Land-use change is an important driver of soil-atmosphere gas exchange, but current greenhouse-gas budgets lack data from urban lands. Field comparisons of urban and non-urban ecosystems are required to predict the consequences of global urban-land expansion for greenhouse-gas budgets. In a rapidly urbanizing region of the U.S. Great Plains, we measured soil-atmosphere exchange of methane (CH4) and nitrous oxide (N2O) for one year in replicated (n = 3) urban lawn, native shortgrass steppe, dryland wheat-fallow, and flood-irrigated corn ecosystems. All soils were net sinks for atmospheric CH4, but uptake by urban, corn, and wheat-fallow soils was half that of native grasslands ($-0.30 \pm 0.04 g C\cdot m^{-2}\cdot yr^{-1}$mean ± 1 SE). Urban ($0.24 \pm 0.03 g N\cdot m^{-2}\cdot yr^{-1}$) and corn ($0.20 \pm 0.02 g N\cdot m^{-2}\cdot yr^{-1}$) soils emitted 10 times more N2O to the atmosphere than native grassland and wheat-fallow soils. Using remotely sensed land-cover data we calculated an upper bound for the contribution of lawns to regional soil-atmosphere gas fluxes. Urban lawns occupied 6.4% of a$1578-km^2$study region, but contribute up to 5% and 30% of the regional soil CH4consumption and N2O emission, respectively, from land-use types that we sampled. Lawns that cover small portions of the landscape may contribute significantly to regional soil-atmosphere gas exchange.