geoclaw-landspill: an oil land-spill and overland flow simulator for pipeline rupture events

The package geoclaw-landspill builds on the geoclaw shallow-water solver to numerically simulate oil land-spills and overland flows that occur during pipeline accidents. It helps understanding how oil flows above ground after accidental release and to study a pipeline’s impact on the environment and economy. By understanding the impact of a to-be-constructed pipeline, one can choose a pipeline route with the least loss if accidents were to happen. On the other hand, understanding how oil flows also helps develop rescue teams’ deployment strategies and remedy plans for potential accidents.


Summary
The package geoclaw-landspill builds on the geoclaw shallow-water solver to numerically simulate oil land-spills and overland flows that occur during pipeline accidents. It helps understanding how oil flows above ground after accidental release and to study a pipeline's impact on the environment and economy. By understanding the impact of a to-be-constructed pipeline, one can choose a pipeline route with the least loss if accidents were to happen. On the other hand, understanding how oil flows also helps develop rescue teams' deployment strategies and remedy plans for potential accidents.
The package provides a numerical solver for the full shallow-water equations, and postprocessing utilities. The solver is an expanded version of GeoClaw (Berger et al., 2011), a parallel shallow-water equation solver for tsunami simulations using adaptive mesh refinement (AMR) and finite-volume methods. We added several new features and modifications to simulate the overland flow of pipeline rupture events, including (citations refer to the details of the adopted models): • point sources with multi-stage inflow rates to mimic rupture points along a pipeline; • Lewis Squires Correlation for temperature-dependent flow viscosity (Mehrotra, 1991); • Darcy-Weisbach friction model with multi-regime coefficient models (laminar, transient, and turbulent regimes) (Yen, 2002) and Churchill's model (Churchill, 1977); • inland waterbody interactions; • Fingas' evaporation models (Fingas, 2004); and • optimizations to improve performance in overland flow simulations.
In addition to the numerical solver, geoclaw-landspill is also able to: • automatically download high-resolution topography and hydrology data, and • create CF-compliant NetCDF raster files for mainstream GIS software (e.g., QGIS, Ar-cGIS).
geoclaw-landspill includes several examples to showcase its capability to simulate the overland flow on different terrain.
We implemented the core numerical solver and new features using Fortran 2008 to better integrate into the original GeoClaw code. Other utilities are in Python. Users can find geoclawlandspill as a Python package on PyPI and install it through pip. The only dependency that pip does not manage is the Fortran compiler. Docker images and Singularity images are also available. They ease the deployment of the solver and simulations to cloud-based highperformance clusters.

Statement of need
In the US, between 2010 and 2017, an average of 388 hazardous liquid pipeline accidents happened per year. Half of accidents contaminate soil, and 41% of accidents affect areas with high consequences in either ecology or economy. Moreover, 85%, on average, of the released oil was not recovered and kept damaging the environment (Belvederesi et al., 2018). From the perspective of risk management, while pipelines are unavoidable in modern days, it is necessary to understand how a pipeline may impact the environment if any accidental release happens. geoclaw-landspill serves this purpose. It provides a free and open-source simulation tool to researchers investigating the danger, risk, and loss posed by potential pipeline accidents.
To our knowledge, geoclaw-landspill is the only open-source high-fidelity flow simulator for oil pipeline rupture events. High fidelity means the results provide more details and accuracy because of high-resolution digital elevation data, fine spatial discretization, and full shallow-water equations. Commercial products with a similar capability to geoclaw-landspill are available (Gin & Davis, 2012;Hydronia, n.d.;RPS Group, n.d.;Zuczek et al., 2008). Other noncommercial software more or less serving a similar purpose usually relies on simplified models, such as 1D open-channel models, diffusive wave approximation, gravity current models, and gradient-based route selection models (Farrar et al., 2005;Guo, 2006;Hussein et al., 2002;Ronnie, 2004;Simmons & Knell, 2003;Su et al., 2017). Moreover, these non-commercial codes are no longer available or are not open-source.
Another value of geoclaw-landspill is that it provides a platform for scholars who study oil flow modeling to implement and test their models. As the main flow solver is under the BSD 3-Clause License, scholars can add their models to geoclaw-landspill freely.

Past or ongoing research projects using the software
The following conference presentations and posters used previous versions of geoclaw-landspill: