Seaward transect

Core 1 (190 cm) was extracted near Laguna Mitla’s western, seaward edge from a water depth of ~50 cm. Two thin, dark brown organic-rich clay layers occur near the core bottom at 183–177 and 173–170 cm (Fig 2). The remaining sediment is mostly gray to brown clay, with low percentages of water (~35–40%), organic (6–12%), and carbonate (1–3%). Ti concentrations increase slightly up-core. With the exception of large spikes in S and Sr in the basal material and in Ca near the core top, elemental concentrations are fairly stable (Fig 2). Water (24–32%) and organic (~5–8%) percentages are depressed from 64–53 cm in the interval containing clay with fine sand. Small gastropods are scattered near the core top.

Core 3 (218 cm) was retrieved from ~50 cm water depth ~0.5 km east of core 1. A thick peat section from 212–164 cm lies on top of 6 cm of clay. Clay, marked by fairly consistent percentages of water (~30–40%), organic (~7–24%), and carbonate (~1–4%) dominates the remainder of the core (Fig 2). S concentration decreases up-core, while Ti, Sr, and Ca concentrations increase up-core (Fig 2).

Core 5 (479 cm), extracted ~3 km east of core 3 from ~1 m water depth, consists of gray sand throughout the bottom 8 cm. The remaining 471 cm is dominated by black peat with high water (~85%) and organic (~80%) values (Fig 2). The clay occurring from ~400–291 cm has a relatively high concentration of Ti, with the exception of four carbonate-rich layers at 294–292 (14%), 309–308 (8%), 327–325 (13%), and 394–392 cm (6%). These carbonate layers are marked by elevated concentrations of S, Sr, and Ca and low concentrations of Ti (Fig 2). Seven Ti-rich clay layers varying in thickness and composition occur within the otherwise homogenous peat section from 290–0 cm.

Landward transect

Core 2 (329 cm) was extracted 0.5 km landward of core 1 at ~50 cm water depth. The bottom section of the core (329–196 cm) is a dark, muddy peat containing thin gray clay layers with low water and organic values (Fig 2). The top 196 cm consists of brown to gray clay, with a dark, organic-rich clay section at 145–115 cm. Ti concentrations are elevated in the low-organic clay sections. At 145–144 cm, the concentration of Ti (5,428 ppm) is the transect maximum (Fig 2). Concentrations of S, Ca, and Sr are lower throughout this clay section than in the underlying muddy peat, with the exception of peaks in S and Ca near the core top, which contains scattered gastropod shells.

Core 4 (216 cm) was extracted 0.5 km east of core 2 at a water depth of ~50 cm. Water and organic percentages, and Ti, Ca, and Sr concentrations are highly variable throughout the bottom section (216–119 cm), which consists of brown to gray clay interlayered with scattered yellow sand layers (Figs 2 and 3). S concentrations are high in this material. The upper 119 cm consists of brown to gray clay with relatively constant water (30–40%), organic (7–10%), and carbonate values (1–2%). Above 119 cm, concentrations of S decrease significantly as Ti increases slightly. Ca concentrations increase near the core top, while Sr concentrations are highly variable throughout the section. Small gastropods are scattered throughout the sediments near the core top.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Elemental concentrations of surface samples.

Surface samples were collected from the coastal beach and adjacent of the coring sites. Concentrations from surface samples belonging to the same coring site were averaged and plotted. Titanium concentrations decrease eastward across the transects. Concentrations of sulfur, strontium and calcium, commonly-used marine indicators, are higher in the landward sites (2, 4, 6) than in the paired seaward sites (1, 3, 5). Beach samples contain relatively low concentrations of sulfur, strontium and calcium.

https://doi.org/10.1371/journal.pone.0161568.g003

Core 6 (129 cm) was extracted ~3 km east of core 4 from a depth of ~1 m. An alternating series of dark gray/black sandy peat and coarse gray sand occurs from 129–88 cm, with higher concentrations of Ti, S, Sr, and Ca in the sand layers (Fig 2). A dark gray sand layer from 88–56 cm is marked by low water, organic, and carbonate values (Fig 2), and higher S concentrations. A dark peat occurs from 56–45 cm. The top 45 cm consists of iterations of reddish, gray, and brown clay. A shell hash section at 45–33 cm is marked by a sharp spike in carbonates (32%) and concentrations of Ca and Sr. Ti concentrations are elevated in the reddish clay near the core top.

All surface samples are clay-dominant with the exception of the sample near site 6, which contains some shell hash and shiny, flaky silicates, possibly mica. S concentrations are higher near the landward sites 2, 4, and 6, than at the seaward sites 1, 3, and 5 (Fig 3). Sr concentrations are nearly an order of magnitude higher at site 6 than at any of the other five sites: site 1, 2, 3, 4, and 5. Ca concentrations are also higher near site 6. Ti concentrations are higher near the western sites, sites 1, 2, 3, and 4, and lower near the eastern sites 5 and 6. S, Sr, and Ca concentrations are relatively low in the beach samples (Fig 3).

Paleoenvironmental reconstruction

The multi-proxy data from core 5 are presented in detail elsewhere [32] to provide a 7000-year reconstruction of the coastal paleoenvironmental evolution and Holocene paleoclimatic changes at Laguna Mitla, and are thus not the main objectives of this paper. However, brief descriptions of regional environmental history and site evolution are presented here to provide a context for understanding the depositional patterns and processes within the lagoon, the focus of this paper. Core 5, the longest and best-dated core, provides a sedimentary record spanning the last 7000 years (Fig 2). At the beginning of the record (~ 6900 years BP), the core site was dominated by Rhizophora (red mangrove) [32]. By ~6200 years BP, the rising seas began depositing a mix of offshore clastics and entrained terrestrial sediments, marking the initiation of the beach barrier as documented along Mexico’s Pacific coast [33]. Clastic input was temporally variable due to varying degrees of wave energy from the spatially discontinuous and highly transient beach barrier, perhaps frequently subjected to perturbations from overwash and hydrodynamic processes. This highly dynamic period lasted until the rate of sea level rise slowed at ~5200 years BP [33], at which time the barrier became more stable and consolidated. From ~5200 years BP to present, the beach ridge plain has prograded seaward, and the site has existed as a backbarrier system isolated from the sea by the beach barrier. The build-up and consolidation of the beach barrier increasingly isolated the backbarrier lagoon from the ocean, so that water level at the coring sites has since been essentially controlled by precipitation, rather than sea level. Throughout the last 5000 years, significant environmental changes at the site were registered sedimentologically, geochemically, and palynologically by the episodic alternations between peat and inorganic sediments. Peat, dominated by Rhizophora and Laguncularia (white mangrove) pollen [32] indicates low water levels and a wetland environment, with salinity levels affected by the seepage of marine water through the barrier. On the other hand, the inorganic sediments, characterized by clay rich in Ti and the regional pollen signal, suggest higher backbarrier water levels and the transformation of the site from a wetland to an open-water lagoon, driven by an increase in precipitation and fluvial discharge to the basin. In core 5, seven such clay bands occurred during the past 4500 years, indicating episodes of wet climate and lagoon phases existing at approximately 4430–4270, 4080, 3950, 3680–3490, 3170–3080, 2990–2870, and 1680–0 years BP. The timing of these wet periods corresponds well with paleoclimatological evidence from the nearby coastal Laguna Tetitlan [34] and the Middle American Trench, located offshore of Oaxaca [35]. These abrupt sedimentary transitions are recorded for matching temporal intervals in cores 2, 3, 4, and 6. Distinct peat/clay transitions occurred at approximately 4500, 4300, and 1900 years BP in core 2, at 4600 years BP in core 3, and at 950 years BP (inferred) in core 6. It should be noted that not all six oscillations recorded in core 5 occur in all cores, and some undated, less distinct organic/inorganic transitions are present in other cores (e.g., the three LOI peaks in the lower half of core 6). The lithological differences among the six cores can be explained by the sites’ different sensitivities in recording water level changes, which are a function of water depth, sediment supply, sedimentation rate, and local habitat and vegetation (e.g., mangrove swamp versus mud flat), all of which vary spatially. Nevertheless, the onset of several of these clay layers, especially the ~4500 years BP event, indicating lagoon phases or wet episodes, are broadly synchronous among cores. It is also remarkable that the uppermost sediments in all six cores are characterized by lagoonal clay, which certainly represents the limnic deposits formed in the depositional environment of the modern shallow lagoon. Radiocarbon dates and inferred ages obtained from cores 2, 5, and 6 suggest that the modern lagoon phase has existed continuously for the last 1900 years.

While we had considered the possibility that the clay intervals in Laguna Mitla result from land movements attributable to either sudden subsidence or uplift from earthquakes or slow-slip events covering longer time periods, this notion has been rejected due to a lack of sedimentological evidence of sudden land movements (e.g., erosive contacts) in the cores. In addition, geochemical evidence of marine incursions is notably lacking; instead, the clay intervals are characterized by the enrichment of terrigenous chemical elements such as Ti. Along the coast of Guerrero, documented land movements from earthquakes range from ~7–15 cm [11] for coseismic uplift and 5–14 cm for slow slip events [36], insufficient to explain the formation of ~1 m deep lagoons, nor their relatively sudden transformation into mangrove-dominated swamps or mud flats.

Sedimentary processes and provenances

The distributions of the elemental concentrations display distinct spatial patterns. The concentrations of Ti, which is commonly applied to identify terrestrial sediments [37] are many times higher in surface sediments collected from the western sites (1, 2, 3, and 4) than the eastern sites (5 and 6) (Fig 3). This suggests that slopewash, associated with large precipitation events and delivered by the ephemeral stream located in the northwest edge of the study site, is the primary source of this element. This spatial relationship remains constant throughout our record as Ti concentrations are higher at all times in core 4 than in the eastern cores 5 and 6 (Fig 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Elemental concentrations of sediment cores.

Titanium, sulfur, calcium, and strontium concentrations vs age for seaward core 5 (black), and landward cores 4 (green) and 6 (red). Titanium concentrations (a.) are higher in core 4, nearest the input stream, than cores 5 and 6, farther east. Concentrations of sulfur (b.), calcium (c.), and strontium (d.) are generally higher in landward cores 4 and 6, than the seaward core 5. Spikes in marine elements are largely synonymous with shells (whole, hash) and/or sand, all largely absent in core 5 with the exception of basal sediments. A major marine event was posited from two landward sites, marked by a star and square in Fig 1 [12, 14]. A matching event is absent in the cores from this study, with no sand occurring at any site from 3000–4000 years BP.

https://doi.org/10.1371/journal.pone.0161568.g004

Due to their high concentrations in seawater and marine sediments, Sr, S, and Ca are commonly used marine indicators. In Laguna Mitla, however, surface sample concentrations of these elements are higher in the landward sites than the seaward sites (Fig 3), with the highest concentrations of all three elements occurring at site 6, ~ 1 km landward of the lagoon’s edge and adjacent to Isla Magueyes. This suggests that these elements do not originate from marine material delivered over the beach barrier, but rather the clastic and carbonate materials eroded from the relict beach ridge in the center of the lagoon. This is further supported by the low Sr, S, and Ca concentrations in the modern beach sand (Fig 3). Many event types can erode and redeposit these materials into the area of the landward transect, including earthquakes, tropical cyclone rain events, increased runoff during climatologically wet periods, and changes in land use.

This spatial pattern extends throughout the entirety of our record. Concentration maxima for S, Ca, and Sr occur in the landward cores 4 and 6, especially in intervals dominated by shells (whole, hash) and/or sand (Fig 4). The spikes in elemental concentrations in the uppermost sections of landward cores 2 and 4 (Fig 2) are not matched by corresponding spikes in seaward cores 3 and 5. This indicates that the material was not transported from a seaward location by marine incursions. Possibly these spikes result from the presence of gastropods, or ancient marine materials eroded from Isla Magueyes and other relict beach ridge remnants in or adjacent to the lagoon.

Revisiting geological evidence attributed to marine incursion events

Large seawater intrusions can be expected to transport coastal sediments (e.g., sand) and marine organisms (shells, diatoms, foraminifera) inland. Marine deposition can vary from isolated overwash lobes to large sand sheets, often tens of centimeters thick [38, 39]. Tsunami waves are highly dynamic and can be extremely powerful, capable of transporting large objects landward [40], and depositing sand and other fine materials many kilometers inland [41]. Common characteristics of tsunami deposits include landward and seaward layering [42], landward fining [5], upward fining [43], rip-up clasts [7], and basal erosion [44].

The occurrence of a large marine incursion at ~3400 years BP for the area has been suggested [12–14], based on evidence found in core ACA 04–06 extracted from wetlands 5 km behind the coast in the northwestern part of the lagoon (red star, Fig 1), and core ACA 03–02 extracted landward of Isla El Conejo and 4 km from the coast (red square, Fig 1). Core ACA 04–06 [12], otherwise dominated by mud and silt, includes a ~60 cm unit of fine to coarse sand that contains rip-up clasts and high concentrations of such marine indicator elements as Na (sodium), Sr, and Ca, as well as a fining-upward stratigraphy. The authors suggest that this layer resulted from a marine incursion associated with an extremely rapid rise in relative sea level. Due to the layer’s sedimentary characteristics, thickness, and distance from the coast, a large tsunami was suggested as the probable delivery mechanism [12]. The downward movement of this event, unlike the uplift normally associated with historical earthquakes along the Guerrero coast, is posited to have been the result of a megathrust event that ruptured the “entire coupled plate interface” [13]. The ~6 m core ACA 03–02 [14] contains a 93 cm thick interval containing marine diatoms, sponge spicules, shells, and high concentrations of marine elements (Na, Sr, S), posited as resulting from marine intrusion with an extrapolated date of ~3500 years BP. This core does not include a sand unit corresponding to that found in ACA 04–06 [12]. The interval posited as resulting from marine intrusion also contains a peat layer [14].

It would be expected that such a marine incursion would have deposited a clear sedimentary signal at our sites, located behind the modern beach ridge plain and ~4 km closer to the coast than the ACA core locations. However, except for the basal sediments, sand layers are absent in cores 2, 3, and 5. Although distinct sand layers do occur in cores 4 and 6, they are much thinner (~1–30 cm) than the posited event deposits in core ACA 04–06, and they fail to exhibit common characteristics of tsunami deposition, including erosional contacts and deposition of marine shells. The elevated concentrations of S, Sr, and/or Ca in the surface samples, and downcore in cores 4 and 6 most likely represent material eroded from nearby Isla Magueyes. Although fine sand occurs from 64–53 cm in core 1, and clay/silt from ~84–79 cm in core 2 (decrease in LOI), core chronologies and stratigraphic correlations indicate that these layers were deposited much more recently than 3400 years BP, as evidenced by a sample dated to 1990±30 ¹⁴C years BP (1940 years BP) taken from 30 cm below the clastic layer in core 2. Furthermore, these sedimentological signatures in the westernmost cores 1 and 2 are not apparent in the eastward located cores 3, 4, 5, and 6, and may therefore suggest erosion and redeposition of remnant beach ridge materials mixed with sediments deposited from the slopes. Our results therefore do not support the identification of a significant marine incursion at 3400 years BP [12]; nor do our findings show definitive evidence of a large megathrust earthquake resulting in significant coseismic subsidence at this time [13, 14]. The lack of definitive stratigraphic evidence of tectonic movement, and the coherent, progressive ecological succession displayed in the multi-proxy record from Laguna Mitla [32], instead support the scenario of relative tectonic stability during the Late Holocene, previously implied by the presence of alluvial and deltaic plains observed by Ramírez-Herrera and Urrutia-Fucugauchi [45].

The presence of shells may explain the high levels of marine elements, particularly Ca, a major component of both terrestrial and marine shells, in core ACA 03–02, extracted landward of Isla El Conejo [14]. Either Isla Magueyes or Isla El Conejo, or other relict beach ridges further inland, are possible sources of the sediments and marine diatoms, which are capable of attaching to sand grains [46].

The low, episodic spikes in S, Sr, and Ca for the seaward cores 1, 3, and 5 in this study may possibly be explained by the episodic breaching of Laguna Mitla’s outlet channel following large precipitation events. Such breaching, a common occurrence for barred inner-shelf lagoons, has been documented from the nearby Laguna Nuxco [15], and can result in the intrusion of marine waters until the breach is closed. Minoura and Nakaya [47] documented saltwater intrusions which, while not severe enough to deposit sediments, were capable of leaving a geochemical signature in the sediments. Given the height, width, and stability of the beach ridge we believe that the most likely sedimentary signature of seismic activity over the last ~5000 years would be increased slopewash following the loosening of sediments associated with violent earth movement rather than deposition of thick sediment layers several kilometers inland.