Hilary Dugan - US grants

The success of machine learning (ML) in many applications where large-scale data is available has led to a growing anticipation of similar accomplishments in scientific disciplines. The use of data science is particularly promising in scientific problems involving processes that are not completely understood. However, a purely data-driven approach to modeling a physical process can be problematic. For example, it can create a complex model that is neither generalizable beyond the data on which it was trained nor physically interpretable. This problem becomes worse when there is not enough training data, which is quite common in science and engineering domains. A machine learning model that is grounded by explainable theories stands a better chance at safeguarding against learning spurious patterns from the data that lead to non-generalizable performance. This is especially important when dealing with problems that are critical and associated with high risks (e.g., extreme weather or collapse of an ecosystem). Hence, neither an ML-only nor a scientific knowledge-only approach can be considered sufficient for knowledge discovery in complex scientific and engineering applications. This project is developing novel techniques to explore the continuum between knowledge-based and ML models, where both scientific knowledge and data are integrated synergistically. Such integrated methods have the potential for accelerating discovery in a range of scientific and engineering disciplines. This project will train interdisciplinary scientists who are well versed in such methods and will disseminate results of the project via peer-reviewed publications, open-source software, and a series of workshops to engage the broader scientific community.

This project aims to develop a framework that uses the unique capability of data science models to automatically learn patterns and models from data, without ignoring the treasure of accumulated scientific knowledge. Specifically, the project builds the foundations of knowledge-guided machine learning (KGML) by exploring several ways of bringing scientific knowledge and machine learning models together using pilot applications from four domains: aquatic ecodynamics, climate and weather, hydrology, and translational biology. These pilot applications were selected because they are at tipping points where knowledge-guided machine learning can have a transformative effect. KGML has the potential for providing scientists and engineers with new insights into their domains of interest and will require the development of innovative new machine learning approaches and architectures that can incorporate scientific principles. Scientific knowledge, KGML methods, and software developed in this project could potentially be extended to a wide range of scientific applications where mechanistic (also known as process-based) models are used.

This project is part of the National Science Foundation's Harnessing the Data Revolution (HDR) Big Idea activity.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

More than 90% of lakes on Earth are covered with ice in the winter. When a lake is covered with ice, this affects many ecosystem processes. The timing of when ice forms and how long it lasts are both important, but so are the thickness and clarity of the ice. Changing climate conditions have affected the duration of ice cover as well as the quality of ice, as warm winter temperatures can cause temporary thawing. There are three different ways to understand lake ice and its impacts, each with strengths and weaknesses. Observations from the shore of the lake are accurate but do not capture many aspects of lake ice timing, extent, and quality. Remote sensing techniques provide new opportunities to expand the scale and scope of research on ice-covered lakes. Lake models can estimate ice parameters that are difficult to measure such as ice thickness. In order to fully understand global lake ice dynamics and how they are changing, stronger connections between limnologists, remote sensing scientists, and modelers are needed. This project will support a workshop to study the dynamics of lake ice formation and the consequences for ecosystem functioning of temperate lakes. Participants in the workshop will join the annual Frozen Assets Festival in Madison, Wisconsin and participate in public events. Researchers studying lake ice from around the world will describe the ecological role and cultural value of lake ice and demonstrate how remote sensing is used to understand lake ice.


The current lake ice database only contains 631 lakes with land-based observations, but there are millions of lakes that freeze every year. The workshop will achieve four main goals: 1) Identify critical questions about lake ice that require cross-disciplinary approaches, 2) Improve cross-disciplinary understanding of how data are acquired and used, 3) Reach consensus on core aspects of a cross-disciplinary lake ice database, and 4) Develop a diverse community of researchers to foster collaborative projects. The objectives will be achieved through a series of interactive sessions and breakout groups during the workshop. A structure for a global lake ice database will also be developed. There are specific activities before, during and after the workshop to initiate and support community-building among scientists from different fields. Ultimately, the workshop will produce guidance and resources for a broad range of researchers working with ice-covered lakes.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

Researchers at the University of Wisconsin Trout Lake Station conduct year-round research on lakes in northern Wisconsin, sampling frozen or near freezing waterbodies for at least six months out of the year. Trout Lake Station is a long-time contributor and leader of winter lake research efforts internationally and prides itself on providing an inclusive size selection of high-quality winter safety gear and hands-on training to researchers entering the field. While the expertise and equipment at Trout Lake Station can accommodate additional winter research, training, and outreach; the station lacks a heated workspace where people can escape the freezing temperatures, work to develop new lake sampling technologies, and where existing gear can be properly stored and serviced. Through the construction of a heated year-round research workshop, Trout Lake Station will increase its research and outreach capacity both seasonally and spatially.

The Winter Instrumentation Limnology Laboratory (WILL) is a ~3,600 sq ft heated multifunctional research outbuilding composed of three main sections: 1) an active gear staging area, 2) a large workshop/toolshop, and 3) a sensor, buoy platform, and limnological gear makerspace. The building will be a stick and frame design with shop heating that maintains workable temperatures throughout all seasons in a north temperate climate. The WILL is proposed to replace an existing 1,200 sqft non-heated tool workshop. It will triple the outbuilding research workspace and complement the existing 11,200 sqft laboratory and conference building. The winter research user base for TLS stretches beyond the UW-Madison campus, and includes the Wisconsin Department of Natural Resources (WDNR), the National Ecological Observatory Network (NEON) staff and scientists, Global Lakes Ecological Observatory Network (GLEON) researchers, Lake Superior National Estuary Research Reserve (LSNERR), and regional universities. Trout Lake Station aims to become a hub that connects water interest groups in the region. Increasing year-round capacity and providing heated workshop space will strengthen links and collaborations between regional partners. For more information, see https://limnology.wisc.edu/trout-lake-station-welcome/.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.