Lake Bonneville and the Wasatch Fault – new theories and new paradigms yield insights into present day hazards in other regions of the world

Three new theories challenge the assumptions underlying 150-years of research regarding Lake Bonneville and extend and redefine the history of this late-Pleistocene/early-Holocene lake. These new theories have relevance to current day hazards in many areas of the globe and are important to our understanding of the climate of the western United States. The lake’s level history and shorelines have presented a confusing array of conflicting data, which has universally and incorrectly been attributed to abrupt and temporary climate oscillations. The Earthquake-induced Surging Theory explains misunderstood lake features, extends the lake level data back to 40kya, and explains the Bonneville Flood, confirming a 17.4kya (cal) date for that event. The Isostatic Rebound Pop Seiche Theory explains the “Intermediate Shorelines” first identified by G.K. Gilbert with a shocking twist regarding timing. This theory teaches us something of importance regarding glacial lakes forming today. The Bear River Exclusion Theory explains the anomalously rapid fall from the Provo Level and resolves the early/late Provo Level controversy. This last theory is going to be important for addressing the future of the Great Salt Lake.


Introduction
Lake Bonneville was a large, late-Pleistocene lake with the eastern edge bordering the Wasatch Fault in the Great Basin of the western United States. (Figure 1) Rhode (2016) studied flora and fauna in the sediment record. He was able to correlate changes in the region's biology with the fall from the Provo level and during the Gilbert episode. The study mentions the well-accepted theories of the Stansbury Oscillation and the existence of other oscillations leading to the last glacial maximum and the Bonneville Flood period but does not provide evidence of changes in flora and fauna that might support the idea that these were dramatic climate events.
The spits and bars in Lake Bonneville were first studied by G.K. Gilbert (1890). Dr. Paul Jewell (2007) studied the spit elevations, angles, and magnitudes to gain insights into climate variations. He found a strong correlation between the angles, the magnitudes, and the fetch (the length of open water in which waves can build). However, the largest spits did not follow the prevailing wind patterns and thus were interpreted as being the result of catabatic winds off the Canadian ice sheets during the 'climate fluctuations' identified by others. There have been several papers published based on these assumptions. Felton, Jewell, Chan and Currey (2006) did a survey of tufa deposits from Lake Bonneville. The tufa deposits vary in nature and chemistry. In studying the tufa deposits, they found that "tufas are prevalent on headlands and windward sides of islands that were exposed to high wave energy" and that "tufa commonly occurs at basin thresholds, where water is moving between a restricted subbasin and the main body of the lake." (Ibid, p. 338) As would be expected, they found that local chemistry was an important element; calcium had to achieve a high concentration for significant deposits to form.
Others have also studied the tufa deposits, and a common theme is acceptance of the Stansbury Oscillation as representing an extremely dry period with extensive evaporation in the lake which raised the concentration of calcium and other minerals in the water (Nelson, et. al., 2005).
Climate, the size of the lake in the Bonneville basin, and snow in the Wasatch are inexorably linked. The last glacial maximum in the Wasatch was the Pinedale glaciation which coincided with the maximum of the Laurentide and Scandinavian ice sheets. At the glacial maximum, icebergs were being calved into Lake Bonneville at the mouth of Little Cottonwood Canyon in the Salt Lake Valley. The glacial till deposited in that area provides a valuable geological record since it forms a blank canvas upon which were etched events from the Wasatch Fault and Lake Bonneville over the last 22ky. Janecke and Oaks (2011) of Utah State University have done extensive research on the Bonneville and Provo level thresholds of Lake Bonneville at the north end of Cache Valley. In a 2011 Geosphere paper they included a section titled "Did an earthquake, overland flow, or sapping, trigger the Bonneville Flood?" They hypothesized that "An earthquake on the Riverdale fault (or on some other Cache Valley fault) could have produced seiche waves that overtopped the Zenda sill with high-velocity waters of sufficient energy to destabilize that dam, to breach part of the Zenda sill, to rapidly incise the length of the dam, or to cause other damage." (Ibid, p. 1387) In a fascinating series of thought experiments, they also allow that a rapidly rising lake level might have triggered seismic activity that caused the flood, or that the flood might have caused seismic activity. They indicate that seismic activity might result in failure of the deltaic deposits of the Zenda threshold due to landslides or due to failure of sediments in the delta weakened by sapping. Qeios, CC-BY 4.0 · Article, May 29, 2023 Qeios ID: G4DAH0 · https://doi.org/10.32388/G4DAH0 6/75 Janecke and Oaks put forth an argument that there was "at least episodic overland flow from the Bonneville level", opening the door for the extended high stand argument. Reasoning that "To argue otherwise would require the unlikely coincidence of Lake Bonneville rising and falling repeatedly to within a few meters of an overflow and stabilizing there as a closed basin for a protracted period of time." (Ibid, p. 1384) They also provide an analysis of the deposits of the Marsh Creek delta at the Zenda outflow and explain why the structure of those deposits might be able to support an extended period outflow.
In a 2020 paper, Oviatt gives an extensive argument for the case of the Zenda threshold failing quickly after being overtopped by the rising Lake Bonneville. The central point to his argument is that the prominent depositional shoreline benches of Lake Bonneville could be built up over time and did not represent an extended high stand.
In a 2020 GSA Connects presentation, I introduced a theory that the well-accepted "graben" at the mouth of Little Cottonwood Canyon in the Salt Lake Valley, Utah was instead a fissure, formed when earthquake-induced surging in Lake Bonneville shifted over 15km² of glacial till deposits in the area. The stable glacial till deposits shifted as large blocks sliding on the underlying lake-transgression sand layer which underwent liquefaction in the event. The point of failure was the intersection of the liquified transgression lakebed and the Wasatch Fault slip plane. The blocks shifted like puzzle pieces and the paper presented how those pieces could be resolved back into a cohesive initial state. The theory predicted fissures at multiple locations between different shifting glacial till masses and field examination proved those predictions correct. A written summary of that work was included in a post-meeting addendum provided to the GSA and that material is available on the acadamia.edu website. That work has not been peer-reviewed.
The 2020 presentation proposed that earthquake-induced, basin-wide surging caused the Bonneville Flood. In a presentation two years later at the 2022 GSA Connects conference, that work was expanded upon. This current paper adds both background and detail to what was presented at the 2022 conference and then expands on that with new findings and theories regarding Lake Bonneville.  The Little Cottonwood Canyon moraine boulder field is an upslope debris field from a tsunami. These types of deposits have been identified along ocean coastlines in other parts of the world. (Scheffers, 2008, 2021, Dewey, et al. 2021) G.K.
Gilbert identified a series of indistinct terminal moraines at the mouth of Little Cottonwood Canyon down-canyon from this point, however there are no boulder-wall-type terminal moraine bands remaining as might be expected. At the 2022 GSA Connects, I presented that when the Bonneville-Flood-event earthquake-induced surge occurred, it swept up the bands of terminal moraine boulders. The surge would have come from the open reach (also, fetch) to the northwest. Looking at the pattern of the boulder field, the most likely scenario is that at the time of the surge the Little Cottonwood glacier had receded well off its maximum depth at the canyon mouth. (Quirk, et al., 2020) A glacial tongue would still have extended down to or close to the shore of Lake Bonneville, but the bottom section of the glacier was a crumbled field of ice. ( Figure   5) When the surge hit, it lifted the lower 0.25km or more of the glacier and drove in underneath it, creating a high-pressure hydraulic blast into the moraine and carrying the terminal moraine boulders into that area. Where the lifted glacier cracked, streams of liquid and rocks were blasted higher into the moraine slope creating vertical spikes of deposits. The point running up the canyon where the glacier was too solid to yield is presented as a very sharp termination line of the boulder field area on the moraine face. Qeios, CC-BY 4.0 · Article, May 29, 2023 Qeios ID: G4DAH0 · https://doi.org/10.32388/G4DAH0 10/75 This boulder field gives us an approximate height of the surge at this point on the Wasatch front: about 50m. The purpose of leading with the boulder field is to convey that this was a massive event; the 2011 tsunami in Japan was estimated at 40m. There is evidence on the Bell Canyon moraine, just south of the Little Cottonwood moraine, that the waves ran about 100m up that slope (Location: 40.56659N, 111.79876W).

The Laabs Boulders
The southern lateral moraine at the mouth of Little Cottonwood Canyon is a very distinctive and high moraine, whereas the northern moraine is quite dispersed. On the north side, the canyon spreads out. Curiously, there are no distinct lines of moraine boulders in this area as one might expect from a glacier advancing and retreating. Instead, the moraine boulders are strewn across the hillside, labeled as the Laabs Boulders in Figure 3.
Dr. Benjamin Laabs and associates have done extensive dating of boulders in the Wasatch to track the glacial advances (Laabs, 2011, 2020, Quirk, et al., 2020. Their work is based on measuring the accumulation of terrestrial cosmogenic nuclides (TCNs) on exposed rocks. Through the years, that technology has advanced, and the assumptions refined. Two papers appeared in 2020, one by Laabs et al. and one by Quirk et al. which included Laabs as a co-author. The Quirk paper was focused just on the Wasatch, whereas the Laabs paper was looking at the dating throughout the western United States. The following discussion will present the Quirk paper data followed by the Laabs paper data in parenthesis.
The differences are not significant since the error margins overlap. The Quirk data will be used in the figures and detailed discussions.
Of particular interest is their dating of the glacial moraine boulders on the north and south sides of the mouth of Little Cottonwood Canyon in the Salt Lake Valley. On the very tall and well-defined southern lateral moraine the boulders were Qeios, CC-BY 4.0 · Article, May 29, 2023 Qeios ID: G4DAH0 · https://doi.org/10.32388/G4DAH0 11/75 dated at 20.8kya +/-2.2 ka (20.8kya +/-4.5 ka). That puts it in the timeframe of the Laurentide maximum. The error range on this data is over twice the typical error range from other areas they studied suggesting that multiple factors are in play at this location.
The dispersed boulders of the north side, however, were found to be deposited three thousand years later: 17.4kya +/-0.3 ka (17.3kya +/-0.7 ka). Note the much tighter error margin on the north side data, suggesting that the deposit was the result of a single, well-defined event, in spite of the fact that these boulders are strewn across the hillside. Quirk et al. attributed this more recent date to a glacial resurgence; however, a major resurgence that could deposit boulders this high up, on the other side of a knoll and around the corner from the low point of the canyon mouth and on the sunny southern exposure side of the canyon would have left evidence high on the other side of the canyon also, which it did not.
Reinterpreting the data using the earthquake-induced surging model of this paper, what the Laabs' team identified was the date of the Bonneville Flood. The northern lateral moraine boulders the team tested are about 15-20m above the Bonneville high stand shore and would have been swept up in the maelstrom of the surge (see Figure 5). That surge would have rolled and sandblasted the glacial boulders in the area, essentially resetting the isotope timestamp on the surfaces.
The 17.4kya (17.3kya) dating of those boulders puts the surge in the same time frame as the Bonneville Flood dating by others of 17.4kya based on features in Idaho and using earlier calibrations of TCN (Janecke and Oaks, 2011). Godsey reported the end of the Bonneville high stand at 17.5kya based on a calendar equivalent of ¹⁴C dating (Godsey, et al., 2005). In his work on Lake Bonneville sediments, Benson determined that the Bonneville Flood occurred before 17.0kya (Benson, et al., 2011).
There is a 'catch 22' in all of this. Cosmogenic nuclide dating requires calibration against a standard time reference of a well-documented event, and the Bonneville Flood has been used as just such a reference standard in this region. In this paper the 17.4kya (TCN) date will be assumed the most accurate date for the Bonneville Flood, with the understanding that as the science advances, dating estimates will continue to be refined. In this paper, cause and effect relationships will be used to determine relative dates; TCN and ¹⁴C data will be used as supporting data or placeholders with the realization that they are approximations.

The Keg Mountain Oscillation
A very distinctive level oscillation during the Lake Bonneville high stand has been documented by several researchers (Currey andBurr, 1988, Milligan andChan, 1998). Their work was based on radiocarbon dating of sediments and on shorelines at Keg Mountain in the southwestern region of the lake basin (see Figure 1). While the researchers agree that there was an oscillation when Lake Bonneville was at or near its high stand, the duration of the Bonneville high stand is a matter of debate. Currey and Burr (1998) presented one scenario: when Lake Bonneville was at or near its high stand (the early Zenda threshold), the outflow inexplicably dropped by 20m to the 'late Zenda threshold', this dropped was immediately follow by Qeios, In my 2020 GSA presentation, I suggested that lake-wide, earthquake-induced surging caused the Bonneville Flood and that it could have been misinterpreted as a climate oscillation. In such an event, the valleys along the Wasatch Fault on the east side of the lake would drop in the fault slip. The full depth of the lake would surge eastward to equalize. Once in motion, momentum would cause the surge to overshoot the high stand on the eastern side of the basin. This overshoot would be exacerbated by the basin topology, analogous to ocean swells approaching a shore where, as the water gets shallower, the energy gets concentrated into a smaller column of water. As with an ocean tsunami, the result is short frequency waves superimposed on a very long wavelength, high amplitude surge. The surge is what creates a long-term impact on the landscape.
On the western side of the basin, the surging or sloshing would draw the lake down from the high stand shoreline by the same amount that the surge on the eastern side moves up. At Keg Mountain in the western deserts, this is evidenced by shallow water sediments being drawn down into the lake and covering deeper water sediments, basically mimicking a climate oscillation lake level drop. As the basin sloshed back a very pronounced shoreline band or bar would be formed at the original shoreline elevation. This bar would have the same features as one formed by a massive storm or a tsunami.
The bar would top out above the normal waterline and the sediments would slope shoreward. If the return surging wave was rushing up a feature such as an alluvial fan, that shoreline band would take the form of an even-sided V-bar; in contrast, a typical a cuspid-foreland V-bar would have more of a cursive shape due to prevailing winds in the area.
On the eastern side, the surge up above the high stand shoreline would capture land-based flora and fauna and then draw it back down into deeper lake sediments. This again would confuse interpretation of the sediments and the use of carbon dating to establish lake levels. Date reversals in the sediments would be a common problem.
The surging in the lake would also have stirred up deep water sediments. As the lake returned to a quiescent state, these suspended solids would have settled out forming a sediment band which would mimic the passage of time. In a surge event such as this, which is essentially sloshing in a basin, at some point the water in the basin will slosh back to equilibrium near the original high stand level. Once the sloshing in the basin subsided, the solids stirred-up in the surging settled out before the lake level had time to fall from the effects of the Bonneville Flood. There is a 1km long, 6m high, horizontal step running across the face of the alluvial fan. The terrain above is smooth, and the terrain below is undulating, exactly as you would expect from a flood disrupted landscape. (Figure 6) Qeios, The Idaho Geological Survey has indicated this anomaly as the only visible section of a suspected normal fault in the area (DeVecchio, 2002). (Figure 7) However, if this were a fault, it might be expected to parallel the ridgeline above, which it does not. Instead, it runs perpendicular to the slope of the alluvial fan, which is a surface feature. This line is an artifact of the surge overtopping the natural dam of the alluvial fan.
From this data point, the main surge ran 30m above the Lake Bonneville high-stand and was about 3km across. The initial surge very quickly focused into a 2km wide erosion channel in the alluvial fan, part of which is still visible today. After the surge receded, the flood to proceeded in earnest in the erosion channel. The Red Rock Pass weir was the limiting factor once the Marsh Creek alluvial fan failed. The initial surge wore the pass down to the Late Zenda Threshold and the peak flow occurred at that point. Even as the surge fell, the ridgeline slopes continued to be undercut and a series of landslides occurred both during and after the flood, widening the pass (Eardley, et al., 1957).
Phase 3 -Post-surge flood. With the pass 27m below the Bonneville high stand, the flood was able to proceed in earnest even after the basin sloshing from the earthquake-induced surge settled out. The basin sloshing would have settled very rapidly due to the number of dashpots in the system such as the Cache and Provo Valleys and the very dispersed western extremes of the system.

The Stansbury Oscillation
Researchers have identified another abrupt 'climate oscillation' during the Lake Bonneville transgression sometime around 25,000 years ago called the Stansbury oscillation and shoreline (Eardley, et al., 1957, Oviatt, Currey andMiller, 1990). (See Figure 2) The widely accepted theory is that in the middle of a steady rise in lake level during the 10,000-year colder and wetter period which saw the Laurentide and other ice sheets grow, there was a brief period where the lake not only stopped rising but actually evaporated at an incredible rate, dropping 47 meters. The climate then immediately shifted to a wetter period and rose back to its previous level before immediately dropping again in another dry spike to the almost the same low shoreline, whereupon the climate turned wet again and returned the lake to the pre-oscillation level, at which point Lake Bonneville continued its steady transgression rate of rise that it had before the oscillation.
The theory also presents that at the two low points, the suspended calcium in this very large freshwater lake concentrated The Stansbury shoreline has been an enigma in Bonneville research, in part because in certain areas of the lake the shorelines are very pronounced and in others they are difficult or impossible to find. What is apparent in traveling the basin is that the shoreline is most pronounced where the surging would have resulted in strong currents resulting in rapid erosion or deposits or there were soft deposits in areas facing long fetches towards the lake's center.
In a study focused on the Stansbury 'climate oscillation', Oviatt, Currey and Miller (1990)  Bonneville at the Stansbury level superimposed shows exactly what the earthquake-induced basin surging model predicts: Stansbury Island near the center of the lake is a node with minimal elevation change during the oscillation, whereas in the western arms of the lake the elevation variation is approximately proportional to the distance to the center.
( Figure 9) The data these researchers presented is difficult to reconcile with a climate oscillation scenario but does support the theory of earthquake-induced-surging in Lake Bonneville.  The Stansbury Oscillation was an earthquake-induced surging event. In the 2022 GSA presentation I made the point that an earthquake-induced surging-type tsunami is different from the earthquake-induced shock-type tsunami. The first is a displacement phenomenon, which can travel at up to 80kph, the second is a shock wave that travels at 800kph. The 47meter swing of the Stanbury event is very close to the 50-meter swing of the later Bonneville Flood/Keg Mountain event. If this were a shock-type tsunami, the depth of the water would be more of a factor, and the two events would be quite different from what occurred. The fact that the oscillation amplitudes are similar also suggests that the fault displacements in the two events were similar.
On Stansbury Island, in the middle of the Bonneville basin, is one of the distinctive features of the Stansbury Oscillations: thick bands of tufaglomerates (cemented beachrock deposits) separated by bands of deep lakebed sediments (see Figure   8).
These bands led researchers to believe the lake had either evaporated to a higher concentration of dissolved minerals where shoreline waves resulted in CO2 degassing and calcium compound precipitation, or that algae contributed to the calcium coming out of solution in the shore zones (Oviatt, 1987). The issue with the first of these theories is that the lake was a very large freshwater lake, so it would take a lot of evaporation to meaningfully concentrate the ions. Also, the lake supported a healthy diversity of freshwater flora and fauna before and after the Stansbury Oscillation, where a dramatic shift in chemistry would be expected to be accompanied by a die-off. Two climate oscillations spanning hundreds of years each would be a significant shock, however in a review of the available literature, no references to appropriate biodiversity-shocks were found. The issue with the second theory is that algae-based tufa formation should be prevalent at many levels in the lake if it was a significant contributor and the evidence of that is missing.
High concentrations of calcium in lakebed sediments occur in areas of the lake where calcium has leached from abovelake deposits in the surrounding terrain. This is prevalent in the middle and western portions of the basin, away from the A surge flow would do a natural size selection. As the surge built towards peak velocity, large stones would be carried into Qeios,.0 · Article, May 29, 2023 Qeios ID: G4DAH0 · https://doi.org/10.32388/G4DAH0 20/75 the gulley and deposited, but the smaller stones, gravel and sand would be carried on, except for what might reside in interstitial spaces. As the surge slowed, the deposits would quickly change to smaller sized materials. This size distribution is quite evident in the Stansbury gulley deposits. In the layers adjacent to the large stone layers are thinner layers of tufa gravel, suggesting that small secondary surges followed the main surges.
Tufa shoreline deposits are a distinctive feature in the lake and define a number of the shorelines at very specific locations. There have been several studies of the tufa deposits (Felton, et al., 2006, Nelson, et al., 2005, Nelson, et al. 2005b). In the 2006 Felton study, they found that the thick tufa deposits tend to be in two types of areas: a. zones where a long fetch distance in the western reaches of the lake would result in high wave energy and b. at subbasin thresholds.
Felton stated that tufa formation at subbasin thresholds were "because reductions of water flow when the lake level dropped may have isolated the waters of subbasins" (Ibid, p. 385), this assumes the result was an increase calcium concentration as the subbasin dried up or influx of fresh water was reduced. However, both long-reach obstructions and pinch-points between subbasins are where there would be highly turbulent and aerated flows during an earthquakeinduced surging event, which would explain tufa deposits at those locations.
The Felton team also found the thick tufa deposits correlated with three primary lake levels: Stansbury, Bonneville, and Provo. The Bonneville and Provo levels were extended occupations, but the Stansbury level was not. What the three levels have in common is that they coincided with major earthquake-induced oscillations (the Provo level will be discussed later).
While tufa can be formed by a number of different processes, an indication of potential earthquake-induced surging is a tufa deposit layer in a sediment sequence where tufa deposits are not the norm.

The Stockton Bar and Spit
Evidence of the Bonneville Flood event surging is prevalent in shoreline features throughout the Bonneville basin.
Gilbert and others have discussed at length the bars and spits in the Bonneville record, attributing these to prevailing winds and monster storms. (Gilbert, 1890, Jewell, 2007 This needs to be re-examined. Large spits which follow potential surge flow and eddy patterns and deviate from prevailing wind patterns are probably from earthquake-induced surges. The Stockton bar and spit formation was first discussed by G. K. Gilbert. (Figure 10) It is being discussed here because not only is it a well-known and extensively studied feature, but it also stands as an excellent example of the type of analysis I am proposing as lake features are reconsidered in light of the earthquake-induced surging theory.
Qeios, CC-BY 4.0 · Article, May 29, 2023 Qeios ID: G4DAH0 · https://doi.org/10.32388/G4DAH0 21/75 Finally, the 2km Stockton spit lacks a tufa cap. Material transport was dominant during its formation and there was no aeration event afterward. The Stockton bar, however, was an existing landform and obstruction, so it ended up with a tufa cap.
In 2003, Smith et al. used ground penetrating radar to study the Stockton spit. The data from that paper supports the concepts presented here. (Smith, et al., 2003) The Stockton bar was a pre-existing feature at the time of the surging.
There were then three principal surges in the event, forming three progressive layers in the Stockton Spit complex. The second was the largest and buried the first in its core. The third was a small, high spit closer to shore formed as the surging was ramping down with smaller turbulence waves. That third spit has now been quarried out of existence, but a conversation with the quarry manager confirms that unlike the main spit, it contained highly striated deposits typical of multiple wave deposition.
There are a lot of spits and bars in the Bonneville record. Some of the smaller ones may be storm event shorelines, however, a lot will be found to provide a record of seismic activity along the Wasatch Fault. Dr. Jewell discussed the presence of large stones at the far reaches of some spits in Lake Bonneville and suggested that these could be used as an indicator of large storm events. Identifying those stones as an anomaly was very astute, here it is just suggested that they indicate something else.

A 45,000-year history of multi-segment earthquakes on the Wasatch Fault and new
insights into the climate record of the Great Basin and The Bear River Exclusion Theory The study of individual earthquakes on the Wasatch Fault has been generally limited to the last 17.4ky, the post-Bonneville-Flood period. The fault scarps from prior to that are obscured by Lake Bonneville sediment (Swan, et al., 1980, McCalpin, 2002. In a 2009 paper, Mayo did some novel work in examining cave sediments east of the fault to understand varying slip rates over the last 750kya, but this type of analysis does not provide information on timing of specific events (Mayo, et al., 2009). In a 2016 article, DuRoss explored the possibility of multi-segment earthquakes on the Wasatch Fault (DuRoss, et al., 2016). While they found no Holocene record of events spanning multiple segments, they also did not find reason to exclude that possibility and suggested further study.

The Lake Bonneville sediment record
In a 2016 study by Rey, Bonneville sediment cores from the Pilot Valley on the Utah/Nevada border were studied (Rey, et al. 2016). In that study, an abrupt and anomalous sand layer showed up in most of the cores. It is best covered by directly quoting the study: The earthquake-induced-surging theory predicts the existence of these types of abrupt transitions in sediments that show reworking. In the course of the current study, this type of transition was treated as a necessary but not sufficient condition.
In trying to identify an earthquake-induced surging event, it is important to look for corroborating evidence across locations and different evidence types.
Deposits need to be considered in the context of how location might affect them. In the Pilot Valley example above, results were dependent on proximity to the shore. Both the swash zone of a beach and earthquake-induced surge strata in deeper water can present similar evidence: A deeper lake would tend to reduce the amount of disruption of the lake bottom sediments but would not be immune in the larger surge events. A very shallow lake might see almost continuous disruption from not only earthquake-induced surging, but also storms.
In an earthquake-induced-surging event, previously deposited sediment layers are swept back up into suspension, basically disrupting the record of the timeline by removing material, mixing the age strata, and then redistributing it. The total thickness of deposited material may not change significantly, but the time sequence within the disrupted zone has changed. Additionally, lightweight organic carbon materials such as shells and wood could be transported in from other levels in the lake and buried in the chronologically wrong sediment.
In an earthquake-induced surging event, as lake-bottom sediment gets resuspended, the chemistry of the lake changes.
Turbulence can result in aeration and precipitation of solids. Rapid settling of these compounds can result in spikes in total inorganic carbon (TIC) in the resulting sediment. Researchers use TIC as a lake-level proxy for trying to understand historic variations in level on the premise that in a closed lake system with no other external factors, TIC increases as the level drops.
Calcium carbonate is a component of TIC. In a surging event, aeration near the shoreline results in degassing of CO2 and CaCO3 (tufa) precipitation. This type of precipitation is accelerated by nucleate precipitation, the result being tufa coating of pebbles and shells. Large storms might cause this type of effect, but it would tend to be limited to zones very close to the shore, and as stated earlier, if storms were the cause, it would be expected to be a common occurrence in the sediment record, and it is not. Anomalous spikes in TIC in the deep-water sediments are an indicator of potential earthquake-induced-surging.
With the Stansbury Oscillation and the Bonneville Flood events as models, a search was conducted for evidence of other earthquake-surging events in the Lake Bonneville record. The Benson stratigraphic study from the Blue Lake Marsh provides just such an opportunity (Benson, et al., 2011). Blue Lake is a spring-fed feature in the very western extreme of the Bonneville basin. Working with the Benson lake-level-proxy graphs and stratigraphic sections permits a search for matching patterns. Figure 12 is the Benson data for TIC (black line). Note that the y-axis represents decreasing concentration, since as a lake level proxy that would correlate with increasing lake levels. Approximate Heinrich Event periods are indicated. The other markings are relevant to the interpretation of the data in the context of the present earthquake-induced surging theory and do not reflect the content of the Benson paper.
Qeios, CC-BY 4.0 · Article, May 29, 2023 Qeios ID: G4DAH0 · https://doi.org/10.32388/G4DAH0 25/75 Figure 13 is the sediment core depiction from the Benson et al. paper. This has been included as a visual reference; the Benson legend has not been included so as not to suggest that any synopsis here is a substitute for reading the Benson paper. The Benson image has been annotated to reflect my analysis of how the sediments would correlate to the lake events identified herein. Only the pre-Younger-Dryas layers of the core are shown because that is the limit of what will be addressed here. In studying the Benson cores, the Stansbury Oscillation was used as a reference feature. The Stansbury Oscillation resulted in two very dramatic TIC concentrations spikes about 1ky apart in the Benson timeline starting at 25kya. If these two TIC spikes are assumed to be level proxies, that would appear to support the climate-based level drop theory.
However, a closer examination of the details of the research opens the door for a completely different interpretation.
The Benson study compared the natural remanent magnetization directional changes (inclination and declination) against the records of other North American lakes and ocean sediments to develop a depth/age model and they checked that against carbon dating where possible. The dates on Figure 13 are interpreted using the Benson correlation curve. Using this correlation for the Stansbury Oscillation timing, we see striated sand layers starting at the appropriate 25kya time, though the Benson stratigraphic depiction shows five independent layers with abundant shells, the last of which is about 1.6kya later. This 20 cm section of the core stands in contrast to the deep lake sediments on either side. This forms our standard model or fingerprint for earthquake-induced surging, though the double spike in TIC would have been expected to be a special case.
What jumps off the page in Figure 12 is that 5ky before the Stansbury Oscillation there was an almost identical double spike TIC event, with two dramatic spikes about 1ky apart. The core sample in this timeframe showed a stratified organic laminate and at the top of the zone a layer that Benson described as "created by wave reworking of older, higher-elevation deposits". (Benson, et al., 2011, p. 65) Using the Stansbury model, this spike in TIC and the wave reworking of deepwater sediments flags this as another earthquake-induced surging event.
Going back another 9ky to about 39kya, there is another spike deviation in TIC concentration. In the sediment record this corresponds with a 15cm thick bed of "wave-reworked, pelleted aragonite", otherwise known as tufa. This yields a third candidate for an earthquake-induced surging event. The TIC timeline is inconclusive on whether this exhibits a double spike, but the apparent duration of the disturbance is about the same as the latter two. The lake was quite shallow in this area at that time and that could have affected the surging and settling in a number of ways.
Evidence of the Bonneville Flood event is also apparent in the lakebed core. The Bonneville Flood event has a tight timeline fix at 17.4kya. Looking at the core sample first, there is a very distinct laminated calcite zone from 18.6kya to 17.0kya in Benson's timeline, though equating timelines between different studies is always a bit problematic. On the TIC graph, there is a sharp drop at the start of this period and then a recovery at the end. This is during the Bonneville highstand and a deep lake, so the lakebed may not see as much disturbance in an event and any TIC would be diluted into higher volumes so the TIC concentration may not spike as much.
Based on an analysis of the Benson sediment core data, five principal candidates for earthquake-induced surging events were identified. These events will be referred to as: S15.6-Provo-1455 S17.4-Bonneville-1551 S24.7-Stansbury-1360

S39-Fremont-1315
Dating techniques and level determinations will continue to be refined in the future and the designations will need to evolve. But for now, this should suffice in keeping the reader oriented to when, what others might have put in the titles of their papers regarding a similar level, and a typical rebound level where evidence might be found. Of course, those levels will vary in different parts of the lake, but the objective is to just keep the reader oriented. Each of these levels and the supporting evidence will be discussed later.
Others have identified a shoreline at the 1305m level and referred to it as the Pilot shoreline or level (Miller and Phelps, 2016). That level is complex and presents evidence as a transgression level and as a regression level, depending on location in the basin. I think this "level" is just a coincidence of factors. The transgression bar may be from a low point of a surge from higher events, and the regression bar may be a product from a Holocene event which will be discussed later.
In either case, I found no evidence to suggest that 1305m was a sustained shoreline level.

The Basalt Ash Deposits
The Bonneville basin has evidence of periodic basalt flows, some as recently as 600 years ago (Utah Geological Survey, 2023, Stahl, 2019. Basalt ash deposits have been found in the Lake Bonneville sediments and used by researchers as timeline markers to correlate between sites (Oviatt and Nash, 1989).

Researchers have identified five principal events in the Lake Bonneville sediments: Hansel Valley, Pony Express, Pahvant
Butte, Tabernacle Hill, and the Lower Basaltic Ash, the first four named after the source locations, the last for where it showed up in the core sample (Oviatt and Nash, 2014, Godsey, et al., 2011, Miller, et al., 2012, Miller, et al., 2008, Oviatt and Nash, 1989, Thompson, et al., 2016. In the course of the present work, it became apparent that there was a timing correspondence between these events and the earthquake-induced surging events. In the sediment record, the Tabernacle Hill eruption is in the timeframe of the Provo oscillation (S15.6-Provo-1455), the Pahvant Butte ash is in the window after the Keg Mountain oscillation and possibly during or just after the Bonneville Flood (S17.4-Bonneville-1551), and the Pony Express basalt ash shows up just after the Stansbury Oscillation (S24.7-Stansbury-1360).
The Hansel Valley eruption in the northern reaches occurred during the initial rise of Lake Bonneville. In a 2007 paper, Miller recounted a 1997 trip with Oviatt where they "traced the ash bed upslope from deepwater lake to shore zone facies, establishing that the ash fell into Lake Bonneville when the lake level was approximated at 1335m altitude." Miller, 1997, Miller, Oviatt, andNash, 2008, p. 239) They dated the deposit as "~28 cal ka" based on carbon dating of marl in the vicinity. Allowing for carbon dating variances, the timing and elevation match what would be expected for the S30- Hansel-1315 event.
The Lower Basaltic Ash was identified by Thompson, et al. in their study on flora in Bonneville sediment cores. In the lowest strata of their sample, they found what they tagged as the 'lower basaltic ash' (Thompson, et al., 2016). It is always problematic trying to compare dating of materials between studies, even between labs, particularly when looking at samples this old. They came up with a ¹⁴C age of 33.38kya, but then they deducted 1.8kya to try to account for what they assumed were reservoir carbon interferences. They ended up with a calibrated date of 35.5kya. The dates used in Figure   12 are based on Benson's work. Benson did not apply a reservoir carbon factor, but he did use the same CAMS lab as The five well-documented basalt ash events in Lake Bonneville have a one-to-one correspondence with the five identified multi-segment-earthquake-induced surging events.
Qeios, CC-BY 4.0 · Article, May 29, 2023 Qeios ID: G4DAH0 · https://doi.org/10.32388/G4DAH0 30/75  before Lake Bonneville and continued after. The interesting thing here is the distribution of the flows in this area. Many of the flows are clustered in areas where there would be the maximum stress in the crust from isostatic deformation of a lake in the basin. There are outliers to this pattern in the south and the extreme north near the ends of the Wasatch Fault displacements and that certainly is a topic for a different discussion.
The Bonneville basin is a very old lake basin and has had a long history of resident lakes, so isostatic deformation in this area has been going on for a long time.
The salient point in regard to Lake Bonneville is that the five known major ash events during the Bonneville period correspond to the five identified multi-segment surging events during that period. A one-to-one correspondence. With both a coincidence of timing and appropriate locations, a correlation between the multi-segment earthquake events, isostatic deformation and the basalt ash eruptions rises to the 'most likely scenario' status.

Lake Bonneville Isostatic Depression, aseismic intervals, and multi-segment earthquakes
The Lake Bonneville time period is a confusing array of controlling factors and level evidence. This paper assumes the reader is familiar with research concerning the structure of the Basin and Range province and the anomalously thin section of stretching crust in this region, along with the body of work surrounding the isostatic deformation of the Bonneville basin by lakes and glaciers (Hetzel and Hampel, 2005, Adams and Bills, 2016, Mayo, et al., 2009. Also pertinent to the current discussion is the variation in slip rate of the Wasatch Fault. In a 2009 paper, Mayo found indications that the slip rate during the late Pleistocene was about half the rate that it is today. (Ibid, 2009) The stresses in the crust and on the Wasatch Fault were a combination of both tectonic factors and isostatic factors, and the isostatic factors were dependent on long-term climate. In this section, the previously unidentified earthquake events in the Bonneville timeline will be discussed along with the supporting evidence.
As mentioned earlier, the spits in the western areas of the basin have long been thought to be the result of prevailing winds and large storms. The V-bars have been interpreted as cuspid forelands due to longshore currents driven by prevailing winds. The numerous bars in the lake have been treated as the products of established lake levels. While there certainly are exceptions, for the most part, none of that is correct. These features were instead each formed in a relatively This type of sloshing in the Bonneville basin would start with a large amplitude surge and then the oscillation would be dampened with each cycle. The type of evidence at the top end of the cycle would depend on the local slope. A steep slope would leave little enduring evidence aside from a slight increase in erosion. A moderate slope would permit a bar to form if the surge was moving soft sediment. A low angle incline such as in Skull Valley would spread the entrained material in the surge over a band. At the bottom end of the cycle, the result would be a bar. This is because as the cycle bottoms out, all the entrained material suddenly hits a low velocity wall of the existing lake and settles out immediately.
That bar would then get spread out a bit by subsequent cycles, but as the cycles moved shallower, the bar would quickly be at a depth where the cycles occurring above no longer affect it, preserving it as evidence of the event.
The material stirred up in these cycles includes the calcium which had previously collected in the lake bottom sediments.
In places like Skull Valley, the waves running up the shore in the initial cycles distribute the calcium compounds over a very wide band, painting images of the surges on the valley floor.
Things changed as the cycle pattern dissolved into noise. What was left were small waves of calcium-rich water smashing against a small band on the shore, still degassing and precipitating calcium carbonate. In Skull Valley, the calcium carbonate resulted in a bathtub ring of white calcium deposits. In many places, lacustrine wave action washed away the bathtub ring. But in other locations, such as Skull Valley, this bathtub ring can be used as an indicator of lake level. well-mannered, damped sine curve, defined by nine data points, supporting a high level of confidence in these findings and the concept that these cycles occurred at a frequency measured in hours, not decades or centuries.
At first glance, there is an apparent first law of thermodynamics violation. The system starts at equilibrium, and then surges away towards the earthquake slip depression in the east. However, when the wave returns, it runs higher than the equilibrium level. There are a couple of things in play here which contribute to this, and it comes down to the position of Eventually the surging cycle dissolves to noise, random large waves crashing on the shore and forming a thick band of tufa. In some places in the lake basin such as on steeper slopes, this appears as a band. In other places only the bottom edge sticks out from under later deposits. Once the noise of the event settles out, the lake still has an elevated calcium concentration and the normal waves on the shore sometimes form a second tufa band, usually about four meters above the first. This is not as frequently visible because it is dependent on more factors and is a more subtle feature.
Qeios, CC-BY 4.0 · Article, May 29, 2023 Qeios ID: G4DAH0 · https://doi.org/10.32388/G4DAH0 33/75  Each successive surge was not as deep as the preceding and not as forceful, but the reflected wave was always 2m higher than the initial wave of the surge.
The event settled down at the 1335m lake level and the residual noise waves left a pronounced calcium bathtub ring at that level.
At other locations in the basin, where the current was perpendicular to the mouth of a bay with soft sediment, the initial drawdown left a distinct bar at about the 1305m level. There would be successively higher bars, but they would be less distinct.
In his 2015 paper on the chronology of Lake Bonneville, Oviatt (2015) indicated in one figure that, at about the time of the Hansel Valley ash event, mollusk shells were deposited at an elevation of 1355-1360m. This would be inconsistent with previous level assumptions for that time but would be well within the potential upward bounds of the S30-Hansel-1335 event surge: 1335m with surging +/-30m.

The S39-Fremont-1315 Event
The Benson core TIC data suggests an earthquake-induced surging event occurred 9ky before the Hansel event. While the sediment record of this event has been securely stored under the depths of Lake Bonneville, the record at the surface has been subjected to 39ky of erosion and surging events, consequently much of the shoreline evidence has been obscured or destroyed. However, there still is evidence of a significant event in the 1312-1318m in many parts of the basin.
The north end of Stansbury Island in the middle of the basin is a location prone to tufa deposits during surging events. Consequently, it provides a record of events identified in this paper. Shelf formations can be formed by a variety of lacustrine occurrences, but the ones identified are very distinctive and at elevations commensurate with other evidence in the basin. (Figure 18) The Stansbury and Provo levels are very prominent in this area. The S30-Hansel-1335 event in not as prominent at this location, but this may be due in part to the location on the steep slope just below the dominant S24.7-Stansbury-1360 event shelf. The S39-Fremont-1315 level is quite distinctive here. Also there appears to be another shelf between the Hansel and Stansbury levels, fainter still, though this corresponds to the elevation of the adjacent ridge, so it may just be an artifact of lacustrine long-shore currents. In a 2022 study, Pigati and Springer reported that the Younger Dryas (12.9-11.7kya) manifested in two stages in western North America.  The first stage ran until about 12.2kya and was consistently a cool wet period, typified by high water-tables in the Death Valley region west of the Great Basin. The second stage, which ran until 11.7kya was unsettled with varying water-table levels but overall was a dryer than normal period.
The Gilbert Level episode in Lake Bonneville corresponds with the Younger Dryas. In a paper on the Gilbert Level by , he analyzed prior work and reported new information regarding this level. He found that the lake achieved a maximum elevation of 1295-1297m at 11.6kya. Others have studied lake levels in the Great Basin from the late Pleistocene through the Younger Dryas (Reheis, et al., 2014, Adams, et al., 2008. In the various Great Basin lakes, the Younger Dryas high-stands can vary by thousands of years between lakes for unknown reasons, but in all cases, there is a bump up in level at a time during or just after the Younger Dryas. This is interesting because of Pigati and Springer's finding of a two-stage Younger Dryas with a wetter than normal first half and a dryer than normal second half. In the Lake Bonneville basin, a resurgence of glaciers in the Wasatch and Uinta Mountains during the first half of the Younger Dryas would have stored water for a continued rise leading into the second half. This same pattern occurred eight thousand years earlier when Lake Bonneville continued to rise after the last glacial maximum in the Wasatch.
The Benson core at Blue Lake in Lake Bonneville is of limited use at this level for resolving this issue since that area was marsh or shallow lake at the time.
Since the Gilbert Level was never threshold controlled, there is no reason other than chance for there to be a sustained Gilbert Level.
The suggestion that there was a seismic event during the occupation of the Gilbert level was introduced by Hyland in a 2012 trench study on the eastern shores of the Great Salt Lake (Hyland, et al. 2012). That trench work also revealed a tufa layer resting on an unconformity above Lake Bonneville deposits; this is consistent with the model of surging presented in this current study.
Oviatt further supported the concept of a Gilbert level earthquake in his 2014 work. In his analysis of a sediment core, he states: "The inclination of these laminations is interpreted to represent wave agitation or earthquake disruption of the Great Salt Lake bottom during the Gilbert episode." (Oviatt, 2014, p. 11) The Hylland and Oviatt data of an earthquake at 11.6kya is accepted as a base assumption in this current work.
For such a recent period, identifying a 'Gilbert Level' has been a problem for researchers. Above and below the level range that Oviatt settled upon are a complex array of level artifacts. This is compounded by the fact that this is the same elevation range of the transition from the basin playa to the rise of the mountains. Identifying the lake level at the time of this earthquake event is of value because it gives a hard date and level fix for the otherwise transient Gilbert episode.
Currey did the initial mapping of the Gilbert shoreline in 1982 and the features he studied ranged from 1293 to 1311m in elevation (Currey, 1982). In his 2014 paper, Oviatt debated the level evidence before settling on the 1295-1297m range as the most likely. However, in his conclusions, he limited the level statement to "An assumption that the lake reached altitudes higher than about 1297 during the Gilbert episode may not be valid." (Oviatt, 2014, p. 18) Without a sustained level, any evidence of a Gilbert high stand would have been washed away by the surging of a Gilbert level event. Not only would a Gilbert surging event create multiple bars and erosion features, but anything up to around 1300m would also see surges from any later Great Salt Lake level events.
The ¹⁴C dating samples reported by Oviatt in developing his Bonneville hydrograph (Figure 2) show multiple potential levels in this time frame running up to almost 1320m. Surging would explain that. In the previously referenced Miller et al. paper on Provo shoreline deposits, they depicted in their Figure 5, a pair of shorelines stacked one on top of the other, one two meters above the other (Miller, et al., 2012, p. 347). In the 2014 paper by Oviatt on the Gilbert level, he depicted, in his Figure 3, two bars at the Magna spit stacked one on top of the other, one two meters above the other (Oviatt, 2014, p. 7). The top bar was found to be at about 1295m, two meters below the 1297m level assigned by Oviatt to the Gilbert level.
The shelf in Figure 18 supports a Gilbert event level of 1300-1301m range. Calcium bathtub rings in the western reaches also support a 1300m elevation (41°24'3.85"N 113°42'9.98"W. The difference between this and Oviatt's Magna spit elevations might be due to Wasatch Fault displacement. Crittenden discusses the tilting of the block just west of Stansbury Island to the Wasatch Fault in the east (Crittenden, 1963, p.E28). The Magma spit is mid-point between the two. 11.6ky of fault displacement would explain a difference. Picking and choosing reference points in the basin could support levels anywhere within the 1295-1302m range. The 1300m level is being used in this paper to be consistent with some of the more prominent features and is being used as the level of the Gilbert earthquake event, not the Gilbert high stand. The Gilbert high stand was probably quite transitory.
What the Gilbert level event is missing is a known basalt ash event in the relevant time frame. At the Gilbert level, isostasy was far less pronounced, supporting speculation that the ash eruptions during the Bonneville period were a byproduct of isostatic depression of the crust in the basin.

The Provo level Shorelines and the Provo Oscillation
In 2011, Godsey et al stated "The Provo shoreline is actually a complex of several coalescing coastal landforms…". (Godsey, et al., 2011, p. 443) That may rank as an understatement. It is difficult to find consensus between researchers.
From a distance, the distinctive shoreline on the mountains makes this level look quite straightforward but interpreting it has been complicated. Researchers have cited issues with isostatic rebound, landslides which repeatedly blocked the outflow and raised the level, earthquakes shifting and dropping the outflow, storms, and an uncertain duration due to a plethora of dating issues. The one thing that is absolutely certain is that at the end of the Provo level was the most significant climate-based level drop in the lake's history, which ended as abruptly as it started. "also found little support for the mid-Provo drawdown". (Miller, 2012, p. 345). Miller coauthored the Godsey paper, so the assumption has to be that this is an accurate paraphrasing of the conclusions of the earlier paper, though the exact phrase in the Godsey paper was "We found no conclusive evidence from the stratigraphic record that a climate-induced drawdown of the lake, and subsequent return to threshold control occurred during mid-Provo time." (Godsey, Oviatt, Miller and Chan, 2011, p. 450) "Little support" and "no conclusive evidence" have different connotations, but the consensus in the literature seems to be that the change in Provo level outflow levels between the Swan Lake sill and the Clifton sill did not translate to the overall Lake Bonneville level.
The Godsey team, which included Oviatt and Miller, did report on an anomalous sand layer near the top of the Provo level marl appearing in several cores (Godsey, 2011, p. 445). This layer has reworked shells and small tufa heads. They suggested that this layer was due to "winnowing" during a storm event. Dating of shells in this layer put some of them just before and some just after their assumed fall from the Provo level at 12.6kya ¹⁴C, but with error margins that overlapped the assumed fall, which they equated to 15kya cal. The Godsey, et al. descriptions are consistent with what would be expected in the earthquake-induced surging event being presented here. Many aspects of the mid-Provo level oscillation described by Janecke and Oaks also fit the pattern of an earthquake-induced surging event.
Miller et. al. surveyed the Provo shoreline at 83 points around the lake basin (Miller, Oviatt and McGeehin, 2012). They did a very detailed analysis and considered many factors, including isostasy. One of the sites they studied was at a railroad cut west of Wendover, Utah (the Ola railroad cut) which runs through the Provo level deposits and the Bonneville Flood sediment (18.1kya in their dating timeframe, where 17.4kya is assumed here). (Figure 19) In these deposits they found three "subhorizontal resistant beds… that are poorly sorted, matrix-rich and typically tufa cemented" (Ibid, p. 349). Though later in the text they report that the tufa is only present in the top and bottom layers and not the middle layer. The description of those top and bottom layers tufa layers matches the proposed indicators of earthquake-induced surging in the sediment record. The bottommost of these tufa beds was just below the Bonneville Flood sediment, making it consistent with the theory of earthquake-induced surging causing the Bonneville Flood. Their dating of that tufa layer was 18.1kya cal, or an offset of 0.7ky from this paper's 17.4kya dating of that event.
The topmost tufa bed is 1-1.5m thick and comprises "flat-bedded, fine pebble gravel with a poorly sorted sand matrix… this bed is resistant owing to cementation by tufa coats on casts" (Ibid, p. 349). There is some evidence of "aquatic plants mats" at the base, which are known to promote tufa formation. However, it would be unusual for aquatic plants to grow for such a brief period in a location and at no other time in the thousand-plus-year time span of the Provo level. The top bed is in the right location to be evidence of a Provo level earthquake-induced disturbance.
At the Ola railroad cut, directly above the upper tufa layer are two shoreline bars identified by the Miller team, the newer bar at a higher elevation on the slope than the older, suggesting a rising lake level. These two The interesting thing about these two bars is how they fit into the broader context of the situation, or more specifically what is missing in the Ola railroad cut cross-section. Between the Bonneville Flood tufa layer and the Provo oscillation tufa layer is 9m of lake-bottom sediment with no evidence of bar formation, so the Provo level went through an extended period with no bar formation in this area. (Miller,et al,p. 350, Fig. 7) Above the Provo oscillation tufa layer are two bars, each about 4-5m thick, with no evidence of static lakebed sediment accumulation. Remember the earlier point that bars are quickly destroyed by normal shoreline wave action unless the level change is quick enough to preserve the bars. The physical evidence points to an extended Provo shoreline after the Bonneville Flood event, then a tufa forming level oscillation, basically an earthquake-induced surging event, followed by the formation of two shoreline bars by extraordinary wave action, followed by an immediate drop in lake level. This analysis results in a 1ky discrepancy between the Miller tufa layer and the Benson sediment core data, but everything else points to the two being the same event. Later in the Miller paper, he presents " Fig. 14. Plot of culled radiocarbon ages for the Provo beach deposits and associated offshore deposits and tufa." (Miller et al. 2012, p. 358).
This is a typical timeline plot of the Provo shoreline ¹⁴C data in the literature and the same data shows up in many papers.
The data ranges from 18kya to 14kya ( Based on the physical data and trying to extract the dating information that appears most consistent, the drop from the Provo level most likely started at about 15.6kya, at the time of Janecke and Oaks disputed "mid-level drop" from the Swan Lake sill to the Clifton sill.
3.5. The fall from the Provo level -The Bear River Exclusion Theory -more than just a cautionary tale.
No one can agree upon when the fall from the Provo level started, or on whether there was an early and a late Provo level based on the Swan Lake and Clifton sills. Strangely enough, on the question of whether there was one Provo level or two, everyone is right, it just depends on the reference point. The Bear River Exclusion Theory presented here that not only embraces all the data, but that also explains previously ignored features.

Researchers have noted a correlation between the high stands of the numerous Pleistocene lakes in the Great Basin with
the Heinrich 1 stadial, though the timing of the high stands varied widely in the region. They also found that the lakes of the region all fell at some point during an extended period between the H1 stadial and the Younger Dryas, though the supporting data is sparse and at times contradictory. (Benson, et al., 1990, Benson and Thompson, 1986, Munroe and Laabs, 2013, Reheis, et al. 2014. Clearly the Great Basin moved into an extended dry period and different drainage basins had different responses. In a 2019 paper studying lake level fluctuations in the Northern Great Basin from the late Pleistocene on, Santi, et al., found that "In many cases, lake transgressions to their high stand levels (from moderate stillstand levels) happened in a relatively short period of time between 17 and 14 ka, while regressions tended to occur over a much longer period." (Santi, et al., 2019, p. 183) However, she and others have consistently remarked on the fact that the fall of Lake Bonneville was quite rapid. (Reheis, 2014 The history of Lake Bonneville's transgression and regression is summarized as follows: Sometime around 55kya, lava flows in the Gem Valley of southwestern Wyoming blocked the Bear River's course to the Pacific Ocean via the Blackfoot River, the Snake River and the Columbia River (Pederson, 2018). The Bear River was diverted into the Bonneville basin and the addition of this major Uinta Mountains drainage was sufficient to start a long, gradual rise of Lake Bonneville. Qeios ID: G4DAH0 · https://doi.org/10.32388/G4DAH0 43/75 level over the same general period and then more or less stabilized, but the data for Lake Bonneville is discrete in nature, or colloquially, like someone threw a switch or closed a valve, shutting off the water one day and then much later just turned it back on.
Stealing a line from the detective novel genre, 'here is what happened'. Lake Bonneville and the Great Salt Lake had and have four principal rivers as sources: the Sevier, the Provo, the Weber, and the Bear. According to the Utah Geological Survey, today the Bear River supplies nearly 40% of the water input to the Great Salt Lake (UGS website). It represents the largest single source of water feeding the Great Salt Lake. That is why the 55kya diversion of the Bear River into the basin was so critical.
The rising Lake Bonneville created lacustrine features throughout the basin. In a 2020 paper, Oviatt suggested that the very prominent Bonneville level benches were depositional in nature and built up over time (Oviatt, 2020). These benches can be observed adjacent to where the Bear River enters the Great Salt Lake Basin at the Cutler Narrows coming out of Cache Valley, and at that point the deposits extend down and cover the Wasatch Fault in the area. As the lake level rose, these benches would have extended across the narrow gap of the Narrows. Earthquake-induced surging events would have carried those sediments up into the Cutler Narrows gap, filling it to within meters of the lake level in any given event.
A single surge can carry an enormous amount of material, as evidenced by the Stockton Spit discussed earlier. The top of the Stockton Spit is about 6m below the Bonneville high stand in that area, so it is reasonable to assume that the Cutler Narrows were filled to the same level.
During the S17.4-Bonneville-1551 surging event, and the initial stages of the Bonneville Flood, the flow would have overtopped not just the obstruction in the Cutler Narrows, but also over 5km of adjacent ridgeline. This wide front would have defeated the formation of any barrier bars between the main Bonneville basin and the Cache Valley. Basically, insufficient material to dam off that broad a front with that great a flow.
As the surging died out and the lake level dropped during the Bonneville Flood, the flow would have channeled down to where it was all flowing through a 0.75km wide sluice in the Cutler Narrows. The flow would have eroded away the soft sediments previously deposited in the gap as it went, but as the flow started to be limited by the Swan Lake sill, the velocity would have dropped until it was just the flow necessary to maintain level between the main basin and the Cache Valley. This flow preserved the connection between the two bodies of water, though it was probably at best a shallow connection or at times even just a river from Cache Valley into the Great Salt Lake basin. The Bear River was a major component of the Lake Bonneville water balance, so the net flow would have been from Cache Valley to the main body of the Lake, with the excess flowing out to the Pacific Ocean through Red Rock Pass.
For over a millennium, the Provo level of Lake Bonneville was controlled by the Swan Lake sill near Red Rock Pass. Then something happened to cut the Swan Lake sill down by 9m to the Clifton sill (Janecke and Oaks, 2011). Janecke and Oaks postulated an earthquake in the Cache Valley, though this paper has presented evidence that the event was much larger: the S15.6-Provo-1447 earthquake and surging event. The initial surge carried up into the Cache Valley and caused the erosion of the Swan Lake sill outflow. The surging would have also carried more bench sediment material into the Cutler Narrows gap. The S15.6-Provo-1447 episode formed two successively higher bars on shorelines throughout the Bonneville basin (Miller, et al., 2011) and there is no reason to believe that this barrier would not have also occurred across the Cutler Narrows gap.
With the new, lower outfall, the level of the Lake in Cache Valley quickly dropped to a level below the obstruction in the Cutler Narrows. Suddenly, the Cache Valley was no longer hydraulically connected to the main body of Lake Bonneville.
With this, the flow of the Bear River was short-circuited directly to the new Clifton sill and down the Portneuf, Snake and Columbia Rivers to the Pacific Ocean. Overnight Lake Bonneville lost 40% of its inflow. The region was already well into the dryer cycle between Heinrich events, and the loss of this flow was catastrophic. Without sufficient inflow to maintain level, Lake Bonneville dropped rapidly and at a very consistent rate dictated by evaporation.
During this same period the newly isolated 'Lake Clifton' in the Cache Valley remained well fed by the Bear River and level limited by the Clifton sill. Janecke and Oaks write about the "landforms of an ancient meandering river" in the area between the Clifton sill and Red Rock Pass to the north (Janecke and Oaks, 2011 fieldguide, p. 204). They further elaborate: "The channel and floodplain of this large river system formed below, and after, the higher Provo shoreline." One other comment of note by Janecke and Oaks: "The average wavelength and width of the meander belt are many times larger than those of the modern Bear River (Fig. 2) but are similar in scale to incised meanders that were cut by the late Pleistocene Bear River…" Oaks 2011b, p. 1384). The reason Janecke and Oaks found that the river going over the Clifton sill was the size of the Bear River of the time, was because it was the Bear River. The other significant point is that the Bear River of that time was larger than the current Bear River, so depriving Lake Bonneville of it would have a significant impact.
This explains why massive Lake Bonneville, which for at least 30ky had a water balance excess, suddenly started to dry up. The other part of the puzzle is that after maybe another half a millennium something happened and both the outflow at the Clifton sill stopped, and the lake level stabilized and started to fluctuate around the Great Salt Lake level.   Figure 22 shows the classic shapes of a burst flood path as is seen in the Missoula Flood landscapes. In Figure 23, the lines of the flood can be seen on the land between the current day meanders. These flood features are prevalent in the meander channel and exceed what occurs with seasonal variations and 100-year storms. To put things into perspective, these features are over 35m above the current river level, and at Corrine, Utah in this same area the largest flood in the last 20 years has been no more than 4m above river level.
Within the steep-walled Cutler Narrows gap, any evidence of the Bonneville-bench sediment dam has been erased by Qeios, CC-BY 4.0 · Article, May 29, 2023 Qeios ID: G4DAH0 · https://doi.org/10.32388/G4DAH0 46/75 14ky of erosion. Within the Cache Valley, evidence of either abrupt or gradual lowering of the lake level is missing, but it would have been obscured by subsequent farming in the area anyway.
The volume of water from draining Lake Clifton would have raised the level of the Great Salt Lake by 3-6m, though any evidence of that level bobble would have been lost in other lake fluctuations, such as the Gilbert episode.

The Fresh Water Events
Two other anomalous events stood out in the Benson study and these events do not fit the earthquake-induced surging model. There were two very dramatic low concentration pulses in TIC (upward spikes on the graph, labeled Fresh Water Events X and V in Figure 12). Benson points to Dansgaard-Oeschger events 5 and 10 as possible causes for these types of spikes (Benson, 2011). I would like to offer another possible explanation for this and some of the other very abrupt and short-term drops in TIC concentration at this specific location. These spikes occur during the earlier and lower-level period of Lake Bonneville where a dilution stream into the Blue Lake area would be inordinately represented in the sediment record. This dilution could come from a flood flow into the lake from a storm or melt water event, but such flows would most likely come from the mountains on the eastern side of the basin, though it is possible that events occurred in the western ranges. An alternate explanation might be a rapid release from a subterranean aquifer. The Blue Lake site is a spring fed marsh today. A rapid release might be due to a seismic event local to that area, or it could be related to the rise of Lake Bonneville. The weight of the lake could have caused periodic collapses of the aquifer, releasing enough water to dilute the Blue Lake cove. In either case, barring additional evidence, these extreme-spike, low concentration events are probably local and not basin-wide phenomena.

A revised Bonneville Hydrograph
The Lake Bonneville Hydrograph has gone through numerous iterations through the years. These timelines are based on ¹⁴C dating of sediments combined with evidence from the traditional "shorelines" and judgement calls. The earthquakeinduced surging theory puts some of those shorelines in question while introducing new time/level anchor points. Qeios ID: G4DAH0 · https://doi.org/10.32388/G4DAH0 50/75 S15.6-Provo-1455 to S11.6-Gilbert-1300 -At the start of this period, the Provo level in the Cache Valley dropped from the Swan Lake sill to the Clifton sill, isolating the Cache Valley from the main body of Lake Bonneville and preventing the Bear River from contributing water to the main basin. Lake Bonneville then dropped to levels near the current Great Salt Lake level. How quickly the level dropped is difficult to determine because of the distribution of the ¹⁴C data.
The failure of the natural dam in the Cutler Narrows spelled the end of the fall and caused Lake Bonneville to rise by the volume of water held in Cache Valley. According to Oviatt, after the fall from the Provo level, the lake was close to the average level for the Great Salt Lake . As to when this occurred, Oviatt only states that the lake fell to that level "before 13,000 calibrated" (Oviatt, 2014, p. 1). Two points which might help in determining that date would be identifying the date when the Clifton sill was abandoned, and then checking that against any evidence of a Cache Valley earthquake in the same time period.

Earthquakes and the Climate
The earthquake-induced surging events identified in this paper are supported by the lake sediment data and aspects of the shoreline deposits. The apparent one-to-one correspondence with ash eruptions in the area adds support to this argument and necessitates the consideration of other seemingly coincidental events.
Those still hanging on to the climate oscillation theory should have picked up on the fact that all the earthquake-induced surging events are within timeframes identified with Heinrich stadials (Figure 24). The exact timing of these stadials varies between researchers. (Hemming, 2004) But the spikes in TIC concentration occur during the Heinrich events and there are no Heinrich events in the time period without TIC spikes, putting the burden of proof on someone arguing against a climate link. I am left with either suggesting the correlation is a coincidence, which I do not believe, or I have to suggest there is a link between the weather and a multi-segment earthquake on the Wasatch Fault, which sounds absurd. I will risk going with that later because that is where the data leads.
The theory of climate-based level oscillations in Lake Bonneville holds that the Heinrich events resulted in dramatic dry spells in the region which resulted in large evaporative water losses from Lake Bonneville. However, the 2018 work by McGee found that during the periods around a Heinrich event the lake levels in the Great Basin rose at a faster rate (McGee, et al., 2018). This opens up a very intriguing possibility that enhances our understanding of isostatic response due to lake load, our understanding of the climate in the region, and our understanding of the Lake Bonneville levels during this period. If a Heinrich event results in an accelerated rise in the lake level, the result would be a rapid rise in stresses in the crust due to increased isostatic deformation. Isostatic effects are often transferred to adjacent faults (Wernicke and Axen, 1988). Arther Conan Doyle gave Sherlock Holmes a great line: "How often have I said to you that when you have eliminated the impossible, whatever remains, however improbable, must be the truth?" (Doyle, 1890) That line consistently applies to Lake Bonneville. Unfortunately, a scientific paper is different from a mystery novel in that you have to lead with the startling revelation that the novel would leave for the last page, thus robbing the reader of the enjoyment of solving the mystery along with the researcher.
There are somewhere around 50 shorelines visible in some of the more protected coves in Lake Bonneville.  (Oviatt, 1997).
These were based on analysis of sediment cores and along with shoreline deposits and features. Nelson expanded this out to six intermediate shorelines in his doctoral dissertation (Nelson, 2012). The last of these was the double shoreline of the late Provo level, previously discussed. In a 2015 paper, Nelson and Jewell studied three shore-zone gravel wedges between finer-grained offshore sediments on an exposed slope of Hogup Mountain in the north central part of the Bonneville basin (Nelson and Jewell, 2015). They interpreted the middle of the three as corresponding to Oviatt U2 and identified the other two as new climate oscillation events. A dating-reversal appeared in their carbon dating of the middle event, but those types of reversals have been common in Bonneville carbon dating. The earlier presented figure of the Oviatt climate oscillation timeline (Figure 2) shows carbon dating results and the scatter of the data makes it difficult to decisively support any conclusion.
Gilbert identified a problem in these intermediate shorelines because he found little to no correspondence between the Qeios, CC-BY 4.0 · Article, May 29, 2023 Qeios ID: G4DAH0 · https://doi.org/10.32388/G4DAH0 52/75 shorelines in one location to those in another other than what might be "referred to fortuitous coincidence" Gilbert, 1890, p.139).
In a later work, Jewell (2016) was able to correlate a few of the shorelines. Schide did an excellent analysis of these shorelines in a 2016 Master's thesis, available at the University of Utah website, and a subsequent peer-reviewed paper. (Schide, 2016, Schide, et al., 2018 Anyone interested in Lake Bonneville should read Schide's Master's thesis.
One of Schide's findings was that: "Lake Bonneville barriers display a wide range of morphologies determined by local sediment supply, wave energy, and other geomorphic conditions. The formation of these barriers cannot be described with one single theory since local factors have greater control on their elevations, shapes, and positions than basin scale water level changes." (Schide, 2016, p.38) 4.1.1. Definition of terms -Intermediate shoreline, bar, closed spit and spit. In this paper the term "intermediate shoreline" refers to the distinctive bars in the elevations between the Bonneville and Provo shorelines, as opposed to the spits and long-shore erosion features in that zone. Spits and long-shore erosion features are often the result of different processes than shoreline bars, and as both Gilbert and Schide point out, in the Bonneville record there is frequently no correlation between the elevations the shorelines and the bars in a region.
In the western arms of Lake Bonneville, spits often take the form of wedding cake layers. This is prevalent in the levels just below the Bonneville high stand. (Thomas, 2014) If this were due to level standstills, you would expect corresponding level formations in other locations, which there are not. If these were storm events, that would require a surprising sequence of decreasing intensity of storms over a short period of time to create such a tiered formation. Earthquakeinduced surging would create this type of tier sequence, in fact it would be expected. The first surge would be the largest and occur at the deepest level and then each succeeding slosh would create a progressively higher and smaller (shorter) spit. The top of each tier would not only be a function of lake depth in the area, but also of the local velocity of the surge, so there would be little correlation between tier elevations at different locations in the basin.
The basic physics of formation of a spit is different from that of a bar, unless the bar is formed by long-shore currents off a point of land (which would make it a closed spit). Spits are formed at eddy lines off of points of land. In this paper, I will restrict bars to what is formed when waves or surges perpendicular to shore draw sediment down into a transition zone to deeper and lower velocity water where the sediments immediately drop out. Under the widely-accepted climate oscillation theory, each of these Intermediate Bars formed during the lake's transgression and each at the peak of an extended dry period where the lake level dropped by over 20-50m followed by a Qeios, CC-BY 4.0 · Article, May 29, 2023 Qeios ID: G4DAH0 · https://doi.org/10.32388/G4DAH0 54/75 wet period which exactly mirrored the dry period and raised the level back up. If these were climate events, variations in the flora and fauna records would be expected, but none have been identified. Such large climate-based oscillations might also be apparent in the sediment cores. In his 1997 paper, Oviatt did chemical analysis of core samples and while there were fluctuations in the CaCO3 and isotope concentrations throughout the transgression (Oviatt, 1997), it would be difficult to pick any potential climate-based fluctuations out from noise in the data. The smaller, more frequent fluctuations in chemical composition are likely due to large storm events or short-term differences in annual rainfall.
Schide did a ground penetrating radar examination of these intermediate shoreline bars in the Matlin Basin. What she found was that the Bonneville high-stand bar was striated, and the deposits sloped towards the open water. Earthquakeinduced surge would have formed this bar very quickly, similar to a storm surge. The smaller waves generated as the event died out would gradually diminish in size so this underwater bar would have layers that towards open water, which is consistent with what Schide found.
In contrast to the Bonneville level bar, Schide found that in all the Intermediate Bars, the strata sloped shoreward, upslope, indicating that the bars were built in an environment of rising water levels. This would support the well-accepted conclusion that these bars were formed during lake transgression. Schide discussed the normal bar formation and destruction process: bars normally form in the large waves associated with a storm, but after the storm passes the smaller, normal ocean waves start to erode away the bar. In a rising level situation, the level rise would have to occur very rapidly to avoid the bar being quickly eroded away in the over-topping process. From this she derives the only logical conclusion and that is that after every one of these level drops or pauses, there was an exceptionally rapid level rise in this massive lake which preserved the bar. The problem with this necessary condition is that a single incident is a fortuitous coincidence, but a recurring pattern suggests that something else is in play.
In Figure 26 is a Google Earth™ elevation profile of the Matlin Basin with each of the shorelines indicated with maximum elevations of each shoreline. It is apparent both visually and in the numbers that these intermediate shorelines occur at regular intervals. A consistent pattern such as this requires a harmonic system or a regulated system. To date, no one has proposed climate cycles which correspond to this frequency of bar formation. Qeios, CC-BY 4.0 · Article, May 29, 2023 Qeios ID: G4DAH0 · https://doi.org/10.32388/G4DAH0 55/75 Everyone from Gilbert on has said that these shorelines were formed during a transgressing lake, and logic and the data certainly seem to support that. These are large bars, and it would take a long-static lake level to form each of them. The bars exhibit sediments consistent with a rising lake. The lake regression through these same levels was during the During the Bonneville Flood, there was a rapid unloading of the crust and isostatic rebound occurred. Rather than the rebound occurring smoothly, the system was sticky. The rebound occurred in "pops". In engineering terms, it was a regulated system. The load of water in the basin depressing the crust was the regulator. The load was being removed at a controlled rate; the Red Rock Pass weir formation was the controller. When enough weight had been removed, the inherent stickiness or friction in the system was overcome and the crust popped upward.
Judging from the evidence, each pop was not a single jolt but a groundwave shaking that set up a seiche in Lake Bonneville. This radiated out and explains the wave energy distribution seen by Nelson in his work; wave energy was independent of fetch and wind direction. (Nelson, 2015) These standing waves created the striated levels in the bars identified by Schide. The rebound pop lifted the center of the lake, displacing lake water outward to the periphery and thus the perimeter shorelines rose during the pop. This resulted in the bars exhibiting the inwards slope of a rapidly rising lake level. But as soon as the pop ended, the ongoing Bonneville Flood quickly lowered the lake level, preserving the bars in pristine form. This explains why the bars are better preserved than you would expect if they had undergone several thousand years of transgression.
In the remainder of this paper, the phenomenon will be referred to as Isostatic Rebound Pop Seiche or IRPS ("urps").
The refuting evidence to this theory would appear to be the carbon dating data associated with these bars. Referring to the climate oscillation timeline presented earlier (Figure 2), the carbon dating evidence in each bar is consistently from the time of the lake transgression, though the distribution has an incredible level of noise. The noise has been explained as supporting evidence of the huge climate-related level swings. While it might seem patently ridiculous to suggest that all this data is an indication of a short-lived event thousands of years later, it is just that.
During the transgression, the shoreline sediments were laid down at the appropriate levels with the appropriate time There are five steps between from the Bonneville level down to the early Provo level. Each step was 19m +/-1m and that was how it sorted out in the first parsing of the data. Even if the steps had varied by a few meters, the consistency between steps is extraordinary. Identifying the controlling variable will provide insight into the behavior of the crust in this area under the effects of isostasy.
First consider the element of time as the potential controlling variable. As stated earlier, the Bonneville Flood was a controlled system, with the decrease in load on the crust being modulated by the flow rate through the Red Rock Pass weir. At the higher levels the flow rate through the pass would have been higher, but level drop is a function of both flow rate and of the volume to be drained at each level. Using the surface areas calculated by Adams and Bills (2016), the volume in the first meter of the Bonneville Flood was 52.11 billion cubic meters and the volume in the last (Provo level) meter of the Bonneville Flood was 38.15 billion cubic meters. While it is possible that the decrease in weir flow exactly matched that, it is unlikely. Lake Bonneville has many shallow arms so there would be a fair amount of variability.
Step changes in flow rate would be expected. Time is most likely not the controlling variable.
The second possible controlling element is the total weight of the lake. This is certainly one of the expected controlling variables. The first meter of level loss at the Bonneville high stand represented about 52 trillion kilograms, whereas at the Provo level the last meter of level loss was 38 trillion kilograms. Total weight lost does not account for the consistency in isostatic rebound pops between the first pop and the last.
The one controlling element which would yield a result just dependent on the level drop is hydrostatic pressure.
Mathematically, weight is the product of density, area, and depth. Hydrostatic pressure is just weight divided by area. Area cancels out and what is left is density times depth. At this location, during the Bonneville Flood, the amount of rebound in each pop was based primarily on hydrostatic pressure.
There is a problem with this finding and that is that in the many studies of isostasy and specifically isostasy in Lake Bonneville, the bowing of the crust is dependent on total weight. Total weight was still a factor during the Bonneville Flood, the reason things resolved to hydrostatic pressure in the Matlin cove was probably due to weight distribution in the basin. At the Bonneville high stand, the lake ran up quite high on the steep sides of the mountain ranges of the Basin and Range region. The shallow arms were all to the south. These arms to the south were sufficiently removed from the main body of the lake to be supported locally by the crust. The change in lake surface area during the Bonneville Flood occurred primarily in the somewhat independent southern arms. The area change in the northern portion was minimal and thus the amount of deformation of the crust in the Matlin Basin is resolved to being mainly, not exclusively, dependent on depth.
That the spacing of the intermediate shorelines can be accounted for through simple engineering principles is the final argument as to why these shorelines are the result of IRPS during the Bonneville Flood.
Reality is rarely that simple. In Matlin Basin there are four Intermediate Shorelines, but in other locations there might be six, eight or even a couple of dozen. Resolving the Lake Bonneville Intermediate Shorelines between different locations has been a perennial issue in Lake Bonneville research. This is because the Intermediate Shorelines are products of wave action during IRPS events, and those were highly dependent on position in the Bonneville basin and on local topography. Specifically, the differences are due to a mix and match of the following: b. The isostatic pop shifted water from the deep center to the perimeter. The effect was that the shorelines in the center of the lake moved lower on the slopes and the shorelines on the perimeter got higher. Matlin Basin experienced a rising shoreline during the pops, hence the bar sediments exhibited the signs of a rising level. Near the center of the lake, any IRPS bars should show evidence of a falling level and slope towards the lake, or as may be the case on the Newfoundland Mountains, the seiche waves acting on a rising shoreline may have eroded away the bars as quickly as they were being formed.
c. The slope matters and may be a second way of forming bars of different sizes. As a slope flattens out, the seiche breaks up into smaller waves forming smaller and more frequent bars. In a very steep slope, the angle of repose of the material comes into play and two pops can combine into one as a new seiche undercuts the platform formed by the previous seiche until a stable platform is formed.
d. Obstructions can break apart a large seiche pattern into smaller waves. Figure

The Provo level Shorelines
The two Provo level shorelines studied by Miller, Oviatt and McGeehin (Miller, et al.,, 2012), occurred in the short time period after the 15.6-Provo-1455 surging, judging by both their position above the tufa layer and the dating of the sediments.
These particular bars were not a result of earthquake-induced surging. Examining the bars throughout the basin, they do not have the cycle node/antinode variation between the center of the lake and the extremes. Another possibility would be that a landslide in the Cutler Narrows allowed Lake Bonneville to rise further after the Bear River Exclusion, however these bars do not present as consistent standstill shoreline elevations between different locations in the basin even after isostatic rebound is accounted for. These bars are not shoreline bars.

The Fall-from-the-Provo-level Shorelines
After the S15.6-Provo-1551 event, when the Bear River stopped replenishing the Bonneville basin, Lake Bonneville started a remarkably steady fall from the Provo level. This is evidenced by the regular spacing of shoreline patterns below the Provo level elevation. In the central part of the Bonneville basin, the shape of the basin changes to that of a shallow bowl where the lake volume per unit of depth increasingly becomes a factor. Since the fall is driven by evaporation, time also becomes an important variable. The Bonneville Flood occurred in a geological instant, measured in weeks or months, the Provo regression probably took hundreds of years. The isostatic stress relief was quite gradual and even if the release was sticky, the expectation would be that the intervals would be smaller and thus the steps of smaller magnitude simply because the system had time to adapt.
The sub-Provo level has examples of large bars, but for the most part the change in the appearance of shorelines below the Provo level is dramatic. A location at the northern end of the Hogup Mountains just west of the northern arm of the Great Salt Lake was chosen for study because it is an area with an approximately constant slope and with well-defined bands. At this location the bars appear in the Google Earth™ view as a series of narrow stripes, almost like ripple marks in the sand. (Figure 28). The shorelines are too fine to be resolved on a Google Earth™ profile, so they were identified visually and plotted. There is a pattern to the shorelines; they are not due to random storm events. Unlike the Bonneville Flood IRPS, the steps are not uniform across the range. The intervals increase as the level drops. This suggests a relationship between these shorelines and lake volume, which equates to the weight depressing the crust: at lower lake levels the basin flattens and a greater step change in elevation is required to provide the same decrease in volume of water.
Qeios, CC-BY 4.0 · Article, May 29, 2023 Qeios ID: G4DAH0 · https://doi.org/10.32388/G4DAH0 61/75 If the shorelines below the Provo level were climate related, then this regular pattern would represent an annual cycle since other climate patterns are just not that consistent. This would mean that the 74 meters studied here dropped in 38 years and at an accelerating rate. That exponentially increasing rate would have had the fall from the Provo level to the Great Salt Lake level occur in less than 100 years just by evaporation. The odds are against these as being climate related shorelines.
It is possible that the bar spacing is simply an erosion feature of a falling lake where a harmonic is set up by how the lake interacts with the shore: waves cut into the slope until it forms a platform and then the platform diffuses the wave energy until the lake drops enough to start to erode a new platform. The problem with that concept is that this type of erosion would remain constant for a constant slope, and that is not true for this location. As with the Bonneville Flood shorelines, the shorelines in the fall from the Provo level appear to be due to the effects of isostatic rebound. The rebound continued to occur in steps during the much slower fall from the Provo level, though the step pattern changed, reflecting the shape of the basin. As the level drops, the basin flattens, so a greater change in elevation is required to achieve the same change in load. Note that the location studied in Figure 28 was selected because it has a relatively constant slope through the elevation range so as to fairly reflect that the increasing spacing is not due to local contour effects.
With these smaller steps, any seiche generated would probably be of limited amplitude.
The fine pattern of shorelines visible here yields some other interesting information about this time period: a. The climate was relatively consistent as far as the direction of prevailing winds and was not much different to what occurs in the region today. The location studied on the north side of the Hogup Mountains has little fetch to the northwest and none to the south. Prevailing winds in this area come from the northwest and large storms are preceded by strong winds from the south. In other areas of the lake that are more exposed to strong winds from the northwest and south, a much coarser pattern of shorelines is all that remains, intermediate shorelines were probably overrun by large storm events.
b. This specific location is a spit deposit from earthquake-induced surging during the Provo level event. This is evident on the western edges (top of Figure 28). This left the location with a clean surface, devoid of conflicting earlier shorelines and with a soft sediment which could be easily molded. The reason the five Lake Bonneville earthquake events named in this paper are probably multi-segment events are these: a. The scale of these events. These were massive, basin-wide events. The Bonneville Flood earthquake event resulted in surging in the far corners of the lake: Little Cottonwood Canyon, Red Rock Pass and Keg Mountain.
b. During the Holocene, a major earthquake occurs somewhere on the Wasatch Fault every few hundred years. But on any given segment (such as the Salt Lake segment) the interval is measured in thousands of years. There may have been small individual segment earthquakes during the Bonneville timeframe, since in the Benson core there were over 50 laminated layers in the sediment core, and these could have come from storm events or from a single segment seismic event. Though the frequency in the thousands of years supports the concept that the major displacements along the fault were collected into single multi-segment events.
c. Others have suggested the possibility of long aseismic intervals during the Bonneville period (McCalpin, 1999, McCalpin andForman, 2002). The tectonic stresses continue to build during these intervals and the greater the accumulated stress, the greater the chances of an event on one segment carrying over to the adjacent segments.
d. Basalt ash eruptions are an integral part of the stress relief system associated with plate movement. The one-to-one correlation between basalt ash eruptions with these large surge events supports the concept that these were extraordinary events.
The larger question is whether the heavily populated Wasatch front is vulnerable to a multi-segment failure today. In a single-segment event, emergency support can be obtained from adjacent areas. A multi-segment event would overwhelm emergency services and shutdown the key infrastructure for the whole region. Electricity, water, natural gas, roads, and communications would all be affected for an extended period.
There are two possibilities: a. The multi-segment earthquakes were a product of the Lake Bonneville isostasy and Lake Bonneville is gone for the relevant time frame.
b. There are two cycles on the Wasatch Fault, the first is a shorter cycle of shallower, single segment earthquakes and the second is a much longer cycle of deeper, multi-segment earthquakes.
In review of the various trench studies of Wasatch Fault scarps, there does not appear to be compelling evidence that there has been a post-Bonneville multi-segment event. This means that at least the last 15ky have not seen a multisegment event, whereas during the Lake Bonneville period the longest interval between events was about 10ky and most were less. One possibility is that the underlying pattern of multi-segment earthquakes is still there, and lake isostasy just forced it into a higher frequency. A supporting data point to this is that as the lake rose, the time between multi-segment events shortened. This is a topic which requires the consideration of a larger group of individuals with a variety of fields of expertise.

The risks of underwater faults
The literature has an unfortunate mixing of terms when it comes to the impact of earthquakes on bodies of water. As an engineer, I am inclined to differentiate these effects by the resolution of forces: Seiche -A harmonic response in a closed body of water produced by an external force. In an earthquake-induced seiche, this is due to the vibration caused by the fault slip. A seiche was reported in Hebgen Lake in the 1959 Montana earthquake, and these occur regularly in swimming pools in California. The larger the body of water, the more intense the event required to set up a seiche. The IRPS events in Lake Bonneville were basin-wide and were the manifestation of an isostatic pop. The first IRPS event during the Bonneville Flood lifted close to 50 trillion kilograms of water possibly around 5 meters, or about the energy of forty-nine first atomic bombs.
Shock-type tsunami -This is the water-hammer effect. When a column of water is dropped, potential energy is converted to kinetic energy and since water is essentially incompressible, this is transmitted as a high velocity shock wave, traveling at 800km/hr. The energy generated is related to the height of the water column dropped and the distance it drops. When talking about the ocean and its great depths, this is a lot of energy. In Lake Bonneville, it would have been a lot less. Thẽ 500km wavelength of a shock-type tsunami is far greater than the width of Lake Bonneville, and that would break up any cycle development. While a shock-type tsunami would be expected in any body of water, it would not have been the dominant effect in Lake Bonneville.
Surge-type tsunami -when one side of a body of water is changed in elevation, the water must flow to achieve a new equilibrium. Momentum is the key. The surge has to be accelerated by gravity, so it is slow to start, but in a large body of water, that is a lot of mass put in motion. In a basin the surge will overshoot the equilibrium point in the first slosh. If the basin gets shallower as the edge is approached, the energy gets concentrated, and the overshoot is exacerbated. There has been a lot of modeling done on shock-type tsunami, but based on what happened in Lake Bonneville, surge-type tsunami may be the more important hazard in certain situations.
The speed of the surge is one tenth that of a tsunami, or around 80km/hr. With acceleration and deceleration at the extremes, the surge cycle in Lake Bonneville would have been measured in hours.
Displacement-type tsunami -this occurs when objects fall into a body of water. Landslides or calving blocks of ice typically cause this type of tsunami. It is only mentioned here to be comprehensive in terminology.

Human occupation of the Bonneville Basin
Qeios, CC-BY 4.0 · Article, May 29, 2023 Qeios ID: G4DAH0 · https://doi.org/10.32388/G4DAH0 65/75 Evidence of early human occupation of an area is easily erased by the elements of time. Consequently, any evidence of humans in a region has to be considered as potential evidence of a more widespread presence. Early humans have consistently proven to be highly mobile.
Human footprints found in the area of White Sands National Park, New Mexico have been dated to between 23kya and 21kya (cal), around the time of the last glacial maximum and while Lake Bonneville was still in transgression (Bennett, et al. 2021). This dating has been questioned because of the risk of carbon reservoirs distorting the results (Madsen, et al., 2022), however Pigati et al. maintain that this dating is robust because they checked their dating against "geologic, hydrologic, stratigraphic, and chronologic evidence" . The work of Pigati et al. supports the concept of a broader presence of humans in the western United States at a time when Lake Bonneville was still in transgression.
Large bodies of water would be attractive areas for early human occupation since they provide water in the winter, and they are a natural draw to game. The shores of Lake Bonneville would have been a logical location for human settlement.
In a 2021 paper Goebel et al. reported on trench studies conducted in the Bonneville Estates Rockshelter, an erosion feature cut into a rock hillside at the Bonneville high stand on the western edge of the Bonneville Basin (Goebel, et al., 2021). They found evidence of human occupation a little less than 15kya cal, at a time when Lake Bonneville had commenced its rapid climate-based fall from the Provo level. At the lowest extent of their excavation, they found an unmodified mammal long-bone fragment and dated it at 18.476kya cal. The fact that the bone was "unmodified" suggests that it was not the product of hunting and butchering. The dating places the limit of their excavation at a time 1ky before the Bonneville Flood, and probably during the last stages of the lake's transgression to high stand. The Rockshelter may have been uninhabitable at high-stand.
A deeper excavation of the Bonneville Estates Rockshelter might be warranted since it is one of the few locations which might have trapped the evidence of pre-high-stand occupation.
Earthquake-induced surging in Lake Bonneville, or other Pleistocene Lakes in the Great Basin, would have been catastrophic for early inhabitants of the region and the S17.4-Bonneville-1551 surge would have destroyed evidence of their presence if they had settled in the logical locations where streams of side canyons and basins emptied into the lake.
The oral history carried by survivors would have delayed resettlement. Once the level reached the Great Salt Lake level, the water was no longer suitable for drinking or fishing, so humans would tend to settle in areas further from the hazards of earthquake-induced surging.

The lessons from the Bear River Diversion
The exclusion of the Bear River from the Bonneville basin during the Provo level occupation resulted in a loss of about 40% of the inflow to the lake. The lake level plummeted at an unprecedented rate and this fall was only arrested when the flow was restored.
A 2016 white paper by researchers from Utah State University estimates that human activities have reduced the net river Qeios, CC-BY 4.0 · Article, May 29, 2023 Qeios ID: G4DAH0 · https://doi.org/10.32388/G4DAH0 66/75 inflows into the Great Salt Lake by 39% since 1847 (Wurtsbaugh, et al. 2016). Agriculture represents 63% of that, with the remainder going to industrial, residential, and commercial consumption (Ibid, 2016). The Great Salt Lake sees anomalously wet years, 2023 for example, but the trend of the last few decades led to a record low in 2022.
Lake Bonneville disappeared after 30ky due to a 40% reduction in inflow. Human activity has diverted 39% of the inflow to the Great Salt Lake and the lake level is dropping.
In a flat basin such as the Great Salt Lake's a small drop in level results in a very large loss in surface area. This exposes more of the mud flats. Those mud flats contain the concentrated heavy metals and other toxins from tens of thousands of years of accumulation. Anyone who has lived in the heavily populated areas to the east of the lake knows that winds from the west can pick up this mud and even on just a hazy day deposit a thin coating of mud on a previously clean car. Air quality is an issue.
Snow on the Wasatch and Uinta Mountains is the principal source of water for this region as well as being the reason for Utah's well-deserved reputation among skiers for the 'greatest snow on earth'. Lake-effect is one of the reasons for the deep snow in these mountains. Lake-effect is dependent on lake surface area. Lake-effect is a type of feedback loop, and a falling lake level puts us on the wrong side of that phenomena.
6. Conclusions / Summary of Findings 6.1. The Wasatch Fault has experienced five multi-segment earthquake series over the last 45 thousand years. There is a one-to-one correlation with the known, major basalt ash eruptions in the basin during this period. There may have been long aseismic intervals between these events. Isostatic deformation by Lake Bonneville and its effect on the Wasatch Fault may have been both the cause of the aseismic intervals and the trigger for the multi-segment events.
6.2. Alternately, there may be a longer-term pattern of multi-segment earthquakes on the Wasatch Fault. While this is unlikely, it would represent a significant concern in this heavily populated area. Studying the long-term frequency of basalt ash eruptions in the basin and a deeper sediment core may provide insight into this pattern.
6.3. The earthquake-induced surging in Lake Bonneville provides a record of this type of surging-type tsunami hazard and this needs to be considered in addition to shock-type tsunami and seiche when reviewing the hazards of underwater faults in other areas of the world. This not only applies to lakes but may also be relevant in ocean locations where a fault parallels a coastline. a. The crust rebound was sticky in nature, not smooth.
b. The rebound occurred in highly regular steps. The pattern of steps was dependent on weight removed and on basin topography.
c. IRPS may explain features in Lake Lahontan and other Pleistocene lakes of sufficient mass to exhibit isostatic effects.
d. Hazards associated with IRPS may be present in glacial lakes forming in regions undergoing deglaciation. IRPS could cause natural dams to fail.
6.6. The climate history of the Bonneville Basin needs to be revisited with the long-held "climate oscillations" removed.
6.7. The Heinrich stadials did not cause dry conditions in the Bonneville Basin and a rapid drop in the lake level as previously assumed. Recent research by others that these were wetter periods is supported by this paper. Surprisingly, there appears to be a link between stadials and multi-segment earthquake events. A rapid increase in isostatic load due to Lake Bonneville's level rising during a wetter period may have been the trigger of an overstressed fault. 6.8. There has been a long-term debate on whether there were early and late Provo levels and there is clear data supporting each theory. This paper resolves that and provides an explanation for the very rapid drop of this massive lake from the Provo level to the Great Salt Lake level. Provo shoreline sediment formed a natural dam in the Cutler Narrows. An earthquake dropped the lake outflow level in Cache Valley, isolating that area from Lake Bonneville and short-circuiting the Bear River flow directly out of the Bonneville Basin, excluding it from the hydraulic balance of the main body. Deprived of 40% of its incoming water, the lake level fell due to evaporation. A later earthquake in Cache Valley resulted in surging and seiche which overtopped and destroyed the Cutler Narrows dam, restoring Bear River flow to the Great Salt Lake basin and stabilizing the level.
6.9. During the Lake Bonneville transgression, the sediment record suggests instances of collapse of subterranean aquifers potentially resulting from isostasy. At Blue Lake in the Bonneville Basin, this would manifest as a dramatical outflow from the existing spring and result in an increased sedimentation and a local drop in the local Total Inorganic Carbon content of those sediments. These have been termed 'Water Events' in this paper.
6.10. The boulder field on the side of the south lateral moraine at the mouth of Little Cottonwood Canyon is evidence of a tsunami in what is now a desert climate area far from any ocean. This warrants a geological marker just to pique the interest of those in the area and to remind everyone that it is important to recognize when the facts do not fit current theory.
6.11. This paper provides additional evidence supporting the author's previous finding that the "grabens" in the area between Big Cottonwood Canyon and Little Willow Canyon in the Salt Lake Valley are actually fissures at the top of an underwater shift of massive areas of glacial till deposits at the time of the S17.4-Bonneville-1551 earthquake-induced surging. The G.K. Gilbert Geological Park at the mouth of Little Cottonwood is eventually going to need the educational panels updated, the actual story is a lot more interesting. That story includes the Bell Canyon terminal moraine splitting Qeios, CC-BY 4.0 · Article, May 29, 2023 Qeios ID: G4DAH0 · https://doi.org/10.32388/G4DAH0 68/75 during the event and the glacial lake at that location draining through the gap leaving a debris wall. A popular hiking trail goes up through that fissure and the hikers are oblivious to the significance of what they are hiking through.