The mountain peaks of the present-day Southern Rocky Mountains are the highest peaks in the Rocky Mountain system. They represent a second generation of mountains, one that originated from a different tectonic mechanism from that of the predecessor Laramide Rockies. Epeirogeny lifted the Laramide ranges in Colorado and New Mexico after their Late Cretaceous–early Cenozoic orogenic creation. The area was lifted tectonically from 1300 m to perhaps as much as 2000 m, the result of heating of the lithosphere stemming from its thinning, as well as inflation of the crust by the intrusion of extensive, relatively low density batholiths and plutons of middle Tertiary age. This uplift produced an elongate north-striking crustal swell that cuts across major structural features in the crust, including the northeast-trending fundamental sutures that resulted from assembly of the North American plate, the northwest-striking trends of the Ancestral Rocky Mountains in Colorado, and the northeast-trending Colorado Mineral Belt. The contemporary Southern Rockies are unique in that their eastern piedmont slope is quite unlike that of other prominent orogenic mountain ranges around the globe owing to the presence of this supporting swell, or epeirogen. The lithosphere beneath the epeirogen's summit is characterized by a coincident geoid anomaly, diminished seismic velocities in the upper mantle, and a north-trending, elevated Curie isothermal surface in the lower crust, all suggestive of elevated temperatures. Surface heat flow on the summit is complex, revealing both shallow crustal heat sources and a much deeper, more profound source that strikes north. Uplift resulting from these factors was initiated in post–middle Eocene time. At the wavelength of topographic smoothing employed here, the epeirogen's regional topography makes it the highest general feature on the North American plate, individual mountain peak elevations, here and elsewhere, excepted. A first-order, re-leveled survey line in southern Colorado suggests that the epeirogen is still rising today.


This paper revisits observations made more than two decades ago (Eaton, 1986, 1987) in light of new data and new paradigms acquired in the past decade and a half. Some of the interpretations originally made are substantiated, some were incorrect and have required discarding, and some are new.

Morgan and Swanberg (1985) and Morgan (2003) investigated the subject of uplift of the Colorado Plateau and southern Rocky Mountains, utilizing some of the same observations and interpretations that are referenced here. They proposed three possible mechanisms as the cause of uplift: (1) phase changes in the lithosphere, (2) physical thickening of the crust, and (3) thermal expansion and thinning of the lithosphere. In examining each of these in rigorous detail, Morgan (2003) chose uplift associated with phase changes in the lithosphere as the preferred model. In this paper I emphasize lithospheric thinning and thermal expansion as the preferred mechanism, though some contribution from phase changes in the upper mantle cannot be ruled out entirely. There is broad agreement between us on certain of the issues involved here, but differences on others, and the interested reader is urged to consult Morgan's (2003) paper in conjunction with this one.

Topography and Elevation

Individual peaks in the Southern Rocky Mountains of Colorado and New Mexico, i.e., 1140 mountains with elevations >3048 m, ~200 with elevations >3960 m, and 54 with elevations >4270 m, are the highest mountains along the crest of the Rocky Mountain system, a complex chain of differing age and deformational style that extends from northern British Columbia to New Mexico. Figure 1 shows the topography of the southern Rocky Mountains region in shaded relief. It consists of two additive components: (1) the topography of the mountains, and (2) outward from the mountains, the topography of a giant swell that lifted the mountains and upon which they sit today.

In the Middle Rocky Mountains and the southeastern part of the Northern Rocky Mountains (the physiographic province names are those of Fenneman, 1931), there are fewer than 24 Rocky Mountain peaks that exceed 3660 m and only a few that exceed 3960 m; most, if not all of the latter, are associated with thermotectonic uplift related to the subsurface presence of the Yellowstone hotspot of Pierce and Morgan (1992). North of the Yellowstone hotspot, however, all the way to northernmost British Columbia, there are no peaks that achieve elevations of 3960 m.

There are two other regions on the North American plate where individual mountain peaks have elevations >4270 m. In California, principally in the Sierra Nevada and White Mountains, there are 12 peaks higher than 4270 m; the elevation of the highest (Mount Whitney) exceeds that of the highest peak in the southern Rockies (Mount Elbert, Colorado) by only 72 m. Even in Alaska, with its many high mountain ranges created by repeated docking of exotic terranes, there are only 14 peaks with elevations >4270 m. Of the latter, however, 7 exceed 4572 m, including Mount McKinley (Denali), with an elevation of 6194 m.

The number of peaks having elevations >4270 m is thus much larger in the southern Rocky Mountains than in the other two regions (54 versus12 versus14), but there are other contrasts between these mountain regions, including differences in surface heat flow, the number of thermal springs, and the regional elevation.

The Alaska Range and Wrangell Mountains and much of their surrounding terrain appear to have uniform surface heat flow values of 80–85 mW/m2 (Blackwell and Richards, 2004; see http://www.smu.edu/eothermal/2004NAMap/2004NAMap.htm.) However, observations are sparse and very widely separated; hence, lateral extrapolations of heat flow values are tenuous at best. Surface heat flow in the Sierras ranges from 30 to 50 mW/m2 and in the Southern Rocky Mountains ranges from 70 to as much as 150 mW/m2 (Decker et al., 1988). Differences in the number and geographic density of thermal springs, evidence of a relatively warm and shallow, fractured crust, are as follows. Alaska has 108 known thermal springs, but they are scattered all over the state (see National Geophysical Data Center listings for Alaska); those in the California region of the Sierra Nevada and White Mountains number ~20 (Waring, 1965), but they occur largely at the east base of these mountains where they adjoin the Basin and Range Province, not on the summits of the ranges; and there are 84 in the Southern Rocky Mountains, where thermal springs occur largely within the range. There are also differences in the elevation of the regional topography from which the peaks rise, as demonstrated here. All these differences set the Southern Rocky Mountains region apart as anomalous.


The Laramide Southern Rocky Mountains underwent severe erosion both during and following the period in which they were created orogenically. The entire Paleozoic and pre-Laramide Mesozoic stratigraphic section, ~1.5–3 km thick, and a not-insignificant part of the uppermost Precambrian basement, was eroded away from the arched basement block uplifts. The resulting high-standing Rocky Mountain erosion surface (Epis and Chapin, 1975; Coleman, 1985; Chapin and Kelley, 1997; Steven et al., 1997; Kelley and Chapin, 2004) has been described and mapped in regional detail. It is overlain locally by volcanic rocks of latest Eocene and Oligocene age. Rounded hills and low rounded mountains are present. Kelley and Chapin (2004) argued that Pikes Peak is a Laramide monadnock on this surface.

Steven et al. (1997) subdivided the erosion surface geographically and described different parts of it in terms of their history of Eocene and Miocene incisional stream drainage. One of the subdivisions drained eastward and south-eastward in late Eocene time, the other drained eastward and northeastward in Miocene time. This evidence is significant. It indicates that the epeirogen was tectonically active. Kelley and Chapin (2004) utilized apatite fission track data and identification of partial annealing zones to determine the degree of denudation of the Laramide southern Rockies. They found that depths of denudation ranged from 1.4 km to as much as 3.85 km. The removal of slabs of rock of these thicknesses would clearly result in epeirogenic uplift, but not as much as today.

The western Great Plains immediately east of the Southern Rockies in southern Wyoming and all of eastern Colorado and New Mexico stand higher at the foot of the mountains than they do both to the north and south. In this part of the high plains, a postorogenic sedimentary cover (the late Eocene–early Oligocene White River; late Oligocene–early Miocene Arikaree; and middle Miocene–early Pliocene Ogallala Formation) contributes locally to higher plains elevations. The geographic distribution of these formations can be viewed on the maps of King and Beikman (1974), Tweto (1979a), Love and Christiansen (1985), and the New Mexico Bureau of Geology and Mineral Resources (2003).

Anomalous Eastern Piedmont Slope of the Southern Rocky Mountains

Figure 2 provides a comparison of the modern eastern piedmont slope of the Southern Rocky Mountains (curve A) with the piedmont slopes of other global mountain ranges (curves B–I). The Southern Rocky Mountains profile was drawn along the interfluve between the Arkansas and Canadian Rivers in southern Colorado. In contrast with an earlier approach (Eaton, 1986, 1987), the profiles here were drawn from the foot of the mountains, rather than from their crest, out to a distance of 1000 km. Elevations were sampled at 25, 50, 100, 200, 300, 400, 500, 750, and 1000 km from the mountain base of each range. The profiles were arbitrarily assigned an elevation value of zero at a distance of 1000 km in order to provide interrange comparison. Actual elevations at the bases of individual ranges are listed in Table 1, along with information regarding the geographic end points of each source profile, the full lengths of which have not been included in Figure 2. They fall into two separate elevation classes, those with mountain base elevations of <600 m and those with mountain base elevations ≥1000 m.

The piedmont slopes of all the ranges illustrated in Figure 2 are concave upward and each has its own age of initial uplift and plate tectonic origin. Every elevated mountain range undergoes the fate of general erosion and, especially, erosional incision, though the rates and degree of uplift of its peaks vary from range to range. What differentiates the Southern Rocky Mountains profile from the others is the much higher elevation at its mountain base. As its elevation declines to the east, this profile has much greater height than that of the others, to a distance of 600–700 km from the mountain front. The difference in mountain base elevation between the Southern Rocky Mountains and the Atlas Mountains, a range whose piedmont slope extends nearly as far from the mountains as does that of the Southern Rocky Mountains, is 1260 m, or nearly 1300 m.

In comparing these profiles it is apparent that profile A is not typical of the piedmont slope of an orogenic mountain range. It is the eastern limb of a mammoth swell or epeirogen of tectonic origin. The focus of this paper is on this epeirogen, uplift of which took place chiefly in post–middle Eocene time and later.

I (Eaton, 1986) referred to this feature as the “Alvarado Ridge” because the topography was essentially identical to that of a spreading ocean ridge. It is here renamed the Southern Rocky Mountain epeirogen because of its gross three-dimensional morphology. The epeirogen is slightly >2750 m high at its highest point and ~2000 km wide at its base. The Southern Rocky Mountains thus have less than one-tenth the width of the epeirogen. The highest mountain peaks, as measured from the top of the epeirogen, have elevations that are lower than the summit of the epeirogen (and its eroded Laramide mountain core) relative to sea level. I (Eaton, 1987) proposed that the epeirogen was created by lithospheric thinning, and Roy et al. (1999) modeled the effect of high, but laterally diminishing, temperatures at the base of the crust. These interpretations have been modeled quantitatively (Cordell et al., 1991; Roy et al., 1999, Fig. 3302303304 therein). In addition to their proposed basal crustal temperature distribution, Roy et al. (1999) also considered, but rejected, for analytical reasons, two other mechanisms: (1) long-wavelength flexural uplift owing to extension in the Rio Grande rift, and (2) thermal buoyancy and passive asthenospheric upwelling driven by extension along the Rio Grande rift.

Roy et al. (2004) suggested the possibility of a wholly different interpretation: the uplift resulted from significant mantle dedensification and thermal expansion during a massive mid-Tertiary magmatic invasion of the crust. These mechanisms may very well have made a significant contribution to the high regional elevation.

A variety of different kinds of observations led to a confirmation of earlier interpretations, extending back more than a half century, to the effect that the Southern Rocky Mountains underwent epeirogenic uplift following the Laramide orogeny (e.g., see Eardley, 1951). This is a geologic certainty. Many investigators have addressed various aspects of this uplift: Cordell (1978), Morgan and Swanberg (1985), Eaton (1987), Roy et al. (1999), McMillan et al. (2002), Sahagian et al. (2002), Morgan (2003), and Roy et al. (2004). Using paleoaltimetry, Sahagian et al. (2002) and Sahagian and Proussevitch (2007) documented what they referred to as “the rise of the Colorado Plateau” during the past 23 m.y., but a number of their collecting sites immediately adjoin or are actually located within the Southern Rocky Mountains, and some are even located in the Rio Grande rift in the interior of the range; hence, both of these elements were lifted, not the Colorado Plateau alone. If their interpretation is valid (discussed here later), it is a key piece of evidence that there was uplift of the epeirogen in Miocene time.

There is no latitudinally continuous, crust-penetrating, active young fault system separating and, therefore, mechanically isolating, the Colorado Plateau from today's Southern Rocky Mountains. As there is likewise no through-going, active, young regional fault system that mechanically separates today's Great Plains from the Southern Rocky Mountains, it can be argued that regional uplift has affected all three provinces simultaneously. As first applied to the Colorado Plateau (Morgan and Swanberg, 1985), it is here applied to the Great Plains.

Prior to the Laramide orogeny, all three of these provinces were part of the North American craton and constituted an essentially laterally uninterrupted lithospheric continuum close to sea level. The crust extended from Utah across both what is now the Colorado Plateau and the Southern Rocky Mountains and continued eastward beneath the Great Plains and Central Lowland. One cannot postulate post-Laramide uplift of one of these provinces without addressing simultaneous uplift of the other two.

The great eastern escarpment of today's Southern Rocky Mountains is both a Laramide structural escarpment and a gigantic east-facing monocline with an east-verging Laramide thrust fault zone at depth, but it is also a later erosional escarpment, where post-Laramide continental sedimentary rocks that unconformably lay across the older, upturned rocks of the structural escarpment have since been stripped away. As steep and abrupt as the eastern escarpment of the southern Rocky Mountains is almost everywhere, it is not a giant fault scarp in the usual sense (Chapin and Cather, 1994; Leonard and Langford, 1994; Kelley and Chapin, 2004).

The younger continental postorogenic Cenozoic sedimentary cover on the Great Plains onlaps the Precambrian basement in southern Wyoming at an elevation of ~2300 m and is preserved within the mountains of central Colorado as scattered remnants overlying the late Eocene erosion surface at an elevation of ~2800 m. A fundamental question is, do differences in elevation of parts of these sedimentary rocks represent deformation of the Ogallala Formation, or do they simply represent an erosionally interrupted, but once continuous, depositional surface on which post-Laramide sediments draped themselves over a then-existing topography? I lean ambivalently toward both answers. The Sahagian et al. (2002) and Sahagian and Proussevitch (2007) data clearly point toward deformation.

Significance of the Evidence of Simultaneous Uplift of the Colorado Plateau and Rocky Mountains

From as far back in North American geologic thought as the work of Dutton (1882), much has been made of the high elevation of the Colorado Plateau relative to sea level and to the general elevation the Basin and Range Province, but little, if anything, seems to have been made of the uniquely high elevation of the western Great Plains, at the foot of today's Southern Rocky Mountains opposite the Colorado Plateau.

The average thickness of the Colorado Plateau crust (40–45 km), even given its variability (Parsons et al., 1996), more or less matches that of much of the rest of the North American craton both east and northwest of the Southern Rocky Mountains. Crustal thickness is generally similar in eastern Washington and Oregon, Idaho, North Dakota, South Dakota, Nebraska, Kansas, Iowa, and Missouri, plus most of Minnesota (Braile et al., 1989, Fig. 4402 therein).

Gao et al. (2004) argued that the thickness of the crust along a northwest-trending, passive seismic line (the LA RISTRA line; Colorado Plateau–Rio Grande rift–Great Plains Seismic Transect) that sampled half of the Colorado Plateau in two dimensions only (x and z), is 45.6 km, making it thicker than the crust of the Great Plains, which they determined to be 44.1 km thick along their transect. Gao et al. (2004) suggested that the extra degree of isostatic buoyancy that would follow from the somewhat greater plateau crustal thickness should result in ~300 m of additional elevation of the plateau, leading to some of the observed asymmetry of the epeirogen. They further determined that the crustal thickness of the Rio Grande Rift is notably less than that of both the plateau and plains, i.e., 35 km.

Gao et al. (2004) identified a low-velocity zone with fairly sharply defined lateral boundaries in the upper mantle. It extends vertically to a depth of 200–250 km. Its Pn and Sn velocities are 5% and 8% slower, respectively, than those of its surroundings. They inferred lower densities in this part of the mantle, attributing them to elevated temperature. Cordell et al. (1991) calculated the thickness of the lithosphere beneath the Colorado Plateau to vary from 75 km near its edges to a little more than 125 km at its center. In contrast, the lithosphere beneath the Great Plains was determined to be >200 km. This would also lead to asymmetry of the epeirogen.

Assuming that Cordell et al.'s (1991) interpretations are valid, the thickness of the lithosphere is notably less beneath the Colorado Plateau than beneath the Great Plains. This should lead to greater buoyancy of the plateau and, hence, to its generally higher elevation. Morgan and Swanberg (1985) noted that low upper mantle seismic velocities beneath the plateau suggest that the upper mantle is anomalously hot and that thermal expansion is to be expected. A significant question is what mechanism is responsible for the significant difference in contemporary lithosphere thickness between the Colorado Plateau and the Great Plains? They are separated by the Southern Rocky Mountains, which are only 200 km wide, a fraction of the east-west width of both the Colorado Plateau and the Great Plains.

My sense is that the crust of the Colorado Plateau was not appreciably thickened during the Laramide orogeny. It was simply lifted and compressionally buckled somewhat as a result of low-angle plate subduction and lithospheric drag by the Farallon plate beneath the continental lithosphere. It was elevated epeirogenically in post–middle Eocene time. Could the mantle portion of the lithosphere of the Colorado Plateau been thinned by passage of the mechanically coupled Farallon plate beneath it, a result of the shearing away of 75 km of its base?

The Colorado Plateau does not require appreciable thickening in order to account for its elevation relative to that of the Great Basin. The extension and collapse of the Great Basin crust appear to account for a major part of the difference in elevation between these two provinces.

I (Eaton, 1987, Fig. 4) and Roy et al. (1999, Fig. 3, profiles 2, 3, 4, and 7) demonstrated the regular westward decrease in elevation from the crest of the Southern Rocky Mountains out toward the center of the Colorado Plateau, the laterally limited western piedmont slope of the Southern Rockies. Topographic profiles drawn west for 250 km from the mountain crest are similar to, but less steep than, the eastern slope of the upper Great Plains. Although the plains extend many hundreds of kilometers east of the mountains, the plateau profiles are interrupted west of the center of the plateau by the presence of the erosional stumps of Early Cretaceous mountains at the western edge of the Colorado Plateau, a product of the Sevier orogeny of Arm-strong (1968), by Tertiary volcanism at the plateau's edge, as well as by rift-flank uplift of the Wasatch Mountains at the eastern boundary of the highly extended terrain of the Great Basin. It is thus not expected that the western piedmont slope of the southern Rocky Mountain epeirogen should match the eastern slope precisely or in its entirety.


Figure 3 presents two topographic profiles and two smoothed topographic maps of the summit region of the epeirogen. A substantial part of the cross-sectional bulk of what is identified here as the epeirogen is, of course, that of the residual Laramide Southern Rocky Mountains, which had not been eroded to sea level or even to levels of a few hundred meters to 1 km above it. The profiles are based on topography along lines of latitude, at 36°N and 33°N. The first one (Fig. 3A) crosses the epeirogen 1° south of the Colorado–New Mexico line. The second one (Fig. 3B) crosses it 1° north of the New Mexico–Mexico line. Dashed arches have been fitted to these profiles by eye, profiles that graze the present-day topographic surface of the Great Plains and eastern part of the Colorado Plateau (as well as the southern Basin and Range Province in the west half of Fig. 3B). A generalized topographic map could be made from a series of such profiles, but to eliminate the inherent subjectivity involved in drawing them, particularly the western slope, a digital elevation model was borrowed from Zoback et al. (1990). The result is shown in Figure 3C. Because all the topography in the region was included in the smoothing process, the elevations of the mountains made a contribution, even as that contribution was diminished by the process of upward projection. For this reason, a second map was prepared in a different manner as a basis for comparison.

The regional topographic map in Figure 3D was derived by inverting a Bouguer gravity map subjected to a 250 km low-pass filter (Hildenbrand et al., 1982), thus eliminating the effects of shallow gravity anomaly sources. Elevations on the epeirogen were determined by utilizing the regional empirical relationship between elevation and Bouguer gravity (illustrated in Eaton, 1987, Fig. 10 therein). The similarity between major features seen on these two maps is obvious, though the derived elevations differ somewhat. Either of these maps could be employed in the series of arguments that follow, but the map in Figure 3C was chosen arbitrarily, simply because the calculations involved in making it were extended to the whole of the North American plate.

Each time the contours from the map in Figure 3C appear on the figures that follow, the reader is asked to keep in mind the following: topographic interference from a lesser regional high at the western edge of the Colorado Plateau obscures part of the western slope of the epeirogen's summit (subjective dashed lines show what the full summit might look like were it not for this contributory interference).

Relationship of the Southern Rocky Mountains, the Rio Grande Rift, and the Southern Rocky Mountain Epeirogen

Figure 4A shows the location of the Southern Rockies and the Rio Grande rift with a superimposed topographic map of the Southern Rocky Mountain epeirogen. The mountains are bordered by deep structural basins filled with post-Laramide Cretaceous and Paleogene sedimentary rocks on both sides. As noted above, the deformed rocks in these basins are unconformably overlain by Oligocene and Miocene continental sedimentary rocks near the mountain front. The total structural relief on the Precambrian surface as measured on a cross section from the deepest part of the Piceance Basin eastward to the crest of Mount Elbert, Colorado, and down to the deepest part of the Denver Basin is >6500 m; see the map of Bump (2003, Fig. 1 therein).

The solid and dashed red line in the southern part of this figure marks the northern boundary of the southern Basin and Range Province. South of this line many mountain ranges strike approximately N55W, but in the southern Rio Grande rift, which continues south of the dashed red line, they strike north-south (Fig. 4B). The southernmost Rio Grande rift appears to be younger than the southern Basin and Range Province, which it adjoins, but because of their general physiological and even structural similarity, Fenneman (1931) identified the ranges within the Rio Grande rift as part of his Basin and Range Province. Although I agree that the southern Rio Grande rift displays basin and range structure, I hesitate to agree with Fenneman's (1931) physiographic province “annexation.” Their creating stress fields set the two apart. Only the structure is similar. The spreading Rio Grande rift is a by-product of development of the Southern Rocky Mountain epeirogen, which, in turn, is a product of lithospheric thinning. The profile in Figure 3B shows that bowing of the epeirogen's crust at that latitude (33°N) includes that of the southern Basin and Range Province, which, in turn, permits the interpretation that the epeirogen is younger than the province.

Figure 4B is a map of normal faults in the area where these two geologic provinces meet. The normal faults are displayed in contrasting colors in order to separate them as to origin. The pattern they display is not dissimilar to that of normal faults near the intersection of the Ethiopian rift system and the Afar rift. Note the flaring or widening of both the Rio Grande rift on the left, and the Ethiopian rift on the right, as they approach areas of transverse rifting. The flared version of the southern Rio Grande rift is shown in Figure 4A by dashed blue lines. It coincides more or less precisely with the southward flaring of high surface heat flow values (as demonstrated in a following figure). The area limited solely to north-trending normal faults is defined by solid blue lines, a more traditional definition of the Rio Grande rift.

A sampling of the uppermost Great Plains elevations at the foot of the east face of the southern Rockies yields an average elevation of 1925 m, but topographic variations result from the fact that the east face of the mountains is an irregular erosional edge. The main point here is that the base of the mountains proper is fairly well circumscribed by, or just a little below, the 2000 m contour on the epeirogen, thus this contour can be used as a reference guide defining the base of the mountains in all the figures that follow.

With respect to the Rio Grande rift in Colorado, Tweto (1979b) made a distinction between what he called “the Rio Grande rift proper” and the “Rio Grande rift system.” The former term refers to the continuous graben of Neogene age that begins at the head of the upper Arkansas River and broadens southward. The latter term includes “the…more extensive belt of Neogene block faulting, including not only the rift proper, but neighboring and north-continuing faults that were tectonically active concurrently with the rift proper” (Tweto, 1979b, p. 34). Although these normal faults continue to the Wyoming state line, they do not represent significant lithospheric rifting, only crustal extension. As demonstrated here, the northern Rio Grande rift system, as defined by Tweto (1979b) for Colorado, and the southern Rio Grande rift in New Mexico, are more or less coincident with (1) an elongate geoid anomaly, (2) an elongate, long-wavelength, deep crustal magnetic anomaly, and (3) a surface heat flow anomaly. These anomalies reflect the existence of the southern Rocky Mountain epeirogen in northern Colorado, and the faults seen there are related to the epeirogen, not to a perceived northward extension of the Rio Grande rift. These geophysical anomalies continue southward to southern New Mexico coincident with the southern Rio Grande rift as defined by Seager and Morgan (1979) and redefined by Baldridge et al. (1984).

The Rio Grande rift proper is the one shown in Figure 4A (in blue). The proposed northward extension of Tweto's Rio Grande rift system is shown in green. The rift proper opens and broadens southward from the headwaters area of the Arkansas River. The greater width of rifted terrain shown at the southern end of the rift, near the Mexican border, has traditionally been shown to be narrower, essentially restricted to a belt of north-trending normal faults flanking the Rio Grande River. To the west, however, contemporaneous lateral spreading includes the northeast-trending Neogene normal faults bordering the Plains of San Agustin graben southwest of Datil, New Mexico, and the San Francisco, Brushy, and Saliz Mountains near Reserve, New Mexico. To the southeast, it includes the north-northwest–trending normal faults bordering the Brokeoff and Guadalupe Mountains and the Salt Basin near El Paso Gap, New Mexico. The full pattern of these splaying normal faults separating mountain ranges and intervening valleys and basins is best seen on the map of Baldridge et al. (1984), where the distribution of middle to late Cenozoic valley fills defines both the broad southern end of the rift and the northern boundary of the Basin and Range Province.

The northern starting point of the rift proper sits at the highest point on the epeirogen and broadens steadily to southern New Mexico, generally following the axis of the epeirogen southward to an elevation of 2000 m. There this axis divides. The division owes to the rather abruptly increased breadth of the rift and consequent wider separation of its high flanking ridges. The 1500 and 1750 m contours define the southern terminus of the epeirogen, crossing the wide southern end of the Rio Grande rift, thereby including part of the Basin and Range Province.

Fenneman (1931) did not recognize the existence or significance of the Rio Grande rift, which was not identified until after 1931. The physiographic province boundary between his Southern Rocky Mountains and Basin and Range Province crosses the rift at the south end of the Sierra Nacimiento and Jemez Mountains on the west and the Sangre de Cristo Mountains on the east. The actual surface topography, as well as the smoothed regional topography, undergoes a significant drop in elevation on the south side of this boundary. The aforementioned three ranges end abruptly at the boundary, and are succeeded southward by the northern end of the laterally offset Sandia Mountains. This east-west physiographic province boundary is a major tectonic boundary of some sort, yet no single fault or distributed fault system is evidenced at the surface. It appears that the southern 30% of the Southern Rocky Mountain epeirogen has either partially collapsed or, more likely, has not been lifted like the area to the north. North-trending Laramide thrust faults and folds continue south and southeast of this boundary well into Mexico, thus the Laramide orogeny clearly deformed the region south of this abrupt boundary. The northern 70% may be a fairly good geologic analog of the Great Basin prior to its collapse.


Figure 5 shows the topography of the Southern Rocky Mountain epeirogen in relation to (1) present-day exposures of Precambrian basement rocks (black), (2) crustal sutures and major faults in the Precambrian basement (maroon) (Karlstrom et al., 2005), (3) outlines of the Late Pennsylvanian ancestral Rocky Mountains (dashed blue lines) (Miller et al., 1992), and (4) the Colorado Mineral Belt (solid green). Although the Colorado Mineral Belt contains mineral deposits of Laramide age, there are more abundant deposits of middle Cenozoic age within it. The second and fourth of these features generally trend northeast; the third trends northwest. The Southern Rocky Mountain epeirogen trends due north, cutting diagonally across the general trend of the North American Cordillera (as illustrated in Eaton, 1986, Fig. 10 therein). It lifted a belt of older, west-northwest– to north-northwest–trending Laramide ranges. In Wyoming and Colorado, the most prominent of these ranges are the southern Front Range, the Gore Range, the Medicine Bow Mountains, the Sangre de Cristo Mountains, the southern San Juan Mountains, the Sawatch Range, the Uncompahgre Plateau, and the Wet Mountains. Their average strike is N40W; the range of strikes is N36W to N48W. Such mountain ranges are few in New Mexico: the Chuska Mountains, the Gallinas Mountains, and the Pinos Altos Range. Elsewhere the trends of most New Mexico mountain ranges is approximately north-south, dominated by east-west crustal spreading associated with the southward widening Rio Grande rift. The orientation of the northwest-southeast ranges is due to Pennsylvanian and Laramide orogeny, but these mountains were later broadly uplifted with the emergence of the north-trending Southern Rocky Mountain epeirogen, a feature uninfluenced by major, first-order geologic structures in the crust created before it. The simplest explanation for this azimuthal discordance is that epeirogenic uplift originated at the lithosphere-asthenosphere boundary, well below the crust and much of the lithosphere.


Figure 6A602 shows the results of extending the smoothed regional topography of the epeirogen's summit to the whole of the contiguous United States and immediately adjoining parts of Canada and Mexico. Figure 6B is a similar map for a large part of Alaska. The latter map includes the regional settings of the Alaska Range (letter A) and the Wrangell–St. Elias Mountains (letter W). Given the wavelength employed in smoothing these maps, they demonstrate that the Southern Rocky Mountain epeirogen has the highest regional elevation on the upper surface of the entire North American plate. The Sierra Nevada, with its 12 peaks exceeding 4270 m, and the 14 high peaks in Alaska that have elevations >4270, both display regional topography that is 500 m lower than that of the Southern Rocky Mountains region.

There are other regions in western North America that exhibit regional elevations >1750 m, for example, central Idaho, central Wyoming, eastern Arizona, and northwestern New Mexico. They are shown by pale green and yellow color intervals. The whole of North America west of the Great Plains and Prairie Provinces in Canada is higher than the rest of the plate.

There are only two regions with smoothed elevations exceeding 2250 m on the maps in Figure 6A. Both are shown by the orange color interval, but one of them, the Southern Rocky Mountain epeirogen, rises higher than the other, to slightly more than 2750 m. It is shown by red and pink color intervals. The first is associated with the Yellowstone hotspot, the second with the region of the Southern Rocky Mountains. Both are first-order anomalous plate-scale features.


The geoid represents an equipotential surface of the Earth approximately coinciding with sea level. Theoretically, a geoid anomaly in a continental region should not correlate well with topography because of subsurface density contrasts associated with isostatic compensation. Part of the National Oceanic and Atmospheric Administration's GEOID 99 map (www.ngs.noaa.gov/GEOID/GEOID99) is shown in Figure 7. The smoothed topography of the epeirogen (white contours) is superimposed on that of the geoid (black contours with colored intervals). In the northwestern corner of this figure (marked by the −10 m contour label) is the southernmost portion of a gigantic geoid anomaly associated with the Yellowstone hotspot (Pierce and Morgan, 1992). The Yellow-stone hotspot geoid anomaly is by far the largest observed in the contiguous U.S. It displays an elevation of −8 m to −10 m at its summit and extends east to both north-central North Dakota and west-central Wyoming.

Immediately east of the southeast corner of the main Yellowstone hotspot anomaly, and essentially filling the rest of the illustration in Figure 7, is a narrower, south-trending, geoid anomaly associated with the Southern Rocky Mountain epeirogen and Rio Grande rift. Its elevation has a maximum value of −11 m at the Wyoming-Colorado border and it declines in elevation in a regular manner down the axis of the epeirogen to southern New Mexico, where it terminates. The 1500 m white elevation contour of the epeirogen locally follows the −22 m black contour of the geoid on the east. The 2750+ m high point on the epeirogen is south of the elevation maximum on the geoid at a level of −13 m. It appears that the regular southward decrease in elevation of the epeirogen's geoid anomaly is influenced by the existence of the regionally much larger Yellowstone hotspot anomaly to its north. Whether or not this is so, isostatic compensation is not yet complete in the Southern Rocky Mountains region and, hence, the assertion to the contrary (Eaton, 1987) was incorrect.

Because deviations of the geoid surface from sea level result from density variations in the subsurface, either lateral density variations associated with contrasts in rock composition or contrasts in physical properties, such as elevated temperature, could be the cause. Such deviations can also result from dynamic and/or flexural support of topography, as occurs at subduction zones.

We turn now to evidence supporting what I believe to be the driver of the epeirogen, lithospheric thinning and resultant elevated temperatures in the lithosphere. This interpretation differs from the preferred interpretation of Morgan (2003), who, after considering several contrasting mechanisms to account for post-Laramide uplift of the Southern Rocky Mountains and Colorado Plateau, chose changes in phase in the upper mantle and lower crust as the most likely explanation, as noted earlier.


Seismic evidence for what is probably elevated upper mantle temperatures in the region was offered in three publications from the LA RISTRA passive seismic experiment: Gao et al. (2004); Wilson et al. (2005a, 2005b); and Lastowka and Sheehan, Snelson et al., and Li et al., cited inKarlstrom and Keller (2005).

Among the several conclusions the LA RISTRA authors drew from their data is that significant velocity anomalies are present in the upper mantle beneath parts of the LA RISTRA line, where notably low upper mantle velocities extend to a depth of nearly 250 km. They concluded that this zone reflects higher solid state temperatures than its surroundings, though the presence of a small amount of partial melt or hydration of the lithosphere (both of which represent phase changes of the sort proposed by Morgan, 2003), were not ruled out entirely as possible causes for depression of the seismic velocities. A partial illustration of LA RISTRA results, along with a profile of the smoothed Southern Rocky Mountain epeirogen topography shown above it, is presented in Figure 8. Note that one of the upper mantle's two separated low-velocity zones (the wider one) is broader than the surface expression of the Rio Grande rift, the object of their study, but, together, the two have a total span narrower than the summit of the Southern Rocky Mountain epeirogen.

Morgan (1983) presented curves of surface uplift values resulting from varying amounts of lithospheric thinning. From these curves, I deduce that a 200-km-thick lithosphere beneath the Great Plains (Cordell et al., 1991), thinned to 50 km, should result in an uplift of 2.0–2.75 km as the outcome of slow and rapid thinning rates. However, not all of the bulk of the Southern Rocky Mountain epeirogen, as defined by the topography of its upper surface, is due to epeirogeny, as emphasized earlier. A lithosphere of 125 km thickness beneath the Colorado Plateau, thinned to 50 km, should result in an approximate uplift of 1.35 km at most.

Figure 9 shows the distribution of surface heat flow values in the region (Blackwell and Richards, 2004) plus an outline of the Rio Grande rift shown by a pale blue dashed line. Surface heat flow values range from <50 mW/m2 to >150 mW/m2 in the map area (Decker et al., 1988). The mean surface heat flow for continents worldwide is 65 mW/m2, based on 10337 measurements (Stein, 1995). Mean heat flow for regions of exposure of Proterozoic basement rocks, as in the Southern Rocky Mountains, is 58.3 mW/m2.

The correlations between areas of high surface heat flow values, high elevation, and exposures of the eroded Precambrian basement are geographically variable. In the Front Range of north-central Colorado, the area of the largest continuous exposure of Precambrian basement rocks, the correlation is positive and strong, as it is across the Uncompahgre Plateau in southwestern Colorado. Elsewhere, elevated surface heat flow is seen above exposures of middle Cenozoic volcanic fields, especially the San Juan field of Colorado and part of the Mogollon-Datil field of New Mexico. The spatial correlation between high surface heat flow and mid-Tertiary igneous rocks is just as striking.

There is high surface heat flow beneath the summit of the Southern Rocky Mountain epeirogen owing to the presence of these post-Laramide rocks at shallow depths, but there is also high reduced heat flow. As an aside, 95% of the thermal springs in Colorado and New Mexico illustrated by Waring (1965) occur in areas where the surface heat flow shown by Blackwell and Richards (2004) is 80 mW/m2 and higher.

The approximate surface heat flow values at locations A, C, F, and G in Figure 9 are illustrative here. They are immediately above the deepest parts of four large Laramide structural basins that flank the Southern Rockies, two on the east side of the range and two on the west. In the basins labeled A, C, and F (the Piceance, Denver, and San Juan Basins, respectively), the Precambrian basement surface is at depths of more than −2000 m. In basin G (the Raton Basin), the basement surface is at a depth of a little more than −1000 m (Bump, 2003). Approximate surface heat flow values above the center of these four basins are as follow: A, 65–70 mW/m2; C, 75–80 mW/m2; F, 65 mW/m2; G, 85 mW/m2. All equal or exceed the mean surface value of continental regions. At the sites labeled B, D, E, and H, the basement surface, though buried, is much shallower than the basin floors. Because Precambrian basement rocks underlie the entire area of Figure 9, there clearly is no relationship between basement surface elevation and surface heat flow.

In a hypothetical model of horizontally layered sedimentary rocks underlain by flat Precambrian basement granites that display radiogenic heat production, vertical offset of the basement surface along a fault will not produce a contrast in surface or reduced heat flow across the fault; hence, the difference in surface heat flow values observed here between mountains and flanking basins is due to some other factor. One possibility is a difference in the depth of erosion of the basement, it being greater in the mountains. Deeper levels of the basement rocks are exposed there, rocks locally displaying penetrative ductile flow. It is possible that the composition of these Precambrian rocks is somewhat different from those underlying the flanking basins, and perhaps they are rocks with higher radiogenic heat production.

In sampling igneous and metamorphic rocks in the Southern Rocky Mountains in order to ascertain levels of radiogenic heat production for calculations of reduced heat flow, Decker et al. (1988, Table 3 therein) found especially high levels in igneous rocks of post-Laramide middle and later Cenozoic age, the preponderance of their samples. These rocks are abundant in the summit area of the Southern Rocky Mountain epeirogen. Their spatial distribution at the surface is shown in a later figure. High values of surface heat flow in the Southern Rockies thus seem to owe significantly, but not exclusively, to heat production in post-Laramide rocks younger than middle Eocene in age, some of them exposed at the surface and some occurring as relatively low density plutons in the shallow crust.

In their concluding remarks, Decker et al. (1988) stated that their study demonstrated unequivocally that the Front Range of the Southern Rocky Mountains in Colorado is characterized by a high reduced heat flow of 54–58 mW/m2 substantially exceeding that of part of its surroundings. Blackwell et al. (1991) agreed, suggesting that the Southern Rocky Mountains and Rio Grande rift have higher heat flow from beneath the radiogenic crustal layer. There is thus a deep source of heat in the area, as well as a shallow one.

(Reiter, 2008) also addressed the topic of heat flow, as well as Bouguer gravity anomalies and teleseismic velocity anomalies in the region, and concluded that heat sources, as well as gravity anomaly sources, occur in both the upper mantle and the crust. There is thus a clear consensus on the existence of high heat flow from beneath the crust.

In discussing the constraints on rift thermal processes, Morgan (1981) offered that the lack of broad regional heat flow anomalies outside the floors of rift zones suggests that thermal anomalies originating in the mantle have not yet reached the surface by conduction. He noted that because the lithosphere thins at a rate faster than a conducted thermal disturbance can propagate through it, it would be only after a very long period of time that the effect of lithospheric thinning would appear as a thermal perturbation at the surface. Morgan (1981) presented a curve that makes possible the determination of the length of time the effect of a 200-km-thick lithosphere thinning to 50 km would take to appear at the surface. In this instance, that period of time is 32 m.y. In the last section herein, I focus on the timing of epeirogenic uplift and show that the initial uplift during the second episode of epeirogenesis in the region probably took place earlier than 32 m.y. ago, thus allowing for the effect of a thermal perturbation resulting from lithospheric thinning to appear at the surface.

In other major international rift zones, where there is an absence of broad heat flow anomalies straddling the rift, as is seen in the northern two-thirds of the Southern Rocky Mountain epeirogen, high heat flow is geographically restricted to the floors of rifts. This would seem to be the case for heat flow anomalies in the southern third of the epeirogen. The more restricted width of the anomaly there and the spatial coincidence of surface heat flow values ranging from 80 to >100 mW/m2 within the rift seem to suggest that heat from the base of the lithosphere has not yet reached the surface or, perhaps, is just now beginning to reach it. It is interesting to note in the data of Sahagian et al. (2002) that the collecting sites they occupied that have been lifted by the least amount are all in the southern third of the epeirogen, where uplift has taken place only in the past 4.7 m.y.

Significance of Pre-Laramide Igneous Activity along the Southern Rocky Mountain Epeirogen Summit

Figure 10 shows the locations of igneous rocks of several different ages emplaced in the summit region of the Southern Rocky Mountain epeirogen prior to Laramide time: (1) granite plutons of Proterozoic age extending the entire length of the southern Rocky Mountain epeirogen, (2) diatremes of the period 800–500 Ma in southern Wyoming and northern Colorado clustered with fossil-bearing Devonian diatremes from the period 404–326 Ma (Lester et al., 2001; Smith et al., 1979), and (3) calciccarbonatites and ferro-carbonatites, syenites, alkali-feldspar granites, and mafic igneous rocks of Cambrian–Ordovician age, their ages overlapping that of the youngest diatremes. These rocks span the period of 574–427 Ma in Colorado and 664–457 Ma in New Mexico. Their composition implies extensional stresses in the crust in early Paleozoic time (McMillan and McLemore, 2004). The chemical compositions of these rocks are consistent with partial melting of upwelling asthenosphere in an extensional environment. Apparently, the lithosphere underwent aborted prerifting here long before development of the Ancestral Rocky Mountains, the Laramide Southern Rockies, the Southern Rocky Mountain epeirogen, and the Rio Grande rift. The episode left an indelible, and what I believe to be a highly significant, north-trending heterogeneity in the lithosphere, one that was tectonically “remembered” in post–middle Eocene time. Because the base of the lithosphere is an isotherm, a period of cooling following this episode might have resulted in recovery of its pre-event thickness, but not its pre-event composition and physical properties. The same is true for the Proterozoic period of granite emplacement. In discussing the evidence and cause of intracontinental uplifts in Africa, Bailey (1972) suggested that during translations and rotations of a continental plate, structures in the crust and irregularities along the base of the plate give rise to irregular stress distributions.

Figure 10 also shows a Proterozoic high-temperature zone (enclosed by a red line) that Shaw et al. (2005) interpreted as an area where Proterozoic temperatures reached as high as 500–550 °C (roughly equivalent to the Barrovian biotite zone). They postulated that the rocks of this zone may represent a deeper level of exposure of the basement than elsewhere in the region. If so, a simple corollary is that these rocks represent the area of greatest basement uplift. As noted earlier, they display penetrative ductile deformation.

As discussed later, magmatic invasions of the crust also took place in both late Eocene– Oligocene and late Miocene time. This is not to suggest that elevated temperatures resulting from any of these magmatic episodes lingered until late Miocene time or to the present, they simply demonstrate that the summit area of this epeirogen has undergone a very long and long-punctuated period of igneous intrusion and volcanic activity, some of it implying aborted rifting, which left its imprint on the lithosphere. A significant heterogeneity was created in both the crust and the upper mantle.

Temperature-Related Seismic Velocities in the Mantle

It was discussed earlier that anomalously low values of Pn and Sn occur along the LA RISTRA line (Gao, 2004). Figure 11 is a map of Pn velocity anomalies at 100 km depth showing their horizontal subsurface distribution. This figure was modified from a digital map (Dueker et al., 2001; Dueker, 2004, personal commun.) by contouring it as an analog map. Elevation contours on the summit area of the Southern Rocky Mountain epeirogen are superimposed. It can be seen that, in general, most areas of low Pn velocities occur within the epeirogen's enclosing 1750 m contour. A single exception occurs in northeastern Colorado.

The area of lowest Pn values is flanked east and west by areas of higher values shown in blue. The same is seen on the cross section of the LA RISTRA line. Given the known correlation between compressional wave velocities and densities, it is mildly puzzling that the fairly abrupt boundaries between depressed and elevated Pn velocities as deep as 100 km are not marked by coincident gravity gradients at the surface. This is especially true for the outer boundaries of the two (blue) zones that straddle the Southern Rocky Mountains.

Dueker (2004) and Dueker et al. (2001) focused on individual Pn anomalies in their interpretation of this map, but the focus here is on the entire cluster of low Pn anomalies because it is spatially coincident with the summit region of the epeirogen.

Elevated Crustal Temperatures Suggested by the Regional Magnetic Field

Figure 12 presents an upward-projected magnetic anomaly map of the region (Mayhew, 1982; Taylor et al., 1983). The contour interval is 5 nT. This map displays a long-wavelength regional magnetic low coincident with the full length of the epeirogen (southern Wyoming to southernmost New Mexico). Areas with values lower than −20 nT are shaded. Taylor et al. (1983) suggested that such an anomaly could arise from any one of several different causes: (1) lateral, intracrustal contrasts in magnetization (lateral rock compositional variations at depth), (2) variations in crustal thickness, or (3) an anomalously shallow depth of the Curie isothermal surface (local diminution in the thick-ness of the magnetized crustal layer). Mayhew (1982) associated the anomaly with the Rio Grande rift and elevated crustal temperatures.

Magnetic minerals lose their ability to maintain magnetization at temperatures above their Curie temperature, which, for magnetite, the most commonly occurring magnetic mineral in nature, is ~575–580 °C, depending on variations in its composition (Carmichael, 1989). Saturation magnetization decays to zero at the Curie temperature.

Although the crust thins laterally across the southern part of the epeirogen, modern maps of crustal thickness (e.g., Keller et al., 1998), especially thickness variations from north to south, reveal a limited to no relationship to the long-wavelength magnetic map. The regional magnetic low in Figure 12 originates at a deep level, probably in the lower crust, owing to a north-trending, arched, and elevated Curie geotherm, as Mayhew (1982) proposed, although rock compositional variations cannot be ruled out entirely, given the distribution of igneous rocks at the surface and, by inference, within the crust.


The dips of postorogenic sediments on the east flank of the Southern Rocky Mountain epeirogen are composite, displaying both primary (depositional) and younger tectonic dips. The Ogallala Formation consists of both fluvial and eolian sediments that once constituted an enormous, gently east dipping, bajada-like sedimentary apron that blanketed the east face of the Southern Rocky Mountains. I believe that it has been tilted since the end of its accumulation, which occurred ~6–5 m.y. ago (Eaton, 1986; McMillan et al., 2002). Its upper surface now defines the eastern surface of the southern Rocky Mountain epeirogen. It is not, however, the only tilted surface of significance in the region.

In studying the denudation history of the Colorado Front Range and Wet Mountains, House et al. (2003) discovered a middle Cenozoic partial annealing zone (PAZ) tilted to the east. It passes eastward from a point above the crest of the Sandia Mountains near Albuquerque, New Mexico, through the present-day topographic surface at Santa Rosa, New Mexico, and to the east it is observed at depth in wells in Oklahoma. In the stretch of its profile illustrated in Figure 10 of House et al. (2003) and in Roy et al. (2004), uplift of the west end of this PAZ surface relative to the east end is ~3250 m. Roy et al. (2004) noted that surface exposures of Triassic sandstone between Santa Rosa and the edge of the Rio Grande rift, nearly 200 km to the west, yield AFT ages of 30–25 Ma; therefore, tilting has occurred since that time. Chronologically distributed volcanic age data described in the following reveal an igneous event-count maximum between 29 and 25 Ma (Chapin et al., 2004, their Fig. 4). The wholesale invasion of the crust by magmas of this age probably marked the initiation of significant epeirogenic uplift, i.e., the birth of the Southern Rocky Mountain epeirogen from pluton emplacement buoyancy, crustal thickening, and thermal expansion. This episode lasted from that time until a notable hiatus in igneous activity in middle Miocene time. The data of McMillan et al. (2002) and Leonard (2002), plus a set of additional new data provided here, suggest that uplift continued after initial deposition of the Ogallala Formation in middle Miocene time. Another example of eastward tilting of the east flank of the mountains is based on the identification of one or more sloping erosion surfaces and the orientation of incised paleostream valleys, as discussed earlier. Steven et al. (1997) postulated that general eastward tilting took place in two episodes, one in middle Miocene time and the other in latest Miocene to Quaternary time.

Figure 13 presents the results of first-order re-leveling on the east flank of the epeirogen in Colorado. The data were drawn from files of the National Oceanic and Atmospheric Administration. The leveled and re-leveled line was arbitrarily anchored east of the town of Las Animas, Colorado, on the Great Plains, 150 km from the mountain base. It was continued west through the city of Pueblo and up into the Southern Rocky Mountains to the town of Salida. This is the route of U.S. Highway 50, which follows the course of the Arkansas River. Although the latter is a modern erosional incisor of the epeirogen today, and though some very small part of the elevation change in the middle 1950s along this line hypothetically could be the result of erosion by the Arkansas River, the lapsed time between leveling runs seems far too short to account for significant erosion, and therefore much, if any, eastward tilting.

Save for two, sharp, single-benchmark downturns on the profile in Figure 13 (both are most likely associated with groundwater withdrawal and respondent local subsidence), the change in elevation along this line increases steadily from east to west. It represents still another tilted surface, one that indicates continued uplift of the epeirogen today. The re-leveling results displayed in Figure 13 represent two separate spans of historic time. The eastern segment was re-leveled in 1985 following an original first-order leveling survey in 1934.The western segment was re-leveled in 1985 along a portion of a line originally run in 1953. The two segments are joined at a shared benchmark (circled) at the bottom of the sharp downturn immediately west of the city of Pueblo.

The resurveys represent two different lengths of historic time, both of them relatively short; the eastern one, a span of 51 yr, the western one, only 32 yr. The general uplift rate suggested by Figure 13, which has a tectonic cause, is a full order of magnitude lower than the rate of isostatic uplift of 1700 m of the Fennoscandian shield following the melting (removal of the load) of its ice cap during the past 10–15 k.y. (Aubrey and Emery, 1993), and thus it is not an entirely unreasonable rate regarding the response of the lithosphere to application of positive or negative crustal loads.

Timing of Epeirogenic Uplift

Given the geographic relationship between broad epeirogenic uplift, lithospheric thinning, elevated mantle temperatures, and crustal rifting, we examine surface manifestations of Cenozoic elevated crustal temperatures, i.e., the presence of volcanic and plutonic rocks exposed at the surface today, in order to gain some understanding of when non-erosion–related epeirogenic uplift began.

Figure 14A1402 is a sketch map of widespread post-Laramide igneous rock distribution in the summit region of the epeirogen. Most of these rocks are shown in black, but two are shown in orange. Older igneous rocks, shown in other colors, are present in the same area. Those in green represent the diatremes, those in red, the 1.4 Ga plutons, and those in purple, the extension-related Cambrian–Ordovician plutons. Outcrops of these older rocks that are within the black areas of post-Laramide rock outcrops are not shown.

Magmatic and volcanic rocks ranging in age from 42 Ma to younger than 1 Ma and in composition from andesite to basalt (and bimodal basalt rhyolite) are included among the rocks shown in black. Despite the age and compositional differences among these rocks and the changing plate tectonic regimes that created such compositional differences, it is seen that the crust in the region has been intruded repeatedly (in Proterozoic, Devonian, Ordovician, and middle Cenozoic time), and subjected to past high-temperature melts that repeatedly brought advective heat and buoyancy to the crust. Some of this activity led to emplacement of simple dikes and some to the development of large volcanic fields, remnants of which are still present in the area, as well as some to large, relatively low density plutons in the crust.

With regard to the timing of epeirogeny, it is necessary to ask what the distribution of these rocks was in time. Figure 14B shows two super- imposed step plots of the ages of these rocks in 3 m.y. intervals. The first data set (represented by the heavier line), was assembled from the compilations of Luedke (1993a, 1993b) and Luedke and Smith (1978a, 1978b). These ages are from igneous rocks in Colorado, New Mexico, and adjoining parts of eastern Utah and eastern Arizona. Those represented by the lighter line were taken from Chapin et al. (2004) and recast from 1 to 3 m.y. intervals. They are limited to New Mexico outcrops only.

A great many rock ages are shared between the two data sets; therefore, it is scarcely surprising that the occurrence of the two major igneous event peaks and the trough between them essentially coincide. In the New Mexico data set, the great bulk of magmatic/volcanic events of the first long igneous episode took place in the time interval 37–22 Ma, peaking between 28 Ma and 25 Ma. It appears that by adding the ages of rocks in Colorado to those in New Mexico, the Eocene–Miocene event peak in the light line plot broadens to a somewhat wider time span, suggesting, but scarcely demanding, that initiation of intense igneous events began somewhat earlier in the north and lasted somewhat longer.

As seen, both plots display a pair of maxima, the one just described and the other between 16 and 0 Ma, and peaking between 4 and 1 Ma. Chapin et al. (2004) suggested that rollback of the low-dip, subducted Farallon plate took place between 36 and 24 Ma. I offer that it actually may have begun as early as 38 Ma. The first evidence of Cenozoic crustal stretching in the region occurred ca. 36–35 Ma, late Eocene time (Cather, 1990; Chapin and Cather, 1994). Crustal extension thus began within 1 m.y. before or after the beginning of sharply increased magmatism. It was well under way when the peak of ignimbrite explosions was reached at 28 Ma. It was followed by an abrupt increase in the number of igneous events in the epeirogen's future summit region, including profound ignimbrite eruptions. De Silva (2008) proposed that ignimbrite flare-ups represent a high magmatic flux driven by an elevated rate of energy transfer from the mantle. In the case of the Southern Rocky Mountains region, the initiation of such a flare-up likely represented an advanced stage of lithospheric thinning, as well as the rise of hot asthenosphere resulting from rollback of the Farallon plate. The uppermost crust was weakened and the brittle-ductile transition moved upward, enabling listric normal faulting in the shallow crust. One of the ignimbrites, the Wall Mountain Tuff, is geographically widespread, extending eastward onto the upper Great Plains. It has an age of 36.7 Ma (McIntosh and Chapin, 2004), and its distribution suggests an east-dipping slope or later eastward tilting. If the subducted Farallon plate was sinking and rolling back a little before this time, the opportunity for hot asthenospheric material to flow in and replace it had to be substantial (Chapin et al., 2004). It doubtless led to heating of the lower lithosphere. It is interesting that the area of the Southern Rocky Mountain epeirogen and its igneous rocks of all ages occupy a significant part of what has been dubbed the “Laramide magma gap,” where no igneous rocks of Laramide age exist, save for those in the Colorado Mineral Belt. It would seem that magmatism was delayed here from Late Cretaceous to late Eocene time.

The younger of the two episodes of post-Laramide igneous activity began immediately following the relative hiatus from 20 to 16 Ma. This younger magmatic event-frequency maximum appears far more pronounced in the New Mexico data set owing in part to more abundant recent sampling and radiometric dating there, much of it along the northeast-trending Jemez volcanic zone, but basalts of this period are widespread across southern New Mexico and southeastern Arizona. Similarly, the highest event-frequency maximum in the heavy-line step plot between 28 and 25 Ma is probably the result of oversampling in the San Juan and Mogollon-Datil volcanic fields of Colorado and New Mexico.

The basaltic rocks of the second event maximum mark a different tectonic regime, as does the development of the Jemez volcanic zone that follows the Precambrian suture between the Yavapai and Mazatzal basement provinces. It is probable that decompression melting of upwelling asthenosphere began to generate basalts, as suggested by Chapin et al. (2004). The Jemez zone saw surface leaking of small floods of basalt, but rocks 16 Ma and younger are sparse along the full length of the Rio Grande rift. Inasmuch as the rift underwent further, but slower, extension at this time, one might have anticipated volcanism of the sort commonly seen in and along other major rift zones. Wilson et al. (2005a, 2005b) suggested that a lack of highly focused mantle upwelling may explain the deficit of rift-related volcanism.

The findings of Sahagian et al. (2002) and Sahagian and Proussevitch (2007) suggest more or less continuous epeirogenic uplift of the surface between 23 and 1.98 Ma. The technique they employed was discussed in Sahagian (1985), Sahagian et al. (1989), Sahagian and Maus (1994), and Sahagian and Proussevitch (2007). It was validated and calibrated on Hawaiian basalt flows taken from a large range of elevations (45 m to 4361 m) on Mauna Loa volcano. The maximum uplift they determined for the Southern Rockies region is 2170 m. It took place at 20 Ma. This episode of uplift began to tilt the oldest Ogallala sediments (17–12 Ma; Eaton, 1986) eastward as they were deposited and seemingly has continued to do so. Figure 15 shows the location of the greatest elevation changes determined by Sahagian et al. (2002) plotted on the topography of the Southern Rocky Mountain epeirogen. Data from their sites at the west edge of the Colorado Plateau, that they plotted separately, are not included here, but they fit the model offered here inasmuch as they occur below the summit of the epeirogen, well down its western slope.

Those collecting sites with the highest elevations are those of the oldest lavas in their collection suite. Those that are youngest are seen at the lowest elevations in Figure 14, having undergone the least amount of uplift. The data permit the interpretation that Miocene and younger uplift began first in the region of the northernmost sites and continued there as it propagated southward. The Miocene climatic optimum, a warm, dry period, began ca. 17.5 Ma, and was followed by significant cooling and a moister climate from 15 Ma to a little after 12.5 Ma. The oldest Ogallala sediments began to be deposited ca. 17.0 Ma, while the oldest lavas in the Sahagian et al. (2002) suite were erupted at 23 Ma.

To reiterate, a massive magmatic invasion of the crust took place following erosional planing and incision of the mountains at the end of the Laramide orogeny. It was initiated between 40 and 37 Ma and peaked between 37 and 22 Ma (late Eocene–early Miocene time). This episode brought abundant, relatively low density magma and advected heat energy to the Laramidethickened crust, resulting in greater lithospheric buoyancy. With what can be assumed to be a hot asthenosphere beneath it, the result of rollback of the Farallon plate, the crust thickened still more than it had from orogenic compression and the lithosphere expanded thermally, bulging upward to an elevation similar to that of the 34 Ma Florissant fossil beds (Colorado), ~2600 m (Gregory and Chase, 1992).

The period of deposition of the Ogallala Formation was followed by its eastward tilting (McMillan et al., 2002). The re-leveling data illustrated in Figure 13 show that tilting is continuing today. Given the post-Laramide chronologic igneous history of the region, it is concluded that the Southern Rocky Mountain epeirogen is partially, if not wholly, an uncollapsed geologic analog of the Great Basin. The crust of the region that was to become the Great Basin was significantly thickened during the Nevadan and Sevier orogenies and had a high elevation. It later collapsed as a result of both its potential energy and a significant episode of extension that thinned the crust.


The site of the present-day Southern Rocky Mountains is one of regional uplift in postorogenic time. A major fraction of the elevation of the highest peaks in today's mountain range results in part from the development of epeirogenic uplift, created in part by lithospheric thinning and crustal heating and in part from stream incision following a Miocene climate change in the region. There thus have been two somewhat different, but colocated, Southern Rocky Mountain ranges in this region, one of Laramide age, the other of post–middle Eocene and younger age. The highest elevations supporting today's Southern Rocky Mountains make the range, on average, the highest region on the North Ameri-can plate. Conducted and advected heat energy and a crust inflated by middle Tertiary plutons and batholiths are responsible for the elevation of the Southern Rocky Mountains region today.

This paper benefited substantially from insightful and challenging reviews by C.E. Chapin and Paul Morgan. Both persuaded me to revisit and alter some of my original interpretations and to tighten or reject others. These two earth scientists, along with S.A. Kelley and S.M. Cather, all of them long students of the region discussed here, provided penetrative questions and patient instruction that were of considerable help. I am also indebted to Richard Saltus for providing a copy of the GEOID 99 map he prepared from the files of the National Oceanic and Atmospheric Administration (NOAA), Renee Shields of NOAA, who directed me to the first-order, level and re-level data on the Great Plains and eastern Southern Rocky Mountains, and Kris Verdin of the EROS (Earth Resources Observation and Science) Data Center, who was technically supportive in a number of ways.