Upper Cretaceous sandstones from 17 localities from California to southeastern Alaska (United States) contain unexpectedly large populations of detrital zircons with Proterozoic U-Pb ages, with age peaks at 1800–1650 and 1380 Ma. These peaks are indicative of a sediment source region in the southern part of the Proterozoic Belt Supergoup basin in central Idaho, which hosts 1800–1650 Ma detrital zircons and which was intruded by rift-related 1380 Ma bimodal plutons and sills. Belt rocks were strongly uplifted and eroded during Late Cretaceous Sevier shortening and fed four paleoriver systems. The Lemhi Pass–Hawley Creek river system flowed east and sourced the Beaverhead-Harebell-Pinyon nonmarine megafan in the Cordilleran foreland basin. The Kione River flowed southwest to northern California, where it sourced a very large, ca. 82–80 Ma, ∼600-m-thick delta and submarine fan complex within the northern Great Valley forearc basin. Considerable Kione detritus also transited the forearc basin to reach the Franciscan trench, sourcing a pulse of deposition and subduction accretion in central California and even part of southern California. The Swakane River flowed northwest out of Idaho into Washington, sourcing the protolith for the high-grade Swakane gneiss. More speculatively, a Yakutat River may have flowed northwest and deposited Yakutat strata in a trench off Washington or British Columbia, before those rocks were translated north to southeastern Alaska. Recognition of a major source area in central Idaho for zircons with an uncommon age of 1380 Ma helps constrain the ca. 85–65 Ma paleogeography and paleotectonics of major sectors of the North American convergent margin orogen.


Through much of the Cretaceous, the Franciscan sector of the North American Cordilleran orogen (Fig. 1) comprised a system of active tectonic elements that included the Franciscan subduction complex, Great Valley forearc basin, Sierra Nevada magmatic arc, Sevier retroarc, Sevier foreland thrust belt, and Cordilleran foreland basin. The establishment of links among these coeval elements is valuable for analyzing a wide range of orogenic processes. This paper uses mainly detrital zircon U-Pb age data to trace out four paleorivers that transported detritus eroded from Sevier thrust zones in central Idaho (United States) to depocenters to the east, southwest, and northwest. Paleoriver pathways can serve as a valuable component of paleogeographic and paleotectonic analyses because they can define a former land surface over long distances across a complex orogen, sometimes with good age control.


Most of the data discussed here are from the Franciscan subduction complex, which consists chiefly of highly deformed Cretaceous to Miocene metasedimentary rocks (e.g., Wakabayashi, 2015). Most of the sediments in the subduction complex and in the generally coeval Great Valley forearc basin were sourced from the nearby Sierra Nevada arc and environs. At 53–50 Ma, there was a strong pulse of sediment influx linked to major extension and erosion in the distant Idaho batholith region (Dumitru et al., 2013, 2015). This paper documents an earlier, ca. 85–80 Ma, pulse of sediment influx from Sevier thrust zones in central Idaho.

Sandstone samples from three new and nine previously published Franciscan and Great Valley sites (Table DR1 in the GSA Data Repository1) show a distinctive detrital zircon age distribution, here termed Type E (adding to Types A–D of Dumitru et al., 2013, 2015). Type E distributions contain mostly Cretaceous and Jurassic zircon grains (Fig. 2; Fig. DR1). The youngest zircon populations are ca. 90 to 84 Ma, setting a maximum for sample depositional ages. Type E is distinguished by a “Lemhi doublet” consisting of two unexpectedly strong Proterozoic age peaks at 1800–1650 Ma and 1380 Ma, dual peaks not seen in other Cretaceous Franciscan or Great Valley sandstones. The 1380 Ma peak is remarkably consistent from sample to sample, as shown on concordia plots (Fig. DR2).


The likely source area for the Proterozoic zircons is in central Idaho in the Lemhi subbasin of the Belt Supergroup. Belt strata were deposited from 1470 to 1400 Ma in a continental rift basin system and are as thick as ∼20 km. Of detrital zircons in Lemhi sandstones, ∼70% yield 1825–1625 Ma ages (Fig. 2A) and were sourced from Yavapai-Mazatzal basement in the southwestern United States, and/or from related basement belts that have since been rifted away (e.g., Ross and Villeneuve, 2003; Link et al., 2007; Lewis et al., 2010). In parts of the Lemhi subbasin, strata were intruded by rift-related bimodal 1380 Ma plutons and sills, now commonly represented by rapakivi granite, augen gneiss, and amphibolite (e.g., Evans and Zartman, 1990; Doughty and Chamberlain, 1996; Lewis et al., 2010, p. 1386; Gaschnig et al., 2013). The 1380 Ma intrusives remain poorly understood (Table DR1), largely because the region has been obscured by metamorphism and erosion and by intrusion of the 98–67 Ma Atlanta lobe of the Idaho batholith and 51–43 Ma plutonic and volcanic rocks of the Challis province (Lewis et al., 2010; Gaschnig et al., 2013). The modern outcrop area of the Lehmi subbasin is ∼250 × 100 km and has been exhumed perhaps 5 to 20 km (not all in the Late Cretaceous), and so represented a major sediment source region.

Other potential source areas in Idaho, Oregon, and Nevada cannot explain the Lemhi doublet. The main part of the Belt basin exhibits additional Proterozoic detrital zircon age groups sourced from Australia and/or Siberia, and also lacks any 1380 Ma intrusive rocks (e.g., Ross and Villeneuve, 2003; Lewis et al., 2010, p. 1386 and 1399). Oregon is underlain by accreted terranes with very different zircon signatures. The retroarc region in Nevada also shows very different signatures (see Dumitru et al., 2015, p. 779). Sources in the Sierra Nevada and environs are also unlikely, as the Lemhi doublet is not seen in pre–85 Ma Great Valley and Franciscan sandstones sourced from those areas (e.g., Snow et al., 2010; Sharman et al., 2015; Dumitru et al., 2015).

Another potential source area is Arizona and southern California, which are underlain by extensive ca. 1800–1650 Ma Yavapai-Mazatzal basement, sparse 1450–1400 Ma granitoids, and a diversity of other units (e.g., Sharman et al., 2015, their appendix DR1). However, we have been unable to identify any bodies of ca. 1380 Ma rock in this region that are substantial enough to have sourced the strong, tight 1380 Ma peak shown by most of the Type E samples. In addition, none of ∼21 previous samples from forearc basin strata in southern California exhibit Lemhi doublets, evidence that 1380 Ma sources were minor (Fig. DR3).


This section traces out a ca. 85–80 Ma Kione paleoriver pathway from Idaho to California, as shown in Figure 1.

Sevier Thrust Zones and Cordilleran Foreland Basin

Janecke et al. (2000) mapped structural culminations in the Sevier thrust belt that deformed Belt rocks in central Idaho. This deformation fed the Lemhi Pass–Hawley Creek paleoriver system, which sourced the Beaverhead-Harebell-Pinyon conglomeratic megafan in the foreland basin that extends toward the east into Wyoming. West of the culminations, major, complex deformation along the western Idaho shear zone and Orofino shear zone also uplifted Belt rocks (e.g., McClelland and Oldow, 2007; Stetson-Lee et al., 2013). The precise ages of these deformations are not known, but some interval of this Sevier-related tectonism probably sourced the sediments in California.

Kione River

We propose that a Kione River transported detritus from Idaho to northern California, adding detritus from Nevada, Oregon, and northernmost California. No fluvial deposits of this river are known and they were probably eroded and/or are now buried beneath extensive younger rocks.

Great Valley Forearc Basin

Williams and Graham (2013, p. 2046) used extensive subsurface data from the northern end of the Great Valley forearc basin (GVFB) to characterize a middle package within the Kione and Forbes formations, a package we propose was mainly sourced from Idaho. This package is ca. 82–80 Ma in age, features southward-prograding clinoforms ∼600 m thick reflecting a large delta and slope system, represents a roughly threefold increase in sediment accumulation rates, accomplished final shoaling of the basin, and had an uncertain source. Summarizing a major unpublished industry study, Imperato et al. (1990, p. 71) noted that the source for the Kione-Forbes system is “…somewhat enigmatic [and it] is unlikely that the Sierra Nevada could have supplied the huge volume of sediment and organic matter present…Petrographic studies remain somewhat unclear, [but]…indicate that the sandstone mineralogy of the Forbes Formation…is distinct from the [slightly older] Rumsey petrofacies…implying a source area other than the Sierra Nevada…”.

Figure 2C shows a Type E sample from outcrop in the central GVFB that is probably the same age as the Forbes Formation (Table DR1). Approximately 44% of its dated zircons are Proterozoic, showing that abundant Idaho detritus reached at least this far south.

The Kione River delivered detritus from at least two distinctive sources, quartzofeldspathic Lemhi rocks and the far northern Sierra Nevada arc. The main Sierra Nevada also shed detritus to coastal California. The proportions of detritus from the various sources must have fluctuated over time, so derivative sandstones should exhibit variable strengths of the Lemhi zircon doublet and variable mineralogy. In our Type E samples, the proportion of total zircons in the Lemhi doublet varies from ∼10% to 47% (Table DR1), and we show this here using the notation “Type E (47%).”

Hornbrook Basin

Figure 2D shows three Type E (30%, 15%, 15%) samples from the Hornbrook forearc basin in northernmost California, one known to be deposited ca. 75–71 Ma (Surpless, 2015). They were apparently partly sourced from Idaho and might indicate that the Kione River shifted ca. 80–75 Ma from feeding the GVFB to feeding the Hornbrook basin.

Main Franciscan Subduction Complex

Seven samples from the Franciscan Complex in central and northern California are Type E (10%–47%) (Wakabayashi, 2015, p. 685; our data). Some of these are from the Franciscan Novato Quarry terrane, which has yielded ca. 85–83 Ma fossils from a few localities (Table DR1).

Franciscan structure is extremely complex with poor age control, so the definition, correlation, and dating of units have been vexing problems. The Type E localities represent a new widespread unit that can be used to help reconstruct parts of the complex’s depositional and accretionary history, as well as postaccretionary displacements along major faults (e.g., Wakabayashi, 2015, p. 685–686; see also Table DR1).

In modern subduction zones, the rate of sediment supply reaching the trench is often the dominant control on the rate of subduction accretion and thus on the overall growth history of the subduction complex. The Type E pulse of accelerated accretion here is an additional illustration of the importance of episodic increases in sediment input from tectonically active, often distant, source areas in controlling the growth and evolution of subduction complexes (e.g., Dumitru et al., 2013, 2015).

Schist of Sierra de Salinas

In southern California an oceanic plateau subducted in the Late Cretaceous, causing a major subduction erosion event that removed the deeper levels of the magmatic arc. The Rand, Pelona, Orocopia, and Sierra de Salinas schist units represent trench sediments that were then underplated beneath the remaining arc rocks (e.g., Barth et al., 2003; Jacobson et al., 2011). The schist of Sierra de Salinas (SSdS) apparently accreted in southern California, before Cenozoic translation ∼390 km northwest as part of the Salinian block. Remarkably, three of five samples from the SSdS are Type E (18%, 22%, 34%). Long-distance turbidite flow of sediment down the trench axis is well documented in some modern subduction zones. Perhaps sediment supply in northern California was sufficiently strong that Idaho detritus could flow downslope along the trench axis to southern California to partly source the SSdS.

Nacimiento Block

The Nacimiento block (NB) is composed of Franciscan and Great Valley–like rocks that were juxtaposed with arc rocks of the Salinian block ca. 75–60 Ma. This juxtaposition involved either ∼150 km of thrust offset, or 500–900 km of left-lateral strike-slip offset (e.g., Jacobson et al., 2011). NB detrital zircon ages generally match southern California source areas and so favor thrust models in which the NB resided in southern California during deposition (Chapman et al., 2015). However, samples from Stanley Mountain, the Cambria slab, and the Pfeiffer slab are Type E (30%, 16%, 20%). The two slabs have been interpreted as trench-slope basins (rather than trench-axis deposits; see Table DR1), so the slab data would seem to favor strike-slip models where the NB resided in central California ca. 85–75 Ma and so could receive detritus from Idaho.

Age Range

The age range of the Kione pathway is not precisely constrained. The best evidence is the ca. 82–80 Ma age of the middle Kione-Forbes system in the GVFB. Very limited fossil data indicate that part of the Novato Quarry terrane is ca. 85–83 Ma and the Cambria slab is Campanian (82–71 Ma; Table DR1). We tentatively interpret that an episode of strong deformation occurred in Idaho ca. 85–80 Ma to generate a rapid influx of sediment into the Great Valley and Franciscan depocenters.


In central Washington, the Swakane Biotite Gneiss is also Type E (∼14%). Its protolith was deposited ca. 91–75 Ma, then metamorphosed at 640–740 °C and 35–40 km depth (Gatewood and Stowell, 2012). Apparently, a northwest-flowing Swakane River supplied detritus from nearby Lemhi rocks.

The Eocene Tyee forearc basin in Oregon exhibits a variant of the Lemhi doublet (30%), with proportionally more zircons between 1650 and 1380 Ma (Fig. DR4). This variant probably represent a mixture of detritus from the Lemhi subbasin (∼25%) plus the main Belt basin (∼75%), that was delivered by the Eocene Tyee River (Dumitru et al., 2013, 2015).

A sample from the Yakutat segment of the vast Chugach subduction complex in Alaska exhibits a Proterozoic pattern (∼14%) identical to the Tyee variant. Garver and Davidson (2015) proposed that ca. 71–65 Ma Yakutat strata were deposited in southern California, then translated ∼3200 km north to southeastern Alaska, a strong positive test for the Baja British Columbia translation hypothesis. However, Belt sources are an excellent match to Yakutat ages, suggesting that a Yakutat River could instead have carried detritus northwest out of Idaho and deposited it in a trench off Washington or British Columbia. That would represent a negative test for Baja–British Columbia (Baja-BC), but only as applied to the Yakutat strata, which have not been central to the Baja-BC debate (Fig. DR4).


Through a confluence of circumstances, Lemhi rocks in central Idaho have generated an unusually distinctive detrital zircon signature (Fig. 2). These circumstances include a narrow age range and limited geographic extent for the 1380 Ma plutonic rocks, the fact that ca. 1380 Ma plutonic rocks are uncommon elsewhere in the western United States, the presence of 1800–1650 Ma zircons to form a diagnostic second age peak, and the strong Sevier deformation and erosion in the area. This in turn has permitted the tracing out of four major paleoriver pathways, and through that, the improved reconstruction of the Late Cretaceous paleogeography of the Cordilleran orogen presented in Figure 1.

These results in turn provide tools to contribute to future investigations of a diverse range of important questions in Cordilleran tectonics. Some examples include the nature of the rift-related 1380 Ma plutonic rocks (which seem to have been more voluminous than currently appreciated), the relationships (if any) of the apparent lowland along the Kione River to the Late Cretaceous elevated Nevadaplano and to the late Cenozoic Snake River Plain, the use of Type E sandstones as markers for constraining both episodic subduction accretion in the Franciscan Complex and Cenozoic motions on the San Andreas and Nacimiento fault systems, tests of the Baja–BC hypothesis, and interpretation of source and reservoir strata in the Kione-Forbes natural gas district.

We are grateful to Andrew Barth, Marty Grove, Raymond Ingersoll, Carl Jacobson, Robert McLaughlin, and Elizabeth Miller for discussions and/or reviews. Partial support came from National Science Foundation grants EAR-0948676 to Grove and EAR-PF1250070 to Chapman; grant EAR-1032156 supported analyses at the Arizona LaserChron Center.

1GSA Data Repository item 2016019, U-Pb sample details, data, and plots, is available online at www.geosociety.org/pubs/ft2016.htm, or on request from editing@geosociety.org or Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301, USA.