PNNL

Abstract

Prior research indicates that land-use and land-cover change (LULCC) in the Central United States (US) has led to significant changes in surface climate. The spatial resolution of simulations is particularly relevant in this region due to its influence on model skill in capturing mesoscale convection systems (MCSs) and on representing the spatial heterogeneity. Recent advances in Earth System Models (ESMs) make it feasible to use variable resolution (VR) meshes to study regional impacts of LULCC, while avoiding inconsistencies introduced by lateral boundary conditions typically seen in limited area models. Here, we present numerical experiments using the Community Earth System Model version 2 - VR to evaluate (1) the influence of resolution and land use on model skill, (2) impacts of LULCC and their resolution dependence over the Central US. These simulations are configured either on the 1° grid or a VR grid with grid refinement to 1/8° over the Contiguous US for the period of 1984-2010 with two alternative land use datasets corresponding to the preindustrial and present day states. Our results show that skill in simulating precipitation over the Central US is primarily dependent on resolution, whereas skill in simulating 2-m temperature is more dependent on accurate land use. The VR experiments show stronger LULCC induced precipitation increases over the Midwest in May and June, corresponding to an increase in the number of MCS-like features and a more conductive thermodynamic environment for convection. Our study demonstrates the potential of using VR ESMs for hydro-climatic simulations in regions with significant LULCC.