Dan Myers Research Showcase: Hydrological model calibration and rain-on-snow

I am currently a third-year PhD candidate at Indiana University Bloomington, USA, in the Ficklin Hydroclimatology Lab. I use hydrological models to better understand how climate change affects rivers and streams in North America, as part of a collaborative National Science Foundation funded project called HydroClim. I focus on model calibration and rain-on-snow melt simulation. 

A person sitting on a big snowy hill wearing snow shoes
Snowshoeing along the Agawa River, Ontario, Canada during snowmelt season (credit: Dan Myers)

Did you know that there are common pitfalls of calibrating hydrological models with certain time periods of observed data? During my research, I have learned that hydrological models are often calibrated and evaluated without considering how the rivers and streams change over time. Changing rainfall patterns, urban developments, and other “nonstationarities” can mean that a model calibrated to one time period may not work well in another period. In an article published in Hydrological Processes in January 2021, my coauthors and I investigated how the arbitrary selection of calibration and validation time periods can have a major influence on simulations of snowmelt, groundwater flow, evapotranspiration, surface runoff, and other components of the water cycle. We found it crucial for hydrological modelers using this approach to choose calibration and validation time periods carefully or use alternative techniques.

Person sits in front of a computer at Indiana University Bloomington, USA office
Modeling hydrology at Indiana University Bloomington, USA (credit: Dan Myers)

Currently, I research a phenomenon known as rain-on-snow melt and how it affects the availability of freshwater in a changing climate. Rain-on-snow melt events, as the name suggests, occur when a warm rain falls on an existing snowpack and (along with other sources of heat) melts it. These can contribute to both extreme floods and decreased water storage. Developing knowledge of how climate change affects rain-on-snow melt patterns is vital for understanding how ecosystems and communities in rain-on-snow prone regions can adapt to them.

My study area is the North American Great Lakes Basin. My models have shown how rain-on-snow melt can increase the severity of winter floods in the Basin, due to the combination of rainfall and meltwater. They also demonstrate how this can decrease the amount of water stored in snowpacks and even in groundwater here. This is because the rain-on-snow melt causes water to quickly leave watersheds through river systems, rather than being stored longer as snowpack or slowly soaking into the ground. This can actually cause lower streamflows in the summer! Further, the models show that the amount of rain-on-snow melt that occurs is very sensitive to warming air temperatures in the Great Lakes Basin.

A satellite image of snow and clouds over the Great Lakes Basin
Snowfall in the Great Lakes Basin (Credit: NOAA NWS)

I have found that getting into the scientific research field is difficult but also very rewarding. There are error messages, modeling mistakes, rejection letters, more rejection letters, even more rejection letters, and other challenges to overcome. All these rejections prompted me to write a post for my university’s science blog. After talking with other students I learned that these rejections are part of the game. The field of hydrology is very fulfilling, particularly because our research can contribute to scientific knowledge and be applied to solve real world problems. I am excited to see where the future takes it.

By Dan Myers, PhD Candidate, Indiana University Bloomington, USA

Leave a Reply