Earthquake clustering inferred from Pliocene Gilbert-type fan deltas in the Loreto basin, Baja California Sur, Mexico

A stacked sequence of Pliocene Gilbert-type fan deltas in the Loreto basin was shed from the footwall of the dextral-normal Loreto fault and deposited at the margin of a marine basin during rapid fault-controlled subsidence. Fan-delta parasequences coarsen upward from marine siltstone and sandstone at the base, through sandy bottomsets and gravelly foresets, to gravelly nonmarine topsets. Each topset unit is capped by a thin shell bed that records marine flooding of the delta plain. Several mechanisms may have produced repetitive vertical stacking of Gilbert deltas: (1) autocyclic delta-lobe switching; (2) eustatic sea-level fluctuations; (3) climatically controlled fluctuations in sediment input; and (4) episodic subsidence produced by temporal clustering of earthquakes. We favor hypothesis 4 for several reasons, but hypotheses 2 and 3 cannot be rejected at this time. Earthquake clustering can readily produce episodic subsidence at spatial and temporal scales consistent with stratigraphic trends observed in the Loreto basin. This model is supported by comparison with paleoseismological studies that document clustering on active faults over a wide range of time scales. Earthquake clustering is a new concept in basin analysis that may be helpful for understanding repetitive stratigraphy in tectonically active basins. Figure 1. A: Location and tectonic setting of Loreto basin. B: Geologic map of Loreto basin, showing location of study area (Fig. 2). *Present address: Dorsey: Department of Geological Sciences, 1272 University of Oregon, Eugene, OR 97403-1272 (E-mail: [email protected]); Falk: Softrock Inc., 591 County Rd. 233, Durango, CO 81301. sey et al., 1995; Falk, 1996). In this paper we call attention to the necessary direct links between any model for faulting controls on fan-delta cyclicity and implications for paleoseismicity. We suggest that temporal clustering of earthquakes is a reasonable hypothesis to explain the observed spatial and temporal patterns of episodic subsidence, which appears to have controlled accumulation of stacked fan deltas in the Loreto basin. GEOLOGIC SETTING AND STRATIGRAPHIC DATA The Pliocene Loreto basin, an oblique halfgraben basin located on the western margin of the Gulf of California, formed by asymmetric subsidence along the dextral-normal Loreto fault (Fig. 1; Umhoefer et al., 1994). The basin was rapidly filled with >1200 m of nonmarine, deltaic, and marine deposits, including a heterolithic assemblage of conglomerate, sandstone, and mudstone that accumulated in Gilbert-type and shelf-type fan deltas (Fig. 1B; Dorsey et al., 1995; Falk, 1996). Pronounced westward thickening of the section and abrupt lateral transitions from nonmarine to marine facies provide evidence for westward tilting and filling of the basin during subsidence on the Loreto fault. Gilbert-type fan deltas lie stratigraphically between tuffs 2 and 3, dated as 2.46 ± 0.06 Ma and 2.36 ± 0.02 Ma, respectively (Umhoefer et al., 1994). Decompacted thicknesses combined with tuff ages show that the Gilbert deltas accumulated during a short period (~100 000 yr) of very rapid subsidence (8 ± 5 mm/yr). This study focuses on gently dipping, wellexposed, footwall-derived Gilbert-type fan deltas and associated marine and nonmarine facies, the stratigraphic architecture of which is revealed in a north-south facies panel (Fig. 2). The panel is ~540 m thick and 3600 m wide, and the south edge of the panel is ~1 km north of the Loreto fault. Stratigraphy in this area consists of nonmarine conglomerate and sandstone in the south that pass laterally through Gilbert-type fan deltas into marine sandstone in the north (Fig. 2). These Gilbert-type fan deltas comprise progradational, coarsening-upward parasequences consisting of (in ascending order): heavily bioturbated, massive, marine siltstone; well-bedded, sandy turbidites (bottomsets); bedded and stratified marine conglomerate and sandstone with primary dips of 15° to 25° (foresets); and marginal-marine to nonmarine conglomerate and sandstone (topsets) (Fig. 2B). Each Gilbert-delta topset unit is capped by a laterally continuous marine shell bed (average thickness = 1 m), which records a hiatus in sediment delivery, rapid marine transgression, and submergence of the underlying deltaic plain to shallow-marine depths (10–30 m) (Fig. 2; Dorsey et al., 1995; Falk, 1996). Parasequences 1–10 are 20–35 m thick, and parasequences 11–16 are 46–78 m thick. Paleocurrent data record overall transport toward the north, parallel to the facies panel (Fig. 2B; Falk, 1996). We can estimate the average frequency (or repeat time) of deltaic parasequences by using the tuff ages described above. The lower tuff (tuff 2 of Umhoefer et al., 1994) is seen near the base of the facies panel (Fig. 2B), and tuff 3 is stratigraphically above the top of the panel. The age difference between the two tuffs is 100 000 ± 80 000 yr. There are 16 parasequences of variable thickness (including 3 more above area of Fig. 2B) preserved between the two tuffs. This gives an average frequency of 6250 ± 5000 yr per parasequence. The actual duration of any parasequence may vary significantly from the average, and if we include subcycles (e.g., 10a, 12a, 12b, and 12c in Fig. 2) the calculated average frequency would be even shorter. This simple analysis reveals very rapid, high-frequency production of Gilbert-type fan deltas in the Loreto basin. INTERPRETATION OF FAN-DELTA CYCLES Each parasequence was produced by rapid progradation of gravelly fan deltas 1–2 km into the shallow-marine basin; each shell bed formed during a sharp reduction in sediment accumulation, submergence of the delta plain, and rapid transgression of the shoreline back toward the Loreto fault. Advance and retreat of fan deltas were controlled by large changes in the balance between the rate of sediment input (which may be internally or externally controlled) and the rate of basin subsidence, possibly modulated by fluctuations in eustatic sea level. In this section we evaluate four processes that could have controlled episodic growth and abandonment of Loreto fan deltas. 1. Repetitive stacking of fan deltas may have resulted from autocyclic channel avulsions and delta-lobe switching superimposed on steady rapid subsidence. Lobe switching is common in delta systems, and detailed stratigraphy in these deltaplain deposits indicates that channel switching was active. This hypothesis predicts that Paleocurrent directions should vary substantially from one parasequence to the next, because the direction of input would have to change in order to shift deltaic lobes laterally to different depocenters through 680 GEOLOGY, August 1997 Figure 2. A: North-south facies panel showing detailed stratigraphic architecture of Gilbert-type fan deltas in southern Loreto basin. SB1 through SB13 are 0.5–1.0-m-thick shell beds that cap topset units.Vertical lines are detailed measured sections. B: Map of nonmarine facies (shaded), marine facies (white),and numbered shell beds that define fandelta parasequences in southern Loreto basin. NRA is north Rodadas Arroyo; SRA is south Rodadas Arroyo. Dashed north-south line is line of projection used for constructing upper part of facies panel (A).