Keywords
frameworks, interface, model-coupling, numerical modeling, python
Start Date
15-9-2020 7:20 PM
End Date
15-9-2020 7:40 PM
Abstract
The Community Surface Dynamics Modeling System (CSDMS) develops and maintains a suite of modeling tools that enable the coupling of numerical Earth-surface models and the rapid creation of new models. Two principal software elements of this suite are the newly released Landlab-2.0 and the Python Modeling Toolkit (pymt), both of which are freely-available Open Source Python packages. Landlab-2.0 consists primarily of three pieces. The first is a gridding engine that allows model developers to quickly create new models in a way that is grid-agnostic. The second element is a rich library of utilities for common operations such as file input and output. The third piece is a collection of modularized components that model single physical processes. Landlab-2.0 brings 31 new models to the collection (for a total of 58) and, importantly, has defined a more streamlined and standardized component interface. Landlab-2.0’s component interface allows its process components to be brought into other frameworks and to be automatically wrapped with a Basic Model Interface (BMI). This allows Landlab components to be included in frameworks such as TerrainBento (a modular landscape evolution modeling package), Umami (a package for calculating landscape metrics for assessing model-data fit), as well as BMI-friendly frameworks such as the pymt. The pymt brings together a collection of legacy models that represent diverse types of environmental systems. Models in the pymt collection are written in a variety of languages, but share a Python front- end with BMI as the standard interface. To operate within pymt, a model must expose a BMI. Landlab-2.0 provides a crosswalk where components can be exchanged between landlab and pymt. Together, Landlab, BMI, and pymt provide a nimble, plug-and-play environment for model building, coupling, and exploration. Collectively, these tools improve efficiency by reducing the time researchers spend wrestling with idiosyncratic programs and their interfaces.
Landlab v2.0: Create and couple Earth surface models in a Python framework
The Community Surface Dynamics Modeling System (CSDMS) develops and maintains a suite of modeling tools that enable the coupling of numerical Earth-surface models and the rapid creation of new models. Two principal software elements of this suite are the newly released Landlab-2.0 and the Python Modeling Toolkit (pymt), both of which are freely-available Open Source Python packages. Landlab-2.0 consists primarily of three pieces. The first is a gridding engine that allows model developers to quickly create new models in a way that is grid-agnostic. The second element is a rich library of utilities for common operations such as file input and output. The third piece is a collection of modularized components that model single physical processes. Landlab-2.0 brings 31 new models to the collection (for a total of 58) and, importantly, has defined a more streamlined and standardized component interface. Landlab-2.0’s component interface allows its process components to be brought into other frameworks and to be automatically wrapped with a Basic Model Interface (BMI). This allows Landlab components to be included in frameworks such as TerrainBento (a modular landscape evolution modeling package), Umami (a package for calculating landscape metrics for assessing model-data fit), as well as BMI-friendly frameworks such as the pymt. The pymt brings together a collection of legacy models that represent diverse types of environmental systems. Models in the pymt collection are written in a variety of languages, but share a Python front- end with BMI as the standard interface. To operate within pymt, a model must expose a BMI. Landlab-2.0 provides a crosswalk where components can be exchanged between landlab and pymt. Together, Landlab, BMI, and pymt provide a nimble, plug-and-play environment for model building, coupling, and exploration. Collectively, these tools improve efficiency by reducing the time researchers spend wrestling with idiosyncratic programs and their interfaces.
Stream and Session
false