Keynote Abstracts
Theme 1: Big data, big dilemma: are we observing nature or nature exposed to our method of questioning?
Dr Giulia Sofia, University of Connecticut
Big Data and machine learning approaches offer previously unheard-of potential to enable place-based studies of landscapes, turning geomorphology into an increasingly data-rich science. This trend has boosted the discipline, but researchers are also facing numerous new technical or management issues in the field, as well as increased research needs. Geomorphology is not a linear, cause-and-effect science, and the inherent complexities and uncertainties of landscape data are challenging to interpret, as much depends upon the experience and training of who is making interpretations, as well as the tools used to process and analyze the data. New tools for big data analytics and possession of (or access to) big data can enable us to make sense of physical landscapes, but they call for generating knowledge with a reliable evidence base. This talk showcases what has been done and what can be achieved in future research in the geomorphology domain towards a data-rich era.
Theme 2: On the ecological footprint of research in Earth an Environmental Science, and potential ways to reduce it!
Dr Odin Marc, Géosciences Environnement Toulouse
After recalling the main rationale for reducing ecological footprint of society, and specifically the CO2 footprints of science, I will detail state of the art methods to evaluate the carbon footprint of research laboratories. First, I will detail the methodology and some examples for 5 Earth, Environmental and Space Science laboratories in Toulouse. Second, I will focus on the quantification of the footprint of internal and external IT infrastructures (computing and storage centers), and on 1st order estimation of the footprint of large research infrastructure producing data, specifically satellite missions. The combination of these approaches yields estimates of the footprint of a typical Earth Scientist in Toulouse, but appear to be fairly similar in other areas. Then I will detail and, attempt to quantify, several strategies aimed at reducing the footprint of Science, including modal report for mobility, flight quotas and diverse measures aiming at reducing or optimising purchases. I will conclude with some more fundamental propositions for rethinking scientific activity and practices and on several arguments on why we should urgently engage toward them.
Theme 3: Best practices in topographic analysis - avenues for the development of TopoToolbox
Dr Wolfgang Schwanghart, University of Potsdam
In the dynamic field of computational geomorphology, topographic analysis stands as a cornerstone for understanding Earth’s evolving landscapes. This talk explores the vital role of software tools in this domain, focusing on best practices in topographic analysis and the development of TopoToolbox, a software for digital terrain analysis. With the advent of globally available high-resolution Digital Elevation Models (DEMs), the demand for robust software has surged. Paradoxically, much of this software is crafted by researchers who lack formal training in software development. TopoToolbox itself emerged from the hands of researchers with no prior programming experience. Yet, over a decade since its inception, it continues to be a valuable resource in the geospatial community. In this presentation, we delve into the lessons learned from the development and evolution of TopoToolbox, and discuss the critical role of the software’s design, and our vision for its future development.
Theme 4: Successful Numerical Modeling = (software + data) * people
Dr Nicole Gasparini, Tulane University
Have you ever had trouble running the computer code that came with a published paper? Have you ever tried to recreate published numerical experiments with a different numerical model and the results didn’t match? Have you ever tried to change a model parameter from the default value and the code exploded? Have you ever wondered why this is so hard? If you answered yes to any or all of these questions, then we have something in common, and this talk might be for you.
I will explore some of the reasons that I think reproducibility with numerical modeling of earth surface processes is often challenging, sometimes impossible, and nearly always undervalued. Reproducibility requires sustainable software and data management. It also requires good communication and time. None of this can happen without committed scientists and resources to support them. The good news is that we can do this if we as a community decide it is important.