In prior EMSL supported work we developed an ab initio molecular dynamics (AIMD) approach to interpret the EXAFS signal of impurities/micronutrients in the structure of Fe(oxyhydroxide) host minerals. This methodology is superior to the usual phenomenological fitting of EXAFS data because AIMD-informed EXAFS constrains the thermal disorder allowing for an independent assessment of structural motifs. Nonetheless, configurational space can be vast and, as our work has shown, impurities adopt unprecedented local coordination environments that often confound our intuition with little to no comparison to their pure phase counterparts making it difficult to navigate configuration space. Coupling such issues to the expensive nature of AIMD has created a bottleneck which is slowing the pace of discovery. Consequently, to accurately and efficiently search the vastness of configuration space that exists for even one impurity in one host phase, we propose developing an automated work-flow scheme that will integrate AI/ML methods, AIMD, multiple data streams, and rare-event techniques. The proposed development represents an important step towards addressing the challenge of interpreting measurements of complex systems in which even modest disorder greatly complicates the straightforward application of modern theoretical methods. The outcome will enable us to go beyond the Fe (oxy)hydroxide system to investigate the environment of metal centers in the organic fraction such as proteins, where the issue of configurational disorder is even greater than for the inorganic fraction. In turn, this will greatly facilitate the community’s ability to predict the fate, transport, and bioavailability of both micronutrients and contaminants in soils and subsurface environments. Further, this work will open possibilities for investigating the role of impurities and their associated defect structures on the stability of adsorbed organic matter with potential implications for carbon cycling. This team is uniquely qualified to implement this plan where Bylaska and Prange are veteran users of EMSL computational resources; Bylaska is the principal architect of the AIMD module in NWChem/NWChemEx and the developer of EMSL Arrows. Prange has contributed to the FEFF code for analyzing EXAFS and is currently a PI on an open call LDRD that is developing ML methods for correlating EXAFS and XANES. Mergelsberg is an expert in X-ray scattering and has worked closely with Ilton, Prange and Bylaska on using AIMD-informed EXAFS and PDF analysis. Ilton leads the domain science aspect of this research and has a deep knowledge of electronic spectroscopy and experimental methods.