The surface of the earth is constantly reshaped by geologic processes. Landslides carry rock and sediment downhill, soil washes off of farm fields, rivers bring sediment from mountains to the sea, and waves constantly reshape coastlines. Understanding how processes like these work is important for mitigating natural hazards, for managing resources, and for understanding how processes in the past helped shape today’s planet. In studying earth-surface processes, geoscientists and engineers rely on computer simulation models. Simulation models make it possible to conduct “digital experiments” with geologic systems that are otherwise too big, too slow, or too remote to fit inside a laboratory. However, working with simulation codes can be expensive and time-consuming. Research-grade computer codes are often written for a specific, one-time purpose, and may not be well documented or user-friendly. Furthermore, obtaining the data needed to run them is often time-consuming, and the steps may be difficult to reproduce when an investigator wants to understand and build on a previous study. In addition, many geoscientists are self-taught in computer programming and therefore may lack skills that are useful in designing and maintaining software that is robust, accessible, and reusable. To address these challenges, the project enacted major three areas of activity: engaging and coordinating the research community, developing and disseminating computing tools and infrastructure, and providing educational events and resources. The project supported the Community Surface Dynamics Modeling System (CSDMS), a facility established by and for the earth-surface processes research community to advance the science by working in these three essential areas of community, computing, and education.

One project goal was to enhance research productivity by developing, disseminating, and coordinating modular software that can be reused and combined in various ways to address emerging new research questions. To meet this goal, the project extended and enhanced software that supports community-based modeling of earth-surface processes. For example, the project developed a new approach, with accompanying open-source tools, for accessing data in a standardized, reproducible way. The method introduces “Data Components”, which are software library tools that allow a researcher to use computer code to automatically fetch certain kinds of data. The project also promoted and extended a formal protocol, known as the Basic Model Interface (BMI), for standardizing simulation-model codes. Standardization makes codes easier to learn and easier to link together to simulate interactions between processes. The BMI standard has now been adopted by several government agencies and research groups. The project also maintained, coordinated, and extended a computer programming library called Landlab, which researchers use to build new simulation codes.

To help engage community, the project hosted four professional conferences (two in-person, two virtual during the height of the COVID-19 pandemic). These conferences enabled attendees to connect with one another and to share ideas and findings on computationally oriented research topics, across a diverse spectrum of geoscience disciplines. The conferences also provided a forum for researchers to learn about the latest software tools to support their work. In addition, the project hosted and maintained a web-based catalog of community simulation codes, where researchers can learn about codes, access them, and share their own codes.

To address the gap in cyberinfrastructure education, the project sponsored a number of hands-on educational workshops (“clinics”) at conferences, developed online learning materials, and hosted a week-long training event for early-career scientists. These events and materials helped to elevate the computational skill level for geoscientists working on earth-surface processes.

Over the duration of the project, the membership in the CSDMS organization grew by about 700, from 1,670 in 2018 to 2,370 by the end of 2023. Similarly, the size of the model repository grew from just under 300 codes in 2018 to over 400 by the end of 2023. The modeling system and its tools supported a wide range of research. As one example indicator of research activity, between 2018 and 2023 the Landlab software library was used to support several dozen published research studies. The topics of these studies ranged from earthquake and landslide hazards to aquatic ecology. Overall, these statistics speak to a thriving research community whose work is increasingly aided by computational tools, and made more efficient by sharing those tools and the science that they support.

Last Modified: 01/17/2024
Modified by: Gregory E Tucker