the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 19 Sep 2025
Meltwater, mud, and the Mississippi: Upper Mississippi River Valley slackwater sediments reveal shifting deglacial meltwater sources associated with the Marquette Readvance of the Laurentide Ice Sheet
Abstract. Meltwater from the Laurentide Ice Sheet changed course frequently over the last deglaciation, with impacts on local landscapes and global climate. However, a scale gap has confounded efforts to trace the sources and pathways of this meltwater: well-dated paleoceanographic records lack the precision of regional geological and geomorphic evidence, but local features are often dispersed and difficult to precisely date, making it challenging to construct a spatially distributed source-to-sink meltwater timeline. Here, we bisect this scale gap by using the upper Mississippi River network as a local integrator of upstream meltwater routing. We analyze a ~400 cm core retrieved from a fluvial terrace near the mouth of the Whitewater River, a tributary to the Upper Mississippi River in southeastern Minnesota, where high water from meltwater floods deposited slackwater sediments. Optically stimulated luminescence (OSL) dates of 11.67±1.51 ka and 11.56±1.44 ka indicate that these slackwater sediments likely coincide with the Marquette Readvance of the Superior Lobe, a period when Laurentide Ice Sheet ice briefly reoccupied the Superior Basin, rerouting deglacial meltwater to the south. We geochemically fingerprint these sediments by applying Principal Component Analysis (PCA) and k-means clustering to X-Ray Fluorescence data, and find that Marquette-age meltwater flowed to the Mississippi first from the Glacial Lake Agassiz basin (rich in Ca) via the modern Minnesota River valley, and then from the Superior basin (rich in Fe, Cu, and Ni) via the glacial-stage St. Croix River. Weathering and post-depositional alteration obscure geochemical indicators of provenance in the top 200 cm of the core, which records the remainder of Marquette-age flood inundation. This study brings the pattern of ice-sheet retreat and associated meltwater routing into sharper focus and highlights the power of alluvial stratigraphy as an intermediary between local glacial geology and the marine sedimentary record.
- Preprint
(12032 KB) - Metadata XML
-
Supplement
(253 KB) - BibTeX
- EndNote
Status: open (until 14 Nov 2025)
- RC1: 'Comment on egusphere-2025-3920', Anonymous Referee #1, 21 Oct 2025 reply
-
RC2: 'Comment on egusphere-2025-3920', Anonymous Referee #2, 28 Oct 2025
reply
COMMENTS ON PENPHRASE ET AL. MANUSCRIPT egusphere-2025-3920
GENERAL COMMENTS
This paper presents some good data but is not ready to be published. Although well-written, with good goals stated (lines 53-58), the paper has numerous shortcomings, the most important being the tenuous chronology of the sequence and its questionable links with external events. With only one OSL date below the weathered zone, the age extrapolation below the OSL date could have a variety of interpretations. I don’t think the authors have chosen the best correlation of sequence with events in the Agassiz and Superior watersheds; I’ve elaborated on this below in SPECIFIC MAJOR CONCERNS.
I think a different, and perhaps shorter, title would be better. Your paper is more than just the recognition of the Marquette readvance, so I’d omit that part. Although the 3 M words in the title Meltwater, Mud, and Mississippi) are cute, I wonder if they are worth including. What about something like “History of late-glacial overflow from the Superior and Agassiz watersheds as recorded in a slackwater sequence in the Upper Mississippi River basin”?
Attributions of previous research on paleohydrological changes related to LIS meltwater flows are not adequate and many important references about continental & marine events should be added; see some suggested additions in the section below on MINOR CONCERNS.
SPECIFIC MAJOR CONCERNS
75-79 I don’t follow what you’re describing. How does a tributary pond breach a hydrological barrier in the main Mississippi River channel?
244 As you indicate, your OSL date at ~150 cm in the “visibly weathered” upper zone is not reliable. You might mention why you concluded this, such as the uncertainty of the value used for water content, root disturbance, etc. So, why did you not get an OSL date from the lower part of the core (e.g. 270-375 cm), where there are a number of changes in the elements and their trends shown in Figure 5?
Section 5 I do not agree with many of your interpretations/correlations in this section that relate zones in the core to the time of specific events in the Lake Agassiz and Lake Superior basins. With only 1 date below the weathered zone, there is no chronological control below the 11.56 +/-1.5 ka at 247 cm depth. My comments below begin with a disagreement on the age of 11.7 ka that you assign to the blocking of the eastern outlets of Lake Agassiz by the Marquette glacial readvance in the Superior basin. Indeed, the ice may have been readvancing in the Superior basin then, as you note by dates on the Gribben Forest Bed, but others have concluded that the blocking of the Agassiz outlets was closer to 11.4 ka (e.g. Clayton, 1983; Fisher, 2020; Fisher & Breckenridge, 2022). The text and Figure 7 should reflect that. Why not begin the story by assigning the clayey sediment just above the locally-derived sandy zone (with a high Si element content) to the first overflow events from Agassiz & Superior? For these reasons, and others noted below, your Phase interpretations need to be changed. Some specifics and suggestions follow:
352-358 I disagree with the words “following the initiation of the Marquette Readvance”. You’ve left out 2500 years of the early phases of Lake Agassiz (the Lockhart and Moorhead). The wording and interpretation in your Phases 2-4 should be changed, or at least an alternative chronology and correlation should be given.
361- It doesn’t seem to me that there is any significant rise or trend in Lake Superior basin major elements in the core just above 350 cm (Fig. 5) as you state, so I wouldn’t conclude that the interval above 350 relates to the Marquette readvance and its blocking of the eastern outlets of Lake Agassiz. However, there is a rise above 310 cm, which I’d argue is the first re-direction of Lake Agassiz overflow into the Superior basin and then into the Mississippi River watershed (i.e. the shift from the Lockhart Phase to Moorhead Phase of Agassiz). Then, around 280 cm, after a peak in Superior basin elements, more Ca-rich waters appear in the core which dates the return of Agassiz overflow thru its southern outlet (River Warren) into the Mississippi basin related to the Marquette ice blocking its eastward overflow into the Superior basin—this interval is also reflected by the low Superior element input that then gradually rises as the Marquette ice retreats (275-250). Following this, there is a diminished Superior element input to the core site that relates to the re-direction of overflow eastward thru the Great Lakes from the Superior watershed.
Using the 11.4 ka (vs 11.7 ka) age for the blocking of the eastern outlets of Lake Agassiz by the Marquette, and using the same assumption that you did about the OSL age of 11.56 +/- 1.5 ka being too old, one could use a date of 10.06 ka at 247 cm depth (i.e. 11.56-1.5 = 10.06 ka) for when major element contributions from Lake Superior last peaked at the core site, as shown in Fig. 5. Below would be my suggested interpretation of Phases at the core site based on the above comments. Whatever your conclusions are, the Phase zones should be shown on Figure 5.
Phase A--below 400 cm: dominated by local sand (high Si) = pre-Agassiz and Superior watershed meltwater contributions (~pre-14 ka).
Phase B—400-315 cm: blend of Ca (Agassiz) and Fe, Cu, Ni (Superior) contributions from 2 watersheds, beginning ~14 ka and ending ~12.9 ka; Lockhart Phase of Lake Agassiz.
Phase C—315-280 cm; decline in Ca (Agassiz), rise in Superior elements with a peak near the end about 280-285 just as Marquette ice invaded Superior basin; Agassiz overflow east into Superior; Moorhead Phase of Agassiz when its waters stopped overflowing south into the Minnesota-Mississippi River basin, 12.9 to 11.4 ka.
Phase D—280-270 cm: jump in Ca (Agassiz), drop in Fe, Cu, Ni (Superior) elements during time when Marquette readvance blocked E outlets of Agassiz for a few hundred years.
Phase E—270-250 cm; decline in Ca, increase in Fe, Cu, Ni as Marquette ice retreats and Agassiz waters resume overflow eastward into Superior basin; Emerson Phase of Lake Agassiz ~10.7-10 ka.
Phase F—250- ; LIS retreats in Agassiz and Superior basins and overflow to core site from them ends.
410-415 I don’t think your conclusions here are the best fit with the major element concentrations shown in Figure 5—above is my interpretation, although there may be alternatives for part of the sequence. And, a more elaborate statement and expansion of your conclusions should be presented here, including some of the rationale and plot of your Phases on Fig. 5.
MINOR CONCERNS (with line references)
By definition, throughout the paper “glacial” Lake Agassiz, should not be Glacial Lake Agassiz.
20 “remapped”?
20-21 Lots of needed references to the shaping of the landscape morphology in Canada such as by Fisher, Teller, etc.
21 Curious short selection of references to how these meltwaters “impacted glacial climate”; Broecker et al, (1989 Nature) as a start and modellers and continental Quaternarists should be added.
24 Lots of continental research exists that is related to the history of Lake Agassiz that spans hundreds of kilometers (e.g. Teller, Fisher, Breckenridge, Colman, Leverington, etc.)—some should be cited.
25 Many “marine sediment cores” have been studied with this in mind from the North Atlantic to the Gulf of Mexico to the Arctic Ocean, and some should be cited here.
40 Fisher (2020 QSR) and/or Fisher & Breckenridge (2022 QSR) would be better to cite here or some of Teller’s articles.
44 other Gulf of Mexico studies should be mentioned (e.g. Aharon 2003 Paleoceanography; Broecker et al., 1989; Flower, etc.).
70-71 not sure this 10.6 ka age is what recent research has concluded, and 10.0 ka may be closer to the time
74 “observations” à studies of stratigraphic records along the river
84 need at least a couple of references of Canadian authors from the Canadian side where the bulk of research on Lake Agassiz has occurred.
85 Gran et al. (2013) citation needs full reference in References section.
98 other references might be worth including here and for following sentences (e.g. Flock, 1983 QSR; Curry et al. 2014 QSR)
103 downstream in? our study area (OR, downstream of our study area in . . . . ?)
Section 4.2 A standard (simple) stratigraphic section description would be helpful here, with OSL ages on it.
240 gray clay laminae (vs “beds”)
283 delete “with depth”
299 use “reflect” for the second word “capture”
318 “markedly different”—describe how
319 “. . . these grains are markedly similar ?in appearance? to local . . . . ??? A little elaboration on how they are similar would be worthwhile (frosted? rounded? stained?).
326 We correlate the locally derived fluvial . . . .
361 AfteràAbove
382 Is the St. Croix route low enough at this time to allow overflow from Superior?
Citation: https://doi.org/10.5194/egusphere-2025-3920-RC2
Data sets
Core Segment Images, X-Radiographic Images, and X-Ray Fluorescence Data of Slackwater Sediments, Whitewater River, Upper Mississippi River Valley, USA S. Penprase et al. https://doi.org/10.5281/zenodo.16779577
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 940 | 40 | 12 | 992 | 16 | 7 | 11 |
- HTML: 940
- PDF: 40
- XML: 12
- Total: 992
- Supplement: 16
- BibTeX: 7
- EndNote: 11
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This is a well-written manuscript on an interesting topic related to meltwater and glacial advances in the Upper Mississippi River Basin.
I do have a few significant concerns and comments (I am also attaching a PDF with detailed comments).
Overall, I would say these are moderate revisions, many of which can be addressed with changes in wording and better explanations. But some of the interpretations may need to be made more lightly, until additional data can be collected in the future.
One of the most significant concerns is that the discussion and conclusions of the manuscript are based solely on one core (the new data). I do think that the geochemical approach is well done and that the contrasting provenance between the sources is quite good. However, the fact that all the geochemical data is from only one 4 m core (and only the lower 2m of it), is somewhat concerning. There should be a caveat about this in the discussion that more cores are needed in the future to test these ideas.
Secondly, the chronology is based solely on 2 OSL ages (with 1.5 ka error). Althought the ages do seem reasonable, it is not clear how the water content estimates were made. The data Table 3 indicates a water content of 25 % was used, but the methods are unclear has this was made from its burial history and the in-situ measurements. Please provide more detail on how calculated. Overall, is the precision of these OSL ages good enough to make a correlation to the Marquette Readvance ? Is the Marquette Readvance the same as the Lakeview Phase that deposited the Miller Creek Fm. in NW Wisconsin (see Syverson et al., 2011. Wisconsin Stratigraphy Lexicon) ?
Third concern: the geochemical data is all on the bulk sample and not a grain size fraction. Thus, it is not clear to what degree the changes in the geochemistry are from source area shifts or from grain size partitioning from the sorting process (with more Si in the sand fraction and more Fe, Ni, Cu in finer fractions --- which is common). It would be important to show the grain size data (if you have it) in the graph next to the elemental data. Please add more on potential pitfalls with the methods related to this. Although the provenance idea may well be the correct interpretation, it is certainly likely that there is also a strong effect from mineral partitioning with grain size. It would be a good test to run a sample with different grain size fractions to prove the case that the elemental trends would hold even if narrow grain size fractions were used rather than bulk core. If unable to do this, then there should be a statement about the possibility that grain size may significantly affect the elemental data and complicate source area interpretations. The provenance data is really nice overall, with a strong contrast between the sources but want to consider and rule out other possibilities.
Fourth, a good core decscription is needed and would be appropriate to put in the Appendix. The photos of the core are great. However, this doesn't replace the need for a standard core description --- especially since only one core is presented here. Please add.
Minor issues or suggestions:
-- use "southern Laurentide Ice Sheet" in title and other areas
-- inconsistent use of LGM term; may want to use last glaciation or Wisconsin Episode in some areas or better define LGM
-- OSL ages could be rounded to one decimal perhaps, given the errors involved
--- some sentences have strange or contradictory wordings; the writing is generally good, but some sentences are confusing and long-winded with too many thoughts and might be better to split into two sentences
Line 3: "well-dated paleoceanographic records" --- make new sentence
Line 12: use "Mississippi River"
Fig. 1 --- show outline of glacial Lake Duluth ? was it present at the time ?
Some additional papers to perhaps cite or investigate:
Flock, M.A., 1983. The late Wisconsinan Savanna Terrace in tributaries to the upper Mississippi River. Quaternary Research, 20(2), pp.165-176. [This paper notes red and gray fine-grained beds in the terrace from perhaps Lake Superior and Lake Agassiz sources --- with late glacial age]
Johnson, M.D., Addis, K.L., Ferber, L.R., Hemstad, C.B., Meyer, G.N. and Komai, L.T., 1999. Glacial Lake Lind, Wisconsin and Minnesota. Geological Society of America Bulletin, 111(9), pp.1371-1386. [another example of glacial lake studies in the region]
Hobbs, H.C. and Breckenridge, A., 2011. Ice advances and retreats, inlets and outlets, sediments and strandlines of the western Lake Superior basin. Field Guide. https://doi.org/10.1130/2011.0024(14) [more details on controversy of Marquette Advance]
Curry, B.B., Hajic, E.R., Clark, J.A., Befus, K.M., Carrell, J.E. and Brown, S.E., 2014. The Kankakee Torrent and other large meltwater flooding events during the last deglaciation, Illinois, USA. Quaternary Science Reviews, 90, pp.22-36. [slackwater lake sediment records on the Illinois River valley]
Hajic, E.R., 1993. Geomorphology of the northern American Bottom as context for archaeology. Illinois Archaeology, 5(1), pp.54-65. [C14 ages on red clay beds in St. Louis region]
Grimley, D.A. , A.C. Phillips, and S.W. Lepley, 2007, Surficial Geology of Monks Mound Quadrangle, Madison and St. Clair Counties, Illinois: Illinois State Geological Survey, Illinois Preliminary Geologic Map, IPGM Monks Mound-SG, 1:24,000 [red clay beds in the American Bottoms of the St. Louis area --- Miss. Valley --- occur at the Holocene-Pleistocene contact and are similar in age to the red clay beds noted in this manuscript]
Porter, D.A. and Guccione, M.J., 1994. Deglacial flood origin of the Charleston alluvial fan, Lower Mississippi alluvial valley. Quaternary Research, 41(3), pp.278-284. [similar timing to the dated events in Minnesota - any relation ?]
Syverson, K.M., Clayton, L., Attig, J.W. and Mickelson, D.M., 2011. Lexicon of Pleistocene stratigraphic units of Wisconsin. Wisconsin geological and natural history survey technical report, 1, p.180. [might be good to correlate unit and phases to the names used in this report --- such as Miller Creek Fm. and Lake View Phase]