The statistical significance of the effects of including different protein sources in poultry diets on the environmental
impacts Global Warming Potential (GWP), Eutrophication Potential (EP) and Acidification
Potential (AP) of typical UK broiler meat and egg production systems was quantified using the Life Cycle
Assessment (LCA) method combined with an uncertainty analysis. The broiler and layer diets compared
in the study were either standard soya-based, or alternative diets based on European-grown protein
crops, including field beans, field peas, sunflower meal and whole rapeseed. Different methods for
accounting for land use change (LUC) in feed crop production were applied, including (1) a weighted
average of ‘‘new’’ and ‘‘mature’’ agricultural land used for soya production (‘‘best estimate’’ scenario),
(2) assuming no LUC in the production of soya used in these diets (‘‘sustainable soya’’ scenario) and
(3) including indirect LUC for all arable crop production (‘‘top-down’’ scenario). Monte Carlo simulations
were used to quantify uncertainties in predicted impacts and to perform statistical comparisons between
the effects of different diet compositions. The results showed that when included at relatively high levels
in the diets (10–30% by mass), peas, beans and rapeseed could slightly reduce the simulated mean value
of GWP (up to 12%) of broiler meat and egg production. However, when uncertainties in the data were
taken into account, these reductions were not statistically significant. Furthermore, the reduction in
GWP strongly depended on the method of LUC accounting applied in the analysis. With the ‘‘sustainable
soya’’ and ‘‘top-down’’ scenarios, only small, non-significant differences between the different diets were
found. In the case of EP, only small non-significant changes could be achieved with the alternative protein
sources. For AP, a significant reduction of more than 20% could be achieved if the crude protein content of
the broiler diet was reduced by using peas in combination with pure amino acids. This study demonstrates
the importance of a holistic approach, coupled with Monte Carlo uncertainty analysis, to evaluate
the environmental impacts of livestock systems. It takes into account the environmental burdens related,
for example, to feed production and transport and differences in emissions from housing and the end use
of the manure. Furthermore, due to the systematic uncertainty analysis, the statistical significance of the
effects of different feeding scenarios can now be evaluated.