From Pliocene to middle Pleistocene time, a large lake occupied most of the San Luis Valley above 2300 m elevation (7550 ft) in southern Colorado. This ancient lake accumulated sediments of the Alamosa Formation (Siebenthal, 1910), for which the lake is herein named. The existence of this lake was first postulated in 1822 and proven in 1910 from well logs. At its maximum extent of nearly 4000 km 2 , it was one of the largest high-altitude lakes in North America, similar to but larger than Lake Texcoco in the Valley of Mexico. Lake Alamosa persisted for ~3 m.y., expanding and contracting and filling the valley with sediment until ca. 430 ka, when it overtopped a low sill and cut a deep gorge through Oligocene volcanic rocks in the San Luis Hills and drained to the south. As the lake drained, nearly 100 km 3 (81 × 10 6 acre-ft or more) of water coursed southward and flowed into the Rio Grande, entering at what is now the mouth of the Red River. The key to this new interpretation is the discovery of ancient shoreline deposits, including spits, barrier bars, and lagoon deposits nestled among bays and in backwater positions on the northern margin of the San Luis Hills, southeast of Alamosa, Colorado. Alluvial and lacustrine sediment nearly filled the basin prior to the lake's overflow, which occurred ca. 430 ka as estimated from 3 He surface-exposure ages of 431 ± 6 ka and 439 ± 6 ka on a shoreline basalt boulder, and from strongly developed relict calcic soils on barrier bars and spits at 2330–2340 m (7645–7676 ft), which is the lake's highest shoreline elevation. Overtopping of the lake's hydrologic sill was probably driven by high lake levels at the close of marine oxygen-isotope stage (OIS) 12 (452–427 ka), one of the most extensive middle Pleistocene glacial episodes on the North American continent. Hydrologic modeling of stream inflow during full-glacial-maximum conditions suggests that Lake Alamosa could fill at modern precipitation amounts if the mean annual temperature were just 5 °C (10 °F) cooler, or could fill at modern temperatures with 1.5 times current mean annual precipitation. Thus, during pluvial epochs the lake would rise to successively higher levels owing to sedimentation; finally during OIS 12, the lake overflowed and spilled to the south. The integration of the upper (Colorado) and lower (New Mexico) reaches of the Rio Grande expanded the river's drainage basin by nearly 18,000 km 2 and added recharge areas in the high-altitude, glaciated San Juan Mountains, southern Sawatch Range, and northern Sangre de Cristo Mountains. This large increase in mountainous drainage influenced the river's dynamics downstream in New Mexico through down-cutting and lowering of water tables in the southern part of the San Luis Valley.

Volcanic clasts incorporated in the lower portion of the Tertiary Santa Fe Group sedimentary rocks of the Culebra graben, San Luis Basin, Colorado, provide constraints on the timing of regional tectonic events by provenance determination. Based on currently exposed volcanic terrains, possible clast sources include Spanish Peaks and Mount Mestas to the east, the San Juan volcanic field to the west, and the Thirtynine Mile volcanic field, a remnant of the Central Colorado volcanic field, to the north and east of the San Luis Basin. Provenance was determined by a variety of geochemical, mineral chemical, and geochronologic data. Large porphyritic Santa Fe Group volcanic clasts are potassic with a wide compositional range from potassic trachybasalt to rhyolite. The whole-rock chemistry of the Culebra graben clasts is similar to that of the Thirtynine Mile and San Juan volcanic fields. Culebra graben amphibole and biotite chemistry is generally consistent with that of rocks of the San Juan volcanic field, but not with Spanish Peaks samples. Trace-element data of Culebra graben volcanic clasts overlap with those of the San Juan and Thirtynine Mile volcanic fields, but differ from those of the Mount Mestas. Thermobarometric calculations using mineral chemistry suggest that many Culebra graben rocks underwent a three-stage crystallization history: ~1120 °C at 7–10 kbar, ~1100 °C at 2.3–4.6 kbar, and hornblende formation ~800 °C at 3 kbar. Within the Culebra graben clasts, zircon rim U-Pb geochronologic systematics as well as amphibole and biotite 40 Ar/ 39 Ar plateau data yield ages ranging from 36 to 29 Ma. These ages are consistent with ages of the Thirtynine Mile volcanic field (36–27 Ma) and the Conejos Formation of the San Juan volcanic field (35–29 Ma), but predate Spanish Peaks (ca. 27–21 Ma) and Mount Mestas (ca. 25 Ma). Based on these data, Spanish Peaks and Mount Mestas are excluded as potential source areas for the Santa Fe Group volcanic clasts in the Culebra graben. The San Juan volcanic field is also an unlikely source due to the distance from the depositional site, the inconsistent paleo-current directions, and the pressure-temperature conditions of the rocks. The most likely scenario is that the Central Colorado volcanic field originally extended proximal to the current location of the Culebra graben and local delivery of volcanic clasts was from the north and northeast prior to the uplift of the Culebra Range and Sangre de Cristo Mountains.

The Sunshine Valley–Costilla Plain, a structural subbasin of the greater San Luis Basin of the northern Rio Grande rift, is bounded to the north and south by the San Luis Hills and the Red River fault zone, respectively. Surficial mapping, neotectonic investigations, geochronology, and geophysics demonstrate that the structural, volcanic, and geomorphic evolution of the basin involves the intermingling of climatic cycles and spatially and temporally varying tectonic activity of the Rio Grande rift system. Tectonic activity has transferred between range-bounding and intrabasin faults creating relict landforms of higher tectonic-activity rates along the mountain-piedmont junction. Pliocene–Pleistocene average long-term slip rates along the southern Sangre de Cristo fault zone range between 0.1 and 0.2 mm/year with late Pleistocene slip rates approximately half (0.06 mm/year) of the longer Quaternary slip rate. During the late Pleistocene, climatic influences have been dominant over tectonic influences on mountain-front geomorphic processes. Geomorphic evidence suggests that this once-closed subbasin was integrated into the Rio Grande prior to the integration of the once-closed northern San Luis Basin, north of the San Luis Hills, Colorado; however, deep canyon incision, north of the Red River and south of the San Luis Hills, initiated relatively coeval to the integration of the northern San Luis Basin. Long-term projections of slip rates applied to a 1.6 km basin depth defined from geophysical modeling suggests that rifting initiated within this subbasin between 20 and 10 Ma. Geologic mapping and geophysical interpretations reveal a complex network of northwest-, northeast-, and north-south–trending faults. Northwest- and northeast-trending faults show dual polarity and are crosscut by north-south– trending faults. This structural model possibly provides an analog for how some intracontinental rift structures evolve through time.

Interpretation of gravity, aeromagnetic, and magnetotelluric (MT) data reveals patterns of rifting, rift-sediment thicknesses, distribution of pre-rift volcanic and sedimentary rocks, and distribution of syn-rift volcanic rocks in the central San Luis Basin, one of the northernmost major basins that make up the Rio Grande rift. Rift-sediment thicknesses for the central San Luis Basin determined from a three-dimensional gravity inversion indicate that syn-rift Santa Fe Group sediments have a maximum thickness of ~2 km in the Sanchez graben near the eastern margin of the basin along the central Sangre de Cristo fault zone, and reach nearly 1 km within the Monte Vista graben near the western basin margin along the San Juan Mountains. In between, Santa Fe Group thickness is negligible under the San Luis Hills and estimated to reach ~1.1 km under the Costilla Plains (although no independent thickness constraints exist, and a range of thicknesses of 600 m to 2 km is geophysically reasonable). From combined geophysical and geologic considerations, pre-rift, dominantly sedimentary rocks appear to increase in thickness from none in the Sanchez graben on the east to perhaps 800 m under the San Luis Hills on the west. The pre-rift rocks are most likely early Tertiary in age, but the presence of Mesozoic and Paleozoic sedimentary rocks cannot be ruled out. Geophysical data provide new evidence that an isolated exposure of Proterozoic rocks on San Pedro Mesa is rooted in the Precambrian basement. This narrow, north-south–trending basement high has ~2 km of positive relief with respect to the base of the Sanchez graben, and separates the graben from the structural depression beneath the Costilla Plains. A structural high composed of pre-rift rocks, long inferred to extend from under the San Luis Hills to the Taos Plateau, is confirmed and found to be denser than previously believed, with little or no overlying Santa Fe Group sediments. Major faults in the study area are delineated by geophysical data and models; these faults include significant vertical offsets (≥1 km) of Precambrian rocks along the central and southern zones of the Sangre de Cristo fault system. Other faults with similarly large offsets of the Santa Fe Group include a fault bounding the western margin of San Pedro Mesa, and other faults that bound the Monte Vista graben in an area previously assumed to be a simple hinge zone at the western edge of the San Luis Basin. A major north-south–trending structure with expression in gravity and MT data occurs at the boundary between the Costilla Plains and the San Luis Hills structural high. Although it has been interpreted as a down-to-the-east normal fault or fault zone, our modeling suggests that it also is likely related to pre-rift tectonics. Aeromagnetic anomalies over much of the area are interpreted to mainly reflect variations of remanent magnetic polarity and burial depth of the 5.3–3.7 Ma Servilleta Basalt of the Taos Plateau volcanic field. Magnetic-source depth estimates are interpreted to indicate patterns of subsidence following eruption of the basalt, with maximum subsidence in the Sanchez graben.

The late Cenozoic extension in the Rio Grande rift of north-central New Mexico was predominantly accommodated by the north-south–trending Pajarito and Sangre de Cristo normal faults and the intervening east-northeast–striking predominantly strike-slip Embudo fault. Using this segment of the rift as our primary example, we have analyzed a series of three-dimensional nonlinear elastic-plastic finite-element models to assess the role of mechanical interactions between pairs of en echelon rift-scale listric normal faults in the evolution of intervening relay zones. The model results demonstrate that under orthogonal extension and an overall plane-strain deformation, relay zones may evolve in a three-dimensional strain field and along non-coaxial strain paths. The extent of non-plane strain and non-coaxial deformation depends on the fault overlap to spacing ratio, the relative orientations of the bounding faults, and the structural position within the relay zone. The model-derived minimum compressive stress vectors within the relay zone are oblique to the regional extension direction throughout the deformation. Within the Rio Grande rift of north-central New Mexico, the occurrence of northerly striking Neogene faults suggestive of east-west extension in the Española and the San Luis Basins, geographic variations in the vertical-axis rotations from paleomagnetic studies, and secondary fault patterns are consistent with the near-surface variations in the strain field predicted by the model. The model suggests that interaction between the Pajarito and the Sangre de Cristo faults may have played a major role in the evolution of this segment of the rift.

The deepest part of the northern Rio Grande rift is just northwest of the Great Sand Dunes, on the eastern side of the San Luis Valley, in south-central Colorado. Approximately 150 km (95 m) of high-quality, 20-fold Common Depth Point seismic data indicate that the basin is filled with approximately 6.4 km (21,000 ft) of Tertiary sediments, mostly Oligocene or younger in age. These data, combined with published data, provide a basis for interpreting the structural geometry of the rift. The internal structural geometry of this part of the rift is surprisingly uncomplicated. Sympathetic and antithetic faulting is not widespread, or major, within the deep part of the basin. The internal geometry of the sedimentary packages, however, indicates a complicated movement history on the faults that are present within the rift. The bounding fault zone between the rift and the Sangre de Cristo Mountains has approximately 9.2 km (30,000 ft) of vertical separation and may be very complicated. Although we were not able to resolve the bounding fault zone with our seismic data, the data constrains the minimum angle of the fault zone to be approximately 45°. Our modeling suggests that a 60° angle is the most likely orientation of the fault zone. The influence of earlier, low-angle extension recognized nearby and elsewhere in the Rio Grande rift is not directly evident in our data. Assuming 60° antithetic shear in the hanging wall, which is supported by the seismic data, a depth of detachment (flattening) is estimated to be approximately 16 km, in the brittle-ductile transition zone estimated from heat-flow data. Cross sections drawn using the constraints of the new seismic data, require only 8 to 12% extension across the rift at this latitude.