November Luncheon

Please RSVP by 3:00pm on Friday, November 13, 2026.

Reservations made after the deadline will incur an additional fee of $10.

LOCATION: In-Person luncheon meeting will be held at the Bush Convention Center. Virtual meeting on Zoom. 

TOPIC: "A Spectral Gamma Ray Machine Learning Model for Clay Speciation in the Wolfcamp D: Filling the Data Gap and Delineating Regional Depositional and Structural Trends" - Calvin Layman*, Ryan Banas, Ben Davis

ABSTRACT: Clay mineralogy influences drilling stability, completion performance, and rock stiffness in the Wolfcamp D, but quantitative clay data are sparse, inconsistently reported, and rarely available at basin scale. We present a machine learning workflow that predicts illite, smectite, and chlorite volumes from spectral gamma ray (potassium, thorium, uranium) and conventional triple-combo logs, which are far more widely available than core-derived mineralogy.

The model was trained on XRD, FTIR, and XRF data from whole core, rotary sidewall cores, and cuttings, with mixed-layer clay volumes allocated to smectite and illite using expandable-layer fraction. Models were built with LightGBM regressors in Microsoft ML.NET. The six-feature model, excluding total GR, gave a mean absolute error of about 6.5% for illite (mean ~25%), under 1% for chlorite, and about 1% for smectite. Sampling resolution, depth alignment, and the variability of XRD as an interpretive product limit the fit, but the model captures overall trends. It was deployed across roughly 5,000 wells in the Midland Basin.

Mapping the outputs by Upper, Middle, and Lower Wolfcamp D reveals clear regional differences in clay speciation. Comparison with log-derived TOC, hydrogen index, and Tmax suggests a possible regional change in kerogen type, and residual structural maps indicate that the proto-Glasscock Nose controlled sedimentation. These results can help high-grade drilling and completion targets by identifying areas with more favorable clay assemblages and more stable rock.

BIO: Calvin Layman is a Senior Geologist at Diamondback Energy in Midland, Texas, with 7 years of experience in the energy industry. His current role focuses on unconventional resource assessment and characterization. Calvin holds a Bachelor of Science in Geology with a minor in Energy from Texas Christian University and a Master of Science in Geology from the University of Oklahoma. In this talk, he presents a spectral gamma ray machine learning model for clay speciation in the Wolfcamp D, addressing a key data gap and delineating regional depositional and structural trends.

Luncheon Generously Sponsored By