1. Burn depths and clinker formation depths.

Transcription

1 1 Supplemental Text, Figures, and Tables GSA Supplemental Data Item GSA Today, v. 21, no. 7, doi: /G107A.1 Clinker geochronology, the first glacial maximum, and landscape evolution in the northern Rockies Peter W. Reiners, Dept. of Geosciences, University of Arizona, Tucson, Arizona 85721, USA, Catherine A. Riihimaki, Biology Dept., Drew University, Madison, New Jersey 07940, USA; Edward L. Heffern, U.S. Bureau of Land Management, 5353 Yellowstone Road, Cheyenne, Wyoming 82009, USA 1. Burn depths and clinker formation depths. Because of the natural combustibility of PRB coal, beds thick enough to sustain combustion (~3 m) are rarely found in surficial exposures. Burning occurs almost exclusively in the subsurface, and unburned beds (generally less than ~6 m thick), which are sometimes found in recently incised gullies or landslide headwalls, have usually experienced devolatilization and oxidation prior to exposure, rendering them inert so they will not burn. The depth of burning is limited primarily by air ventilation from the surface and the water table. Variation in these factors is expected to cause variation in local burn depths. Burn depth may be shallower near valley bottoms, where the water table is closest to the surface, and greater for thicker coal beds, whose burning produces greater collapse and fissuring, facilitating deeper ventilation. In general, observations of modern and historical burn depths and coalclinker contacts in the PRB suggest that burn depth is usually about m (Shellenberger and Donner, 1979; Woessner et al., 1980; Coates and Naeser, 1984), which serves as a useful approximation for the typical maximum depth at which clinker is formed. Most of our samples were collected close to the base of clinker units, as close as possible to coal ash layers, although in a few locations only ventilation chimneys could be sampled. In such cases, clinker formation depth may be shallower. 2. Potential effect of groundwater table on burn depth. Given that most clinker is found in regions where groundwater is deeper than the modern burn front, in general air ventilation is likely to be the most important control on burn depth. In some locations, however, burn front depth may be much shallower than the m that we assume for many of the calculations in this paper. In particular this may be the case near valley bottoms occupied by perennial streams. In such cases, a reasonable assumption is that burn depth would be shallowest at low elevation near valley bottoms and deepest at high elevation near ridges. This would lead to younger clinker formation ages near valley bottoms, even if incision and lateral backwasting rates were equal and erosion rates were uniform. For example, assuming burn depth, or Z c, varies linearly from 0 to 30 m the bottom of a valley to the top, the general effect on most of the model clinker age curves in Figure 4 is to simply shift the low-elevation intercept to zero-age, while preserving the general form of the rest of the curve. The exception to this is the two model curves in Fig. 4b with solid lines that represent finite lateral backwasting rates but zero rates of incision or background erosion. For these cases, the predicted clinker ages are uniform (and inversely proportional to backwasting rate) over the entire elevation range. These potential effects are illustrated in more detail (Fig. S2) using the simpler erosion model described in the next section.

2 2 3. Alternative landscape evolution model for interpreting clinker age distributions. In the main text we use a model that assumes a vertical incision rate at the channel bottom (I) and a horizontal (lateral) backwasting rate at the valley rim or ridge (L). An alternative model, which generates very similar results, can be constructed that does not involve a horizontal component of erosion and simply assumes differing vertical erosion rates at the channel bottom (I) and valley rim or ridge (R). In this model, the local vertical erosion rate as a function of normalized elevation above base level (Z/Z b ) is simply I(1-Z/Z b )+R(Z/Z b ). As before, the predicted clinker age is burn depth Z c divided by this local erosion rate. As in the other model, ages are invariant with Z/Z b when I=R, show an inverse concave-up correlation with Z/Z b when I<R, and a positive concave-up one when I>R (Fig. S2). For a given I and when when R=L, these predicted trends are extremely similar to those of the model involving lateral backwasting as shown in Figure 4. The only difference is that for I R, the model described here produces slightly more concave-up correlations, because of the quadratic distribution of erosion rates in the model involving backwasting; this is shown in a comparison in Figure 5 of the main text. One complication of this simple model involving only vertical erosion is that when I<R, the topographic profile gradually inverts over time from a valley to a ridge. This occurs over the duration c/(r-i), where c is the initial valley depth. Obviously a condition involving I<R must be transient to prevent topographic inversion. To illustrate a possible, though likely extreme, effect of groundwater on clinker formation depths described in the previous section, Figure S2 shows predictions of clinker age as a function of EABL in the vertical-only erosion model, but for Z c varying linear from zero to 30 m for samples taken at the channel bottom to ridge top (blue trends in Fig. S2). These trends simply reinforce the conclusions of the previous section.

3 Fig. S2. Predicted clinker formation ages as a function of normalized elevation above base level, using simplified vertical-only erosion model described here. These predictions are analogous to those for the similar model described in the main text and shown in Figure 4. I and R are vertical erosion rates at channel bottom and valley rim, respectively. Z c is clinker formation (burn) depth. 3

4 4 4. Detrital clinker U/Pb and (U-Th)/He ages Fig. S1. Upper panel: (U-Th)/He and U/Pb ages of double-dated zircon grains (single grains dated by both U/Pb and He methods) from the matrix of 2.6-Ma terrace sample 06PRB34. Most grains with ages less than 85 Ma fall on or close to the first-cycle trend, and are likely derived from volcanic or hypabyssal plutonic sources. Older grains have cooling ages significantly younger than crystallization ages, indicating slow cooling or reheating following crystallization, as would be expected for grains derived from crystalline basement. Symbols with elongated lower error bars represent grains with subhedral to euhedral morphology whose He ages have been corrected for alpha-ejection (Reiners, 2005) assuming full alpha-ejection correction; fully corrected age is an overestimate if all or part of the grain has lost a significant portion of the outermost 18 μm due to natural abrasion during transport. Symbols with elongated upper error bars represent grains with anhedral morphology whose He ages have not been corrected for alpha-ejection; this age is an underestimate if all or part of the grain retains a significant portion of the outermost 18 μm. Lower panel: Logarithmic U/Pb age probability density spectrum of the 2.6-Ma terrace sample shown above as well as a nearby sandstone from the Tongue River Member of the Fort Union Formation. Ages of probability maxima are labelled for significant peaks in each spectrum (in Ma for peaks of 102 Ma and younger, and in Ga for older peaks). See Tables S3 and S4 for data and methods.

5 5 5. Basinwide clinker age distributions. Figure S3. Zircon (U-Th)/He ages of in-place clinker units as a function of distance across PRB (A: from west to east for southern groups (in Wyoming); B: from south to north for northern groups (in Montana); see Figure 1 for sample groupings). Southern group samples are predominantly on the western and eastern edges of PRB, and show a general trend of decreasing age towards the center of the basin. Northern group samples (Ashland and Otter Creek groups) are near the northern edge of the PRB and show a general trend of decreasing age towards the south (towards the center of the basin). 6. Reference cited in Electronic Supplement and not in Main Article. Reiners, P.W., 2005, Zircon (U-Th)/He Thermochronometry, in Reiners, P.W. and Ehlers, T.A. (Eds.), Thermochronology: Reviews in Mineralogy and Geochemistry, v. 58, p

6 Little Thunder Creek Group Table S1. Weighted average zircon (U-Th)/He ages and supporting data from in-place clinker units with fully reset ages Sample Northing a Easting a Elev (NED) b (km) c Dist to BL Northern Subgroup Elev of BL (m) d Age 2σ error # of aliquots analyzed CLK CLK SPL SPL PRB CLK Southern Subgroup 06PRB PRB PRB PRB PRB PRB Central Subgroup 06PRB PRB CLK PRB PRB PRB PRB PRB PRB EABL (m) g East Central Powder River Group Sample Northing a Easting a Elev (NAD) Dist to BL (km) c Elev of BL (m) d Age 2σ error # of aliquots analyzed EABL (m) g 06PRB PRB PRB PRB PRB PRB Western Powder River Group Sample Northing a Easting a Elev (NAD) Northern Subgroup Dist to BL (km) c Elev of BL (m) d Age 2σ error # of aliquots analyzed 07PRB EABL (m) g

7 07PRB PRB PRB PRB PRB PRB Central Subgroup CLK PRB PRB PRB PRB Southern Subgroup 07PRB PRB CLK PRB PRB Padlock Ranch Group Sample Northing a Easting a Elev (NAD) Dist to BL (km) c Elev of BL (m) d Age 2σ error # of aliquots analyzed EABL (m) g 07PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB Otter Creek Group Sample Northing a Easting a Elev (NAD) Dist to BL (km) c Elev of BL (m) d Age 2σ error # of aliquots analyzed EABL (m) g 03NPRB PRB PRB

8 06PRB PRB Ashland Region Group Sample Northing a Easting a Elev (NAD) Dist to BL (km) c Elev of BL (m) d Age 2σ error # of aliquots analyzed EABL (m) g 03NPRB NPRB PRB NPRB PRB NPRB PRB PRB PRB PRB PRB PRB PRB NPRB NPRB NPRB NPRB NPRB NPRB PRB Notes: a UTM zone 13N, North American Datum 1983 b NED is National Elevation Dataset, 1 arcsecond cell size, projected to UTM zone 13 and cell size of 27 m c Distance to closest stream channel with a drainage area >100 km 2 d Elevation of the closest stream channel with a drainage area >100 km 2, here defined as the local base level g Elevation of the sample site above local base level

9 Table S2. Single grain zircon (U-Th)/He ages and supporting data Sample name fmol 4 He ng U ng Th a F T Mass (μg) half-width (μm) U (ppm) Th (ppm) Th/U corrected age In-place clinker 03NPRB1zA NPRB1zB * weighted mean NPRB3z_EH NPRB3z_EH NPRB3z_EH NPRB3z_EH weighted mean NPRB4zB NPRB5zA NPRB5zB weighted mean NPRB6zA NPRB14zA NPRB14zB weighted mean NPRB16zA NPRB18zF NPRB18zA NPRB18zB weighted mean NPRB19zC NPRB19zD NPRB19zA NPRB19zB weighted mean NPRB22zA NPRB22zC NPRB22zD weighted mean NPRB25zB zA zB est 2σ error

10 weighted mean PRB20zA PRB20zB weighted mean PRB18zA PRB18zB weighted mean PRB17zA PRB17zB weighted mean PRB19zA PRB19zB weighted mean PRB11zA PRB11zB weighted mean PRB12zA PRB12zB weighted mean PRB15zA PRB15zB weighted mean SPL1zb CLK1c weighted mean CLK5a CLK5c CLK5d weighted mean SPL2zA SPL2zB weighted mean SPL3zA SPL3zB weighted mean CLK6b

11 CLK7c CLK7d weighted mean CLK8b CLK8c weighted mean CLK9a CLK9b weighted mean zA zB weighted mean PRB12zA PRB12zB weighted mean PRB13zA PRB13zB weighted mean PRB16zA PRB16zB PRB16zAprime weighted mean PRB26zr PRB26zA PRB26zB weighted mean PRB27z PRB01zA PRB01zB weighted mean PRB02zA PRB02zB weighted mean PRB03zA PRB03zB weighted mean PRB04zB

12 06PRB24zB PRB24zB weighted mean PRB25z PRB25z PRB25zB weighted mean PRB05zA PRB05zB weighted mean PRB06zA PRB06zB weighted mean PRB07zA PRB07zB weighted mean PRB08zA PRB08zB weighted mean PRB09zA PRB09zB weighted mean PRB10zB PRB11zA PRB11zB weighted mean PRB47zr PRB47z weighted mean PRB19z PRB20z PRB21z PRB29z PRB29z PRB29zr

13 weighted mean PRB38z PRB38z weighted mean PRB44z PRB44z weighted mean PRB49z PRB49z weighted mean PRB50z PRB50z weighted mean PRB51z PRB51z weighted mean PRB3zA PRB3zB weighted mean PRB4zA PRB4zB weighted mean PRB5zA PRB5zB weighted mean PRB8zA PRB8zB weighted mean PRB15z PRB15z PRB15z weighted mean PRB20zA PRB20zB weighted mean PRB11z PRB11z

14 weighted mean PRB19zA PRB21z PRB21z PRB21z weighted mean PRB22z PRB22z PRB22Az PRB22Az weighted mean PRB13z PRB13z weighted mean PRB10z PRB10z weighted mean PRB7zA PRB7zB weighted mean PRB24z PRB24z weighted mean PRB14z PRB14z weighted mean PRB16z PRB17z PRB17z weighted mean PRB26z PRB26z weighted mean PRB28z PRB29z PRB29z

15 weighted mean PRB6zA PRB6zB weighted mean PRB23z PRB32z PRB35z PRB35z weighted mean PRB36z PRB36z weighted mean PRB37z PRB37z weighted mean PRB39z PRB39z weighted mean PRB40z PRB40z weighted mean PRB41z PRB41z weighted mean PRB44z PRB44z weighted mean PRB45z PRB45z PRB45z weighted mean PRB47z Detrital clinker in Pliocene gravel terrace 06PRB33KzA PRB33KzB weighted mean

16 PRB33Azr PRB33Dzr PRB33DzrB PRB33DzrB weighted mean PRB31z PRB35z PRB35z weighted mean PRB36z PRB36z weighted mean Matrix (nonclinker) of Pliocene conglomerate terrace 06PRB34z PRB34z Notes: a F T is alpha ejection correction for zircon (Reiners, 2005) * grain surface coated in clinker matrix >15 μm thick

17 Table S3. U/Pb ages, supporting data, and methods for detrital zircon grains from Pliocene terrace and Paleocene sandstone. Isotope ratios Apparent ages Zircon grain U 206Pb U/Th 206Pb* ± 207Pb* ± 206Pb* ± error 206Pb* ± 207Pb* ± 206Pb* ± Best age ± (ppm) 204Pb 207Pb* (%) 235U* (%) 238U (%) corr. 238U* 235U 207Pb* 06PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB

18 06PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB

19 06PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB

20 06PRB PRB PRB PRB PRB PRB PRB PRB PRB PRB Notes: Sample 06PRB34: Pliocene alluvium atop strath terrace: Location (NAD1983): Northing: ; Easting: Sample 06PRB30: Massive Paleocene sandstone in Tongue River Member of Fort Union Formation: Location (NAD1983): Northing: , Easting: Methods: All zircons were mounted as whole grains on double sided tape and analyzed by LA ICP MS following methods in Gehrels et al. (2006; 2008). A subset of grains from 06PRB34 were then analyzed for (U Th)/He ages using methods in Reiners (2005). References Cited Gehrels, G., Valencia, V., and Pullen, A., 2006, Detrital Zircon Geochronology by Laser Ablation Multicollector ICPMS at the Arizona LaserChron Center, in Olszewski, T., Ed., Geochronology: Emerging Opportunities: Paleontology Society Papers, Volume 12, p Gehrels, G.E., Valencia, V., Ruiz, J., 2008, Enhanced precision, accuracy, efficiency, and spatial resolution of U-Pb ages by laser ablation-multicollector-inductively coupled plasma-mass spectrometry: Geochemistry, Geophysics, Geosystems, v. 9, Q03017, doi: /2007gc Reiners, P.W., 2005, Zircon (U-Th)/He Thermochronometry, in Reiners, P.W. and Ehlers, T.A. (Eds.), Thermochronology, Reviews in Mineralogy and Geochemistry, v. 58, p

21 Table S4. (U-Th)/He ages and supporting data of zircon grains also dated by U/Pb (Table S3). Sample name pmol 4 He ng U ng Th a F T mass (μg) half-width (μm) U (ppm) Th (ppm) Th/U corrected age 2σ analytical err positive 2σ uncertainty negative 2σ uncertainty 06PRB34Z PRB34Z PRB34Z PRB34Z PRB34Z PRB34Z PRB34Z PRB34Z PRB34Z PRB34Z PRB34Z PRB34z PRB34z PRB34z PRB34z PRB34z Notes: a F T is alpha ejection correction for zircon (Reiners, 2005). Ages of grains with subhedral to euhedral morphology have been corrected for alpha ejection (Reiners, 2005) assuming full alpha ejection correction. Ages of grains with anhedral morphology have not been corrected for alpha ejection (F T =1), assuming complete loss of alpha ejection affected profile during natural abrasion. b The uncorrected age is an underestimate if if all or part of the grain retains a significant portion of the outermost 18 μm. Grains with no alpha ejection corrections have positive uncertainties extending to their fully alpha ejection corrected ages. Grains with alpha ejection correction have assumed positive uncertainties of 8% (2σ). c The fully corrected age is an overestimate if all or part of the grain has lost a significant portion of the outermost 18 μm due to natural abrasion during transport. Grains with full alpha ejection corrections have negative uncertainties extending to their raw, uncorrected ages. Grains with no alpha ejection correction have assumed negative uncertainties of 8% (2σ).