Nitrous oxide (N₂O) emissions from semi-arid arable soils are small and highly variable. Consequently, any difference between agronomic treatments can be difficult to evaluate. However, accurate measurement of the difference in emissions in response to fertiliser application is critical when determining appropriate fertilizer emission factors for greenhouse gas inventories. We successfully measured N₂O emissions from an alkaline vertisol in semi-arid south-eastern Australia growing winter wheat (Triticum aestivum), either with or without urea fertilizer (50 kg N.ha^-1) applied at sowing. Continuous gaseous monitoring using a tuneable diode lazer, measured N₂O flux over 30 minutes every 90 minutes in nine automated chamber enclosures. The measured flux was highly variable but covariance among the large number of measurements made each day enabled the fitting of cubic smoothing splines and the evaluation of treatment differences by linear mixed model analysis. The application of urea fertilizer at sowing increased N₂O emissions for approximately two months, over two consecutive cropping seasons (measured from sowing to harvest). The increased emissions occurred in two generalised periods that were thought to correspond to nitrification based emissions in the first month following application of the fertilizer, which was then followed by a period of denitrification based emissions. The estimated emissions attributed to the fertilizer ranged between 41 and 111 g N₂O-N ha^-1, with an average of 75 g N₂O-N ha^-1, which was 0.15% of the added fertilizer, confirming that a reduced fertilizer emissions factor is appropriate for N fertilizer applied to semi-arid rain-fed grain crops in Australia.