Fracture Network Connectivity Controls on CO2 Mineral Trapping Efficiency in Basalt: A Stochastic Reactive Transport Study

This deposit accompanies the manuscript "Fracture network connectivity controls on CO2 mineral trapping efficiency in basalt: A stochastic reactive transport study" by Chen, Xie, Kang, and Regenauer-Lieb, submitted to Water Resources Research (2026). The study couples stochastic discrete fracture network (DFN) generation with reactive transport modeling to quantify how fracture network topology controls CO2 mineral trapping efficiency in basalt. Twenty-five three-dimensional DFN realizations are simulated across five fracture intensity levels under representative reservoir conditions (50°C, 5 MPa, continuous injection of carbonated brine for 50 years), using a full basalt mineralogy and experimentally derived kinetic rate laws. The principal finding is that at fixed fracture intensity, carbonate precipitation varies by four orders of magnitude, and that a flow-weighted topological metric — the flow-reactive co-location index — predicts the rank order of trapping outcomes (Spearman ρ = 0.865, p < 0.001). Contents of this deposit: - dfn_library.zip: 25 DFN meshes generated with dfnWorks (.uge, .inp, boundary .ex files, and dfn_summary.json per realization)- pflotran_results.zip: 25 PFLOTRAN HDF5 simulation outputs and corresponding simulation_params.json files- paper_figures.zip: Nine published figures in PNG and PDF format- stress_test_results.zip: Backbone, dead-end, and finite-size topological metrics for Table 8- betweenness_results.csv and betweenness_results.json: Flow-reactive co-location index values per realization (Table 8) Source code is maintained on GitHub at:https://github.com/Yongqiang100/co2-basalt-dfn-ReactiveTransport The eight Python and shell scripts in the GitHub repository handle DFN generation (using dfnWorks), PFLOTRAN input preparation and simulation execution, post-processing, and figure generation. The figures can be reproduced from the supplied HDF5 outputs without rerunning PFLOTRAN. See the README in the GitHub repository for full reproduction instructions. Simulations were performed using PFLOTRAN (Lichtner et al., 2015) for multiphase reactive transport, and dfnWorks (Hyman et al., 2015) for stochastic DFN generation. Network analysis used NetworkX (Hagberg et al., 2008). Released under the Apache License, Version 2.0, applying to both code and data in this deposit.