LCA Food 2012 in Saint Malo, France! - Manifestations et colloques ...

PLENARY SESSION 2: METHODOLOGICAL CHALLENGES FOR ANIMAL PRODUCTION SYSTEMS8 th Int. Conference on LCA in the
Agri-Food Sector, 1-4 Oct 2012
Modelling GHG emissions of dairy cow production systems differing
in milk yield and breed – the impact of uncertainty
216
Monika Zehetmeier *1 , Markus Gandorfer 2 , Alois Heißenhuber 1 , Imke J.M. de Boer 3
1 Institute of Agricultural Economics and Farm Management, Technische Universität München, 85350 Freising, Germany
2 Chair Group Economics of Horticulture and Landscaping, Technische Universtät München, 85350 Freising, Germany
3 Animal Production Systems Group, Wageningen University, Wageningen, The Netherlands
Corresponding author. E-mail: monika.zehetmeier@tum.de
ABSTRACT
The main objective of our study was to incorporate uncertainty in modelling greenhouse gas (GHG) emissions of dairy cow production
systems differing in milk yield and breed. Stochastic simulation was undertaken to account for uncertainty of main model assumptions.
The developed stochastic model accounts for two different methods for handling co-products of dairy farming (beef,
surplus calves): economic allocation and system expansion. Whereas the choice of method for co-product handling depends on the
scope of GHG modelling the stochastic model approach gave an insight into robustness and variation of model outcomes within each
method for handling co-products. The method of system expansion is recommended if the consequences of changes or mitigation
options in dairy cow production need to be evaluated. In that case stochastic models offer the advantage of predicting not just an
outcome, but also the likelihood of this outcome. This is of special importance identifying cost-effective GHG mitigation options.
Keywords: greenhouse gas emissions, uncertainty, stochastic modelling, dairy cow production
1. Introduction
Dairy cow production contributes to about 23 to 70% of total agricultural GHG emissions in different
countries within the EU-27 (Lesschen et al., 2011). Thus a growing interest can be observed in modelling
GHG emissions from dairy cow production systems and identifying cost effective GHG abatement options.
As milk is the main output of dairy farms most studies express GHG emissions produced per kg milk delivered.
However, beef can be considered as an important co-product of dairy farming (beef from culled
cows and surplus calves sold to fattening systems) especially within dual purpose dairy cow production systems.
To account for co-products from dairy farming different methods can be observed in literature (Flysjö
et al., 2011). Two main approaches can be distinguished: economic allocation and system expansion. In case
of economic allocation GHG emissions are allocated between milk and co-products at the dairy farm gate
according to their economic value. This approach is mainly used in the calculation of carbon footprints. It
identifies GHG emissions at the dairy farm gate caused by milk production and allocates GHG emissions
based on the value of milk and beef to the consumer. In case of system expansion allocation between milk
and co-products is avoided by expanding the system and accounting for the alternative way of beef production
(i.e. sucker cow production). It is assumed that the beef derived from culled cows and fattening of surplus
calves replaces beef from suckler cow production. The avoided GHG emissions are credited to the dairy
farm. The method of system expansion is recommended by the International Organisation for Standardization
(ISO, 2006). This approach is especially important if the consequences of changes or mitigation options
in dairy cow production need to be evaluated (Flysjö et al., 2011).
Recent determinist studies showed that the choice of method for co-product handling has a major impact
on GHG emissions outcomes of dairy cow production systems (Flysjö et al., 2011, Zehetmeier et al., 2012).
Despite the impact of choice of method for co-product handling it has to be considered that assumptions and
input data modelling GHG emissions from dairy cow production have known uncertainties.
Many guidelines and scientific studies point out the importance of incorporating uncertainty in GHG and
economic modelling (ISO, 2006; IPCC, 2006; Pannell, 1997).
The inclusion, the discussion and the reporting of model changes due to uncertainties can be important to
identify robustness and variation of model outcomes and sensitive or important variables (Pannell, 1997). To
show the impact of uncertainty on GHG emission outcomes a deterministic model developed to calculate
GHG emissions of confinement dairy farm systems differing in milk yield and breed (Zehetmeier et al.,
2012) was further developed. A stochastic model was established that accounts for uncertainty in various
components. Compared with deterministic models, stochastic models offer the advantage of predicting not
just an outcome, but also the likelihood of this outcome. Thus, stochastic modelling was undertaken to answer
the following questions:
- Does the inclusion of uncertainty influence the ranking of modelled dairy cow production systems in
terms of GHG emissions? (6000, 8000, 10000 kg milk/cow per year)
- which uncertainties have the highest impact on variation of GHG emission outcomes?