Agricultural best management practices (BMPs) intended to solve one environmental challenge may have unintended climate impacts. For example, manure injection is often promoted for its potential to reduce runoff and N loss as NH3, but the practice has been shown to increase N2O, a powerful GHG, compared to surface application. Urease inhibitor application with N fertilizer is another BMP that can enhance N retention by reducing NH3 emissions, but its impact on N2O emissions is mixed. Thus, we measured N2O, CO2, soil mineral N availability, soil moisture, soil temperature, and yield in a two‐year perennial hayfield trial with four fertilization treatments (manure injection, manure broadcast, synthetic urea, and control) applied with or without a urease inhibitor in Alburgh, VT. We used linear models to examine treatment effects on daily and cumulative N2O emissions and a boosted regression tree (BRT) model to identify the most important drivers of daily N2O fluxes in our trial. While fertilization type had a significant impact on N2O fluxes (p<0.05), our treatments explained an unexpectedly small amount of the variation in emissions (R2 = 0.042), and urease inhibitor had no effect. Instead, soil moisture was the most important predictor of daily N2O fluxes (39.7% relative influence in BRT model), followed by CO2 fluxes, soil inorganic N, and soil temperature. Soil moisture and temperature interacted to produce the largest daily N2O fluxes when both were relatively high, suggesting that injecting manure during dry periods or during wet but cool periods could reduce its climate impacts.This article is protected by copyright. All rights reserved