Urea is frequently surface applied to winter wheat (Triticum aestivum L.) during cold weather months (October–April) in the semiarid northern Great Plains. This study was conducted to quantify NH3 volatilization loss from surface-applied urea (100 kg N ha−1) and urea amended with N-(n-butyl) thiophosphoric triamide (NBPT) during this period. Ammonia emissions were quantified by a micrometeorological integrated horizontal flux method with samplers placed on a mast in the center of circular plots (20-m radius). Cumulative NH3 losses from urea varied but averaged 20.5% of applied N across 12 trials. The largest losses (30–44% of applied N) occurred after urea was applied to high-water-content soil surfaces, followed by a period of slow drying with little or no precipitation. Emissions occurred for a prolonged period often lasting >42 d. Periods (1–2 wk) of high NH3 flux (>30 g N ha−1 h−1) frequently occurred when mean daily soil temperatures (1-cm depth) were –2 to 5°C. In one trial, 24.3% of applied N was lost after urea was applied to a 140-mm snowpack. Ammonia losses were moderated by applying urea to dry soil surfaces. If precipitation events that followed were light (<8 mm), losses were reduced to 10 to 20% of applied N. If the events were heavy (>18 mm), then losses were <10%. Coating urea with NBPT (1 g kg−1) reduced cumulative NH3 losses by 66%. Volatilization protection lasted 2 to 3 wk on acidic soils, and >7 wk on an alkaline soil. This study demonstrated that significant NH3 losses from surface-applied urea could occur during cold weather months.