Where we are working
Ventura Basin evolution
Point Loma Fm submarine lobe deposits
Rosario Formation submarine levee deposits
Cloridorme Fm. basin plain turbidites
Chinle Fm. meander belt architecture
Atoka Fm. turbidite deposits
Bone Spring Fm. carbonate slope deposits
Machine learning from core photos, Schiehallion field, UK
Lewis Shale petrophysics machine learning
The dimensions and stacking patterns of depositional systems are records of past Earth-surface change and control reservoir heterogeneity in the subsurface. CoRE investigates stratigraphic architecture from outcropping ancient systems using traditional (mapping, logging section) and advanced (XRF data, digital outcrop models) methodologies. A major focus is on submarine depositional environments, but we also have active projects in fluvial and eolian environments.
Lateral Variability of Event-Beds in Submarine Depositional Environments
A major focus for CoRE is to understand the fine-scale heterogeneity of turbidite beds. Conceptual models of facies variability in channels (axis-to-margin) and lobes (axis-to-fringe) are often qualitative and are rarely quantified at the event‐bed scale. Our work quantifies the ‘thinning rates’ of sand and mud beds in different environments to learn how sediment gravity flows react to seafloor topography. In particular, understanding the lateral and vertical changes of low net-to-gross event-bed deposits is important for reservoir modeling and lateral facies predictions when drilling horizontal wells.
Rosie Fryer’s MS project focused on quantifying event-bed thinning rates using a large data compilation – this paper is published in open-access journal The Depositional Record.
Point Loma Formation
Rosie Fryer, M.S. (2018) also studied submarine lobe deposits in the Point Loma Formation (San Diego, CA), constraining event-bed architecture in medial-to-distal lobe deposits. This work has been pre-printed here, but is yet to be published in a journal.
Kaci Kus’ MS project is currently focused on nearby deposits of the Point Loma Formation, where sand beds are thinner and more laterally variable.
Other Active projects in Stratigraphic Architecture
Carbonate Slope deposits in the Delaware Basin
Wylie Walker (M.S., 2019) studied the Bone Spring Formation in the Delaware Basin to constrain the partitioning of siliciclastic and carbonate sediment in a muddy slope environment. This work is in review in Geosphere, and is pre-printed here.
Eolian Bounding surface architecture
In collaboration with Mary Carr, Evan Gross (MS student) is constraining the architecture of 1st and 2nd order flow-inhibiting bounding surfaces in a wet eolian system. We pair traditional field methods with advanced technologies (e.g, high precision GPS data and drone-derived outcrop models) to help resolve the detailed surface morphologies and the three-dimensional (3D) stratigraphic architecture of the Middle Jurassic Entrada Sandstone that crops out on Rone Bailey Mesa, 60 km south of Moab, Utah.
Scaling relationships are very useful for predicting scales and geometries of depositional elements – for example, the size of a river is determined by its catchment area. CoRE investigates these scaling relationships for submarine depositional systems, utilizing both seafloor and outcrop data to better constrain the dimensions and architecture of turbidite event-beds and the hierarchically larger elements like submarine channels and lobes.
Current projects (2019):
Scaling relationships in submarine channel deposits
Lauren Shumaker (CoRE postdoc) recently published a paper in Geosphere (link to open-access paper) demonstrating for the first time a rigorous scaling between channel width and depth, and explored how aspect ratios change with channel scale and thalweg gradient.
Scaling relationships in submarine channel-lobe deposits
Luke Pettinga (PhD student) recently published a paper in Geology (link to open-access paper) demonstrating that lobes proportionally scale to their concomitant channels, and thus to the volume of sediment supplied prior to an avulsion.
Avulsion criteria for rivers and submarine channels
Zane Jobe and colleagues have discovered fundamental differences in the way that rivers and submarine channels evolve, controlling the rate at which they avulse. This paper is currently in review in Frontiers in Earth Science, but can be accessed on EarthArxiv.
Sediment dispersal patterns
We utilize detrital zircon geochronology, heavy mineral analysis through automated mineralogy, and sandstone petrography to constrain sediment flux and sedimentary basin evolution. Most applicable for exploration-scale reservoir quality questions, although can sometimes be used to differentiate production-scale sand-bodies.
Current projects (2019):
Ancient sediment routing into the Ventura strike-slip basin, California
Clark Gilbert (PhD student) is using detrital zircon and mineral abundance mixture modeling to determine the strike-slip history along the San Gabriel fault. The relative abundances of Proterozoic, Triassic, Jurassic and Cretaceous age peaks in the Ventura Basin sediments record the erosion of distinct source terranes in the San Gabriel Mountains as the basin is translated northward along the fault.
Detrital provenance of isolated marine sand bodies, Western Interior Seaway, Colorado
Luke Pettinga (PhD student) is investigating spatial and/or temporal variability in provenance signatures of isolated Cretaceous marine sandbodies in Colorado to elucidate how they relate to interpreted sediment sources from fluvial-deltaic sediments of the Book Cliffs in Utah.
Sediment routing in the Clare Basin, Ireland
Zane Jobe is collaborating with Martin Nauton and Shane Tyrrell (Univ. Galway) to investigate the effects of climate and tectonic forcing on sediment maturity in the Clare Basin using detrital zircon geochronology and heavy mineral abundance data.
Machine Learning for facies prediction
CoRE is also focused on providing efficient use of large datasets created by the other themes, with the aim to use deep-learning models to help predict facies variability in subsurface deposits. Ross Meyer, CoRE data scientist, and Thomas Martin, PhD student, are instrumental in this effort.
Current projects (2019):
Automated Interpretation of Depositional Environments Using Graphic Logs
Zane Jobe is overseeing the digitization of >50,000 turbidite event-beds in order to build a model that can be used to classify core descriptions from unknown environments and assess uncertainty for a given interpretation.
Automated Lithology Prediction from Core Images and Well Log Data
Thomas Martin (PhD student) and Ross Meyer (CoRE data scientist) are using subsurface core images from the North Sea to develop a deep-learning model to automatically predict lithology and facies. This model can unlock warehouses full of high-resolution data, and can be applied to areas where large amounts of legacy data are available (e.g., Permian Basin).
Automated Lithology Prediction from Digital Outcrop Models
Zane Jobe and colleagues are also utilizing machine-learning models to extract lithologic and architectural information digital outcrop models derived from drones and other Structure-from-Motion techniques. See our GitHub repo and preliminary results. We are also collaborating with Gabe Walton at Mines on this project.