Mitigation of agricultural greenhouse gas emissions is a target area for the Australian Government and the pork industry. The present study measured methane (CH4), nitrous oxide (N2O) and ammonia (NH3) from a deep-litter piggery and litter stockpile over two trials in southern New South Wales, to compare emissions from housing pigs on deep litter with those of pigs from conventional housing with uncovered anaerobic effluent-treatment ponds. Emissions were measured using open-path Fourier transform infrared spectrometry, in conjunction with a backward Lagrangian stochastic model. Manure excretion was determined by mass balance and emission factors (EFs) were developed to report emissions relative to volatile solids and nitrogen (N) input. Nitrous oxide emissions per animal unit (1 AU ≤ 500 kg liveweight) from deep-litter sheds were negligible in winter, and 8.4 g/AU.day in summer. Ammonia emissions were 39.1 in winter and 52.2 g/AU.day in summer, while CH4 emissions were 16.1 and 21.6 g/AU.day in winter and summer respectively. Emission factors averaged from summer and winter emissions showed a CH4 conversion factor of 3.6%, an NH3-N EF of 10% and a N2O-N EF of 0.01 kg N2O-N/kg N excreted. For the litter stockpile, the simple average of summer and winter showed an EF for NH3-N of 14%, and a N2O-N EF of 0.02 kg N2O-N/kg-N of spent litter added to the stockpile. We observed a 66% and 80% decrease in emissions from the manure excreted in litter-based housing with litter stockpiling or without litter stockpiling, compared with conventional housing with an uncovered anaerobic effluent-treatment pond. This provides a sound basis for mitigation strategies that utilise litter-based housing as an alternative to conventional housing with uncovered anaerobic effluent-treatment ponds. © CSIRO 2016.