Paper No. 12
Presentation Time: 3:45 PM

LANDLAB – A NEW, OPEN-SOURCE, MODULAR, PYTHON-BASED TOOL FOR MODELING LANDSCAPE DYNAMICS


HOBLEY, Daniel E.J.1, TUCKER, Gregory E.2, ADAMS, Jordan M.3, GASPARINI, Nicole M.4, HUTTON, Eric5, ISTANBULLUOGLU, Erkan6 and SIDDHARTHA NUDURUPATI, Sai6, (1)CIRES and Dept of Geological Sciences, University of Colorado, UCB 399, 2200 Colorado Avenue, Boulder, CO 80309-0399, (2)CIRES and Department of Geological Sciences, University of Colorado, Campus Box 399, 2200 Colorado Avenue, Boulder, CO 80309-0399, (3)Dept of Earth & Environmental Sciences, Tulane University, 101 Blessey Hall, New Orleans, LA 70118, (4)Department of Earth and Environmental Sciences, Tulane University, 101 Blessey Hall, New Orleans, LA 70118, (5)Instaar, University of Colorado, campus Box 450, 1560 30th St, Boulder, CO 80303, (6)Civil & Environmental Engineering, University of Washington, 201 More Hall, Box 352700, Seattle, WA 98195-2700, daniel.hobley@colorado.edu

Landlab is a novel numerical landscape evolution model, which we have designed for maximum usability amongst the wider community. The model is targeted both at end users, and at developers who may wish to add their own functionality to the model. This rationale has led to the following features:

  • Open source code and community access through the CSDMS model repository.
  • Truly modular design, so a user can select which processes they wish to model in a landscape in a “plug and play” fashion. Each module is known as a component.
  • Python coding, for maximum usability for the minimum effort from the user. Components can be called readily ad-hoc from a Python interactive environment.
  • Full compatibility with the CSDMS Modeling Tool, and associated compatibility with other models under that framework.

The model is designed to run on both regular and irregular (e.g., Voronoi) model grids, and includes visualization tools. Model drivers handle the sharing of state variables amongst components, so that multiple components can easily update the same state variables. We will shortly implement routines to automate sharing of grid data between components, further simplifying the task of the end user. Dynamic regridding and multiple grid algorithms are under development.

The flexible design of Landlab allows its use over a wide variety of timescales, surface processes, and scientific hypotheses. We will present some examples of landlab in action, including coupled 2D sediment-hydrological systems, cratered planetary surfaces, long timescale fluvial dynamics, and soil moisture modeling.