The experiment was conducted at the J. Bjelke Petersen Research Station of the Department of Agriculture, Fisheries and Forestry (DAFF). The research site is located in Kingaroy (26°58’16,8’’ Latitude South, 151°82’85.3’’ Longitude East, altitude 441 m a.s.l), in the southern inland Burnett region of southeast Queensland, Australia. The climate is subtropical (Cfa, according to Köppen climate classification) with warm, humid summers and mild winters. Daily mean maximum and minimum temperatures range from 20.1°C to 4.0°C in winter and from 29.6°C to 16.5°C in summer, respectively. Local mean annual precipitation is 776.2 mm and varies from a minimum of 28.6 mm in August to a maximum of 114.1 mm in January (Australian Bureau of Meteorology). The soil is a Brown Ferrosol (Australian Soil Classification of Isbell (2002)), characterized by a relatively slow permeability and a high clay content (50-65% clay). The effective rooting zone reaches 1.4 m of and the water holding capacity is 100 mm. The physical and chemical soil properties of the field site are listed in Table 1.

The experiment was arranged as a randomized complete block design with three replicates per treatment. Each plot measured 12 m in length x 7.2 m wide and included eight crop rows. Buffer zones of 1.8 m and 1.5 m separated the parcels along their width and length, respectively.N2O emissions were measured in plots planted with sorghum (Sorghum bicolor L.) following two distinct cropping histories. One crop rotation (hereafter called legume ley pasture-cereal) included two seasons of alfalfa pasture (Medicago sativa, L., summers 2009/2010 and 2010/2011), one season of maize (Zea mays, L., summer 2011/2012) and one season of sulla pasture (Hedysarum coronarium L., winter 2012) prior to sowing sorghum. The other crop rotation (hereafter called grass ley pasture-cereal) consisted of two seasons of a mixed rhodes grass (Chloris gayana, K.) and alfalfa pasture (summers 2009/2010 and 2010/2011), one season of maize (Zea mays, L., summer 2011/2012) and one season of wheat (Triticum aestivum L., winter 2012). Even though the mixed pasture alfalfa was sown in consociation with rhodes grass, rhodes grass took rapidly over and by the end of the first season the pasture wall almost completely composed by rhodes grass. All crops in both rotations were unfertilised. Sulla and wheat were direct drilled in August 2012 and managed as forage crops. Both crops were terminated in November 2012 with all residues returned to the soil as a mulch before being incorporated into the soil with four shallow cultivations (20 cm). The incorporation of sulla residues (2.3 t dry mater ha-1, 1.57% N) was estimated to supply to the soil approximately 36 kg N ha-1, while wheat residues (1.24 t dry mater ha-1, 0.75% N) about 9 kg N ha-1. The whole field trial was irrigated with 20 mm three days before plots were planted to sorghum on 12 December 2012. Sorghum (cultivar Pioneer G22) was planted with a plant density of 7 plants/m2 and an inter-row space of 98 cm.

Treatments G100 and L70 were base dressed at planting banding 20 kg N ha-1 as urea. On 15 January 2013, at growth stage 3 (eight leaf stage), both treatments were inter-row cultivated and side dressed with banded urea, receiving 80 kg N ha-1 (G100) and 50 kg N ha-1 (L70). The N application rate for G100 was designed to achieve maximum yield potential and was representative of farming practices of the region. The synthetic N rate used in L70 was reduced compared to G100 to assess whether the estimated 30 kg N ha-1 resulting from the mineralisation of the sulla residues would have been available to sorghum.To prevent water stress limiting the potential yields, the trial was irrigated three times over the season (25 mm on 18 December 2012, 40 mm on 4 January 2013, 40 mm on January 18 January). On each event all treatments were irrigated with sprinklers using surface stored dam water.Sorghum was harvested on 18 June 2013. The trial area was left fallow until being cultivated on 6 August 2013 (offset disc and chisel plough to a depth of 20 cm) and on 19 September 2013 (offset disc to a depth of 20 cm) to prepare the seedbed for the next crop. On 27 and 29 August 2013 two rain events were simulated irrigating all plots with 30 and 40 mm, respectively to assess whether significant amounts of N were still available for denitrification after harvest.