Paleoseismology of the Zia Fault and Its Impact on Seismic Hazard for Albuquerque, New Mexico, USA Open Access

This study reconstructed the chronology of the latest surface-faulting events on the Zia Fault following the abandonment of the ca. 1.5 Ma Llano de Albuquerque (LdA) geomorphic surface. The project used trenching to measure mid- to late Quaternary slip rates, the age of the most recent paleoearthquake, and recurrence intervals. Those data were then compared to similar data for the County Dump (McCalpin et al., 2006), Calabacillas (McCalpin et al., 2011), and East Paradise (Personius and Mahan, 2000) Faults to characterize the Quaternary activity of all western rift-margin faults trenched to date in the northern Albuquerque Basin. We spent 3 weeks excavating, logging, and sampling three trenches on the northern end of the Zia Fault. Subsequent to this, the U.S. Geological Survey (USGS) incorporated the fault slip rates into the 2023 update of the U.S. National Seismic Hazard Map (NSHM). The paper closes with a discussion of the impact of these slip rates on the seismic hazard assessment of the Albuquerque metropolitan area.

The Albuquerque-Belen Basin is one of the largest structural depressions in the Rio Grande Rift, measuring 135 km in N-S dimension and up to 60 km in width (E-W dimension) (Kelley, 1977). At the latitude of Albuquerque (35°N; Figure 1), the basin is bounded on the east by the steep range front of the Sandia Mountains and on the west by a low mesa (the LdA or West Mesa). Basin asymmetry indicates it is an east-tilted half graben with the master fault on the east and dipping west. West of the Rio Grande river, numerous east-dipping, antithetic normal faults are expressed by Quaternary fault scarps on the LdA; from west to east, these faults are the Calabacillas segment of the San Ysidro Fault, Centipede segment of the Zia Fault, main segment of the Zia Fault, and East Paradise Fault (Figure 2). Figure 3 shows an upper-crustal cross section of the Zia Fault reconstructed from borehole data and geophysics. At this latitude in the western rift margin, the Zia Fault displays the most vertical displacement, flanked by secondary normal faults in the Calabacillas sub-basin and the Montoyas graben.

Prior to our trench study, no detailed paleoseismic work had been performed on the Zia Fault (Machette et al., 1998, p. 157), although the fault was mapped as early as 1998 by Cather et al. (1997). Based on early mapping, Machette et al. (1998, p. 157) cited the slip rate on the Zia Fault as “unknown; probably <0.02 mm/yr” and the timing of the most recent paleoearthquake as “middle and late Quaternary (<750 ka).” The four major faults of the LdA were trenched in the late 1990s to early 2000s (Table 1), with paleoearthquake dating conducted via thermoluminescence (TL), infrared stimulated luminescence (IRSL), and accumulation rates of secondary soil carbonate. Few trenches contained datable carbon. With published slip rates, the four faults qualified for inclusion in the 2023 NSHM (Table 1).

Early calculations of Zia Fault slip rates were complicated by an error in the numbering of the luminescence samples, which led to stratigraphically inverted ages. While this paper fixes that problem, the luminescence ages (TL, IRSL) are still much older than ages predicted from the accumulation rate of secondary soil carbonate (discussed later).

Our trench site at the northern rim of the LdA is drained by shallow ephemeral streams that trend southeast and east at an average slope of 1.5°. Streams have been beheaded by the rapid southward migration of the rim, driven by relatively rapid down-wasting of the soft Santa Fe Group in the badlands north of the LdA. On the LdA surface, the Zia Fault is expressed as a 5 m-high east-facing scarp and an adjacent lowland that we interpret as a filled graben (Figure 4). The main fault scarp trends roughly N10°W and forms the margin of a 100-m-wide, 300-m-long graben. Antithetic scarps of the graben are poorly preserved and lie just east of the Sec. 31/32 fence and the road that parallels this section. The main fault scarp is transected by two gullies that have incised into the upthrown block and have built vague alluvial aprons (Qal on Figure 4) into the graben. Both gullies head only 50–75 m west of the scarp and thus carry only locally derived runoff.

The scarp is narrower and steeper than the broad and high scarps of the County Dump and Calabacillas Faults. We initially thought that scarp steepness was caused by an indurated calcrete soil on the LdA surface. However, trenching revealed only a moderately developed, relatively friable K horizon. The real reason for scarp steepness here is that slivers of very hard, fine-grained Pantadeleon Formation (Pleistocene to Pliocene?) had been brought to the surface along the main fault plane.

The northern rim of the LdA lies roughly 300 m north of our trenches, and its steep north-facing slopes expose an ∼20-m-thick stack of syn-tectonic deposits and paleosols on the hanging wall (HW) of the Zia Fault (Figure 5). These deposits were named the Pantadeleon Formation by Connell et al. (1999), and they represent fluvial, marsh, and eolian sediments that accumulated in the fault-angle depression caused by repeated down-to-the-east slip on the Zia Fault. The syn-tectonic wedge reaches its maximum thickness (∼20 m) against the Zia Fault and thins to zero about 1 km to the east. The wedge is very similar to that exposed on the HW of the County Dump Fault (McCalpin et al., 2006), which is 18 m thick.

We excavated three trenches at this site (Figure 6), each for a different reason. The middle trench was excavated across the steepest, highest part of the fault scarp, and it was the largest trench. Its purpose was to reveal the colluvial wedge stratigraphy that represents the effects of the latest several paleoearthquakes, and to yield a glimpse of the thickness and character of the graben fill. The southern trench was excavated in the axis of the southern gully that cuts through the fault scarp. Its purpose was to expose the alluvial fill that unconformably overlies the fault plane, which might yield a minimum age on latest faulting. The northern trench was excavated on a portion of the main fault scarp that exhibited a 1-m-high, unvegetated band that looked like a young free face. It was unclear if this free face represented very young faulting or resulted from erosion unrelated to faulting. The north trench exposed no faults beneath this “free face,” so the band must have been created by erosion of the original scarp back into the footwall (FW). Paleoseismic interpretation is based almost entirely on the middle trench.

The middle trench cut the fault scarp where it is best preserved, and this trench was 1 m wide, up to 6 m deep, and 36 m long (Figure 7). The scarp crosscuts an ambient slope of 4° and has 4 m of vertical separation. The trench extended westward to the crest of the fault scarp, in an attempt to expose the relict soil developed on the LdA surface (petrocalcic horizons 1K1/1K2). Therefore, the western 17 m section of the trench was only 1.5 m deep and unshored. Once the fault zone was exposed by the backhoe, the trench was deepened to 6 m (Figure 8).

We defined 90 map sub-units in the middle trench, of which 51 are stratigraphic sub-units and 39 are soil horizons (Figure 9).

The fault FW is composed of Santa Fe Group (major unit 1), subdivided into 3 sub-units (units 1a–1c). These sub-units are textural types (1a, stratified sandy gravel; 1b, stratified sand; 1c, unstratified gravelly sand), so multiple beds carry the same sub-unit designation. These beds are inferred to represent mainly channel alluvium.

The fault zone is composed of at least six major fault-bounded “slivers,” and correlation of the stratigraphy between slivers (units 2–8) is somewhat ambiguous. In general, we could not correlate strata between adjacent slivers, which implies that the vertical displacement on the faults is greater than the vertical exposed thickness in the slivers.

Accordingly, we typically assigned a new major unit number at the bottom of the exposed stratigraphic sequence within each fault sliver. In addition, we assumed that all faults are normal faults, and thus the deposits are youngest in the easternmost fault sliver and oldest in the westernmost fault sliver (abutting the Santa Fe Group).

Most units in the fault slivers are hard to very hard massive sands, inferred to be either fluvial or eolian components of the Pantadeleon Formation (sensu Connell et al., 1999). The Pantadeleon Formation disconformably rests on strongly developed petrocalcic soils (paleosols) correlated by Connell et al. (1999) to the LdA surface. Later, Connell (2008b) recommended abandonment of the Pantadeleon Formation because it could not be mapped regionally. However, for the purposes of this trenching study, it is a useful stratigraphic unit.

In addition to sandy deposits, units 3g, 3i, 3k, 3n, 4b, 4d, 4e, 4f, 4i, and 4k (Figures 7 and 9) are extremely hard clays, similar to clays exposed in the Pantadeleon badlands exposures on the rim of the LdA. These units are interpreted as sag pond or marsh (ciénega) deposits, laid down in standing water in the closed part of a tectonic depression up against the Zia Fault.

The HW exposes an old compound soil (from top to bottom, horizons 9bBtk1, 9bBtk2, 9bBtk3, 9aBt, 9aBk, and 9aCk) in the bottom of the trench, unconformably overlain by a 1.5–2-m-thick sequence of coarse fluvial gravels (units 10–12). Coarse gravels grade upward into finer, thinner-bedded fluvial gravels and sands (units 20–21), which in turn grade upward into thin sheets of eolian sand (units 22–24). Overlying the sheets, there are two wedge-shaped deposits of sand (units 25 and 26) that we interpret as colluvial wedges dominantly composed of reworked eolian sand. Finally, the entire fault scarp is mantled by thin slope-wash colluvium (units 30–33 and 40).

The compound soil developed on a sandy parent material, the original texture of which is now obscured by strong B horizon development. The soil underlies basal gravels of units 10–12, both in the HW and in the fault zone (that is, west of fault F6), but correlation across fault F6 is uncertain. Thus, we define the two old soils to be developed on units 9a and 9b on the HW, but in the fault zone, we number the soil beneath the gravels as unit 8. This change in numbering reflects the possibility that the soil beneath the gravel in the fault zone developed on deposits of a deeper stratigraphic level than soil on the HW.

Units 10–12 form a nested sequence of relatively thick (0.5–1 m) channel gravels, with units to the east being incised into units to the west. This pattern, as well as the consistently east-dipping cross-beds in unit 11, indicates that the gravels were deposited by a distant stream that initially flowed up against the scarp (unit 10) and then migrated progressively eastward. This pattern is expected if the stream was initially diverted into the tilted topographic depression at the scarp base and then flowed parallel to the fault scarp, but then migrated farther east onto the HW.

The package of three units (12b–12d) represents a stream that was diverted westward from its earlier position on the HW (unit 12a) to flow once more up against the fault zone. Such a scarp-ward migration may have been caused by a faulting event that tilted the HW toward the fault scarp. A major change in deposition occurs with unit 12g, which is much thicker and more poorly stratified than units 12b–12f. The western parts of units 12b–12e have been disrupted by faulting (gray hachured area on log), but unit 12g overlies this disrupted area, is undeformed, and can be traced intact over the deformed beds to the next fault sliver to the east. This geometry suggests that a faulting event on fault F4 affected units 12b–12e but did not affect unit 12g.

Units 20–21 comprise a transition from the purely fluvial deposits of unit 12 to the purely eolian deposits of units 22 and higher. Units 20–21 are composed of very thin sandy beds with thin intercalated soil horizons.

Units 23–24 form tabular bodies of massive sand and are interpreted as eolian blankets that may have once mantled the entire fault scarp.

Units 25 and 26 are wedge-shaped deposits of massive sand that reach a maximum thickness against the fault zone (fault F6). Both units are very soft and friable and contain only very weak calcareous soil horizons (the lower horizons of the surface soil). Unit 25 is clearly faulted by strands of fault F7 (deposited after the PE), whereas unit 26 is unfaulted (deposited after the most recent event).

The entire ground surface of the scarp is mantled by a thin reworked deposit formed by short downslope transport and creep of subjacent units, with a heavy admixture of eolian sand (unit 31). Unit 32 comprises fissure fill within the fault zone. Unit 40 is the youngest deposit exposed by the trench, and it comprises an eastward-thickening wedge of eolian/fluvial sand that laps onto the toe of the fault scarp.

The major structural features of the trench are, from east to west: (1) an antithetic fault (F15) and associated small graben at the toe of the fault scarp, (2) the HW sequence of colluvial wedges and alluvium (units 9–25), now tilted 7° toward the fault, (3) a zone of small-displacement synthetic faults in HW strata (F7a through F7e), and (4) a complex zone of fault slivers of Pantadeleon Formation (F6–F1), the westernmost sliver of which (F1) is in fault contact with the Santa Fe Group. The small graben at the toe of the trench has only 0.5 m of cumulative down-to-the-west displacement across it, measured on the top of unit 9. The graben is interpreted as an extensional hinge that separates gently dipping HW strata (to the east) from 7°-tilted HW units (to the west). Eolian/colluvial units 21–26 pinch out at the graben, supporting its function as a hinge and accommodating HW rotation since deposition of unit 21.

The westernmost part of the HW stratigraphic sequence is cut by a zone of five small-displacement faults (fault F7a through F7e) that dominantly dip steeply west. Because these faults all display down-to-the-east displacement, their dip gives them the appearance of reverse faults. However, the main fault (F1) juxtaposing Quaternary against Tertiary deposits is clearly a normal fault, as is the Zia Fault at crustal scale (Figure 3). These five west-dipping faults are inferred to be splays off of major fault F6 that refracted to vertical and then to “overhanging” dips as they propagated upward through the relatively soft deposits of units 9–25. This type of “overhanging” fault structure is commonly observed near the surface along normal faults (McCalpin et al., 2006) and is sometimes interpreted as toppling of fault-bounded slivers toward the HW. All faults are inferred to merge with the east-dipping fault F6 several meters beneath the trench floor.

The main deformation zone is composed of six east-dipping normal faults (F6–F1) in a zone 2–2.5 m wide. These faults have relatively consistent dips (65°–80° east) and are spaced 0.25–0.75 m apart (Figure 10). The faults thus bound fault “slivers,” each of which contains a different part of the stratigraphic section of the Pantadeleon Formation. In the F6–F4 sliver, units 10–22 are easily correlated across F6. Beneath unit 10, there is a >1.6-m-thick complex paleosol (units 8Bk2 down through 5Bk) that may correlate (wholly or partly) with the >1-m-thick HW paleosol at the bottom of the trench (units 9bBtk1 down through 9aCk). Because the paleosol is not in direct contact across F6, we did not number it unit 9 on the F6 FW, but rather denote it as unit 8, admitting the possibility that it might be older than units 9a and 9b on the HW. The F4-F3 sliver contains 17 mappable strata, the upper six of which are correlated with the lowest beds in the F6–F4 sliver (units 8Bk2 to 5Bk), based on horizon development and thickness. Beds in the lower two thirds of the sliver (units 4j down through 4a) predate unit 5Bk and are not observed in the HW or in the other fault slivers.

Strata in the F3-F2 sliver (units 3n down to 3a) do not contain any recognizable paleosols and are denser (from diagenesis?) and more deformed than sliver units to the east, dipping 35°E at the top to as much as 65°E at the bottom. This implies that unit 3 is much older than unit 4 (deformed by more displacement events), but how much older cannot be determined. We cannot measure its vertical displacement across the sliver zone, except to note that it must be >4.2 m, since unit 3 is not exposed on the HW. Mapped deposits in the F2-F1 sliver are even more deformed (east dips up to 80°), to the point it is unclear if they are stratigraphic units or shear zones. Overall, this is the pattern one would expect if all the faults displayed normal slip, and thus we infer that the stratigraphic section within each fault sliver becomes older toward the FW.

In general, the dip of the faults decreases toward the FW, being the lowest for the westernmost fault (F1), which places basal (?) Pantadeleon Formation against the Santa Fe Group. If faults F2–F6 are projected below the floor of the trench, they would all intersect fault F1 within 4 m of the trench floor. Thus, all the faults in the trench can be seen as upward splays off of fault F1.

The trench exposes a surface soil (developed in units 31–26) and six buried soils (paleosols) on the HW developed (from top to bottom) in units 25–24, 23, 22, 20, 9b, and 9a. Each paleosol is composed of multiple horizons, which are labeled on Figure 9. It is unclear if the complex paleosol in slivers F6 to F3 is the same as the soils developed in units 9a and 9b. The surface soil is developed in different deposits on various parts of the fault scarp, following the general pattern of soil development on normal fault scarps (McCalpin and Berry, 1996). On the FW, the upper horizons (A, Bwk1, Bwk2) are developed in unit 31, which is mainly a young (Holocene?) eolian sand blanket that lies unconformably on the Santa Fe Group gravels. The lower horizons (K1, K2) are developed on Santa Fe Group and constitute the relict soil on the LdA geomorphic surface that developed over many hundreds of thousands of years.

On the upper scarp face, the surface soil is developed in unit 31 and into the underlying colluvial wedges, but traced downslope, these wedges pinch out, and thus the lower part of the soil is developed in progressively older deposits (units 24 through 20) that approach the surface. These deposits had thin soils of their own beneath the colluvial wedge (buried soils 1–4), and the surface soil is thus progressively superimposed onto these soils as they approach the surface at the scarp toe. At the extreme eastern end of the trench, east of the antithetic fault, the surface soil is developed in unit 40, which directly overlies unit 12, with all intervening deposits having been pinched out or eroded away. Again, this soil catena is typical of normal fault scarps (McCalpin and Berry, 1996).

Buried soils in units 20–25 are all relatively weak, developed on eolian or transitional eolian/fluvial sands. These soils are typified by Bwk/Ck or Ck horizons, but never a true Bt horizon. They are inferred to represent short hiatuses in eolian deposition on the fault scarp. In contrast, buried soils developed in units 9a and 9b are stronger, with the 9b soil composed of thin Btk1/Btk2/Btk3 horizons, whereas the 9a soil is composed of thicker Bt/Btk/Ck horizons. These two buried soils reach an aggregate exposed thickness of ∼1 m on the HW, and they are possibly correlative with the uppermost buried soil in the fault zone (horizons 8Bk1 through 8Ck). The three lower buried soils (in the F6-F4 sliver, developed in units 7, 6, and 5) are less well developed, in that none of them has a true Bt horizon.

We calculated the mass of secondary (pedogenic) calcium carbonate in each soil horizon as the difference between total carbonate and assumed original carbonate, following standard methods (i.e., Machette, 1985; see Supplemental Material Appendix S1). The surface soil and buried soils 1–4 (defined in Figure 11) contain a total of 11.34 g of secondary carbonate per square centimeter column through the soil. We assumed that the 1.5 m thickness of units 10–12 contains no secondary carbonate because no signs of soil development were observed in these units. Buried soils 5 and 6 in units 9b and 9a contain an additional 10.69 g of secondary carbonate per square centimeter column, bringing the total to 22 g/cm² (Figure 11). Carbonate amounts in the older buried soils in fault slivers were not measured.

Two forms of age estimates are available from this trench: luminescence dating and soil carbonate accumulation. The luminescence age estimates are given in Table 2 and range from 57 ka to 416 ka. The sample numbering system has been corrected from that of the original Final Technical Report (McCalpin and Harrison, 2001). We emphasize IRSL dating of feldspar (7 of 9 samples) rather than optically stimulated luminescence (OSL) on quartz because the latter tends to saturate in deposits older than 100 ka (Zhang and Li, 2020), which we would expect in this 5-m-deep trench.

The two largest time jumps in our IRSL age series (Figure 9) are the 137 k.y. between units 26Bwk2 (57 ka) and 26 (194 ka) and the 76 k.y. between units 22 and 20b. Based on the County Dump and Calabacillas trenches (McCalpin et al., 2006, 2011), such long time jumps should be represented by ∼1-m-thick thick stage 4 K horizons (Machette, 1985). In our Zia trench, however, there is not even a sign of an unconformity at those two stratigraphic locations (Figure 9), much less a K horizon. The next three IRSL dates below (212, 252, and 328 ka) are also incompatible with an absence of any strong soils above the thick fluvial gravel deposits (unit 12). Thus, there is a strong disparity between the soil stratigraphy of the trench and the IRSL ages. Therefore, we calculated cumulative soil age as a reality check (below).

The mass of secondary carbonate in the soils can be converted into an age estimate if we assume a rate of carbonate accumulation. The calibrated rate derived from the 1999 LdA trenches on the Calabacillas Fault (McCalpin et al., 2011) is from the closest and most similar site to the Zia trench. At the Calabacillas trenches, a series of concordant OSL dates back to ca. 200 ka yielded a long-term rate of carbonate accumulation on the HW of 0.35 g/k.y. Applying this rate to the cumulative carbonate mass value in Figure 11 yielded an age estimate of 32.4 ka for the surface soil through buried soil 4. The same rate indicates an additional 30.5 k.y. of soil development time would be needed to form buried soils 5 and 6. Thus, the stacked paleosols in the middle trench would have required at least 62.9 kyr. However, erosion associated with the coarse fluvial gravels (units 10–12) has probably removed some deposits and soils between buried soils 4 and 5. Therefore, the top of buried soil 5 (unit 9bBtk1) is at least as old as the bottom of buried soil 4 (unit 20b, 32.4 ka), but it could be several tens of thousands of years older. The age estimates cited in Figure 11 are thus minima for buried soils 5 and 6.

The sequence of faulting, deposition, erosion, and soil formation was reconstructed by retro-deforming the trench log (Supplemental Material Appendix S2). We used the simplified-incomplete method of McCalpin (2009, p. 260). We defined five event horizons (V [oldest] through Z [youngest]) based on upward terminations of faults (e.g., faults F8–F11 on Figures 7 and 9), angular unconformities in HW strata, or by downdip displacement increases on a fault. Table 3 lists fault throw across five event horizons by faults F7 (zone of five faults), F6, and F4. Throw on F7 does not increase with depth, indicating these “overhanging” faults developed in the most recent event. On faults F6 and F4, throw increases with depth (age) of the event horizon, culminating with 5.1 m total throw across event horizon V. This is larger than the 4.0 m of vertical separation across the fault scarp, a result of both dip-slip displacement and additional rotation of the HW toward the FW. Prior to event X, we have indirect evidence via unconformities W and V that one or two faulting events occurred during the deposition of unit 12. The earliest event (V) is the one that tilted the top of unit 9 toward the fault and led to the initial diversion of the unit 10 stream up against the fault scarp.

Event Z occurred after the formation of buried soil 1 (units 25Ckb1 to 24a) and prior to the deposition of the overlying, unfaulted colluvial wedge of unit 26. That isochron (unconformity Z) is represented by the contact between units 26 and 25Ckb1, which is dated by soil carbonate at ca. 10 ka and by IRSL at between 194 ka (sample ZIA-6) and 212 ka (sample ZIA-5) (see Figure 9). For reasons described previously, we favor the soil carbonate date as more accurate. Thus, event Z resulted in a net 0.9 m of throw on fault zones F7 and F6b, and possibly some unmeasurable throw on faults F1–F5, at about 10 ka.

Event Y occurred after the formation of buried soil 2 (units 23BCkb2 to 22) and prior to the deposition of the penultimate colluvial wedge of units 24–25. That isochron (unconformity Y) is represented by the contact between units 24 and 23BCkb2, which is dated by soil carbonate at ca. 21 ka and by IRSL at between 212 ka (sample ZIA-5) and 252 ka (sample ZIA4) (see Figure 9). For reasons described previously, we favor the soil carbonate date as more accurate. Thus, event Y resulted in a net 1.9 m of throw on fault zones F7 and F6b, and at least 0.2 m of throw on fault F4, at about 21 ka.

The displacement analyses of Table 3 and Supplemental Material Appendix S2 suggest that fault displacement has migrated toward the HW with time, which implies that earlier faulting events (X, W, V) mainly caused displacement on faults closer to the FW (F1 through F5).

When our suite of IRSL ages turned out to be discordant with nearby trench sites, we turned to carbonate accumulation as a reality check (Figure 12). Compared to the degree of soil development in three other trenches in the Rio Grande Rift (East Franklin Mountains [McCalpin et al., 2021]; County Dump [McCalpin et al., 2006]; and Calabacillas, [McCalpin et al., 2011]), the luminescence ages from the Zia trench are extremely telescoped. Normally, IRSL ages of 200 ka or more are found only at depths of 5–8 m below the surface, and beneath strong K horizon soils (caliche or calcrete). At the Zia trench, such ages came from less than 1 m below the surface, with no K horizon development. There are two likely explanations. First, the Zia samples may have been contaminated with clastic particles eroded from the FW relict soil profile, which does indeed have K horizons (Figure 7). Second, subsequent to our luminescence dating, IRSL dating fell into disfavor due to age overestimation after correction for anomalous fading (see references in Zhang and Li, 2020). A replacement method to avoid overestimation (post-infrared IRSL dating) was later developed (see Ishii, 2024). The inherent overestimation of IRSL may help to explain why the Zia luminescence ages are much older than carbonate accumulation ages for the same units.

Carbonate accumulation ages suffer from two major limitations. First, they only yield time required for soil formation; they do not include any time estimate for the deposition of sediments. Second, if there has been any erosion of sediments or soils from the HW, that time will also not be represented. So, we need to estimate how past erosion of parts of the HW sequence (by gravel units 10–12) has affected the age estimate for older deposits (unit 9). One assumption is that the non-calcareous units 10–12 may have eroded an equal thickness of preexisting eolian and colluvial sediments. Those sediments would have contained secondary soil carbonate, which is now gone and not accounted for in Figure 11. Figure 13 shows graphically how to “restore” the missing carbonate that may have been in the eroded units, so we can make a better estimate of the age of unit 9 (the oldest deposits in the trench). Assuming the stream may have eroded a thickness equal to units 10–12, that eroded material may have contained a low carbonate percent similar to units 20–30 or a high carbonate percent like the unit 9 paleosols. The former assumption would add 13 k.y. to the 63 ka age of our oldest HW unit (9aCk), while the latter assumption would add 60 k.y. The midpoint of these estimates is 100 ka for the bottom of the 5.3-m-thick HW sequence. This compares with 82 ka at that same depth in the County Dump trench, and ca. 200 ka at the Calabacillas trench.

A wide range of slip rates can be derived from observed displacements and carbonate accumulation age estimates:

• (1) Two short-term closed-cycle slip rates can be computed. Between event X and event Y, 2.6 m of strain accumulated and was released, yielding a slip rate of 0.22 mm/yr. Between event Y and event Z, 0.9 m of strain accumulated and was released, yielding a slip rate of 0.07 mm/yr.

• (2) A longer-term slip rate includes the two cycles cited above plus the open cycles before event X and after event Z. Total throw of >5.1 m over the past 76 k.y. to 123 k.y. yields an apparent slip rate of 0.041 to 0.067 mm/yr.

• (3) A million-year-long slip rate can be derived from Figure 3 from a throw of 70 m on the base of the lower Pleistocene–Pliocene(?) Ceja Formation, which has an age of ∼3–4 Ma. This yields slip rates of 0.023 to 0.035 mm/yr.

• (4) Finally, a post-rifting slip rate can be derived from the total Neogene throw of 350 m on the top of the pre-rift Diamond Tail and Galisteo Formations, presumably deposited as early as the mid-Miocene, ca. 14 Ma. This yields a slip rate of 0.025 mm/yr.

The present U.S. NSHM calculates fault slip rates by averaging the published geologic slip rate with four estimates of geodetic slip rates (Hatem et al., 2022). However, the geodetic slip rates must fit within the uncertainty range of the geologic slip rates. For faults with small variance in geologic slip rates, the geodetic rates are more tightly constrained by the geologic rates. However, for faults with large variance in geologic slip rates, the geodetic slip rates are free to vary within a wider range and may dominate the composite slip rate used by the 2023 NSHM. Thus, it is in the interest of earthquake geologists to minimize the range of slip rate estimates, if possible. The first task would be to refine slip rates by minimizing measurement errors in displacement amounts and ages; these are epistemic uncertainties that can be reduced by additional data collection and analysis (such as Figure 13). Conversely, apparent variations in slip rates through time might reflect real variations among a fault’s past seismic cycles, a property intrinsic to the fault itself (or to Coulomb stress changes from neighboring faults). For example, our trench lies in the overlap between the Albuquerque-Belen Basin and the Santo Domingo Basin. The latter basin is notorious for its “seesaw subsidence” in which the master fault switches from the east side to the west side (Smith et al., 2001). Since the middle Pleistocene, fault activity has switched to the west side, where our Zia trench lies. Such aleatory variability cannot be reduced by more or better dating.

The Zia Fault at the northern edge of the Llano de Albuquerque has formed a narrow, steep fault scarp with an adjacent graben. The middle trench across the steepest, highest part of the scarp revealed evidence for four to five faulting events. Displacement in the latest three events ranged from 0.9 to 2.6 m, with an average of 1.8 m.

We also estimated the ages of the surface soil and six buried soils by calculating the time required to accumulate the measured mass of secondary calcium carbonate. Our preferred accumulation rate is 0.35 g/k.y., based on calibration results from our 1996 County Dump Fault trenches (McCalpin et al., 2006) and 1999 Calabacillas Fault trenches (McCalpin et al., 2011). Soil carbonate age estimates suggest that soils in the upper 3 m of the trench required 32 k.y. to form, and the strong soils at the base of the trench required an additional 30.5 k.y. However, there are several unconformities between these soil sections, so almost surely some soils and deposits were removed by erosion in the sequence. Thus, the age estimates of the lower soils are minima. One model of accounting for soil carbonate removed from HW strata by fluvial erosion suggests that the lowest soil profile may be 76–123 ka.

According to Wells and Coppersmith (1994), historic normal fault earthquakes that have created fault scarps with an average height of 0.9–2.6 m range in magnitude from M 6.75 to M 7.06. The average displacement of the five inferred events in the middle trench (1.8 m) implies a magnitude of M 6.95. This magnitude is very close to that associated with a 32-km-long fault (Machette et al., 1998).

Due to the conflict between the IRSL and soil carbonate dates, and uncertainties about erosion of HW deposits, any conclusions about recurrence interval must be tentative. If one believes the soil carbonate dates from the surface soil and younger buried soils are accurate, then the recurrence interval between events Z and Y was 12 k.y. The ages of the older event horizons (X, W, and V), and thus the length of previous recurrence cycles, cannot be estimated from soil carbonate due to erosion and missing section. However, a long-term average recurrence can be estimated if we assume that the inferred five events have occurred since the formation of buried soils 5 and 6, which have a soil carbonate date of ca. 76 to 123 ka. If the latest event occurred at 12 ka, then the four preceding recurrence cycles spanned 65–112 k.y., which yields an average recurrence interval of 16–28 k.y.

The four-to-five events inferred from the middle trench have a cumulative down-to-the-east throw of ∼5.1 m. If that throw accumulated over the past 76–123 k.y., then the average long-term slip rate is 0.041–0.067 mm/yr. This slip rate is greater than that inferred for the two other major faults on the LdA, the County Dump Fault (0.016–0.018 mm/yr) and the Calabacillas Fault (0.0053–0.0072 mm/yr), but this may result from additional Coulomb strain coming from the Santo Domingo Basin.

This trenching study would not have been possible without landowner permission granted by The State of New Mexico Land Board. Trenches were logged by J. P. McCalpin, L. C. Allen Jones, Deborah J. Green, and Nicole Bailey. Bruce Harrison and Nelia Dunbar (New Mexico Tech) respectively defined the soil profiles and performed carbonate analyses, and examined volcanic ash (?) from the middle trench. Dr. Glenn Berger (Desert Research Institute, Reno, NV) collected the luminescence samples and dated them. John Hawley blessed the trenches with his presence. Sean Connell (New Mexico Bureau of Mines & Geology) put up with our bothersome presence at the bureau’s Albuquerque office. The trenches were excavated and backfilled by Mark’s Backhoe Service (Corrales, NM), and shoring was provided by Southwest Safety Services (Albuquerque, NM). Comments by three anonymous reviewers improved the manuscript.

Supplemental material associated with this article can be found online at https://www.aegweb.org/e-eg-supplements.