., 1999; Hick- man and Stuart, 2001). Thus, the depth profile across the Turkmenistan margin consisted of a shelf, slope, and basin, distinctly different from the ramp off Azerbaijan. No deltaic system was present at the western margin of the South Caspian Basin during most of the Productive Series time (Figure 17). Onlapping Pereryva strata indicate that the Cen- tral Caspian Basin was bounded on the north in this region by the Kara-Bogaz high(Figures 16, 17). Onlap of Pereryva is also observed along the western mar- gin of the South Caspian Basin. Pereryva sediments are present onshore Turkmenistan (Figure 17), and the eastern onlap margin is unknown. This distri- bution of the Pereryva suite across the Central and South Caspian, and its total absence in the North Caspian, indicates that the Caspian Sea was elon- gated in an eastwest direction (Figure 17) in the early Pliocene. This pattern appears to have persisted at least throughout the deposition of the Balakhany (Figures 16, 17). During the deposition of the Surakhany, the Cas- pian Sea flooded over the Kara-Bogaz high and de- veloped the northsouth orientation that it has today (Figure 17). It is also possible that a passage- way existed between the South Caspian and the Black Sea through Azerbaijan during the deposition of the Pereryva and Balakhany suites, as indicated by the presence of a thick sedimentary package in the Kura basin, onshore Azerbaijan (Figure 16). From Pereryva to Balakhany, the paleo-Volga delta de- creased in size, and the paleo-Amu Darya delta expanded. The Caspian Sea continuously expanded during the lower Pliocene, as indicated by the coastal MioceneQuaternary Sequence Stratigraphy of the South and Central Caspian Basins / 83 onlap in the Azerbaijan margin and in the Kara- Bogaz high. During the Surakhany deposition, lake level rose above the Kara-Bogaz high, connecting the southern andnorthernCaspianbasins inthis region(Figure 17). This flooding event caused the backstepping of the paleo-Volga, pushing the delta to the northern Caspian. In the South Caspian Basin during the late Pliocene, the paleo-Amu Darya delta was most- ly aggrading the Turkmenistan shelf, and the paleo- Kura delta was probably being formed onshore Azerbaijan. The paleo-Kura and paleo-Amu Darya deltas were mostly regressive after the uppermost Surakhany (Figure 17). A major backstep of these systems oc- curred close to the top of the Akchagylian. During the Akchagylian transgression, the Caspian Sea was probably linked to the Black Sea via a North Caspian connection. There was a strong basinward shift of coastal on- lap during the Apsheronian in the South and Cen- tral Caspian. A series of deltaic systems were formed in the western margin of the Caspian (Figure 17), which maybe related to an uplift in the Caucasus FIGURE 16. Position of the paleo-Volga incised valley and delta from Pereryva to middle Balakhany in the Central and South Caspian basins. The figure also shows the onlap of Pereryva and Balakhany strata in the Central Caspian Basin (A. Milton and P. Ware, 1998, personal communication, respectively). 84 / Abreu and Nummedal Mountains. Along the eastern margin, the paleo- Amu Darya was prograding at a fast rate, reaching a similar size and position as it achieved during the deposition of Balakhany. CONCLUSIONS Sequence-stratigraphic interpretation of well logs at the Megastructure (ACG) oil field indicates that the depositional systems in the South Caspian Basin were basically controlled by orbitally driven climatic cycles. Sequences resolved in the seismic data ana- lyzedinthis paper represent cycles of 100 and400 k.y. duration. Coastal onlap and relative lake-level curves for the Caspian Basin were built to address the relative importance of sediment sources for the different deltaic systems in the Central and South Caspian basins. A continuous trend exists toward higher lake level from the lower productive series to the Ak- chagylian. This relative rise in lake level caused the backstepping of the paleo-Volga delta. However, the high sedimentation rate of the paleo-Amu Darya delta overwhelmed the rising lake level, forming a prograding wedge across the Turkmenistan shelf during the upper productive series. The paleo-Kura delta aggraded on the southern margin of the west- ern South Caspian Basin during the deposition of the Pereryva and Balakhany suites and prograded during the deposition of the upper Surakhany and the Ak- chagylian. The major deltaic systems in the Caspian Basin backstep toward the top of the Akchagylian Suite. The paleo-Kura and paleo-Amu Darya deltas were mostly regressive in the Quaternary, whereas the paleo-Volga experienced a major transgression. Patterns of lithologic change and onlap termina- tions across the Central and South Caspian basins FIGURE 17. Schematic paleogeographic reconstruction of the Caspian Sea for the Pliocene and Quaternary, showing the change in the position of the three main deltaic systems in the South and Central Caspian, as well as the location of the possible connections between the Caspian Sea and the Black Sea. MioceneQuaternary Sequence Stratigraphy of the South and Central Caspian Basins / 85 show that during the deposition of the Pereryva suite, the paleo-Volga delta was the most important sediment source for the Caspian Sea. Detailed seis- mic and well-log interpretation indicates that the paleo-Volga incised valley was active during the Pe- reryva and Balakhany, sourcing a broad delta front localized at the Apsheron Peninsula region. During the Balakhany, the paleo-Volga was backstepping, and the paleo-Amu Darya delta was prograding, to eventually become the most important sediment source for the South Caspian during Surakhany time. The Kara-Bogaz high in the Central Caspian was an effective barrier for lake expansion until the deposi- tion of the Surakhany. Thus, the lake was mostly re- stricted to the South Caspian during the deposition of the Pereryva and Balakhany, expanding to the east in onshore Turkmenistan and to the west in onshore Azerbaijan. Therefore, the lake was elongated in an eastwest directionduring the depositionof the lower productive series and turned to its current north south orientation during the Surakhany deposition. In summary, large-scale transgressive and regres- sive trends in the different margins in the Central and South Caspian basins were strongly controlled by local sediment supply, which was controlled by tectonic uplift of source areas, local subsidence, and climate. ACKNOWLEDGMENTS We are thankful to the Unocal Corporation for the permission to publish this work. We express our gratitude to Andrew Crossley, Marek Kacewicz, Dan Self, Art Trevena, Roger Witmer, and Eileen Wil- liams for suggestions and discussions. This work al- so greatly benefited from discussions with Kirmaku Valley Project team, mainly Ed Clifton (formerly of Conoco), Greg Riley, and Jeff Stein (formerly of Amoco). The authors are also very thankful to Jack Neal (ExxonMobil) for reviewing a later version of the manuscript. REFERENCES CITED Berggren, W. A., D. V. Kent, C. C. Swisher III, and M.-P. Aubry, 1995, A revised Cenozoic geochronology and chronostratigraphy, in W. A. Berggren, D. V. Kent, M.-P. Aubry, and J. Hardenbol, eds., Geochronology, time scales and global stratigraphic correlation: SEPM Special Publication 54, p. 129212. Hickman, R. G., and C. J. Stuart, 2001, Linked extension- compressional structures in deltaic settings (abs.): AAPG Annual Convention Program, p. A87. Hickman, R. G., V. Abreu, and D. Self, 1999, Detached extensional Compressional structures, paleo-Amu Darya delta, Turkmenistan shelf, South Caspian Sea (abs.): AAPGAnnual ConventionProgram, v. 8, p. 5960. Mitchum, R. M., P. R. Vail, and S. Thompson III, 1977, Seismic stratigraphy and global changes of sea level: Part 2. The depositional sequence as a basic unit for stratigraphic analysis, in C. E. Payton, ed., Seismic stratigraphy Applications to hydrocarbon explora- tion, AAPG Memoir 26, p. 5362. Nummedal, D., et al., 2000, Sedimentological response to climate cycles during the Pliocene of the SouthCaspian Basin, in AAPG Inaugural Regional International Con- ference, Istanbul, Turkey, July 912, 2000, Program Book, p. 223. Shackleton, N. J., S. Crowhurst, T. Hagelberg, N. G. Pisias, and D. A. Schneider, 1995, A new late Neogene time scale: Application to Leg 138 sites, in N. G. Pisias, L. A. Mayer, T. R. Janecek, A. Palmer-Julson, and T. H. van Andel, eds., Proceedings of the Ocean Drilling Pro- gram, Scientific Results, v. 138, p. 73101. Vail, P. R., R. M. Mitchum Jr., and S. Thompson III, 1977, Seismic stratigraphy and global changes of sea level: Part 3. Relative changes of sea level fromcoastal onlap: Section 2. Application of seismic reflection configura- tiontostratigraphic interpretation, in C. E. Payton, ed., Seismic stratigraphy Applications to hydrocarbon exploration: AAPG Memoir 26, p. 6381. 86 / Abreu and Nummedal