SPGM: A Scalable PaleoGeomorphology Model

Numerical models of landscape evolution are playing an increasingly important role in providing an improved understanding of geomorphic transport processes shaping Earth's surface topography. Improving theoretical underpinnings coupled with increasing computational capacity has led to the development of several open source codes written in low-level languages. However, adapting these codes to new functionality or introducing greater flexibility often requires significant recoding. Here we present a multi-process, scalable, numerical model of geomorphological evolution, built with a modular structure and geared toward seamless extensibility. We implement recent algorithmic advances that reduce the computational cost of flow routing - a problem that typically scales quadratically with the number of unknowns - to linear in time while allowing for parallel implementations of geomorphic transport processes. Our scalability tests demonstrate that such parallelizations can achieve an order of magnitude speedup on a typical desktop computer, making large-scale simulations more tractable.