Environmental Engineering Reference
In-Depth Information
option by including estimated cost functions for dairy products from the European
Dairy Industry Model (EDIM) to parameterize dairy processing. Furthermore,
such a statistical approach is also applied to represent the relevant part of
the simulation behaviour of the bio-physical model DNDC (DeNitrification-
DeComposition [DNDC] model, Li et al.
1994)
into the spatially explicit CAPRI
results (see further below).
The second option has been used to link economic models working at
different
spatial scales
, such as the breakdown of results at NUTS 2 regional or farming
system level in CAPRI, by using regional or farm accounting statistics (for CAPRI
farm type layer, see Britz et al.
2007
: 107-114); at
different time scales
, such as the
calibration of agricultural production to econometric projections or expert knowledge
at a future point in time (e.g. baseline calibration to DG-AGRI medium-term
market outlooks, see European Commission
2007b
: 38-45), and
different scope
,
such as the link between the CAPRI aggregated agricultural sector and GTAP as
attempted in SEAMLESS-IF (see Fig. 11.
2
). By doing so, the aggregation bias of
single model components is reduced, more flexibility in methodology is achieved
and different data sets can be better exploited without forcing the development of a
'one-fits-all model'. Nevertheless, within this type of framework, it is important to
be consistent from an ontology perspective (Chapter 9), i.e. with a clear data structure,
so that similar scenarios, model variables and parameters are correctly exchanged.
In Fig.
11.2
, the main interactions between model components, in both the CAPRI
and SEAMLESS-IF model chains, are presented.
SEAMLESS-IF consists of a vertical concatenation of models whereas the CAPRI
model chain presents a more 'spider-like' structure, with vertical (e.g. DNDC) and
horizontal links (e.g. factor markets and bioenergy processing linked to the supply
module) to models and databases. The part of CAPRI included in SEAMLESS-IF,
i.e. the SEAMCAP model component, is comprised in a green box. All model
components in the CAPRI model chain outside this box are not included in
SEAMLESS-IF. The arrows represent flows of information between modules
(one or both directions). The dotted lines connecting model components between
both model chains indicate a significant similarity of functionality within the
respective modelling chain (for example, DNDC and APES, both dealing with
environmental externalities at lower level).
The Farm Type Models in Both Systems
NUTS 2 level regions are often characterized by a mix of rather diverse farming
systems, so that simulating the average regional impacts of policies, e.g. on the
environment, may hide important differences between farms of different specialization.
As a response to a request of DG Environment, the CAPRI network developed a
farm type layer in 2003 where the database and the supply model were consistently
broken down up to eight different farm types, based on FADN (Farm Accounting
Data Network) farm typology and data. These single farm type models are based
