Comparing the environmental impacts of UK turkey production systems using analytical error propagation in uncertainty analysis

Abstract

The aim of this study was to (1) quantify the environmental impact of UK turkey systems and (2) develop a methodology for analytical uncertainty analysis, as currently error propagation methods for such analyses of environmental impacts of agricultural commodities rely on time consuming Monte-Carlo approaches. The turkey systems considered were: 1) Stags (males) with controlled ventilation, 2) Hens (females) with controlled ventilation, 3) Stags with natural ventilation, and 4) Hens with natural ventilation, all being the main UK turkey production systems. An LCA modelling framework, based on a system approach and mechanistic sub-models was applied to assess several environmental impact categories, expressed per unit of live weight, and their associated uncertainties. For the first time, detailed production data and their variations from the industry, including slaughter age and weight, feed composition and consumption, mortality, and farm energy use, were used as input. The statistical significance of the differences between the systems was analysed using an analytical “top–down” method for uncertainty analysis, developed in this study. The results show that there were only small, mainly non-significant differences in impacts between the systems, affected mainly by their feed conversion ratio and slaughter weight, both of which were generally higher in the stag systems than the hen systems. A significant difference (P < 0.05) between the systems was found only in Acidification Potential, for which the stag system with controlled ventilation had a higher impact (88 ± 4.5 kg SO2 equivalent per 1000 kg live weight at farm gate) than the hen system with natural ventilation (72 ± 6.3 kg SO2 equivalent). For the other impacts, the average Primary Energy Use varied from 18,000 to 21,600 MJ, Global Warming Potential from 4000 to 4600 kg CO2 equivalent and Eutrophication Potential from 26 to 31 kg PO43− equivalent per 1000 kg live weight at farm gate, depending on the system (without any statistically significant differences). As a central outcome of this study, the development of the novel uncertainty analysis method makes it possible to precisely quantify the overall uncertainties of outputs of complex systems models, without the need for time-consuming Monte Carlo simulations, thus allowing statistical comparisons between different systems and scenarios.