- Email: [email protected]
An optical age chronology of Late Pleistocene fluvial deposits in the northern lower Mississippi valley
Quaternary Science Reviews 22 (2003) 1105–1110
An optical age chronology of Late Pleistocene fluvial deposits in the northern lower Mississippi valley Tammy M. Rittenour*, Ronald J. Goble, Michael D. Blum Department of Geosciences, University of Nebraska-Lincoln, 214 Bessey Hall, Lincoln, NE 68588-0340, USA
Abstract The lower Mississippi valley (LMV) contains many large braided channel belts that are preserved west of the Holocene floodplain. Previous efforts to establish geochronologic control on channel-belt construction have been hindered by the lack of organic material for radiocarbon dating. Luminescence techniques provide a burial date for the sediment itself and may prove useful in this context. Samples from three channel belts in the northern LMV were analyzed using the single aliquot-regenerative technique on 90–125 mm quartz. Optical ages (19.7–17.8, 16.1–15.0 and 12.5–12.1 ka) are consistent with geomorphic relationships and indicate that channel belts were formed in the late Pleistocene under glacial conditions. These optical ages provide the first detailed chronology of LMV channel-belt formation and are the first step towards developing a chronology for the entire LMV. r 2003 Elsevier Science Ltd. All rights reserved.
1. Introduction and background The lower Mississippi valley (LMV) is a broad alluvial valley that contains large braided channel belts with dimensions of the order of 10’s of km wide and 100’s of km long. The Mississippi River currently drains North America between the Appalachian and Rocky Mountains, but during the late Pleistocene its drainage basin was enlarged to include meltwater from the Laurentide ice sheet and large glacial lakes such as Lake Agassiz (Fig. 1) (Teller, 1990; Licciardi et al., 1999). In addition, the lower Mississippi River experienced high-amplitude variations in eustatic sea-level during the last glacial cycle (Chappell et al., 1996). Braided channel belts in the LMV have been examined by three generations of researchers over the past 65 years. Conflicting interpretations of channel belt ages and the importance of sea level and meltwater discharge in channel belt formation have been proposed. For example, Fisk (1944) suggested that the entire LMV was deeply incised during the last glacial eustatic lowstand. Coarse-grained braided channel belts formed in response to early to middle Holocene sea-level rise, with transition to a fine-grained meandering stream during the late Holocene sea-level highstand. Saucier *Corresponding author. Tel.: +1-402-472-2663; fax: +1-402-4724917. E-mail address: [email protected] (T.M. Rittenour).
(1994a, b, 1996) in contrast, proposed that braided channel belts formed during deglaciation and the transition to a meandering regime occurred in the early Holocene due to a loss of glacial sediment and meltwater discharge. The chronology proposed by Saucier (1994a, b) places channel belt formation during deglaciation in marine isotope stage (MIS) 4 and 2. Blum et al. (2000) concurred with Saucier’s glacial forcing mechanism but proposed that major channel belt surfaces represent MIS 6 and 2 deglaciations, with MIS 4 channel belts only preserved in isolated locations. All previous studies of LMV channel belts were based on relative age relationships, with a limited number of radiocarbon age estimates. Only a few reliable radiocarbon ages have been obtained after years of study by various groups, most of which only constrain the minimum timing of channel belt abandonment (see Blum et al., 2000). Collection of material for radiocarbon dating has been difficult in the LMV because the braided channel-belt deposits often lack organic material or are too old for the radiocarbon technique. In addition, bulk organic sediments are commonly contaminated by lignite, producing age overestimates. Hence, channel belt ages and their correlation to forcing mechanisms have remained poorly resolved. Because of the difficulty in obtaining reliable radiocarbon ages, optically-stimulated luminescence (OSL) were applied to the LMV to develop a chronology of channel belt formation. Since the deposits are primarily
0277-3791/03/$ - see front matter r 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0277-3791(03)00041-6
1106
T.M. Rittenour et al. / Quaternary Science Reviews 22 (2003) 1105–1110
Fig. 1. The Mississippi River and its drainage basin. Also shown is the last glacial maximum extent of the Laurentide ice sheet, the maximum extent of glacial lakes and major drainage routes (arrows). The location of Fig. 2 is indicated.
composed of sand with few gravel lenses, the selection of suitable sample localities was much easier than for radiocarbon and allowed multiple ages to be obtained for each channel belt surface. This paper reports initial results from these efforts and presents the first detailed chronology of LMV channel belts.
2. Methods Samples for optical dating were collected from three previously mapped, cross-cutting channel belts in the northern LMV. The channel belts sampled, listed from the highest (oldest) to lowest (youngest) surface, are Pve, Pvl 2 and Pvl 1, using the map unit names of Saucier (1994a) (Fig. 2). These channel belts have the best radiocarbon age control and offer the best opportunity to test previous age models, as well as test the suitability of optical dating in this setting. All samples were processed at the Luminescence Geochronology Lab at the University of Nebraska-Lincoln. 2.1. Sample collection and preparation Samples were collected in aluminum tubes pushed into sediment exposures or by coring with a vibracorer. Sample depth, elevation, and latitude/longitude were noted for calculation of cosmic contribution (Prescott and Hutton, 1994). Representative samples for the determination of water content and dose rate were collected from within 30 cm of the sample tube in sediment exposures, or from within the core tube from samples collected by vibracorer. In situ water content was measured by drying the samples at 100 C. The bulk
Fig. 2. Field area showing the location of sample sites for optical ages. The lightly shaded channel belts are Pve, Pvl 2 and Pvl 1 using Saucier’s (1994a) map units. Darker shaded channel belts were not sampled. A cross-section through the channel belts (A-A0 ) is provided at the bottom of the figure.
sediment concentration of K, Rb, U and Th was measured using ICP-MS and ICP-AES techniques. For this analysis, the sample was ground to 75 mm and dissolved in a hydrofluoric–nitric–perchloric acid digestion (performed by CHEMEX Labs, Sparks, Nevada). Representative duplicates were checked for complete dissolution using a fusion flux dissolution technique. Finally, dose-rates incorporating water content, chemistry and cosmic contribution were calculated using the methods of Aitken (1998) and Prescott and Hutton (1994) (Table 1). Samples were opened in the laboratory under dim amber light and wet sieved to separate the 90–125 mm fraction. To isolate the quartz component, the samples were treated with 10% hydrochloric acid to remove carbonates and floated in 2.7 g/cm3 sodium polytungstate to isolate heavy minerals. To remove feldspars and etch the quartz grains, the samples were first treated with 10% hydrofluoric acid for one hour to remove the highly reactive minerals. This solution was then
T.M. Rittenour et al. / Quaternary Science Reviews 22 (2003) 1105–1110
1107
Table 1 ICP-MS and ICP-AES chemistry and dose-rate data for samples. Site numbers refer to sample localities in Fig. 2 Site no.
Sample no.
UNL Lab no.
Depth (m)
H2Oa (wt%)
Th (ppm)
U (ppm)
K2O (wt%)
Rb2O (ppm)
Cosmic (Gy/ka)
Dose rate (Gy/ka)
Pve 22 8 20
MIS 06–24–22 MIS 03–27–8 MIS 06–23–20
UNL-120 UNL-168 UNL-346
1.8 1.5 1.8
16.0 5.9 5.0
5.270.5 3.170.3 2.070.2
1.470.1 1.170.1 1.070.1
1.6670.04 1.6370.04 1.4170.04
54.072.2 81.573.3 45.771.8
0.17 0.17 0.17
1.870.08 1.870.05 1.670.04
Pvl 2 47 44 45 51
MIS MIS MIS MIS
07–03–47 06–31–44 06–31–45 07–05–51
UNL-125 UNL-343 UNL-124 UNL-353
2.0 2.1 2.1 1.8
5.2 2.0 7.5 8.1
3.570.3 2.070.2 3.570.3 3.270.3
0.970.1 1.070.1 1.170.1 1.070.1
1.6070.04 1.5870.04 1.5470.04 1.5770.04
80.073.2 48.371.9 77.073.1 45.471.8
0.16 0.16 0.16 0.17
1.870.05 1.870.04 1.870.05 1.770.05
Pvl 1 37 39 16 18
MIS MIS MIS MIS
06–28–37 06–29–39 06–22–16 06–22–18b
UNL-169 UNL-123 UNL-119 UNL-335
2.3 2.7 2.1 2.7
15.0 7.7 8.2 0.7
3.570.3 6.070.5 3.170.3 4.870.4
1.070.1 1.170.1 1.070.1 1.270.1
1.5070.04 1.3670.03 1.7570.04 1.5470.04
75.073.0 68.072.7 87.573.5 51.072.0
0.16 0.15 0.16 0.15
1.670.06 1.870.06 1.870.05 2.070.05
a
In situ water content, assume 30% error in measurement.
decanted and replaced with 48% hydrofluoric acid for a minimum of 50 min, followed by 23% fluorosilicic acid for 48 h. This was followed by one hour in 47% hydrochloric acid to remove any fluorides produced. A mechanical agitator was used to ensure continual stirring during the acid treatments. The rinsed and dried sample was then re-sieved to remove the o 75 mm fraction. Infrared stimulation was used to detect the presence of feldspar and to assure the purity of the quartz separate. If necessary, the sample was treated a second time with concentrated hydrofluoric acid and fluorosilicic acid to remove residual feldspars. 2.2. Optical measurement setup The 90–125 mm quartz fraction was mounted in a 2mm diameter region on aluminum disks using Silkospray (a silicone based spray). Optical measurements were performed on a RISO TL/OSL-DA-15B/C reader with blue-green light stimulation (470 nm, Hoya U340 filter), using the single aliquot regenerative (SAR) protocol of Murray and Wintle (2000). To determine the proper preheat temperatures, preheat plateau tests were measured between 200 C and 300 C at 20 C increments, using 5 aliquots per temperature step. Equivalent dose values were not affected by preheat temperatures between 220 C and 260 C, but showed greater scatter above 260 C. Therefore, a preheat temperature of 240 C was selected. Cutheat temperature was 160 C, with optical measurements conducted at 125 C. Dose recovery tests were performed by applying a beta dose of 21 Gy to samples after bleaching. Dose recovery, using regenerative doses of 11 Gy (regenerative dose 1 and 4), 21 Gy (regenerative dose 2) and 43 Gy
(regenerative dose 3), and 5 Gy test doses, was within 0.5–3% of the applied dose, suggesting that the SAR method with the above preheat and cutheat temperatures is suitable for LMV channel belt samples.
3. Equivalent dose distributions Previous workers have shown that partial bleaching can be a problem in fluvial environments (Murray et al., 1995; Olley et al., 1998, 1999), especially in proglacial systems (Hutt and Jungner, 1992; Majdal and Funder, 1994; Gemmell, 1999). Partial bleaching was of some concern for this study because the Mississippi River is known to have received glacial meltwater and sediment from the Laurentide ice sheet during deglaciation. To help identify partial bleaching, small quantities of sand (B100 grains) were used in each aliquot in order to approach single-grain distributions (Olley et al., 1999). This caused the data from some aliquots to be discarded due to low response, but increased sensitivity to partially bleached populations. Equivalent dose (De ) distributions in histograms and cumulative probability plots were examined for evidence of asymmetry symptomatic of partial bleaching (Olley et al., 1998, 1999). Most De distributions were roughly Gaussian in nature, with some distributions slightly skewed toward higher values (Fig. 3). In addition, standardized De -signal intensity plots (Colls, 1999; Colls et al., 2001) had low correlation (r2 ¼ 020:2) and De ðtÞ plots (Bailey et al., 2002) showed no increase in De with measurement time, both indicating a lack of evidence for partial bleaching. This was surprising, given the glacio-fluvial environment and high suspended sediment concentrations in the river system, as
T.M. Rittenour et al. / Quaternary Science Reviews 22 (2003) 1105–1110
1108
MIS 06-28-37 av = 19.4 +/- 0.4, n = 32
MIS 06-23-20 av = 28.1 +/- 0.7, n = 27
8
12
Frequency
Frequency
16
8 4 0
6 4 2 0
0
10
20
30
0
10
Cumulative Probability Curve
20
30
40
50
Cumulative Probability Curve 4
3 Arbitrary units
Arbitrary units
6
4
2
2
1
0
0 0
10
20
0
30
10
De (Gy)
20
30 De (Gy)
40
50
Fig. 3. Histograms (upper) and cumulative probability curves (lower) for two samples with individual De and associated errors for aliquots plotted in rank order. Bin width for the histograms is 2 Gy. For the cumulative probability curves, the sum of the normal distributions (73s) from the aliquots is plotted in black with the distribution calculated from the mean and standard deviation in grey. Units are arbitrary and reflect increasing probability. Table 2 Optical age estimates and De values from LMV channel belt samples, De errors reported as standard error 14
C age (cal. ka)a
Site no.
Sample no.
UNL Lab no.
Dose rate (Gy/ka)
No. of aliquots
De (Gy)
s
Optical age (ka)
Pve 22 8 20
MIS 06–24–22 MIS 03–27–8 MIS 06–23–20
UNL-120 UNL-168 UNL-346
1.870.08 1.870.05 1.670.04
28 45 27
36.470.8 33.670.6 28.070.7
4.1 4.1 3.8
19.771.1 18.570.8 17.870.8
19.7570.65b
Pvl 2 47 44 45 51
MIS MIS MIS MIS
07–03–47 06–31–44 06–31–45 07–05–51
UNL-125 UNL-343 UNL-124 UNL-353
1.870.05 1.870.04 1.870.05 1.770.05
28 27 30 34
28.970.5 27.870.4 26.670.5 25.870.4
3.4 2.2 2.6 2.4
16.170.7 15.870.6 15.170.7 15.070.7
>13.1370.46c >12.9370.30d
Pvl 1 37 39 16 18
MIS MIS MIS MIS
06–28–37 06–29–39 06–22–16 06–22–18b
UNL-169 UNL-123 UNL-119 UNL-335
1.670.06 1.870.06 1.870.05 2.070.05
32 29 20 43
19.470.4 21.970.6 22.570.7 24.270.5
2.5 3.1 2.9 3.2
12.570.7 12.470.6 12.270.6 12.170.5
>10.2270.51e >9.8870.33f
Radiocarbon ages were collected by previous workers from other sites and represent the age constraint on the channel belt as a whole. Site numbers refer to sample localities in Fig. 2. a Radiocarbon ages converted to calendar years by CALIB 4.3 (Stuiver et al., 1998a, b). b Blum et al. (2000). c Wesnousky and Leffler (1992). d R.W. Graham, personal communication reference in Blum et al. (2000). e Guccione et al. (1988). f King and Allen (1977).
indicated by thick slackwater deposits (Robnett, 1997; Blum et al., 2000). Despite these conditions, long transport distances on the order of several hundred kilometers appear to have been sufficient for complete sediment bleaching.
Due to the limited evidence of partial bleaching, the mean De was used to calculate the optical ages (Table 2). This was justified by the nearly Gaussian De distributions and close agreement between the mean De ; median De and the peak cumulative probability De (see Fig. 3).
T.M. Rittenour et al. / Quaternary Science Reviews 22 (2003) 1105–1110
1109
Table 3 Comparison of ages proposed by different researchers. All ages in calendar ka Surface
Fisk (1944)
Saucier (1994a)
Blum et al. (2000)
Optical age
Pve Pvl 2 Pvl 1
12–5 12–5 12–5
60–80 25–12 earlier 25–12 later
ca. 21 ca. 13.5–12.5 ca. 12.5–10.7
19.7–17.8 16.1–15.0 12.5–12.1
4. Optical ages
6. Conclusions
Optical ages and radiocarbon age constraints for the three channel belt surfaces sampled are shown in Table 2. The highest channel belt surface, mapped as Pve, produced optical ages between 19.7 and 17.8 ka, whereas samples from Pvl 2 were between 16.1 and 15.0 ka, and the lowest channel belt, Pvl 1, produced ages between 12.5 and 12.1 ka. The spread in ages is consistent with a braided system that has occupied a channel belt for a period of time. Age estimates from these surfaces are in agreement with geomorphic relationships and previously determined radiocarbon ages. Of special note is the good agreement between the optical ages for the Pve channel belt and the radiocarbon age from a surface thought to correlate with Pve (Table 2). Other published radiocarbon ages are minimum age estimates for the surfaces; nevertheless the optical chronology agrees with these ages.
The application of optical luminescence techniques has provided the first detailed chronology for the large braided channel belts of the lower Mississippi alluvial valley. Previous chronologies have been based on relative age relationships and a limited number of radiocarbon ages. The main deterrent to obtaining age estimates in the past was the lack of datable organic material in the channel belt sediments. By contrast, the luminescence technique is ideal for the LMV since the channel belts are primarily composed of quartz-rich sand, and provide a nearly limitless selection of suitable sample localities. Samples from three previously mapped cross-cutting braided channel belts were collected to test previous age models developed for LMV evolution. Optical ages suggest that the highest channel belt, mapped as Pve by Saucier (1994a) is 19.7–17.8 ka, the Pvl 2 channel belt is 16.1–15.0 ka, and the lowest channel belt, Pvl 1, formed 12.5–12.1 ka. Despite glacial meltwater conditions in the Mississippi during the late Pleistocene, the samples show minimal evidence of partial bleaching. Optical ages agree with relative geomorphic relationships and the available radiocarbon ages, and suggest the channel belts formed in the late Pleistocene during MIS 2 under conditions of high meltwater discharge. These new data support the glacial forcing hypothesis of Saucier (1994a) and Blum et al. (2000) for channel belt formation in the northern LMV. The new optical chronology presented here has allowed previous models of LMV landform evolution to be tested. The ages obtained here are the first step towards a larger research objective to develop an optical chronology for the entire LMV. Once completed, it will be possible to resolve the spatial and temporal response of this continental-scale river system to adjustments in hydrology and sea level during a full glacial-interglacial cycle.
5. Comparison to previous channel belt chronologies The optical ages presented here allow previous models for LMV channel belt formation to be re-examined. To reiterate from above, under Fisk’s (1944) model, all braided channel belts formed during the early to middle Holocene in response to sea-level rise (Table 3). Saucier (1994a, b; 1996) in contrast suggested that the channel belts formed in response to glacial discharge. In relationship to the specific channel belts sampled here, Saucier (1994a) placed the Pve channel belt in MIS 4 and Pvl 2 and Pvl 1 in MIS 2 (Table 3). Finally, Blum et al. (2000) proposed that the Pve surface formed during the last glacial maximum (ca 21 ka) and concurred with Saucier (1994a, b) that Pvl 2 and Pvl 1 formed during MIS 2 deglaciation between ca 13.5 and 10.7 ka (Table 3). The optical ages presented here support a MIS 2, last glacial maximum, age for the Pve channel belt, as suggest by Blum et al. (2000). Furthermore, new optical ages greatly refine age estimates for Pvl 2 and Pvl 1, suggesting they are slightly older than what has been inferred from minimum ages obtained by the radiocarbon technique (Table 2). By providing the first numerical age estimates for LMV channel belts, these new optical ages provide an opportunity to move towards linking periods of channel belt construction and abandonment with various forcing mechanisms.
Acknowledgements This research was funded by a Student Research Grant from the Geological Society of America and a Geological Society of America J. Hoover Mackin Research Grant awarded to T. Rittenour, a NASA
1110
T.M. Rittenour et al. / Quaternary Science Reviews 22 (2003) 1105–1110
Earth System Science Student Fellowship and a National Science Foundation Doctoral Dissertation Grant (SES 0082226) awarded to T. Rittenour and M. Blum, and a National Science Foundation Research Grant (EAR 0107028) awarded to M. Blum.
References Aitken, M.J., 1998. An Introduction to Optical Dating, The Dating of Quaternary Sediments by the Use of Photon-stimulated Luminescence. Oxford Press, Oxford, 267pp. Bailey, R.M., Singarayer, J., Ward, S., Stokes, S., 2002. Identification of partial bleaching using De as a function of illumination time. 10th International Conference on Luminescence and Electron Spin Dating, Reno, NV. Programs and Abstracts, p.2. . Blum, M.D., Tornqvist, T.E., 2000. Fluvial response to climate and sea-level change: a review and a look forward. Sedimentology 47, 2–48. Blum, M.D., Guccione, M.J., Wysocki, D.A., Robnett, P.C., Rutledge, M., 2000. Late Pleistocene evolution of the lower Mississippi Valley, Southern Missouri to Arkansas. Geologic Society of America Bulletin 112, 221–235. Chappell, J., Omura, A., Ezat, T., McCulloch, M., Pandolfi, J., Ota, Y., Pillans, B., 1996. Reconciliation of late Quaternary sea-levels derived from coral terraces at Huon Peninsula with deep sea oxygen isotope records. Planetary Science Letters 141, 227–236. Colls, A.E.L., 1999. Optical dating of fluvial sediment from the Loire Valley, France. Unpublished M.Sc. Thesis, University of Oxford, Oxford. Colls, A.E., Stokes, S., Blum, M.D., Straffin, E., 2001. Age limits on the Late Quaternary evolution of the upper Loire River. Quaternary Science Reviews 20, 743–750. Fisk, H.N., 1944. Geological Investigation of the Alluvial Valley of the Lower Mississippi River. US Army Corps of Engineers Mississippi River Commission, Vicksburg, MS, 78pp. Gemmell, A.M.D., 1999. IRSL from fine-grained glacifluvial sediment. Quaternary Science Reviews 18, 207–215. Guccione, M.J., Prior, W.L., Rutledge, E.M., 1988. Crowley’s Ridge, Arkansas. In: Hayward, O.T. (Ed.), South-Central Section of the Geological Society of America, Centennial Field Guide. Geological Society of America, Boulder, CO, pp. 225–230. Hutt, G., Jungner, H., 1992. Optical and TL dating on glaciofluvial sediments. Quaternary Science Reviews 11, 161–163. King, J.E., Allen Jr., W.H., 1977. A Holocene vegetation record from the Mississippi River valley, southeastern Missouri. Quaternary Research 8, 307–323.
Licciardi, J.M., Teller, J.T., Clark, P.U., 1999. Freshwater routing by the Laurentide ice sheet during the last deglaciation. In: Clark, P.U., Web, R.S., Keigwin L.D. (Eds.), Mechanisms of Global Climate Change at Millennial Time Scales. American Geophysical Union, Washington, DC, Geophysical Monograph 112, 177–201. Mejdahl, V., Funder, S., 1994. Luminescence dating of late Quaternary sediments from East Greenland. Boreas 23, 525–535. Murray, A.S., Wintle, A.G., 2000. Luminescence dating of quartz using an improved single aliquot regenerative-dose protocol. Radiation Measurements 32, 57–73. Murray, A.S., Olley, J.M., Caitcheon, G.C., 1995. Measurement of equivalent doses in quartz from contemporary water-lain sediments using optically stimulated luminescence. Quaternary Science Reviews 14, 365–371. Olley, J., Caitcheon, G., Murray, A., 1998. The distribution of apparent dose as determined by optically stimulated luminescence in small aliquots of fluvial quartz: implications for dating young fluvial sediments. Quaternary Science Reviews 17, 1033–1040. Olley, J., Caitcheon, G., Roberts, R.G., 1999. The origin of dose distributions in fluvial sediments, and the prospect of dating single grains from fluvial deposits using optically stimulated luminescence. Radiation Measurements 30, 207–217. Prescott, J.R., Hutton, J.T., 1994. Cosmic ray contributions to dose rates for luminescence and ESR dating: large depths and long-term time variations. Radiation Measurements 23, 497–500. Robnett, P.C., 1997. Late Quaternary occupation and abandonment of the Western Lowlands valley train course: Mississippi River, southeastern Missouri. M.S. Thesis, Southern Illinois University, Carbondale. Saucier, R.T., 1994a. Geomorphology and Quaternary Geologic History of the Lower Mississippi Valley. US Army Corps of Engineers Mississippi River Commission, Vicksburg, MS, 364pp. Saucier, R.T., 1994b. Evidence of late-glacial runoff in the Lower Mississippi Valley. Quaternary Science Reviews 13, 973–981. Saucier, R.T., 1996. A contemporary appraisal of some key Fiskian concepts with emphasis on Holocene meander belt formation and morphology. Engineering Geology 45, 67–86. Stuiver, M., Reimer, P.J., Bard, E., Beck, J.W., Burr, G.S., Hughen, K.A., Kromer, B., McCormac, F.G., van der Plicht, J., Spurk, M., 1998a. INTERCAL98 Radiocarbon age calibration 24,000–0 cal. BP. Radiocarbon 40, 1041–1083. Stuiver, M., Reimer, P.J., Braziunas, T.F., 1998b. High-precision radiocarbon age calibration for terrestrial and marine samples. Radiocarbon 40, 1127–1151. Teller, J.T., 1990. Volume and routing of late-glacial runoff from the southern Laurentide Ice Sheet. Quaternary Research 34, 12–23. Wesnousky, S.G., Leffler, L., 1992. The repeat time of the 1811 and 1812 New Madrid earthquakes: a geological perspective. Seismological Society of America Bulletin 82, 1756–1785.
An optical age chronology of Late Pleistocene fluvial deposits in the northern lower Mississippi valley Tammy M. Rittenour*, Ronald J. Goble, Michael D. Blum Department of Geosciences, University of Nebraska-Lincoln, 214 Bessey Hall, Lincoln, NE 68588-0340, USA
Abstract The lower Mississippi valley (LMV) contains many large braided channel belts that are preserved west of the Holocene floodplain. Previous efforts to establish geochronologic control on channel-belt construction have been hindered by the lack of organic material for radiocarbon dating. Luminescence techniques provide a burial date for the sediment itself and may prove useful in this context. Samples from three channel belts in the northern LMV were analyzed using the single aliquot-regenerative technique on 90–125 mm quartz. Optical ages (19.7–17.8, 16.1–15.0 and 12.5–12.1 ka) are consistent with geomorphic relationships and indicate that channel belts were formed in the late Pleistocene under glacial conditions. These optical ages provide the first detailed chronology of LMV channel-belt formation and are the first step towards developing a chronology for the entire LMV. r 2003 Elsevier Science Ltd. All rights reserved.
1. Introduction and background The lower Mississippi valley (LMV) is a broad alluvial valley that contains large braided channel belts with dimensions of the order of 10’s of km wide and 100’s of km long. The Mississippi River currently drains North America between the Appalachian and Rocky Mountains, but during the late Pleistocene its drainage basin was enlarged to include meltwater from the Laurentide ice sheet and large glacial lakes such as Lake Agassiz (Fig. 1) (Teller, 1990; Licciardi et al., 1999). In addition, the lower Mississippi River experienced high-amplitude variations in eustatic sea-level during the last glacial cycle (Chappell et al., 1996). Braided channel belts in the LMV have been examined by three generations of researchers over the past 65 years. Conflicting interpretations of channel belt ages and the importance of sea level and meltwater discharge in channel belt formation have been proposed. For example, Fisk (1944) suggested that the entire LMV was deeply incised during the last glacial eustatic lowstand. Coarse-grained braided channel belts formed in response to early to middle Holocene sea-level rise, with transition to a fine-grained meandering stream during the late Holocene sea-level highstand. Saucier *Corresponding author. Tel.: +1-402-472-2663; fax: +1-402-4724917. E-mail address: [email protected] (T.M. Rittenour).
(1994a, b, 1996) in contrast, proposed that braided channel belts formed during deglaciation and the transition to a meandering regime occurred in the early Holocene due to a loss of glacial sediment and meltwater discharge. The chronology proposed by Saucier (1994a, b) places channel belt formation during deglaciation in marine isotope stage (MIS) 4 and 2. Blum et al. (2000) concurred with Saucier’s glacial forcing mechanism but proposed that major channel belt surfaces represent MIS 6 and 2 deglaciations, with MIS 4 channel belts only preserved in isolated locations. All previous studies of LMV channel belts were based on relative age relationships, with a limited number of radiocarbon age estimates. Only a few reliable radiocarbon ages have been obtained after years of study by various groups, most of which only constrain the minimum timing of channel belt abandonment (see Blum et al., 2000). Collection of material for radiocarbon dating has been difficult in the LMV because the braided channel-belt deposits often lack organic material or are too old for the radiocarbon technique. In addition, bulk organic sediments are commonly contaminated by lignite, producing age overestimates. Hence, channel belt ages and their correlation to forcing mechanisms have remained poorly resolved. Because of the difficulty in obtaining reliable radiocarbon ages, optically-stimulated luminescence (OSL) were applied to the LMV to develop a chronology of channel belt formation. Since the deposits are primarily
0277-3791/03/$ - see front matter r 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0277-3791(03)00041-6
1106
T.M. Rittenour et al. / Quaternary Science Reviews 22 (2003) 1105–1110
Fig. 1. The Mississippi River and its drainage basin. Also shown is the last glacial maximum extent of the Laurentide ice sheet, the maximum extent of glacial lakes and major drainage routes (arrows). The location of Fig. 2 is indicated.
composed of sand with few gravel lenses, the selection of suitable sample localities was much easier than for radiocarbon and allowed multiple ages to be obtained for each channel belt surface. This paper reports initial results from these efforts and presents the first detailed chronology of LMV channel belts.
2. Methods Samples for optical dating were collected from three previously mapped, cross-cutting channel belts in the northern LMV. The channel belts sampled, listed from the highest (oldest) to lowest (youngest) surface, are Pve, Pvl 2 and Pvl 1, using the map unit names of Saucier (1994a) (Fig. 2). These channel belts have the best radiocarbon age control and offer the best opportunity to test previous age models, as well as test the suitability of optical dating in this setting. All samples were processed at the Luminescence Geochronology Lab at the University of Nebraska-Lincoln. 2.1. Sample collection and preparation Samples were collected in aluminum tubes pushed into sediment exposures or by coring with a vibracorer. Sample depth, elevation, and latitude/longitude were noted for calculation of cosmic contribution (Prescott and Hutton, 1994). Representative samples for the determination of water content and dose rate were collected from within 30 cm of the sample tube in sediment exposures, or from within the core tube from samples collected by vibracorer. In situ water content was measured by drying the samples at 100 C. The bulk
Fig. 2. Field area showing the location of sample sites for optical ages. The lightly shaded channel belts are Pve, Pvl 2 and Pvl 1 using Saucier’s (1994a) map units. Darker shaded channel belts were not sampled. A cross-section through the channel belts (A-A0 ) is provided at the bottom of the figure.
sediment concentration of K, Rb, U and Th was measured using ICP-MS and ICP-AES techniques. For this analysis, the sample was ground to 75 mm and dissolved in a hydrofluoric–nitric–perchloric acid digestion (performed by CHEMEX Labs, Sparks, Nevada). Representative duplicates were checked for complete dissolution using a fusion flux dissolution technique. Finally, dose-rates incorporating water content, chemistry and cosmic contribution were calculated using the methods of Aitken (1998) and Prescott and Hutton (1994) (Table 1). Samples were opened in the laboratory under dim amber light and wet sieved to separate the 90–125 mm fraction. To isolate the quartz component, the samples were treated with 10% hydrochloric acid to remove carbonates and floated in 2.7 g/cm3 sodium polytungstate to isolate heavy minerals. To remove feldspars and etch the quartz grains, the samples were first treated with 10% hydrofluoric acid for one hour to remove the highly reactive minerals. This solution was then
T.M. Rittenour et al. / Quaternary Science Reviews 22 (2003) 1105–1110
1107
Table 1 ICP-MS and ICP-AES chemistry and dose-rate data for samples. Site numbers refer to sample localities in Fig. 2 Site no.
Sample no.
UNL Lab no.
Depth (m)
H2Oa (wt%)
Th (ppm)
U (ppm)
K2O (wt%)
Rb2O (ppm)
Cosmic (Gy/ka)
Dose rate (Gy/ka)
Pve 22 8 20
MIS 06–24–22 MIS 03–27–8 MIS 06–23–20
UNL-120 UNL-168 UNL-346
1.8 1.5 1.8
16.0 5.9 5.0
5.270.5 3.170.3 2.070.2
1.470.1 1.170.1 1.070.1
1.6670.04 1.6370.04 1.4170.04
54.072.2 81.573.3 45.771.8
0.17 0.17 0.17
1.870.08 1.870.05 1.670.04
Pvl 2 47 44 45 51
MIS MIS MIS MIS
07–03–47 06–31–44 06–31–45 07–05–51
UNL-125 UNL-343 UNL-124 UNL-353
2.0 2.1 2.1 1.8
5.2 2.0 7.5 8.1
3.570.3 2.070.2 3.570.3 3.270.3
0.970.1 1.070.1 1.170.1 1.070.1
1.6070.04 1.5870.04 1.5470.04 1.5770.04
80.073.2 48.371.9 77.073.1 45.471.8
0.16 0.16 0.16 0.17
1.870.05 1.870.04 1.870.05 1.770.05
Pvl 1 37 39 16 18
MIS MIS MIS MIS
06–28–37 06–29–39 06–22–16 06–22–18b
UNL-169 UNL-123 UNL-119 UNL-335
2.3 2.7 2.1 2.7
15.0 7.7 8.2 0.7
3.570.3 6.070.5 3.170.3 4.870.4
1.070.1 1.170.1 1.070.1 1.270.1
1.5070.04 1.3670.03 1.7570.04 1.5470.04
75.073.0 68.072.7 87.573.5 51.072.0
0.16 0.15 0.16 0.15
1.670.06 1.870.06 1.870.05 2.070.05
a
In situ water content, assume 30% error in measurement.
decanted and replaced with 48% hydrofluoric acid for a minimum of 50 min, followed by 23% fluorosilicic acid for 48 h. This was followed by one hour in 47% hydrochloric acid to remove any fluorides produced. A mechanical agitator was used to ensure continual stirring during the acid treatments. The rinsed and dried sample was then re-sieved to remove the o 75 mm fraction. Infrared stimulation was used to detect the presence of feldspar and to assure the purity of the quartz separate. If necessary, the sample was treated a second time with concentrated hydrofluoric acid and fluorosilicic acid to remove residual feldspars. 2.2. Optical measurement setup The 90–125 mm quartz fraction was mounted in a 2mm diameter region on aluminum disks using Silkospray (a silicone based spray). Optical measurements were performed on a RISO TL/OSL-DA-15B/C reader with blue-green light stimulation (470 nm, Hoya U340 filter), using the single aliquot regenerative (SAR) protocol of Murray and Wintle (2000). To determine the proper preheat temperatures, preheat plateau tests were measured between 200 C and 300 C at 20 C increments, using 5 aliquots per temperature step. Equivalent dose values were not affected by preheat temperatures between 220 C and 260 C, but showed greater scatter above 260 C. Therefore, a preheat temperature of 240 C was selected. Cutheat temperature was 160 C, with optical measurements conducted at 125 C. Dose recovery tests were performed by applying a beta dose of 21 Gy to samples after bleaching. Dose recovery, using regenerative doses of 11 Gy (regenerative dose 1 and 4), 21 Gy (regenerative dose 2) and 43 Gy
(regenerative dose 3), and 5 Gy test doses, was within 0.5–3% of the applied dose, suggesting that the SAR method with the above preheat and cutheat temperatures is suitable for LMV channel belt samples.
3. Equivalent dose distributions Previous workers have shown that partial bleaching can be a problem in fluvial environments (Murray et al., 1995; Olley et al., 1998, 1999), especially in proglacial systems (Hutt and Jungner, 1992; Majdal and Funder, 1994; Gemmell, 1999). Partial bleaching was of some concern for this study because the Mississippi River is known to have received glacial meltwater and sediment from the Laurentide ice sheet during deglaciation. To help identify partial bleaching, small quantities of sand (B100 grains) were used in each aliquot in order to approach single-grain distributions (Olley et al., 1999). This caused the data from some aliquots to be discarded due to low response, but increased sensitivity to partially bleached populations. Equivalent dose (De ) distributions in histograms and cumulative probability plots were examined for evidence of asymmetry symptomatic of partial bleaching (Olley et al., 1998, 1999). Most De distributions were roughly Gaussian in nature, with some distributions slightly skewed toward higher values (Fig. 3). In addition, standardized De -signal intensity plots (Colls, 1999; Colls et al., 2001) had low correlation (r2 ¼ 020:2) and De ðtÞ plots (Bailey et al., 2002) showed no increase in De with measurement time, both indicating a lack of evidence for partial bleaching. This was surprising, given the glacio-fluvial environment and high suspended sediment concentrations in the river system, as
T.M. Rittenour et al. / Quaternary Science Reviews 22 (2003) 1105–1110
1108
MIS 06-28-37 av = 19.4 +/- 0.4, n = 32
MIS 06-23-20 av = 28.1 +/- 0.7, n = 27
8
12
Frequency
Frequency
16
8 4 0
6 4 2 0
0
10
20
30
0
10
Cumulative Probability Curve
20
30
40
50
Cumulative Probability Curve 4
3 Arbitrary units
Arbitrary units
6
4
2
2
1
0
0 0
10
20
0
30
10
De (Gy)
20
30 De (Gy)
40
50
Fig. 3. Histograms (upper) and cumulative probability curves (lower) for two samples with individual De and associated errors for aliquots plotted in rank order. Bin width for the histograms is 2 Gy. For the cumulative probability curves, the sum of the normal distributions (73s) from the aliquots is plotted in black with the distribution calculated from the mean and standard deviation in grey. Units are arbitrary and reflect increasing probability. Table 2 Optical age estimates and De values from LMV channel belt samples, De errors reported as standard error 14
C age (cal. ka)a
Site no.
Sample no.
UNL Lab no.
Dose rate (Gy/ka)
No. of aliquots
De (Gy)
s
Optical age (ka)
Pve 22 8 20
MIS 06–24–22 MIS 03–27–8 MIS 06–23–20
UNL-120 UNL-168 UNL-346
1.870.08 1.870.05 1.670.04
28 45 27
36.470.8 33.670.6 28.070.7
4.1 4.1 3.8
19.771.1 18.570.8 17.870.8
19.7570.65b
Pvl 2 47 44 45 51
MIS MIS MIS MIS
07–03–47 06–31–44 06–31–45 07–05–51
UNL-125 UNL-343 UNL-124 UNL-353
1.870.05 1.870.04 1.870.05 1.770.05
28 27 30 34
28.970.5 27.870.4 26.670.5 25.870.4
3.4 2.2 2.6 2.4
16.170.7 15.870.6 15.170.7 15.070.7
>13.1370.46c >12.9370.30d
Pvl 1 37 39 16 18
MIS MIS MIS MIS
06–28–37 06–29–39 06–22–16 06–22–18b
UNL-169 UNL-123 UNL-119 UNL-335
1.670.06 1.870.06 1.870.05 2.070.05
32 29 20 43
19.470.4 21.970.6 22.570.7 24.270.5
2.5 3.1 2.9 3.2
12.570.7 12.470.6 12.270.6 12.170.5
>10.2270.51e >9.8870.33f
Radiocarbon ages were collected by previous workers from other sites and represent the age constraint on the channel belt as a whole. Site numbers refer to sample localities in Fig. 2. a Radiocarbon ages converted to calendar years by CALIB 4.3 (Stuiver et al., 1998a, b). b Blum et al. (2000). c Wesnousky and Leffler (1992). d R.W. Graham, personal communication reference in Blum et al. (2000). e Guccione et al. (1988). f King and Allen (1977).
indicated by thick slackwater deposits (Robnett, 1997; Blum et al., 2000). Despite these conditions, long transport distances on the order of several hundred kilometers appear to have been sufficient for complete sediment bleaching.
Due to the limited evidence of partial bleaching, the mean De was used to calculate the optical ages (Table 2). This was justified by the nearly Gaussian De distributions and close agreement between the mean De ; median De and the peak cumulative probability De (see Fig. 3).
T.M. Rittenour et al. / Quaternary Science Reviews 22 (2003) 1105–1110
1109
Table 3 Comparison of ages proposed by different researchers. All ages in calendar ka Surface
Fisk (1944)
Saucier (1994a)
Blum et al. (2000)
Optical age
Pve Pvl 2 Pvl 1
12–5 12–5 12–5
60–80 25–12 earlier 25–12 later
ca. 21 ca. 13.5–12.5 ca. 12.5–10.7
19.7–17.8 16.1–15.0 12.5–12.1
4. Optical ages
6. Conclusions
Optical ages and radiocarbon age constraints for the three channel belt surfaces sampled are shown in Table 2. The highest channel belt surface, mapped as Pve, produced optical ages between 19.7 and 17.8 ka, whereas samples from Pvl 2 were between 16.1 and 15.0 ka, and the lowest channel belt, Pvl 1, produced ages between 12.5 and 12.1 ka. The spread in ages is consistent with a braided system that has occupied a channel belt for a period of time. Age estimates from these surfaces are in agreement with geomorphic relationships and previously determined radiocarbon ages. Of special note is the good agreement between the optical ages for the Pve channel belt and the radiocarbon age from a surface thought to correlate with Pve (Table 2). Other published radiocarbon ages are minimum age estimates for the surfaces; nevertheless the optical chronology agrees with these ages.
The application of optical luminescence techniques has provided the first detailed chronology for the large braided channel belts of the lower Mississippi alluvial valley. Previous chronologies have been based on relative age relationships and a limited number of radiocarbon ages. The main deterrent to obtaining age estimates in the past was the lack of datable organic material in the channel belt sediments. By contrast, the luminescence technique is ideal for the LMV since the channel belts are primarily composed of quartz-rich sand, and provide a nearly limitless selection of suitable sample localities. Samples from three previously mapped cross-cutting braided channel belts were collected to test previous age models developed for LMV evolution. Optical ages suggest that the highest channel belt, mapped as Pve by Saucier (1994a) is 19.7–17.8 ka, the Pvl 2 channel belt is 16.1–15.0 ka, and the lowest channel belt, Pvl 1, formed 12.5–12.1 ka. Despite glacial meltwater conditions in the Mississippi during the late Pleistocene, the samples show minimal evidence of partial bleaching. Optical ages agree with relative geomorphic relationships and the available radiocarbon ages, and suggest the channel belts formed in the late Pleistocene during MIS 2 under conditions of high meltwater discharge. These new data support the glacial forcing hypothesis of Saucier (1994a) and Blum et al. (2000) for channel belt formation in the northern LMV. The new optical chronology presented here has allowed previous models of LMV landform evolution to be tested. The ages obtained here are the first step towards a larger research objective to develop an optical chronology for the entire LMV. Once completed, it will be possible to resolve the spatial and temporal response of this continental-scale river system to adjustments in hydrology and sea level during a full glacial-interglacial cycle.
5. Comparison to previous channel belt chronologies The optical ages presented here allow previous models for LMV channel belt formation to be re-examined. To reiterate from above, under Fisk’s (1944) model, all braided channel belts formed during the early to middle Holocene in response to sea-level rise (Table 3). Saucier (1994a, b; 1996) in contrast suggested that the channel belts formed in response to glacial discharge. In relationship to the specific channel belts sampled here, Saucier (1994a) placed the Pve channel belt in MIS 4 and Pvl 2 and Pvl 1 in MIS 2 (Table 3). Finally, Blum et al. (2000) proposed that the Pve surface formed during the last glacial maximum (ca 21 ka) and concurred with Saucier (1994a, b) that Pvl 2 and Pvl 1 formed during MIS 2 deglaciation between ca 13.5 and 10.7 ka (Table 3). The optical ages presented here support a MIS 2, last glacial maximum, age for the Pve channel belt, as suggest by Blum et al. (2000). Furthermore, new optical ages greatly refine age estimates for Pvl 2 and Pvl 1, suggesting they are slightly older than what has been inferred from minimum ages obtained by the radiocarbon technique (Table 2). By providing the first numerical age estimates for LMV channel belts, these new optical ages provide an opportunity to move towards linking periods of channel belt construction and abandonment with various forcing mechanisms.
Acknowledgements This research was funded by a Student Research Grant from the Geological Society of America and a Geological Society of America J. Hoover Mackin Research Grant awarded to T. Rittenour, a NASA
1110
T.M. Rittenour et al. / Quaternary Science Reviews 22 (2003) 1105–1110
Earth System Science Student Fellowship and a National Science Foundation Doctoral Dissertation Grant (SES 0082226) awarded to T. Rittenour and M. Blum, and a National Science Foundation Research Grant (EAR 0107028) awarded to M. Blum.
References Aitken, M.J., 1998. An Introduction to Optical Dating, The Dating of Quaternary Sediments by the Use of Photon-stimulated Luminescence. Oxford Press, Oxford, 267pp. Bailey, R.M., Singarayer, J., Ward, S., Stokes, S., 2002. Identification of partial bleaching using De as a function of illumination time. 10th International Conference on Luminescence and Electron Spin Dating, Reno, NV. Programs and Abstracts, p.2. . Blum, M.D., Tornqvist, T.E., 2000. Fluvial response to climate and sea-level change: a review and a look forward. Sedimentology 47, 2–48. Blum, M.D., Guccione, M.J., Wysocki, D.A., Robnett, P.C., Rutledge, M., 2000. Late Pleistocene evolution of the lower Mississippi Valley, Southern Missouri to Arkansas. Geologic Society of America Bulletin 112, 221–235. Chappell, J., Omura, A., Ezat, T., McCulloch, M., Pandolfi, J., Ota, Y., Pillans, B., 1996. Reconciliation of late Quaternary sea-levels derived from coral terraces at Huon Peninsula with deep sea oxygen isotope records. Planetary Science Letters 141, 227–236. Colls, A.E.L., 1999. Optical dating of fluvial sediment from the Loire Valley, France. Unpublished M.Sc. Thesis, University of Oxford, Oxford. Colls, A.E., Stokes, S., Blum, M.D., Straffin, E., 2001. Age limits on the Late Quaternary evolution of the upper Loire River. Quaternary Science Reviews 20, 743–750. Fisk, H.N., 1944. Geological Investigation of the Alluvial Valley of the Lower Mississippi River. US Army Corps of Engineers Mississippi River Commission, Vicksburg, MS, 78pp. Gemmell, A.M.D., 1999. IRSL from fine-grained glacifluvial sediment. Quaternary Science Reviews 18, 207–215. Guccione, M.J., Prior, W.L., Rutledge, E.M., 1988. Crowley’s Ridge, Arkansas. In: Hayward, O.T. (Ed.), South-Central Section of the Geological Society of America, Centennial Field Guide. Geological Society of America, Boulder, CO, pp. 225–230. Hutt, G., Jungner, H., 1992. Optical and TL dating on glaciofluvial sediments. Quaternary Science Reviews 11, 161–163. King, J.E., Allen Jr., W.H., 1977. A Holocene vegetation record from the Mississippi River valley, southeastern Missouri. Quaternary Research 8, 307–323.
Licciardi, J.M., Teller, J.T., Clark, P.U., 1999. Freshwater routing by the Laurentide ice sheet during the last deglaciation. In: Clark, P.U., Web, R.S., Keigwin L.D. (Eds.), Mechanisms of Global Climate Change at Millennial Time Scales. American Geophysical Union, Washington, DC, Geophysical Monograph 112, 177–201. Mejdahl, V., Funder, S., 1994. Luminescence dating of late Quaternary sediments from East Greenland. Boreas 23, 525–535. Murray, A.S., Wintle, A.G., 2000. Luminescence dating of quartz using an improved single aliquot regenerative-dose protocol. Radiation Measurements 32, 57–73. Murray, A.S., Olley, J.M., Caitcheon, G.C., 1995. Measurement of equivalent doses in quartz from contemporary water-lain sediments using optically stimulated luminescence. Quaternary Science Reviews 14, 365–371. Olley, J., Caitcheon, G., Murray, A., 1998. The distribution of apparent dose as determined by optically stimulated luminescence in small aliquots of fluvial quartz: implications for dating young fluvial sediments. Quaternary Science Reviews 17, 1033–1040. Olley, J., Caitcheon, G., Roberts, R.G., 1999. The origin of dose distributions in fluvial sediments, and the prospect of dating single grains from fluvial deposits using optically stimulated luminescence. Radiation Measurements 30, 207–217. Prescott, J.R., Hutton, J.T., 1994. Cosmic ray contributions to dose rates for luminescence and ESR dating: large depths and long-term time variations. Radiation Measurements 23, 497–500. Robnett, P.C., 1997. Late Quaternary occupation and abandonment of the Western Lowlands valley train course: Mississippi River, southeastern Missouri. M.S. Thesis, Southern Illinois University, Carbondale. Saucier, R.T., 1994a. Geomorphology and Quaternary Geologic History of the Lower Mississippi Valley. US Army Corps of Engineers Mississippi River Commission, Vicksburg, MS, 364pp. Saucier, R.T., 1994b. Evidence of late-glacial runoff in the Lower Mississippi Valley. Quaternary Science Reviews 13, 973–981. Saucier, R.T., 1996. A contemporary appraisal of some key Fiskian concepts with emphasis on Holocene meander belt formation and morphology. Engineering Geology 45, 67–86. Stuiver, M., Reimer, P.J., Bard, E., Beck, J.W., Burr, G.S., Hughen, K.A., Kromer, B., McCormac, F.G., van der Plicht, J., Spurk, M., 1998a. INTERCAL98 Radiocarbon age calibration 24,000–0 cal. BP. Radiocarbon 40, 1041–1083. Stuiver, M., Reimer, P.J., Braziunas, T.F., 1998b. High-precision radiocarbon age calibration for terrestrial and marine samples. Radiocarbon 40, 1127–1151. Teller, J.T., 1990. Volume and routing of late-glacial runoff from the southern Laurentide Ice Sheet. Quaternary Research 34, 12–23. Wesnousky, S.G., Leffler, L., 1992. The repeat time of the 1811 and 1812 New Madrid earthquakes: a geological perspective. Seismological Society of America Bulletin 82, 1756–1785.