A model that treats nesting of gas, size- and composition-resolved aerosol, radiative, and meteorological parameters from the global through urban scales (<5-km grid spacing) was developed. The model treats multiple one-way-nested layers and multiple air quality and meteorological domains in each layer between the global and the urban scales. This latter feature allows forecast of air pollution and weather at several urban or regional sites during the same simulation. Regardless of the number of domains used during a single continuous simulation, the central memory required never exceeds 1.5 times and 2.1 times that of the largest domain for gas and gas/aerosol simulations, respectively. A submodule was developed for all domains to treat ground temperatures, latent heat fluxes, and sensible heat fluxes over subgrid soil types (with and without vegetation), water, sea ice, and urban areas. Urban areas are divided into road surfaces, rooftops, vegetation, and bare soil. Snow is treated over all surface types. The global-through-urban model is applied in a companion paper to study elevated ozone, ozone in national parks, and weather during a field campaign in northern and central California.