The Relevance of Coupled Climate Model WRF-ELM for Atmosphere-Coastal-Urban-Rural Interaction Analysis

The land component (ELM) of the Energy Exascale Earth System Model (E3SM) is a cutting-edge model under active development with several state-of-the-art capabilities and features, including ecosystem-based hydrology, subsurface multiphase flow, and topography-based subgrid representation. ELM can be run either offline or online coupled within the E3SM framework. Being part of an Earth system model that is usually run at coarser scales leads to difficulties when simulating fine-scale energy budget and water cycle in coupled configurations since the atmospheric forcing is identical over all subgrids within the coarse model grid. Given the importance of capturing the heterogeneity of land-atmosphere interactions, especially along coastal interfaces and across land cover transition zones, here we develop this capability by incorporating ELM within a regional climate modeling framework by coupling it with the Weather Researcher and Forecasting (WRF) model through the Lightweight Infrastructure for Land Atmosphere Coupling (LILAC) from National Center for Atmospheric Research (NCAR). The LILAC coupler comprises two components: a top-level driver for variable handling and memory allocation from WRF, and the ESMF-capping for ELM workflow control, encompassing initialization, execution, and finalization. Remarkably, this coupling framework preserves the core ELM structure, allowing external models to be readily integrated with WRF-ELM.

We systematically assess the capability of WRF-ELM in capturing land-atmosphere interactions within complex domains by conducting numerical simulations at a spatial resolution of 4 km over the Great Lake Region (GLR). These simulations include 5 ensemble members for the 2018 summer, starting with initial conditions 12 hr apart between 0000 UTC on 12 May 2018 and 1200 UTC on 14 May 2018 and ending on 0000 UTC 1 September 2018. Essential energy budget and water cycle components, such as surface albedo, latent/sensible heat fluxes, precipitation, near-surface air temperature, and near-surface wind, are chosen for comprehensive model skill analysis. We compare the performance of WRF-ELM against WRF-CTSM (WRF coupled with the Community Terrestrial Systems Model). Initial findings reveal consistent latent heat flux simulations from both models over land and lake surfaces, with WRF-ELM exhibiting lower albedo and latent heat over urban and rural areas surrounding the Great Lakes. Notably, WRF-ELM displays finer spatial heterogeneity and smoother gradients compared to WRF-CTSM. This distinction arises from WRF-ELM's retention of subgrid land unit descriptions and the plant functional types (PFTs), while WRF-CTSM employs a single dominant land unit and PFT. This coupling effort establishes a platform for applying the most advanced ELM in regional climate modeling and potentially expands the utilization of new land surface parameterizations in regional studies.