This study focuses on Paleozoic rocks of the Turtleback Complex (TBC) and East Sound Group (ESG) of northwestern Washington. These assemblages record a long and complex history of magmatism, sedimentation, and volcanism as part of the Chilliwack composite terrane. This study investigates the tectonic setting and magmatic evolution of the Chilliwack terrane by analyzing U-Pb geochronologic and Hf isotopic data from zircons extracted from igneous and sedimentary rocks of the TBC and ESG. Igneous rocks of the TBC yield U-Pb ages of ~420-~370 Ma for zircon rims, which are interpreted to record pluton crystallization, and ages of ~500-370 Ma for inherited cores and antecrysts. Zircons older than 410-400 Ma yield typical igneous values of U/Th and U concentration and somewhat more evolved εHft values (+5 to +10), whereas younger grains yield values of U/Th and U concentrations that record ongoing metamorphism and have more juvenile εHft values of +10 to +15. This evolution is interpreted to have occurred in a long-lived island arc that experienced a transition at ~410-400 Ma from crustal thickening to regional crustal extension. Detrital zircon grains from the ESG yield dominant ages of ~421-~366 Ma and εHft values of +8 to +15, both of which overlap with values from igneous rocks of the underlying TBC. Strata of the ESG also contain abundant older zircon grains, with a dominant age peak of ~1860 Ma and subordinate peaks at 2673, 2570, 1485, 1113, 736, and 553 Ma. When compared with similar results from other assemblages and terranes in the northern Cordillera, our results suggest possible connections with the Yellow Aster Complex of the nearby North Cascade Range. Our results also suggest general similarities with the Alexander terrane, Wrangellia terrane, and southeast Alaska portion of the Yukon-Tanana terrane. These similarities are consistent with previous interpretations that the Chilliwack composite terrane formed in association with the Alexander, Wrangellia, and southeast Alaska portion of the Yukon-Tanana terrane along the paleo-Arctic margin of Laurentia during early Paleozoic time. Our work also illustrates the importance of recycling of older magmatic rocks over millions of years within a juvenile island arc system.

The San Juan Islands (SJI) of northwestern Washington are part of a collage of terranes that underlie much of the North American Cordillera and extend from California to Alaska (Figure 1) [13]. Rocks exposed on these islands occupy a critical position in the Cordillera, separating rocks of the Insular superterrane (Wrangellia and Alexander terranes) to the west from rocks of the Intermontane superterrane (Stikine and Yukon-Tanana terranes) to the east [46]. One of the outstanding questions about the geology of the San Juan Islands is whether the oldest rocks, of early to mid-Paleozoic age, formed as part of the inboard Intermontane superterrane or the outboard Insular superterrane. This distinction is significant because Paleozoic rocks of the Intermontane terrane formed and evolved along the western margin of Laurentia (the North American continent), whereas rocks of the Insular terrane are thought to have originated in the paleo-Arctic ocean basin, perhaps in proximity to Baltica, the Caledonian orogen, or Siberia [7]. Establishing connections with inboard versus outboard terranes is also important for understanding the Mesozoic evolution of the Cordillera, as Sigloch and Mihalynuk [8, 9] and Clennett et al. [10] suggest that an ocean basin separated the Insular and Intermontane terranes until Cretaceous time, but it is not clear whether this cryptic suture traces inboard or outboard of the San Juan Islands. This study utilizes U-Th-Pb geochronology and Lu-Hf isotope geochemistry of early-mid Paleozoic rocks of the San Juan Islands in an effort to place better constraints on the tectonic evolution of the northern Cordillera during Paleozoic and Mesozoic time.

The San Juan Islands consist of four nappes that are stacked in a northwest-vergent thrust system ([11]; Figure 2). The nappes are several kilometers thick, up to one hundred kilometers in width, and consist, from lowest to highest, of

  • (1)

    Devonian through Permian island-arc volcanic and sedimentary rocks of the East Sound Group

  • (2)

    Ordovician through Devonian arc basement plutonic rocks of the Turtleback Complex and overlying(?) volcanic and subordinate sedimentary rocks of the East Sound Group

  • (3)

    Permian through Jurassic ribbon chert and pillow basalt of the Orcas Chert and Deadman Bay Volcanics

  • (4)

    Jurassic and Cretaceous pillow basalt, ribbon chert, and graywacke that belong to a variety of ocean floor assemblages

Our study focuses on the Turtleback Complex (TBC) and the East Sound Group (ESG), which are exposed in the northwestern San Juan Islands (Figure 3). An apparent continuation of this nappe is located in the Northern Cascade Mountains and is represented by rocks of the Yellow Aster Complex (YAC) and Chilliwack Group (CH) (Figure 2). The TBC and ESG are interpreted to be related to these rocks based on structural relationships and the occurrence of similar ages, lithologies, and oceanic and continental fauna in both localities [1, 12].

The TBC and the ESG are interpreted to have evolved in an oceanic arc setting during Silurian to Permian time [1, 2, 1114]. These rocks were accreted to the western edge of Laurentia in a subduction zone setting during Mesozoic time, but the location of accretion is not well constrained. Following initial accretion, the nappes were emplaced, exhumed during oblique plate convergence, and finally displaced northward to their present setting between about 110 Ma and 93 Ma [12, 1518].

The TBC consists of a variety of small intrusive bodies that are moderately deformed and of greenschist facies [11]. Compositions range from gabbro to leucotonalite with dominant compositions of diorite, tonalite, and leucotonalite. Igneous rocks of the TBC generally consist of many small plutons, dikes, and sills that form an injection complex. There are some locations where mafic rocks intrude felsic rocks, and at other locations felsic rocks intrude mafic rocks.

The ESG consists mainly of sedimentary and volcanic rocks with low-grade metamorphism [11]. Most abundant are volcanic rocks, volcanic-derived sedimentary rocks, turbidites, and limestone. Also present is a distinctive sedimentary breccia that consists of angular to well-rounded blocks of vesicular basalt, tonalite, argillite, sandstone, and pillow basalt. In some locations, the limestone forms large, discontinuous blocks surrounded by basalt. Turbidites of the ESG consist of sandstone, siltstone, mudstone, and shale.

Sedimentary rocks in the ESG are locally intruded by igneous rocks that are interpreted to belong to the TBC. Igneous rocks of felsic to intermediate composition occur as dikes and flows where they are associated with sedimentary rocks of the ESG [1]. The contact between the ESG and TBC is interpreted to be depositional, with strata of the ESG resting conformably on the TBC, and locally intrusive [1, 12]. This relationship is difficult to demonstrate, however, because of widespread brittle deformation.

The initial geochronologic study of rocks of the TBC was conducted by Mattinson [19], who reported U-Pb results from one multigrain zircon fraction each from two samples of quartz diorite. The analyses were highly discordant, with 206Pb/238U ages of 357 and 379 Ma and 206Pb/207Pb ages of 420 and 448 Ma. Discordance was interpreted to be due to Pb loss, and a crystallization age of ~460 Ma was inferred. Whetten et al. [14, 20] reported U-Pb results of multigrain zircon fractions, but the results were so highly discordant that no age interpretation was possible. Whetten et al. [14] also reported a K-Ar age of ~554 from a pegmatitic gabbro.

More recently, Brown et al. [1] reported U-Pb ages of 417.9±5.4Ma (TB-1), 374.6±5.0Ma (TB-2), and 388.9±4.6Ma (TB-3) from three different plutonic rocks of the TBC (Figure 3). These ages demonstrate that the TBC is significantly younger than the ~460 Ma age reported by Mattinson [19] and the ~554 Ma K-Ar age reported by Whetten et al. [14]. This study also reported the additional presence of 510-440 Ma ages from zircon grains in two of their three samples. These older ages were interpreted to record either (1) inclusion of igneous rocks of several different ages in the samples or (2) inheritance of slightly older material in some zircon grains.

Brown et al. [1] also reported U-Pb ages of detrital zircons from a sample of Devonian sandstone from the ESG, which was collected from O’Neal Island (ES-1; Figure 3). The lithology sampled is a quartzose turbidite from a sequence of interbedded sandstone and argillite. Most grains are Precambrian in age, with peak ages of 1856 and 2671 Ma. There also are smaller groups of younger grains with peak ages of 553, 429, 420, and 368 Ma [1].

U-Pb geochronologic and Hf isotopic analyses were conducted on ten samples as part of this study (Table 1). Most samples are from shoreline exposures on Orcas, Jones, and O’Neal Islands (Figure 3), where details of contact relations are well exposed due to wave action. Samples TB-6, TB-11, and TB-17 were collected from different compositional members of the TBC. We also present additional analyses from samples TB-1 and TB-3, which were collected and analyzed by Brown et al. [1], and samples BM 92-41 and BM 92-57, which were collected by Bill McClelland. Samples TB-8 and TB-9 were collected from sedimentary rocks of the ESG for detrital zircon analysis (Figure 3). We also present additional data from sample ES-1, which was collected and analyzed by Brown et al. [1]. Locations of all samples are shown on Figure 3, and representative rock types of most samples are shown on Figure 4.

4.1. Igneous Samples

TB-1 is a quartzose tonalite from Ship Peak in the Turtleback Range on Orcas Island (Figure 3) that was originally analyzed for U-Pb geochronology by Brown et al. [1]. TB-2 is from a gabbro pegmatite that intrudes gabbro on the east side of East Sound, which was analyzed for U-Pb geochronology by Brown et al. [1] and for K-Ar by Whetten et al. [14]. TB-3 is from a tonalite body on the northern end of O’Neal Island, which was analyzed by Brown et al. [1].

TB-6 is from the east shore of Deer Harbor (Figure 3). The locality is within a complex of dikes and small intrusive bodies of tonalite, quartz diorite, fine-grained gabbro, and basalt. Sample TB-6 is from a foliated tonalite dike that intrudes a strongly foliated diorite body (Figure 4(a)).

TB-11 and TB-17 were collected from compositionally homogeneous, coarse-grained, leucotonalite bodies with large (3-8 mm) quartz and plagioclase phenocrysts. TB-11 contains abundant muscovite and was collected from Deer Harbor (Figures 3 and 4(b)). TB-17 was collected from West Sound (Figures 3 and 4(c)).

BM 92-41 is a massive, coarse-grained, plagioclase-porphyritic tonalite that was collected from the same locality on Orcas Island as geochronology sample #9 of Mattinson [19]. BM 92-57 is a massive, leucocratic quartz- and feldspar-bearing dike that intrudes fine-grained tonalite on Orcas Island.

4.2. Detrital Samples

TB-8 consists of greyish-tan, medium-grained sandstone that was collected from the east side of Jones Island (Figure 3). The outcrop consists mainly of 20 cm thick limestone layers with subordinate 10 cm thick layers of sandstone (Figure 4(d)). The sequence is intruded by numerous basalt dikes. The sandstone consists largely of mafic volcanic lithic grains of variable size (1-5 mm) and includes ~30% angular quartz grains. The limestone is reasonably interpreted as reefoidal and locally has yielded Devonian fossils [11, 21, 22].

TB-9 (Figures 3 and 4(e)) is a coarse-grained sandstone collected from northwestern Orcas Island. The sequence at this locality consists of 10-20 cm thick graded beds of sandstone interlayered with finer-grained clastic strata and fragmental mafic volcanic horizons. The sandstone consists of rounded to angular lithic grains of sandstone and mafic to felsic volcanic rocks as well as ~20% angular quartz grains.

ES-1 is a quartzose turbidite that consists of chert (50%), monocrystalline quartz (~25%), siltstone (~20%), and volcanic lithic grains (~5%) [1]. At this locality on eastern O’Neal Island (Figure 3), the ESG consists of interbedded sandstone, argillite, and conglomerate (Figure 4(f)) that are intruded by dikes and sills of igneous rocks of felsic to intermediate composition. These rocks were interpreted by Danner [13, 22] and Brown et al. [1] as resting unconformably on igneous rocks of the TBC.

5.1. Sample Processing

Zircons were separated at the Arizona LaserChron Center (ALC) by traditional methods of crushing and grinding followed by separation using a water table, Frantz magnetic separator, and heavy liquids (https://www.laserchron.org). In spite of careful processing, most samples yielded only a small number of zircon grains, most of which were of relatively small size (e.g., <150 microns in length). We accordingly processed all sample materials a second time, but few additional grains were recovered. As described below, the low yield of relatively small zircon grains posed challenges in generating and interpreting the U-Pb and Lu-Hf data.

Grains were mounted in a 1 epoxy mount along with fragments or grains of various standard reference materials. Mounts were polished to a 1-micron finish and imaged using a Hitachi back-scatter electron detector (BSE) and Gatan cathodoluminescence (CL) detector. CL and BSE images were used to select analytical points in an attempt to avoid inclusions and fractures. The images also provided information about grain textures, morphology, internal zoning patterns, and variations in CL color response. These images revealed that zircons from igneous rocks of the TBC are quite complex. As shown in Figure 5 (and in detail in Supplementary Files 17), nearly all grains contain distinct cores and rims, and some samples contain domains with very high U concentrations (dark shade in CL images). Mounts were cleaned with a 2% HNO3 and 1% HCl solution prior to isotopic analysis.

U-Pb geochronologic analyses were conducted by laser ablation multicollector inductively coupled plasma mass spectrometry (LA-ICPMS). Analyses were calibrated using an in-house Sri Lanka zircon as the primary standard and R33 as the secondary standard. The analytical methods used are described by Gehrels et al. [23] and Gehrels and Pecha [24], with details of the analytical parameters described in Supplementary Table 1.

Data reduction and reporting protocols utilized in this study follow the recommendations of Horstwood et al. [25], with details presented by Gehrels et al. [23] and Gehrels and Pecha [24]. All measured isotope ratios and calculated ages (and their uncertainties) for both unknowns and standards are reported in Supplementary Table 2. The reported uncertainties include contributions from the measurement of 206Pb/238U, 206Pb/207Pb, and 206Pb/204Pb. Also reported are external (systematic) uncertainties that include contributions from the in-run analyses of standards, assumed ages of the standards, assumed composition of common Pb (from [26]), and decay constants for 238U and 235U.

5.2. U-Pb Analysis of Igneous Samples

Igneous samples were analyzed during several different sessions. The first session consisted of one analysis per zircon grain, with the analysis pit located in either a core or a rim of a zircon grain where possible. In many cases, however, due to the small size of the grains and the complexity of age zonation, analysis pits commonly overlapped both core and rim materials. These analyses were conducted with a 30-micron laser beam diameter to clean the analysis spot on the polished surface of the crystal and a 20-micron laser beam diameter for the analysis. Each analysis consisted of 12 s with the laser firing and 20 s with the laser off (to measure background counts and for signal washout) and utilized faraday collectors to measure 206Pb, 207Pb, 208Pb, 232Th, and 238U and ion counters to measure 202Hg and 204(Hg+Pb). The ablation pit was roughly 12 microns deep.

Unfortunately, due to the widespread age zonation, the first set of CL images and geochronologic analyses was not sufficient to determine separate ages for cores and rims of the samples. We accordingly acquired higher-resolution CL images of the largest and highest quality grains from each sample and conducted additional sets of analyses utilizing a smaller laser beam and multiple analyses of each grain. These analyses were conducted with a 12- or 15-micron beam diameter and pits of ~6 microns depth. Faraday collectors were used for 232Th and 238U and ion counters were used to measure 204Pb, 206Pb, 207Pb, and 208Pb.

All resulting ages are reported in Supplementary Table 2, with core and rim analyses reported separately (where possible). All analyses are reported in Supplementary Table 2, but the first set of analyses is not used for age interpretations because they could not be confidently assigned to cores or rims. As recommended by Horstwood et al. [25], uncertainties of the individual analyses include only internal (measurement or random) uncertainty components, whereas uncertainties of final (pooled) ages include the quadric sum of the internal and external uncertainties. Four different methods are used to analyze the distributions of ages and uncertainties.

  • (1)

    The weighted mean function of Isoplot [27] weighs each analysis according to the square of its uncertainty and reports the resulting weighted mean and uncertainty (as standard error of the mean). Analyses that are statistical outliers are excluded from the calculation. This function also generates an MSWD (Mean Square of the Weighted Deviates), which compares the degree to which the analytical uncertainties are consistent with the scatter of ages

  • (2)

    Distributions of U-Pb ages from each igneous sample are also shown on a relative age-probability plot (or probability density plot (PDP)), which sums all ages and uncertainties into a continuous probability distribution [27]. For comparison of numerous samples, the distributions are normalized (divided by the number of analyses) such that each curve contains the same area. For samples with high (>1) MSWD, relative age-probability curves provide a means of evaluating whether the distribution contains more than one age population. For example, the presence of two separate peaks in an age distribution suggests the presence of two populations, with ages that may be represented by the probability maxima. In such a situation, the weighted mean age would be between the two age components, with large uncertainty and high MSWD

  • (3)

    The TuffZirc age is based on the age extractor function in Isoplot [27]. This routine identifies the largest cluster of ages in a distribution that overlap to an acceptable degree (probability-of-fit>0.05), reports the median value as the most likely age, and uses the range of included ages to calculate an asymmetric uncertainty of the median value. Excluded ages are interpreted to predate the selected cluster (if older) or to be compromised by Pb loss (if younger)

  • (4)

    For pooled ages, we use the Unmix function of Isoplot [27] to calculate the maximum likelihood age and uncertainty of each age group. This method uses a maximum likelihood analysis to determine the Gaussian distribution that best fits each age group

The likelihood that grains within an age group may be genetically related is evaluated by comparing the U concentrations and U/Th values of the constituent grains. Useful guides in comparing these values are provided by Rubatto [28], who report that most igneous zircon has U/Th values<10, whereas metamorphic zircon commonly yields U/Th values>10. Plots of U concentration and U/Th versus age are provided for each data set in Supplementary Table 2.

5.3. U-Pb Analysis of Detrital Samples

Detrital samples were analyzed during a single session utilizing a 20-micron laser beam and faraday collectors. CL images were used to locate analysis pits in homogeneous domains in central portions of zircon crystals. For two samples, TB-8 and TB-9, all available grains were analyzed. For sample ES-1, we analyzed 191 different crystals. Age distributions are shown with relative age-distribution (or probability density) plots. The maximum depositional age (MDA) for each sample is assigned based on the youngest significant U/Pb age peak.

5.4. Lu-Hf Analysis

Lu-Hf isotopic analyses were conducted by laser ablation multicollector inductively coupled plasma mass spectrometry (LA-ICPMS) at the ALC utilizing methods described by Cecil et al. [29], Gehrels and Pecha [24], and Gehrels [30]. Details of the analytical parameters are described in Supplementary Table 1. Hf isotope data were reduced using an in-house data reduction system (HFcalc) that has been described by Cecil et al. [29] and Gehrels and Pecha [24]. This system corrects for interferences of 176Yb and 176Lu using the method of Woodhead et al. [31] and isotope compositions reported by Vervoort et al. [32] and Patchett [33]. The 176Hf/177Hf at time of crystallization is calculated from measurement of present-day 176Hf/177Hf and 176Lu/177Hf, using the decay constant of 176Lu (λ=1.867e11) from Scherer et al. [34] and Söderlund et al. [35]. Standards used for calibration include zircon standards Mud Tank, 91500, R33, Temora2, Plesovice, Sri Lanka, and FC-1. Generating and interpreting the Lu-Hf data from our samples were challenging because of the low number of available grains from most samples, their small to medium (<150 μm) size, and the ubiquitous age zonation (e.g., Figure 5). One strategy used to address these complexities was to place the Lu-HF analysis pit on top of the U-Pb pit (e.g., [24]), thereby increasing the likelihood that the U-Pb age is relevant for the Hf isotopic information. Unfortunately, as shown in Figure 5, in many cases the U-Pb and/or Lu-Hf analysis pit overlapped multiple age domains, in which case it is uncertain whether the isotopic information is relevant for the rim or the core of the grain. Because of this uncertainty, the measured U-Pb age, rather than the pooled age of the analyzed material (rim or core), is used to reduce and display the Lu-Hf isotopic information for each analysis.

A second strategy was to conduct the Lu-Hf analyses with low laser energy in an effort to reduce the likelihood that an age boundary was encountered at depth during an analysis. This appears to have been somewhat effective in reducing complexities from zonation, although variations in Lu-Hf isotope intensities and ratios were observed during many analyses. The downside of this strategy, however, was that it resulted in analyses with low ion yields (average of 2.8 V total Hf) and therefore analyses with relatively low precision (average uncertainty of 2.2 epsilon Hf units, at 2-sigma).

A third strategy was to analyze every grain of sufficient size and quality from every sample in an effort to generate as much information as possible and to analyze a large number of standards to provide a reliable calibration of the measurement of unknowns. As shown in Supplementary Table 3, this resulted in a variable number of analyses from each sample (e.g., 49 from an igneous sample but only 11 from a detrital sample) and a large number of standard measurements (n=534) relative to unknown measurements (n=182).

Resulting data for both unknowns and standards are presented in Supplementary Table 3, following the data reporting protocols recommended by Fisher et al. [36]. In describing the results, we follow Vervoort and Patchett [37] in referring to more negative εHft values as evolved and more positive values as juvenile.

This section summarizes the geochronologic results presented by Brown et al. [1] and presents the results of our new U-Pb and Hf isotope analyses. For igneous samples, final ages are interpreted from consideration of the weighted mean ages, peak ages on probability distribution plots, and TuffZirc ages, as well as CL textures, U concentrations, and U/Th values. Age plots are shown on Figures 68; details of the U-Pb isotopic analyses are presented in Supplementary Table 2. U concentrations and U/Th values for the igneous samples are shown in Figure 9. Figure 10 presents a summary of the igneous ages. The results of Lu-Hf analyses of the igneous samples are shown on Figure 11; details are presented in Supplementary Table 3. For detrital samples, U-Pb and Lu-Hf results are shown on Figure 12. CL images and analysis locations for igneous samples are presented in Supplementary Files 17.

6.1. Igneous Samples

TB-1: Brown et al. [1] reported 206Pb/238U ages ranging from ~507 to ~406 Ma for sample TB-1 and offered an interpreted age of 417.9±5.4Ma (MSWD=2.7) based on twenty-two overlapping analyses (Supplementary Table 2). Because of the high MSWD, Brown et al. [1] suggested that the sample may contain multiple igneous components or that some zircon grains contain inherited components. The latter interpretation was preferred, and the age of 417.9±5.4Ma was reported on the basis of the weighted mean of the youngest set of ages.

In an effort to refine the interpretations of crystallization and inherited ages from this sample, we conducted two additional sets of analyses (as described above). The CL images from this sample, and the locations of core and rim analyses, are shown in Supplementary File 1.

Our new data from TB-1 consist of 230 U-Pb analyses, 40 utilizing faraday collectors and a larger beam diameter, and 190 utilizing ion counters, a smaller beam diameter, and multiple analyses per grain (Supplementary Table 2). The latter set of analyses from rims yields a PDP peak age of 394.7 Ma, a TuffZirc age of 396.1 Ma, and a weighted mean age of 394.0±1.8Ma (2-sigma; MSWD of 1.1) (Figure 6(a)). The low MSWD of this set of ages suggests that 394.0±4.2Ma is a reliable crystallization age for sample TB-1. Cores from this sample yield ages that range from >480 Ma to <395 Ma, with a peak age of 411.2 Ma, TuffZirc age of 410.9 Ma, and weighted mean age of 415.4±1.7Ma (MSWD of 1.9) (Figure 6(b)). The broad range of ages with high MSWD value, combined with the significant difference between core and rim ages (Figure 10), suggests that the cores represent inherited components which are mainly in the 420-400 Ma range, but with some grains as old as ~480 Ma. U concentrations and U/Th values of both cores and rims are typical for igneous rocks (Figure 9).

Lu-Hf analyses were conducted on eighteen grains from TB-1. Resulting εHft values are between +8.0 and +15.3 (Figure 11; Supplementary Table 3). As shown on Figure 11, core analyses yield slightly lower εHft values than rims.

TB-2: sample TB-2 was analyzed by Brown et al. [1], who reported a cluster of analyses with a weighted mean age of 374.6±5.0Ma (MSWD=0.9), as well as several younger ages compromised by Pb loss and older ages that record inheritance of older grains. The age of 374.6±5.0Ma was interpreted as the crystallization age. We have not conducted additional analyses on this sample.

TB-3: for sample TB-3, Brown et al. [1] reported that analyses from 26 grains yield ages ranging from ~399 to ~376 Ma, with a weighted mean age of 388.9±4.6Ma and an MSWD of 2.1. We conducted an additional 309 analyses, 216 with larger analysis pits and 49 analyses of rims and 44 analyses of cores utilizing smaller analysis pits (Supplementary Table 2). The latter are shown on CL images in Supplementary File 2.

Rim analyses from this sample yield a PDP peak age of 386.3 Ma, a TuffZirc age of 385.4 Ma, and a weighted mean age of 386.3±1.3Ma with MSWD of 0.8 (Figure 6(c)). The low MSWD of this set of ages suggests that 386.3±4.4Ma is a reliable crystallization age for sample TB-3. Core ages from this sample overlap with and slightly predate rim ages, with a PDP peak age of 395.7 Ma, a TuffZirc age of 394.3 Ma, and a weighted mean age of 394.4±1.5Ma (MSWD of 1.1) (Figure 5(d)). The low MSWD value for the core ages suggests that most grains belong to a dominant age group of 394.4±4.4Ma, although some grains are somewhat older (Figure 6(d)). U concentrations and U/Th values of the core analyses are slightly higher than rim analyses (Figure 9), but still suggestive of an igneous origin.

Forty-nine grains from sample TB-3 were analyzed for Hf, with εHft values between +10.7 and +16.1 (Figure 11; Supplementary Table 3).

TB-6: analyses from sample TB-6 include 18 conducted with a larger beam diameter, as well as 26 from rims and 57 from cores using a smaller laser beam (Supplementary Table 2). The latter are shown on CL images in Supplementary File 3. Rim analyses yield a PDP peak age of 383.8 Ma, a TuffZirc age of 386.5 Ma, and a weighted mean age of 385.2±2.4Ma (MSWD of 0.8), which we interpret as the crystallization age (Figure 6(e)). Cores yield mostly older ages, with some as old as ~438 Ma, PDP peak ages of 422.7 and 399.1 Ma, a TuffZirc age of 398.2, and a weighted mean age of 407.2±3.5Ma (MSWD=6.1) (Figure 6(f)). The broad distribution of ages, presence of two separate PDP peak ages, and MSWD of 6.1 all suggest that the cores represent inherited zircons with a broad range of ages. The two PDP peaks suggest dominant ages of ~423 and ~399 Ma for the inherited components. The U concentrations and U/Th values for cores and rims of all analyses are typical for igneous zircon. εHft values for thirteen of these grains range from +7.0 to +12.4 (Figure 11; Supplementary Table 3).

TB-11: because the zircon grains for sample TB-11 are small and contain very thin rims, all analyses were conducted utilizing a 15-micron beam diameter and ion counters (Supplementary Table 2). Analysis positions on CL images are presented in Supplementary File 4. Three analyses are available from rims, with a PDP peak age of 386.1 Ma and a weighted mean age of 386.1±4.6Ma (MSWD of 0.3) (Figure 7(a)). Twenty-seven analyses of cores yield a PDP peak age of 384.6 Ma, a TuffZirc age of 384.6 Ma, and a weighted mean age of 384.5±3.4Ma (MSWD of 1.0). Given the apparent overlap of the rim and core ages and the low MSWD values from both sets, we interpret the weighted mean age of 384.5±4.9 (from cores) as the crystallization age for sample TB-11 (Figure 7(b)). All zircons from this sample yield extremely high U concentrations, with some over 6000 ppm, and also extremely high U/Th values up to 6000. According to the criteria presented by Rubatto [28], both rims and cores are interpreted to have grown in the presence of abundant metamorphic fluids. Nine grains were analyzed for Hf, with resulting εHft values that range from +10.3 to +1.7 (Figure 11; Supplementary Table 3).

TB-17: four sets of analyses were conducted on sample TB-17 (Supplementary Table 2). The first set consisted of 52 analyses with a large laser beam diameter and faraday collectors. The second and third sets were conducted with a smaller laser beam diameter and ion counters. For the fourth set, four large grains with well-developed zonation were reimaged at higher resolution and analyzed with a small laser beam diameter and ion counters. All of the images and analysis locations for sets two through four are shown in Supplementary File 5.

As shown on Figures 7(c) and 7(d), both rim and core analyses yield a broad range of ages, with the youngest ages (mostly from rims) that likely are compromised by Pb loss and the oldest ages (mostly from cores) that likely are inherited components. Rim analyses yield a PDP peak age of 398.3 Ma, a TuffZirc age of 397.3 Ma, and a weighted mean age of 394.4±1.8Ma (with MSWD of 2.7). The weighted mean age is interpreted to be unreliable given that it incorporates a tail of young ages which likely have been compromised by Pb loss. Core analyses yield a PDP peak age of 399.7 Ma, a TuffZirc age of 399.8 Ma, and a weighted mean age of 400.7±1.5Ma (with MSWD of 2.0). Although the weighted mean age matches the PDP peak and TuffZirc ages, suggesting a dominance of ~400 Ma ages, it is likely compromised by the presence of both significantly younger and older ages. Collectively, the available data from rims suggest a crystallization age of approximately 397±5Ma for this sample. Most cores overlap with or slightly predate crystallization, although some are significantly older. The U concentrations and U/Th values for cores and rims of these grains are similar to each other and typical for igneous zircon.

Hf isotope analyses were conducted on twenty-five grains and yield εHft values that range from +4.1 to +8.9 (Figure 11; Supplementary Table 3).

BM 92-41: sixty analyses were conducted on sample BM 92-41, with 24 using a larger beam diameter and 36 utilizing a smaller beam diameter and ion counters (Supplementary Table 2). Locations of analysis pits for the latter are shown on CL images on Supplementary File 6. Rims from sample BM 92-41 yield a PDP peak age of 404.4 Ma, a TuffZirc age of 403.8, and a weighted mean age of 405.0±3.4Ma (MSWD of 1.5) (Figure 8(a)). This high MSWD value results mainly from two younger grains that may be compromised by Pb loss; the other analyses form a coherent cluster which we interpret to record an igneous crystallization age of 405.0±4.9Ma. Cores from this sample yield a PDP peak age of 430.9 Ma, a TuffZirc age of 430.8 Ma, and a weighted mean age of 428.3±5.5Ma (MSWD of 2.4). These ages are interpreted to represent inherited components that significantly predate crystallization, with some ages as old as ~487 Ma. As shown on Figure 9, both cores and rims yield U concentrations and U/Th values that are typical of igneous processes. Fifteen grains analyzed for Hf isotopes yield εHft values that range from +4.3 to +9.8 (Figure 11; Supplementary Table 3).

BM 92-57: we conducted 100 analyses on zircon crystals from sample BM 92-57, 33 using a larger laser beam and faraday detectors and 77 using a smaller laser beam diameter and ion counters. The latter are shown on CL images in Supplementary File 7. Rim analyses yield a PDP peak age of 381.2 Ma, a TuffZirc age of 382.6 Ma, and a weighted mean age of 383.3±2.3Ma (MSWD of 1.4). The latter age of 383.3±4.8Ma is interpreted as the crystallization age (Figure 8(c)). Core analyses yield a PDP peak age of 393.2 Ma, a TuffZirc age of 395.2 Ma, and a weighted mean age of 393.8±2.0Ma (MSWD of 1.1) (Figure 8(d)). The low MSWD value of the weighted mean age, combined with the range of observed ages, suggests that some of the core ages are indistinguishable from rim ages, many were inherited from source rocks of ~394 Ma, and some range back to ~430 Ma. The U concentrations and U/Th values for cores and rims of these grains are typical for igneous zircon. Analysis of fifteen grains for Hf isotopes yields εHft values that range from +8.0 to +16.2 (Figure 11; Supplementary Table 3).

6.2. Detrital Samples

Sample TB-8 yielded eight U-Pb analyses that range from ~442 to ~392 (Supplementary Table 2). The probability density plot of these ages yields main age peaks at 410 and 396 Ma and a subordinate age peak at 442 Ma (Figure 12). The youngest age peak of ~396 Ma is interpreted as the maximum depositional age (MDA) for this sample. Unfortunately, grains from this sample are too small for Hf analysis.

TB-9 yielded thirty-one U-Pb analyses that range from ~1.8 Ga to ~363 Ma (Supplementary Table 2). A probability density plot of these ages yields dominant age peaks at 420 and 369 Ma and a subordinate age peak at 540 Ma (Figure 12). The maximum depositional age for TB-9 is interpreted to be 369 Ma. εHft values for ten grains analyzed from this sample range from +10.5 to +14.5 (Figure 12), with one grain with an age of ~458 Ma yielding an εHft value of -27.5 (which we interpret to have been compromised analytically).

Sample ES-1 was initially analyzed by Brown et al. [1], who reported U-Pb analyses of 99 grains (Figure 12). Most grains are Precambrian in age, with a dominant age peak at 1858 Ma and subordinate peaks at 1988, 553, 424, and 369 Ma. In our current study, 191 U-Pb analyses and 28 Hf analyses were conducted on zircon crystals from this sample. The age distribution from our analyses ranges from ~2.9 Ga to ~358 Ma, with a dominant age peak at 1860 Ma and subordinate peaks at 2673, 2562, 2313, 2000, 1490, 1180, 1106, 736, 556, 430, and 362 Ma (Figure 12). With results of all 289 analyses combined, the dominant age peak is at 1860 Ma, with subordinate peaks at 2672, 2570, 2313, 1998, 1480, 1112, 736, 555, 430, and 362 Ma. ES-1 has an interpreted maximum depositional age of ~362. εHft values for ES-1 range from –4.6 to +14.4 (Figure 12; Supplementary Table 3).

7.1. Igneous Samples

As described above, Brown et al. [1] reported U-Pb ages of 417.9±5.4Ma (TB-1), 374.6±5.0Ma (TB-2), and 388.9±4.6Ma (TB-3) from three different plutonic rocks of the TBC. Inherited components, with ages as old as ~506 Ma, were also reported from these samples. Our results confirm this general age range for igneous rocks of the TBC and also document the presence of inherited components in every sample analyzed.

We herein explore the complexities of these samples in more detail utilizing CL images that document growth patterns (Figure 5; Supplementary Files 17), patterns of U concentration and U/Th (Figure 9), and ages of cores versus rims (Figure 10). Following the terminology and conceptual framework of Miller et al. [38], samples with cores that significantly predate rims (e.g., samples TB-1, TB-6, and BM 92-41) are interpreted to consist of rims that record zircon growth during magmatic crystallization and older cores that were inherited from preexisting igneous rocks. These core domains are accordingly referred to as xenocrysts. In contrast, samples with cores that overlap with or only slightly predate rims (e.g., samples TB-3, TB-11, TB-17, and BM 92-57) are interpreted to consist of rims that record zircon growth during magmatic crystallization and coeval to slightly older cores that were in part derived from coexisting magma chambers. Such cores are referred to as antecrysts. Regardless of origin, cores of the zircons analyzed from our igneous samples record the presence of igneous activity in the TBC from ~500 Ma to ~370 Ma, with dominant ages of ~425 and ~397 Ma based on maximum likelihood analysis (Figure 10). The U concentrations and U/Th values of most zircon cores record formation due to typical igneous processes (Figure 9).

The ages of zircon rims are interpreted to record crystallization of the sampled plutons or smaller igneous bodies. As shown on Figure 10 (all rims curve), rim ages range from ~410 to ~370 Ma or as young as ~360 Ma including analyses from sample TB-2. Maximum likelihood analysis suggests the presence of dominant age groups of ~401 and ~385 and a minor age group of ~364 Ma (Figure 10). Samples TB-1, TB-17, and BM 92-41 belong to the ~401 phase, TB-3, TB-6, TB-11, and BM 92-57 belong to the ~385 Ma group, and TB-2 represents the ~364 Ma phase. There is an apparent compositional trend among the sampled plutons, with the older set consisting mainly of leucotonalite, the middle set including more tonalite, and the youngest body consisting of gabbro.

The combination of rim and core ages (upper curve of Figure 10) is interpreted to provide a record of the history of magmatism in the TBC, with rim ages indicating the ages of sampled plutons and core ages indicating the ages of igneous bodies at depth that were involved in magma generation. The full range of ages (Supplementary Table 2) documents components as old as ~500 Ma and as young as ~360 Ma, with scattered <360 Ma ages interpreted to record Pb loss during younger greenschist-facies metamorphism. Maximum likelihood analysis suggests the presence of two dominant phases of magmatism at ~401 Ma and ~386 Ma and subordinate phases at ~424 and ~363 Ma (upper curve of Figure 10).

High U/Th values indicate that metamorphic processes also contributed to zircon crystallization between ~410 and ~370 Ma (Figure 9). The extremely high U/Th values for sample TB-11, and the abundance of muscovite in the sampled pluton, document the role of metamorphic fluids during crystallization at ~385 Ma. The occurrence of moderately high U/Th values in other samples, especially TB-3 and BM 92-57 (Figure 9(b)), suggests that metamorphic processes may also have been active during their formation at ~386 and ~384 Ma (respectively). The patterns on Figure 9(b) suggest an onset of metamorphism at ~410 Ma and continuation through ~370 Ma. This metamorphism is interpreted to be related to pluton generation and emplacement processes and distinct from the younger greenschist-facies metamorphism of all rocks of the TBC, which may have been responsible for the <370 Ma Pb loss recorded by some analyses.

Hafnium data from our samples are consistently very juvenile, with most εHft values between +5 and +15 (Figure 11; Supplementary Table 3). These values demonstrate that there is little if any older (e.g., Precambrian) continental material in the crust of the TBC. Although nearly all εHft values are juvenile, there is an apparent trend toward more evolved (less positive) εHft values from ~450 to ~410 Ma and then a stronger trend toward more juvenile (more positive) εHft values from ~410 to ~370 Ma (mostly in rims) (Figure 11). The evolution toward more evolved values may record the involvement of a greater proportion of crust during melting, perhaps in response to crustal thickening, whereas the following trend toward more juvenile values may record crustal extension, providing an increasing contribution of mantle-derived melts through time. Crustal extension during younger (e.g., ~410 to ~370 Ma) magmatism may also be recorded by the trend toward more mafic compositions through time.

Collectively, the patterns of crystallization and inherited ages, εHft values, and U/Th values record a protracted phase of evolution within the Turtleback magmatic system. Inherited cores record the existence of igneous source rocks ranging from ~500 to ~370 Ma in the crust within the TBC (all cores’ curve of Figure 10). The age distribution of cores suggests the presence of only scattered remnants between ~500 and ~440 Ma and progressively increasing volumes of crust from ~440 to ~395 Ma. These crustal components are represented in TBC plutons as >410 Ma xenocrysts (derived from the melting of preexisting arc crust) and as <410 Ma antecrysts (derived from mixing with coeval magma bodies). Some samples (e.g., TB-1, TB-6, and BM 92-41) are dominated by older components and others (e.g., TB-3, TB-11, TB-17, and BM 92-57) record significant mixing with coeval magma bodies, but most record both processes.

Variations in εHft and values of U concentration and U/Th provide insights into the tectonic processes that were operating during this protracted history of magma generation, mixing, and crystallization. Variations in εHft values are interpreted to record a progressive increase in crustal melting between ~450 and ~410 Ma, perhaps in response to crustal thickening, followed by an increasing contribution of mantle-derived magma from ~410 to ~370 Ma, possibly due to crustal extension. U/Th values record significant interaction with metamorphic fluids between ~410 and ~370 Ma, suggesting that metamorphism played an important role in generating the magmas that formed TBC plutons (and especially TB-3, TB-11, and BM 92-57).

7.2. Detrital Samples

The three detrital zircon samples analyzed for this study all yield early Paleozoic zircons, and sample ES-1 also contains a significant proportion of Precambrian grains (Figure 12(a)). Early Paleozoic grains in all three samples are mainly 445-360 Ma in age, similar to the age range of igneous rocks in the TBC (Figure 12(a)). As shown in Supplemental Table 2, U/Th values and U concentrations of the detrital grains are also similar to the values for most igneous samples from the TBC.

The overlap of the 445-360 Ma ages with the ages from igneous rocks of the TBC (Figure 12(a)) and their similar U/Th values and U concentrations suggest that Paleozoic detrital zircons from the ESG were likely shed from igneous rocks of the TBC. It is interesting, however, that compilations of all igneous ages and all detrital ages yield slightly different patterns, with igneous ages having dominant age peaks at ~401 and ~386 Ma and detrital ages having dominant age peaks of ~421 Ma and ~366 Ma (Figure 12(a)).

The overlap of εHft values for early Paleozoic zircon grains in our detrital and igneous samples also supports the interpretation of Brown et al. [1] that strata of the ESG were derived, at least in part, from the Turtleback Complex (Figure 12(b)). The occurrence of predominantly Precambrian grains in sample ES-1 demonstrates, however, that sources including Precambrian igneous rocks (or sedimentary rocks containing Precambrian grains) also contributed detritus to strata of the ESG.

7.3. Tectonic Implications

Possible connections between the TBC/ESG and other northern Cordilleran assemblages can be evaluated by comparison of the U-Pb geochronologic and Lu-Hf isotopic results from our study with data from the nearby Yellow Aster Complex (YAC), as well as the Alexander terrane, Wrangellia terrane, and southeast Alaska portion of the Yukon-Tanana terrane (Figure 1). Figure 13 compares results from the TBC and ESG with available data from the YAC in northwestern Washington [39], whereas Figure 14 compares both sets with the southern Alexander terrane in southeast Alaska [40], northern Alexander terrane in the Saint Elias Mountains [41, 42] and coastal British Columbia [43], Yukon-Tanana and Taku terranes in southeast Alaska [44], and Wrangellia terrane on Vancouver Island [45].

These comparisons show that the TBC/ESG has strong similarities, but also significant differences, with the Yellow Aster Complex. As shown on Figure 13(a), the age of early Paleozoic magmatism recorded in detrital zircons of the YAC is very similar to the older (~420-400 Ma) group of igneous rocks in the TBC. Both suites also record an increasing occurrence of high U/Th values (dashed lines with downward-increasing values in Figure 13(a)) during this magmatism. As shown in Figure 13(b), however, the εHft values are very different, with significantly more evolved values in the YAC. It is interesting to note that although the εHft values are quite different, the highly negative values in the YAC occur at the same time as a decrease in values in the TBC. For Precambrian grains, which are present in the ESG but not the TBC, there are strong similarities with the YAC for both U-Pb ages and εHft values.

Collectively, the available data support previous suggestions by Brown et al. [1] of connections between the TBC/ESG and the YAC. The TBC and YAC are interpreted to have formed as different parts of an arc system, with separate oceanic (TBC) and continental (YAC) segments. A reasonable scenario is that magmatism commenced in both assemblages at ~500 Ma, progressively increased until ~410 Ma, experienced a major flare-up in crustal melting and metamorphism between ~410 and ~400 Ma, and then transitioned to juvenile arc magmatism (only in the TBC) from ~400 to ~370 Ma. These changes are interpreted to have occurred in a regime of crustal thickening prior to ~400 Ma, followed by an episode of crustal extension (e.g., [46]). The YAC experienced ductile deformation and metamorphism prior to and during plutonism; such evidence is not recorded in the TBC. Magmatism ceased in both assemblages by ~380-360 Ma, after which strata of the ESG received detritus from both the TBC (Paleozoic grains) and YAC (mainly Precambrian grains). On the basis of these age similarities, we follow Brown et al. [1] in suggesting that the TBC, ESG, and YAC may all belong to the Chilliwack composite terrane, but suggest that differences in arc basement and deformation history record along-strike variations in the arc system.

The patterns of U-Pb ages and εHft values shown on Figure 14 can be used to evaluate potential connections with the Alexander terrane, Wrangellia terrane, and southeast Alaska portion of Yukon-Tanana terrane. Connections among these terranes are suggested by the occurrence of early Paleozoic magmatism in all, although the timing of magmatism is quite variable (Figure 14(a)). As shown in Figure 14(b), the εHft signature of Paleozoic rocks in each terrane is also quite different, with some assemblages recording mostly juvenile or mostly evolved magmatism, and some experiencing both. In particular, the variation from juvenile to evolved signatures in southern versus northern assemblages of the Alexander terrane is similar to that observed in the composite Chilliwack terrane. Although the differences in magmatic history and isotopic affinity prohibit direct correlations among the various terranes, it is intriguing that all appear to share coordinated patterns of εHft values through time. As described by Gehrels and McClelland [47], the εHft values from all of these terranes follow a series of pull ups (increasing values) and pull downs (decreasing values). This pattern is shown by the dashed line in Figure 14(b), which is the sliding window average of all values from all terranes. Maxima in εHft values occur at ~432, ~383, and ~332 Ma; minima occur at ~402 and ~357 Ma. In much the same way that Kemp et al. [48] have interpreted variations in εHft values from southeast Australia, this pattern is interpreted to record an orogenic cyclicity of crustal thickening (pull downs) and crustal thinning (pull ups).

The similarity of εHft values shown in Figure 14(b) supports the conclusion of Brown et al. [1] and Schermer et al. [39] that the TBC, ESG, and YAC most likely formed in association with the Alexander and Wrangellia terranes. Although direct connections among these terranes cannot be established, we follow Gehrels and McClelland [47] in suggesting that all terranes formed in a complex convergent margin/island arc system that included regional variations in basement rocks, magmatic histories, and proximity to continental margins (e.g., [47]). This convergent margin system likely originated in the paleo-Arctic during early Paleozoic time and then migrated to the Cordilleran margin during the mid- to late Paleozoic [1, 7, 3942, 44, 45, 4953]. Connections with Alexander and Wrangellia also have important implications for the Mesozoic accretionary history of the northern Cordillera in that the TBC/ESG and Yellow Aster Complex would all lie outboard of the hypothesized Intermontane-Insular suture, as proposed by Sigloch and Mihalynuk [8, 9] and Clennett et al. [10].

As noted on Figure 14, rocks of the Yukon-Tanana terrane in southeast Alaska yield somewhat similar patterns of age and εHft values as the TBC, ESG, YAC, and Alexander and Wrangellia terranes, which is problematic given that rocks of the Yukon-Tanana terrane are interpreted to have formed along the Cordilleran margin as part of the Intermontane terrane (e.g., [7]). These rocks are also interpreted to lie inboard of the hypothesized Insular-Intermontane suture [8, 9], which would not be the case if the southeast Alaska portion of Yukon-Tanana has connections with the Insular terrane. Possible explanations of these problematic relations are that the Yukon-Tanana terrane (1) originated in the paleo-Arctic (as part of the Insular terrane) but moved into the Cordilleran realm (and became part of the Intermontane terrane) during early Paleozoic time (e.g., [7]), (2) formed as (and remained) part of the Insular terrane [47], or (3) consists of two separate components, a southeast Alaska assemblage (characterized in Figure 14) that belongs to the Insular terrane and a northern assemblage that belongs to the Intermontane terrane (e.g., [44]). These scenarios require additional research given their important implications for the Paleozoic displacement history and Mesozoic accretionary history of northern Cordilleran terranes.

Our U-Pb geochronologic and Lu-Hf isotopic analyses yield several first-order conclusions regarding the TBC and ESG:

  • (1)

    The TBC yields a history of magmatism revealed through detailed reconstruction of the growth history of complex zircon crystals. Nearly all the zircon crystals contain cores that record igneous activity beginning at ~500 Ma and continuing through ~370 Ma, with dominant age modes of ~425 Ma and ~397 Ma. These cores occur as xenocrysts and antecrysts that are surrounded by rims with dominant ages of ~405 to ~375 Ma

  • (2)

    εHft values from cores and rims are consistently quite juvenile, but record a transition toward more juvenile values at ~410-400 Ma, which is the same time that the igneous rocks become more mafic in composition. Collectively, these relations are interpreted to record a transition from crustal thickening to crustal thinning

  • (3)

    U/Th values and U concentrations of the cores and rims also show a change from typical igneous values to much higher values that record an influx of metamorphic fluids beginning at ~400 Ma, presumably related to high heat input during crustal thinning

  • (4)

    Early Paleozoic detrital zircons in overlying strata of the ESG yield age and εHft values that resemble values for the TBC, thereby confirming previous suggestions (e.g., [1]) that these rocks were derived in large part from the TBC. ESG strata also yield abundant Precambrian detrital zircons which were likely shed from the nearby Yellow Aster Complex or from similar source terranes

  • (5)

    Comparison of U-Pb/Hf data from the TBC and ESG with similar information from the Yellow Aster Complex, Alexander terrane, Wrangellia terrane, and southeast Alaska portion of the Yukon-Tanana terranes precludes direct correlations, but suggests that all of these assemblages experienced a coordinated history of early Paleozoic magmatism and crustal evolution. As shown on Figure 14, all show εHft pull ups at ~432, ~383, and ~332 Ma and pull downs at ~402 and ~357 Ma, which are interpreted to record an orogenic cycle of crustal thickening and thinning [47]

  • (6)

    Connections with the Alexander and Wrangellia terranes demonstrate that rocks of the San Juan Islands are exotic components in the North American Cordillera. It is likely that these rocks originated in the paleo-Arctic during early Paleozoic time and migrated to the Cordilleran margin during the mid- to late Paleozoic (e.g., [52]) through displacement along a major transform system active on the northern Cordilleran margin [54]

  • (7)

    Potential connections with the Yukon-Tanana terrane age are problematic, as these rocks are interpreted to have formed and resided along the Cordilleran margin (e.g., [7]). Further investigations of these relations will have important implications for the Paleozoic displacement history and Mesozoic accretionary history of northern Cordilleran terranes (e.g., [79])

All of the data presented in this paper are available from the tables and figures within the text, as well as from several supplementary files and tables.

The authors are not aware of any conflicts of interest with the information presented herein.

We thank Dan Alberts, Mark Pecha, Nicky Giesler, Chelsi White, Clay Kelty, Gayland Simpson, Ken Kanipe, Kurt Sundell, Martin Pepper, and Natalie Speaks of the Arizona LaserChron Center for their assistance in generating the data presented herein. We also thank J.D. Mizer for his assistance with this project. The reference list in the revised manuscript was modified at the request of the journal. The National Science Foundation is acknowledged for support of this project (grant numbers EAR-1242939 and NSF EAR-1249115) and for support of the Arizona LaserChron Center (grant numbers EAR-1338583 and EAR-1649254).

Exclusive Licensee GeoScienceWorld. Distributed under a Creative Commons Attribution License (CC BY 4.0).