Great Salt Lake: A globally-important ecosystem threatened by water overuse

Saline lakes around the world are facing a “desiccation crisis”: threatened by water overuse and climate change, with wide-ranging consequences to regional and hemispheric ecosystems, air quality, weather patterns, economic activities, and more [1–3]. In this paper, we present a case study of the largest terminal lake in the Western hemisphere: Great Salt Lake, in northern Utah within the arid Great Basin, which has experienced a sustained decline in lake level driven by human overconsumption in its watershed (Fig 1) [1].

Download:

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

Fig 1. Great Salt Lake field sites and hydrograph.

(A) Map of Great Salt Lake showing the north (N) and south (S) arms of the lake with major sites described in this paper: USGS lake elevation sites 1001000 (ES1, Saltair site) and 10010024 (ES2, Causeway site), weather station sites KUTSYRAC22 (WS1) and KUTSYRAC27 (WS2), Buffalo Point microbialite reef sites (BP), and Ladyfinger Point (LFP). Left inset shows the location of Great Salt Lake (blue) in northern Utah, USA. Right inset shows the northern tip of Antelope Island. Map baselayers are from the U.S. Geological Survey, National Geospatial Program (https://basemap.nationalmap.gov/arcgis/rest/services/USGSTopo/MapServer and https://www.sciencebase.gov/catalog/item/52c78623e4b060b9ebca5be5). (B) Satellite images of Great Salt Lake from October 29, 2012 and (C) October 28, 2022 showing the shrinking shoreline of Great Salt Lake (MODIS corrected reflectance images from NASA Worldview). (D) Map of Great Salt Lake (at 1280 m elevation) showing the approximate location and extent of submerged vs. exposed microbialite reef areas in summer 2022, after Baskin et al. [13]. (E) Lake hydrograph from 1848 to 2022; area highlighted in gray is expanded in (F). The dashed gray line in both figures shows the historical (1963) lake lowstand. (G) Detail of field sites at Buffalo Point, with logger sites as vertical bars (B and B3), recovery experiment sites as horizontal bars (RA–RC), and the microbialite monitoring quad with monitored (M1–M3) and cored (C1–C3) microbialites. The 2020 shoreline is also shown as a dashed line. The underlying aerial view (from Google Earth) shows the site in May 2022, with bright areas showing exposed, desiccated microbialites.

https://doi.org/10.1371/journal.pwat.0000100.g001

Great Salt Lake comprises not only the hypersaline open water but also distinct habitats along a salinity gradient, including fresh- to brackish-water estuaries and wetlands where rivers enter the lake, and expansive mudflats and playas. The main body of the lake is segmented by a rail causeway, which isolates the salt-saturated north arm (Gunnison Bay) from the south arm (Gilbert Bay), which encompasses our study site. The south arm supports a relatively simple but significant food web (Fig 2A); Great Salt Lake is a hemispherically important ecosystem [4] that supports millions of resident and migratory birds [5, 6] and a brine shrimp industry that harvests cysts used as feed in global aquaculture [7]. The lake is a hypersaline Cl-Na-SO₄-Mg-dominated system [8, 9] with chemical and biological factors contributing to its “carbonate factory” [10] despite moderate modern pH values; carbonate deposits include oolitic sand, organic-rich carbonate mud, and mounded reef-forming microbialites [11, 12].

Download:

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

Fig 2. Great Salt Lake’s microbialite-supported ecosystem.

(A) Simplified ecosystem diagram of Great Salt Lake’s south arm, illustrating the importance of the lake’s microbialites and the effect of microbialite exposure. Dashed arrows represent life cycle stages, solid arrows represent consumption. After Baxter (2018) [63] & Belovsky et al. (2011) [35]. (B) Stereo-photomicrograph of a Great Salt Lake microbialite piece imaged at 10x showing a healthy periphyton community, with three-dimensional clumps of Euhalothece bound by extracellular polymers, and white points of carbonate highlighted with arrows. Sample collected from Site B on July 7, 2020. (C) Phase contrast photomicrograph of a healthy microbialite periphyton community sample imaged at 400x magnification. The greenish mass is a clump of Euhalothece. Also visible in association with the Euhalothece mat are a pennate diatom (arrow 1), filamentous organism (arrow 2), and green alga (arrow 3). Sample was collected October 10, 2019 from Site B. (D) Positive phase and differential interference contrast photomicrograph imaged at 1000x magnification of Euhalothece culture from a Great Salt Lake microbialite sample collected in 2019 at Antelope Island State Park [14].

https://doi.org/10.1371/journal.pwat.0000100.g002

Great Salt Lake is currently threatened by a rapid decline in lake levels and consequent increase in salinity. Layered onto normal decadal cycles in precipitation [15], its watershed has been impacted by the megadrought that has gripped the western United States since 2000, which has been worsened by anthropogenic climate change [16]. An even greater threat to the lake, however, has been the overuse and diversion of the waters that would otherwise feed Great Salt Lake for agricultural, industrial, and municipal uses. Such consumptive water diversions are estimated to have reduced the lake’s volume by >60% [17]. As a result, the lake has shrunk to historic low levels in the past decade (Fig 1E and 1F), following a pattern of water overuse leading to lake demise seen in ecologically-important saline lakes around the world [1, 5]. Water overuse in Great Salt Lake’s watershed has substantially impaired the lake’s resilience to future changes in regional hydroclimate [18]. It has become clear that, without an overhaul of water use policy and practice in the watershed, the lake could soon be lost [19].

Great Salt Lake’s ecosystem-critical microbialites in peril

Low lake levels and consequent shoreline shift has exposed hundreds of square kilometers of microbialite mounds, which occur in extensive reefs [13] in Great Salt Lake’s near-shore environments and benthos. Microbialites, carbonate mounds formed by interactions of microbes with the lake’s chemical environment, are of academic interest as analogues for economically important hydrocarbon reservoirs [11] and paleoenvironmental records [12, 20, 21]. Microbialites are not unique to Great Salt Lake; indeed, other lakes threatened by modern lake level decline are home to similar structures, including microbially-influenced carbonate mounds in several other lakes in the Western flyway, Australia’s Lake Clifton in Australia, Ethiopia and Djibouti’s Lake Abhe, and the giant microbialites of Turkey’s Lake Van [22–24].

Although typically viewed as relics—geobiologic curiosities that provide a window into Earth’s deep past and life’s evolution—microbialites are beginning to gain appreciation for their role in supporting modern ecosystems (e.g., [25, 26]). Perhaps the best-studied example of this is Great Salt Lake, where it has become apparent that the lake’s microbialites serve a critical function in the lake’s overall ecosystem. Robust photosynthetic microbial mats (periphyton, adopting for this paper the broad definition that acknowledges the presence of a diverse community, cf. [27]) exist on the surfaces of microbialites in Great Salt Lake when conditions—primarily submergence and salinity—are favorable. The periphyton is dominated by a single genus of halophilic, coccoidal cyanobacteria, Euhalothece (Fig 2), although other phototrophs, including diatoms, green algae, and flagellates, can be seen associated with the mats, and the mat community is taxonomically and functionally diverse [12, 28, 29]. Microbialite periphyton conservatively contribute one third of the primary production in Great Salt Lake [30, 31].They are the primary food source for brine fly (Ephydra spp.) larvae [32, 33], as well as a seasonally important food source for the lake’s economically important brine shrimp (Artemia franciscana) [7, 33]. Additionally, microbialites, which are presumably built by carbonate production facilitated by their periphyton [28], offer stable oases in the shifting oolitic sands and carbonate mud that compose most of the Great Salt Lake benthos. This makes them crucial habitat for Ephydra larvae, which depend on the microbialites for both food and pupation habitat [32, 33]. The organisms that microbialite periphyton support feed in turn feed millions of birds that depend on the lake ecosystem [5, 30, 32–36] (Fig 2A). Lake level fall is subjecting microbialites and their periphyton to desiccation.

In autumn 2022, roughly 40% of the lake’s microbialites were subaerially exposed (Fig 1D, [13, 37]) and desiccated (Figs 1G and 3), representing a substantial loss of productivity. A critical question for the management of Great Salt Lake and its associated watershed is how effectively and under what conditions the microbialite periphyton communities persist. If policies are enacted that allow lake levels to rebound, do microbialite periphyton communities recover their ecosystem function? We investigated these questions during two successive summers where historic lake lowstands were reached and exceeded (2021 and 2022); this paper presents the findings along with other recent data on Great Salt Lake’s microbialite periphyton communities.

Download:

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

Fig 3. Time series photographs from Buffalo Point field sites showing the effects of lake level fall.

(A) Site B instrument pipe in July 2019 (left; top of pipe highlighted with arrow) vs. August 2021 (right). (B) Site B3 microbialite site in July 2021 (left) vs. July 2022 (right).

https://doi.org/10.1371/journal.pwat.0000100.g003

Field sites, time series data logging, and sample collection

The work described in this study focused on a microbialite reef on the northern end of Antelope Island in Great Salt Lake (Fig 1). GPS coordinates of all measurement and sample locations are provided in S1 Table. Time series water pressure, temperature, and downwelling irradiance were measured every 15 minutes collected using data loggers (temperature/pressure: HOBO U20L, light/temperature: HOBO MX2202, Onset Computer Corporation) attached to a PVC pipe anchored to the lake bed. In March 2019, the instrument site was placed in 0.9 m deep water in a microbialite reef ~75 m from that date’s shoreline (Site B; Fig 1G). In August 2021, Site B became subaerially exposed (Fig 3A) and the logger pipe was moved to a deeper site ~150 m farther lakeward from the 2019 shore (Site B3; Fig 1G). In addition, manual water depth, visibility, salinity (using a handheld 0–28% refractometer with automatic temperature compensation; measurements are reported as a % by mass), density (using a brewing hydrometer), and temperature (using a digital aquarium thermometer) measurements were collected monthly to seasonally, along with microbialite surface observations. Lake elevation data were obtained from two U.S. Geological Survey monitoring sites in the lake’s South Arm (Fig 1A) in order to provide a continuous record of lake elevation during the period of our study: one near Saltair (ES1: Station 10010000, the standard site for Great Salt Lake south arm elevation measurements, which had an interruption in data collection from 2022-09-28 to 2022-12-14 due to historic lake level fall), and the other on the railroad causeway (ES2: Station 10010024, which operated from 2020-06-08 onwards) [38]. Multiple manual field measurements of site water depths at each site were then used to determine depth offsets vs. lake level (i.e., site elevation). During dates when both ES1 and ES2 recorded data, daily mean values at the two sites were averaged for inferring water depth changes at our field sites.

Weather data

Weather data for 2019–November 2020 was obtained from a station on Antelope Island, located 4 km from the field site and operated by Antelope Island State Park (WS1: KUTSYRAC22, Ambient Weather WS-2090; Fig 1A); the station was non-operational beginning in November, 2020. Data for nearby stations available on WeatherUnderground (wunderground.com) were analyzed to find a new station with values consistent with those measured at KUTSYRAC22; for the full analysis see the file in the Open Science Framework data archive for this study (hereafter referred to as OSF archive). The station with the best coverage and closest similarity to KUTSYRAC22 was determined to be a private station located 14 km from the field site (WS2: KUTSYRAC27, Ambient Weather WS-2902; Fig 1A), with publicly available data retrieved and used in this study with permission from the station owner. For analytical purposes, measured weather values were averaged when data from both sites were available.

Microbialite field monitoring & core sampling

In addition to general observations collected during the long-term monitoring work, detailed systematic monitoring of microbialites at the study sites was conducted from July 27–August 17, 2021 and July 12–August 2, 2022 as part of the Weber State University GETUP (Geoscience Education Targeting Underrepresented Populations) Summer Research Experience program. Microbialites monitored in summer 2022 were additionally visited and sampled sporadically until October 20, 2022. For this work, microbialites were flagged for repeat photography and sampling. In 2022, monitored microbialites (in addition to logger Site B3 and recovery experiment Site RC) were located within a roped-off rectangle (quad) to protect them from foot traffic, as the lake’s microbialites are currently unprotected. Each flagged microbialite was photographed during each visit, and the location of different colored bands on the surface of each microbialite were measured using a homemade surveying device (Fig S1.1 in S1 Appendix): the vertical distance of the band from the water surface, corrected for fluctuating lake elevation, was used to compare band height across dates. The vertical distance from the sediment/water interface was also measured, however the soft and mobile nature of the sediment surrounding the microbialites made measurement from the water surface the more reliable measurement.

In 2022, core samples from microbialite tops were collected using 50 mL syringes with the tapered tip cut off, producing a coring tube that could be pushed by hand directly into the incompletely-lithified surfaces of the microbialites, to a depth of up to 4–7 cm. Cores were then extracted using the syringe plunger onto core cradles with a scale, and photographed. The 3–7 cm cores were sectioned in the field into three roughly equal 1–2 cm top, middle, and bottom (deep) sections using a sterile scalpel, then stored in sterile 15 mL centrifuge tubes on ice for transport to the lab. For downstream analyses, collected sections were coded based on their original depth within the microbialite as belonging to one of several horizons, with horizon T capturing the uppermost 0–1 cm, horizon M capturing 1–2.5 cm, and horizon B capturing 2.5–4 cm (see S1 Appendix for details). Back in the lab, core subsections were ground to a paste with a sterilized mortar & pestle to homogenize, then aliquoted for microscopy and chlorophyll extraction as described below.

Recovery experiments

Recovery experiments involved submerging pieces of a desiccated microbialite back into lakewater, incubating for varying lengths of time, and recovering them for measurements of community regrowth. The desiccated microbialite (t₀ control) was collected in October 2016 from the beach at Ladyfinger Point (Fig 1A) at roughly 1278.6 m elevation, indicating that it had been subaerially exposed for at least two years at the date of collection, after which it sat undisturbed and dry on a laboratory windowsill until its use in this experiment. The t₀ control microbialite was broken into 1–7 g pieces, which were placed into individual nylon mesh bags (mesh size = 350 μm) that were attached to submerged PVC anchors (Fig S1.3 in S1 Appendix) such that the samples hung suspended in the middle of the water column at the start of the experiment. In 2021 (September 27–November 12), experiments were run at two sites: Site RA (near the edge of a microbialite reef with significant water movement), and Site RB (in the middle of a reef surrounded by microbialites with healthy periphyton), both of which had water depths to sediment of ~ 20 cm, and samples were suspended at ~ 10 cm beneath the water surface. Triplicate samples were collected at timepoints from 10–40 days. The experiment was repeated in 2022 (September 27–November 12) at Site RC (adjacent to logger instrument Site B3 in the middle of a different, initially healthy microbialite reef), at a water depth to sediment of ~ 40 cm, and samples were suspended at ~ 30 cm beneath the water surface. Triplicate samples at Site RC were collected at timepoints from 0–100 days. All three sites were hydrologically connected (open) to south arm lake water at the beginning of the experiments, however, Site RC became hydrologically closed during the course of the experiment due to rapid lake level decline.

In 2021, samples collected at each timepoint were subsampled in the field for pigment and DNA extractions. Pigment subsamples were collected with surrounding lakewater in sterile 1.5 mL centrifuge tubes wrapped in electrical tape to minimize light exposure. DNA subsamples were collected by first swabbing an area of each sample and mixing the swab in DNeasy PowerSoil kit (Qiagen, Cat. # 12888–50) bead tubes that had been prepared with kit lysis solution. Then, a ~ 5 mm solid piece of each sample was also broken off using sterile tweezers and added to the same bead tube. Both pigment and DNA subsamples were flash-frozen on dry ice and stored frozen for transport. Back in the lab, DNA samples were thawed, vortexed for 5 minutes, and stored frozen at -20°C until extracted following kit protocols (see S1 Appendix for additional details). Pigment samples were processed immediately following the protocols described below.

In 2022, instead of processing samples in the field, samples were stored in sterile centrifuge tubes placed on ice for transport, and processed in the lab. Whole samples were then ground and processed following the pigment extraction and microscopy protocols in the same manner described for the core samples, and DNA extractions were performed using the same protocol as for the 2021 recovery experiment samples.

Laboratory desiccation experiments

A submerged microbialite was collected near Site B3 in July, 2022, then placed in an incubator at 30°C with one full-spectrum LED lamp (24W 3500 K full-spectrum lamp, Juhefa, Cat. # B08S7VSX6) and one UVA/UVB CFL lamp (23W 6500 K UVA/UVB lamp, Lucky Herp, Cat. # B082DYBQLL); spectra are shown in Fig S1.4 in S1 Appendix. The microbialite was allowed to desiccate for several days to weeks, and rinsed with distilled water at intervals between 6–52 days to simulate rain events. Microbialite surface coloration was measured immediately before and after rinsing events and periodically thereafter. Photographs were taken alongside a color card under standardized light conditions in order to measure surface coloration using the method described below. In addition, the dry mass of the microbialite was measured to assess removal of surface carbonate (weathering) during rinsing events.

Color analysis

Microbialite surface coloration—specifically, the relative amount of green—was measured as an indicator of surface pigmentation for field microbialites and the microbialite in the laboratory desiccation experiment, with lack of green indicating surface bleaching. To quantify coloration, photographs were taken of microbialite surfaces alongside a standard color card (Pixel Perfect 24-Color Standard Calibration Chart). Color thresholding was done using the public-domain image processing software ImageJ (https://imagej.nih.gov), and the thresholded images were then used to quantify the green pixels in an image, with a full protocol described in S1 Appendix.

Microscopy

Core and recovery experiment samples collected for microscopy were weighed (~0.2 g dry mass), vortexed with 1 mL of 2% PBS-buffered formaldehyde to fix, and stored at 4°C. For both phase contrast and confocal laser scanning microscopy (CLSM), wet mount slides were prepared from ~60–100 μL of sample that had been vortexed to suspend solid material, using clear nail polish to seal coverslips to prevent water evaporation and salt precipitation. Brightfield and phase contrast microscopy photomicrographs were collected using an Accu-Scope EXC-500 with an Excelsis MPX-16C camera and CaptaVision software. For confocal laser scanning microscopy (CLSM), 4’,6-diamidino-2-phenylindole (DAPI, 1 μL of 1 μg⋅mL-1 stock in sterile, nuclease-free water; Thermo ScientificTM Cat. # 62248) and calcein (1 μL of 100 μg⋅mL-1 stock in sterile, nuclease-free water; Invitrogen Cat. # C481) fluorescent probes were added to the fixed sample prior to fixing cover slips, and fixed slides were stored in a dark box prior to analysis to prevent photobleaching. CLSM photomicrographs were imaged using an Olympus FV3000 with the following channels: DAPI (for staining DNA), ex = 405 nm, em = 430–480 nm; calcein (highlighting bound Ca²⁺ and Mg²⁺, which can be used to image polymers and biogenic carbonate), ex = 488 nm, em = 500–540 nm; and chlorophyll, ex = 514 nm, em = 550–600 nm (full settings used for imaging are in S1 Appendix). For both phase contrast microscopy and CLSM, ten random photomicrographs were collected at 200x magnification for relative color/fluorescence analysis, and interesting features were photographed at various magnification; protocols for color/fluorescence analysis are described in detail in S1 Appendix. The Euhalothece culture in Fig 2D was imaged using an Olympus BX51 microscope using a DP27 5 MP capture card mounted on a U-TV1XC adaptor; positive phase contrast and DIC were achieved by pairing an Olympus CX-PCD Phase Contrast condenser on Ph3 setting and Nomarsky DIC filter with a 100x objective.

Chlorophyll extractions

Chlorophyll extraction protocols used in 2021 and 2022 differed slightly in response to changes in equipment availability. In 2021, pigment samples were extracted by grinding solid samples to a powder using a sterile mortar and pestle, weighing the powder, then extracting overnight in 5 mL chilled, 100% acetone at 4°C in 9 mL glass test tubes. The next day, the glass tubes were centrifuged inside 15 mL polypropylene centrifuge tubes for 5 minutes at 3000 g to sediment solids. The supernatant, containing polar pigments, was scanned in a 3.5 mL quartz cuvette (Vernier) from 500–800 nm at 1 nm spectral resolution and 0.1 s⋅λ-1 averaging using a UV-VIS spectrophotometer (Cary 60, Agilent).

In 2022, solid samples were weighed, added to 4°C chilled 90% acetone in 1.5 mL opaque black polypropylene tubes (Argos Technologies), vortexed 5 seconds to mix, and polar pigments were extracted overnight at 4°C. Prior to measurement, samples were again vortexed, then centrifuged at 1000 g to sediment solids. Supernatant containing pigments was then scanned in a 400 μL quartz cuvette from 350–1000 nm at 1 nm spectral resolution and 0.1 s averaging using the UV-VIS spectrophotometer. Samples with dense concentrations of pigments (with peak absorbance readings > 2.0) were diluted with chilled 90% acetone prior to reading.

Chlorophyll a concentrations were then quantified using the equation described in Ritchie (2008), correcting for dilutions and normalizing to extracted sample mass. Values are reported as mg extractable chlorophyll a per gram of dry microbialite sample (mg/g).

DNA extraction, sequencing, and analysis

DNA extractions were performed using a DNeasy PowerSoil kit (Qiagen, Cat. # 12888–50), and extracted DNA was quantified fluorometrically (Qubit 2.0 Fluorometer, Life Technologies) with a high-sensitivity, double-stranded DNA analysis kit (ThermoFisher Scientific Cat # Q32851). Sample barcoding, amplification, and pooling for both 16S and 18S SSU rRNA genes was performed per Earth Microbiome Project protocols (16S: [39]; 18S: [40, 41]) by Wright Labs (Huntington, PA, USA). Amplicon sequencing was done, also by Wright Labs, using Illumina MiSeq v2 chemistry with paired-end 250 base pair reads for 16S, or 150 base pair reads for 18S. Sequences were deposited to the National Institute of Health’s Sequence Read Archive (SRA) under BioSample accession SAMN32545145–SAMN32545168. Sequence analysis was performed using the Qiime2 pipeline [42]: quality-controlled amplicon sequence variants (ASVs) were taxonomically classified against the SILVA 138 database [42], then filtered prior to diversity analysis: mitochondria, chloroplasts, or unassigned sequences were removed from the 16S dataset; and bacteria, and non-microbial or contaminant sequences (vertebrata, arthropoda, and Entomophthora muscae) were removed from the 18S dataset. Pipeline steps are described in detail in the S1 Appendix. Raw sequence data with linked sample metadata were archived with NCBI’s SRA (https://www.ncbi.nlm.nih.gov/sra) under BioSample accession SAMN32545145–SAMN32545168. Diversity analyses done with both filtered and unfiltered data are available in the OSF archive.

Depth, temperature, and salinity trends at the study site

Water depth, temperature, and salinity trends are summarized in Fig 4. Daily time series data are available in S2 Table, and raw time series data are available in the OSF archive.

Download:

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

Fig 4. Environmental trends at microbialite reef field sites, 2019–2022.

(A) Great Salt Lake elevation measurements (m-asl) at two USGS sites in the South Arm (left axis, lines), with manual depth measurements (right axis, circles) at Sites B and B3. (B) Air temperature (dashed lines) measured at KUTSYRAC22 (WS1) and KUTSYRAC27 (WS2), water temperature (solid lines) logged at the field site by three different data loggers, and manually measured temperatures (circles) at different water depths (water surface = top; middle depth = mid; sediment/water interface = bottom). (C) Daily mean downwelling (top) and sidewelling (side) irradiance logged at the field sites. (D) Daily accumulated precipitation measured at station KUTSYRAC22 (WS1) and KUTSYRAC27 (WS2). (E) Water salinity measured at different water depths (water surface = top; middle depth = mid; sediment/water interface = bottom); the October 2022 point was converted from density. In all plots, the shaded gray area indicates the field site move from Site B (left) to Site B3 (right, gray).

https://doi.org/10.1371/journal.pwat.0000100.g004

Changes in manual field site water depth measurements correlated closely with lake elevation changes when field sites remained hydrologically connected to the greater lake (Fig 4A). We were therefore able to estimate site elevations of 1277.26 m at Site B, 1277.0 at Sites RA and RB, and 1276.55 at Site B3 and RC. Using logged lake elevation data, we could then estimate site water depth over time during periods when sites were hydrologically open. Manual measurements were used when sites became hydrologically closed, which resulted in faster evaporation compared to the lake as a whole. Water depth at Site B declined by 0.95 m from July 2019 to July 2021, and declined an additional 0.10 m during the three weeks of microbialite desiccation monitoring in 2021, at which point the site became dry and the instrument pipe was moved to Site B3. Although some depth variability was recorded, there was no significant decline in water depth during the 2021 recovery experiment (Sites RA and RB), with a median water depth to sediment of ~ 20 cm at both sites (with samples suspended at ~ 10 cm). Water depth declined 0.11 m at Sites B3 and RC during the three weeks of microbialite desiccation monitoring in 2022, and an additional 0.27 m during the remaining duration of the 2022 recovery experiments such that by the end of the experiment at 100 days, sample bags were at the water surface. The roped-off quad area containing instrument pipe B3, recovery experiment RC, monitored microbialites M1–M3, and cored microbialites C1–C3 became hydrologically closed toward mid-August, 2022, cut off from the greater lake by exposed microbialites.

Water temperature was strongly seasonal and mimicked trends in air temperatures (Fig 4B), ranging from -3.0°C (measured January 3, 2022, when air temperature dropped as low as -7.9°C) to +37.8°C (measured July 31, 2022). Water temperatures were somewhat warmer during the 2022 microbialite monitoring season compared to the 2021 season (median temperature of 27.4°C and 31.2°C, respectively). Downwelling and sidewelling irradiance were also strongly seasonal (Fig 4C), and increased from 2021–2022, likely due to decreased water depth.

Total precipitation during the 2021 summer monitoring season was greater than during the 2022 monitoring season: 1.8 vs. 0.7 cm, respectively, with most of the rain during the 2021 monitoring season falling between the first and second monitoring weeks (Fig 4D). Total precipitation was also greater during the 2021 vs. 2022 recovery experiments: 9.6 vs. 5.0 cm, respectively, despite the substantially longer duration of the 2022 experiment.

Open-water salinity in the lakeward portion of the microbialite reef increased from 11.7% in July 2019 to 18.2% in July 2022, then to 19.5% in October 2022 (Fig 4E). Salinity proximal to the monitored microbialites was 16–17% during the 2021 monitoring season, 13–17% during the 2021 recovery experiments, 18–21% during the 2022 monitoring season, and 18–27% during the 2022 recovery experiments. Salinity in extensive, hydrodynamically restricted portions of the microbialite reef (including Sites B3 and RC) reached halite saturation in July 2022 (Fig 5), which became more widespread with lake elevation fall. Salinity measured using different techniques were closely correlated (Fig S1.5 in S1 Appendix); we therefore use refractometer measurements (% by weight) for the remainder of this paper except values ≥ 25%, which were converted from lab density measurements using the equation of state from Naftz et al. [43]. For more discussion on the salinity and density measurements, see S1 Appendix.

Download:

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

Fig 5. Field photographs showing halite saturation in closed microbialite reef areas in 2022.

(A) Halite lenses (arrows) forming on a closed portion of microbialite reef within 20 m of Site B3 in July, 2022. (B) Halite crystals coating a mesh bag from the recovery experiments at Site RC in September, 2022. (C) Photograph taken August 17, 2021 showing mm-scale halite crystals (white) covering the tops of microbialites following a windy (wave action) period. Abundant brine flies and fly pupal casings (dark) are also visible in the image.

https://doi.org/10.1371/journal.pwat.0000100.g005

Microbialite field monitoring and core sampling

Prolonged out-of-water (subaerial) exposure results in surficial bleaching, a process by which microbialite surfaces change color from dark green to white, and display signs of weathering (Figs 3 and 6, and S2 Fig). We observed initial stages of bleaching during both the 2021 and 2022 three-week intensive monitoring seasons (Fig 6, S1 and S2 Figs). Bleaching was quantifiable using image color analysis (Fig 6B). Quantitative results of surface monitoring are available in S3 Table. Over several-week timescales, bleaching was observed to be superficial, temporary, and influenced by water-soluble salts (especially halite); monitored microbialites re-greened following rain events (Fig 6 and S1 Fig), and samples collected were usually green just beneath the surface (< 1 cm). Indeed, microscopy revealed no obvious changes in surface samples, with Euhalothece clumps persisting in surface samples even in microbialites that appeared superficially bleached (S3 Fig).

Download:

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

Fig 6. Visible changes of microbialite surface over 100 days of subaerial exposure.

(A) Photographs of a single microbialite at Site B3 showing exposure, bleaching, and weathering during summer–autumn, 2022. Exposure times of each image align with dates notedin the graph below. (B) Surface coloration measurements of three monitored microbialites (green lines, left axis); points represent averages of three measurements in defined rectangular areas of interest (AOI) from photographs taken on each monitoring date, with standard deviations represented as error bars. Also shown are daily precipitation amounts (blue bars, right axis). Note the bright discoloration (salt precipitation) at day 14, followed by re-greening on day 21 after a light rain event, then bleaching and weathering after heavier rain on day 72, followed by continued bleaching and weathering to day 100. Tic marks are days, with labels at each sampling date.

https://doi.org/10.1371/journal.pwat.0000100.g006

Beneath the surface, microbialite communities persisted following subaerial exposure, with no statistically significant (ANOVA p ≤ 0.1) changes in extractable chlorophyll through time at any core depth (Fig 7A). In contrast, DAPI channel fluorescence decreased over the initial 21-day monitoring period at all core depths (Fig 7B), and calcein channel fluorescence decreased somewhat in the bottom core horizon (Fig 7C). A summary of microbialite core measurements can be found in S1 File, with the raw dataset available in the OSF archive.

Download:

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

Fig 7. Microbialite core series results.

Measurements from core top (T), middle (M), and bottom (B) horizons; colors represent different timepoints: 0, 7, 14, 21, 72, and 100 days. (A) Extractable chlorophyll a. (B) DAPI, and (C) calcein channel fluorescence in CLSM random photomicrographs as percent of total pixels. Due to resource constraints, CLSM fluorescence measurements were not done on samples after day 21. For each measurement, horizons identified as having significant change through time in ANOVA (p ≤ 0.1) are indicated with *.

https://doi.org/10.1371/journal.pwat.0000100.g007

Lab desiccation experiments

Lab desiccation experiments accelerated, but otherwise mimicked field observations: surficial bleaching occurred rapidly, due at least in part to the accumulation of surface salts, and was immediately reversed upon rinsing with distilled water (Fig 8). Bleaching did, however, have an impact on microbialite weathering: during each rinse, surface material (which itself re-greened upon exposure to distilled water) sloughed off, and the microbialite lost dry mass. Material that was previously resistant to weathering was easily removed following each additional period of bleaching. Surface color measurements and masses for this experiment are available in S4 Table.

Download:

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

Fig 8. Lab desiccation experiment results.

(A) Photograph series showing surficial changes to the microbialite, with days of exposure shown at bottom center of each image. The images for day 6 show the microbialite before and after (denoted with *) a rinse event. (B) Surface green area analysis results (green line) with points representing the average measured fraction of green pixels in three rectangular areas of interest (AOI) for each date, and standard deviations shown as error bars, overlaid on measured microbialite mass (dashed line). Disturbance events are indicated along the top: the timing of rinse events is denoted with *, and a measurement following the removal of sandy surface material on day 49 is denoted with +. Note the loss of color over time, significant regreening following all rinse events, and weathering of surface material during disruption events.

https://doi.org/10.1371/journal.pwat.0000100.g008

Recovery experiments

Results in 2021.

In 2021, we observed exponential increases in extractable chlorophyll and DNA over the 40-days the experiment ran (Fig 9A and 9B; exponential best fit line r² values > 0.96 and 0.83, respectively), despite salinity levels exceeding 17% and falling water temperatures (Fig 4). In both cases, the growth rate was greater at the channel site (RA) vs. the reef interior site (RB); for chlorophyll, the doubling time was 9 days at RA, and 16 days at RB, which is on the slow end of rates reported for periphyton/biofilm recolonization of substrates in both freshwater and marine systems (e.g., [44, 45]). Extrapolating the exponential best fit lines to levels of chlorophyll a and DNA comparable to that of surrounding microbialites with healthy periphyton suggested a possible time to recovery in these measurements of 60–80 and 100–130 days for RA and RB, respectively (Fig 9C and 9D).

Download:

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

Fig 9. Recovery experiment results.

The top row shows measured values of extractable chlorophyll a (A), and DNA (B) with exponential best-fit lines shown along with r² values (in parentheses next to legend entries). Exponential recovery projections (dashed lines) extrapolated from exponential best fit lines for chlorophyll a and DNA are shown in (C) and (D), respectively (log scale). The lower edge of the green shaded region in C & D indicates the mean of concentrations measured from healthy microbialite periphyton. RA and RB are from the experiments conducted in 2021, in a flowing channel vs. microbialite interior, respectively. RC is from the experiment conducted in 2022 in a reef interior site where salinity reached halite saturation by day 72.

https://doi.org/10.1371/journal.pwat.0000100.g009

Community analysis from DNA sequencing.

DNA sequencing results from the 2021 experiment indicated a consistent shift in community composition over time driven largely by a relative increase in Proteobacteria and Bacterioidota in bacterial 16S sequences (Fig 10A), and an increase in diatoms in eukaryotic 18S sequences (Fig 10B).

Download:

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

Fig 10. Recovery experiment DNA sequencing results.

Results of 16S and 18S amplicon sequencing for the 2021 experiment at sites RA and RB, showing changes in microbial community over incubation time (numbers indicate incubation days) compared to lakewater and healthy microbialite periphyton (HM) samples. (A) Relative abundance of 16S sequencing ASVs belonging to different taxonomic groupings within the Bacteria and Archaea. *Most ASVs classed as Dactylococcopsis had closest sequence similarity to Euhalothece. (B) Relative abundance of 18S sequencing ASVs belonging to different taxonomic groupings within the Eukarya after contaminant, vertebrata, and arthropoda sequences were removed. (C) 16S and (D) 18S principal coordinate analysis (PCoA) plots of unweighted UniFrac sample distances showing the grouping of like samples (green = recovery experiment samples, blue = lakewater samples, red = healthy microbialite periphyton samples, gray = desiccated t₀ control samples) and the evolution of the recovery experiment communities over time (green arrows) away from the healthy periphyton community. The percentages of sample variance explained by each of the three displayed axes are shown next to the respective axis. For interactive versions of these plots, the reader is directed to the qiime2 visualization files in the OSF archive, which can be viewed using the QIIME 2 View web interface (https://view.qiime2.org.

https://doi.org/10.1371/journal.pwat.0000100.g010

The community composition of control samples from a microbialite with a healthy periphyton was largely consistent with compositions found at other times and locations [12, 14], and distinctly different from the composition found on the t₀ control microbialite used to seed the experiments (Fig 10). The difference between healthy periphyton and t₀ control was especially apparent in the 18S results, where the healthy periphyton control samples were dominated by Artemia sequences, which were absent from the bleached t₀ control microbialite and early recovery experiment samples, as well as from water compositions at different points during the experiment. Principal components ordinate analysis using unweighted UniFrac sample distances indicated that the shift in recovery community was away from the healthy periphyton community, and more similar to water than healthy periphyton (Fig 10C and 10D). Amplicon sequence variant (ASV) tables of sequence abundance in each sample can be found in the OSF archive.

Microbialite bleaching is initially superficial, with some endoliths persisting over months of subaerial exposure

Although decreases in surface coloration were observed over a several week period (Fig 6), with a white zone of bleaching extending gradually downward on the microbialites (Fig 3 and S2 Fig), pigment extracts and microscopy from core samples revealed that bleaching was initially superficial, with substantial pigmented cellular material, including large Euhalothece clumps, persisting even in the upper ~1 cm samples (Fig 7 and S3 Fig).

Both our lab experiments and field observations indicate that some apparent bleaching, especially during early weeks of exposure, is attributable to the precipitation of reflective evaporite minerals on microbialite surfaces during subaerial exposure and evaporation. This is particularly obvious during periods of higher wave action, which can lead to the growth of mm-scale halite crystals on microbialite surfaces (Fig 5C). These minerals can dissolve during rain events, leading to a significant re-greening of microbialites that previously appeared bleached (Figs 5 and 8, and S2 Fig).

Maintenance of chlorophyll a cm within the sampled microbialites (Fig 7A) may indicate an endolithic survival strategy during subaerial exposure. Survival of ordinarily surficial cyanobacteria and affiliated communities in the subsurface of permeable rock during periods of subaerial exposure has been documented in other systems, including desiccated lakes [57], and extracellular polymeric substances (EPS) can aid in desiccation and freezing survival in endolithic communities [58]. This is the first time to our knowledge that this growth habit has been documented in the Great Salt Lake microbialites; indeed, we are unaware of other literature documenting the impacts of subaerial exposure on microbialites over time periods longer than tidal cycles (e.g., [59]).

However, the fraction of DAPI-fluorescent cells decreased significantly over time at all sampled depths in the microbialite (up to the 4 cm depth monitored; Fig 7B), indicating that endolithic survival was not universal. Although it is not possible to tell from microscopy alone which taxa persisted or were lost, it appears from the chlorophyll results to be primarily non-photosynthetic members of the community, e.g., heterotrophs, sulfur cyclers, protists, fungi, and other diverse organisms, some of which may contribute to EPS production, overall biofilm structure, and nutrient cycling.

Microbialite bleaching enhances weathering

In longer-term observations of individual microbialites, substantial weathering was observed following periods of bleaching (Figs 3 and 6, S1 and S2 Figs). Upper portions of the microbialites lose their robust, gooey/blubber-like quality over periods of prolonged exposure and become loose and sandy, likely the direct consequence of declines in the surface microbial community and its affiliated EPS. Observable decreases in calcein fluorescence (Fig 7C), which binds to extracellular calcium and can therefore highlight degrading EPS, is consistent with this interpretation. The sandy carbonate material left on the desiccated surface is then easily weathered, as observed in the field and as illustrated in the observed mass loss following rinse events during the lab desiccation experiments (Fig 8). Enhanced weathering of the microbialites may be one of the longer-lasting consequences of microbialite exposure, as microbialite growth rates are not yet known in Great Salt Lake, but assumed to be quite slow [21].

Partial recovery of bleached microbialites occurs even at high salinity levels

The results of our recovery experiments in summer/autumn 2021 were broadly encouraging, showing an exponential rate of recovery when desiccated samples were re-submerged in lakewater at a salinity between 13–17%, as measured by extractable chlorophyll a and DNA (Fig 9A and 9B). If recovery continued along the same exponential trajectory observed over the initial 40 days of the experiment, a full recovery to “healthy” values could occur within 130 days of being re-submerged, i.e., a single summer growth season (Fig 9C and 9D). Also encouraging was the appearance of Euhalothece sequences on the recovery samples after just 10 days incubation, as well as sequences of eukaryotic phototrophs (Fig 10A and 10B). The combined results imply that a recovery of the ecosystem-critical microbialite primary producers is possible if microbialites—even badly bleached microbialites—are re-submerged in healthy lakewater. However, as discussed in the next section, while Euhalothece (and, thus, chlorophyll and presumably primary productivity) return to the microbialites, we did not see evidence of the return of the full, healthy microbial community in our analysis of the DNA sequencing data; thus, the time and conditions to true recovery (full-community and robust microbial mat) remains uncertain.

When we repeated the experiment in the higher-salinity water of summer–autumn, 2022 (18–27%), recovery was markedly slower (with a calculated doubling time from chlorophyll recovery of 31 days vs. 9 and 16 days at the two 2021 sites, although the exponential best fit was poor in 2022) but did eventually appear to occur, with extracted chlorophyll a values reaching those seen in 2021 by 72 days incubation, and increasing further in the following month (Fig 9A). Euhalothece clumps were visible in microscopic analyses in samples collected after 2–3 months incubation (S3 Fig), supplying additional evidence of recovery. This is remarkable, considering the salinity was at or near halite saturation when this apparent recovery was occurring. It is possible that the apparent recovery was simply a gradual accumulation of cells from the surrounding water (supported by the faster recovery rates of microbialites exposed to more lakewater flow; Fig 9A and 9B), which could potentially be preserved for some time by encrusting halite [60]. Additional work is required to determine whether the observed Euhalothece are actually photosynthetically active, or simply preserved.

Microbialite recolonization likely depends on the health of the lakewater community

Recovery of the microbialite communities in our 2021 experiment (no DNA sequencing results are available for the 2022 experiments) appears to be based on recolonization from the surrounding water, rather than a ‘resurrection’ of dormant endolithic communities. Community distance analysis indicates that the microbial community present in recovery samples were more similar to lakewater than to either the desiccated microbialite t₀ community or the community present on microbialites with healthy periphyton (Fig 10C and 10D). The recovery sample communities that are most similar to lakewater were unsurprisingly those incubated for the shortest periods, and evolved away from lakewater over time, suggesting an initial seeding by lakewater. It is also notable that the 2021 recovery experiment site located in a channel away from the greater reef and exposed to more water flow (Site RA) experienced recovery rates that were nearly twice as fast as was seen at the reef-interior site (Site RB) despite similar salinity levels and hydrologic connection.

The recovery sample community also evolved over time to become less similar to that seen in microbialites with healthy periphyton, despite the persistence of Euhalothece. Most notable is the relative dearth of Desulfobacterota in recovery experiment samples and high abundance of Flavobacteriales, Rhodobacterales, and Gammaproteobacteria, especially Marinobacteraceae and Oceanospirillales (Fig 10A and 10B). The low abundance of Desulfobacteria can be explained by the high salinity levels present during the recovery experiments; even halotolerant strains are not known to tolerate salinity levels above 13% [61, 62]. It could, however, also be a consequence of an immature microbial mat lacking the anoxic zones required to support sulfur-reducing metabolisms. What is unclear is whether—given enough time and development of a robust biofilm, then microbial mat—the community would begin to shift to more closely resemble that seen in microbialites with a healthy periphyton.

Primary seeding from lakewater indicates that microbialite recolonization is dependent on the health of the lakewater microbial ecosystem and the presence of viable Euhalothece and other organisms in lakewater. While microbialite communities may be able to survive, or at least be preserved, through periods of subaerial exposure and at high salinity, this is not true of lakewater communities. In the north arm of Great Salt Lake, where salinity values routinely exceed 24%, Euhalothece sequences are absent [29], and the community composition is markedly different than that in the lake’s south arm. Thus, a high-salinity lake would not be able to re-seed healthy microbialite periphyton communities. It appears that the lake’s south arm is the ultimate reservoir for the organisms and metabolisms that support the broader Great Salt Lake ecosystem, including the microbialite microbial communities.