The atmospheric concentration of nitrous oxide (N2O), a potent greenhouse gas and ozone-depleting chemical, continues to increase, due largely to the application of nitrogen (N) fertilizers. While nitrite (NO2-) is a central regulator of N2O production in soil, NO2- and N2O responses to fertilizer addition rates cannot be readily predicted. Our objective was to determine if quantification of multiple chemical variables and structural genes associated with ammonia (NH3)- (AOB, encoded by amoA) and NO2--oxidizing bacteria (NOB, encoded by nxrA and nxrB) could explain the contrasting responses of eight agricultural soils to five rates of urea addition in aerobic microcosms. Significant differences in NO2- accumulation and N2O production by soil type could not be explained by initial soil properties. Biologically-coherent statistical models, however, accounted for 70-89% of the total variance in NO2- and N2O. Free NH3 concentration accounted for 50-85% of the variance in NO2- which, in turn, explained 62-82% of the variance in N2O. By itself, the time-integrated nxrA:amoA gene ratio explained 78 and 79% of the variance in cumulative NO2- and N2O, respectively. In all soils, nxrA abundances declined above critical urea addition rates, indicating a consistent pattern of suppression of Nitrobacter-associated NOB due to NH3 toxicity. In contrast, Nitrospira-associated nxrB abundances exhibited a broader range of responses, and showed that long-term management practices (e.g., tillage) can induce a shift in dominant NOB populations which subsequently impacts NO2- accumulation and N2O production. These results highlight the challenges of predicting NO2- and N2O responses based solely on static soil properties, and suggest that models that account for dynamic processes following N addition are ultimately needed. The relationships found here provide a basis for incorporating the relevant biological and chemical processes into N cycling and N2O emissions models. Published by Elsevier Ltd.