The paragenesis of carbonate pseudomorphic textures in the rock record that are inferred to represent replaced metastable ikaite (CaCO3·6H2O), which forms at frigid temperatures, is uncertain. Petrographic analysis of Mono Lake (California, USA) Pleistocene tufas allowed recognition of a distinctive calcite microtexture, termed guttulatic calcite, that forms during carbonate dehydration and is diagnostic for precursor ikaite. The texture is characterized by pseudo-hexagonal or spherical low-Mg cores, which likely formed initially as vaterite, with an ellipsoidal overgrowth, and a secondary high-Mg sparry or micritic cement. Observations of Mono Lake ikaite pseudomorphs, combined with a review of more ancient examples, indicate that guttulatic texture records carbonate dehydration of precursor ikaite and can be used to infer frigid paleotemperatures.

Accurate reconstruction of past oscillations in seawater composition requires correctly identifying primary mineralogy and diagenetic overprints in the marine carbonate record based on carbonate textures (e.g., Grotzinger and Reed, 1983; Stanley and Hardie, 1998). Ikaite (CaCO3·6H2O) is a hydrated carbonate that was initially discovered in Ikka Fjord, Greenland, (Buchardt et al., 2001) and subsequently reported in Antarctic and Arctic marine environments (Suess et al., 1982) and alkaline lakes of Patagonia and western North America including Mono Lake, California (Bischoff et al., 1993b; Council and Bennett, 1993). Although experimental results show that high concentrations of Mg2+ and PO43– or high pH may enable ikaite nucleation at up to 12–35 °C (Purgstaller et al., 2017; Stockmann et al., 2018; Tollefsen et al., 2020), ikaite is metastable at these temperatures due to an increase in solubility with temperature (Bischoff et al., 1993a). Ikaite crystals have been observed to form naturally only at <9 °C (Huggett et al., 2005; Field et al., 2017) even in highly alkaline and P-rich water (e.g., Mono Lake, pH ~10; Bischoff et al., 1993b; Council and Bennett, 1993). When heated, ikaite loses its structural waters and transforms into a variety of Ca-carbonate polymorphs, including intermediate products of monohydrocalcite, vaterite, and—ultimately—calcite (Tang et al., 2009; Sánchez-Pastor et al., 2016; Purgstaller et al., 2017; Stockmann et al., 2018; Tollefsen et al., 2020).

Putative ikaite pseudomorphs of predominantly low-Mg calcite, known as “glendonite,” are argued to be a proxy for past cold depositional water environments (De Lurio and Frakes, 1999; Swainson and Hammond, 2001; Selleck et al., 2007). For example, Neoproterozoic glendonite has been used as evidence of near-freezing marine temperatures during, between, and after global glaciations (James et al., 2005; Dempster and Jess, 2015; Wang et al., 2020). Glendonite texture has no formal definition, although it is typically inferred to be an ikaite pseudomorph that is sometimes marked by centimeter-scale stellate crystals (e.g., Selleck et al., 2007). Some of these may be pseudomorphs after evaporite minerals that also form stellate or bipyramidal crystals (e.g., Jafarzadeh and Burnham, 1992). Modern ikaite forms a broad range of morphologies including stellate, bi-pyramidal, prismatic, radial bladed, or “thinolite” crystals in addition to massive microcrystalline textures in porous tufas, and speleothem crusts. Many of these macrotextures are undifferentiable from anhydrous carbonate tufa (see later sections herein). Therefore, a refined understanding of pseudomorphic microtextural paragenesis from ikaite dehydration is required to confidently identify former ikaite and infer primary cold-water temperatures. In this study, we show that calcite pseudomorphs of former ikaite display a characteristic microtexture—here termed “guttulatic”—which makes it a stronger proxy for recognizing ikaite dehydration within the carbonate record that relies neither on preservation of primary mineralogy nor on precursor ikaite having formed centimeter-scale stellate crystals.

Water column carbonate precipitation is negligible in hyperalkaline Mono Lake due to depletion of Ca (~4 ppm) (Fig. 1; Dunn, 1953). However, carbonates precipitate rapidly as either shoreline deposits or tufa towers where Ca-rich streams or groundwater mix with lake waters (Dunn, 1953; Bischoff et al., 1993b). Tufas are classified based on macroscopic textures: (1) thinolite tufa comprises radiating, centimeter-scale crystals regarded as ikaite pseudomorphs (Shearman et al., 1989); (2) dendritic tufa forms branching structures that have colloform or shrubby appearance; and (3) lithoid tufa has a massive appearance (Dunn, 1953). Notably, fine-grained ikaite forms along the shoreline of Mono Lake during winter months and subsequently decomposes to anhydrous Ca-carbonate during seasonal warming (Bischoff et al., 1993b; Council and Bennett, 1993).

The three tufa textures characterize shoreline and tufa tower deposits that formed during high-stands of Pleistocene Lake Russell above present-day lake level (Benson et al., 1990). We studied thinolite tufa, dendritic tufa, and fabrics of thinolite-dendritic mixtures (Fig. 1) from both Pleistocene high-stand shorelines (n = 19) and towers (n = 5; Fig. 1). The carbonates described were confirmed as calcite via X-ray powder diffraction. In addition, the carbonate samples were prepared as polished thin sections and studied by optical microscopy and field emission scanning electron microscopy (SEM).

Thinolite tufas from Mono Lake consist of elongate crystals of 1–20 cm in length and several millimeters to centimeters in width. Each centimeter-scale crystal is pseudomorphed by a mosaic of beige or brown calcite crystals (Fig. 1F). These calcite pseudomorphs have a ~10–100 μm pseudo-hexagonal or spherical core with a spherical-, ellipsoidal-, or anhedral-shaped syntaxial calcite overgrowth (Figs. 13). Rhombohedral cores are rarely present. The syntaxial overgrowth records zones that express hexagonal crystal morphology close to the core and progressively evolve to ellipsoidal morphology further away from the core and often with accumulation of fluid and solid inclusions at the growth zone boundaries (Fig. 3). No chemical difference was identified between cores and overgrowth through SEM backscattered electron (BSE) imaging and energy dispersive spectroscopy (EDS) analyses (Fig. 2; see the Supplemental Material1 for tables of SEM-EDS data). Core and overgrowth calcite generally have low Mg mole% of ≤7.6% with an average of 3.5 ± 4.0% (n = 46) based on semiquantitative calculations from EDS spectra (see the Supplemental Material). In between overgrowth calcite, thinolite crystals are filled with high-Mg micritic or sparry calcite cements that have a Mg mole% of 10.6% to 50.0%, with an average of 29.2 ± 9.6% (n = 27; Figs. 2 and 3).

The edges of the centimeter-scale thinolite crystals are coated with a micron-thick, dark, texturally heterogenous, optically opaque, microcrystalline or possibly amorphous layer that is enriched in phosphorus (Fig. 2). This yet undetermined P-bearing phase appears to be partly composed of silicates with P/Si of 1.0 ± 0.5 (n = 28). The P-bearing phase contains significant Ca with Ca/Si ratios of 2.5 ± 1.2 (n = 28) and Ca/P ratios of 2.7 ± 1.4 (n = 28) but only minor Mg. Furthermore, these intercrystalline spaces may be filled with micritic low-Mg calcite and/or microcrystalline Mg-silicates (Fig. 2). The Mg-silicates measured with SEM-EDS are consistent with Mg-clays (see the Supplemental Material).

Paragenetic Sequences

Ancient and sub-recent ikaite pseudomorphic textures identified in published photomicrographs show a microtexture similar to that observed in Mono Lake thinolite, including cores, overgrowth, and cement in a calcite mosaic (Fig. 3). Petrographic analysis revealed that cores are predominantly pseudo-hexagonal, and overgrowths have pseudo-hexagonal to spherical/ellipsoidal zones (Fig. 3). We propose a detailed paragenetic model for this uncommon but distinctive microtexture through comparison to recent experimental results (Fig. 4B). We provide two likely mechanistic transformation pathways that support models proposed in previous studies (e.g., Vickers et al., 2018).

Ikaite heating (at 5°C to 35°C) experiments have shown that dehydration and transformation of ikaite into spherical and rhombohedral crystals of vaterite and calcite, respectively, occurs within minutes (Tang et al., 2009; Sánchez-Pastor et al., 2016; Purgstaller et al., 2017; Tollefsen et al., 2020). Vaterite is a highly unstable, hexagonal, anhydrous Ca-carbonate polymorph that forms nano- to micron-scale spherical and hexagonal crystals in experiments and natural cold spring environments (Tang et al., 2009; Konopacka-Łyskawa, 2019). Vaterite transforms to calcite within minutes to days (Tang et al., 2009; Sánchez-Pastor et al., 2016; Purgstaller et al., 2017; Konopacka-Lyskawa et al., 2019). The unique pseudo-hexagonal and spherical cores within ikaite pseudomorphs are most consistent with former vaterite pseudomorphs that are subsequently replaced by calcite (Fig. 4B). Rare rhombohedral cores would represent transformation to calcite directly from ikaite. The syntaxial overgrowth could form originally as vaterite growing through ostwald ripening of amorphous calcium carbonate into spherical/ellipsoidal structures (Konopacka-Łyskawa, 2019) with subsequent single-crystal replacement into calcite. Alternatively, the syntaxial overgrowth may be later calcite cement growing in stages. The former scenario is consistent with low-Mg calcite composition of cores and overgrowth in this study as both would be sourced from the low-Mg ikaite precursor. In other studies, overgrowths and cores differ in chemical composition, suggesting overgrowths could be products of later cementation (Vickers et al., 2018). Mg- and Fe-rich sparry calcite cements were observed more broadly within Mono Lake tufa facies and in several previous studies. These sparry cements have a later diagenetic origin that likely plays an important role in preserving the initially porous and friable pseudomorph crystal (Huggett et al., 2005; Vickers et al., 2018).

The association between ikaite pseudomorphs and microcrystalline Mg-silicates in Mono Lake tufas has not been previously observed, while the P-bearing phases have only recently been studied (Huggett et al., 2005; Ingalls et al., 2020). These phases postdate ikaite crystal growth as they coat the sides of the original ikaite crystals (Fig. 1F) and are not regarded as significant in the context of early ikaite transformation to calcite. Mg-silicates commonly are associated with carbonates that rapidly precipitate from high alkalinity water (e.g., Gomes et al., 2020), while the P-bearing phase is interesting due to phosphate's proposed role in promoting ikaite formation (Bischoff et al., 1993a).

Guttulatic Calcite: A Distinctive Microtexture

We propose the term “guttulatic calcite” to capture the characteristic pseudomorphic ~10–100 μm pseudo-hexagonal or spherical calcite core, hexagonal and ellipsoidal zoned syntaxial overgrowth, and cement paragenetic sequence associated with neomorphic inversion of ikaite to vaterite and subsequently to calcite (Figs. 2 and 3). The term guttulatic is adapted after guttula, the Latin word for “very small droplet,” and is an appropriate descriptor for the droplet-like shape of calcite cores with overgrowth. A review of petrographic images or descriptions of microtextures associated with reported ikaite pseudomorphs shows that many examples do contain guttulatic calcite, which was originally described as a “microgranular mosaic,” “oolitic and sparry,” or “type 1–3 calcite” (Fig. 3). We establish a correlation and distinction between guttulatic calcite and glendonite fabrics for many reported examples. We propose that guttulatic calcite is diagnostic specifically for carbonate dehydration and that its recognition helps to confidently establish former frigid conditions of primary mineral precipitation.

Significantly, several proposed Neoproterozic ikaite pseudomorphs do not exhibit guttulatic microtexture (e.g., James et al., 2005; Dempster and Jess, 2015; Wang et al., 2020). Extensive recrystallization may have erased the guttulatic texture given that guttulatic calcite is an early diagenetic feature that may not always be preserved. Other studies that report glendonite stellate crystal macrotextures and infer frigid paleotemperatures could also be substantiated through the petrographic search for guttulatic texture (e.g., De Lurio and Frakes, 1999; Price and Nunn, 2010). The guttulatic microtexture is unique in the context of carbonate mineral paragenesis and can be used to identify precursor ikaite and vaterite precursor in the many cases where precursor minerals did not form stellate or bipyramidal crystals (Fig. 4A).

Thinolite tufas at Mono Lake preserve a characteristic microtexture—guttulatic calcite—that is composed of ~10–100 μm-sized pseudo-hexagonal or spherical cores, ellipsoidal zoned syntaxial overgrowth, and a secondary high-Mg sparry or micritic cement. These crystal pseudomorphs were later coated with Mg-silicates associated with alkaline lake conditions and a P-bearing amorphous phase; this later step may be particular to alkaline, lacustrine carbonates such as those of Mono Lake. Guttulatic microtexture is recognized in many, but not all, rocks described with ikaite pseudomorph macrotextures. We propose that guttulatic calcite forms during dehydration diagenesis through warming. Within minutes to months, the low-Mg calcite hexagonal and rounded cores are formed through transformation of the precursor hydrated carbonate into metastable vaterite that is subsequently transformed to calcite. Ellipsoidal calcite overgrowths form as part of initial vaterite recrystallization or as a later cement. Subsequently, high-Mg calcite cement precipitates in pores from pore waters with different chemistry. Recognition of guttulatic calcite provides evidence that hydrated carbonate precursors underwent dehydration diagenesis induced by heating after forming at frigid conditions regardless of whether the precursor formed stellate macrocrystals.

This work was initiated under the Simons Collaboration for the Origin of Life (Simons Foundation, New York, USA) with support to J.P. Grotzinger. E.L. Scheller was supported by NASA Earth and Space Science Fellowship #80NSSC18K1255. Original authors are credited for figure panels in Figures 3E3I and Figure 4A. We thank Nathan Stein for his contributions in the field. We humbly acknowledge the Kutzadika’a, the communities of Mono Basin, and other indigenous peoples whose identities may have been lost to colonialism, on whose land this work relies.

1Supplemental Material. Full SEM-EDS datasets, further descriptions of the methodology, and full references for the figure panels. Please visit https://doi.org/10.1130/GEOL.S.16589939 to access the supplemental material, and contact editing@geosociety.org with any questions.
Gold Open Access: This paper is published under the terms of the CC-BY license.