TAPHOFACIES OF COASTAL LAKES BASED ON MOLLUSK DEATH ASSEMBLAGES: A CASE STUDY OF MIRIM LAKE

– Mirim Lake is the second-largest lacustrine body in Brazil, stretching for some 185 km and covering about 375 thousand hectares of water surface. This lake is part of the Patos-Mirim System, one of the world’s largest complexes of coastal lagoons, which is the result of the rise and fall of sea level over transgressive-regressive cycles triggered by glacio-eustasy during the Late Pleistocene and Holocene. Its characteristics and dynamics give it singular importance for studies on the genesis of shell accumulation in lagoons. This study aims to improve the knowledge of how the taphonomic signatures reflect the environmental characteristics and dynamics of Mirim Lake. Categorization of the arrangement and packing of death assemblages, as well as the orientation of shells, supported the biostratinomical approach. The taphonomic signatures included disarticulation, fragmentation, and corrosion of shells. Corrosion is the primary damage observed in the bioclasts, varying from a single loss of color to complete degradation. Differences in the characteristics of death assemblages observed in the dune fields and the lacustrine plain led to the recognition of two taphofacies. A more significant number of whole shells was observed in the dune field taphofacies, while sharp fragments and open-articulated bivalves in hydrodynamic unstable positions characterize the lacustrine plain taphofacies. The genesis of death assemblages and the taphonomic signatures of bioclasts were linked to three sedimentary dynamics attributed to the environment of Mirim Lake.

Extending for some 185 km and covering about 375 thousand hectares of water surface, Mirim Lake (Figure 1A) is the second-largest lacustrine body in Brazil (Oliveira et al., 2015;Valentini et al., 2020;Lopes et al., 2021), which gives it singular importance for the study of the genesis of shell accumulations in freshwater bodies.This lake is part of one of the world's largest complexes of coastal lakes and lagoons, the Patos-Mirim System, which consists of Patos Lagoon, Mirim Lake, Mangueira Lake, and smaller lakes and marshes (see Lopes et al., 2021).This extensive complex results from the rise and fall of sea level along the transgressive-regressive cycles triggered by glacio-eustasy during the Pleistocene and Holocene.
During the transition from the Late Pleistocene (ca.325 ka to 125 ka) to the Early Holocene (ca.8 ka), the landscape was flooded by marine transgression, and four barriers were formed (Figure 1B) (Rosa et al., 2017).Afterward, the following regression resulted in the formation of the lakes and lagoons in the back-barrier zone (Figure 1A) (Villwock & Tomazelli, 1995;Tomazelli & Villwock, 2000;Rosa et al., 2017).Consequently, observing fossils and modern marine and freshwater mollusks in death assemblages of lakes and lagoons is possible.Also, a complex deposit containing fossils of continental mammals along with marine mollusks has already been documented by Lopes & Buchmann (2008) from Passo da Lagoa.These characteristics make such deposits a singular window to observe the taphonomic bias on the bioclastic accumulations of large lakes of coastal plains that arise after the sea level falls.
Mirim Lake has a beach profile comprising a welldeveloped dune field and a lacustrine plain (Figures 1C  and D).Due to the changes in the sedimentary dynamic along the beach profile, differences in the taphonomic signatures of bioclasts are expected to be observed, leading to the recognition of different taphofacies.Thus, this study aims to understand how the taphonomic signatures reflect the environmental characteristics and dynamics of Mirim Lake and to characterize the taphofacies of this lake.Death assemblages observed on the lacustrine plain and the dune fields of Mirim Lake were studied to fulfill this objective.Based on the distinction of taphonomic signatures along Mirim Lake, the classification of the shell deposits into taphofacies models is proposed herein.

MATERIAL AND METHODS
Field data and samples were obtained in October 2022 as a part of an undergraduate field trip in Capilha Beach (Figure 1).Mollusk death assemblages (DAs) were sampled in the lacustrine plain and the dune field (Figures 1C and D).Eight square kilometers were surveyed to record, characterize, and photograph the DAs.
A total of 4,333 mollusk bioclasts from thirteen DAs arranged in the lacustrine plain and the dune field were studied.DAs were described according to their shape (e.g., narrow, broad, wavy, straight) and concentration of the shell accumulations (e.g., densely to loosely concentrated, scattered accumulations).Taphonomic analysis was carried out by surveying the following signatures: fragmentation degree (whole, fragmented); articulation of bivalves (closedarticulated, open-articulated, disarticulated); orientation of valve axis in plain view (i.e., unimodal, bimodal, randomly oriented); hydraulic stability of bivalve valves (e.g., convex/ concave upward); dissolution, bioerosion, and encrustation on bivalve valves.The bioclasts were qualified as whole (more than 90% of the valve preserved) and fragmented (less than 90% preserved).Fragments were classified as sharp or abraded (rounded).Corrosion was assessed only on bivalves and classified as initial (loss of luster, color, and minor corrosion of the surface), moderate (surface partially corroded), and severe (complete dissolution of external layers, full corrosion of umbo, exposure of the pearly layer).
To determine the azimuth orientation, the longest axis of the bioclasts was directed using a semicircular protractor.The angles were recorded in a Microsoft Excel spreadsheet.Two hydrodynamic shapes, triangular and elliptical, were recognized and used to measure the orientation of bioclasts.For bioclasts with elliptical shapes, the data were mirrored at the opposite angle to that observed in the protractor (e.g., 90º-270º).Lastly, rose graphs were made to illustrate the axis direction of the bioclasts.
A Motorola Moto g31 cell phone photographed the DAs in the field.In the laboratory, a digital camera, Canon EOS Rebel XT, with a lens of 100 mm, was used to take photographs of the specimens.A few shells were sampled in the lacustrine plain and the dune field to record the taxa present in the DAs and take photos of taphonomic signatures observed in the bioclasts.
All taphonomic data from 13 DA samples were pooled into a Microsoft Excel spreadsheet for preview and general analysis.The comparison between the type of sedimentary environment (lacustrine plain and the dune field) and the state of each taphonomic signature were assessed using contingency tables (Zar, 2010).The difference in the taphonomic data for each portion (i.e., the dune field and  Tomazelli & Villwock (2000).The black arrow indicates the relative location of Mirim Lake concerning the current Barrier-Lagoon System IV. C, shore of Mirim Lake delimiting the lake, lacustrine plain, and barrier system (dune field).D is a barrier system subdivision represented by dune fields (front dunes and dunes).E, frontal view of a 7-m-high dune.F, vegetated and non-vegetated areas of the dune field.
lacustrine plain) of the Mirim Lake was used as a criterion to separate them into distinct taphofacies.Moreover, differences in the results observed in vegetated and non-vegetated zones of the dune field were also noted, which led to the proposal of sub-taphofacies to represent each one.
The samples collected were deposited in the Museu da História da Vida e da Terra of the Universidade do Vale do Rio dos Sinos (Unisinos), São Leopoldo, Brazil.They were identified by initial ULVG followed by the ID number in the collection.

GEOLOGICAL SETTING AND THE STUDY AREA
The Pelotas Basin is located in the southernmost continental margins of Brazil and Uruguay.It is characterized as a subsided marginal and open basin filled with continental and marginal marine sediments (Barboza et al., 2008).Its genesis is related to the erosion of Precambrian, Paleozoic, and Mesozoic rocks after the opening of the Atlantic Ocean, with sediment accumulation between the Cretaceous and the Quaternary periods (Villwock & Tomazelli, 1995;Barboza et al., 2008).In their onshore portion, two depositional systems represented by alluvial fans and barrier-lagoon deposits are identified (Figure 1B).Alluvial fan deposits are considered Miocene age, while barrier-lagoon systems developed in the Quaternary (Tomazelli & Villwock, 2005).The transgressive-regressive events that reworked the alluvial fan system also generated the four Quaternary-aged barrierlagoon depositional systems that have been developed and shaped by high-frequency, 4th order (100 ka) glacio-eustatic cycles (Villwock & Tomazelli, 1995;Tomazelli & Villwock, 2005;Rosa et al., 2017;Lopes et al., 2020).Each barrier was probably formed during the transgression and preserved by forced shoreline regression (Tomazelli & Villwock, 2005).
Three sets of depositional sequences (i.e., progradationalaggradational, retrogradational, and progradationaldegradational) were proposed by Rosa et al. (2017) as a result of transgressive-regressive pulses that formed the barrier-lagoon system of the Coastal Plain of the state of Rio Grande do Sul (CPRS).The barrier-lagoon systems have been developed parallel to each other, following the coastline orientation.These include long sand barriers extending for more than 600 km, lagoons, lakes, and wetlands arranged on the back-barrier zone (Figure 1A).Their deposits are represented by the shallow marine and eolian sand facies (e.g., barriers) and mud facies of the bottom of lagoons and coastal lakes.Siliciclastic, fine to very fine-grained sands, with small amounts of feldspars, biogenic carbonates (from shell degradation), and heavy minerals characterize the sediments of barriers (Lopes et al., 2020).
The Mirim Lake originated in the marine transgression of the Middle Holocene and became a brackish lagoon about 7.6 ka ago.However, it had a fully saltwater condition ca.6-5 ka ago during the sea-level highstand, and the marine influence reduced ca. 4 ka ago with the fall of the sea level and closure of the connection with the ocean after the evolution of the sand barrier that originated the current shoreline (Lopes et al., 2021).
At the margin of the Mirim Lake, sand barriers are present as dunes covered by typical shoreline plants (Figures 1D-F).
Bivalve and gastropod shells are found along the beach profile (ca.180 m of extension), with the shell deposits occurring from the lacustrine plain to the front dunes (Figures 1C and  D).The dune fields are about 117 m distant from the lake, widespread parallel to the coast, and reach around 7 m above the lake level (Figures 1E).Mollusk and vertebrate fossils have been reported in several marginal zones, occurring at a few centimeters down in the subsurface or on the surface as exhumed materials (Lopes et al., 2021, references therein).

RESULTS
A total of 4,333 bioclasts arranged into 13 DAs were observed along the dune field and the lacustrine plain of Mirim Lake.They were arranged as scattered, loosely, and densely concentrated shell deposits (Figures 2 and 3).Of these, 57.44% of the bioclasts were whole (n = 2,489), 7.6% were abraded fragments (n = 532), and 27% were sharp fragments (n = 1,312).Many bioclasts identified as Corbicula fluminea Müller, 1774 showed a pattern of fragmentation irradiating from the center to the margin of the valve (Figure 2D); 8.7% of all bioclasts (n = 379) were articulated (closed = 7.4%, open = 1.3%;Figures 2C and F), and 35.5% of convex bioclasts (n = 644) were oriented in a hydrodynamic unstable position (Figure 2C); 87% of bioclasts (n = 3,756) showed initial stages of corrosion, while bioclasts with no damage (n = 275) and moderate damage (n = 249) corresponded to 6% each.Severe corrosion damage was observed in 1% of bioclasts (n = 53).A synthesis of the distribution of bioclasts according to the taphonomic variables is shown in Table 1.
Five DAs (n = 2,949, 68% of the bioclasts) were recorded in the lacustrine plain, while eight DAs (n = 1,384, 32% of the bioclasts) were recorded in the dune field.Dense concentrations of shell accumulations were recorded only in the vegetated portions of the dune field.In non-vegetated areas of the dune field and the lacustrine plain, the shell deposits were arranged as scattered and loose concentrations (Figures 2B and 3).Concerning taphonomic variables, the dune field and lacustrine plain taphofacies are recognized, differing from each other in all taphonomic signatures evaluated (Table 1, Figure 5).

Dune field taphofacies (front dune zone)
In the dune field taphofacies, the wind is the most effective agent in the transportation and remobilization of sediments (Tomazelli, 1993).This environment is subdivided into dunes and front dunes (Figure 1D).Front dunes are characterized by vegetated and non-vegetated areas (Figure 1F).The DAs were observed in this zone of the dune fields.Shell accumulations formed patches of densely to loosely concentrated and scattered shapeless shell deposits (Figures 2A and B).Whole bioclasts are more abundant than fragments (p < 0.0001) and totalize 75% of the sample (n = 1,040 of 1,384) (Table 1, Figure 5A).The ratio between abraded and sharp-edged fragments is significantly lower than in the lacustrine plain taphofacies (p < 0.0001) (Table 1).In contrast, the frequency of articulated shells and the subcategories of open and closed-articulated shells were significantly higher (p < 0.0001).A total of 258 articulated valves, between close (n = 255) and open (n = 3), were recorded in the dune field taphofacies (Figure 5B, Table 1).
Inspection of the plot of Figure 6A does not allow for the recognition of any preferential orientation of shells in the plan view.A total of 269 valves (f = 20%) were oriented in a hydraulic unstable position and are significantly different from the lacustrine plain taphofacies (p < 0.0001) (Figure 5C, Table 1).Corrosion classified as initial was recorded in 73% (n = 1,012) of the bioclasts, while 16% (n = 215 of 1,384) did not exhibit signs of dissolution (Figure 5D, Table 1).
The taphonomic data from two areas of the front dunes (i.e., vegetated and non-vegetated dune fields) were significantly different in almost all taphonomic signatures (Figure 7, Table 2).Non-vegetated dune-field sub-taphofacies show a higher percentage of valves in hydrodynamically stable than unstable positions (p < 0.0001).Based on these data, the classification of the dune field taphofacies into two sub-taphofacies was proposed.A description of each subtaphofacies is presented below.
In addition, sub-taphofacies of non-vegetated dunes correspond to paths devoid of vegetation that intersect the Table 2. Relative abundance of taphonomic signatures of the vegetated and non-vegetated sub-taphofacies of the dune field and the result of chi-square (x 2 ) test to similarity for each category.Chi-square significance: if p ≥ 0.05 = no significant difference; if p < 0.05 = significant difference.

Vegetated dune
Non vegetated areas of the front dunes (Figure 1F), with dispersed to scattered shapeless shell deposits.

Lacustrine plain taphofacies
The lacustrine plain is an emerged, flat, and smoothly sloped sand portion of the lake closer to the water body (Figure 1C).In this zone of the lake, the shell deposits were distributed as narrow and laterally elongated accumulations forming a sinuous string geometry (Figure 3).In these DAs, bioclasts were arranged as loose concentrations and scattered shapeless shell deposits (Figures 3A-C).
As observed in the dune field, the inspection of Figure 6B does not allow for the recognition of any preferential orientation of bioclasts.One thousand one hundred thirty-two convex bioclasts (f = 75%) were oriented in a hydrodynamic stable position (Figure 5C, Table 1).The rate of convex-up valves in the lacustrine plain is significantly higher than in the dune fields (p < 0.0001) (Table 1).Of the 2949 bioclasts, 2889   A, dominance of whole bioclasts in the dune field followed by the higher abrasion of bioclasts in the vegetated sub-taphofacies.B, frequency of abraded and sharp bioclasts considering only fragments.C, higher frequency of disarticulated valves than articulated ones to both sub-taphofacies of Mirim Lake.D, higher occurrence of hydrodynamic stable positioned valves only in the non-vegetated sub-taphofacies.E, higher occurrence of corroded bioclasts only in the vegetated sub-taphofacies.F, a similar proportion of corrosive damage between the two sub-taphofacies.
Considering the low energy gradient of lake environments, mechanical reworking is not expected to be the primary damaging agent (Cate & Evans, 1994).Nearly all bioclasts (94%) exhibit some degree of corrosion, resulting in the weakness of the valves and making them more susceptible to breaking.Furthermore, some bioclasts show signs of trampling (Figure 2D) that could be attributed to the presence of cattle grazing and local fishermen at the shore.
Corrosion levels vary from slight (initial = loss of luster and color) to severe damages (total deterioration of the affected zone of the valve, Figure 8).In one specimen of Diplodon ellipticus, it is possible to observe stains of dissolution (Figure 8D).At the same time, another has its pearly layer exposed due to severe dissolution (Figure 8B).Although the umbo is a thicker structure of the shell, some valves show the most accentuated damage in this portion (Figure 8A).A combined effect of corrosion and abrasion could have caused the complete loss of the umbo of the specimen illustrated in Figure 8A.Diverse factors and features (e.g., taxon, size, weight) are involved in the corrosion rates (Flessa & Brown, 1983).
Despite the small number, the articulated bivalves provide information about the sedimentary dynamic of the lake.At least 7.5% of all bioclasts were closed-articulated, indicating brief exposure and no reworking after dying.The opening of the valves is relatively fast.It is due to the functional anatomy of the ligament that forces the aperture of the bivalve after the decay of the soft tissues (adductor muscle).Moreover, the disarticulation (partial or total) is directly related to the sedimentation rate and energy gradient (besides the oxygen content of the environment and the exposure time and reworking of the articulated skeleton) (Allen, 1990;Brett, 2003;Gendy et al., 2015;Ilarri et al., 2015), which reinforces the hypotheses of brief exposure and no reworking.
Differences in the taphonomic signatures along the beach profile of Mirim Lake led to the recognition of two taphofacies: the dune field taphofacies and lacustrine plain taphofacies.
Densely concentrated shell deposits, with a higher proportion of both whole (75%) and closed-articulated bivalves (18%) deposited in the front dune zone, characterize the field dune taphofacies.In addition, at least 16% of bioclasts did not exhibit loss of color or luster.Dense concentrations were only observed in the vegetated zones of the dune fields, suggesting that the higher accumulation of the shells may result from their trapping in the vegetation combined with a low sedimentation rate.On the other hand, the better preservation of bioclasts (i.e., whole and articulated shells, preservation of color, and absence of corrosion signs) is usually attributed to low exposure in the taphonomic active zone (Meldahl & Flessa, 1990;Olszewski, 2004), which suggests that this shell assemblages probably represent recent deposits.
Some birds show a behavior known as avian preydropping behavior, which may produce concentrations of shell fragments (Maron, 1982;Cristol & Switzer, 1999;Switzer & Cristol, 1999).This behavior involves birds dropping the shells on hard ground to break them.However, the dune field at Mirim Lake is characterized by a soft surface that is inadequate for this behavior.
Concentrations of shells of Corbicula fluminea as artificial deposits made by fisherman activities have been described in Argentina (Labaut et al., 2021).Despite the occurrence of fisherman activity in Mirim Lake, this hypothesis is discarded because the bioclasts that form the shell concentrations on the dune field show different sizes, which refutes the idea that they could be used as bait.Moreover, the shell deposits show sinuous geometry, suggesting wavy reworking.Therefore, it seems more plausible to consider that the bioclasts were transported during flood periods (when the level of the lake rises) and were deposited about 117 m away (Figures 1C-D) when the lake's level fell.
Flooding of lacustrine plain and front dunes in the dune field is a standard process in Mirim Lake, ranging from 2 to 3 m (Oliveira et al., 2015).Since Mirim Lake has no connection to the sea, the floods are linked to the weather (seasonal rain and storms), while low-water periods are closely related to the use of the lake's water in the rice fields (Motta Marques et al., 2002).A graphic summarizing the seasonal level variation of the Taim region over several decades can be found in Motta Marques et al. (2002).
Bivalve shells show good floatability and thus are easily transported (Lever et al., 1964;Brenchley & Newall, 1970;Allen, 1984;Dent & Uhen, 1993).As a result, bioclasts would be deposited along the beach profile.Thus, it is plausible to infer the transport of bivalves to the front dunes during the maximum flooding periods and their subsequent trapping in the vegetation during the drought periods.
Wind-induced sediment dynamics are the primary process related to the transport of sand in dune environments (Tomazelli, 1993).Fast burials may result from wind-induced sediment dynamics, which would favor the preservation of bioclasts as observed here (whole and closed-articulated bivalves).Furthermore, sharp fragments are more abundant in non-vegetated than vegetated areas of the dune fields, which suggests that the bioclasts were differently affected in each zone.In summary, there is a difference in the time exposure of bioclasts in each dune field area.
The lacustrine plain taphofacies is characterized by shell accumulations varying from loosely concentrated to scattered.They are arranged as wavy, laterally extended deposits following the parallel line of the lake.Distinct characteristics of lacustrine plain taphofacies are the higher proportion of disarticulated and open-articulated bivalves, sharp fragments, and corroded bioclasts.
Despite the higher frequency of closed-articulated bivalves recorded on the dune field (18.7%,Table 1), the occurrence of open-articulated bivalves was higher in the lacustrine plain (1.8%) (Figure 2C, Table 1).The bivalves were observed in both convex down and convex up positions (Figure 2C).Due to the instability of the convex down orientation, they are quickly turned over to convex up by currents (Allen, 1990;Brett, 2003).In this sense, open-articulated bivalves oriented in an unstable hydrodynamic position (i.e., convex down, Figure 2C) may open after being stranded concomitantly with the fall of the water level.Open-articulated articulated bivalves in a stable hydrodynamic position (Figures 2C and  3C) suggest that lake currents can capsize them.
Although previous works have already demonstrated that convex-down valves are frequent in tidal flats during periods of fair weather (Brenchley & Newall, 1970;Allen, 1984;Dent & Uhen, 1993), a relationship between the occurrence of convex-down open-articulated bivalves and predator activities cannot discarded.
The sinuosity of DA margins (Figure 3A) is evidence of the action of waves that reworked the shells.At the same time, the abundance of sharp fragments points to the low residence time of these bioclasts in contact with abrasive elements.Studies on shallow lakes (e.g., Bengtsson & Hellström, 1992;Lövstedt & Bengtsson, 2008) have demonstrated that waves induced by winds are common, and the waves and currents generated during higher kinetic events such as storms can resuspend and transport the sediment of the whole lake.Higher water turbulence due to constant and strong winds associated with a shallow condition is attributed to Mirim Lake (Goulart & Saito, 2012).In this sense, the reworking of bioclasts by wind-induced waves can be expected on the lake shore.

CONCLUSIONS
Distinct taphonomic signatures were observed in the death assemblages along all beach profiles from the dune field to the lacustrine plain.This led to recognizing two taphofacies: dune field and lacustrine plain taphofacies.
Flood events and wind-induced sediment dynamics were associated with the genesis of the death assemblages observed in the dune field taphofacies, which resulted in a lower reworking of bioclasts (higher proportion of closed-articulated bivalves) and a higher abundance of abraded fragments, which highlights their exposure to abrasive agents.
A higher frequency of abraded and a low number of bioclasts showing no signs of corrosion was observed in the vegetated areas of the dune fields.In contrast, opposite results were observed in non-vegetated dunes, allowing to classify these as sub-taphofacies from the dune field taphofacies.
Wind-induced waves and low exposure to abrasive agents were conditions attributed to the lacustrine plain, evidenced by the wavy shape of death assemblages deposited on the shore and the abundance of sharp-edged fragments.
Finally, three sedimentary dynamics were recognized.The first two correspond to flooding and wind-induced sediment dynamics linked to the genesis of death assemblages in the dune field.The third corresponds to wind-induced wave dynamics related to the genesis of death assemblages in the lacustrine plain.

Figure 1 .
Figure 1.Map of the location of Mirim Lake and characteristics of the study area.A, Google Earth© image of the Patos-Mirim System and location of Capilha Beach at Mirim Lake.B, schematic profile of the barrier system of the Coastal Plain of Rio Grande do Sul (CPRS) with ages from systems I to IV. Adapted fromTomazelli & Villwock (2000).The black arrow indicates the relative location of Mirim Lake concerning the current Barrier-Lagoon System IV. C, shore of Mirim Lake delimiting the lake, lacustrine plain, and barrier system (dune field).D is a barrier system subdivision represented by dune fields (front dunes and dunes).E, frontal view of a 7-m-high dune.F, vegetated and non-vegetated areas of the dune field.

Figure 2 .
Figure 2. Field photographs of death assemblages in the dune field and lacustrine plain.A, densely to loosely concentrated death assemblage forms patches in vegetated front dunes (dune field taphofacies).B, scattered shell deposits in the non-vegetated area of the front dune zone (dune field taphofacies).C, shells scattered in the lacustrine plain with open-articulated bivalves (op.art.).D, loosely concentrated death assemblage from the lacustrine plain composed of whole and fragmented bioclasts and trampled valves (tra.).E, local gastropods (G) accumulation in the dune field taphofacies.F, scattered closed-articulated bivalves (clos.art.) deposited in the dune field taphofacies.C, D, and F, monospecific death assemblages composed of Corbicula fluminea; E, gastropod shells were identified as Cochliopidae.Not scale.

Figure 3 .
Figure 3. Geometry characterization of death assemblages in the lacustrine plain taphofacies.A, loosely packed concentration showing string geometry.B and C, scattered deposits.The dotted line indicates the wavy arrangement observed in the death assemblages of the lacustrine plain produced by wave reworking.Not scale.

Figure 6 .
Figure 6.Graphic showing the relative frequency of hydrodynamic orientation of the long axis of bioclasts in plain view.Mirim Lake is located to the west, and the dune field is to the east of the compass rose.A, orientation of bioclasts in the Dune field taphofacies.B, orientation of bioclasts in the Lacustrine plain taphofacies.

Figure 7 .
Figure 7. Mosaic plot showing the significantly different frequency (p ≤ 0.0001) of taphonomic signatures between vegetated and non-vegetated sub-taphofacies.A, dominance of whole bioclasts in the dune field followed by the higher abrasion of bioclasts in the vegetated sub-taphofacies.B, frequency of abraded and sharp bioclasts considering only fragments.C, higher frequency of disarticulated valves than articulated ones to both sub-taphofacies of Mirim Lake.D, higher occurrence of hydrodynamic stable positioned valves only in the non-vegetated sub-taphofacies.E, higher occurrence of corroded bioclasts only in the vegetated sub-taphofacies.F, a similar proportion of corrosive damage between the two sub-taphofacies.

Figure 8 .
Figure 8. Different degrees of corrosion were observed in the bioclasts of Mirim Lake.A, closed-articulated specimen of Corbicula (ULVG 14292) showing intense surface corrosion resulting in the complete degradation of the umbo, loss of color, and formation of dissolution pits.B, severe dissolution damage on a Diplodon valve (ULVG 14305) (complete loss of the superficial layers and subsequent exposure of the pearly layer).C, a valve of Corbicula sp.(ULVG 14294) showing severe corrosion (loss of color, formation of dissolution pits, and partial degradation of the umbo).D, a specimen of Diplodon ellipticus (ULVG 14304) showing dissolution stains and peeled umbo (initial level of corrosion).Scale bars: A-B = not scale; C-D = 10 mm.

Table 1 .
Relative abundance of taphonomic signatures of the lacustrine plain and the dune field taphofacies and result of chi-square (x 2 ) test to similarity for each category.Chi-square significance: if p ≥ 0.05 = no significant difference; if p < 0.05 = considerable difference.