Master of Science

thesisThe Stochastic Time-Inverted Lagrangian Transport (STILT) model is comprised of a compiled Fortran executable that carries out advection and dispersion calculations as well as a higher level code layer for simulation control and user interaction, written in the open source data analysis language R. We introduce modifications to the STILT-R codebase with the aim to improve the model's applicability to fine-scale trace gas measurement approaches. The changes facilitate placement of spatially distributed receptors and provide high level methods for single and multinode parallelism. We present a kernel density estimator to calculate influence footprints and demonstrate improvements over previous methods. This framework provides a central source repository to reduce code fragmentation between STILT user groups as well as a systematic, well-documented workflow for users. We apply the modified STILT to lightrail measurements in Salt Lake City, UT and discuss how results from our analyses can inform future fine-scale measurement approaches and modeling efforts