Comparing GHG emissions of dairy cow production systems and
exploring their potential of GHG emission leakage is a three-step
process. First, to identify “hot spots” (i.e. parameters that
contribute most to GHG emissions), methods such as “all GHGs to
milk” should be used to calculate GHG emissions based on the system
boundary of the dairy farm gate (Moran et al., 2011). In the
second step, systemexpansion can be used to ensure that production
systems with lower farm-gate GHG emissions do not inadvertently
increase overall GHG emissions due to shift of GHG emissions to
other food production sectors or countries (GHG emission leakage).
In the third step, stochastic models are particularly useful because
they provide insight into the robustness of model predictions
(Pannell, 1997). This study demonstrates the importance of taking
into account epistemic and variability uncertainties and one possibility
of GHG emission leakage (i.e. shift of GHG emissions from dairy
beef production to suckler beef production systems).
However, differences in milk yield are likely to lead to leakage
effects not only in beef production, but also e.g. in land use. Future
studies should explore additional leakage effects and epistemic and
variability uncertainties, e.g. of feed intake, cattle fattening systems
and manure management.
There is a big interest in quantifying carbon footprints. However,
our study shows that current LCA methods are not very precise
because of large epistemic and variability uncertainties. The
implementation of a single carbon footprint for different dairy cow
production systems is problematic because of these uncertainties
but also due to various other environmental impacts (e.g. biodiversity,
nitrogen leaching).