TraceSeis, Inc. delivers rock physics, seismic inversion, and model validation tools built on 40+ years of field-tested expertise.
TraceSeis is a privately owned geophysical consulting and software company headquartered in Houston, Texas. Founded in 2003 by Alvaro Chaveste, TraceSeis delivers quantitative reservoir analysis to oil and gas operators through a combination of specialized consulting services and proprietary software tools.
The company's work is grounded in a single discipline: every estimate must be independently verifiable before it informs a decision. That standard originated in seismic inversion — where a wrong call sends a drilling program at the wrong target — and it now extends to AI model validation through CONFIRM, which applies the same principle to machine learning classification systems.
Before any inversion is run, TraceSeis verifies two things: that the rocks are sensitive to the property you care about, and that the seismic data faithfully represents the earth's reflectivity. Only then do we quantify.
A rock physics study evaluates whether changes in porosity, lithology, or fluids would actually be observable in seismic data before budget is committed to inversion.
Pre-stack data is conditioned and verified against synthetics, ensuring the seismic response represents true earth reflectivity rather than acquisition or processing artifacts.
Optimum rock properties are computed through seismic inversion, verified at well locations, and transformed into reservoir and geomechanical properties suited for exploration and development decisions.
Six-module platform covering LAS loading, fluid substitution (Gassmann), AVO modeling, well-log inversion, 3D property cubes, and statistical analysis. Built on the Relative Rock Properties methodology published in SEG's The Leading Edge.
Statistical audit for any system that classifies things into categories. Uses chi-square testing and Cramér's V to grade model consistency A–F across every domain it claims to handle. Applied to 174 reward models on the Allen Institute for AI's RewardBench 2 benchmark.
Wireline data preprocessing platform. Denoising, intelligent gap-filling, outlier detection, mnemonic standardization, and physics-informed QC for LAS files before rock physics analysis. Garbage in, garbage out — POLISH exists so it's never garbage in.
TraceSeis's methods are grounded in peer-reviewed research published in SEG's The Leading Edge, the AAPG Explorer, and presented at AGU. Selected publications:
Full citation record: Google Scholar — Alvaro Chaveste
Shale analysis, producibility prediction, relative flow-rate estimates, and synthetic curve generation.
Prospect generation, independent evaluation of prospects and assets, business development, and commercialization.
Integrated enterprise seismic imaging — Kirchhoff PSTM/PSDM, tomography, RTM, and 5-D interpolation.
Whether it's a feasibility study, a full inversion workflow, a SeisTool® license, or an independent CONFIRM™ validation — start with a conversation.
Get in Touch Explore CONFIRM™