Generation and Application of the Next Generation of High-Resolution Climate Projections for the Great Lakes Region

Project Summary

The UW-RegCM4 model is the primary source of future projections in GLISA’s GL Ensemble and a vital source of climate change guidance for many practitioners, boundary organizations, and government agencies in the region (e.g., Alliant Energy, GL Indian Fish and Wildlife Commission, Ducks Unlimited, WI Initiative on Climate Change Impacts, MI Department of Natural Resources). Co-PI Notaro and colleagues dynamically downscaled this dataset for the GL region by driving the 25-km Regional Climate Model Version Four (RegCM4) with output from six CMIP5 global climate models (GCMs). The value of this downscaled product lies in the interactive coupling of the regional climate model (RCM) with a 1D lake model, addressing critical lake-atmosphere interactions. In contrast, many other GCM and RCM efforts applied more rudimentary approaches to treating the GL by either neglecting the lakes, including a handful of awkwardly placed water grid cells, or extrapolating Atlantic or Hudson Bay water temperatures across the lakes (Briley et al. 2015 and 2021a, Notaro et al. 2015, and Xue et al. 2017). However, this product has limitations that urgently need addressing. In this project, we seek to attain significantly greater regional reliability and confidence in climate projections in order to improve usability. Due to its coarse grid spacing (25-km) and use of convective parameterization, meteorological extremes in UW-RegCM4 are often dampened in their intensity and exhibit imprecise morphologies. The ensemble’s coarse and 1D lake representation produces: 1) insufficiently resolved seasonal to interannual variability in lake surface temperature (LST); 2) an underestimated historical lake warming trend; 3) pronounced seasonal biases in LST, stratification, and evaporation; and 4) excessive ice cover projection. Given that the GL region is increasingly a hotspot of rapid climatic and limnological changes, it is critical to develop and apply a state-of-the-art RCM at a finer spatial resolution that includes complex 3D lake processes and lake-atmosphere exchanges of heat and moisture. This is key to providing the most dependable data and guidance to regional stakeholders. 

We will produce a new dataset that provides better projections of regional climate extremes and GL hydrology, including precipitation, evaporation, and runoff, with implications for future lake levels. This will yield, for the first time, reliable data to generate future projections of precipitation IDF curves (1A, p.7), a key need for local and regional planning. Through a recent NASA grant (Notaro, PI), Dr. Pengfei Xue (Michigan Technological University) interactively coupled a 3-km non-hydrostatic version of the NASA-Unified Weather Research and Forecasting (NU-WRF) model to the Finite Volume Community Ocean Model (FVCOM) to represent 3D lake circulation and ice motion. Adopting this higher spatial resolution and inclusion of 3D lake processes will address the limitations of the UW-RegCM4 product. Initial results demonstrate a vast reduction in LST and ice cover biases and improvement in lake-atmosphere interactions (which regulate regional climate, Figure 7).