<?xml-model href='http://www.tei-c.org/release/xml/tei/custom/schema/relaxng/tei_all.rng' schematypens='http://relaxng.org/ns/structure/1.0'?><TEI xmlns="http://www.tei-c.org/ns/1.0">
	<teiHeader>
		<fileDesc>
			<titleStmt><title level='a'>Timing of Pedogenic Carbonate Formation in Fine‐Grained Soils: Decoupled T(Δ &lt;sub&gt;47&lt;/sub&gt; ) and δ &lt;sup&gt;18&lt;/sup&gt; O &lt;sub&gt;w&lt;/sub&gt; Seasonal Bias</title></titleStmt>
			<publicationStmt>
				<publisher>American Geophysical Union</publisher>
				<date>02/01/2025</date>
			</publicationStmt>
			<sourceDesc>
				<bibl> 
					<idno type="par_id">10585196</idno>
					<idno type="doi">10.1029/2024PA005071</idno>
					<title level='j'>Paleoceanography and Paleoclimatology</title>
<idno>2572-4517</idno>
<biblScope unit="volume">40</biblScope>
<biblScope unit="issue">2</biblScope>					

					<author>Rachel Havranek</author><author>Kathryn Snell</author><author>Sarah Brookins</author><author>Brett Davidheiser‐Kroll</author>
				</bibl>
			</sourceDesc>
		</fileDesc>
		<profileDesc>
			<abstract><ab><![CDATA[<title>Abstract</title> <p>Historically, clumped isotope thermometry (T(∆<sub>47</sub>)) of soil carbonates has been interpreted to represent a warm‐season soil temperature based dominantly on coarse‐grained soils. Additionally, T(∆<sub>47</sub>) allows the calculation of the oxygen isotope composition of soil water (δ<sup>18</sup>O<sub>w</sub>) in the past using the temperature‐dependent fractionation factor between soil water and pedogenic carbonate, but previous work has not measured δ<sup>18</sup>O<sub>w</sub>values with which to compare to these archives. Here, we present clumped isotope thermometry of modern soil carbonates from three soils in Colorado and Nebraska, USA, that have a fine‐to‐medium grain size, contain clay, and are representative of many carbonate‐bearing paleosols preserved in the rock record. At two of the three sites, Briggsdale, CO and Seibert, CO, T(∆<sub>47</sub>) overlaps with mean annual soil temperature (MAST), and the calculated δ<sup>18</sup>O<sub>w</sub>overlaps within uncertainty with measured δ<sup>18</sup>O<sub>w</sub>at carbonate bearing depths. At the third site, in Oglala National Grassland, NE, mean T(∆<sub>47</sub>) is 8–11°C warmer than MAST, and the calculated δ<sup>18</sup>O<sub>w</sub>has a significantly higher isotope value than any observations of δ<sup>18</sup>O<sub>w</sub>. At all three sites, even in the fall season, δ<sup>18</sup>O<sub>w</sub>values at carbonate bearing depths overlap with spring rainfall δ<sup>18</sup>O<sub>w</sub>, and there is little to no evaporative enrichment of δ<sup>2</sup>H<sub>w</sub>and δ<sup>18</sup>O<sub>w</sub>values. These data challenge long‐held assumptions that all pedogenic carbonate records a warm‐season bias, and that δ<sup>18</sup>O<sub>w</sub>at carbonate‐bearing depths is affected by evaporative enrichment.</p>]]></ab></abstract>
		</profileDesc>
	</teiHeader>
	<text><body xmlns="http://www.tei-c.org/ns/1.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xlink="http://www.w3.org/1999/xlink">
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="1.">Introduction</head><p>Pedogenic carbonate nodules are a commonly used material for paleoenvironment reconstructions, because multiple stable isotope measurements (e.g., &#948; 13 C, &#948; 18 O, &#8710; 47 , &#8710; 17 O, etc.) can be used as proxy measurements for many environmental conditions of interest, such as vegetation composition <ref type="bibr">(Cerling, 1984;</ref><ref type="bibr">Cerling &amp; Quade, 1993;</ref><ref type="bibr">Passey, 2012)</ref>, ancient pCO 2 (e.g., <ref type="bibr">Breecker et al., 2010;</ref><ref type="bibr">Da et al., 2019;</ref><ref type="bibr">Ji et al., 2018;</ref><ref type="bibr">Zhang et al., 2018)</ref>, changes in the hydrologic system <ref type="bibr">(Kelson et al., 2023;</ref><ref type="bibr">Levin et al., 2011;</ref><ref type="bibr">Passey, 2012;</ref><ref type="bibr">Passey et al., 2014)</ref>, and temperature change associated with both climatic changes and tectonic evolution (e.g., <ref type="bibr">Carrapa et al., 2014;</ref><ref type="bibr">DeCelles et al., 2007;</ref><ref type="bibr">Garzione et al., 2006;</ref><ref type="bibr">Ghosh et al., 2006;</ref><ref type="bibr">Huntington &amp; Lechler, 2015;</ref><ref type="bibr">Ingalls et al., 2018;</ref><ref type="bibr">Rugenstein &amp; Chamberlain, 2018;</ref><ref type="bibr">Snell et al., 2013)</ref>. To improve our understanding of the relationship between the stable isotope compositions of pedogenic carbonates and environmental conditions in the geologic past, the geoscience community has studied 'modern' carbonate formation in Holocene soils, which serve as an approximation of modern conditions (e.g., <ref type="bibr">Breecker et al., 2009;</ref><ref type="bibr">Burgener et al., 2016;</ref><ref type="bibr">Cerling, 1984</ref>;</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>RESEARCH ARTICLE</head></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>10.1029/2024PA005071</head><p>Key Points:</p><p>&#8226; This work provides further evidence that fine-grained soils will often have a different seasonal bias than coarsegrained soils</p><p>&#8226; In fine-grained, carbonate bearing soils there is very little annual-scale variation in the isotopic composition of soil water at depths &gt;50 cm &#8226; We infer that the difference in seasonal bias is due to slower dry-down in finegrained soils, delaying pedogenic carbonate formation into the fall Supporting Information: Supporting Information may be found in the online version of this article. <ref type="bibr">Cerling &amp; Quade, 1993;</ref><ref type="bibr">Hough et al., 2014;</ref><ref type="bibr">Huth et al., 2019;</ref><ref type="bibr">Kelson et al., 2020;</ref><ref type="bibr">Peters et al., 2013)</ref>. These modern calibration studies are useful for constraining the strengths and limitations of pedogenic carbonate as a proxy for paleoenvironmental conditions.</p><p>Most modern calibration studies on pedogenic carbonate done over the last 10-15 years have capitalized on the development of clumped isotope thermometry to both understand the relationship between a clumped isotope temperature (T(&#8710; 47 )) and air temperature and to improve our understanding of the timing and mechanism of pedogenic carbonate formation <ref type="bibr">(Burgener et al., 2016;</ref><ref type="bibr">Gallagher &amp; Sheldon, 2016;</ref><ref type="bibr">Hough et al., 2014;</ref><ref type="bibr">Hudson et al., 2024;</ref><ref type="bibr">Huth et al., 2019;</ref><ref type="bibr">Peters et al., 2013)</ref>. An additional benefit of clumped isotope measurements is that it is possible to calculate the isotopic composition of the soil water that the carbonate precipitated from, which offers additional insight into formation mechanism, timing and soil hydrology processes. However, few studies to date have made use of the additional potential insight from the &#948; 18 O of soil water (&#948; 18 O w ). In modern calibration studies, soils have generally been chosen as part of climatic or elevation gradients and have broadly been conducted in immature, medium to coarse grained soils <ref type="bibr">(Burgener et al., 2018;</ref><ref type="bibr">Hough et al., 2014;</ref><ref type="bibr">Huth et al., 2019;</ref><ref type="bibr">Oerter &amp; Amundson, 2016;</ref><ref type="bibr">Peters et al., 2013)</ref>. In a synthesis of these studies, soil grain size appears to be an important control on the temperature (i.e., mean annual temperature vs. warm-season biased temperature) that was recorded by pedogenic carbonate, with coarser grained soils more biased toward warmer temperatures <ref type="bibr">(Gallagher &amp; Sheldon, 2016;</ref><ref type="bibr">Kelson et al., 2020)</ref>. This synthesis also demonstrated that there was limited data from fine-grained and clay-rich soils as compared to other grain size classes. The observation of a grain-size influenced temperature bias, alongside the limited amount of temperature and soil water isotope data from fine grained and clay-rich soils, is important because many terrestrial paleoclimate records are based on pedogenic carbonate datasets from fine-grained and clay-rich paleosols (e.g., <ref type="bibr">Bowen et al., 2015;</ref><ref type="bibr">Da et al., 2019;</ref><ref type="bibr">Driese et al., 2016;</ref><ref type="bibr">Salazar-Jaramillo et al., 2022)</ref>. While the geologic community clearly understands the sets of processes that influence carbonate mineral formation, it is unclear which (if any) dominate pedogenic carbonate mineral formation, and how these are affected by (or vary with) soil texture.</p><p>To fill the critical gap in our understanding of carbonate formation in fine grained and clay-rich soils, we studied three Holocene fine-to-medium grained, clay-containing soils in Colorado and Nebraska, USA. We used clumped isotope thermometry, climatological data and soil data to learn about carbonate formation in these soils. Additionally, in this study, we include measurements of precipitation, regional surface water and soil water stable isotope geochemistry to gain a deeper understanding of how the oxygen isotope composition of carbonate records information about the hydrologic cycle.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="2.">Background</head></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="2.1.">Stable Isotope Geochemistry of Pedogenic Carbonate</head><p>The &#948; 18 O value of pedogenic carbonate (&#948; 18 O C ) is a function of both the temperature at which the mineral formed and the &#948; 18 O w from which the mineral precipitated (e.g., <ref type="bibr">Cerling, 1984;</ref><ref type="bibr">Kim &amp; O'Neil, 1997)</ref>. The &#948; 18 O w value is initially set by the &#948; 18 O value of meteoric water, but can be modified by evaporation (e.g., <ref type="bibr">Cerling &amp; Quade, 1993;</ref><ref type="bibr">Hough et al., 2014;</ref><ref type="bibr">Quade et al., 2007)</ref>; the extent to which evaporation modifies the recorded &#948; 18 O c value is unknown at most sites (e.g., <ref type="bibr">Breecker et al., 2009)</ref>.</p><p>T(&#8710; 47 ) provides an estimate of pedogenic carbonate formation temperature through a relationship between temperature and the degree of bond ordering of the heavy, rare isotopes 13 C and 18 O <ref type="bibr">(Eiler, 2007)</ref>. At low temperatures there is a greater degree of 13 C -18 O ordering, or clumping, in the carbonate ion (CO 3 2-) compared to a purely random, stochastic distribution of the rare heavy isotopes, leading to an inverse relationship between &#916; 47 and temperature <ref type="bibr">(Eiler, 2007)</ref>. T(&#8710; 47 ) can additionally be used with the simultaneously measured &#948; 18 O C to estimate the &#948; 18 O w that the carbonate formed from, using an equation for the temperature-dependent fractionation of oxygen isotopes between carbonate minerals and water (e.g., <ref type="bibr">Da&#235;ron et al., 2019;</ref><ref type="bibr">Kim &amp; O'Neil, 1997)</ref>.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="2.2.">Pedogenic Carbonate Formation</head><p>There are three primary levers that control carbonate mineral formation in soils. First, pCO 2 of the soil environment, which is controlled by primary productivity and soil permeability, acts as a control on soil water pH; carbonate mineral formation is favored at lower soil pCO 2 (e.g., <ref type="bibr">Breecker, 2013;</ref><ref type="bibr">Cerling &amp; Quade, 1993)</ref>. Second, soil temperature controls carbonate mineral formation because CO 2 becomes less soluble in water as temperatures increase, and therefore carbonate solids are favored to form at higher temperatures. Soil temperature is controlled by the surface temperature and the thermal properties of the soil which is influenced by texture, porosity, water content, bulk density, and mineralogy (e.g., <ref type="bibr">de Vries, 1952;</ref><ref type="bibr">Hillel, 2004)</ref>. Lastly, increased calcium availability can drive mineral formation. Calcium cation availability depends on the delivery of calcium either from the soil substrate and/or from detrital material, both of which are influenced by local geology and water movement through the soil column (e.g., <ref type="bibr">Birkeland, 1999;</ref><ref type="bibr">Zamanian et al., 2016)</ref>. Under drying soil conditions, Ca 2+ can become increasingly concentrated, driving supersaturation (e.g., <ref type="bibr">Breecker et al., 2009;</ref><ref type="bibr">Hough et al., 2014)</ref>. Over the last 40 years, there has been considerable effort to understand which of these levers dominates pedogenic carbonate formation and controls the timing and style of pedogenic carbonate formation, in part to improve interpretation of paleoenvironmental proxies <ref type="bibr">(Breecker et al., 2009;</ref><ref type="bibr">Burgener et al., 2018;</ref><ref type="bibr">Cerling, 1984;</ref><ref type="bibr">Cerling &amp; Quade, 1993;</ref><ref type="bibr">Hough et al., 2014;</ref><ref type="bibr">Huth et al., 2019;</ref><ref type="bibr">Kelson et al., 2020;</ref><ref type="bibr">Oerter &amp; Amundson, 2016;</ref><ref type="bibr">Peters et al., 2013)</ref>. Recent work also suggests that the timing and relative magnitude of carbonate mineral dissolution strongly influences the preserved isotope record <ref type="bibr">(Breecker et al., 2025;</ref><ref type="bibr">Huth et al., 2019)</ref>.</p><p>In general, there are three approaches that fall along a continuum that have been used to study pedogenic carbonate formation in Holocene soils. The first approach is to create a timeseries of the &#948; 13 C value of soil CO 2 alongside timeseries of relevant soil conditions, including soil temperature, soil moisture, and/or soil pCO 2 (e.g., <ref type="bibr">Breecker et al., 2009;</ref><ref type="bibr">Oerter &amp; Amundson, 2016)</ref>. These studies either measure &#948; 18 O w once to establish equilibrium, or assume equilibrium between &#948; 18 O of CO 2 and the &#948; 18 O w to interpret the timing of carbonate formation. Then, theoretical &#948; 13 C C and &#948; 18 O C are calculated from the timeseries data, and are compared to the measured &#948; 13 C C and &#948; 18 O C of pedogenic carbonate. The second approach omits measurements of the &#948; 13 C value of soil CO 2 , and instead adds carbonate clumped isotope data. These studies create a time series of environmental parameters, such as soil temperature, soil moisture, air temperature, precipitation amounts, and &#948; 18 O value of precipitation (&#948; 18 O p ). Then clumped isotope data (T(&#916; 47 ), &#948; 13 C C , &#948; 18 O C , and calculated &#948; 18 O w ) are used to constrain the timing of pedogenic carbonate formation from the same, or nearby sites <ref type="bibr">(Burgener et al., 2016</ref><ref type="bibr">(Burgener et al., , 2018;;</ref><ref type="bibr">Hough et al., 2014;</ref><ref type="bibr">Peters et al., 2013)</ref>. The final approach, taken by <ref type="bibr">Huth and others (2019)</ref>, was to add soil water isotopes taken from discrete samples a few times throughout the study to the second approach, thereby circumventing assumptions of isotopic equilibrium. While all of these studies display nuance related to their particular field sites, the current consensus is that at low to moderate elevations (&lt;4000 m), across a broad range of environmental conditions, pedogenic carbonate clumped isotope temperatures are typically biased towards warm season temperatures, and the dominant hypothesis is that soil dry-down drives pedogenic carbonate formation <ref type="bibr">(Breecker et al., 2009;</ref><ref type="bibr">Burgener et al., 2016</ref><ref type="bibr">Burgener et al., , 2018;;</ref><ref type="bibr">Hough et al., 2014;</ref><ref type="bibr">Huth et al., 2019;</ref><ref type="bibr">Kelson et al., 2020;</ref><ref type="bibr">Oerter &amp; Amundson, 2016;</ref><ref type="bibr">Peters et al., 2013)</ref>. However, the majority of the studies highlighted above were conducted in poorly developed (e.g., inceptisol, entisol, aridisol), medium to coarse-grained soils. Soil texture is a primary control on the rate and style of soil dry-down because in a coarse-grained soil, the pores will empty quickly and become nonconductive even at high suction, whereas in a fine-grained soil the pores will empty more slowly and conductivity will not decrease as quickly even as suction increases <ref type="bibr">(Hillel, 2004)</ref>. Therefore, it makes sense that there may be a grain size bias in recorded clumped isotope temperatures <ref type="bibr">(Kelson et al., 2020)</ref>. This highlights a critical gap for paleoclimate studies that apply T(&#916; 47 ) fine-grained or clay-rich paleosols.</p><p>In addition, despite the importance for understanding both carbonate formation and interpreting the &#948; 18 O w archive, few of the above-mentioned studies characterize &#948; 18 O w , which highlights an additional critical data gap. In a synthesis of the studies that incorporate clumped isotope thermometry data, the calculated mean &#948; 18 O w of all the datasets together was not biased towards summer precipitation values, as one might expect with the assumption that carbonate typically forms in the summer, but was instead similar to an integrated mean annual &#948; 18 O p value <ref type="bibr">(Kelson et al., 2020)</ref>. However, individual datasets within the compilation showed considerable scatter about the mean, and in some cases, a coherency that resulted in individual dataset means differing significantly from mean annual precipitation values. This result leaves open the question about how to interpret soil water &#948; 18 O w values derived from clumped isotope studies.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="3.">Methods</head><p>The general approach of this study was to compare pedogenic carbonate formation across three different soil textures within the same soil moisture regime (Ustic). From north to south, the three field sites are located in</p><p>Paleoceanography and Paleoclimatology 10.1029/2024PA005071 </p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="3.1.">Field Methods</head></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="3.1.1.">Meteorological Data From Oglala National Grassland, NE</head><p>The ONG site was instrumented with air temperature, relative humidity and precipitation quantity sensors in June 2021 (Onset, S-THC-M002 and Davis Instruments Rain gauge sensor, S-RGE-M002) (Figure <ref type="figure">2b</ref>). However, due to significant livestock interference, this dataset is discontinuous. Therefore, we also incorporate data from the Alliance, Nebraska MesoNet site (Latitude: 42.18&#176;N/Longitude: 102.92&#176;E/Elevation: 1,241 m) that is 103 km to the southeast of the ONG site. The Nebraska MesoNet is supported by the state of Nebraska and operated by the University of Nebraska Lincoln. At the Alliance Mesonet site, air temperature, relative humidity, and precipitation quantity are recorded on an hourly basis, as well as soil moisture and temperature at 10, 25, 50, and 100 cm that are recorded at 1-minute intervals.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="3.1.2.">Meteorological Data From Briggsdale, CO and Seibert, CO</head><p>Both field sites in Colorado are co-located with CoAgMET sites, which are operated by the Colorado Climate Center, 2022 (Colorado State University, Fort Collins, CO). At both sites, the Colorado Climate Center measures air temperature, relative humidity, wind speed and direction, precipitation quantity, and soil temperature at 5 and 15 cm depths (Figure <ref type="figure">2a</ref>). The Briggsdale site has been instrumented as a CoAgMET site since 31 July 2002 and the Seibert site has been instrumented since 2 April 2015. </p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>Paleoceanography and Paleoclimatology</head><p>10.1029/2024PA005071</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="3.1.3.">Soil Conditions</head><p>At each site, we excavated two holes (Figure <ref type="figure">2b</ref>). We dug out the first hole once at the beginning of the study to install soil temperature and moisture loggers as well as water vapor sampling probes. We report soil temperature data at 25, 50, and 100 cm soil depth, and soil moisture at depths of 10, 25, 50, 75, and 100 cm. We also have soil temperature data from 10 to 75 cm from some sites, but to stay consistent in the comparison across sites, we excluded them. The excluded data do not affect any interpretations. We installed water vapor sampling probes into the intact sidewall of the pit at soil depths of 25, 50 and 75 cm. After logger and probe installation, we carefully backfilled the first hole with soil from approximately the same depths, taking care to compact the soil appropriately and maintain horizon depths as much as possible. We also excavated the second hole at the start of the study, and we maintained this hole throughout the study to store the soil water isotope storage system (SWISS) and associated components. We chose to store the SWISS units and associated components belowground to limit temperature variability experienced by the units, better protect against precipitation, weather, and animal On the left is the instrumented hole where the soil temperature and moisture loggers as well as the vapor permeable probes were installed. On the right is the hole where the SWISS and power components were stored. Vapor impermeable tubing that connects the vapor permeable probes with the SWISS was buried at a depth of approx. 15 cm in a PVC pipe.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>Paleoceanography and Paleoclimatology</head><p>10.1029/2024PA005071</p><p>interference, and to make the sites more discreet to avoid human interference. Lastly, we connected the SWISS to the water vapor sampling probes via Bev-A-Line impermeable tubing, which was run through a PVC pipe buried at approximately 15 cm depth. Burying the impermeable tubing reduces the effect of diurnal temperature variability, and limits condensation as water vapor travels from the soil to the SWISS.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="3.2.">Soil Characterization</head><p>We excavated soil pits to a depth of 100 cm for soil description and sample collection. We collected bulk soil samples for soil texture and mineralogical analyses at depths of approximately 10, 25, 50, 75 and 100 cm. In some cases, we collected additional bulk soil samples from horizons or transitions of interest. For example, we collected extra samples where the character (size and/or hardness) of the pedogenic carbonate nodules changed or where there was a large change in soil color without a change in carbonate characteristics. At this time, we also collected pedogenic carbonate nodules for isotope geochemistry from at least two levels in each pit (Figure <ref type="figure">1b</ref>).</p><p>Soil texture and mineralogical analyses were performed in the Sediment/Plant Analysis and Processing Lab and XRD/Soil Processing Lab, respectively, which are both in the Institute of Arctic and Alpine Research (INSTAAR) at the University of Colorado Boulder. We measured soil particle size using a Malvern Mastersizer 3000. Broadly, we defined soils that were &lt;50% sand as fine-to medium-grained. For the purposes of this study, we define silt loam as fine-grained, sandy-loam as medium-grained and loamy sand as moderately coarse-grained. We characterized soil mineralogy using Siemens D5000 XRD with ROCKJOCK6 processing software. See Supporting Information S1 for greater detail.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="3.3.">Carbonate Nodule Characterization</head><p>Using thin sections of individual carbonate nodules, we used optical and cathodoluminescence (CL) microscopy to identify fabrics and textures consistent with primary carbonate formation as well possible additional, later stages of carbonate formation within the nodules. See Supporting Information S1 for greater detail.</p><p>To determine the approximate age of carbonate nodule formation, whole nodules from 2 to 3 soil levels from each site were dated using 14 C radiocarbon dating. Samples were prepared in the INSTAAR Laboratory for AMS Radiocarbon Preparation and Research. See Supporting Information S1 for greater detail.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="3.4.">Carbonate Stable Isotope Geochemistry</head><p>To produce sample powder for stable isotope geochemistry, we either (1) drilled nodules mounted in epoxy from which paired thin section were cut, or (2) crushed whole nodules and homogenized the powder using mortar and pestle.</p><p>We collected carbonate clumped isotope data during six sessions in 2022 and 2024 on a Thermo Scientific 253 Plus dual-inlet isotope ratio mass spectrometer with an on-line, custom-made vacuum extraction line in the University of Colorado Boulder Earth System Stable Isotope Laboratory (CUBES-SIL; RRID SCR-019300, <ref type="bibr">Fetrow et al., 2022)</ref>. The calcium carbonate standards ETH 1, 2, and 3 were used to correct the data to the community accepted Intercarb-Carbon Dioxide Equilibrium Scale (ICDES) <ref type="bibr">(Bernasconi et al., 2021)</ref>. In addition, when available in a given session the additional standards ETH-4, IAEA-C1, IAEA-C2 were added. The carbonate standard IAEA-C2 was used as a monitoring standard as it is a low-temperature standard with similar &#948; 13 C c and &#948; 18 O c values to our samples. Two samples from Briggsdale (B50 and B110) required a double-clean or bleach-clean protocol during measurement to sufficiently remove isobaric contaminants from the sample gas in order to determine a reliable T(&#916; 47 ) value (see Supporting Information S1). Prior to double cleaning, both samples yielded unreasonably cold temperatures (approx. -45&#176;C), which can happen when there is contamination of mass-47 with non-CO 2 compounds. Additionally, to test if the high temperatures from the ONG site were due to contamination effects, we treated 3 additional samples. After an initial replicate of each of those samples, the &#916; 47 values from the ONG were not significantly from other values from that site, so we did not pursue further replicates.</p><p>Temperatures were estimated from the mean &#8710; 47 values using the <ref type="bibr">Anderson et al. (2021)</ref> calibration line, and temperature uncertainties were estimated by propagating the &#916; 47 standard error for every sample <ref type="bibr">(Da&#235;ron, 2020)</ref>. See Supporting Information S1 for greater detail. From the measured T(&#916; 47 ) and &#948; 18 O c values, we calculated the &#948; 18 O w value that the carbonate mineral precipitated from using the equilibrium fractionation between calcite and</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>Paleoceanography and Paleoclimatology</head><p>10.1029/2024PA005071</p><p>water established by <ref type="bibr">Da&#235;ron et al. (2019)</ref>. The &#948; 18 O w value are reported as the per mil (&#8240;) deviation relative to VSMOW/SLAP <ref type="bibr">(Coplen, 2011)</ref>. Repeatability of all &#916; 47 (anchors + unknowns) was 0.0183&#8240;.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="3.5.">Water Stable Isotope Geochemistry</head></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="3.5.1.">Precipitation and Surface Water Stable Isotope Geochemistry</head><p>We collected integrated monthly precipitation samples for stable isotope analysis (&#948; 18 O p and &#948; 2 H p ) following <ref type="bibr">Scholl et al. (1996)</ref>. We measured the precipitation in the INSTAAR Stable Isotope Laboratory at CU Boulder using a Picarro L2130-i Isotope and Gas Concentration Analyzer paired with an autosampler and high precision vapourizer unit. We measured 2-3 replicates of each sample. The isotope values were corrected for instrument drift during the run period and memory effects between samples, and &#948; 18 O p and &#948; 2 H of precipitation (&#948; 2 H p ) values are reported as the per mil (&#8240;) deviation relative to VSMOW/SLAP. Final sample &#948; 18 O p and &#948; 2 H p values were calculated as the mean value of replicates; error is reported as 2 sigma standard deviation. Precision varied by run, but for &#948; 18 O was between 0.02 and 0.03&#8240;, and for &#948;2H was between 0.09 and 0.14&#8240;.</p><p>To complement precipitation isotope data from the study sites, we also collated precipitation isotope data from nearby National Ecological Observation Network (NEON) sites that are relatively proximal to both field sites in Colorado (Water Isotopes Database, 2017). There was no existing proximal precipitation isotope data available in the Nebraska panhandle region, and so we instead use data from the US Network of Isotopes in Precipitation (US NIP) site located in Newcastle, WY (Water Isotopes Database, 2017).</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="3.5.2.">Soil Water Isotope Geochemistry</head><p>We collected two soil water isotope time series from all three field sites using Soil Water Isotope Storage System (SWISS) units <ref type="bibr">(Havranek et al., 2020</ref><ref type="bibr">(Havranek et al., , 2023))</ref>. The SWISS units are capable of autonomously collecting 15 soil water vapor samples in the field over a &#8764;1 month interval (and longer in some cases) before returning the units back to the laboratory for stable isotope analysis <ref type="bibr">(Havranek et al., 2023)</ref>. As noted in Section 3.1.3, water vapor probes for vapor collection were installed at depths of 25, 50 and 75 cm at all three field sites. We collected soil water vapor samples at intervals of 5-7 days; the SWISS was under development during this interval, and so while the units were deployed continuously from June 2021 to September 2021 and April 2022 to October 2022, only two &#8764;1 month soil water isotope time series from 2022 yielded reliable data for each site. Maximum storage time for all SWISS units was 41 days.</p><p>We measured soil water vapor samples in the INSTAAR Stable Isotope Laboratory using a continuous flow Picarro L2130-i Isotope and Gas Concentration Analyzer. We measured samples using the dry-air carrier gas method to enhance measurement stability; see Supporting Information S1 for greater detail <ref type="bibr">(Havranek et al., 2023)</ref>. Water vapor sample values were converted to liquid water values using a correction specific to the vapor permeable membrane tubing <ref type="bibr">(Rothfuss et al., 2013)</ref>. We then corrected for isotope scale compression, and report all &#948; 18 O w and &#948; 2 H of soil water (&#948; 2 H w ) values as the per mil (&#8240;) deviation relative to VSMOW/SLAP. Finally, an offset correction of 1.0 and 2.6&#8240; for &#948; 18 O w and &#948; 2 H w , respectively, associated with the storage and measurement of water vapor via the SWISS, was applied <ref type="bibr">(Havranek et al., 2023)</ref>. For all soil water samples, we report an uncertainty of 0.9 and 3.5&#8240; for &#948; 18 O w and &#948; 2 H w , respectively. Both the offset correction and assessment of uncertainty are based on a series of experiments in which water vapor of known composition was introduced into the SWISS units, stored for 14 days, and then measured <ref type="bibr">(Havranek et al., 2023)</ref>.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="4.">Results</head></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="4.1.">Meteorological Data</head><p>For each site below, we focus on changes in soil conditions at a depth of 50 cm, because for two of the three sites pedogenic carbonate nodules are first observed near that depth. Further, many paleosol carbonates are collected roughly from this depth (30-50 cm) because of considerations about atmospheric CO 2 , radiative heating of the soil surface, evaporative enrichment of oxygen isotopes, and so on.</p><p>In Alliance, NE, soil temperature varies sinusoidally throughout the year, and at 50 cm depth, soil temperature peaks in August (Figure <ref type="figure">3a</ref>). At the ONG site, soil moisture varied sinusoidally through the year at depths &#8805;50 cm,</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>Paleoceanography and Paleoclimatology</head><p>10.1029/2024PA005071</p><p>and was not responsive to intense rain events like soil moisture at depths of 10 and 25 cm. Precipitation is summer biased, with peak precipitation amounts in June (Figure <ref type="figure">S1a</ref> in Supporting Information S1).</p><p>Soil temperature varies sinusoidally at the Briggsdale site and maximum soil temperatures at 50 cm depth are reached in August (Figure <ref type="figure">3a</ref>). Volumetric soil moisture at a depth of 50 cm varied widely from December 2020 to October 2022 between 0.05 and 0.22 m 3 /m 3 (Figure <ref type="figure">3b</ref>). Precipitation at this site is summer biased, and is typically highest in May (Figure <ref type="figure">S1b</ref> in Supporting Information S1). At the Seibert site, soil temperature varies sinusoidally and at 50 cm depth, peaks in August (Figure <ref type="figure">3a</ref>). At soil depths of 25 cm and below, soil moisture varies sinusoidally (Figure <ref type="figure">3b</ref>), while at a depth of 10 cm, soil moisture is responsive to snow melt and rain fall events (Figure <ref type="figure">3b</ref>). Precipitation dominantly falls in the summer months, with the most precipitation in May and July (Figure <ref type="figure">S1c</ref> in Supporting Information S1).</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="4.2.">Soil and Carbonate Nodule Characterization</head><p>The soil at the ONG site has a silt loam texture (fine-grained) (Figure <ref type="figure">S2</ref> in Supporting Information S1), and has an argillic horizon, where there is an accumulation of smectite clays. The presence of smectitic clays allows for Paleoceanography and Paleoclimatology 10.1029/2024PA005071</p><p>cracking that, in the dry season, extends to the surface of the soil. In the argillic horizon, soil peds displayed small slickensides, consistent with shrink-swell behaviors associated with the presence of smectite clays. The clay size fraction reaches a maximum in the Bt2 horizon, at 24% (Figure <ref type="figure">1c</ref>), and carbonate formation in this soil begins below the Bt2 horizon. Depth to carbonate formation at this site is 52 cm, with the presence of nodules starting at 75 cm, and the carbonate bearing horizon continues to the bottom of the described pit. Two nodules from the ONG site that were collected at depths of 75 and 100 cm yield 14 C dates of 7,220 &#177; 20 and 8,765 &#177; 25 radiocarbon years BP, respectively.</p><p>The upper portion of the soil profile at the Briggsdale site (&lt;45 cm depth) has a loamy sand (moderately coarsegrained) texture and the lower portion of the soil profile has a sandy loam (medium-grained) texture (Figure <ref type="figure">S2</ref> in Supporting Information S1). Clay particle size fraction in the soil profile slowly increases throughout the profile, and the clay size fraction reaches a maximum in the Bk horizon, at 9.8% (Figure <ref type="figure">1c</ref>). Depth to carbonate formation is 50 cm, and nodule formation begins at the same level. Two carbonate nodules from the Briggsdale site that were collected at depths of 64 and 84 cm yield 14 C dates of 6,980 &#177; 20 and 8,200 &#177; 25 radiocarbon years BP, respectively.</p><p>The uppermost 30 cm of the soil at the Seibert site has a sandy loam (medium-grained) texture, and below 30 cm has a silty loam (fine-grained) texture (Figure <ref type="figure">S2</ref> in Supporting Information S1). The clay size fraction reaches a maximum of 12% in the Bk horizon, at a depth of 90 cm (Figure <ref type="figure">1c</ref>). The depth to carbonate formation is the shallowest of the three sites at 30 cm but between 30 and 75 cm carbonate is extremely diffuse and is not present as discrete masses. Carbonate nodule formation starts at a depth of approximately 75 cm. The carbonate nodules at this site also have a slightly different morphology than the other sites -they incorporate large fragments (&gt;250 &#956;m length) of other minerals like feldspar and quartz (Figures S4 and S5 in Supporting Information S1). Two carbonate nodules from the Seibert site collected at depths of 75 and 100 cm depth yield 14 C dates of 9,680 &#177; 30 and 7,655 &#177; 25 radiocarbon years BP, respectively.</p><p>Across all three sites, carbonate was typically non-luminescent to dully luminescent under CL (Figure <ref type="figure">S4</ref> in Supporting Information S1). In the nodules from the ONG site, there were brightly luminescent clasts that had the mottling and shape of earthworm calcite granules <ref type="bibr">(Prud'homme et al., 2018)</ref>. All of the nodules were multimineralic and included clasts of quartz, feldspars and clays (Figure <ref type="figure">S5</ref> in Supporting Information S1).</p><p>Carbonate nodule stable isotope results are presented in Figure <ref type="figure">4</ref> and Table <ref type="table">1</ref>. In general, &#948; 13 C c and &#948; 18 O c are positively correlated, and grouped by field site. Clumped isotope temperatures for all three sites generally cluster Paleoceanography and Paleoclimatology 10.1029/2024PA005071</p><p>between 11.3 &#177; 2.7&#176;C and 22.2 &#177; 2.2&#176;C, except for one nodule from Briggsdale which yielded a temperature of 4.0 &#177; 2.0&#176;C.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="4.3.">Water Isotope Geochemistry</head><p>For all three sites, the integrated monthly precipitation isotope samples fall within the population of nearby precipitation or stream water samples (Figure <ref type="figure">S6</ref>, Table <ref type="table">S2</ref> in Supporting Information S1). In general, at all three sites, soil water from depths of 50 and 75 cm depth cluster together, and there is very little change in isotope value from summer (June-July) into the fall (September to October) (Figure <ref type="figure">5</ref>). At the Briggsdale site, the soil water data from 50 to 75 cm depths in July 2022 fall above the global meteoric water line, while soil water isotope values from September 2022 fall below the global meteoric water line (Figure <ref type="figure">S6</ref> in Supporting Information S1). At both the ONG and Seibert sites, soil water isotope values from 50 to 75 cm depth overlap with the local and global meteoric water lines. For all three sites, soil water isotope values from 25 cm in the fall months were offset to higher values.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="5.">Discussion</head><p>First, we discuss the assumptions associated with measuring modern environmental parameters and comparing those to pedogenic carbonate nodules that likely formed during the Holocene Climatic Optimum. Second, we discuss our confidence in the soil water isotope data. Then, we integrate all of the environmental and isotopic data and discuss the implications these data have for pedogenic carbonate formation mechanisms, and how that can and should inform future paleoclimate work.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="5.1.">Holocene Climate</head><p>One common challenge for modern calibration studies is that we are comparing instrumental data cover the last 1-10 years, while soil carbonates record conditions over 100-1,000 years. We, therefore, briefly explore how and if the climatic differences over the interval that the soil carbonate formed are meaningfully different than our modern observations. The 14 C dates from the carbonate nodules from all three sites yielded dates between 6,980 &#177; 20 and 9,680 &#177; 30 years before present. These dates are consistent with carbonate nodule formation during the Holocene Climatic Optimum, which persisted from &#8764;9 Ka to-&#8764;5Ka <ref type="bibr">(Shuman &amp; Marsicek, 2016;</ref><ref type="bibr">Steig, 1999)</ref>.</p><p>In North America, temperatures during the Holocene Climatic Optimum were 0-1&#176;C warmer than pre-industrial conditions, similar to today (IPCC, 2022; <ref type="bibr">Shuman &amp; Marsicek, 2016)</ref>. Given the relatively small shift in temperatures relative to the current precision of the clumped isotope data (approx. &#177;3-5&#176;C), the modern soil temperature structure is likely an appropriate approximation of soil temperature conditions at the time of carbonate formation. a Calculated using <ref type="bibr">Anderson et al. (2021)</ref>, &#916; 47 = 0.0391 &#177; 0.0004 x 10 6 /T 2 + 0.154 &#177; 0.004. b Calculated using <ref type="bibr">Da&#235;ron et al. (2019)</ref>. c Indicates sample that was treated <ref type="bibr">(Fiebig et al., 2024)</ref>. d Indicates sample that was double-cleaned (See Text S1 for details).</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>Paleoceanography and Paleoclimatology</head></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>10.1029/2024PA005071</head><p>Precipitation decreased in the mid-latitudes in North America during this time, likely as a consequence of a northward shift of westerly storm tracks <ref type="bibr">(Routson et al., 2019;</ref><ref type="bibr">Shuman &amp; Serravezza, 2017)</ref>. Despite the drying, the &#948; 18 O p likely only varied by &#8764;0.5&#8240; <ref type="bibr">(Liu et al., 2014)</ref>. Given the magnitude of uncertainty on our modern soil water oxygen isotope values (&#177;1.0&#8240;), and the soil water oxygen isotope values calculated from a clumped isotope measurement (approx. &#177;0.6&#8240;), modern observations of soil moisture patterns, especially the magnitude of soil water isotope change at carbonate bearing soil depths, is likely a reasonable approximation of soil water during the time of carbonate formation.</p><p>Lastly, one significant challenge about the Briggsdale site is that the land use history of the site over the last 150 years is not well known. The American dustbowl of the 1930s strongly affected this portion of northeastern Colorado, and in this region, generally, the topsoil was over-tilled and often stripped away, which could explain the absence of the mollic epipedon at the site. So, it is possible that the upper horizons at this site have been altered within the last 100 years.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="5.2.">Modern Soil Water Isotope Data</head><p>At ONG, soil water &#948; 18 O w and &#948; 2 H w values largely cluster together, except for two 25 cm depth samples in September 2022; those data are offset to significantly higher &#948; 18 O w and &#948; 2 H w values (Figure <ref type="figure">5a</ref>). The water vapor concentrations for those two samples were also considerably lower than other samples (i.e., &lt;15,000 ppm), coinciding with a time when the volumetric water content (VWC) of the soil fell below 0.1 m 3 /m 3 . The significantly higher uncertainty and non-linearity of isotope value measurements made below 10,000 ppm on Picarro L2130-I instruments could account for these high values, or they could reflect highly evaporated water related to the dry conditions. We cannot determine which is the case, however, the shallower depth of these samples relative to the carbonate bearing depths means the uncertainty of these values has little impact on our interpretations of the carbonate data. In contrast, the &#948; 18 O w and &#948; 2 H w values at carbonate bearing depths (50 and 75 cm) fall within error of the GMWL and do not show significant variability between June and September 2022. This is consistent with other observations of limited soil water isotope variability at depths 50 cm and greater <ref type="bibr">(Oerter &amp; Bowen, 2019;</ref><ref type="bibr">Quade et al., 2018)</ref>. The annual average value of &#948; 18 O p from Newcastle, WY (110 km from the ONG site) is -11.2&#8240;, and our soil water isotope values from 50 to 75 cm show a slight winter bias that is similar to regional groundwater <ref type="bibr">(Jasechko et al., 2014</ref>). There is one potential meteorological explanation for this bias; soil water isotope monitoring was done during a La Ni&#241;a event. It is challenging to know how different the soil water isotopes are year to year, but based on soil moisture data from the Alliance, NE mesonet site, it is likely that the soil was drier than previous years, with less summer precipitation.</p><p>Soil water isotope data at Briggsdale extends July-October 2022 (Figure <ref type="figure">5b</ref>). There are several complexities to this soil water isotope dataset. First, all stable isotope values from 25 cm depth are more uncertain because water vapor concentrations were below 15,000 ppm during measurement due to low soil moisture (&lt;0.1 m 3 /m 3 VWC). Despite having similar &#948; 18 O w values, the samples collected in July 2022 from depths of 50 and 75 cm have &#948; 2 H w values that are 10-15&#8240; higher than for soil water samples collected in September-October; as a result, the samples from July to August fall above the GMWL. Previous work has shown that the &#948; 18 O w values and &#948; 2 H w values from SWISS measurements can be decoupled <ref type="bibr">(Havranek et al., 2023)</ref>. We hypothesize that this decoupling is a result of pressure gradients inadvertently created by introducing dry air into the flasks during the measurement of gas samples that more easily bias hydrogen isotope results (Steen-Larsen &amp; Zannoni, 2024). We therefore interpret the oxygen isotope values from all dates at this site with greater confidence than the &#948; 2 H w . The calculated &#948; 18 O w value from the clumped isotope measurements (Figure <ref type="figure">5b</ref>) overlaps within uncertainty of the measured &#948; 18 O values from 50 to 75 cm depth.</p><p>All of the soil water isotope data from Siebert had sufficiently high water concentrations for robust measurement, with the exception of the September-October 2022 soil water samples from 25 cm depth which had low VWC. Data from 50 to 75 cm depth are relatively invariant across the nearly 4-month timespan. Similarly to the other two sites, the soil water isotope values from this site cluster together, except for the 25 cm depth samples September-October 2022 with low LWC, which again have higher &#948; 18 O w and &#948; 2 H w values. As a result, data from 9/14/22 and 9/28/22 fall off the global meteoric water line. The return to the GMWL on 10/5/22 and 10/12/22 may be a product of the infiltration of precipitation from 9/30/22 and 10/4/22 at this site.</p><p>The overlap of the measured &#948; 18 O w with the GMWL and the observation of very little variation of &#948; 18 O w from summer into the fall at carbonate bearing depths from all three sites (Figure <ref type="figure">5</ref>) is consistent with many observations from the ecohydrology literature. First, fine grained soils have higher dispersivity than coarse-grained soils, and so it is expected that at depths &#8805;50 cm, mobile and immobile pore water is very effectively homogenized <ref type="bibr">(Kabeya et al., 2007;</ref><ref type="bibr">Sprenger, Leistert, et al., 2016;</ref><ref type="bibr">Sprenger, Seeger, et al., 2016)</ref>. Second, one dimensional models of soil water isotopes predict very little temporal variability in soil water isotopes at depths &gt;25 cm. The model predicts low temporal variability because as the thickness of the dry soil layer increases, the evaporation rate decreases as diffusional resistance increases <ref type="bibr">(Barnes &amp; Allison, 1983)</ref>. Finally, recent advances in soil water isotope sampling strategies have enabled the creation of temporal records of &#948; 18 O w and &#948; 2 H w that also show very little variation in isotope values at depths of 50 and 75 cm <ref type="bibr">(Oerter &amp; Bowen, 2019;</ref><ref type="bibr">Seeger &amp; Weiler, 2021)</ref>. Given the lack of change in &#948; 18 O w across the summer-fall at carbonate bearing depths, changes in soil moisture at these depths must be driven by processes that produce minimal isotope effects such as hydraulic</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head>Paleoceanography and Paleoclimatology</head><p>10.1029/2024PA005071 redistribution, root water uptake, preferential flow, and hydrodynamic dispersion <ref type="bibr">(Barnes &amp; Allison, 1983;</ref><ref type="bibr">Hillel, 2004;</ref><ref type="bibr">La Follette et al., 2023;</ref><ref type="bibr">Orlowski et al., 2015;</ref><ref type="bibr">Sprenger, Leistert, et al., 2016)</ref>.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="5.3.">Pedogenic Carbonate Formation</head><p>One reason to study the impact of grain size and clay content on soil carbonate formation is because dewatering proceeds quickly in coarse-grained soils, whereas, in fine-grained and clay-rich soils, dewatering proceeds at much slower rates <ref type="bibr">(Gaur &amp; Mohanty, 2016;</ref><ref type="bibr">Hillel, 2004)</ref>. As a result, in fine-grained and clay-rich soils, soil drydown continues later into the fall season in summer-dominated precipitation regimes. It is logical, therefore, that if soil dewatering is the primary driver of carbonate mineral supersaturation, there may be a grain-size bias in preserved clumped isotope temperatures <ref type="bibr">(Kelson et al., 2020)</ref>. Here, we examine the hypothesis of soil dewatering as the primary driver of carbonate supersaturation in fine-medium grained soils from two perspectives: T (&#916; 47 ) and &#948; 18 O w .</p><p>At the Briggsdale and Siebert sites, carbonate clumped isotope temperatures overlap with mean annual soil temperature (MAST) (Figure <ref type="figure">3a</ref>). This result is consistent with other studies of fine to medium grained soils <ref type="bibr">(Kelson et al., 2020)</ref>. A temperature approximately equivalent to MAST points to net carbonate formation in either the fall or spring season. Carbonate formation in the spring would suggest that carbonate mineral formation is being driven primarily by Ca 2+ ion delivery by an advective flux of water from the surface; carbonate formation in the fall would be more consistent with a combination of soil-dry down and a concomitant decrease in soil pCO 2 as the driving mechanisms of carbonate formation. At both the Briggsdale and Siebert sites, the calculated value of &#948; 18 O w from clumped isotope measurements overlaps within uncertainty of modern measurements of &#948; 18 O w (Figures <ref type="figure">5b</ref> and <ref type="figure">5c</ref>), and the lack of seasonal variability in &#948; 18 O w at carbonate bearing depths means that soil water isotopes on their own cannot be used to distinguish between either Ca 2+ ion delivery or soil dewatering with decreasing pCO 2 as the driver of carbonate supersaturation. Geochemical modeling has demonstrated that infiltration commonly results in net dissolution, and so it is unlikely that net carbonate precipitation occurs during the spring <ref type="bibr">(Huth et al., 2019)</ref>. At Briggsdale, data gaps don't allow us to evaluate the style of dry-down at &#8805;50 cm depth in the fall. However, it is typical at these depths that soil moisture variability is somewhat sinusoidal. At Seibert, we observe a sinusoidal change in soil moisture, and the low variability in overall soil moisture at depths &#8805;50 cm (Figure <ref type="figure">3b</ref>) signals that if dewatering is the ultimate driving mechanism in fine grained soils, then the degree of variation needed to reach super saturation is very small. Other studies have emphasized the role of soil CO 2 in driving pedogenic carbonate formation in arid soils <ref type="bibr">(Breecker et al., 2009;</ref><ref type="bibr">Oerter &amp; Amundson, 2016)</ref>, as well as its role in conjunction with soil moisture <ref type="bibr">(Huth et al., 2019)</ref>, and so an important next avenue of research in these, and other fine grained soils, would be to simultaneously monitor soil pCO 2 at &#8805;50 cm <ref type="bibr">(Huth et al., 2019)</ref> to constrain the magnitude of variability of that parameter.</p><p>In contrast to the other two sites, at the ONG site, clumped isotope temperatures are &#8764;10&#176;C higher than MAST and overlap with soil temperatures at carbonate bearing depths between late-June and October (Figure <ref type="figure">3a</ref>). The calculated &#948; 18 O w is similar in value to late spring -summer &#948; 18 O p (Figure <ref type="figure">5a</ref>). At face value, the T(&#916; 47 ) and &#948; 18 O c data together require that pedogenic carbonate formed during a period of peak soil temperatures either in equilibrium with summer precipitation or with evaporatively 18 O-enriched soil water. However, the calculated &#948; 18 O w is &#8764;6&#8240; higher than the measured &#948; 18 O w at carbonate bearing depths during the summer-fall (Figure <ref type="figure">5a</ref>). Throughout the period of peak soil temperatures, the measured &#948; 18 O w at carbonate bearing depths (&#8805;50 cm) is more similar to a winter (snow-melt) or early spring isotope value than a summer precipitation value (Figure <ref type="figure">5a</ref>). Moreover, the soil moisture data do not indicate that there either is an infiltration pulse to depths &#8805;50 cm during the summer that would accommodate the delivery of isotopically heavier water, or enough evaporation to drive a 6&#8240; change in &#948; 18 O w (Figure <ref type="figure">3b</ref>). The in situ soil water isotope data and soil moisture data, therefore require that we consider alternative carbonate formation mechanisms.</p><p>One potential hypothesis to explain the clumped isotope data from the ONG site is that cracking of the soil from the shrinking of 2:1 clays caused rapid soil CO 2 degassing, which in turn caused pedogenic carbonate to form out of isotopic equilibrium with the DIC pool, creating anomalously high T(&#916; 47 ) and &#948; 18 O w <ref type="bibr">(Affek &amp; Zaarur, 2014;</ref><ref type="bibr">Tripati et al., 2015)</ref>. There are two ways to test this hypothesis. The first is to do paired soil organic matter and pedogenic carbonate &#948; 13 C analysis, which could indicate if pedogenic carbonate is forming in equilibrium under very high soil CO 2 (water-logged) conditions, or under low-productivity conditions <ref type="bibr">(Monta&#241;ez, 2013)</ref>. The second way to test this hypothesis is with dual-clumped isotope thermometry (Bajnai et al., 2020; Fiebig Paleoceanography and Paleoclimatology 10.1029/2024PA005071 <ref type="bibr">et al., 2021)</ref>. This tool compares the temperature calculated from the mass-48 CO 2 isotopologue with the temperature generated simultaneously from mass-47 to assess if a T(&#916; 47 ) value reflects isotopic equilibrium (i.e., the &#916; 47 and &#916; 48 temperatures are the same) or is the result of kinetic processes such as rapid CO 2 degassing <ref type="bibr">(Bajnai et al., 2020;</ref><ref type="bibr">Fiebig et al., 2021)</ref>. There were two reasons we did not evaluate &#916; 48 in this study; first, the precision of the data in study are not high enough to evaluate &#916; 48 , and second, very small amounts of contamination can drive large variation in mass-48 <ref type="bibr">(Fiebig et al., 2024)</ref>, and this was of extra concern in modern soils near (&lt;200 m) from active agricultural sites. A second potential explanation for the mismatch between the observed and calculated &#948; 18 O w is the sorption of atmospheric vapor to clay minerals during deep cracking, leading to disequilibrium between the total soil water pool and the liquid water that carbonate forms in equilibrium with. While our data don't allow us to rigorously evaluate this hypothesis, the influence of vapor sorption is minimal above gravimetric water contents greater than 10%, and at the time of measurement, volumetric water content at this site was between 8% and 14% <ref type="bibr">(Oerter &amp; Bowen, 2017)</ref>. Overall, the high clumped isotope temperature and calculated &#948; 18 O w are challenging to explain without one of these mechanisms for disequilibrium and points to a need for future work at this site, and in other carbonate bearing clay-rich soils. Data from the ONG site indicate the need to include and consider soil water isotope datasets in future modern calibration efforts.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="5.4.">Implications for Paleoclimate Studies</head><p>One important implication of this study for the paleosol carbonate proxy is that this study substantiates previous findings that in fine grained and clay-rich soils, pedogenic carbonate clumped isotope temperatures can reflect mean annual air temperatures <ref type="bibr">(Kelson et al., 2020)</ref>. This calls into question previous attempts to estimate seasonality based on T(&#916; 47 ) of paleosol carbonates from fine-grained, clay-rich soils, as these studies assumed a warm season bias of carbonate formation in their seasonality estimates (e.g., <ref type="bibr">Hyland et al., 2018;</ref><ref type="bibr">Snell et al., 2013)</ref>. This work also highlights the utility of creating clumped isotope records of ancient examples of climate change (e.g., Paleocene-Eocene Thermal Maximum, middle Miocene Climatic Optimum, etc.) to understand phenomena such as polar amplification and terrestrial amplification of warming, because fine grained, clay rich soils may give us a view into mean change, not just seasonal change.</p><p>The second implication of this work is that in fine grained and clay-rich soils, at carbonate bearing depths, the &#948; 18 O w has a value that is set by the largest pulse of precipitation. In many parts of the western United States, the largest pulse of precipitation is often during the late spring from a combination of snow melt and larger spring season precipitation events. There is very little seasonal evolution of &#948; 18 O w at depths &gt;50 cm after the value is initially set <ref type="bibr">(Oerter &amp; Bowen, 2019;</ref><ref type="bibr">Seeger &amp; Weiler, 2021;</ref><ref type="bibr">Stern et al., 1999)</ref>, and our data is consistent with these previous studies. Most paleoclimate studies using paleosol carbonates assumed that T(&#916; 47 ) and the calculated &#948; 18 O w have the same seasonal bias, but the lack of variation in measured &#948; 18 O w through the year leads to a decoupling between the season that T(&#916; 47 ) records (fall) and the precipitation season recorded (spring). These data support the need for thoughtful interrogation of the dominant precipitation season in the geologic past to interpret both T(&#916; 47 ) and the calculated &#948; 18 O w , and consideration of the role of snowmelt in ancient soils under different climatic regimes <ref type="bibr">(Hudson et al., 2024)</ref>.</p><p>A final consideration regards the conventional wisdom over the last 30 years that the &#948; 18 O w calculated from paleosol carbonate T(&#916; 47 ) values are most often evaporatively 18 O-enriched relative to the precipitation that sourced the water. Another implication of this work is that pedogenic carbonate in fine grained and clay rich soils do not in fact record an evaporatively enriched signal in the calculated &#948; 18 O w . This aligns with recent developments in and observations of triple oxygen isotopes on pedogenic carbonates that show the degree of evaporative enrichment in pedogenic carbonate is highly variable <ref type="bibr">(Kelson et al., 2023)</ref>. More modern calibration work is needed to understand what controls if and when carbonate records evaporatively enriched soil water so that &#948; 18 O w can be reliably used for paleoclimate and paleoaltimetry applications as well as to calibrate isotopically enabled global circulation models in the past and future <ref type="bibr">(Campbell et al., 2024;</ref><ref type="bibr">Winnick et al., 2015)</ref>. This conclusion also supports the hypothesis that we can also design experiments to exploit soil depth to target evaporative signals more deliberately <ref type="bibr">(Kelson et al., 2023)</ref>.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="6.">Conclusions</head><p>In this study, we present data on carbonate formation in three modern soils from Colorado and Nebraska, USA. These three soils span soil textures from loamy sand (medium grained) to silty loam (fine grained) with up to 24%</p><p>Paleoceanography and Paleoclimatology 10.1029/2024PA005071 clay. To understand the timing of pedogenic carbonate formation in these soils, we compared modern observational environmental data, including observations of soil water isotopes and historical climate data from three mesonet sites, with pedogenic carbonate clumped isotope temperature and &#948; 18 O w data.</p><p>Clumped isotope temperatures from pedogenic carbonate overlap with MAST at both sites in Colorado, but clumped isotope temperatures are 8-11&#176;C warmer than MAST at the most clay-rich field site in Nebraska.</p><p>Similarly, the &#948; 18 O w calculated from the clumped isotope temperatures overlap within uncertainty of the measured &#948; 18 O w from carbonate bearing depths at both sites in Colorado. In contrast, at the Nebraska site, the calculated &#948; 18 O w is about 6&#8240; higher than observations of &#948; 18 O w at carbonate bearing depths. At all three sites, observed &#948; 18 O w was nearly invariant from summer-fall, and overlapped with spring &#948; 18 O p . These data support the interpretation that, in most cases, in fine-grained, clay-rich soils it is appropriate to interpret pedogenic carbonate clumped isotope temperatures as representative of mean annual soil temperatures. The combination of soil water isotope data and clumped isotope data also demonstrate that, in fine-grained soils, pedogenic carbonate from &#8805;50 cm may record a &#948; 18 O w value that is set by the largest pulse of precipitation and has very little summer evaporative enrichment.</p></div><note xmlns="http://www.tei-c.org/ns/1.0" place="foot" xml:id="foot_0"><p>25724525, 2025, 2, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2024PA005071 by Universit&#228;tsbibliothek T&#252;bingen, Wiley Online Library on [25/04/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License</p></note>
		</body>
		</text>
</TEI>
