Cold Season Performance of the NU-WRF Regional Climate Model in the Great Lakes Region

Michael Notaro aNelson Institute Center for Climatic Research, University of Wisconsin–Madison, Madison, Wisconsin

Search for other papers by Michael Notaro in
Current site
Google Scholar
PubMed
Close
,
Yafang Zhong bSpace Science and Engineering Center, University of Wisconsin–Madison, Madison, Wisconsin

Search for other papers by Yafang Zhong in
Current site
Google Scholar
PubMed
Close
,
Pengfei Xue cDepartment of Civil, Environmental, and Geospatial Engineering, Michigan Technological University, Houghton, Michigan

Search for other papers by Pengfei Xue in
Current site
Google Scholar
PubMed
Close
,
Christa Peters-Lidard dHydrosphere, Biosphere, and Geophysics Earth Science Division, NASA Goddard Space Flight Center, Greenbelt, Maryland

Search for other papers by Christa Peters-Lidard in
Current site
Google Scholar
PubMed
Close
,
Carlos Cruz eNASA Goddard Space Flight Center, Greenbelt, Maryland

Search for other papers by Carlos Cruz in
Current site
Google Scholar
PubMed
Close
,
Eric Kemp eNASA Goddard Space Flight Center, Greenbelt, Maryland

Search for other papers by Eric Kemp in
Current site
Google Scholar
PubMed
Close
,
David Kristovich fIllinois State Water Survey, University of Illinois at Urbana–Champaign, Champaign, Illinois

Search for other papers by David Kristovich in
Current site
Google Scholar
PubMed
Close
,
Mark Kulie gNOAA/National Environmental Satellite, Data, and Information Service, Madison, Wisconsin

Search for other papers by Mark Kulie in
Current site
Google Scholar
PubMed
Close
,
Junming Wang fIllinois State Water Survey, University of Illinois at Urbana–Champaign, Champaign, Illinois

Search for other papers by Junming Wang in
Current site
Google Scholar
PubMed
Close
,
Chenfu Huang cDepartment of Civil, Environmental, and Geospatial Engineering, Michigan Technological University, Houghton, Michigan

Search for other papers by Chenfu Huang in
Current site
Google Scholar
PubMed
Close
, and
Stephen J. Vavrus aNelson Institute Center for Climatic Research, University of Wisconsin–Madison, Madison, Wisconsin

Search for other papers by Stephen J. Vavrus in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

As Earth’s largest collection of freshwater, the Laurentian Great Lakes have enormous ecological and socioeconomic value. Their basin has become a regional hotspot of climatic and limnological change, potentially threatening its vital natural resources. Consequentially, there is a need to assess the current state of climate models regarding their performance across the Great Lakes region and develop the next generation of high-resolution regional climate models to address complex limnological processes and lake–atmosphere interactions. In response to this need, the current paper focuses on the generation and analysis of a 20-member ensemble of 3-km National Aeronautics and Space Administration (NASA)-Unified Weather Research and Forecasting (NU-WRF) simulations for the 2014/15 cold season. The study aims to identify the model’s strengths and weaknesses; optimal configuration for the region; and the impacts of different physics parameterizations, coupling to a 1D lake model, time-variant lake-surface temperatures, and spectral nudging. Several key biases are identified in the cold-season simulations for the Great Lakes region, including an atmospheric cold bias that is amplified by coupling to a 1D lake model but diminished by applying the Community Atmosphere Model radiation scheme and Morrison microphysics scheme; an excess precipitation bias; anomalously early initiation of fall lake turnover and subsequent cold lake bias; excessive and overly persistent lake ice cover; and insufficient evaporation over Lakes Superior and Huron. The research team is currently addressing these key limitations by coupling NU-WRF to a 3D lake model in support of the next generation of regional climate models for the critical Great Lakes Basin.

Significance Statement

Climate change poses a serious threat to the vital natural resources of the Laurentian Great Lakes region. Complex lake–atmosphere interactions and limnological processes are a challenge for regional climate models. To address the threat of climate change, there is a clear need to further evaluate and develop modeling tools for the Great Lakes Basin. Here, we evaluate the regional performance of the National Aeronautics and Space Administration’s regional climate model at high spatial resolution in support of ongoing efforts to develop the next generation modeling tool for the Great Lakes region.

© 2021 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Yafang Zhong, yafangzhong@wisc.edu

Abstract

As Earth’s largest collection of freshwater, the Laurentian Great Lakes have enormous ecological and socioeconomic value. Their basin has become a regional hotspot of climatic and limnological change, potentially threatening its vital natural resources. Consequentially, there is a need to assess the current state of climate models regarding their performance across the Great Lakes region and develop the next generation of high-resolution regional climate models to address complex limnological processes and lake–atmosphere interactions. In response to this need, the current paper focuses on the generation and analysis of a 20-member ensemble of 3-km National Aeronautics and Space Administration (NASA)-Unified Weather Research and Forecasting (NU-WRF) simulations for the 2014/15 cold season. The study aims to identify the model’s strengths and weaknesses; optimal configuration for the region; and the impacts of different physics parameterizations, coupling to a 1D lake model, time-variant lake-surface temperatures, and spectral nudging. Several key biases are identified in the cold-season simulations for the Great Lakes region, including an atmospheric cold bias that is amplified by coupling to a 1D lake model but diminished by applying the Community Atmosphere Model radiation scheme and Morrison microphysics scheme; an excess precipitation bias; anomalously early initiation of fall lake turnover and subsequent cold lake bias; excessive and overly persistent lake ice cover; and insufficient evaporation over Lakes Superior and Huron. The research team is currently addressing these key limitations by coupling NU-WRF to a 3D lake model in support of the next generation of regional climate models for the critical Great Lakes Basin.

Significance Statement

Climate change poses a serious threat to the vital natural resources of the Laurentian Great Lakes region. Complex lake–atmosphere interactions and limnological processes are a challenge for regional climate models. To address the threat of climate change, there is a clear need to further evaluate and develop modeling tools for the Great Lakes Basin. Here, we evaluate the regional performance of the National Aeronautics and Space Administration’s regional climate model at high spatial resolution in support of ongoing efforts to develop the next generation modeling tool for the Great Lakes region.

© 2021 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Yafang Zhong, yafangzhong@wisc.edu

1. Introduction

The Laurentian Great Lakes are the Earth’s largest collection of freshwater and an invaluable resource to society and wildlife (Botts and Krushelnicki 1988). The Great Lakes megaregion is home to over 55 million people (Todorovich 2009). The lakes critically support the United States’ and Canadian economies through impacts on shipping, drinking water, power production, manufacturing, fishing, and recreation (Vaccaro and Read 2011). The basin contains a rich diversity of fish, animals, and plants (Crossman and Cudmore 1998) and ecologically valuable wetlands.

The Great Lakes exert a prominent effect on regional climate due to their large thermal inertia, variability as a moisture source to the atmosphere, and contrasts in moisture, heat, friction, and radiation compared to adjacent land (Changnon and Jones 1972; Scott and Huff 1997; Chuang and Sousounis 2003; Notaro et al. 2013a). Heat and moisture fluxes destabilize and moisten the boundary layer during autumn–winter (Bates et al. 1993; Blanken et al. 2011). The lakes’ relative warmth and resulting enhanced low-level convergence make the basin a preferred region of wintertime cyclogenesis (Petterssen and Calabrese 1959; Colucci 1976; Eichenlaub 1979). Lake-induced precipitation peaks during September–March when cloud cover and precipitation are enhanced downwind of the lakes (Niziol et al. 1995; Scott and Huff 1996; Kristovich and Laird 1998). Overlake turbulent fluxes and lake-effect precipitation are dampened by mid- to late winter (February–March) as ice cover becomes extensive (Niziol et al. 1995; Brown and Duguay 2010).

The Great Lakes region has experienced dramatic climatic and limnologic changes (Kling et al. 2003; Wuebbles and Hayhoe 2004; Wuebbles et al. 2010; Sharma et al. 2018), including a regime shift in lake-surface temperature (LST) and ice cover (Van Cleave et al. 2014). During 1900–2010, annual air temperatures rose by 0.88°C in the Midwest United States (Kunkel et al. 2013; Schoof 2013; Pryor et al. 2014; Zobel et al. 2017, 2018). Due to mutual surface–atmosphere warming (Manabe and Wetherald 1967) and resulting earlier lake stratification, Lake Superior’s surface water temperatures increased by 2.5°C during July–September of 1979–2006, exceeding the regional atmospheric warming rate (Austin and Colman 2007; Zhong et al. 2016; Ye et al. 2019). The lakes’ ice cover declined by 71% during 1973–2010 due to the aforementioned mutual surface–atmosphere warming (Wang et al. 2012; Mason et al. 2016). Rising lake temperatures, ice cover reductions, and increased frequency of intense cyclones supported a long-term positive trend in lake-effect snowfall (Burnett et al. 2003; Ellis and Johnson 2004; Kunkel et al. 2009), which locally reversed over portions of the Great Lakes Basin in recent decades (Bard and Kristovich 2012; Hartnett et al. 2014; Suriano and Leathers 2017; Clark et al. 2020). Heavy precipitation events have become more frequent (Kunkel et al. 2003, 2012; Easterling et al. 2000; Winkler et al. 2012), with an invigorated hydrologic cycle generating extreme lake level variations (Gronewold et al. 2013).

Given the importance of lake–atmosphere interactions and pronounced climate change in the Great Lakes Basin, there is a need to generate, evaluate, and improve climate modeling for the region. Large lakes and their regional climate influence are poorly resolved in coarse global climate models (Mallard et al. 2014, 2015; Briley et al. 2017). The Great Lakes’ representation across the Coupled Model Intercomparison Project global climate models varies broadly among land, wet soil, ocean, or inland lake grid cells, with the most advanced representation in the Coupled Model Intercomparison Project global climate models based on 1D lake models (none are coupled to 3D lake models) with inappropriate assumptions for deep lakes (Roeckner et al. 2003; Briley et al. 2017). One rudimentary regional climate modeling approach consists of extracting sea surface temperatures from the initial and lateral boundary conditions datasets over the Atlantic Ocean, Pacific Ocean, or Hudson Bay and applying those oceanic sea surface temperature values as LST boundary conditions for the Great Lakes (Mallard et al. 2015; Spero et al. 2016; Sharma et al. 2018). Such erroneous LSTs, retrieved from oceans rather than lakes, can negatively impact simulated pressure and air temperature regionwide (Spero et al. 2016). Alternatively, regional climate models that apply historical, remotely sensed or reanalysis-based LSTs, rather than a coupled lake model, neglect hydrodynamic feedbacks and are impractical tools for developing climate projections (Sharma et al. 2018).

Regional climate models have been employed in an array of Great Lakes studies. Zhong et al. (2012) demonstrated the ability of select regional climate models to capture the lakes’ impacts on regional climate and outperform global climate models. The Regional Climate Model version 4, coupled to a 1D lake model, was applied to examine the lakes’ influence on atmospheric circulation, stability, moisture, and temperature; highlight model skill in capturing variability and trends in air temperature, ice cover, and snowfall; elucidate the mechanisms behind recent lake warming; and formulate winter severity projections (Notaro et al. 2013a,b, 2014, 2015, 2016; Zhong et al. 2016). Applying the “Providing Regional Climates for Impacts Studies” regional climate model, Zhang et al. (2020) projected that wintertime precipitation in the Great Lakes Basin would increase during this century. The Weather Research and Forecasting (WRF; Skamarock et al. 2008) Model is a commonly used regional climate model for the Great Lakes Basin. According to Shi et al. (2010), the nested WRF Model with 1-km grid spacing accurately simulated snowfall and cloud patterns from Canadian snowstorms. Wright et al. (2013) revealed a close association between Great Lakes’ ice cover distribution and resulting snowfall pattern in WRF and concluded that coarse models cannot capture local water–ice–atmosphere interactions that regulate snowband intensity and distribution. Insua-Costa and Miguez-Macho (2018) estimated that, during lake-effect snowstorms in November 2014, 30%–50% of WRF-simulated precipitation downwind of the lakes originated from lake evaporation, similar to those estimated from observed water and ice fluxes (e.g., Kristovich and Braham 1998) Applying nested WRF with 3-km grid spacing, Shi and Xue (2019) determined that resolving LST spatial variations enhances surface wind convergence, vertical motion, and lake-effect snowfall on the lee sides of the Great Lakes. The WRF-based findings of Sharma et al. (2019) included enhanced skill due to spectral nudging (Rockel et al. 2008; Wang and Kotamarthi 2013), better performance during winter than summer, and successfully simulated lake-effect precipitation at both 12- and 4-km grid spacing. Complex lake–atmosphere interactions and lake-effect snowfall morphology require high-resolution modeling (Notaro et al. 2013a,b, 2015; Wright et al. 2013; Briley et al. 2017; Xiao et al. 2018; Shi and Xue 2019). Future climate projections for the Great Lakes Basin were developed by Gula and Peltier (2012) and Peltier et al. (2018) using WRF either uncoupled or coupled to the Freshwater Lake Model (Mironov 2008). Peltier et al. (2018) identified a wintertime cold bias in WRF coupled to the Freshwater Lake Model across the Great Lakes Basin.

More advanced regional climate models typically represent the Great Lakes using 1D lake models, which incorporate coupled lake–atmosphere interactions and can generally capture the broad spatiotemporal patterns of LSTs and ice cover (Gula and Peltier 2012; Notaro et al. 2013b), but are characterized by serious limitations. These shortcomings for large lakes include the lack of dynamic lake circulation, explicit horizontal mixing, or ice motion; an oversimplified stratification process; assumed instantaneous mixing of instabilities; and deficient treatment of eddy diffusivity (Martynov et al. 2010; Stepanenko et al. 2010; Bennington et al. 2014; Mallard et al. 2014, 2015; Gu et al. 2015; Sharma et al. 2018). Such regional climate models, coupled to a 1D lake model, generate excessive ice cover due to the absence of horizontal mixing and ice movement (Bennington et al. 2010; Notaro et al. 2013b; Xiao et al. 2016). One-dimensional lake models commonly produce an anomalously early stratification and positive bias in summertime LST (Bennington et al. 2014). Charusombat et al. (2018) revealed that WRF coupled to a 1D lake model, adapted from the Community Land Model version 4.5 (Subin et al. 2012; Oleson et al. 2013), produces excessive sensible and latent heat fluxes, compared to Great Lakes Evaporation Network measurements, that can be largely resolved by modifying the roughness length scales. One common approach to reduce vertical temperature profile errors in 1D lake models is to artificially enhance the vertical eddy diffusivity of deep lakes to imitate the neglected dynamic circulation and vertical mixing processes (Subin et al. 2012; Bennington et al. 2014; Lofgren 2014; Gu et al. 2015;