Abstract
The Rio Grande Rift system tapers northward into the center of the southern Rocky Mountains. How rift initiation and evolution are related to development of the Rocky Mountains is not well understood. The Gore Range and adjacent Blue River Valley of central Colorado are the northernmost significant fault-related manifestations of the rift. New apatite (U-Th)/He data from the range record only middle and late Cenozoic exhumation phases, unlike neighboring Rocky Mountain uplifts that record a single phase of Late Cretaceous–early Tertiary unroofing. Middle Tertiary dates require that Gore Range basement remained covered by sedimentary rocks until at least the Middle Eocene. Normal faulting initiated during the Oligocene, inducing exhumation of basement to the surface and deposition of synrift fill in the Blue River Valley. Major cooling and unroofing in the Miocene were restricted to the southern Gore Range and continued until at least 7 Ma, with ∼2.3 km of total displacement along the Blue River normal fault. This middle to late Tertiary unroofing history is strikingly similar to that inferred for other rift flank uplifts to the south, and suggests broad synchroneity in the middle Tertiary onset and subsequent evolution of >700 km of the Rio Grande Rift. The results highlight that the notion of a northward-propagating rift, as suggested by its northward-tapering geometry, is incorrect, and preclude models invoking rift propagation to explain late Cenozoic elevation gain of the Rocky Mountains.
INTRODUCTION
The Rio Grande Rift is a zone of intracontinental extension that extends >1000 km from Mexico into central Colorado (Fig. 1). The rift bisects the southern Rocky Mountains, and for most of its length separates the Colorado Plateau to the west from the Great Plains to the east. There is no consensus on the geodynamic mechanisms responsible for the origin of the rift. Models include extensional collapse of the southern Rocky Mountains (e.g., Cordell, 1978; Eaton, 1986), clockwise rotation of the Colorado Plateau (e.g., Hamilton, 1981; Cordell, 1982; Steiner, 1986), and mantle upwelling associated with sinking of the Farallon slab (Moucha et al., 2008). How the rift is related to the evolution of the Rocky Mountains is similarly controversial; there are inconsistencies in the literature as to whether the rift propagated spatially through time, and how this bears on broader uplift in the southern Rocky Mountains region. Studies suggest that onset of rifting was broadly coeval ca. 29–26 Ma from at least central New Mexico through southern Colorado (e.g., Chapin and Cather, 1994). This synchroneity in rift development has been used to support Colorado Plateau rotation as the cause of rift initiation (Chapin and Cather, 1994). However, the rift’s northward-tapering character has led to the persistent notion that rifting initiated in the south and then propagated north into Colorado. This idea is most notable in work inferring late Cenozoic elevation gain of the southern Rocky Mountains based on tilting of late Cenozoic units on the Great Plains. These studies specifically invoked northward rift propagation as a plausible cause of surface uplift (McMillan et al., 2002; Leonard, 2002; Heller et al., 2003; Duller et al., 2012). In addition, reference to a northward-propagating rift is relatively common in studies located in the vicinity of the rift where the point is only peripherally relevant to the investigation (e.g., Penn and Lindsey, 1996; Porreca et al., 2006; Reiter and Chamberlin, 2011), but where the mention demonstrates the extent to which this concept is entrenched in the literature.
Although the Upper Arkansas Basin is the northernmost of the major Rio Grande Rift basins containing deep Tertiary fill (Fig. 1; Tweto, 1979), basins lacking that fill but sharing other structural similarities with the southern rift basins extend north to the Wyoming border (Chapin and Cather, 1994; Kellogg, 1999). The evolution of these less-studied rift basins to the north is not as well constrained as the history of those to the south. Determining whether the northern basins evolved over a time frame similar to that of their southern counterparts, or are somewhat younger and developed as the rift expanded northward, is key for evaluating hypotheses of a propagating rift and the proposed link with late Tertiary Rocky Mountain uplift.
The Gore Range in central Colorado is the most significant of the northern Rio Grande Rift flank uplifts (Fig. 2). The Blue River normal fault separates the Blue River Valley from the range and accommodates ≥1.4 km of vertical displacement at its southern end (Kellogg et al., 2011). The ∼1.4 km of relief in the southern Gore Range stands out against the neighboring and lower relief western Front Range and Park Range. The range is also distinctive when evaluating patterns of apatite fission track (AFT) dates in the southern Rocky Mountains (Fig. 1). Middle to late Tertiary AFT dates for the Gore Range (Naeser et al., 2002) contrast with Late Cretaceous and early Tertiary results from adjacent areas (Bryant and Naeser, 1980; Naeser et al., 2002; Kelley and Chapin, 2004; Kelley, 2005). Apatite (U-Th)/He (AHe) thermochronometry is sensitive to lower temperatures (∼30–90 °C) than the AFT technique, making it a potentially ideal tool to better constrain the middle and late Cenozoic cooling history of the region. We use AHe data (100 analyses, 22 samples) from the Gore Range to decipher the Cenozoic cooling and unroofing history of this northern segment of the Rio Grande Rift. This history is then compared with the evolution of rift flank uplifts to the south, and the first-order implications for the overall development of the Rio Grande Rift system are considered.
GEOLOGIC SETTING
Regional Geologic Setting
The southern Rocky Mountains are a north-northwest–trending set of ranges that extend from New Mexico through Colorado and into southern Wyoming. The highest elevations are in central Colorado. The mountains are between two geologic provinces that were less disrupted by Phanerozoic deformation, the Colorado Plateau to the west and the Great Plains to the east. The Rio Grande Rift bisects the southern Rocky Mountains and is a product of the change from the contractional environment of the Laramide orogeny to an extensional regime that has dominated the region during the late Cenozoic (e.g., Chapin and Cather, 1994). The rift extends from Chihuahua, Mexico, north through Texas and New Mexico into central Colorado (Chapin and Cather, 1994; Keller and Baldridge, 1999). Some have restricted use of the term “Rio Grande Rift” to the extensional basins containing deep Tertiary fill that occur as far north as the Upper Arkansas Basin in Colorado (Fig. 1; e.g., Tweto, 1979). However, normal fault–bounded valleys lacking that fill extend north into Wyoming (Tweto, 1979; Chapin and Cather, 1994; Kellogg, 1999). We follow Tweto (1979) and Kellogg (1999) and use the term “Rio Grande Rift system” to include all late Cenozoic extensional basins and flanking ranges, including those north of the Upper Arkansas Basin.
The onset of rifting and rates of extension have been relatively well studied in the major basins of the rift in central New Mexico and southern Colorado (e.g., Chapin and Cather, 1994; Brister and Gries, 1994; Lozinsky, 1994; Lewis and Baldridge, 1994; Cather et al., 1994; Miggins et al., 2002). Geochronology on volcanic rocks interbedded with synrift sediments and uplifted on rift flanks suggests initial basin development and sedimentation in the Late Oligocene, with the onset of rifting broadly synchronous from at least the Socorro Basin northward through the San Luis Basin (Chapin and Cather, 1994; Fig. 1). The sedimentary record suggests rapid extension and the development of significant topographic relief along rift flanks in Miocene time, followed by the slowing of extension and a shift from aggradation to incision by the Late Miocene (Chapin and Cather, 1994; Cather et al., 1994; Ingersoll, 2001; Miggins et al., 2002; Chapin, 2008). At approximately the time of rift initiation, widespread intermediate volcanism in the region ceased and was replaced by bimodal volcanism, with a peak of activity occurring ca. 28 Ma and continuing episodically thereafter; volcanism continued into the Pliocene in the San Luis Valley and into the Pleistocene in the Rio Grande Rift basins of New Mexico (Lipman and Mehnert, 1975; Chapin and Cather, 1994; Miggins et al., 2002; Chapin et al., 2004).
The first-order geophysical characteristics of the rift have been imaged along a transect including the rift and the adjacent Great Plains and Colorado Plateau in central New Mexico (West et al., 2004; Wilson et al., 2005). These studies indicate that the rift is characterized by a high heat flow of ∼90 mW/m2 and crust that is 10–15 km thinner than adjacent regions. Lithospheric thickness is ∼50 km as opposed to ∼200 km under the Great Plains and ∼135 km under the Colorado Plateau, with a low-velocity signature in the upper mantle twice as wide as the surface exposure of the rift. Geodetic analysis indicates that this broad zone currently undergoes evenly distributed extension of ∼0.12 mm yr−1 per 100 km (Berglund et al., 2012).
Blue River Valley and Gore Range Region
The Gore Range and adjacent Blue River Valley of central Colorado (Fig. 2) represent one of the northernmost fault-related manifestations of the Rio Grande Rift system north of the major extensional basins of the rift. The Blue River Valley is a 5–9-km-wide half-graben bounded by north-trending basement-cored uplifts that typify the southern Rocky Mountains. The valley is floored by thousands of meters of Mesozoic sedimentary deposits cut and offset by east-dipping normal faults (Kellogg, 1999), and contains a limited quantity of Oligocene sandstones, conglomerates, and volcanic rocks. The Blue River normal fault has ≥1.4 km of offset and separates the valley from the southern Gore Range, where only Precambrian crystalline basement is exposed. A cross fault divides the range into northern and southern segments (Naeser et al., 2002). The rolling hills north of the cross fault have a relief of no more than 700 m and typically <200 m, with no more than ∼200 m of offset across the Blue River fault based on the local juxtaposition of Jurassic units (Fig. 2; Kellogg et al., 2011). The northern Gore Range retains a partial cover of Paleozoic to Early Cretaceous sedimentary rocks and middle Tertiary volcanic rocks. The Gore reverse fault bounds the western edge of the Gore Range. On the eastern side of the Blue River Valley, the Williams Range thrust separates the valley from the crystalline rocks of the Front Range. The Middle Park Basin to the north is similar to the Blue River Valley in that it is floored by Mesozoic sedimentary fill at its western boundary with the basement-cored Park Range. The narrow, crystalline Tenmile Range is directly south of the Gore Range and is bounded by the Mosquito normal fault at its western edge.
The Late Cretaceous and younger tectonic history of the region is constrained by several sedimentary units and a widespread Middle to Late Eocene unconformity. At the northeastern end of the Blue River Valley and in the Middle Park Basin, the Upper Cretaceous to Paleocene Middle Park Formation consists of volcanic rocks and coarse sedimentary units containing clasts shed from the Front Range. This formation is interpreted to record the initiation of crustal shortening and relief development associated with the Laramide orogeny (Cole et al., 2010; Kellogg et al., 2011). In basins on the eastern side of the Front Range the youngest synorogenic strata are overlain by the 36.7 Ma Wall Mountain Tuff (Raynolds et al., 2007; Kelley, 2002). The tuff was deposited on a low-relief erosional surface that truncates older structural features (Epis and Chapin, 1975). The presence of this surface and the overlying tuff suggest that Laramide deformation had ceased in central Colorado by the Middle Eocene (Dickinson et al., 1988).
A single 500-m-thick Tertiary sedimentary section is preserved in a fault block in the Blue River Valley, and imposes key constraints on the post-Laramide tectonic and geomorphic history of the Gore Range (Tweto et al., 1970; Kellogg et al., 2011). A welded tuff overlying the basal conglomerate in the Blue River Valley section has been dated as 26.9 ± 0.06 Ma and imposes a minimum age for the initiation of sedimentation and relief development in the study area (Naeser et al., 2002). A 23.6 ± 0.6 Ma trachyandesite flow near the top of the section and flows of similar age overlying Jurassic–Cretaceous outcrops farther north in the valley (Naeser et al., 2002) indicate that there was little post–24 Ma net erosion in the Blue River Valley. Other middle to late Tertiary igneous activity in the region includes a 31.5 ± 0.1 Ma intrusive trachyte complex in the Blue River Valley (Naeser et al., 2002) and undated basalt flows overlying the State Bridge Formation, Morrison Formation, and crystalline basement in the northern Gore Range. Similar basalts west of the Gore Range are 23–24 Ma (Larson et al., 1975; Kunk et al., 2001, 2002).
During the Quaternary much of the Blue River fault trace was covered by glacial deposits (Kellogg et al., 2004). Most movement along the Blue River fault is thought to predate these units (West, 1978), but small fault scarps cut till that has been interpreted as Bull Lake age, indicating that faulting may have continued into the middle Pleistocene (Kellogg et al., 2002). The Blue River currently flows northward through the Blue River Valley to its confluence with the Colorado River, which drains Middle Park. The Colorado River flows westward across Mesozoic sediments of the western Middle Park valley floor and incises Gore Canyon in the northern Gore Range, adding another 500 m to the vertical bedrock relief in that location (Fig. 2).
PREVIOUS LOW-TEMPERATURE THERMOCHRONOMETRY IN THE SOUTHERN ROCKY MOUNTAINS AND RIO GRANDE RIFT
AHe and AFT techniques are complementary methods that are sensitive to temperatures from ∼30–90 °C and ∼60–120 °C, respectively. Thermal histories inferred from these data can be used to decipher unroofing through the upper few kilometers of the crust, provide insights into geotherm evolution, and enable inferences regarding elevation change histories. In this paper, unroofing refers to the removal of rocks at the Earth’s surface due to erosion or extensional tectonism, and uplift is the increase in elevation of the Earth’s surface.
Extensive AFT data sets have been acquired in the Rio Grande Rift system and southern Rocky Mountains. These data are plotted in Figure 1 and reveal several distinct patterns. At the highest elevations, such as the top of Pikes Peak and Mount Evans in the Colorado Front Range, middle Mesozoic and older AFT dates are preserved, and are interpreted to reflect rocks that were at temperatures <110 °C prior to the Laramide orogeny (Bryant and Naeser, 1980; Kelley and Chapin, 2004). Much of the Colorado Rocky Mountains and certain uplifts in northern and central New Mexico preserve Late Cretaceous and early Tertiary dates, reflecting passage through the ∼110 °C isotherm during exhumation caused by latest Cretaceous and early Tertiary mountain building (Bryant and Naeser, 1980; Kelley et al., 1992; Kelley and Chapin, 2004; Kelley, 2005). Rocks with middle Tertiary or younger AFT results are found almost entirely in rift flank uplifts along the main Rio Grande Rift system basins in southern Colorado and New Mexico, where dates young from the tops of ranges toward the basin-bounding faults (Lindsey et al., 1986; Kelley and Duncan, 1986; Kelley et al., 1992; House et al., 2003; Kelley and Chapin, 2004). An exception to this is the Gore Range of central Colorado, which contains the northernmost middle and late Tertiary AFT dates yet published in the southern Rocky Mountains, with dates as young as 5 Ma (Naeser et al., 2002).
In contrast, despite sensitivity to lower temperatures than the AFT method, published AHe data sets are sparse in the southern Rocky Mountains and Rio Grande Rift. No detailed AHe data set has been published for any location within Colorado. The only such published AHe data in New Mexico are Early to Middle Miocene dates from the Sandia Mountains outside Albuquerque (House et al., 2003).
APATITE (U-Th)/He DATA FROM THE GORE RANGE
We acquired AHe dates for 100 individual apatite crystals from 22 samples collected from elevations spanning 2138–4139 m in and around the Gore Range. The sample suite consists of 21 samples of Proterozoic basement and 1 sample of red siltstone from the Permian–Triassic State Bridge Formation. Half of these samples were newly collected during the course of this project, and half are mineral separates shared by C. Naeser for which AFT data were previously published (Naeser et al., 2002; Supplemental Fig. 11).
All AHe data were acquired at the California Institute of Technology following the methods described in Flowers and Kelley (2011). Analytical results are listed in Table 1. Analytical uncertainties for individual analyses are based on the propagated error from the U, Th, and He analyses and grain length measurements. In Figure 3 and elsewhere in the text, sample results are reported as the sample mean and 1σ sample standard deviation for the 17 samples with <20% sample standard deviation. For the two samples with standard deviation >20%, the range of dates is reported. The individual dates are noted for the three samples with only one or two analyzed grains.
Mean AHe dates range from 6.9 ± 2.8 to 41.8 ± 5.3 Ma. Figure 3 displays these results on a geologic map, color coded by mean sample date. Supplemental Figure 22 displays the same results on a digital elevation model. Figure 4 plots the results versus elevation for basement samples with three or more grains analyzed. To facilitate discussion of results and interpretations, we subdivide the data into four groups based on location: (1) eight samples from the southeastern Gore Range, west of the Blue River fault and east of the range crest, characterized by late Tertiary dates (7–19 Ma), (2) two samples from the southern Gore Range, interpreted to be east of the Blue River fault, with 31–33 Ma dates, (3) six samples from the southwestern Gore Range, on or west of the range crest, that yielded 30–42 Ma dates, and (4) six samples from the northern Gore Range, four of which yielded middle Tertiary dates (32–39 Ma) and two from Gore Canyon with slightly younger results (17–29 Ma). The northern Gore samples were all newly collected for this project. Four of these samples were targeted because of the stratigraphic control provided by the preservation of Mesozoic sedimentary cover in the northern Gore Range, and provide important constraints for our subsequent interpretations of the larger data set. One sample from a siltstone in the Permian–Triassic State Bridge Formation yielded a mean date of 40 Ma but a large range of dates for individual apatite crystals, from 13.6 to 115.5 Ma, that shows no correlation with grain size or eU (effective U). We speculate that the large dispersion in this sample may be partly attributable to heterogeneous zonation in the detrital apatite population, as recent work has shown that strongly variable apatite zonation may add dispersion to a data set without significantly shifting the mean sample date (Ault and Flowers, 2012).
PALEODEPTH RECONSTRUCTIONS
Reconstructing preextensional paleodepths in the Gore Range is key for accurately interpreting the significance of thermal histories inferred from the AHe data set. Only Precambrian crystalline basement is exposed in the southern Gore Range, such that paleodepths cannot be inferred based on observations in this region alone. We therefore exploit the stratigraphic control provided by Mesozoic sedimentary cover preserved in the northern Gore Range to interpret paleodepths for the highest elevations of the crystalline southern Gore Range. We then evaluate viable structural tilts for the southern Gore Range and quantify the displacement magnitude across the main trace of the Blue River fault. Together these steps allow us to convert our age-elevation profile (Fig. 4) to an age-paleodepth profile.
Figure 5A depicts a modern northwest-southeast cross section along the long axis of the range (B–B′ of Fig. 2), and shows the results for the three basement samples from the northern Gore Range and the highest elevation samples of the southwestern Gore Range. The northern Gore samples are 5–200 m below the sedimentary section and yield Late Eocene to Early Oligocene mean AHe dates that are similar to those for samples at high elevations in the southwestern Gore Range. Therefore, the two sets of samples were likely derived from comparable paleodepths, indicating that the high-elevation southwestern Gore Range was close to the base of the sedimentary section prior to dip-slip movement along the cross fault. Secondary faults displace the basement throughout the Gore Range (Kellogg et al., 2011), and likely account for some of the age variability in the southern Gore Range. Thickness estimates of Paleozoic–Mesozoic cover based on strata preserved in the Gore Range region are ∼3.1 km, consisting of an upper ∼2.8 km section of Cretaceous marine shales of the Benton, Niobrara, and Pierre Formations, and a lower ∼300 m of Permian–Cretaceous section still preserved locally in the northern Gore Range (Kellogg et al., 2011). From this we infer basement paleodepths of ∼3.1 km in the northern Gore Range and the high-elevation southwestern Gore Range (Fig. 5B).
Samples in the southeastern Gore Range have dates up to ∼20 m.y. younger than those at equivalent elevations in the western part of the range because of the effects of fault-block tilting (e.g., Fitzgerald et al., 1991; Miller et al., 1999; Fosdick and Colgan, 2008). A 10° tilt angle is the minimum reasonable value because it places the highest eastern samples of Miocene age and the lowest western samples of middle Tertiary age at comparable paleodepths (Supplemental Fig. 33). In order to find a maximum tilt value we compared the southern Gore Range to other flank uplifts of the Rio Grande Rift where tilting is well constrained by sedimentary and volcanic units. Extension in Rio Grande Rift basins decreases northward, with estimates of ∼55% in the Soccorro and La Jencia Basins (Cather et al., 1994), ∼28% in the southern Albuquerque Basin (Russell and Snelson, 1990), ∼17% in the northern Albuquerque Basin (Russell and Snelson, 1990), and 8%–12% in the San Luis Basin (Kluth and Schaftenaar, 1994). The Lemitar Mountains between the Socorro and La Jencia Basins underwent 35°–45° of post–16 Ma tilting (Cather et al., 1994), while the Culebra Range bounding the San Luis Basin was tilted <25° after 15 Ma (Miggins et al., 2002). The continued reduction in graben width in the Upper Arkansas Basin and Blue River Valley indicates still less extension in these northern rift basins. Tilting in the Gore Range should therefore not approach the 25°–45° of stratal tilting in the Lemitar and Culebra Ranges, suggesting that a maximum structural tilt of 20° is reasonable.
In the following discussion (and illustrations) we assume a structural tilt of 15° in the southern Gore Range and use the base of the eroded sedimentary section projected from the western edge of the Gore Range as our paleohorizontal datum (Fig. 6). This places the youngest samples in the data set at ∼2.4 km beneath the base of the sedimentary section and thus at ∼5.5 km paleodepth during the Late Cretaceous (Figs. 6 and 7). Varying the tilt by ±5° yields an uncertainty estimate of ±400 m for paleodepths at the deepest structural levels (Fig. 7). The paleodepth reconstruction also allows an estimate of offset across the Blue River fault to be made using basement sample L09-BL4 in the hanging wall of the fault. This sample is <200 m below the Morrison Formation and yields a mean date of 33.1 ± 2.6 Ma, compatible with dates for samples in the footwall inferred to be similarly proximal to the base of the stratigraphic section. Assuming a 15° tilt and restoring the hanging-wall sample to comparable footwall paleodepths (Fig. 6) indicates ∼2.3 km of fault offset, which is nearly 1 km greater than that inferred from stratigraphic relationships alone.
DISCUSSION
Cenozoic History of the Gore Range
Late Cenozoic Unroofing
The AHe paleodepth profile suggests an episodic cooling history (Fig. 7A), with characteristics similar to previously published AFT data from these same samples (Fig. 7B; Supplemental Fig. 1 [see footnote 1]; Naeser et al., 2002). In the southeastern Gore Range, a profile of four samples from Keller Mountain is characterized by a steep section of 13–20 Ma dates nearly invariant within error over a paleodepth range of ∼1 km, overlapping within uncertainty the previously published AFT dates reported for a subset of these same samples (Figs. 3 and 7B; Naeser et al., 2002). This pattern indicates a rapid cooling episode during the Early to Middle Miocene, likely caused by unroofing associated with normal motion along the Blue River fault. The lowest elevation sample from this vertical transect yields a younger AHe date of 10.2 ± 0.9 Ma and is characterized by the deepest inferred paleodepth of the entire data set (5.5 ± 0.4 km, sample L09-BL2). This result imposes a maximum age of ca. 10 Ma for the most recent timing of major movement along the Blue River fault at this location, and can be used to constrain the minimum unroofing magnitude since Late Miocene time.
We carried out thermal history simulations to address the amount of post–10 Ma unroofing by employing the inverse modeling capabilities of the HeFTy thermal modeling program (Ketcham, 2005) and using the radiation damage accumulation and annealing model for apatite He diffusion kinetics (Flowers et al., 2009). Thermal history simulations end at the modern mean annual temperature of 0 °C, and commence with a Late Cretaceous temperature of 130 °C because AFT data indicate complete annealing of fluorapatites in the Gore Range prior to Tertiary unroofing (Naeser et al., 2002). A precise Late Cretaceous temperature estimate is not important for the results. For each simulation, 10,000 random histories satisfying the constraints were compared against the mean date, eU, and grain size of the sample, with paths that yielded statistically significant “good” and “acceptable” fits to the data considered viable (Ketcham, 2005). The key result is that all good-fit paths indicate that the sample must have been at temperatures >50 °C prior to 10 Ma (dashed lines, Fig. 8A). Applying the modern Gore Range geothermal gradient of ∼35 °C/km (Berkman and Watterson, 2010) implies ∼1.4 km of unroofing along the eastern margin of the southern Gore Range since 10 Ma. A 50 °C /km geotherm, which is the highest sustained over wide areas in the modern Rio Grande Rift in Colorado (Berkman and Watterson, 2010), would require at least 1 km of post–10 Ma unroofing. We consider these results representative of the central part of the southern Gore Range because of this sample’s continuity with the Keller Mountain transect and its overlap in date with another 11.8 Ma sample in the same area. In contrast, the youngest AHe result of the data set (6.9 ± 2.8 Ma, sample GRF1) was yielded by an isolated sample at the far southeastern corner of the range with an estimated paleodepth of 5.4 ± 0.4 km. This younger sample places an upper bound of 7 Ma for the most recent timing of major fault movement at this locality, and using the same thermal history constraints as above requires ≥1.4 km and ≥1.0 km of post–7 Ma unroofing assuming the 35 °C/km and 50 °C/km geotherms, respectively. Although these results clearly have local relevance, this sample’s isolated occurrence leads us to consider the constraints from the Keller Mountain transect as more representative of the history throughout the southern Gore Range. Along-strike variation in the slip history is indicated by these results, but cannot be further resolved without the addition of more along-strike samples to the data set.
In normal fault settings, the flow of groundwater in the shallow crust can lower isotherms over the range crest and raise isotherms around the boundary between the range and the basin (Ehlers, 2005). However, borehole temperature profiles and noble gas measurements of a small watershed in the nearby Front Range indicate that groundwater in the metamorphic bedrock is only circulating and raising the expected conductive geothermal gradient by 1–3 °C within 200 m of the surface, such that below this depth temperature gradients are consistent with expected conductive geotherms (Manning and Caine, 2007). Given the similarity of the Gore Range and the Front Range in both topography and rock type, this suggests that groundwater flow does not greatly distort geotherms in the Gore Range and therefore will not have an impact on our major interpretations of thermochronometry data from the region.
Implications for Middle Tertiary Unroofing and Elevated Heat Flow
The occurrence of middle Tertiary AHe dates at inferred Cretaceous paleodepths from ∼3.1 to ∼3.9 km in the southwestern Gore Range, northern Gore Range, and east of the Blue River fault indicates cooling in middle Tertiary time. This result is compatible with evidence in the Blue River Valley for middle Tertiary removal of Gore Range sedimentary cover. A section of middle Tertiary sedimentary rocks in the Blue River Valley contains a basal boulder conglomerate overlain by mudstone, siltstone, pebbly sandstone, and a 27 Ma tuff; clasts in the conglomerate are made of the granite and migmatitic biotite gneiss found in the surrounding ranges. Above the tuff, the section is composed of shale and siltstone capped by 24 Ma flows of trachyandesite (Naeser et al., 2002; Kellogg et al., 2011). The conglomerate requires that some degree of relief had been generated between the Blue River Valley and a neighboring range prior to 27 Ma. In addition, basalts overlie Proterozoic basement and Paleozoic–Mesozoic strata on Elliot Ridge in the northern Gore Range. Although these basalts are undated, they have been linked to the 24–20 Ma set of basalt flows to the west of the Gore Range in the Yarmony Mountain–State Bridge–Piney Ridge area (Larson et al., 1975; Kunk et al., 2002) and to the 24 Ma basalts in the Blue River Valley (Naeser et al., 2002). This correlation would demand that the rocks now exposed in the northern Gore Range were near the surface by Late Oligocene–Early Miocene time.
To constrain the temperatures required by the middle Tertiary AHe dates, inverse thermal history modeling was performed for northern Gore Range sample L10-ER3, having a mean AHe date of 38.6 ± 3.1 Ma, collected ∼5 m below the base of the sedimentary section and characterized by Cretaceous paleodepths of ∼3.1 km. We use the same approach and similar constraints as described for the southern Gore Range samples, but require that the sample was at near-surface temperatures by 24 Ma, consistent with the age estimate for basalt flows overlying basement in the northern Gore Range. The good-fit paths from the simulation results indicate that the sample must have been at temperatures >50 °C prior to 45 Ma (dashed lines, Fig. 8B), with subsequent cooling to near-surface temperatures by 24 Ma as demanded by the simulation constraints.
We next reconstructed hypothetical paleotemperatures in the Gore Range sedimentary section and basement for Late Cretaceous and Middle Eocene time to quantitatively assess the timing and magnitude of sedimentary cover denudation and associated geotherms necessary to explain Gore Range basement middle Tertiary temperatures (Fig. 9), in an analysis similar to that in House et al. (2003) and Kelley and Chapin (2004). The equation T(z) = T(s) – (qb/k) · z was used, where T(z) is the temperature at depth, T (s) is the mean annual surface temperature, qb is basal heat flow, and k is thermal conductivity. We used the measured thermal conductivity of 1.5 W/mK for the Pierre Shale (Kelley and Chapin, 2004), and assigned values for other formations based on averages for their lithologies (from compilation in Ehlers, 2005). Surface temperatures of 20 °C were applied. Cretaceous reconstructions use the average continental basal heat flow value of 63 mW/m2 (as in House et al., 2003; Kelley and Chapin, 2004). Middle Eocene reconstructions considered both a 63 mW/m2 heat flow value assumed to have remained unchanged since end-Cretaceous time, as well as an elevated basal heat flow value of 110 mW/m2 comparable to heat flow interpreted at the surface of the Gore Range area today (Berkman and Carroll, 2007).
The Late Cretaceous reconstruction is characterized by temperatures >140 °C in the crystalline basement immediately beneath the sedimentary section, consistent with the AHe and AFT thermochronometry data indicating complete resetting of both systems prior to Tertiary cooling (Fig. 9A). The Middle Eocene temperature profile, assuming the same basal heat flow value as in the Cretaceous (63 mW/m2), demonstrates that maintenance of temperatures >50 °C at 45 Ma at the nonconformity as demanded by the AHe data requires preservation of at least 0.9 km of sedimentary section (Fig. 9B). Even assuming the modern heat flow of 110 mW/m2 during Middle Eocene time requires that ≥0.6 km of sedimentary cover remained on the basement until 45 Ma (Fig. 9C). It is clear that maintenance of temperatures >50 °C through 45 Ma in the Gore Range basement precludes the Late Cretaceous unroofing history of most of the Front Range, characterized by denudation to basement depths by the end of the Cretaceous during Laramide tectonism. Rather, part of the sedimentary package, including >300 m of marine shale, must have been preserved above the Gore Range crystalline basement during the Middle Eocene. The denudation of this sequence by 24 Ma, as indicated by basaltic flows likely of that age overlying the northern Gore basement, demands an additional phase of middle Tertiary denudation that is apparently not recorded in most of the adjacent ranges, with the exception of the western Front Range near the Blue River Valley (Naeser et al., 2002). The spatial localization of this middle Tertiary unroofing episode around the Blue River Valley suggests that denudation is associated with the initial phase of Rio Grande Rift activity in Late Oligocene time, as recorded in other rift basins to the south. This might further suggest a contribution to middle Tertiary Gore Range paleotemperatures from elevated heat flow, as interpreted elsewhere in the rift (House et al., 2003; Kelley and Chapin, 2004; Roy et al., 2004). Elevated heat flow in the Gore Range prior to ca. 30 Ma was also interpreted by Naeser et al. (2002) based on the Gore Range AFT data. Although there is a discrepancy in the thermochronometry data for the western Gore Range samples, with AHe results slightly older than the AFT results, this difference would not change the primary interpretations made from these datasets, resulting only in a shift in the inferred timing of the middle Tertiary unroofing episode.
Model for Gore Range Evolution
In Figure 10 we propose a model for evolution of the Gore Range since Late Cretaceous time that integrates the thermochronological results with sedimentary and other geochronological data. The cross-sectional reconstructions are drawn in a southwest-northeast orientation from west of the Gore fault eastward into the Blue River Valley. Evolution of the northern and southern Gore Ranges is similar until at least 24 Ma, such that the 70 Ma, 45 Ma, and 24 Ma reconstructions are applicable to both segments of the range (Figs. 10A–10C). After 24 Ma the histories of the northern and southern Gore Ranges diverge. The white line on the 24 Ma reconstruction represents the modern northern Gore Range profile, showing limited post–24 Ma erosion of the northern Gore Range (Fig. 10C). The subsequent evolution of the southern Gore Range is represented separately by depiction of the modern southern Gore Range cross section in Figure 10D.
In Late Cretaceous time, central Colorado was inundated by the Western Interior Seaway and ∼3.1 km of sedimentary rocks buried Precambrian basement east of the Gore fault. Only the lower 1.1 km of this sequence is shown in Figure 10A. The Gore fault, which was active in Pennsylvanian–Permian time, separates Precambrian basement to the east from central Colorado trough sedimentary units to the west. The Late Permian–Triassic State Bridge Formation was nonconformably deposited over the beveled surface of these older rocks, and thins toward what is now the eastern Gore Range (Kellogg et al., 2011). Formations deposited during the Jurassic and Cretaceous are of a relatively consistent thickness across the region.
During the Laramide orogeny, contractional movement on a reactivated Gore fault caused anticlinal warping of the basement and overlying section along an axis near the western edge of the modern Gore Range. The reconstruction depicts ∼2 km of unroofing by 45 Ma (Fig. 10B), with the entire range still buried by ∼1.1 km of the sedimentary section, including ∼700 m of Cretaceous Pierre Shale. At least 1.8 km of the same shale package remained in the Blue River Valley and eastern central Colorado trough, based on the sedimentary section that is preserved there today (Izett et al., 1971; Kellogg et al., 2011). Only minor relief likely existed across the study area at this time, as dramatic relief would have induced complete denudation of the highly erodable shale sequence. This relief may have been similar to that envisioned for the Rocky Mountain erosion surface during Late Eocene time (Chapin and Kelley, 1997), inferred to have a minimum age of 36.7 Ma based on 40Ar/39Ar dates for the Wall Mountain Tuff overlying the surface in the eastern Front Range (McIntosh and Chapin, 2004).
By the Late Oligocene, most of the remaining sedimentary cover on the Gore Range was unroofed, as indicated by the AHe data and volcanic rocks of probable 24 Ma age overlying both Precambrian basement and Permian–Jurassic formations in the northern Gore Range (Fig. 10C). Precambrian basement clasts in the Blue River Valley boulder conglomerate, which is overlain by a 26.89 ± 0.06 Ma tuff and 23.62 ± 0.6 trachyandesite flow (Naeser et al., 2002), are also consistent with initiation of this Gore Range unroofing episode by 27 Ma. The variable erosion depths into both the basement and Cretaceous units imply that some relief existed within the range. Additional relief between the range and neighboring valleys is indicated by uncomformable relationships of volcanic rocks with the sedimentary units in the region. For example, volcanic rocks dated as 23–24 Ma directly overlie the Cretaceous Niobrara Formation and Pierre Shale to the west of the Gore Range (Larson et al., 1975; Kunk et al., 2001), and basalts dated as 23.62 ± 0.6 Ma and 23.62 ± 0.06 Ma overlie the Jurassic Morrison Formation through the Cretaceous Pierre Shale and middle Tertiary section in the Blue River Valley (Naeser et al., 2002; Kellogg et al., 2011). The middle Tertiary Gore Range denudation phase can in part be explained by initiation of normal movement along the Blue River fault, which in the northern Gore Range locally has a post-Mesozoic offset of no more than ∼200 m (Fig. 2; Kellogg et al., 2011).
The northern Gore Range landscape appears to have been relatively stable since 24 Ma, with perhaps 100–200 m of additional erosion until the present, as shown by the white line depicting the modern northern Gore topographic profile in Figure 10C. In contrast, the higher relief and deeper exposure levels of the southern Gore Range require substantial post–24 Ma unroofing and normal fault movement (Fig. 10D). The eastern edge of the southern Gore Range underwent ∼2.5 km of unroofing since the beginning of the Miocene, coincident with ∼2.1 km of additional displacement along the Blue River fault, as indicated by the AHe data and paleodepth reconstructions. Exhumation rates, and presumably fault movement, accelerated from 20 to 13 Ma. At least 1 km of unroofing and associated fault movement occurred after 10 Ma. Synrift fill was likely deposited in the Blue River Valley during late Tertiary denudation of the Gore Range, on top of the older boulder conglomerate and interbedded 24–27 Ma volcanic rocks. Subsequent establishment of the modern Blue River drainage flowing northward into the Colorado River would have removed most of these synrift deposits, leaving only the 500-m-thick remnant preserved today. Pleistocene glaciations carved additional relief into the southern Gore Range, but did not affect the lower elevation and more subdued topography of the northern segment of the range.
Implications for Rio Grande Rift Propagation Models
Our data clearly indicate that the post-Cretaceous evolution of the Gore Range differed from that of neighboring uplifts in two important ways. First, the AHe data demand that a minimum of 0.6–0.9 km of Mesozoic cover remained on the Gore Range through middle Tertiary time. In contrast, the basement in the adjacent ranges of the Rocky Mountains was likely at or near the surface by the early Tertiary. In most of the Front Range to the east, AFT dates preserve a history exclusively of early Tertiary and older cooling, with Proterozoic clasts shed from the Front Range appearing in the Denver Basin by 68 Ma (Bryant and Naeser, 1980; Naeser et al., 2002; Kelley and Chapin, 2004; Raynolds et al., 2007). To the west of the Gore Range in the White River uplift, and to the north in the Park Range, AFT results indicate major unroofing nearly to the modern surface during Late Cretaceous–early Tertiary time (Naeser et al., 2002; Kelley, 2005). Second, the Gore Range AHe data record multiple middle and late Tertiary unroofing episodes that stripped the sedimentary section and exposed the Precambrian basement. With the exception of the western edge of the Front Range bordering the Blue River Valley, neither of these younger unroofing phases are recorded in ranges to the east, west, or north of the Gore Range (Naeser et al., 2002; Kelley and Chapin, 2004; Kelley, 2005).
Instead, the middle and late Tertiary Gore Range history is consistent with Rio Grande Rift evolution to the south. The middle Tertiary cooling signature in the Gore Range is similar to that identified in rift flank uplifts bounding the major basins of the Rio Grande Rift in southern Colorado and New Mexico (Lindsey et al., 1986; Kelley and Duncan, 1986; Kelley et al., 1992; House et al., 2003; Kelley and Chapin, 2004). Moreover, the Late Oligocene sedimentary section preserved in the Blue River Valley resembles more extensive synrift units of comparable age preserved elsewhere in the rift system. Volcanic rocks interbedded with synrift deposits in the San Luis, Española, Albuquerque, La Jencia, and Socorro Basins suggest rift initiation, basin development, normal fault displacement, and associated relief generation prior to 26 Ma, (Chapin and Cather, 1994; Miggins et al., 2002; Chapin, 2008), entirely consistent with the histories we infer for the Gore Range and Blue River Valley by 27 Ma. The restricted use of the term “Rio Grande Rift” by some to refer only to extensional basins with deep Tertiary fill would exclude basins such as the Blue River Valley north of the Upper Arkansas Basin. However, outcrops similar to the Oligocene Blue River Valley synrift section occur near Climax, Colorado (Tweto, 1979), and basal conglomerates bracketed between 27.3 ± 0.1 Ma and 11.0 ± 0.05 Ma (Cole et al., 2010; Izett and Obradovich, 2001) are present in the Troublesome Formation of Middle Park. These observations suggest that Oligocene fill of this type may once have been continuous from the Upper Arkansas Basin through the Blue River Valley and into Middle Park. Subsequent integration of the Colorado River and Blue River drainages would have denuded most of these synrift deposits, such that the absence of this sedimentary record may be a consequence of erosion rather than lack of deposition. These observations suggest that the Blue River Valley could more appropriately be considered part of the Rio Grande Rift proper.
In addition, the Early to Middle Miocene pulse of cooling along the eastern edge of the Gore Range overlaps with intervals of increased cooling and extension rates recorded in the southern rift basins. There, the stratigraphic record and thermochronologic data indicate that after the Oligocene initiation of rifting, rates of extension increased during the Middle to Late Miocene before slowing again to the present (e.g., Chapin and Cather, 1994). For example, AFT dates from the Sangre de Cristo Range bordering the San Luis Basin indicate rapid cooling ca. 15 Ma (Kelley et al., 1992), and 40Ar/39Ar geochronology from offset volcanic rocks in the San Luis Basin and southern Sangre de Cristo Range reveal the same Middle Miocene unroofing event (Miggins et al., 2002). In the Sandia Mountains flanking the Albuquerque Basin, AFT and AHe thermochronology data record accelerated cooling from 22 to 17 Ma followed by slower cooling to 16 Ma, a short increase in cooling rate at 14 Ma, and then slower cooling to the present (House et al., 2003). In the Gore Range, the decrease in cooling rate after ca. 13 Ma is coeval with the slower extension and cooling rates seen throughout the rift in Late Miocene to Pliocene time (Chapin and Cather, 1994; Lozinsky, 1994; Cather et al., 1994; House et al., 2003).
There is the intriguing possibility that the middle Tertiary thermochronology dates in the Gore Range and western Front Range that differ notably from results in adjacent ranges reflect a contribution from elevated heat flow in and around the Blue River Valley (Naeser et al., 2002). Farther south in the Rocky Mountains and High Plains of southern Colorado and northern New Mexico, thermochronology data for both vertical transects and borehole samples were used to infer heat flux during Oligocene–Early Miocene time that was more than 25 mW/m2 higher than today (House et al., 2003; Kelley and Chapin, 2004; Roy et al., 2004). A link between an increased crustal geotherm and widespread middle Tertiary magmatism was proposed for these areas (e.g., Roy et al., 2004). Although middle Tertiary magmatism was somewhat less voluminous in central Colorado, and near the Gore Range is mostly related to the Colorado Mineral Belt (Chapin, 2012), the distinctive middle Tertiary cooling phase in the Gore Range is compatible with a post-Cretaceous increase in heat flow by middle Tertiary time, as inferred elsewhere in the rift.
The geological observations and thermochronology data establish a middle and late Tertiary history for the Gore Range that is similar to that of the more significant basins of the Rio Grande Rift to the south. The evidence indicates the generally coeval initiation and subsequent evolution of this northern segment of the rift with basins as far south as Socorro, New Mexico, together representing a >700 km segment of this major intracontinental rift system. The results therefore support models such as rotation of the Colorado Plateau for the origin of the rift system. The relationships in turn contradict the idea of a rift that progressively propagated northward during Tertiary time, thus precluding rift propagation as an important contributor to sedimentary unit tilting on the Great Plains and as the cause of the postulated late Tertiary (after 10 Ma) surface uplift of the southern Rocky Mountains (e.g., McMillan et al., 2002; Leonard, 2002; Heller et al., 2003; Duller et al., 2012).
CONCLUSIONS
New AHe thermochronology data better constrain the Cenozoic cooling and unroofing history of the Gore Range, a northern Rio Grande Rift flank uplift in central Colorado. The rugged southern Gore Range has 1.4 km of vertical relief, is bounded on its east side by the Blue River normal fault, and is separated from the more subdued rolling hills of the northern range by a cross fault. In contrast to the entirely crystalline southern range, the northern Gore Range retains a partial cover of Mesozoic sedimentary and middle Tertiary volcanic rocks. Paleodepth reconstructions are enabled by the stratigraphic control imposed by these sedimentary units in the northern Gore Range, the distribution of AHe dates throughout the range, and tilting estimates. The results yield Cretaceous paleodepth estimates of ∼3.1 km for the high-elevation southern Gore Range and ∼5.4 km for the structurally deepest levels, with ∼2.3 km of total normal offset along the Blue River fault.
AHe dates reveal multiple cooling and unroofing episodes during middle Cenozoic and younger time. Middle Tertiary dates for basement samples proximal to the nonconformity demand temperatures >50 °C prior to 45 Ma, requiring the preservation of ≥0.6 km of the Mesozoic sedimentary section through the Middle Eocene. Normal faulting initiated by the Late Oligocene, with an associated phase of unroofing and relief development. The small quantity of Oligocene synrift sediment preserved in the Blue River Valley may have been part of more extensive fill that was later removed by integration of the Colorado River system. Significant Miocene unroofing and Blue River normal fault displacement was restricted to the southern Gore Range, as indicated by the Miocene AHe dates in this area and their absence in the northern range. The cross fault dividing the northern and southern Gore Range accommodated down-to-the-north extension as movement on the Blue River normal fault caused rift flank uplift and unroofing of the remaining sedimentary and volcanic cover in the southern Gore Range. In contrast, the northern Gore Range retained its low-relief character and carapace of late Paleozoic and younger sedimentary and volcanic rocks.
The AHe data demonstrate that the Gore Range history diverged from that of the surrounding ranges from Laramide time onward, indicating that the Gore Range and Blue River Valley originated and evolved as part of the Rio Grande Rift system. Unlike the thermochronology data from nearby ranges that dominantly preserve a significant Late Cretaceous–early Tertiary erosional phase that denuded the basement to near-surface conditions, the AHe data from the Gore Range record younger unroofing episodes that exposed the basement. These middle and late Tertiary unroofing phases are coeval with those of the more southern basins of the rift system in southern Colorado and New Mexico, differing mainly in the magnitude of extension and thicknesses of preserved synrift fill. The results demonstrate synchroneity in the initiation and evolution of much of the Rio Grande Rift system, and show that the idea of northward rift propagation in Cenozoic time is a misconception.
Acknowledgment is made to the Donors of the American Chemical Society Petroleum Research Fund, for support of this research through grant 47476-G8 to Flowers and by student research grants from ExxonMobil, the Geological Society of America, and Amherst College to Landman. We thank Charles Naeser for his generosity in sharing his Gore Range mineral separates, and Karl Kellogg and Craig Jones for their perceptive comments. Helpful reviews by Chris Andronicos and an anonymous reviewer, as well as comments by Terry Pavlis, improved the paper. We thank Ryan Nell, Nathan Rogers, and Andrew Wickert for their help in the field.