Greg Ewing

Greg Ewing

Hydrologist, Environmental Engineer, Taco Enthusiast; University of Iowa & University of Michigan

Talks and presentations

See a map of all the places I've given a talk!

A Client-Side Web Framework for Rainfall-Runoff Modelling in a Hybrid Learning Context

June 23, 2022

Oral Presentation, AGU Frontiers in Hydrology, San Juan, Puerto Rico

Using hydrological models in classroom settings face numerous barriers, many of which stem from the limited time to teach both the scientific material and vocational tools (i.e., a modelling software,) during a course. For example, data preparation and model development are integral, time consuming tasks when using models in professional settings, yet may not provide meaningful educational opportunities as part of a hydrology curriculum. Towards the goal of using less class time teaching tools and more time exploring hydrologic concepts, we present a client-side web framework for rainfall-runoff modelling in hybrid learning contexts. At the core of our application is a client-side rainfall-runoff model, HLM-Web. To address the barriers to modelling in education, we designed the web application to extend HLM-Web such that it minimizes many time-intensive modelling tasks. For example, we have prepared input data for five basins of varying size (from 10 km2 to 12,000 km2,) and allow users to seamlessly select from these with the click of a button. Likewise, users can then interactively select between constant or variable rainfall-runoff processes and alter parameter values of the model using slider elements. The application also includes a rainfall forcing generator, which allows for a variety of inputs to be quickly generated. Once a model is prepared and the simulation is complete, hydrograph visualizations appear along with data export capabilities. Notably, all features and computation of the application are performed client-side. The application was developed with iterative feedback from course instructors such that it would supplement university-level hydrology course curricula. The application framework was piloted at the University of Iowa and University of Manitoba. Finally, we demonstrate many educational exercises enabled by the framework.

Device-Agnostic Environmental Modelling Framework for the Web

June 21, 2022

Oral Presentation, AGU Frontiers in Hydrology, San Juan, Puerto Rico

Recently, web-enabled tools for data processing, storage, computation, and visualization have proliferated to support web-based environmental modelling. Most of these tools rely on server resources for computation and data tasks. Yet, the continued advancement of in-browser, client-side compute performance presents an opportunity to further leverage client-side compute resources. Modern browsers also offer a rich and flexible platform on which to build interactive, user-centric tools making it a strong candidate to be the interface of choice for next generation modelling. Toward the goal of device agnostic web-based modelling, we propose an environmental modelling framework developed using modern web standards only. First, we present the development of the modelling framework using a Service Oriented Architecture. Next, we demonstrate the capabilities of client-side compute to simulate environmental processes, such as the rainfall-runoff generation process and advection-reaction chemical transport process. Finally, we discuss the application areas for this technology such as operational forecasting and community-level decision making.

Applying Interaction Design to Improve Decision Support in Urban Watershed Management

June 21, 2022

Oral Presentation, AGU Frontiers in Hydrology, San Juan, Puerto Rico

Over the past decade, advances in urban drainage modeling, continuous monitoring, and active control have transformed the way engineers and decision makers think about the future of urban watersheds. These new technologies have the potential to reduce flooding and improve water quality, but only if they are adopted in the real world. Yet, as urban water systems are outfitted with new sensors, actuators, and modeling tools, human operators are faced with the task of leveraging increasingly large amounts of data to take actions in a complex, growing decision space. As a result, opportunities for intelligent control and cross-system cooperation are being missed, even where the technology is in place to make them possible.