Topographic Thresholds and Soil Preservation along the Southern High Plains Eastern Escarpment, Northwest Texas, USA

The eastern escarpment of the Southern High Plains (USA) is today a semi-arid erosional landscape delineated by canyon breaks and topographic relief. A series of buried soils were identified, described, and sampled at 19 soil profile localities exposed along terraces of the South Fork of the Double Mountain Fork of the Brazos River (South Fork) and two associated tributaries (Spring Creek and Macy 285 drainage). Radiocarbon dating revealed late-Pleistocene to early Holocene (~12,580–9100 14C B.P.), middle-Holocene (~6025–4600 14C B.P.), and late-Holocene (~2000–800 14C B.P.) buried soils. The late-Pleistocene to middle-Holocene soils were preserved only at higher elevations within the upper section of the South Fork and Spring Creek. A topographic position analysis was conducted using GIS to identify and examine the impacts of a soil topographic threshold on the preservation and distribution of buried soils within this geomorphic system. Above the identified ~810 m threshold, lateral migration of channels was constrained. Extensive channel migration below the threshold removed older terraces that were replaced with late-Holocene terraces and associated buried soils. Landscape topography constraints on geomorphic processes and soil formation impacted the preservation of archaeological sites in this semi-arid region.


Introduction
The threshold concept is commonly used for discerning the impact of geomorphic processes on landscape evolution [1][2][3]. Thresholds represent the point on the landscape in which the system of erosion or deposition changes [3]. Common threshold-based models involve the influences of topography [4] or erosion [1,5,6]. Topographic thresholds are typically identified through GIS analysis of digital elevation models DEMs [2,5]. How topographic thresholds effect gully formation, erosion, and stream dynamics has been of particular interest worldwide [7][8][9].
The use of thresholds to examine soil genesis, soil geomorphology, and the preservation of those soils across the landscape contributes to the understanding of the dynamics of landscape development and evolution [10][11][12]. Delineating thresholds across landscapes [13,14] provides a model for examining differences in soil formation processes, age of buried soils, and soil preservation. This landscape approach is particularly important in determining the relationship between buried soils and past hunter-gatherer occupations across the landscape [15,16].
Geomorphic systems are affected by thresholds. Geomorphic thresholds involve intrinsic and extrinsic aspects [17]. The geomorphic thresholds concept is one of abrupt landform change and that the landscape is not always in balance or equilibrium [17] (p. 93). The concept regards unstable landforms and that landform instability is inherent. Abrupt changes in erosion and deposition are part Headward erosion of the Southern High Plains' eastern escarpment during the Pleistocene has differentially exposed Triassic-age mudstones, sandstones and conglomerates (Dockum Group) [24], and the gravels, aeolian sediments, and Caprock caliche layer that comprises the Miocene to Plioceneage Ogallala Formation within the research area [25,26]. In contrast to the flat Southern High Plains, below the eastern escarpment, the breaks are rough, broken land formed from incision and erosion of the differentially resistant Ogallala Formation sediments and Triassic sandstone bedrock [24,27]. In the research area, the breaks are drained by numerous tributaries of the upper Brazos River basin.
The Post research area consists of ~335 km 2 of ranchland that encompasses portions of the Southern High Plains uplands, escarpment breaks, and westernmost Rolling Plains. The modern climate is continental and semi-arid. Precipitation varies greatly on an annual basis, ranging from 130 mm to 1010 mm, but with a mean annual precipitation of 480 mm. Precipitation occurs mainly as spring and summer thunderstorms producing short episodes of heavy rain that can cause extensive erosion [28,29]. Groundwater springs emanating from the escarpment breaks remain a constant source of water feeding the tributaries [30].
The South Fork of the Double Mountain Fork of the Brazos River (hereafter South Fork) flows from northwest to southeast through the research area (  The Post research area consists of~335 km 2 of ranchland that encompasses portions of the Southern High Plains uplands, escarpment breaks, and westernmost Rolling Plains. The modern climate is continental and semi-arid. Precipitation varies greatly on an annual basis, ranging from 130 mm to 1010 mm, but with a mean annual precipitation of 480 mm. Precipitation occurs mainly as spring and summer thunderstorms producing short episodes of heavy rain that can cause extensive erosion [28,29]. Groundwater springs emanating from the escarpment breaks remain a constant source of water feeding the tributaries [30]. The South Fork of the Double Mountain Fork of the Brazos River (hereafter South Fork) flows from northwest to southeast through the research area ( Figure 2) with numerous spring-fed tributaries. These tributaries provide the main discharge of water and contribute to the formation of the landscape. Past work in the research area examined a cross-section of buried soils from the Southern High Plains surface to the escarpment breaks [22]. Key results were: (1) late-Pleistocene to early Holocene (~11,000 to 8000 14 C B.P.) soils were exposed and occurred predictably at higher elevations near the escarpment and within inset terraces of reentrant valleys; (2) late-Pleistocene and early Holocene soils were absent at lower elevations away from the escarpment; and (3) intensive erosion had occurred during the middle Holocene that resulted in a period of net sediment removal, followed by net sediment storage with buried soils in low order drainages during the late Holocene. A late-Holocene series of soils (~2800 to 600 14 C B.P.) were exposed and predictable within inset terraces at lower elevations. Past work in the research area examined a cross-section of buried soils from the Southern High Plains surface to the escarpment breaks [22]. Key results were: (1) late-Pleistocene to early Holocene (~11,000 to 8000 14 C B.P.) soils were exposed and occurred predictably at higher elevations near the escarpment and within inset terraces of reentrant valleys; (2) late-Pleistocene and early Holocene soils were absent at lower elevations away from the escarpment; and (3) intensive erosion had occurred during the middle Holocene that resulted in a period of net sediment removal, followed by net sediment storage with buried soils in low order drainages during the late Holocene. A late-Holocene series of soils (~2800 to 600 14 C B.P.) were exposed and predictable within inset terraces at lower elevations.

Materials and Methods
Field reconnaissance yielded 19 soil profile localities based on buried soils and sediments in alluvial parent materials and landscape positions. These localities were natural cut-bank exposures of terrace remnants associated with stream channels and floodplains of the South Fork and Macy 285 drainage. Soils were described and sampled using standard soil survey methods [31,32]. Pedon descriptions were recorded to indicate soil morphology and characterize soil properties. They were documented from the surface to the bottom of the Quaternary-age deposit, or until no dateable material was left to relate the age of the landform.

Soil and Sediment Analysis
A combination of particle size, calcium carbonate equivalence (CCE), total organic carbon content (OCC), and thin-section analyses were conducted at 16 of the 19 profiles to confirm field descriptions and to delineate episodes of parent material deposition.
Particle size descriptions were determined by sieving and the pipette method (NRCS method 3A1a1; [31,32]). Samples were pretreated to remove organic matter, and then dispersed in solution.

Materials and Methods
Field reconnaissance yielded 19 soil profile localities based on buried soils and sediments in alluvial parent materials and landscape positions. These localities were natural cut-bank exposures of terrace remnants associated with stream channels and floodplains of the South Fork and Macy 285 drainage. Soils were described and sampled using standard soil survey methods [31,32]. Pedon descriptions were recorded to indicate soil morphology and characterize soil properties. They were documented from the surface to the bottom of the Quaternary-age deposit, or until no dateable material was left to relate the age of the landform.

Soil and Sediment Analysis
A combination of particle size, calcium carbonate equivalence (CCE), total organic carbon content (OCC), and thin-section analyses were conducted at 16 of the 19 profiles to confirm field descriptions and to delineate episodes of parent material deposition.
Particle size descriptions were determined by sieving and the pipette method (NRCS method 3A1a1; [31,32]). Samples were pretreated to remove organic matter, and then dispersed in solution. Sand fractions (2-0.05 mm) were separated by wet sieving and dry sieving into their respective sand size classes: very coarse (2-1 mm); coarse (1-0.5); medium (0.5-0.25 mm); fine (.25-0.125 mm); and very fine (0.125-0.05 mm) [26]. Coarse fragments (>2 mm) were reported by a visual volume estimate determined in the field. Total sand fraction (0.05-2.0 mm) was used to analyze soil stratigraphy in order to eliminate the influence of translocated clay during pedogenesis.
The CCE has been determined using a total modified pressure method outlined by Soil Survey Staff [33] adapted by Clark and Hudnall [34]. The method has been modified to prevent the oxidation of organic matter and ensure complete digestion of calcium carbonate, limestones, and dolomites. This method utilizes a Tensimeter to measure the pressure generated in a sealed container as a result of a reaction of the carbonates present in soil with hydrochloric acid (HCl). The reaction that takes place is: The OCC is determined by the loss on ignition method [33]. Loss on ignition is a taxonomic criterion for organic soil materials [33]. Additional laboratory analysis of total carbon and total nitrogen has been determined by a LECO truspec CN analyzer.

Thin-Section Analysis
Samples were prepared for the petrographic microscope ensuring that intact peds were impregnated with an epoxy without disturbing the soil fabric. The impregnated samples were cut into billets before being mounted onto a slide and ground to a thickness of 30 microns. The Geosciences Department at Texas Tech University performed the necessary preparations of the samples. The thin sections were examined using a Nikon Eclipse LV100POL polarizing light microscope equipped with a digital camera and monitor.

Radiocarbon Dating
Seventeen of the 19 localities have been sampled for radiocarbon dating, with 48 radiocarbon samples taken from the top and bottom 10 cm of each buried soil as feasible ( Table 1). All of the samples were radiocarbon dated at the University of Arizona's AMS laboratory. Both residue and humate fractions have been dated for each sample. Dating both fractions limits the possibility for CaCO 3 contaminants, of particular importance in semi-arid regions, to obtain reliable results [30,31]. The reported age for each buried soil is the soil fraction yielding the oldest date as contamination by younger carbon is more likely than by older carbon [35][36][37][38]. When a fraction from a paired sample could not be dated, the dated fraction is reported but discounted as the older age could not be determined. Both uncalibrated radiocarbon years and calendar years are reported (Table 1). OxCal v4.2.3 using the IntCal 13 atmospheric curve (1 standard deviation) was used to calibrate radiocarbon ages to calendar.

Topographic Threshold Analysis
A topographic position analysis [39,40] was conducted to determine whether a soil system's topographic threshold was identifiable, what that threshold was, and its relationship to landscape position and the late-Quaternary stratigraphy across the study area. The deviation from mean elevation (DEV) raster calculation was used in this analysis as previous studies found it to be a more suitable measurement in heterogeneous landscapes [39]. The DEV "calculates the relative topographic position of the central point (z0), as the difference from the mean elevation divided by the standard deviation of the elevation, within a predetermined neighborhood (R)" [40] (p. 3438). Neighborhood focal statistics were used to create mean and standard deviation DEMs (digital elevation model) with a search radius of 300, 600, and 1200 m. Experimentation with different search radii was important to determine which neighborhood search radius was most appropriate for this analysis [40] (p. 3438). A DEM with 5 m resolution and Grass GIS 7.0 were used in these calculations. Results then were classified into five large-scale landscape morphological classes (ridge > 1.00; upper slope 1.00 to 0.50; middle slope 0.50 to −0.50; lower slope −0.50 to −1.00; and valley < −1.00) from the DEV raster based on Weiss [41]. These values indicated relative topographic variation in the landscape. The distribution of landscape classes was examined to determine their relationship to the distribution of dated soils and to aid in identifying a topographic threshold. Each raster was then visualized in Grass GIS with the 3D module.

Pedological Analysis
Radiocarbon results (Table 1) ranged from 12,580 to 800 14 C B.P., with three soil forming periods reflected in the buried soils. The late-Pleistocene to early Holocene soils dated between~12,580 and 9100 14 C B.P. Middle-Holocene soils dated between~6025 and 4600 14 C B.P. The last soil forming period dated to the late Holocene between~2000 and 800 14 Figures A1-A3). The informal name UU soil is used to identify a soil that formed between~12,580 and 9100 14 C B.P. The UU soil represents a time-transgressive surface and is a marker for the minimum age of late-Pleistocene-to early Holocene-age valley fill deposits. The range in stable carbon isotopes from radiocarbon assays (−15.5% to −17.5% ; Table 1) reflects a mixed grassland of C 3 and C 4 plant communities that formed under cooler and more moist conditions than today.

Pedological Analysis
Radiocarbon results (Table 1) ranged from 12,580 to 800 14 C B.P., with three soil forming periods reflected in the buried soils. The late-Pleistocene to early Holocene soils dated between ~12,580 and 9100 14 C B.P. Middle-Holocene soils dated between ~6025 and 4600 14 C B.P. The last soil forming period dated to the late Holocene between ~2000 and 800 14 C B.P.

Late-Pleistocene to Early Holocene Soils
The oldest soil profiles are located in the upper portion of the South Fork (UU localities 3, 4, and 6; Figures 1-3 and A1-A3). The informal name UU soil is used to identify a soil that formed between ~12,580 and 9100 14 C B.P. The UU soil represents a time-transgressive surface and is a marker for the minimum age of late-Pleistocene-to early Holocene-age valley fill deposits. The range in stable carbon isotopes from radiocarbon assays (−15.5‰ to −17.5‰; Table 1) reflects a mixed grassland of C3 and C4 plant communities that formed under cooler and more moist conditions than today.
A fining upwards sequence in particle size indicated that the vertical accretion of alluvial sediments occurred slowly, keeping pace with the rate of pedogenesis. A relatively thick (~75-100 cm) organic-rich deposit with dark colors (~10YR/7.5YR3/3) characterized the UU soil. Cumulic soil development in organic-rich sediments had a comparatively thick, buried mollic epipedon, with Stage I carbonate development. Visible secondary carbonates were identified throughout the UU soil. Depending on local conditions, these carbonates developed both pre-burial and post-burial. The distribution of carbonates throughout this buried soil suggested that soil welding had occurred. Micromorphological analyses (Figure 4) of this soil show unconsolidated alluvium of predominantly Ogallala Formation sediments. Clastic textures of the sediments reveal a matrix supported, fine, moderately sorted, sub-rounded grains. Quartz is the dominant identifiable grain, with feldspars, lithic fragments, and mica grains present in lower abundances. Lithic fragments, A fining upwards sequence in particle size indicated that the vertical accretion of alluvial sediments occurred slowly, keeping pace with the rate of pedogenesis. A relatively thick (~75-100 cm) organic-rich deposit with dark colors (~10YR/7.5YR3/3) characterized the UU soil. Cumulic soil development in organic-rich sediments had a comparatively thick, buried mollic epipedon, with Stage I carbonate development. Visible secondary carbonates were identified throughout the UU soil. Depending on local conditions, these carbonates developed both pre-burial and post-burial. The distribution of carbonates throughout this buried soil suggested that soil welding had occurred.
Micromorphological analyses (Figure 4) of this soil show unconsolidated alluvium of predominantly Ogallala Formation sediments. Clastic textures of the sediments reveal a matrix supported, fine, moderately sorted, sub-rounded grains. Quartz is the dominant identifiable grain, with feldspars, lithic fragments, and mica grains present in lower abundances. Lithic fragments, however, occur in a higher abundance than the other identified buried soils. Pedogenic fine-grained calcite occurs as~8-12% of the total sample, and is identified generally as filling void spaces. Detrital carbonate grains show signs of weathering as evident by quartz inclusions and diffuse boundaries. Clay films are identified on grain surfaces. The clay films are oriented, but non-continuous, covering~15-20% of grain surfaces. Clay films coat grain surfaces, with carbonate cement occurring after the development the clay films. Carbonates are identified mainly filling void spaces. Clay films generally do not occur where the majority of carbonates fill void spaces.
Geosciences 2020, 10, x FOR PEER REVIEW 10 of 38 however, occur in a higher abundance than the other identified buried soils. Pedogenic fine-grained calcite occurs as ~8-12% of the total sample, and is identified generally as filling void spaces. Detrital carbonate grains show signs of weathering as evident by quartz inclusions and diffuse boundaries. Clay films are identified on grain surfaces. The clay films are oriented, but non-continuous, covering ~15-20% of grain surfaces. Clay films coat grain surfaces, with carbonate cement occurring after the development the clay films. Carbonates are identified mainly filling void spaces. Clay films generally do not occur where the majority of carbonates fill void spaces.

Middle-Holocene Soil
Middle-Holocene-age soils are documented in the upper section of Spring Creek (Macy Locality 10, Figure A4) at the confluence of Spring Creek and Macy Fork, and at two South Fork terrace localities (UU localities 4 and 6; Figures 2, 3, A2 and A3). ULS (U-Lazy-S) is used here as an informal name to correlate middle-Holocene soils across the research area. These middle-Holocene soils form at the end of alluvial sedimentation and were buried by colluvium or aeolian deposition. This truncated boundary suggests increased aridity after middle-Holocene pedogenesis. These soils range in age from ~6025 to 4600 14 C B.P. (Table 1).
At the UU localities, the absence of clay films suggests an increase in drying during soil development. The presence of carbonates and their accumulation is derived from the aeolian sand deposited on top. Soil organic matter and stable isotope values (−16.6‰, −17.5‰, and −17.6‰) indicate that prior to burial, a middle-Holocene grassland environment was present over the landscape. That grassland has carbon contributions derived from both C3 and C4 plant communities.
Macy Locality 10 ( Figure A4) represents a ponded alluvial deposit likely developed by slower flowing water. Clay films are present, and carbonate accumulations were secondary features from translocated carbonates. This situation suggests soil welding to some degree. Carbonate

Middle-Holocene Soil
Middle-Holocene-age soils are documented in the upper section of Spring Creek (Macy Locality 10, Figure A4) at the confluence of Spring Creek and Macy Fork, and at two South Fork terrace localities (UU localities 4 and 6; Figures 2, 3, A2 and A3). ULS (U-Lazy-S) is used here as an informal name to correlate middle-Holocene soils across the research area. These middle-Holocene soils form at the end of alluvial sedimentation and were buried by colluvium or aeolian deposition. This truncated boundary suggests increased aridity after middle-Holocene pedogenesis. These soils range in age from 6025 to 4600 14 C B.P. (Table 1).
At the UU localities, the absence of clay films suggests an increase in drying during soil development. The presence of carbonates and their accumulation is derived from the aeolian sand deposited on top. Soil organic matter and stable isotope values (−16.6% , −17.5% , and −17.6% ) indicate that prior to burial, a middle-Holocene grassland environment was present over the landscape. That grassland has carbon contributions derived from both C 3 and C 4 plant communities.
Macy Locality 10 ( Figure A4) represents a ponded alluvial deposit likely developed by slower flowing water. Clay films are present, and carbonate accumulations were secondary features from translocated carbonates. This situation suggests soil welding to some degree. Carbonate accumulation increases with depth, indicating a zone of illuviation associated with pre-burial. The carbon isotope values indicate C 4 vegetation populated the landscape.

Late-Holocene Soils
Soils and landscapes during the late Holocene develop through cumulic processes as denoted by the remaining 15 soil profile localities. These soils are expressed through a multistory or stacked vertical sequence (pedocomplex) rather than a singular buried soil (Figures 5 and A5, Figures A6-A16). Variations delimited at profile localities influence the number of buried soils and their characteristics during this formational period. Characteristics of the multistory sequence represent a marker for this period of landscape stabilization during the late Holocene and the soils that developed. The informal name of South Fork pedocomplex is used to correlate late-Holocene soils across the research area.
Geosciences 2020, 10, x FOR PEER REVIEW 11 of 38 accumulation increases with depth, indicating a zone of illuviation associated with pre-burial. The carbon isotope values indicate C4 vegetation populated the landscape.

Late-Holocene soils
Soils and landscapes during the late Holocene develop through cumulic processes as denoted by the remaining 15 soil profile localities. These soils are expressed through a multistory or stacked vertical sequence (pedocomplex) rather than a singular buried soil (Figures 5 and A5-A16). Variations delimited at profile localities influence the number of buried soils and their characteristics during this formational period. Characteristics of the multistory sequence represent a marker for this period of landscape stabilization during the late Holocene and the soils that developed. The informal name of South Fork pedocomplex is used to correlate late-Holocene soils across the research area. The Macy 285 drainage soil profile localities ( Figure 2) provide a transect through sloping landforms leading to the main South Fork channel. As no discernable differences in sediment characteristics or timing of soil formation were identified, the Macy 285 drainage and lower South Fork terraces are grouped together.
A major soil forming period occurred from ~2000 to 800 14 C B.P. The valley surface was abandoned as a frequently active alluvial depositional environment around ~800-600 14 C B.P., then buried by aeolian and slopewash sediments that blanket the entire landscape. The range in stable carbon isotopes from radiocarbon assays (−14.6‰ to −21.0‰; Table 1) reflected a mixed grassland of C3 and C4 plant communities.
Similar pedogenic features are common through these late-Holocene sediments. The buried soils are in alluvial sediments distinguished by fining upwards sequences that express lateral variation of The Macy 285 drainage soil profile localities ( Figure 2) provide a transect through sloping landforms leading to the main South Fork channel. As no discernable differences in sediment characteristics or timing of soil formation were identified, the Macy 285 drainage and lower South Fork terraces are grouped together.
A major soil forming period occurred from~2000 to 800 14 C B.P. The valley surface was abandoned as a frequently active alluvial depositional environment around~800-600 14 C B.P., then buried by aeolian and slopewash sediments that blanket the entire landscape. The range in stable carbon isotopes from radiocarbon assays (−14.6% to −21.0% ; Table 1) reflected a mixed grassland of C 3 and C 4 plant communities.
Similar pedogenic features are common through these late-Holocene sediments. The buried soils are in alluvial sediments distinguished by fining upwards sequences that express lateral variation of overbank deposition. Soils generally exhibit cumulic soil development through vertical accretion of sediments that kept pace with the rate of pedogenesis. Higher values of organic matter, with darker colors (~10YR/7.5YR 3/3), characterize the soils in the pedocomplex. Localized conditions determine the amount of organic matter input from vegetation, and soil horizons are either mollic or ochric epipedons. Visible secondary carbonates are identified as soft masses and root tracings throughout these soils, with Stage I carbonate development. Carbonates have developed both pre-burial and post-burial, with variation of intensities determined by local conditions. The distribution of carbonates through the pedocomplex suggests that soil welding had occurred.
Micromorphological analysis (Figure 4) shows these late-Holocene soils are composed of an unconsolidated mix of alluvium from Ogallala Formation and Triassic sediments. Clastic textures of the sediments reveal matrix supported, very fine, well-sorted, rounded grains. Quartz is the dominant identifiable mineral grain, with feldspars, lithic fragments, and mica grains present in smaller abundances. Pedogenic fine-grained calcite is relatively non-existent. Detrital carbonate grains, however, are more prevalent when compared to the older late-Pleistocene and early Holocene sediments. These detrital carbonate grains do not show a high degree of weathering, demonstrating relatively high relief and sharp boundaries. Occasionally, detrital carbonate grains show signs of weathering as evident by quartz inclusions, however, it is not known if the degree of weather occurred post-deposition. Clay films occur on~5-15% of the grains, evident on grain surfaces. The clay films are oriented, but non-continuous covering~10-15% of grain surfaces. The clay matrix of these late-Holocene-age sediments contains a much higher amount of iron oxide clays, representing sediments sourced with more inputs from reworked Triassic alluvial sediments.

Radiocarbon Dating
The radiocarbon ages for most of the profiles are in chronological order from top to bottom (Table 1). Only five dates are out of sequence, and of those, only two were from separate soil horizons [UU Locality 4 Profile A (CPUU4A-02, 03) and UU Locality 6 Profile A (CPUU6A-03, 04), Table 1]. The other three are from the same soil horizon sampled within 10 cm of each other [Macy Locality 45 Profile B (CPMACY45B-02, 03), Macy Locality 283 (CPMACY283A-01, 02), and UU Locality 3 Profile A (CPUU3A-02, 03)]. In addition, only at Macy Locality 45 Profile B do out-of-sequence radiocarbon ages not overlap within their error ranges (A15783, 1070 ± 80 14 C; A15784, 1800 ± 110 14 C B.P., Table 1). Overlap in most of the out-of-sequence radiocarbon ages suggests continuous pedogenesis during the time-frame to produce overlapping radiocarbon ages within a 20 cm vertical span.
The younger of the out-of-sequence dates at Macy Locality 45 Profile B (A15784.1, 800 ± 110 14 C B.P., Table 1), UU Locality 4 Profile A (A16098.1, 12,120 ± 290/280 14 C B.P., Table 1), and UU Locality 6 Profile A (A16106, 9100 ± 205/200 14 C B.P., Table 1) are discounted because they are younger in age than the dates immediately above them. It is more difficult to introduce older materials to bias the results in comparison to more recent sources of carbon contamination [35].

Topographic Position Analysis
The differences in the neighborhood search radii (300, 600, and 1200 m) used in the analysis had a big impact in landscape classification (Figure 6). At 300 m, much of the Southern High Plains surface was classified as middle slope. This scale, then, masked an overall pattern in landscape classification and was not used. Patterns in landscape classification at 600 m became apparent and more so at 1200 m. These two search radii were used in the analysis. The topographic threshold at which the topography impacted the valley landscape classification is at an elevation of ~810 m between the UU2 and UU3 localities above where Spring Creek meets the South Fork (Figure 7). The slope gradient within the South Fork changes from 0.577 above the topographic threshold to 0.385 below this threshold. The continued valley classification of the South Fork above the ~810 m threshold indicates a higher degree of topographic relief and confined vertical channel incision that typifies the escarpment breaks. Moving down valley at lower elevations, the South Fork channel becomes less confined with more lateral movement. The late-Pleistocene to middle-Holocene soils are found only above this ~810 m threshold associated with the 1200 m search radius valley classification. These older buried soils have a much better chance of preservation with less lateral movement of the South Fork above the ~810 m threshold. At 1200 m, the part of the landscape classified as valley, i.e., within the South Fork and Spring Creek axes, was concentrated within the escarpment breaks and at higher elevations ( Figure 6). At lower elevations away from the breaks along the South Fork axis, the valley classification ceases and instead the South Fork is classified as lower slope. All buried soil profile localities along the South Fork and Spring Creek were classified within the valley class at 600 m. At 1200 m, however, only buried soil profiles at Macy Locality 126 and higher elevations were within the valley classification along the South Fork and Spring Creek axes. Both the 600 m and 1200 m search radius placed the buried soil profiles within the Macy 285 drainage (Figure 6) within the lower to middle slope classes.
The topographic threshold at which the topography impacted the valley landscape classification is at an elevation of~810 m between the UU2 and UU3 localities above where Spring Creek meets the South Fork (Figure 7). The slope gradient within the South Fork changes from 0.577 above the topographic threshold to 0.385 below this threshold. The continued valley classification of the South Fork above the~810 m threshold indicates a higher degree of topographic relief and confined vertical channel incision that typifies the escarpment breaks. Moving down valley at lower elevations, the South Fork channel becomes less confined with more lateral movement. The late-Pleistocene to middle-Holocene soils are found only above this~810 m threshold associated with the 1200 m search radius valley classification. These older buried soils have a much better chance of preservation with less lateral movement of the South Fork above the~810 m threshold.

Discussion
Based on soil stratigraphic records and radiocarbon ages, the alluvial terraces of the South Fork represented a transect through the study area. Nineteen profiles were examined to identify formational characteristics and the effect of topography on the distribution and preservation of buried soils in the divergent Southern High Plains eastern escarpment breaks. Profile localities of a singular buried soil indicated landscape stability persisting through time. Multistory profile localities provided a range of soil development periods separated by episodes of deposition and erosion (K cycle). Localized factors in soil formation determined the number and thicknesses of these buried soils. These soils, however, exhibited similar pedogenic characteristics that were differentiated by time.
Radiocarbon dating the buried soils produced a range of burial times from ~12,580 to 700 14 C B.P. ( Table 1). Most of the radiocarbon ages were within chronological order for each profile. Overlaps in radiocarbon age error ranges between samples collected within 20 cm of each other indicated continuous pedogenesis for many of the soil sequences.
Late-Holocene (~2000 to 800 14 C B.P.; Table 1) buried soils were the most common and traceable across both the terraces of the South Fork and the Macy 285 drainage [42]. The predominance of late-Holocene sequences of buried soils also was well documented ~40 km downstream (near Justiceburg; Figure 1) in the westernmost region of the Rolling Plains [43]. In contrast, the UU-age soil (~12,580 to 9100 14 C B.P.; Table 1) and ULS-age soil (~6025 to 4600 14

Discussion
Based on soil stratigraphic records and radiocarbon ages, the alluvial terraces of the South Fork represented a transect through the study area. Nineteen profiles were examined to identify formational characteristics and the effect of topography on the distribution and preservation of buried soils in the divergent Southern High Plains eastern escarpment breaks. Profile localities of a singular buried soil indicated landscape stability persisting through time. Multistory profile localities provided a range of soil development periods separated by episodes of deposition and erosion (K cycle). Localized factors in soil formation determined the number and thicknesses of these buried soils. These soils, however, exhibited similar pedogenic characteristics that were differentiated by time.
Radiocarbon dating the buried soils produced a range of burial times from~12,580 to 700 14 C B.P. ( Table 1). Most of the radiocarbon ages were within chronological order for each profile. Overlaps in radiocarbon age error ranges between samples collected within 20 cm of each other indicated continuous pedogenesis for many of the soil sequences.
Late-Holocene (~2000 to 800 14 C B.P.; Table 1) buried soils were the most common and traceable across both the terraces of the South Fork and the Macy 285 drainage [42]. The predominance of late-Holocene sequences of buried soils also was well documented~40 km downstream (near Justiceburg; Figure 1) in the westernmost region of the Rolling Plains [43]. In contrast, the UU-age soil (~12,580 to 9100 14 C B.P.; Table 1) and ULS-age soil (~6025 to 4600 14 C B.P.) were found in only four of the 19 profiles. These profiles were all located in the upper reaches of the South Fork where the valley margin is much narrower.
The UU and ULS soils were found only above the~810 m threshold (Figure 7). Results from the topographic position analysis found that above this threshold, the surface was classified as a valley, even with larger neighborhood search ranges. This situation indicated that above the threshold, the valley margin is well defined in comparison to below the threshold.
A late-Pleistocene and early Holocene soil (UU soil) developed through cumulic soil processes in the canyon-confined areas in the upper reaches of the South Fork drainage. These processes resulted in the development of a thick, organic-rich soil. Small quantities of alluvium were deposited slowly that allowed soil development to keep pace with alluviation.
A middle-Holocene soil (ULS soil) also is preserved above the threshold within upper South Fork and Spring Creek. These soils represent a period of stability before an erosional event marked by an erosional disconformity and then burial by aeolian or colluvial sediments.
In contrast, downstream morphology of the South Fork drainage revealed a late-Holocene series of multistoried buried soils or a catena (the South Fork pedocomplex). The development of multistoried buried soils along this section of the South Fork channel and Macy 285 drainage was a product of fluctuating rates of alluviation. Accretionary sedimentation periodically slowed or completely ceased, resulting in the formation of a series of stable land surfaces during the late Holocene. These buried soils were composed from various inputs of reworked Ogallala Formation and Triassic sediments and showed less effects from weathering due to more recent deposition than the older UU and ULS soils ( [42]; Figure 8). This catena formed a landscape-wide, late-Holocene pedocomplex.
The topographic positioning index defined valleys predominantly in the upper reaches of the South Fork. More valleys were classified through the neighborhood search radii. Increasing the radii (600 to 1200 m) produced a larger variance in elevations that provided an expanded classification of valley and lower slopes. A soil topographic threshold at~810 m elevation was identified just above the confluence of the South Fork channel and Spring Creek drainage (i.e., moving downstream out of the canyon-confined area). It was at this point where the first initial outcroppings of the Triassic redbeds are exposed (Figure 7). Above this threshold, the valley channel was more confined by bedrock with less lateral movement from the South Fork in comparison to below the threshold when the valley begins to widen.
Sometime prior to~13,000 14 C B.P., the upper reaches of the South Fork incised and created a terrace in which the UU and ULS soils formed. During the later Holocene, erosion removed the late-Pleistocene to Middle-Holocene-age terraces containing the UU and ULS soils below the threshold because of more lateral movement of the South Fork in its less confined valley setting. In contrast, above the threshold, the lack of or minimal lateral migration in the upper reaches preserved the soils and their terraces. That situation would explain why at 1200 m search radius, a valley classification occurs above~810 m elevation but not below that elevation. The South Fork drainage valley axes above the~810 m elevation threshold were more confined. This confinement was due to bedrock and a steeper gradient.
Topographic position analysis is a common technique used to subdivide landscapes into geomorphic classes for research [39] (p. 40). The advantage of this approach is that it provides an objective and replicative method to classify the landscape. The spatial relationship between landscape classes, derived from topographic position analysis, and the age of buried soils has been used to identify a soil threshold. Based on these results, the spatial areal extent of valley classification across the landscape has been used to identify the~810 m threshold within the research area (Figure 7). The~810 m threshold marks a crossover in pedogenesis from the canyon breaks to the lower lying South Fork Valley. Topography in this research represents a primary driving extrinsic factor in soil development effecting the matrix of parent material, and then soil burial and preservation. or soil/landscape preservation at a certain elevation, with the late-Pleistocene to early Holocene soil and middle-Holocene ULS soil restricted to the upper reaches of the drainage system, and late-Holocene soils found in both the upper and lower reaches. The late-Holocene multistoried soils are documented across the lower reaches. Topography drives the differential burial and preservation of the soils. This model is strengthened by the integration of a previous study adjacent to and slightly overlapping the current research area. Murphy et al. [22] examined profiles along Spring Creek, Middle Creek, and smaller unnamed tributaries of the South Fork (Figures 2 and 7). Soils in the upper reach of Spring Creek dating to the The Triassic redbeds highly influenced the South Fork's channel dynamics (Figure 8). The channel changed from highly incised and dissected landforms to broader floodplains with larger amounts of net sediment supply. In the lower portions of the South Fork channel further downstream, the topographic relief was less and resulted in landforms classified as middle and lower slopes. The late-Holocene catena dominated and transformed the landscape. The Macy 285 drainage was intended to provide a transect across the middle and lower slopes outside of the canyon-confined areas higher in the drainage network. Pedological and radiocarbon analyses, however, identified that buried soils in this drainage developed at the same time as landforms associated with the lower, broader channel of the South Fork. This situation further strengthened the topographic soil threshold as occurring upstream.
The general soil threshold model developed from this study indicates a crossover in pedogensis or soil/landscape preservation at a certain elevation, with the late-Pleistocene to early Holocene soil and middle-Holocene ULS soil restricted to the upper reaches of the drainage system, and late-Holocene soils found in both the upper and lower reaches. The late-Holocene multistoried soils are documented across the lower reaches. Topography drives the differential burial and preservation of the soils. This model is strengthened by the integration of a previous study adjacent to and slightly overlapping the current research area.
Murphy et al. [22] examined profiles along Spring Creek, Middle Creek, and smaller unnamed tributaries of the South Fork (Figures 2 and 7). Soils in the upper reach of Spring Creek dating to the early Holocene ranged in age from~10,730 to 8155 14 C B.P. (Figure 7). Late-Pleistocene lacustrine muds were dated from cores, the oldest ages being~22,300 and~15,000 14 C B.P. Applying the threshold model, these soils and muds were found above the~810 m threshold (Figure 7). Middle-Holocene soils documented in the upper reaches of Spring and Middle creeks ranged in age from~6025-4600 14 C B.P. These soils also were found only above the~810 m threshold (Figure 7). Packages of late-Holocene-age sediments and soils were found both below and above the~810 m threshold [22].
Integration of the two datasets demonstrates that the model holds for this segment of the escarpment-plains interface. Specific to this research area, the late-Pleistocene to early Holocene soils documented by Murphy et al. [22] are interpreted as the UU soil. In turn, their middle-Holocene soils are grouped as the ULS soil. It also is clear that the late-Holocene multistoried soils of Murphy et al. [22] are part of the South Fork pedocomplex. These informally named buried soil groupings express and are part of the soil threshold model, each reflecting soil threshold elevation, topographic position, and age.
An important threshold found in stream channels is the transition from upland valley to distal floodplain valleys [44] (p. 53). In upland valley settings, the adjacent hillslopes are coupled to the stream and sediment is removed directly from hillslopes and through the streams. The downstream floodplain below the threshold is decoupled from neighboring hillslopes and is a place of sediment accumulation within a broader floodplain setting.
A wide range of sediment sizes are poorly sorted within sediments in channels within the uplands above the threshold [44]. In contrast, below the threshold, smaller materials are moved downstream and the sediments are finer and well-sorted. In the Post study area, a transition occurs in sediment texture from fine and moderately sorted to very fine and well-sorted sediment matrix from the UU soil to the ULS soil to the late-Holocene soils of the South Fork pedocomplex. In addition, above thẽ 810 m threshold, the sediment matrix is dominated by Ogallala Formation sediments, while below the threshold, soils formed in a mixture of both Ogallala and Triassic sediments [42]. This situation further demonstrates that above the threshold, the channel sediments are coupled with the exposed Ogallala Formation sediments on the adjoining hillsides. Below the threshold where the valley broadens out, a change takes place where sediment storage is both Ogallala and Triassic material. This change indicates that the stream channel in this location is decoupled from the adjoining hillsides.
Archaeological surveys and excavation in the region since 2005 have identified over 350 localities ranging in age from Paleoindian~11,000 14 C B.P. to Anglo-American settlement of the region (AD~1877-1950). Late-Holocene-age localities are recorded both above and below the threshold in the research area. For example, Macy Locality 126 is located at the confluence of Spring Creek and the South Fork below the threshold. Survey and excavation at this site have uncovered a multi-occupational Ceramic-to Protohistoric-age (700-400 14 C B.P) campsite locality [45]. Above the threshold, a Ceramic-age (~2000 to 1000 14 C B.P.) occupation has been documented at Macy Locality 16 [46].
In contrast, Paleoindian to Middle-Archaic occupations have been documented mainly above thẽ 810 m threshold. Paleoindian research has focused on deposits exposed along Spring Creek located above the~810 m threshold. To date, a Clovis projectile point (~11,000 14 C B.P.) comes from the surface (Macy Locality 10), and excavation has revealed a Folsom-age (~10,400 14 C B.P.) bison kill/processing feature (Macy Locality 349). In addition, a Middle-Archaic Bell projectile point (~6000 to~5000 14 C B.P.) found on survey at Macy Locality 10 likely is associated with the ULS soil documented at that locality. Research in Spring Creek also has focused on several Quaternary paleontological localities (e.g., Macy Locality 100) that have provided important insights into late-Pleistocene to early Holocene (~11,500-8000 14 C B.P.) faunal communities [47,48].
Channel migration below the topographic threshold has altered the landscape and removed most Paleoindian to Middle Archaic-age deposits and left behind a late-Holocene record. Above the topographic threshold, constrained channel migration has preserved a Paleoindian-to Middle Archaic-age record. Exceptions are surface late-Paleoindian (~10,000-8600 14 C B.P.) localities that have been found below the threshold. For example, at Macy Locality 15, a chalcedony Plainview point (~10,000 14 C B.P.) associated with a few tools and a basal thinning flake have been recovered [49].
The eastern Southern High Plains escarpment [27], draws of the Southern High Plains [35], and stream valleys in the westernmost Rolling Plains [27,43] contain buried landscapes expressing the late Pleistocene through Holocene geoarchaeological record. A number of Southern Plains studies have identified and examined the lateral distribution of buried soils and past stable landscapes through time [27,43,50]. The Post research area, however, is different than many of the study areas identified on the adjoining Southern High Plains and Rolling Plains because the topographic influences are much higher in comparison. Within the escarpment breaks, spatial and temporal discontinuities reflect the evolution in these landscapes in response to natural formation processes. They generally relate to the conditions present in the regional environment. The escarpment breaks are defined by lateral erosion of the Southern High Plains, and erosion and downcutting by the headwaters of the South Fork. Due to the high topographic relief, the escarpment breaks are a source of sediment that supplies the South Fork.
In the adjacent Southern High Plains and Rolling Plain regions, the late-Pleistocene to middle-Holocene sediments are deeply buried within draws or along older terrace settings. In the research area, these early sediments and associated buried soils (i.e., UU soil and ULS soil), however, are exposed only in the upper reaches within the escarpment breaks. Along the South Fork~40 km downstream from the Post research area, late-Pleistocene through middle-Holocene soils are preserved only in the higher elevations of tributaries feeding into the South Fork ( Figure 1). In addition, terraces of late-Holocene age along the South Fork are well preserved and laterally traceable [43].
The regional soils record of the Southern High Plains eastern escarpment and the westernmost Rolling Plains, then, indicates climate change, with a shift from the cooler and humid late Pleistocene to the arid and semi-arid conditions of the middle and late Holocene. This shift has caused the stripping of late-Pleistocene to middle-Holocene soils throughout the area. Preservation of these older soils on the landscape, then, is dictated by local topography.
Determining a topographic threshold, as revealed in this study, is useful for quantifying the effects of the landscape on the preservation and distribution of buried soils and archaeological sites. Results of this study, may be comparable to other regions with plateaus and steep escarpment breaks like the Southern High Plains. Most researchers examining the effects of topographic thresholds along escarpments, however, have focused on identifying the initiation and rate of gully erosion within these regions [51][52][53][54]. Identifying the effects of a topographic threshold on the distribution and preservation of buried soils in these regions will also be important for developing research and preservation strategies to study and protect associated archaeological sites.

Conclusions
Current research has found a soil topographic threshold at~810 m elevation and a change in the behavior of the South Fork that determines the amount of lateral migration of the South Fork. That change in behavior most likely is caused by the underlying geology and slope. The increased migration of the South Fork below the~810 m elevation threshold has eroded away the late-Pleistocene to middle-Holocene terraces and their UU and ULS soils. In contrast, above the~810 m elevation threshold, these terraces and soils are still intact due to the limited lateral migration within the upper reaches of the South Fork drainage. These areas along the escarpment breaks, then, are a good place to find intact archaeological sites dating to the late Pleistocene through middle Holocene.
The understanding of these spatial and temporal patterns, and processes controlling erosion, is particularly significant in assessing site formation and preservation within semi-arid regions.
The determination of soil thresholds, and particularly topographic threshold, that relates the preservation of buried soils and landforms provides a better understanding of the geographic distribution of the geoarchaeological record. Isolating the parent materials in this study that contributed to the formation of landforms over different time periods has allowed for independent viewing of the topographic influences on landscape preservation. This study provides the foundation or first step towards quantifying erosion and examining any potential geomorphic biases through identification of a soil topographic threshold in this highly modified area.
Author Contributions: Investigation, methodology, and writing-original draft preparation, T.C. Investigation, methodology, supervision, visualization, and writing-editing, S.H. Funding acquisition, investigation, methodology, project administration, resources, supervision, and writing-review and editing, E.J. All authors have read and agreed to the published version of the manuscript.