Mesoscale convective systems (MCSs) occur frequently during spring and summer in the US east of the Rocky Mountains. With heavier rain rates and larger area coverage compared to isolated deep convection (IDC), MCSs are important contributors to flooding with costly consequences. Observations show increasing trends in MCS occurrence frequency and rainfall amount in the past two decades, motivating the need to understand potential future changes in MCSs and their characteristics. As spring MCSs in the US are strongly driven by large-scale environments associated with synoptic fronts and the Great Plains low-level jet (GPLLJ), future changes in spring MCS may be inferred from changes in the large-scale environments projected by global climate models. For summer MCSs which are less influenced by the large-scale environments, cloud-resolving models provide an important tool for projecting convective storm changes in the future. Here changes in convection populations are simulated using a global model with regional refinement at ~3 km resolution over the US. Tracking of MCS in the simulations allows us to isolate the changes in MCS vs. IDC, including their frequency of occurrence, duration, rainfall amount, and other properties. Simulations show a potential shift between MCS and IDC populations and contrasting changes in inland vs. coastal regions. To isolate the thermodynamic influence on the convection populations, conceptual models of IDC and MCS based on single column and multi-column parcel models, respectively, are used to simulate the changes in IDC and MCS as driven by the changes in the thermodynamic environments simulated by the cloud-resolving models. Results indicate important role of thermodynamic changes in explaining the contrasting changes in MCS and IDC and in inland vs. coastal regions while changes in atmospheric circulations cannot be neglected, underscoring the complexity of how convection populations are impacted by environmental changes over land.