Previous Article in Journal
Interannual Variability of Sea Ice Dirtiness in the East Siberian Sea Based on Satellite Data
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Geomatics
Volume 5
Issue 4
10.3390/geomatics5040067
geomatics-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Track by Track: Revealing Sauropod Turning and Lateralised Gait at the West Gold Hill Dinosaur Tracksite (Upper Jurassic, Bluff Sandstone, Colorado)

by
Anthony Romilio
1,2,*,
Paul C. Murphey
3,4,
Neffra A. Matthews
5,
Bruce A. Schumacher
6,
Lance D. Murphey
7,
Marcello Toscanini
8,
Parker Boyce
8 and
Zach Fitzner
9
1
School of the Environment, The University of Queensland, St. Lucia, QLD 4072, Australia
2
School of the Veterinary Science, The University of Queensland, Gatton, QLD 4343, Australia
3
Department of Paleontology, San Diego Natural History Museum, 1788 El Prado, Balboa Park, San Diego, CA 92101, USA
4
Paleo Resource Specialists, Paleontological and UAS Consulting Services, 2923 Umatilla Street, Denver, CO 80211, USA
5
Prehistoric Impressions, Lakewood, CO 80227, USA
6
U.S Forest Service, 1617 Cole Blvd, Building 17, Lakewood, CO 80401, USA
7
The Story Speaks, 809 Owl Drive, Louisville, CO 80027, USA
8
Dinosaur Ridge, 16831 W Alameda Pkwy, Morrison, CO 80465, USA
9
Paleo Resource Specialists, 1737 Palisade Street, Grand Junction, CO 81503, USA
*
Author to whom correspondence should be addressed.
Geomatics 2025, 5(4), 67; https://doi.org/10.3390/geomatics5040067
Submission received: 15 September 2025 / Revised: 8 November 2025 / Accepted: 17 November 2025 / Published: 20 November 2025
(This article belongs to the Topic 3D Documentation of Natural and Cultural Heritage)
Browse Figures
Versions Notes

Abstract

Drone photogrammetry and per-step spatial analysis were used to re-evaluate the West Gold Hill Dinosaur Tracksite (Bluff Sandstone, Colorado), which preserves an exceptionally long sauropod pes trackway. Building on earlier segment-based descriptions, we reconstructed the entire succession at millimetre-level resolution and quantified turning and gait asymmetry within an integrated digital workflow (UAV photogrammetry, Blender-based landmarking, scripted analysis). Of 134 footprints previously reported, 131 were confidently identified along a mapped path of 95.489 m that records 340° cumulative anticlockwise reorientation. Traditional end-point tortuosity (direct distance/trackway length; DL/TL) yields a moderate ratio of 0.462, whereas our incremental analysis isolates a fully looped subsection (tracks 38–83) with tortuosity of 0.0001 (DL 0.005 m; TL 34.825 m), revealing extreme local curvature that global (end-to-end) measures dilute. Gauge varies substantially along the trackway: the traditional metric (single pes width) averages 32.2% (wide gauge) with numerous medium-gauge representatives, while footprint-specific (‘incremental’) gauge spans 23.1–71.0% (narrow/medium/wide gauges observed within the same trackway). Our tests for asymmetry quantified that left-to-right paces and steps are longer (p = 0.001 and 0.008, respectively), central trackway width is greater (p = 0.043), and pace angulation is lower (p = 0.040) than right-to-left. Behaviourally, these signals are consistent with right-side load-avoidance but remain speculative (alternative explanations may include habitual laterality, local substrate heterogeneity). The study demonstrates how UAV-enabled, fully digital, sequential analyses can recover intra-trackway variability and enhance behavioural understanding of extinct trackmakers from fossil trackways.

1. Introduction

Unmanned aerial vehicles (UAVs) have transformed dinosaur ichnology by enabling safe, rapid, and repeatable acquisition of georeferenced imagery at expansive, intermittently exposed, or otherwise difficult-to-access sites [1,2,3,4]. Complementing ground-based photogrammetry and drone-derived imagery has underpinned 3D modelling of the Broome Sandstone tracksites, where intertidal surfaces are subaerially exposed only a few times per year [5], and has supported comprehensive mapping on steeply inclined, track-bearing horizons in China [6,7]. These applications demonstrate that drone workflows yield high-resolution orthomosaics and digital elevation models adequate for fine-scale, quantitative analyses along entire trackway sequences—precisely the resolution required to assess the West Gold Hill Dinosaur Tracksite (WGHDT), which preserves the purported longest continuous succession of sauropod pes impressions worldwide [8,9].
Situated within the Junction Creek Member of the Bluff Sandstone (Colorado), the WGHDT preserves an extensive trackway of more than 130 footprints attributed to a single sauropod that records a near-complete (~360°) directional change and alternating long and short steps. Previous work focused on five acetate-traced segments (two sequences of three tracks, one of four, one of five, and one of nine) [8], an approach that lacked the spatial continuity needed to resolve fine-scale behavioural variation along the full trackway.
Here, we present a full-coverage analysis of the continuous WGHDT trackway, integrating drone photogrammetry, spatial modelling, and custom Python (v. 3.11.7) scripting in Blender (v. 4.1) to derive conventional ichnological metrics (e.g., pace and stride lengths) alongside non-traditional parameters such as side-symmetry, incremental gauge, localised tortuosity, and polar plots [10,11]. This spatially explicit reanalysis overcomes the methodological constraints of earlier segmented studies and advances ichnological practice for quantifying dinosaur trackmaker behaviours.

2. Location and Geology

Until 2024, when it was acquired by the U.S. Forest Service, the WGHDT was situated on privately owned land and had been known to locals since the late 1950s. The site lies at an elevation of 2835 m near the terminus of the Silvershield Trail, close to the town of Ouray in the San Juan Mountains of Colorado (Figure 1A,B). Access requires a steep and strenuous hike of approximately 3.2 km each way.
The track-bearing surface is an orangish-brown, near-horizontal, approximately 5-metre-thick silicified sandstone locally referred to as the ‘Lower Quartzite’ [12]. The surface exhibits Pleistocene glacial polishing that enhances the visibility of the footprints [8]. The sandstone is well-indurated, fine- to medium-grained, and typically well-sorted; primary sedimentary textures are variably preserved where polishing is subdued [13,14]. Sedimentologically, the unit reflects deposition on intermittently wet sand flats in an overall arid-to-semi-arid system. Subsequent subaerial exposure created firm-ground conditions suitable for track formation, while later diagenetic silicification enhanced the rock’s competence and weathering resistance, promoting long-term preservation of the track-bearing surface [15].
This unit has been variously correlated with the basal Tidwell Member of the Morrison Formation [12,14], the Junction Creek Member of the Morrison Formation [16], and the Junction Creek Member of the Bluff Formation [17,18]. The geological setting was reviewed by Goodell et al. [8], who placed the track-bearing horizon at the top of the Junction Creek Member of the Bluff Sandstone. In their framework, the Junction Creek is overlain by the Recapture Member, and the Bluff Sandstone as a whole overlies the Summerville Formation. Based on lithostratigraphic context, the WGHDT is regarded as Late Jurassic in age, most likely Oxfordian (Figure 1C).
The Junction Creek Member of the Bluff Sandstone records a mosaic of fluvial, overbank, lacustrine and locally aeolian facies deposited on seasonally dry floodplains [15]. Within this framework, the WGHDT sandstone—quartz-rich, well-sorted and cross-bedded—most plausibly represents sand-flat to marginal-aeolian deposition subject to intermittent wetting. Periodic inundation generated low-amplitude symmetrical ripples, and subsequent subaerial exposure created firm-ground conditions suitable for footprint formation and long-term preservation [15].
Geomatics 05 00067 g001
Figure 1. Geographic and stratigraphic context of the West Gold Hill dinosaur tracksite. (A) Regional map of the western United States showing the location of the tracksite (yellow dot) in southwestern Colorado; (B) aerial view of Ouray, Colorado, with the West Gold Hill tracksite positioned on the slopes to the west of the township. (C) Generalised stratigraphic column showing the sedimentary succession exposed at the tracksite, comparing formation and member-level units adapted from previous studies [8,19]. The track-bearing horizon indicated by the sauropod silhouette occurs within the Junction Creek Sandstone Member of the Bluff Sandstone. Map images: copyright Google Earth.
Figure 1. Geographic and stratigraphic context of the West Gold Hill dinosaur tracksite. (A) Regional map of the western United States showing the location of the tracksite (yellow dot) in southwestern Colorado; (B) aerial view of Ouray, Colorado, with the West Gold Hill tracksite positioned on the slopes to the west of the township. (C) Generalised stratigraphic column showing the sedimentary succession exposed at the tracksite, comparing formation and member-level units adapted from previous studies [8,19]. The track-bearing horizon indicated by the sauropod silhouette occurs within the Junction Creek Sandstone Member of the Bluff Sandstone. Map images: copyright Google Earth.
Geomatics 05 00067 g001

3. Materials and Methods

3.1. Digital Mapping of the Track Surface

Photogrammetry was employed to document and visualise the topography of the track surface at the West Gold Hill Dinosaur Tracksite (WGHDT), following established workflows [20,21,22]. Prior to imaging, the site was cleared of debris, drained (for the most part) of water, and allowed to (mostly) dry. A DJI Mavic Enterprise (M3E) drone equipped with an integrated digital camera (5280 × 3956 px, 12.29 mm focal length; 3.36 × 3.36 µm pixel size) was flown manually at an average altitude of 4.67 m under natural light. The flight path comprised continuous passes across the 1680 m2 target area, maintaining a slow velocity (<1 m/s). Image capture was conducted in 2 s interval mode using the M3E’s mechanical shutter (exposure 1/1000 s), which prevents the rolling-shutter effect. The UAV did not pause between photographs, minimising vibration and motion-blur artefacts.
A total of 925 images were captured, including 275 nadir and 650 oblique photographs to ensure complete imaging of deep tracks. Images were colour-balanced in Adobe Lightroom Classic prior to processing. Structure-from-Motion (SfM) reconstruction was performed in Agisoft Metashape Professional Edition (v. 2.1.2, build 18548) [23,24,25]. Both nadir and oblique photographs were aligned within a single processing chunk. The alignment produced 7,323,943 initial tie points, subsequently reduced to 1,191,047 high-quality tie points following an error-reduction workflow (Reconstruction Uncertainty = 15, Projection Accuracy = 30, Reprojection Error = 0.29, main/tiepoint_weighting_factor = 0.085). Image aligment resulted in 513,824 successful image-to-feature projections. Alignment parameters included Accuracy = High; Generic preselection = Yes; Reference preselection = Source; Key point limit = 80,000 (1000 per Mpx); Tie point limit = 0; Optimisation parameters = f, b1, b2, cx, c_y, k1–k4, p1, p2; Fit additional corrections = Yes. A final mean reprojection error of 0.161 px was achieved, with the optimised camera-location error averaging 8.8 cm (7.3 cm X, 4.2 cm Y, 2.7 cm Z).
Spatial accuracy was refined using a DJI RTK 2 base station and two Propeller AeroPoint 2 PPK check points (WGS 84; EPSG:4326). Four calibrated scale bars validated internal scaling, giving a total internal error of 0.000246 m. Horizontal (RTK + PPK) accuracy was within the manufacturer’s stated tolerances; a vertical offset of ~5 m was detected between AeroPoint and RTK heights, but this did not affect relative measurements or x–y georeferencing.
Depth maps were generated at High quality with Aggressive filtering. The final dense point cloud contained 186,076,621 points (16.3 points cm−2; 2.48 mm point spacing). Extraneous vegetation points were removed manually. A digital surface model (DSM) was constructed on High quality without mesh generation, yielding a final grid resolution of 2.48 mm/pixel and ground sampling distance of 1.41 mm/pixel. The polygonal mesh comprised 27.4 million faces and 13.7 million vertices, textured at 8192 × 8192 px. For the orthomosaic, only the nadir imagery was used to avoid edge distortion from oblique views. The resulting model provides millimetre-scale relative precision and centimetre-scale absolute accuracy across the 1.68 × 103 m2 track surface.
The resulting 3D model was imported into MeshLab (version 2023.12) [26], where it was rotated and re-centred to align the track surface with the XY plane of the spherical coordinate system. While the track surface was topographically flat in the field, it was not horizontally level in 3D space due to the natural tilt of the exposure. This reorientation ensured that the surface was positioned at a 0° inclination (i.e., XY plane), facilitating a consistent nadir perspective for visualisation and subsequent analyses. Visualisations including diffuse, ambient occlusion, and false-colour elevation images, generated using ParaView (version 5.10.1) [27] through a series of filters (Figure 2).

3.2. Digital Landmarking and Analysis

A false-colour elevation image of the WGHDT was imported into the open-source software Blender. The image was scale-corrected using the scalebar associated with the image and oriented with north aligned to Blender’s +Y axis. While the original documentation of the WGHDT reported 134 footprints, our analysis could not resolve three footprints: the first, the penultimate, and the final footprint.
Most trackway parameters were derived from a constructed multi-vertex polyline, with vertices manually placed at the centre of each track. These placements were determined by visual inspection of the false-colour elevation image generated from the 3D model, ensuring accurate identification of track landmarks (Figure 3). An exception was to calculate inner and outer trackway widths (ITW and OTW, respectively), where polylines were created with vertices manually placed as medial pes margin and lateral pes margin, respectively. Widths are referenced to the trackway midline where track landmarks are positive values if on their respective side of the trackway (i.e., left position on the left side of the trackway, and vice versa) providing a positive value, whereas negative values denote crossover (i.e., left over to the right and vice versa).
Inter-footprint distances and all subsequent calculations for trackway parameters and trackmaker biometrics were performed using custom Python script executed within Blender.

3.3. Trackway Parameter Calculations

For the previously reported track sequence of 134 footprints [8]—with a maximum of 133 paces and 132 measurements each for stride, pace-angulation, and width—eighteen paces (from 133, 13.5%), six strides (from 133, 4.5%), thirteen pace-angulations (from 132, 9.8%), ten outer widths (from 132, 7.6%), and nine inner-widths (from 132, 6.8%) had been reported [8]. In the current study, we did not restrict the evaluation of trackway parameters to specific segments; instead, we took measurements of every footprint along the trackway, thus reducing sampling bias while maintaining sensitivity to variations throughout the entire trackway.
Length measurements (i.e., pace and stride) measured the distance between the point at the start (x1, y1) and the end (x2, y2) of the respective trackway parameter using Equation (1).
Length = ((x2 − x1)2 + (y2 − y1)2)0.5
In dinosaur ichnology, pace length refers to the distance between successive placements of opposite feet (e.g., from the heel of a left footprint to the heel of the next right footprint), whereas step length measures the distance between opposite footprints when aligned parallel to the stride axis [10,11,28]. So, step length focuses on forward progression along the direction of movement, while pace length captures forward and lateral displacement. Notably, prior research [8] labels certain parameters as ‘step’ but correspond exclusively to pace length.
Step length (not to be mistake with pace length) is the distance between successive footprints when measured parallel to the stride, using Equation (2).
Step length = (pace2 − trackway width2)0.5
The path or trackway length was defined as the distance covered by the trackmaker along the trackway trajectory and was calculated by summing step lengths.
Pace angulation (Equation (3)) is the angle between two successive pace lengths. It is widely used as a proxy for trackway gauge: lower values indicate greater lateral spacing between foot placements (wide gauge), whereas values approaching 180° indicate narrowing (narrow gauge). Less commonly reported is the step angle (Equation (4); equivalent to the ‘angle of divergence’ sensu [10]), the angle between the pace and the stride. We regard step angle as more intuitive, because larger values reflect greater distances between foot placements, and vice versa. Although many workers report both pace angulation and step angle as unsigned quantities (0–180°), this convention obscures lateral sign information during crossover intervals. In this study, we adopt a crossover-aware convention, permitting pace angulations > 180° and signed step angles (negative during crossover).
Cos (pace angulation) = ((pace1)2 + (pace2)2 − (stride)2)/(2 × pace1 × pace2)
Cos (step angle) = ((pace1)2 + (stride)2 − (pace2)2)/(2 × pace1 × stride)
The orientation of the stride lengths was also measured both for azimuth and cardinal directions along the trackway.
Trackway width was considered the distance between footprints measured perpendicular to the stride orientation, using Equation (5).
Trackway width = pace × sin (step angle)
The following assessment of gauge used Equation (6) and the established categories [29]: narrow gauge equating to values of 50% and greater, medium gauge indicated by values between 35% and 50%, and wide gauge being 35% or less.
Gauge = 100 × (PW/OTW)
Here, we first calculated gauge using a single representative pes width (PW = 0.2778 m) [8] that we term ‘traditional gauge’. Since the width of the pes track may vary across the trackway, we measured the pes width (PWi) for each footprint i as the distance between the medial and lateral pes margins from the ITW and OTW polylines (respectively). This incrementally measured pes width was used with the associated outer trackway width to assess ‘incremental gauge’.
Building on the observation of an alternating long–short pace sequence reported by earlier research [8,9], we tested for lateralised locomotion by partitioning the entire trackway into right-to-left and left-to-right footfall sequences and comparing standard parameters between these two classes. Here, we assessed: pace length, step length, stride length, trackway width (WAP = central, ITW, OTW), pace angulation, step angle, and gauge (global, incremental). For each, we performed paired two-tailed t-tests across corresponding left and right measurements. Statistical significance was determined based on a two-tailed threshold of p < 0.05, with exact p-values reported in Table 1.
Given that the WGHDT sauropod trackway is notable for its looped pattern, we quantified path deviation using the tortuosity ratio [30,31,32]. This metric is traditionally an expression of the direct-line distance between the first and last footprints (DL) divided by the total trackway length (TL). Values approaching 1.0 indicate near-linear paths, and lower values indicate more sinuous or looping paths.
Applied in this traditional, end-to-end manner, prior research of the WGHDT trackway reported a ‘moderate tortuosity ratio (DL/TL) of 0.62’ [8]. However, we and others [9] recognise that a single global value masks the highly looped portion and underestimates its contribution to the overall movement path.
To better capture these localised variations, we developed an incremental approach using a custom Python script. This method evaluates tortuosity across all possible subsequences of the trackway. Specifically, for every pair of footprints (i, j) with i < j, we calculate the straight-line distance (DLi,j) and divide it by the corresponding trackway path length (TLi,j) to yield a local tortuosity value (Equation (7)).
Tortuosity = DLi,j/TLi,j
This produced a matrix of tortuosity values (in excess of 8000) ranging from short to long path segments (e.g., track 1-to-2, 1-to-3, …, 1-to-113; track 2-to-3, track 2-to-4, …, track 2-to-131, etc.), enabling the identification of where and over what spatial extent the trackway is most curved. The matrix was visualised as a heat map, where each cell represents the tortuosity of a given path segment. A red–white–blue colour gradient was applied, with higher curvature (i.e., lower tortuosity) segments–particularly those near the looped section—appearing as warmer (red) cells, and straighter segments (higher tortuosity) shown in cooler (blue) tones. In doing so, the method preserves sensitivity to the looped portion while remaining directly comparable to the conventional first and last footprint approach.
To visualise directional changes along the trackway, we created a polar plot of stride orientations (azimuths) using a custom Python script within Blender. Each point on the polar plot represents the compass direction of a single stride orientation, calculated from successive pes print positions. The first stride is plotted at the innermost ring, with each subsequent stride plotted at an evenly spaced radial distance outward. The outermost points, therefore, represent the final strides in the trackway. Concentric rings are drawn at every 10 stride intervals to aid visual interpretation, but are not proportional to physical distance. This method allows directional shifts and curvature patterns to be assessed where traditional rose diagrams are unable.

3.4. Trackmaker Biometric Inferences

Previous research attributed the WGHDT trackway to a sauropod trackmaker but did not assign to a specific a body-fossil taxon [8], an approach we also adopt here. Initially, we considered identifying plausible trackmakers based on the Bluff Sandstone’s slightly older age (Oxfordian) relative to the body-fossil-rich Morrison Formation (Kimmeridgian and early Tithonian, with some localities extending into the latest Oxfordian). Within the Morrison Formation, Camarasaurus is the most abundant sauropod (4.96%), followed by Diplodocus (3.00%) and Apatosaurus (2.47%), according to Foster’s census [33]. Biomechanical modelling by Henderson [34] suggested that Diplodocus, with its posteriorly positioned centre of mass, was best suited to narrow-gauge trackways, while taxa with more anterior centres of mass (e.g., brachiosaurids) were stable on wide-gauge trackways. Apatosaurus has also been inferred to have an anterior centre of mass and likely produced wide-gauge trackways [34].
Such approaches include the assumption that of single gauge use from trackmakers, which has been shown to may not always be the case [35,36], and is certainly not the case here (see Section 4). Instead, we used a conventional baseline that does not rely on any specific body-fossil reconstruction. Hip height of the trackmaker was estimated using a standard multiplier four times the footprint length (FL), which for an FL of 0.369 m yielded a hip height of 1.476 m. Here, footprint size (small) used the categories of tiny (FL < 25 cm), small (25 cm < FL < 50 cm), medium (50 cm < FL < 75 cm), and large (PL > 75 cm) [37].
Trackmaker velocity was estimated using Alexander’s [38] original equation (Equation (8)).
v = 0.25 × (g0.5) × (SL0.67) × (HH−1.17)
Here, v is velocity (m/s), g = 9.81 m/s2, SL is stride length (m), and HH is hip height (m). The method assumes that shorter inter-footprint distances indicate slower movement, and longer distances indicate faster movement. Additionally, because a single stride comprises two consecutive steps—one from each foot—the equation presumes that both are of equal length, reflecting symmetrical forward progression between the left and right sides of the body. However, in this study, we observed that steps/paces were not consistent within strides, indicating asymmetrical gait. Since such asymmetry violates the core of these assumptions [10,39], we reported Alexander-derived speeds [38] for completeness but advise caution when interpreting values from segments showing alternating long–short step/pace sequences (see Section 4).

4. Results

4.1. Trackway Parameters

Of the 134 footprints reported by Goodell et al. [8], three—the first, penultimate, and final—could not be confidently identified in the field, in derived photographs, nor in our digital model. At the start of the trackway, there is a broad depression (approximately 1 m wide) that may have been interpreted as the first print. Near the end of the trackway, a pile of rubble may have obscured the penultimate footprint described by Goodell et al., but the location of the final print is free of debris and shows no indication of a footprint. We therefore analyse 131 sequential footprints spanning a path length of 95.489 m. The trackway orientation begins towards the northeast, rotates through to north, then sweeps westward, then south, before continuing to turn and finishing east–northeast. The first and last measured stride orientations are 37° and 57° (respectively) consistent with a cumulative anticlockwise turning of ~340° (Figure 2).
In contrast to the original investigation’s measurement of only segments, we achieved complete trackway coverage: 130/130 for each of the paces and steps, 129/129 of strides, pace angulations, step angles, stride orientations, and measurements each for inner (ITW), central (WAP), and outer (OTW) trackway widths (Figure 4 and Figure 5; Table 1; Supplementary Table S1). Paces varied from 0.520 to 1.609 m (mean 0.901 m; median 0.881 m; SD 0.151 m). Step lengths ranged from 0.218 to 1.520 m (mean 0.737 m; median 0.712 m; SD 0.180 m). Stride lengths ranged from 0.988 to 2.500 m (mean 1.470 m; median 1.416 m; SD 0.251 m).
Patterns of locomotor asymmetry were detected across the entire trackway in forward progression measurements (see Supplementary Table S2). Pace length differed significantly (t = 3.358, p = 0.001), with left-to-right paces longer (mean 0.944 m, SD 0.155 m) than right-to-left paces (mean 0.858 m, SD 0.135 m). Step length showed the same pattern (t = 2.708, p = 0.008), with longer left-to-right steps (mean 0.779 m, SD 0.189 m) than right-to-left steps (mean 0.695 m, SD 0.163 m). By contrast, stride length did not differ significantly between directions (t = −0.922, p = 0.358; left-to-right mean 1.450 m, SD 0.245 m; right-to-left mean 1.489 m, SD 0.248 m).
Based on vertices placed at the centre of each track, pace angulations spanned 75–161° (mean 109°; median 108°; SD 14°), and step angles spanned 10–65° (mean 35°; median 35°; SD 8°). Trackway width (WAP) ranged 0.202–0.734 m (mean 0.502 m; median 0.501 m; SD 0.084 m), whilst when measured from the medial margin, the ITW varied from −0.201 to 0.409 m (mean 0.185 m; median 0.190 m; SD 0.089 m), and those taken from the lateral margin had an OTW ranging between 0.576 and 1.083 m (mean 0.863 m; median 0.863 m; SD 0.095 m).
Patterns of locomotor asymmetry were also detected across the entire trackway in the measurements of lateral placement (see Supplementary Table S2). Trackway width measured from the centre of the tracks (WAP) was greater during left-to-right progression (mean 0.517 m, SD 0.076 m) than right-to-left (mean 0.487 m, SD 0.090 m; t = 2.043, p = 0.043), a pattern mirrored by lower pace angulation in left-to-right sequences (mean 106.769°, SD 13.757°) relative to right-to-left (mean 111.891°, SD 14.326°; t = −2.070, p = 0.040).
Taken together, these findings indicate a modest but consistent asymmetry: during left-to-right progression, the trackmaker adopted slightly longer pace/step lengths and a wider lateral foot placement (larger WAP with lower pace angulation), whereas right-to-left progression was characterised by shorter pace/step lengths and a relatively narrower WAP with higher pace angulation.
When assessing gauge with the traditional approach of using as single pes width, in this case a width of 0.278 m, we found that gauge ranged from 25.7 to 48.3% (mean 32.6%; median 32.2%; SD 3.9; i.e., wide gauge) with no narrow-gauge widths, 25 as medium, and 104 as wide. When measured incrementally, gauge classified 13 as narrow, 84 as medium, and 32 as wide, with incremental values having ranged from 23.1 to 71.0 (mean 40.1; median 39.5; SD 8.3; i.e., medium gauge), based on footprint-specific pes widths that ranged 0.195–0.495 m (mean 0.343 m; median 0.333 m; SD 0.064 m). Neither the traditional (p = 0.069) nor the incremental gauge (p = 0.112) differed significantly in symmetry (Figure 6; Table 2).
For traditional tortuosity, the direct-line distance (DL) between tracks 1 and 131 is 44.108 m and the total trackway length (TL) is 95.489 m, producing a tortuosity index of 0.462. Using incremental tortuosity, the segment between footprints 38 and 83 exhibited the largest disparity between the direct distance (0.005 m) and the cumulative trackway length (34.825 m). This segment corresponded to the lowest tortuosity value within the trackway, calculated as 0.0001, indicative of a completely looped path (Figure 7A; Supplementary Table S3).
The polar plot revealed a gradual and consistent change in movement direction along the sauropod trackway. Initial strides were oriented nearly due north, with subsequent steps progressively veering westward. This anticlockwise deviation formed a smooth arc, indicating a broad turning manoeuvre rather than abrupt direction changes. The absence of erratic shifts in azimuth suggests deliberate, stable locomotion during the turning sequence (Figure 7B).

4.2. Trackmaker Biometrics

Trackmaker biometrics were calculated using the conventional hindfoot length-to-hip height ratio of 4.0, as widely applied in dinosaur ichnological studies (Supplementary Table S4). For the WGHDT trackmaker, the estimated hip height was 1.47 m, based on a footprint length of 0.369 m and a footprint length multiplier of 4. Relative stride lengths ranged from 0.7 to 1.7 (mean: 1.0; median: 1.0; SD: 0.2), indicating a consistent walking gait during track registration. Calculated stride velocities varied between 1.75 and 8.26 km/h, with an average of 3.46 km/h (median: 3.20 km/h; SD: 1.07 km/h). A trackmaker with a hip height of 1.47 m is estimated to transition from walking to trotting (i.e., relative stride value of 2) at 10.9 km/h and from trotting to running (i.e., relative stride value of 2.9) at 20.2 km/h.

5. Discussion

5.1. Integrating Original Mapping with Continuous, Centreline Morphometrics

Our study is in agreement with findings of others [8,9] that the WGHDT preserves an extensive single sauropod trackway with a broad change in orientation during the registration of footprints. Our photogrammetric model resolved 131 consecutive pes impressions (three fewer than originally reported) along a path length of 95.489 m, which is slightly shorter than the ~96.3 m trackway length reported previously. Our analysis places the initial orientation as 37° (northeast) and final direction as 57° (east–northeast) that when considering in between has a looped section, represents a net change in direction is 340°, i.e., nearly a full rotation (Figure 2 and Figure 7B). This loop occurs between tracks 38–83, a segment constituting approximately 37.4% of the overall trackway.
The discrepancy between our count (131 footprints) and the 134 reported by Goodell et al. [8] likely reflects methodological differences. Whereas their study describes using tracings of small trackway segments to derive parameters, we conducted a continuous digital reconstruction of the entire trackway using high-resolution photogrammetry. This approach enabled objective footprint identification and reduced the risk of counting artefacts or ambiguous features. In our survey, the first footprint may have been previously inferred from a broad surface depression, while the penultimate print was likely obscured by rubble. The final reported footprint, however, shows no indication of a track in the field or model. These observations highlight the value of digital documentation for reproducible ichnological analyses, particularly in complex or partially obscured sites.
Earlier work assessed selected segments and a restricted suite of parameters [8]. Our whole-trackway analysis is broadly consistent with those measurements, while extending the metrics and resolving additional fine-scale behavioural variation. Specifically, our mean pace (0.901 m) closely matches their reported ‘step’ mean (0.908 m; ‘step’ is equivalent to ‘pace’ in their usage) for selected tracks. Our mean stride length (1.470 m) exceeds their 1.234 m (perhaps reflecting our inclusion of longer intervals within the continuous sequence), and mean pace-angulation is similar (109° versus 114°; Table 1; Figure 4).
Inner trackway-width means are comparable (0.180 m previously versus 0.185 m herein), whereas outer trackway width differs (means of 0.730 m and 0.863 m, for the previous [8] and present study, respectively). Beyond the use of means values, reporting full ranges provides additional insight into limb kinematics and/or substrate effects. For inner width, values cross the midline by up to 0.201 m and extend laterally to 0.409 m (range span 0.608 m). For outer width, the range is 0.576–1.083 m (span 0.507 m; Figure 5).
Prior research [8] assessed gauge using the traditional ratio of a single pes width to outer trackway width and obtained 38%, a value within the medium-gauge class (although they noted it was near the 35% wide-gauge boundary). In that original study [8], the researchers were appropriately cautious about ichnotaxonomic assignment, noting Parabrontopodus (narrow gauge) and Brontopodus (wide gauge) as possibilities; medium gauge, poor preservation, and the absence of manus impressions precluded a confident designation.
Our comprehensive analysis along the entire WGHDT sauropod trackway raises the question whether gauge is sufficiently constant for ichnotaxonomic attribution. Using the traditional gauge metric, we obtained a mean of 32.2% (wide-gauge class) but observed considerable variability, with 25 of 129 samples falling in the medium-gauge class. Measured incrementally, the distribution was still more heterogeneous (13 narrow, 84 medium, 32 wide; n = 129), with incremental gauge spanning 23.1–71.0% (Table 2; Figure 6). Collectively, these results indicate that gauge can vary substantially along a single trackway, cautioning against its use as a sole or decisive ichnotaxonomic criterion.
Most published dinosaur trackways are relatively short segments, limiting opportunities to quantify intra-trackway variability and encouraging reliance on single summary values for gauge. By contrast, the WGHDT trackway is long and continuous, permitting comprehensive, statistically informed evaluation. In this context, we show that gauge is not a fixed attribute but varies substantially along the trackway, just as parameters like pace, step, and stride exhibit within-trackway heterogeneity. While we cannot assume that the magnitude of variability observed at WGHDT occurs across all sauropod trackways, the present analysis illustrates what becomes resolvable when continuous, long trackways are available. Notwithstanding, even relatively short trackways usually comprise multiple consecutive steps, permitting several gauge measurements and at least a preliminary assessment of within-trackway variation.
Comparable intra-trackway gauge variability has been reported elsewhere, with multiple, non-exclusive drivers proposed. At Barkhausen (Germany), Diedrich [35] argued that trackway gauge depends on size and speed, cautioning that gauge is not a reliable ichnotaxonomic discriminator and noting specifically that gauge narrows with faster speeds. Castanera et al. [36] (Berriasian, Spain) likewise document gauge shifts within single trackways and show that the relation between speed and gauge is not uniform: in trackway LCU-I-37, a narrower phase coincides with a directional change and decrease in speed, whereas in SS1-R1, a wider phase accompanies a decrease in speed (the reverse pattern). Castanera et al. [36] further infer an abnormal gait on a deformable substrate for parts of SS1-R1 (overprinting; strong sub-surface deformation), emphasising that substrate effects can modulate apparent gauge independent of limb kinematics as suggested previously [37].
Against this backdrop, the WGHDT ‘loop’ helps decouple local speed peaks from gauge change. Tracks 78–83, at the exit of the loop, record the highest stride lengths in the sequence (Figure 4C), and thus imply locally increased speed under standard stride-based estimators. Yet in this same interval, the traditional gauge shows only a modest shift from wide- to medium-gauge and not the greatest narrowing observed along the trackway (Figure 6A). Likewise, the incremental gauge distribution in tracks 78–83 includes medium and narrow values that are above the mean (40.1%; Figure 6B) but not distinctly separated from the even wider variation documented elsewhere. Taken together, these observations suggest that the WGHDT gauge pattern around the loop is not governed primarily by speed or turning. Instead, substrate heterogeneity as reported by Castanera et al. may play a role.
A further consideration is gait symmetry. As Castanera et al. [36] observed, abnormal gait on deformable substrates can modify apparent gauge independently of true stance width. In the WGHDT trackway, statistically significant left–right differences in pace, step, pace-angulation, and central trackway width are consistent with unilateral load-avoidance (see Section 5.3 below). It is therefore plausible that the interaction of gait asymmetry and local substrate conditions contributed to the gauge variability documented here; however, this remains interpretative rather than demonstrated. Systematic application of the same comprehensive, segment-wise methodology across additional sauropod trackways—both long and short—will be required to test the generality and mechanisms of these effects.

5.2. Resolution of Turning Data

A distinguishing feature of the WGHDT trackway is the looped section that occupies approximately one third of the trackway. The first publication of this trackway [8] reported a turn greater than 180°, but—using the traditional end-to-end tortuosity ratio (DL/TL)—derived a ‘moderate’ value (0.62) across the whole trackway. As noted previously [8,9], those instances where a loop exists, tortuosity values should approach zero, but when restrained through the use of traditional end-to-end tortuosity techniques, it will mask the tightest curvatures that lay between the first and last tracks.
To enhance comparability across sites, Lockley et al. [9] undertook a global assessment of non-linear dinosaur trackways and formalised turning by binning net directional change into 45° increments (e.g., ‘towards half-circle’, ‘beyond three-quarter’, ‘towards full circle’). While valuable, the approach remains primarily descriptive and, by their own discussion, involves subjective choices—whether to reference proximal or distal orientations, or to connect start and end points—because a non-linear trackway has no single, unique orientation.
To identify the location and spatial extent of turning, we employ two complementary, fully sequential digital methods: (1) stride–orientation polar plots; and 2) incremental tortuosity (Figure 7). For (1), we plot the azimuth of each consecutive stride, enabling any trajectory to be quantified and placed in context with its neighbour: for WGHDT, this records 340° of cumulative anticlockwise turn (Figure 2). For (2), incremental tortuosity calculates the end-to-end ratio DL/TL for all subsequences of footprints i…j (i < j), producing over 8000 subsequence ratios (Table S3) that can be visualised as a heatmap to localise curvature (Figure 5). When applied to the looped subsection (tracks 38–83), this analysis gives DL = 0.005 m over TL = 34.825 m (DL/TL 0.0001), exposing the extreme curvature that the traditional, single end-to-end tortuosity masks.

5.3. Behavioural Inferences from Asymmetry Data

Speed estimates of the WGHDT sauropod trackmaker do differ between ours and that of the original assessment. Goodell’s [8] average (2.21–2.58 km h−1) derives from Alexander-based formulae [38] with HH = 4.0–4.45 × FL, and even though we used one of these multipliers (i.e., HH = 4.0 × FL), our assessment across the entire trackway averages higher at 3.46 km h−1. By extending speed calculations across the entire trackway, we calculated that the stride velocities varied from as low as 1.75 to as high as 8.26 km/h. Based on calculations of relative stride (between 0.7 and 1.7, respectively), the trackmaker appears to have been consistently walking.
Although speed estimators are commonly used in dinosaur ichnology, they assume steady level-ground locomotion with symmetrical alternation between limbs—meaning both steps in a stride are of equal length. Under this model, shorter stride lengths suggest slower speeds, and longer strides suggest faster movement. However, in the case of limping or asymmetric gait, these assumptions are violated. Alternating long–short step/pace sequences can arise from load-avoidance behaviour rather than actual variation in forward speed [10,39]. In such instances, shorter steps may not reflect slower movement, but rather a rapid transfer of body weight away from a compromised limb. Consequently, standard speed equations can misinterpret these gaits. While earlier studies at the West Gold Hill Dinosaur Tracksite noted localised asymmetry in selected trackway segments [8], the present study provides the first statistically robust, trackway-wide evidence of alternating step asymmetry and its association with low pace-angulation values. These findings reinforce the need for caution when interpreting speed estimates from asymmetrical trackways,
In practical terms, an injured or load-avoided limb may elicit a locomotor response that limits stance time and propulsion on that side, producing a shorter distance from the injured to the contralateral (non-injured) limb and, conversely, a longer distance from the non-injured to the injured limb. The West Gold Hill trackway exhibits this pattern at trackway scale: left-to-right steps and paces are consistently longer than right-to-left, accompanied by wider lateral placement (larger WAP and lower pace angulation) in the left-to-right direction. Speculatively, if the right limb were compromised at the time of registration, these asymmetries are consistent with right-side load-avoidance and a relative preference for the left limb. We emphasise that this remains a hypothesis; alternative, non-pathological drivers (habitual lateral preference, local substrate heterogeneity, or the left limb consistently serving as the inner limb during an overall left-turning action) have the potential to generate similar signatures. Accordingly, stride-based speed estimates should be interpreted cautiously for this trackway, as asymmetry should be considered when inferring locomotor dynamics.

5.4. Trackmaker Attribution Challenges

The WGHDT preserves an extraordinarily long pes-only sauropod trackway in which we, as in the original study [8], found no evidence of manus impressions. Goodell et al. [8] proposed that this absence reflected pes overprinting of manus traces. However, given the length of the succession and the curved nature of several sections—where manus and pes would have had to register in differing positions to negotiate turns [40,41]—we consider this explanation unlikely. The poor morphology of many tracks further suggests that portions of the WGHDT trackway may represent undertracks.
The occurrence of pes-only trackways can be explained by load partitioning linked to the trackmaker’s centre of mass. A posteriorly situated centre of mass yields a greater hindlimb load fraction, whereas an anteriorly located centre of mass increases forelimb load [34]. Although manus and pes impressions frequently co-occur [40,41,42,43,44,45,46,47,48], simulations demonstrate that the more heavily loaded autopodium applies greater mean pressure, and under certain substrate shear-strength conditions only this autopodium deforms the surface sufficiently to register a track [49]. Such models reproduce manus-only trackways for anteriorly loaded taxa and pes-only trackways for those with posteriorly located centres of mass.
The Oxfordian Bluff Sandstone has not produced sauropod body fossils but is slightly older than the mostly Kimmeridgian and early Tithonian Morrison Formation, which preserves a diverse sauropod fauna dominated by Camarasaurus, Diplodocus, and Apatosaurus [33]. Henderson [34] demonstrated that Diplodocus, with its posteriorly positioned centre of mass, was biomechanically suited to narrow-gauge trackways, whereas taxa with more anterior centres of mass (e.g., brachiosaurids) were stable on wide-gauge trackways. Given the pes-only preservation at WGHDT, a narrow-gauge trackway might be expected; however, our analysis shows that most footprints fall within the wide-gauge class using traditional metrics, or medium gauge using incremental analyses, with only a few footprints classed as narrow gauge.
The original investigation [8] argued that sauropods with posteriorly positioned centres of mass, such as Diplodocus, were biomechanically suited to narrow-gauge trackways, reflecting modest lateral limb excursions and load distributions consistent with manus–pes proportions. By contrast, taxa with anteriorly located centres of mass, such as brachiosaurids, were considered stable only on wide-gauge trackways, exemplified by the ichnogenus Brontopodus. They further proposed that narrow-gauge locomotion represented the primitive sauropod condition, with wide-gauge walking evolving as a functional necessity among very large forms (>12.5 tonnes).
At present, we do not attribute the WGHDT trackway to a specific body-fossil taxon. The relatively small size of the pes impressions indicates that the trackmaker was diminutive by sauropod standards, and it is possible that ontogenetic variation in body proportions, track registration, and gait may have influenced both the formation and gauge classification of the trackway. This lies beyond the scope of the current investigation; however, further work is required to test these possibilities and to evaluate whether ontogenetic factors may have contributed to the unusual pes-only preservation observed at WGHDT.

Supplementary Materials

The following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/geomatics5040067/s1, Table S1: Trackway parameters; Table S2: Asymmetry analyses; Table S3: Tortuosity; Table S4: Trackmaker biometrics. Figure S1: Trackway map–diffuse image; Figure S2: Trackway map–false-colour image.

Author Contributions

Conceptualization, A.R. and P.C.M.; methodology, A.R. and P.C.M.; software, A.R. and P.C.M.; formal analysis, A.R.; investigation, A.R.; resources, A.R. and P.C.M.; writing—original draft preparation, A.R. and P.C.M.; writing—review and editing, A.R., P.C.M., N.A.M., B.A.S., L.D.M., M.T., P.B. and Z.F.; visualisation, A.R. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Data Availability Statement

Data is contained within the article or Supplementary Materials.

Acknowledgments

We thank the anonymous reviewers for their constructive comments. ChatGPT (OpenAI; GPT-5 Thinking) was used solely to assist with code refinement, figure generation workflows, and language editing. No analytical conclusions were accepted without independent verification by the authors, who take full responsibility for the accuracy and integrity of this publication.

Conflicts of Interest

Authors include individuals employed in UAV-based field assessments of palaeontological sites and others are employed by governmental agencies; however, none of the authors received payment, equity, or other financial benefit for this study or for publication of the manuscript. These affiliations did not influence the design, data collection, analysis or interpretation of the results.

Abbreviations

The following abbreviation is used in this manuscript:
WGHDTWest Gold Hill Dinosaur Tracksite
HHHip Height
FLFootprint Length
DLDirect Length/distance between two footprints
TLTrackway Length/distance between two footprints

References

  1. Matthews, N.A.; Noble, T.A.; Breithaupt, B.H. Close-Range Photogrammetry for 3-D Ichnology: The Basics of Photogrammetric Ichnology. In Dinosaur Tracks: The Next Steps; Falkingham, P., Marty, D., Richter, A., Eds.; Indiana University Press: Bloomington, IN, USA, 2016; pp. 29–55. [Google Scholar]
  2. Salisbury, S.W.; Romilio, A.; Herne, M.C.; Tucker, R.T.; Nair, J.P. The Dinosaurian Ichnofauna of the Lower Cretaceous (Valanginian–Barremian) Broome Sandstone of the Walmadany Area (James Price Point), Dampier Peninsula, Western Australia. J. Vertebr. Paleontol. 2016, 36, 1–152. [Google Scholar] [CrossRef]
  3. Petti, F.M.; Petruzzelli, M.; Conti, J.; Spalluto, L.; Wagensommer, A.; Lamendola, M.; Francioso, R.; Montrone, G.; Sabato, L.; Tropeano, M. The use of aerial and close-range photogrammetry in the study of dinosaur tracksites; Lower Cretaceous (upper Aptian/lower Albian) Molfetta ichnosite (Apulia, southern Italy). Palaeontol. Electron. 2018, 21, 1–18. [Google Scholar] [CrossRef]
  4. Murphey, P.C.; Matthews, N.A.; Breithaupt, B.H.; Lockley, M.G. Thirty-six years of tracking and mapping at Dinosaur Ridge, Morrison, Colorado, USA. Vertebr. Paleontol. A Tribut. Martin Lockley New Mex. Mus. Nat. Hist. Sci. Bull. 2024, 95, 327–347. [Google Scholar]
  5. Romilio, A.; Hacker, J.M.; Zlot, R.; Poropat, G.; Bosse, M.; Salisbury, S.W. A multidisciplinary approach to digital mapping of dinosaurian tracksites in the Lower Cretaceous (Valanginian-Barremian) Broome Sandstone of the Dampier Peninsula, Western Australia. PeerJ 2017, 5, e3013. [Google Scholar] [CrossRef]
  6. Xing, L.D.; Lockley, M.G.; Romilio, A.; Klein, H.; Zhang, J.Q.; Chen, H.Y.; Zhang, J.J.; Burns, M.E.; Wang, X.L. Diverse sauropod-theropod dominated track assemblage from the Lower Cretaceous Dasheng Group of Eastern China: Testing the use of drones in footprint documentation. Cretac. Res. 2018, 84, 588–599. [Google Scholar] [CrossRef]
  7. Yang, Q.; Xing, L.; Du, F.; Chen, Q.; Klein, H.; Romilio, A.; Jin, Y.; Liu, L.; Qi, J.; Zhao, M.; et al. A new sauropod-dominated tracksite from the Lower Cretaceous in central Ningxia, northwestern China, and the implications on palaeoenvironments. Swiss J. Palaeontol. 2025, 144, 35. [Google Scholar] [CrossRef]
  8. Goodell, Z.; Lockley, M.G.; Lucas, S.G.; Schumacher, B.A.; Smith, J.A.; Trujillo, R.; Xing, L. A High-Altitude Sauropod Trackway Site in the Jurassic of Colorado: The Longest Known Consecutive Footprint Sequence Reveals Evidence of Sharp Turning Behavior. New Mex. Mus. Nat. Hist. Sci. Bull. 2021, 82, 101–112. [Google Scholar]
  9. Lockley, M.G.; Xing, L.; Kim, K.S.; Meyer, C.A. Tortuous trackways: Evidence and implications of deviations, turns and changes in direction by dinosaurian trackmakers. Hist. Biol. 2021, 33, 3326–3339. [Google Scholar] [CrossRef]
  10. Romilio, A.; Shao, C. Analysing trackway-based speed calculations to infer dinosaur locomotive capabilities and behaviours. Hist. Biol. 2024, 36, 2244–2253. [Google Scholar] [CrossRef]
  11. Romilio, A.; Park, R.; Nichols, W.; Jackson, O. Dinosaur footprints from the Lower Jurassic (Hettangian–Sinemurian) Precipice Sandstone of the Callide Basin, Queensland, Australia. Hist. Biol. 2025, 1–12. [Google Scholar] [CrossRef]
  12. O’Sullivan, R.B. Jurassic Wanakah and Morrison Formations in the Telluride-Ouray-Western Black Canyon Area of Southwestern Colorado; US Government Printing Office: Washington, DC, USA, 1992; Volume 1927, pp. 1–24.
  13. Burbank, W.S.; Luedke, R.G. Geology of the Ouray Quadrangle. In U.S. Geological Survey Geological Quadrangle; U.S. Geological Survey: Washington, DC, USA, 1962; Volume 152. [Google Scholar]
  14. Burbank, W.S.; Luedke, R.G. Geology and Ore Deposits of the Uncom-Pahgre (Ouray) Mining District, Southwestern Colorado; U.S. Geological Survey: Reston, VA, USA, 2008; Volume 1753, p. 107.
  15. Kirkland, J.I.; DeBlieux, D.D.; Hunt-Foster, R.K.; Foster, J.R.; Trujillo, K.C.; Finzel, E. The Morrison Formation and its bounding strata on the western side of the Blanding Basin, San Juan County, Utah. Geol. Intermt. West Utah Geol. Assoc. 2020, 7, 37–195. [Google Scholar] [CrossRef]
  16. Kelley, V.C. Geology of Ouray and environs. In New Mexico Geological Society Guidebook; New Mexico Geological Society: Socorro, NM, USA, 1957; Volume 8, pp. 203–207. [Google Scholar]
  17. Lucas, S.G.; Heckert, A.B. Mesozoic stratigraphy at Durango, Colorado. In New Mexico Geological Society, 56th Field Conference Guidebook, Geology of the Chama Basin, NM, USA, 21–24 September 2005; Lucas, S.G., Zeigler, K.E., Lueth, V.W., Owen, D.E., Eds.; New Mexico Museum of Natural History and Science: Albuquerque, NM, USA, 2005; pp. 160–169. [Google Scholar]
  18. Lucas, S.G. Triassic-Jurassic stratigraphy in southwestern Colorado. In New Mexico Geological Society, Guidebook; New Mexico Geological Society: Socorro, NM, USA, 2017; Volume 68, pp. 149–158. [Google Scholar]
  19. Goldman, M.I.; Spencer, A.C. Correlation of Cross’ La Plata Sandstone, southwestern Colorado. Am. Assoc. Pet. Geol. Bull. 1941, 25, 1745–1767. [Google Scholar]
  20. Falkingham, P.L.; Bates, K.T.; Avanzini, M.; Bennett, M.; Bordy, E.; Breithaupt, B.H.; Castanera, D.; Citton, P.; Díaz-Martínez, I.; Farlow, J.O.; et al. A standard protocol for documenting modern and fossil ichnological data. Paleontology 2018, 61, 469–480. [Google Scholar] [CrossRef]
  21. Romilio, A. An Instructional Guide to Visualising Dinosaur Tracks; Amazon: Columbia, SC, USA, 2020. [Google Scholar]
  22. Lallensack, J.N.; Buchwitz, M.; Romilio, A. Photogrammetry in ichnology: 3D model generation, visualisation, and data extraction. J. Paleontol. Tech. 2022, 22, 1–18. [Google Scholar] [CrossRef]
  23. Szeliski, R. Computer Vision: Algorithms and Applications, 1st ed.; Springer Nature: London, UK, 2010; p. 957. [Google Scholar]
  24. Westoby, M.J.; Brasington, J.; Glasser, N.F.; Hambrey, M.J.; Reynolds, J.M. ‘Structure-from-Motion’ photogrammetry: A low-cost, effective tool for geoscience applications. Geomorphology 2012, 179, 300–314. [Google Scholar] [CrossRef]
  25. Falkingham, P.L. Acquisition of high resolution 3D models using free, open-source, photogrammetric software. Palaeontol. Electron. 2012, 15, 1–15. [Google Scholar]
  26. Cignoni, P.; Callieri, M.; Corsini, M.; Dellepiane, M.; Ganovelli, F.; Ranzuglia, G. MeshLab: An open-source mesh processing tool. In Proceedings of the Sixth Eurographics Italian Chapter Conference, Salerno, Italy, 2–4 July 2008; pp. 129–136. [Google Scholar]
  27. Ahrens, J.; Geveci, B.; Law, C. 36—ParaView: An End-User Tool for Large-Data Visualization. In Visualization Handbook; Hansen, C.D., Johnson, C.R., Eds.; Butterworth-Heinemann: Burlington, ON, Canada, 2005; pp. 717–731. [Google Scholar]
  28. Romilio, A.; Dick, R.; Skinner, H.; Millar, J. Uncovering hidden footprints: Revision of the Lower Jurassic (Sinemurian) Razorback Beds—Home to Australia’s earliest reported dinosaur trackway. Hist. Biol. 2025, 37, 596–603. [Google Scholar] [CrossRef]
  29. Romano, M.; Whyte, M.A.; Jackson, S.J. Trackway ratio: A new look at trackway gauge in the analysis of quadrupedal dinosaur trackways and its implications for ichnotaxonomy. Ichnos 2007, 14, 257–270. [Google Scholar] [CrossRef]
  30. Benhamou, S. How to reliably estimate the tortuosity of an animal’s path: Straightness, sinuosity, or fractal dimension? J. Theor. Biol. 2004, 229, 209–220. [Google Scholar] [CrossRef]
  31. Nams, V.O. Using animal movement paths to measure response to spatial scale. Oecologia 2005, 143, 179–188. [Google Scholar] [CrossRef] [PubMed]
  32. Bustos, D.; Jakeway, J.; Urban, T.M.; Holliday, V.T.; Fenerty, B.; Raichlen, D.A.; Budka, M.; Reynolds, S.C.; Allen, B.D.; Love, D.W.; et al. Footprints preserve terminal Pleistocene hunt? Human-sloth interactions in North America. Sci. Adv. 2018, 4, eaar7621. [Google Scholar] [CrossRef]
  33. Foster, J. Relative abundances of the Sauropoda (Dinosauria, Saurischia) of the Morrison Formation and implications for Late Jurassic paleoecology of North America. Mesa Southwest Mus. Bull. 2001, 8, 47–60. [Google Scholar]
  34. Henderson, D.M. Burly gaits: Centers of mass, stability, and the trackways of sauropod dinosaurs. J. Vertebr. Paleontol. 2006, 26, 907–921. [Google Scholar] [CrossRef]
  35. Diedrich, C.G. Upper Jurassic tidal flat megatracksites of Germany—Coastal dinosaur migration highways between European islands, and a review of the dinosaur footprints. Senckenberg. Lethaea 2011, 91, 129–155. [Google Scholar] [CrossRef]
  36. Castanera, D.; Pascual, C.; Canudo, J.I.; Hernández, N.; Barco, J.L. Ethological variations in gauge in sauropod trackways from the Berriasian of Spain. Lethaia 2012, 45, 476–489. [Google Scholar] [CrossRef]
  37. Marty, D. Sedimentology, taphonomy, and ichnology of Late Jurassic dinosaur tracks from the Jura carbonate platform (Chevenez—Combe Ronde tracksite, NW Switzerland): Insights into the tidal-flat palaeoenvironment and dinosaur diversity, locomotion, and palaeoecology. GeoFocus 2008, 21, 1–278. [Google Scholar]
  38. Alexander, R.M. Estimates of speeds of dinosaurs. Nature 1976, 261, 129–130. [Google Scholar] [CrossRef]
  39. Lockley, M.G.; Hunt, A.P.; Moratalla, J.J.; Matsukawa, M. Limping dinosaurs? Trackway evidence for abnormal gaits. Ichnos 1994, 3, 193–202. [Google Scholar] [CrossRef]
  40. Xing, L.D.; Lockley, M.G.; Marty, D.; He, J.J.; Hu, X.F.; Dai, H.; Matsukawa, M.; Peng, G.Z.; Yong, Y.; Klein, H. Wide-gauge sauropod trackways from the early Jurassic of Sichuan, China: The oldest sauropod trackways from Asia. Eclogae Geol. Helv. 2016, 109, 415–428. [Google Scholar]
  41. Xing, L.D.; Marty, D.; Wang, K.B.; Lockley, M.G.; Chen, S.Q.; Xu, X.; Liu, Y.Q.; Kuang, H.W.; Zhang, J.P.; Ran, H.; et al. An unusual sauropod turning trackway from the early cretaceous of Shandong province, China. Palaeogeogr. Palaeoclim. Palaeoecol. 2015, 437, 74–84. [Google Scholar] [CrossRef]
  42. Lockley, M.G. Dinosaur tracks. Ichnos 1992, 2, 79–83. [Google Scholar] [CrossRef]
  43. Moratalla, J.J. Sauropod tracks of the Cameros Basin (Spain): Identification, trackway patterns and changes over the Jurassic-Cretaceous. Geobios 2009, 42, 797–811. [Google Scholar] [CrossRef]
  44. Myers, T.S.; Fiorillo, A.R. Evidence for gregarious behavior and age segregation in sauropod dinosaurs. Palaeogeogr. Palaeoclim. Palaeoecol. 2009, 274, 96–104. [Google Scholar] [CrossRef]
  45. dos Santos, V.F.; Moratalla, J.J.; Royo-Torres, R. New sauropod trackways from the Middle Jurassic of Portugal. Acta Palaeontol. Pol. 2009, 54, 409–422. [Google Scholar] [CrossRef]
  46. Meyer, C.A. The hitch-hikers guide to the Late Jurassic and Early Cretaceous-Dinosaur tracks from the Swiss and French Jura Mountains in a sequence stratigraphic context. In Dinosaur Track Symposia 2011 Obernkirchen, Abstracts Volume; Richter, A., Hübner, T., van der Lubbe, T., Eds.; Universitätsverlag Göttingen: Göttingen, Germany, 2012; p. 43. [Google Scholar]
  47. Vildoso, C.A.; Calvo, J.O.; Porfiri, J.D. The highest sauropod tracksite in the World, Lower Cretaceous, Peru. In Dinosaur Track Symposia 2011 Obernkirchen, Abstracts Volume; Richter, A., Hübner, T., van der Lubbe, T., Eds.; Universitäts-Verlag Göttingen: Göttingen, Germany, 2011; p. 35. [Google Scholar]
  48. Kim, J.Y.; Lockley, M.G. New sauropod tracks (Brontopodus pentadactylus ichnosp. nov.) from the Early Cretaceous Haman Formation of Jinju area, Korea: Implications for sauropods manus morphology. Ichnos 2012, 19, 84–92. [Google Scholar] [CrossRef]
  49. Falkingham, P.L.; Bates, K.T.; Margetts, L.; Manning, P.L. Simulating sauropod manus-only trackway formation using finite-element analysis. Biol. Lett. 2011, 7, 142–145. [Google Scholar] [CrossRef]
Geomatics 05 00067 g002
Figure 2. The sauropod trackway at the West Gold Hill Dinosaur Tracksite (Upper Jurassic, Bluff Sandstone, CO, USA). Nadir views: (A) diffuse orthophotograph; (B) false-colour digital elevation image, showing elevation from low (blue) to high (red). Arrows indicate the travel direction of the trackmaker, including a looping turn near the centre of the trackway. All images are oriented with north to the top. Scale bar = 30 m.
Figure 2. The sauropod trackway at the West Gold Hill Dinosaur Tracksite (Upper Jurassic, Bluff Sandstone, CO, USA). Nadir views: (A) diffuse orthophotograph; (B) false-colour digital elevation image, showing elevation from low (blue) to high (red). Arrows indicate the travel direction of the trackmaker, including a looping turn near the centre of the trackway. All images are oriented with north to the top. Scale bar = 30 m.
Geomatics 05 00067 g002
Geomatics 05 00067 g003
Figure 3. Segment of sauropod trackway at the West Gold Hill Dinosaur Tracksite (Upper Jurassic, Bluff Sandstone, CO, USA). Nadir views: (A) diffuse orthophotograph; (B) ambient occlusion image; (C–E) false-colour elevation image showing semi-automatic approach to trackway parameter (see text for details). North is oriented to the top. Scale bar = 1 m.
Figure 3. Segment of sauropod trackway at the West Gold Hill Dinosaur Tracksite (Upper Jurassic, Bluff Sandstone, CO, USA). Nadir views: (A) diffuse orthophotograph; (B) ambient occlusion image; (C–E) false-colour elevation image showing semi-automatic approach to trackway parameter (see text for details). North is oriented to the top. Scale bar = 1 m.
Geomatics 05 00067 g003
Geomatics 05 00067 g004
Figure 4. Trackway parameters for the West Gold Hill Dinosaur Tracksite sauropod trackway (Upper Jurassic, Bluff Sandstone, CO, USA). Each panel pairs a sequential scatterplot (left) with a side-specific violin plot (right): (A) pace length (m), (B) step length (m), (C) stride length (m), and (D) pace angulation (°). In scatterplots, grey shading marks the looped section of the trackway (tracks 38–83), and points are colour-coded by foot (left or right). Violin plots summarise the distribution of values for each foot: the width of each shape corresponds to data density (i.e., wider regions indicate a higher concentration of values), with embedded points and lines showing individual measurements and the median, respectively. Left–right values differ significantly for pace, step, and pace angulation (two-tailed paired t-tests), whereas stride shows no significant asymmetry.
Figure 4. Trackway parameters for the West Gold Hill Dinosaur Tracksite sauropod trackway (Upper Jurassic, Bluff Sandstone, CO, USA). Each panel pairs a sequential scatterplot (left) with a side-specific violin plot (right): (A) pace length (m), (B) step length (m), (C) stride length (m), and (D) pace angulation (°). In scatterplots, grey shading marks the looped section of the trackway (tracks 38–83), and points are colour-coded by foot (left or right). Violin plots summarise the distribution of values for each foot: the width of each shape corresponds to data density (i.e., wider regions indicate a higher concentration of values), with embedded points and lines showing individual measurements and the median, respectively. Left–right values differ significantly for pace, step, and pace angulation (two-tailed paired t-tests), whereas stride shows no significant asymmetry.
Geomatics 05 00067 g004
Geomatics 05 00067 g005
Figure 5. Trackway width parameters for the sauropod trackway at the West Gold Hill Dinosaur Tracksite (Upper Jurassic, Bluff Sandstone, CO, USA). Panels A–C pair a sequential scatter plot (left) with a side-specific violin plot (right): (A) central trackway width (m), (B) inner trackway width (m), and (C) outer trackway width (m). In scatterplots, grey shading marks the looped section of the trackway (tracks 38–83), and points are colour-coded by foot (left or right). Violin plots summarise the distribution of values for each trackway width parameter: the violin plot width corresponds to data density (i.e., wider regions indicate a higher concentration of values), with embedded points and lines showing individual measurements and the median, respectively. Left–right differences are significant for central width only (two-tailed t-tests); inner and outer widths show no significant left–right difference. Panel (D) plots central trackway width against stride length (m).
Figure 5. Trackway width parameters for the sauropod trackway at the West Gold Hill Dinosaur Tracksite (Upper Jurassic, Bluff Sandstone, CO, USA). Panels A–C pair a sequential scatter plot (left) with a side-specific violin plot (right): (A) central trackway width (m), (B) inner trackway width (m), and (C) outer trackway width (m). In scatterplots, grey shading marks the looped section of the trackway (tracks 38–83), and points are colour-coded by foot (left or right). Violin plots summarise the distribution of values for each trackway width parameter: the violin plot width corresponds to data density (i.e., wider regions indicate a higher concentration of values), with embedded points and lines showing individual measurements and the median, respectively. Left–right differences are significant for central width only (two-tailed t-tests); inner and outer widths show no significant left–right difference. Panel (D) plots central trackway width against stride length (m).
Geomatics 05 00067 g005
Geomatics 05 00067 g006
Figure 6. Trackway gauge for the sauropod trackway at the West Gold Hill Dinosaur Tracksite (Upper Jurassic, Bluff Sandstone, CO, USA). Each panel pairs a sequential scatter plot (left) with a side-specific violin plot (right). (A) Gauge (%) calculated with the traditional method using a single constant pes width; (B) gauge (%) calculated incrementally using footprint-specific pes widths. In scatterplots, shaded bands denote gauge classes (wide, medium, narrow). Violin plots summarise the distribution of values for each gauge parameter: the violin plot width corresponds to data density (i.e., wider regions indicate a higher concentration of values), with embedded points and lines showing individual measurements and the median, respectively. Left–right symmetry does not differ significantly for either method ((A): p = 0.069; (B): p = 0.112). ‘ns’ denotes not significant.
Figure 6. Trackway gauge for the sauropod trackway at the West Gold Hill Dinosaur Tracksite (Upper Jurassic, Bluff Sandstone, CO, USA). Each panel pairs a sequential scatter plot (left) with a side-specific violin plot (right). (A) Gauge (%) calculated with the traditional method using a single constant pes width; (B) gauge (%) calculated incrementally using footprint-specific pes widths. In scatterplots, shaded bands denote gauge classes (wide, medium, narrow). Violin plots summarise the distribution of values for each gauge parameter: the violin plot width corresponds to data density (i.e., wider regions indicate a higher concentration of values), with embedded points and lines showing individual measurements and the median, respectively. Left–right symmetry does not differ significantly for either method ((A): p = 0.069; (B): p = 0.112). ‘ns’ denotes not significant.
Geomatics 05 00067 g006
Geomatics 05 00067 g007
Figure 7. Turning patterns of the sauropod trackway at the West Gold Hill Dinosaur Tracksite (Upper Jurassic, Bluff Sandstone, CO, USA). (A) Incremental tortuosity heat map where each cell reports tortuosity (direct distance ÷ cumulative trackway length) for the subsequence from start footprint i to end footprint j. While traditional tortuosity yields a single value (first to last prints; equivalent to the final cell in the first row), the incremental approach resolves over 8000 values across this trackway. Colours run from low (red) to high (blue) tortuosity; scale truncated at 0.30+; (B) polar plot of stride orientations along the trackway. Each point represents the azimuth of an individual stride, plotted at evenly spaced radial intervals progressing outward from the centre. The innermost points correspond to the earliest strides, and outermost to the latest. Concentric rings denote every ten strides to aid interpretation (not distance-scaled), illustrating the gradual curvature and turning behaviour of the trackmaker.
Figure 7. Turning patterns of the sauropod trackway at the West Gold Hill Dinosaur Tracksite (Upper Jurassic, Bluff Sandstone, CO, USA). (A) Incremental tortuosity heat map where each cell reports tortuosity (direct distance ÷ cumulative trackway length) for the subsequence from start footprint i to end footprint j. While traditional tortuosity yields a single value (first to last prints; equivalent to the final cell in the first row), the incremental approach resolves over 8000 values across this trackway. Colours run from low (red) to high (blue) tortuosity; scale truncated at 0.30+; (B) polar plot of stride orientations along the trackway. Each point represents the azimuth of an individual stride, plotted at evenly spaced radial intervals progressing outward from the centre. The innermost points correspond to the earliest strides, and outermost to the latest. Concentric rings denote every ten strides to aid interpretation (not distance-scaled), illustrating the gradual curvature and turning behaviour of the trackmaker.
Geomatics 05 00067 g007
Table 1. Trackway parameters and associated symmetry assessment for the West Gold Hill Dinosaur Tracksite sauropod trackway (Upper Jurassic, Bluff Sandstone, CO, USA). ‘L2R’ and ‘R2L’ denote left-to-right and right-to-left, respectively. ‘t’ denotes t-statistic and p the corresponding two-sided p-value; statistical significance is assessed at 0.05.
Table 1. Trackway parameters and associated symmetry assessment for the West Gold Hill Dinosaur Tracksite sauropod trackway (Upper Jurassic, Bluff Sandstone, CO, USA). ‘L2R’ and ‘R2L’ denote left-to-right and right-to-left, respectively. ‘t’ denotes t-statistic and p the corresponding two-sided p-value; statistical significance is assessed at 0.05.
Pace
(m)		Step
(m)		Stride
(m)		Pace
Ang (°)		Step
Angle (°)		Trackway Width
Central (m)		Trackway Width
Inner (m)		Trackway Width
Outer (m)
Mean: 0.9010.7371.470109350.5020.1850.863
Median: 0.8810.7121.416108350.5010.1900.863
(SD 0.151)(SD 0.180)(SD 0.251)(SD 14)(SD 8)(SD 0.084)(SD 0.089)(SD 0.095)
Min–Max: 0.520–1.6090.218–1.5200.988–2.50075–16110–650.202–0.734−0.201–0.4090.576–1.083
L2R: 0.9440.7791.450107340.5170.2000.878
(SD 0.155)(SD 0.189)(SD 0.245)(SD 14)(SD 8)(SD 0.076)(SD 0.088)(SD 0.090)
R2L: 0.8580.6951.489112360.4870.1690.848
(SD 0.135)(SD 0.163)(SD 0.249)(SD 14)(SD 9)(SD 0.090)(SD 0.090)(SD 0.099)
t: 3.3582.708−0.922−2.070−0.9442.0431.9401.799
p: 0.0010.0080.3580.0400.3470.0430.0550.074
Table 2. Gauge and symmetry assessment for the West Gold Hill Dinosaur Tracksite sauropod trackway (Upper Jurassic, Bluff Sandstone, CO, USA). ‘L2R’ and ‘R2L’ denote left-to-right and right-to-left, respectively. ‘t’ denotes t-statistic and p the corresponding two-sided p-value; statistical significance is assessed at 0.05.
Table 2. Gauge and symmetry assessment for the West Gold Hill Dinosaur Tracksite sauropod trackway (Upper Jurassic, Bluff Sandstone, CO, USA). ‘L2R’ and ‘R2L’ denote left-to-right and right-to-left, respectively. ‘t’ denotes t-statistic and p the corresponding two-sided p-value; statistical significance is assessed at 0.05.
Gauge (Traditional)Gauge (Incremental)Pes Width (m)
Mean: 32.640.10.343
Median: 32.239.50.333
(SD 3.9)(SD 8.3)(SD 0.064)
Min–Max: 25.6–48.323.1–71.00.195–0.495
L2R: 32.039.0
(SD 3.5)(SD 6.9)
R2L: 33.341.3
(SD 4.3)(SD 9.5)
t: −1.8−1.6
p: 0.0690.112
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Romilio, A.; Murphey, P.C.; Matthews, N.A.; Schumacher, B.A.; Murphey, L.D.; Toscanini, M.; Boyce, P.; Fitzner, Z. Track by Track: Revealing Sauropod Turning and Lateralised Gait at the West Gold Hill Dinosaur Tracksite (Upper Jurassic, Bluff Sandstone, Colorado). Geomatics 2025, 5, 67. https://doi.org/10.3390/geomatics5040067

AMA Style

Romilio A, Murphey PC, Matthews NA, Schumacher BA, Murphey LD, Toscanini M, Boyce P, Fitzner Z. Track by Track: Revealing Sauropod Turning and Lateralised Gait at the West Gold Hill Dinosaur Tracksite (Upper Jurassic, Bluff Sandstone, Colorado). Geomatics. 2025; 5(4):67. https://doi.org/10.3390/geomatics5040067

Chicago/Turabian Style

Romilio, Anthony, Paul C. Murphey, Neffra A. Matthews, Bruce A. Schumacher, Lance D. Murphey, Marcello Toscanini, Parker Boyce, and Zach Fitzner. 2025. "Track by Track: Revealing Sauropod Turning and Lateralised Gait at the West Gold Hill Dinosaur Tracksite (Upper Jurassic, Bluff Sandstone, Colorado)" Geomatics 5, no. 4: 67. https://doi.org/10.3390/geomatics5040067

APA Style

Romilio, A., Murphey, P. C., Matthews, N. A., Schumacher, B. A., Murphey, L. D., Toscanini, M., Boyce, P., & Fitzner, Z. (2025). Track by Track: Revealing Sauropod Turning and Lateralised Gait at the West Gold Hill Dinosaur Tracksite (Upper Jurassic, Bluff Sandstone, Colorado). Geomatics, 5(4), 67. https://doi.org/10.3390/geomatics5040067

Article Metrics

Back to TopTop