The ICEland-1 curated database is compiled from existing peer-reviewed publications, book chapters, PhD and MSc theses, radiocarbon date lists previously compiled by the Institute of Arctic and Alpine Research (University of Colorado Boulder), and our unpublished data. Dates and metadata (see Table 1) obtained from compilations are cited as well as the original source. In many cases, geographical coordinates were not provided in the original publication. Where maps were provided with site locations, we georeferenced these figures in Google Earth to derive estimated coordinates. All site locations have been reformatted and included as decimal degrees (° N, ° E, WGS84). Where neither coordinates nor maps locations were given, we exclude the data from our database. Each terrestrial date's “Region” follows the eight currently recognized regions of Iceland: Northeast, Northwest, Westfjords, West, Capital Region, Southern Peninsula, South, and East.

The ICEland-1 curated database covers the marine and terrestrial realms of Iceland and includes geochronological data derived from 14C (conventional and AMS), select Early Holocene marker tephra layers of known age, and TCN exposure ages (36Cl and 3He). In this first version of ICEland-1, we elected to omit most tephra layers and all paleomagnetic secular variation correlation tie points (e.g., Geirsdóttir et al., 2013; Ólafsdóttir et al., 2013). As both methods rely on user correlation, there is an inherent degree of subjectivity in assigning an age and our primary aim is to focus on the most objective, radiogenic toolsets. However, we do include two marker tephra layers because they feature diagnostic stratigraphical and compositional attributes and are relevant for delineating Early Holocene ice sheet limits: the Askja S tephra layer (10 887–10 773 calyrBP, Bronk Ramsey et al., 2015) and the G10ka Series (10 400–9900 calyrBP, Óladóttir et al., 2020). As the G10ka Series was the product of up to 13 eruptions from the Grímsvötn volcanic system that were dispersed in varying directions (Harning et al., 2025b), there is some regional variability in limiting ages and are noted as such in the database where relevant (e.g., Harning et al., 2018a, 2019b, 2025b; this study). While the ∼12kaBP Vedde Ash has previously been included in Icelandic deglacial stratigraphies, particularly in the marine realm (e.g., Geirsdóttir et al., 2002, 2022; Eiríksson et al., 2004), we omit this marker as recent evidence from Iceland and abroad (Lane et al., 2012; Harning et al., 2024) demonstrate that multiple tephra layers with indistinguishable geochemical composition were produced in the Late Glacial and Early Holocene, leading to high temporal uncertainty without supporting chronological control. As the G10ka Series and Askja S tephra layers are both dated using 14C, they are included alongside 14C analyses in ICEland-1.

To ensure that our curated database follows FAIR (Findable, Accessible, Interoperable, and Reusable) data management principles (Wilkinson et al., 2016), we followed data synthesis approaches and standards first established for other ice sheet geochronologies (DATED-1, Hughes et al., 2016). This includes both the metadata and quality control criteria (Tables 1 and 2), with some minor modifications to suit Iceland's specific datasets (e.g., tephra and TCN exposure ages from basalt), so that future efforts can integrate ICEland-1 with other ice sheet databases and improve interoperability. Our database is hosted on Zenodo and Ghub (Tulenko et al., 2025), ensuring that our database is findable and accessible by the research community and public. Finally, to ensure reusability, we provide all raw data for 14C and TCN datasets so that ages can be recalibrated pending future developments in calibrations and calculators. The ICEland-1 has a census date of 1 January 2026, meaning that calibrated ages reflect the state-of-the-art at the time of publication as described in the following sections. Dates and associated metadata published after 1 January 2026, will be included in subsequent iterations to be formally described and submitted to a peer-reviewed data journal (e.g., Earth System Science Data).

In addition to published 14C ages, we report 32 new 14C dates from marine sediment and 17 new 14C from lake sediment and peat, which include 10 new 14C dates from Trjáviðurlækur constraining max/min ages of the Hekla 5 tephra and a minimum regional age for the G10ka Series tephra in south Iceland. New marine 14C ages were measured on shells and foraminifera and new terrestrial 14C ages were measured on plant macrofossils and humic acids. All samples were extracted (for humic acids, Abbott and Stafford, 1996) and graphitized at the Laboratory for AMS Radiocarbon Preparation and Research (NSRL), University of Colorado Boulder, and measured by AMS at the W.M. Keck Carbon Cycle AMS Laboratory, University of California Irvine. We recalibrated all 14C ages, including those previously published, using the latest terrestrial (IntCal20, Reimer et al., 2020) and marine calibration curves (Marine20, Heaton et al., 2020), with no corrections for variable reservoir age (ΔR), in OxCal Version 4.4 (Bronk Ramsey, 2009) and report uncertainty to 1σ (68 %). Compared to conventional 14C dates, AMS 14C dates are more reliable due to requiring less sample material and greater precision and accuracy within 14C calibration windows (Bronk Ramsey et al., 2004; Heaton et al., 2020; Reimer et al., 2020).

Marine20 differs from prior marine 14C calibrations in that it attempts to take large scale ΔR effects into account (Heaton et al., 2022) and therefore does not require local corrections for ΔR. However, Marine20 was also not designed for subpolar regions, such as Iceland, where sea ice extent, ocean upwelling and air–sea gas exchange may cause larger changes in ΔR through time. Evidence from an absolutely dated, annually resolved marine shell chronology on the North Iceland Shelf show centennial-scale ΔR variability over the last 1300 years (Wanamaker et al., 2012), similar to 14C-tephra layer comparisons from marine sediments over the last 4500 years (Eiríksson et al., 2004). This centennial-scale ΔR variability is consistent with Late Holocene water mass variability inferred from paleoceanographic reconstructions on the North Iceland Shelf (Kristjánsdóttir et al., 2017; Harning et al., 2021). Recent PSV-dated sediment records from Iceland's shelf demonstrate that when using Marine20 for the Holocene, a ΔR correction of 0 yields age estimates within uncertainty of Holocene tephra layers and annual marine shell chronologies (Reilly et al., 2023). However, as previously noted, some tephra layers are not well-constrained (e.g., Vedde Ash), leaving large uncertainties in past variation of ΔR on Iceland's shelf, particularly during the Late Glacial. As a result, 14C-dated marine sediments from glacial periods are likely too old, possibly by millennia (Heaton et al., 2023), and therefore, reflect maximum ages. 14C dates from sediment-feeding molluscs (e.g., Yoldia) may also be stratigraphically too old due to the update of old carbon in the sediment compared to suspension feeders (e.g., Mya truncata, England et al., 2013).

Bulk and humic acid 14C dates from the terrestrial realm may also be older than their stratigraphic position due to the incorporation of pre-existing organic carbon on the landscape, particularly during the Late Holocene (e.g., Geirsdóttir et al., 2009b). One way to test for potential offsets is through the comparison of bulk 14C dates with tephra layers of known age, which for some settings show basal bulk 14C dates consistent with the tephra ages, suggesting that potential offsets may be less of an issue shortly after deglaciation due to removal of pre-existing organic carbon by the IIS (e.g., Harning et al., 2024). However, in some instances where carbon content is low, basal bulk 14C deglaciation can yield substantially older ages than the expected timing of deglaciation (Brader et al., 2015), suggesting that not all carbon was removed from the landscape during the last glacial period. Hence, bulk 14C ages are generally less reliable and require careful consideration alongside other stratigraphic and dating constraints.

There is current debate and preference over which calculators yield the most accurate TCN exposure ages from 36Cl and 3He. For 36Cl, most dates included in ICEland-1 have used the development version of CREp (Martin et al., 2017; https://crep-dev.otelo.univ-lorraine.fr/#/init, last access: 1 February 2026), although there is generally no a priori reason stated as to why a given calculator was chosen over the others. Therefore, to be the most objective and maintain consistency across the database, we recalculated all ages where possible using the CRONUS-Earth online calculator v.3 (Balco et al., 2008), CRONUScalc (Marrero et al., 2016), and the development version of CREp (Martin et al., 2017; https://crep-dev.otelo.univ-lorraine.fr/#/init, last access: 1 February 2026). All metadata and geochemical constraints from the literature are provided in ICEland-1 required to recalculate ages pending future advances in the field. For most dates, we made some basic assumptions about samples, such as year of collection, rock density, formation age, and major and trace element concentrations and uncertainties that were not provided in the original publication. For example, we use the publication date for the year of collection, a rock formation age of 10±1Ma, and density of 2.7 gcm-3 following assumptions in the literature. Major and trace element concentrations and uncertainties not reported were assumed to be 0. To maintain consistency with other recent ice sheet chronology curated databases (e.g., DATED-1, Hughes et al., 2016), we do not make corrections for post exposure uplift, erosion, or snow cover. For 3He exposure ages measured on tuyas, or tabletop mountains, we report rebound age adjustment correction factors (range 0.6 %–3.0 %, Licciardi et al., 2007) under the “Rebound_corr_perc” column, that if applied, will render a slightly older apparent exposure age. All recalculated age errors are reported as both internal analytical precision and total uncertainty that includes those from production rates and scaling schemes, and both reflect 1σ (68 %). All dates also include the originally reported age (“Orig_age”), in addition to information on the production rate (e.g., “Ca_spall_ref”) and calculator (“Calc_ref”) references used for that calculation. Only three studies did not provide enough data to recalculate 36Cl exposure ages (Principato et al., 2006; Schomacker et al., 2012; Hout, 2016), and therefore, were flagged for reduced reliability with brief discussion provided in the corresponding “Comment” column. We note that Licciardi et al. (2006, 2008) did not report 36Cl and 3He exposure ages as those samples were independently dated and used for production rate calculations.

Each ICEland-1 date is first classified in terms of its “Context”: Glacier, Relative sea level, and/or Paleoclimate. This classification is largely taken from the source publication's original interpretation; however, we do use some continuous sediment records (e.g., marine and lake) that recovered complete postglacial sediment sequences and have basal ages near the glaciomarine/glaciolacustrine to postglacial mud transition. These records may have originally been published as paleoclimate records, but we consider there stratigraphies and dates to also be useful for “Glacier context” (i.e., Deglacial). This first-order classification will allow the user to immediately understand the relevant purposes of the data entry and how it can be used in our three focus areas. Glacier dates are then classified in terms of their “Glacial context”, i.e., how it constrains ice growth, retreat, or the past position of the margin based on its stratigraphy: Advance, Margin, Deglacial, Ice free, and/or Glacial. In addition, some TCN dates constrain more complex glacier dynamics and processes requiring several additional “Glacier context” classifications based on interpretations from their original publications: Rock glacier stabilization, Debris-covered glacier stabilization, Debris-covered glacier collapse, and Jökulhlaup drainage. The Ice-free classification is also important for most relative sea level data as well as continuous postglacial sediments that provide targets for paleoceanographic and paleoclimate proxy records. Our classification broadly follows that used for the Eurasian ice sheets (DATED-1, Hughes et al., 2016) as described with minor modifications below. Sediment descriptions and terminologies used here and in the “Comments” column of the database follow those previously used to describe marine sediments on the Iceland shelf (Syvitski et al., 1999).

To assess the reliability of ICEland-1, we leverage the data quality control criteria first developed for DATED-1 (Hughes et al., 2016) with some minor modifications relevant for tephra layers (Davies et al., 2020) – see Table 2. These quality control criteria include those specific to individual dating techniques that address analytical considerations as well as criteria that apply for all dating methods, such as coordinate location and description of geologic setting. We assign each date included in ICEland-1 a qualitative quality control (QC) score on a scale of 1–3, where 1 is most reliable and 3 is least reliable. For dates to receive a QC rating of 1, they must satisfy all criteria. Taking a step forward from DATED-1 (Hughes et al., 2016), for every criterium that is not met, the date's QC rating is reduced by 1, meaning that dates with a QC rating of 3 have two or more unsatisfied criteria. While this approach may result in fewer “reliable” dates, it is in our opinion the most objective and conservative. However, we also acknowledge that while some dates with QC rating of 3 may not date ice sheet processes closely, they can still provide relevant information for where the IIS was in the past and its basal temperatures. For example, reworked ice-contact sediments on the marine shelf signify past ice sheet presence at the site, likely during the LGM, and inherited 36Cl exposure ages of LGM age likely indicate past cold-based ice. In this sense, we urge users to take a careful look at the “Comments” column for recommended use of individual data points as described in the original publication.

ICEland-1 provides a detailed inventory of geochronological data from an array of geologic sites across Iceland's marine and terrestrial realms (Fig. 1). In terms of the primary context classifications, the database includes 567 glacier dates, 257 relative sea level dates, and 1119 paleoclimates, with some dates fitting two or three of the context classifications. In the marine realm, ICEland-1 includes 70 different sediment core locations, with 672 individual 14C and tephra layer data points. For the terrestrial realm, chronological data include 14C/tephra layers from sedimentary archives and glacier margins and TCN exposure ages from bedrock, erratics, and boulders. Terrestrial 14C/tephra layer data are included from 249 sites, totalling 830 data points, and TCN data come from 123 sites, totalling 242 data points. Combined, ICEland-1 covers 442 sites and includes 1744, quality-assessed data points from 165 sources that can be used to constrain glacier, relative sea level, and paleoclimate patterns during the Late Quaternary.

The temporal distribution of ICEland-1 dates varies between the marine and terrestrial realms and with dating techniques. In the marine realm, the majority of 14C dates are younger than 16 kaBP (Fig. 2). The distribution of these younger dates is bimodal with peaks in the number of dates occurring around 12–10 kaBP and then during the Late Holocene (last 4.2 kaBP) (Fig. 2). The oldest marine 14C dates are >52kaBP, and dates spanning 16->52kaBP are relatively few per 1000 year bin (Fig. 2). Many of these older dates received QC ratings below 1; the oldest reliable (QC 1) marine 14C date is 41.9 kaBP, from site V30-130 on the Iceland Plateau that signifies, along with its stratigraphy and position, ice never reached this location during the LGM (this study). On land, the majority of 14C and tephra layer dates are younger than 15 kaBP, with the highest number of dates occurring in the 11–10 kaBP bin (Fig. 3). The oldest terrestrial 14C dates span up to 39.2 kaBP, measured on whale bone from Rauðamelur (Fig. 3). Some of the dates older than 15 kaBP are from humic acids (20.4 and 17.5 kaBP) and have been previously deemed unreliable due to likely contamination from older carbon (Brader et al., 2015). The oldest reliable (QC 1) date is 14.7 kaBP, from the basal, fossiliferous sediments in Melasveit, a coastal marine sediment section in West Iceland (Sigfúsdóttir and Benediktsson, 2020). For TCN, most exposure ages are younger than 16 ka, with the highest number of dates occurring in the 11–10 ka bin (Fig. 4). The oldest exposure age is >62ka, although this date as well as many over 15 ka likely incorporate some degree of nuclide inheritance that reduce the date's reliability (e.g., Principato et al., 2006; Brynjólfsson et al., 2015b). The oldest reliable exposure age is from an erratic on the Hornstrandi peninsula, northwest Iceland (Brynjólfsson et al., 2015b), and ranges from 14.2–17.6 ka, depending on choice of calculator (Balco et al., 2008; Marrero et al., 2016; Schimmelpfennig et al., 2022).

ICEland-1's data coverage is variable across the marine and terrestrial realms. For the marine realm, while 14C data are limited to the western and northern shelf, the coverage is relatively dense. No chronological data currently exist from the eastern and southern shelf (Fig. 1). For the terrestrial realm, 14C and tephra layer data exist from all eight regions: Northeast (158), Northwest (114), Westfjords (149), West (155), Capital Region (65), Southern Peninsula (12), South (132), and East (45) (Fig. 1). For TCN, exposure data are largely restricted to the Northeast (43), Northwest (82), and Westfjords (43). The bulk of TCN data in South Iceland constrains Ca spallation and 3He production rates, rather than glacial processes explicitly (Licciardi et al., 2006, 2007) (Fig. 1). While not included in ICEland-1's chronological database, a variety of geomorphological data exist from bathymetric surveys along Iceland's shelf that constrain the spatial footprint of past ice limits (e.g., Ólafsdóttir, 1975; Boulton et al., 1988; Syvitski et al., 1999; Spagnolo and Clark, 2009) and can be used alongside ICEland-1 dates to reconstruct past ice sheet patterns (Fig. 5).

ICEland-1 provides quality control (QC) assessment of each date and detailed information as to its reliability and interpretation in the context of glacier and paleoclimate history. For marine 14C dates, 496 dates received a QC 1 rating (74 %), 151 dates for QC 2 (22 %), and 25 dates for QC 3 (4 %) (Fig. 2). For terrestrial 14C dates and tephra layers, 329 dates received a QC 1 rating (40 %), 449 dates for QC 2 (54 %), and 52 dates for QC 3 (6 %) (Fig. 3). For TCN dates, 184 dates received a QC 1 rating (76 %), 49 dates for QC 2 (20 %), and 7 dates for QC 3 (3 %) (Fig. 4). Combined, dates with a QC 1 rating represent 58 % of the entire dataset, with QC 2 dates representing 37 %, and QC 3 dates representing 5 %. We suggest dataset users rely on the QC 1 as these have been assessed to be the most reliable, along with QC 2 dates at the user's discretion (Fig. 5). QC 3 dates should not be used for chronological constraint. However, some QC 3 dates, such as ice-contact sediments on the marine shelf that have been reworked (e.g., MD99-2264) or not (e.g., B997-322), provide useful spatial constraints on potential limits of the IIS during the LGM (Fig. 5).

Where possible, temporal uncertainties are explicitly quantified and reported in ICEland-1. These uncertainties primarily relate to the analytical measurements used for 14C and TCN and their calibrated and calculated ages, respectively. All quantified temporal uncertainties are reported at the 1σ (68 %) confidence level. However, it is important to remember that there is also unquantified uncertainty in dates as well. For example, due to unconstrained ΔR corrections, 14C dates in glacial age marine sediments are likely too old, possibly by millennia (Heaton et al., 2023). Similarly, bulk and humic acid 14C dates from the terrestrial realm may also be older than their stratigraphic position due to the incorporation of pre-existing organic carbon on the landscape (e.g., Geirsdóttir et al., 2009b). In some cases, the relative impact of old 14C bias can be accounted for in terrestrial sediments by comparing bulk 14C dates with tephra layers of known age that have been independently dated elsewhere (e.g., Harning et al., 2024). Apparent TCN exposure ages can be too old or too young due to inheritance or subsequent burial, but these factors cannot be quantified for current entries in ICEland-1 due to the single isotope approach used. Moreover, different TCN exposure age calculators yield a spectrum of apparent ages most likely due to the choice of various production rates used in age calculations. While CREp permits the selection of productions rates for 36Cl (e.g., Ca spallation, K spallation, etc.), both CRONUScalc and CRONUS Earth calculator v.3 (a developmental version at the time of this writing) are coded to use specific production rates that may differ from each other, thus resulting in the range of exposure observed between calculators. Furthermore, the utilization of CREp permits the input of specific production rates allowing users to incorporate values calibrated from source material more representative of their data (e.g., Ca spallation from Icelandic rocks, Licciardi et al., 2008) thus increasing reliability of the resulting TCN exposure ages. While previous studies have relied on a single calculator for simplicity (e.g., CREp), we suggest that all ages are treated as possibilities and reported as a range.

In addition to temporal, the other major source of uncertainty and bias is in the spatial metadata. First, 248 dates (14 %) have coordinate locations estimated from georeferenced maps in Google Earth. While this reduces the reliability of the date in our quality control assessment, it is challenging to accurately quantify the error associated with these inferred coordinates. Generally, the datum's map location point from the original source publication yields a diameter up to 5 km, meaning that ±2.5km is a reasonable spatial uncertainty estimate for map-inferred coordinates. Second, robust relative sea level data, such as sea level index points, require specific indicative meaning that describe where the sea-level indicator formed with respect to tide levels, reference water levels, and elevation metadata (Hijma et al., 2015). In Iceland, sea level index points are generally found from isolation basins with varying degrees of uncertainty reported for key metadata (Lloyd et al., 2009; Brader et al., 2015, 2017). All pertinent relative sea level metadata provided in original publications is reported in ICEland-1's “Comment” column. The final spatial uncertainty is for dates from sediments that have been mobilized or reworked from their initial deposition site. We cannot provide spatial uncertainty estimates for these dates, and therefore, substantially reduces their reliability for dating purposes.

ICEland-1 provides detailed acknowledgement of spatial and temporal uncertainties for each data point. For example, spatial uncertainties related to inferred coordinates and reworked stratigraphies and temporal uncertainties related to bulk 14C and expected TCN inheritance/burial are incorporated into the quality control assessment and noted in the “QC_comment” column. Temporal uncertainties for analytical measurements and age calculations are provided to the extent that the data were available in the original source publication. We also provide detailed notes in the “Comment” column where dates should be treated as maximum and minimum as it relates to both 1) dating techniques (e.g., glacial age marine sediment and terrestrial bulk sediment 14C), and 2) stratigraphy. For the latter, it is important to remember that in most cases, 14C dates do not date the exact moment of glacier or relative sea level change at the site. Using deglacial 14C dates as an example, the organic material dated is deposited after the date the glacier retreats from the site and/or catchment, leaving an unconstrained amount of time between the two. Hence, these are limiting ages.

The curated database was compiled using the following sources: Alsos et al. (2021); Andersen et al. (1989); Andrés et al. (2019, 2025); Andresen et al. (2005); Andrews et al. (2000, 2001, 2002a, b, 2009, 2021a, b); Andrews and Giraudeau (2003); Andrews and Helgadóttir (2003); Ardenghi et al. (2024); Ásbjörnsdóttir and Norðdahl (1995); Ashwell (1967, 1975); Axford et al. (2007, 2009); Bender (2020); Bendle and Rosell-Melé (2007); Bergþórsdóttir (2014); Björck et al. (1992); Black (2008); Blair et al. (2015); Brader et al. (2015, 2017); Brynjólfsson et al. (2015a, b); Caseldine et al. (2003, 2006); Castañeda et al. (2004); Coquin et al. (2016); Decaulne et al. (2016); Doner (2003); Dugmore (1989); Dunhill et al. (2004); Eddudóttir et al. (2015, 2016); Einarsson (1961, 1964); Eiríksson et al. (1997, 2000a, b, 2004); Erlendsson and Edwards (2009); Erlendsson et al. (2009); Farnsworth et al. (2025); Fernández-Fernández et al. (2019, 2020); Gathorne-Hardy et al. (2009); Gehrels et al. (2006); Geirsdóttir and Eiríksson (1994); Geirsdóttir et al. (1997, 2002, 2009b, 2022); Gudmundsdóttir et al. (2011); Gunnarson (2017); Håkansson (1987); Hannesdóttir (2006); Hansom and Briggs (1991); Hardardóttir et al. (2001); Harning et al. (2016a, b, 2018a, b, 2019a, b, 2023, 2024, 2025a, b c); Helgadóttir (1984); Hellqvist et al. (2020); Hjartarson (1989, 1993); Hjartarson and Ingólfsson (1988); Hjort et al. (1985); Holmes et al. (2016); Hout (2016); Hunt (1992); Ingólfsson (1985, 1987, 1988); Ingólfsson and Norðdahl (2001); Ingólfsson et al. (1995); Jennings et al. (2000); Jiang et al. (2015); Jóhannesson et al. (1994, 1997); Jóhannsdóttir (2007); John (1974); Jónsdóttir et al. (2015); Kaldal (1993); Karlsdóttir et al. (2012, 2014); Knudsen and Eiríksson (2002); Kirkbride et al. (2006); Kjartansson (1966); Kjartansson et al. (1964); Larsen et al. (2012); Larsen et al. (2024); Licciardi et al. (2006, 2007); Lloyd et al. (2009); Magnúsdóttir and Norðdahl (2000); Maizels (1991); Manley and Jennings (1996); Mercier et al. (2013, 2017); Norðdahl (1991); Norðdahl and Hjort (1987, 1993); Norðdahl and Ásbjörnsdóttir (1995); Norðdahl and Sæmundsson (1999); Norðdahl and Einarsson (2001); Norðdahl and Pétursson (2005); Norðdahl et al. (2019); Óladóttir et al. (2020); Ólafsdóttir et al. (2010); Olsson et al. (1969); Palacios et al. (2021); Pétursson (1986, 1991, 1997); Phillips and Plummer (1996); Phillips et al. (1996, 2001); Principato (2003, 2008); Principato et al. (2006); Quillmann et al. (2009, 2010); Richardson (1997); Riddell et al. (2018, 2024); Roy et al. (2018); Rundgren (1995, 1998); Rundgren et al. (1997); Sæmundsson (1995); Sæmundsson and Jóhannesson (2005); Sæmundsson et al. (2012); Santo-González et al. (2025); Schimmelpfennig et al. (2009, 2019); Schomacker et al. (2003, 2012, 2016); Sigfúsdóttir and Benediktsson (2020); Sigurgeirsson (1993, 2016); Sigurgeirsson and Leósson (1993); Sigvaldason (2002); Símonarson and Leifsdóttir (2002); Smith and Licht (2000); Stone (2000); Stoner et al. (2007); Stone (2000); Stötter (1991); Striberger et al. (2011); Sveinbjörnsdóttir and Johnsen (1991); Sveinbjörnsdóttir et al. (1993, 1998); Swanson and Caffee (2001), Tanarro et al. (2021); Thorarinsson (1956); Thors and Helgadóttir (1991); van der Bilt et al. (2021); Vilmundardóttir et al. (1979); Wanamaker et al. (2012); Wastl (2000); Wastl et al. (2001); Wells et al. (2025).