Watershed land use change indirectly dominated the spatial variations of CH4 and N2O emissions from two small suburban rivers

[Display omitted] •Suburban rivers are important CH4 & N2O emitters to the atmosphere.•Suburban rivers presented spatial non-continuity of CH4 & N2O emissions.•Urban expansion dominated the spatial pattern of CH4 & N2O fluxes.•CH4 and N2O fluxes functions based on urban land use were established.•Increasing urban pollutants were powerful driver of CH4 and N2O fluxes. Urban rivers have been proved to be the hot spots of atmospheric methane (CH4) and nitrous oxide (N2O) emissions. However, the rivers across rural-urban interface, which are undergoing significant environmental changes in the process of urbanization, received little attention. In this study, we conducted seasonal investigations in two small suburban rivers in mountainous southwest China, to clarify the spatial-temporal characteristics of CH4 and N2O concentrations and fluxes in relation to watershed land use change and their potential controls. The results showed that, the small suburban rivers acted as strong CH4 and N2O emitters to atmosphere with averaged fluxes of 960.2±966.6 and 205.1±240.8 µmol·m−2·d−1, respectively. The concentrations and fluxes of CH4 and N2O presented a sharp increase and spatial variability when the rivers cross the rural-urban interface. Such discontinuous spatial pattern of CH4 and N2O emission in river continuum were closely related to watershed land use, and could be modeled as functions of the urban land use proportion. The urban land proportion in the basins could explain 70% and 59% of the total spatial variations of CH4 and N2O emissions, respectively. Stepwise multiple regression models (SMRM) based on water environment parameters were established in our study and indicate that DTC, NH4+, NO3−, TP and conductivity play key controls on CH4 and N2O emissions in the two studied mountainous suburban rivers. Moreover, SMRM and urban land use based functions were tested to have good predictions based on constrained mountainous samples, could provide a possible path to extrapolate CH4 and N2O emissions in mountainous urbanized-rivers at the whole regional network scale. We also found evidences for the sewage-dominated tributary with abnormally high CH4 and N2O concentrations, which was partially responsible for the sharp increase of CH4 and N2O fluxes in urban reaches. The results of structural equation model provided strong evidence for direct and indirect controls of urban expansion on longitudinal differentiation of riverine CH4 and N2O emissions crossing rural-