Analysis of Dutch farm location and MAFF data in relation to CORINE

1
Summary of the ELPEN project results, May 2003
2
Introduction to the ELPEN project
3
The
Consortium
4
What is ELPEN?

• A network that
• can answer specific questions about the economic,
  environmental and social impacts of policy
  related to livestock systems
• uses state of the art technology the ELPEN
  system

5
The ELPEN system was developed on the basis of
the End User Group requirements, which were

• Integrating economic, environmental and social
  impacts (DG Agri Env)
• Spatially explicit (DG Agri Env) regional
  scale
• Focus on land dependent systems initial focus
  on dairy systems (DG Agri)
• A flexible expert tool, not a desktop tool

6
The ELPEN system is

• A decision support system containing
• EU statistical and geographical data
• An ELPEN farm typology
• A reference farm database
• Expert knowledge (simple rules and models)
• Meta data explaining data and knowledge
• Fast procedures to process large amounts of data
• Procedures to display output maps and tables

7
The ELPEN system user is

• An expert who supports policy makers to answer
  specific questions by
• Adding and combining the stored EU statistical
  and geographic data
• Adding relevant expert knowledge and meta data
  for each new policy question
• Generating output maps and tables
• Reporting on the results
• Maintaining and improving the utility of the
  system for policy impact assessment

8
Policy Impact Assessment with the ELPEN system
Policyquestion
Spatially explicit socio-economic and
environmental impacts
9
Explanation of The ELPEN System
10
System data (e.g units)
Explanation of The ELPEN System
Statistical data
Geographic data
ELPEN Farm typology
Reference Farms (examples of ELPEN farm types)
Schemes (knowledge rules)
Cases (results, obtained by applying schemes on
data)
11
ELPEN system User interface
Mapview
Meta data view and technical details
Browser
Table view
12

• Statistical DataA part of the following EU
  data sets are incorporated in the ELPEN system
• Farm Accountancy Data Network (FADN)
  Commission of the European Communities, DG
  Agriculture in ELPEN (March 2003) 1990-91,
  1997-98, 1999-2000
• Farm Structure Survey (FSS/Eurofarm)
  Eurostat in ELPEN (March 2003) 1990, 1997
• Regional data bank (REGIO) Eurostat
  in ELPEN (March 2003) 1980 - 2001
• Relation tables Stored relations between
  different data sets

13
Statistical Data Relation tables
Stored relations between different data
sets e.g The geographic relations between
the different statistical data sets is
established by Harmonised regions (Harm)
14
Geographic relations harmonised regions
HARMonised regions make it possible to integrate
data from different statistical sources by
applying them to the same geographic entities
15
Geographic relations harmonised regions
HARMonised regions make it possible to integrate
data from different statistical sources by
applying them to the same geographic entities
16
Geographic relations harmonised regions
ELPEN regions a clustering of Harm regions, a
pragmatic solution to enlarge regions in order
toprevent disclosure problems when displaying
data concerning less than 15 farmsin one
region
Country12 15
EU
Other geographic regions in ELPEN
17
Example data
18
(No Transcript)
19
Example data
20
(No Transcript)
21
Administrative Regions Designated areas Land
cover Corine Pelcom Elevation Soil, Climate
Environment (Nitrogen Vulnerable Zones,
N-leaching) Landscape (not yet available)
22
(No Transcript)
23
(No Transcript)
24
An extract of the meta data on Corine Land Cover
(CLC) CLC was elaborated based on the visual
interpretation of satellite images (SPOT, LANDSAT
TM and MSS). Ancillary data (aerial photographs,
topographic or vegetation maps, statistics, local
knowledge) were used to refine interpretation and
the assignment of the territory into the
categories of the CORINE Land Cover nomenclature.
The smallest surfaces mapped (mapping units)
correspond to 25 hectares. Linear features less
than 100 m in width are not considered. The scale
of the output product was fixed at 1100.000.
Thus, the location precision of the CLC database
is 100 m.
25
An extract of the meta data on Pelcom The
Pelcom land cover database is calculated from
Earth Observation images using an algorithm that
computes the first and second minimum distances
for each AVHRR image pixel based on the spectral
signatures, and as a result, it derived the first
best class ('highest probability') and the second
best class ('second highest probability') for
each pixel.
26
(No Transcript)
27
This classification corresponds with the FADN
data
28
ELPEN Farm Typology why?

• to differentiate farms according to
• responses to policy
• impact on the social and bio-physical environment
• and
• to aggregate FADN farm level data into farm types
  that
• can be associated with reference farms in the
  field

29
ELPEN Farm Typology how?

• Selection of classifying variables using expert
  knowledge of systems in the field(e.g.
  CEAS-study 2000)
• Cluster analysis on FADN-data to determine
  usefulness of variables and threshold values

30
Sector Grazing livestock farms(gt50 production
value from grazing livestock)
ELPEN Farm typology result
31
How many and what ELPEN farm types are located
in what regions?
32
Sector Grazing livestock
Production type Meat
Land use type Permanent grass
Intensity Low input
Size Medium scale
Livestock type Cattle
What ELPEN farm types are in what regions?
33
example time-series
34
Example time seriesChange in nr of dairy cattle
farms 1990 - 1999 (Index 19901)
decreaseIncrease lt 15 dairy cattle farms or
missing data
35
For each (group of) farm type(s) in a region,
several profiles can be computed that
characterise these farm types

• Profiles
• Environmental
• Structural
• Economic
• Social
• Regional

36
(No Transcript)
37
Total livestock units present on all farms of
this type in the region of Acquitaine, being 1.2
of the total LU of all farms of this type in
the EU
38
Stocking density on grassland, being 6.6 under
EU average for this farm type
Total livestock units present on all farms of
this type in the region of Acquitaine, being 1.2
of the total LU of all farms of this type in
the EU
39
Reference farms areReal example farms of ELPEN
farm types
A) General questions B) Land related
questions C) Grass and Fodder D) Mechanisation E)
Dairy cows F) Beef cattle G) Sheep H) Goats J) Oth
er grazing livestock K) Livestock housing /
welfare L) Questions about Agricultural Policy
Changes
The data are gathered using a questionnaire
with 93 questions about 10 aspects ofthe farm
and 7 policy questions
40
Reference farms areReal example farms of ELPEN
farm types
A) General questions B) Land related
questions C) Grass and Fodder D) Mechanisation E)
Dairy cows F) Beef cattle G) Sheep H) Goats J) Oth
er grazing livestock K) Livestock housing /
welfare L) Questions about Agricultural Policy
Changes
The data are gathered using a questionnaire
with 93 questions about 10 aspects ofthe farm
and 7 policy questions
41
Data on reference farms can be viewed as follows
42
Data on reference farms can be viewed as follow
43
Data on Sheep
Data on reference farms can be viewed as follow
44
Schemes Schemes contain the knowledge rules that
are used to compute results Scheme-displays can
be generated automatically by the system e.g.
Rules to allocate livestock types per grid cell
45
Schemes Schemes contain the knowledge rules that
are used to compute results Scheme-displays can
be generated automatically by the system e.g.
Rules to allocate livestock types per grid cell
FSS data nr. lu dairy cows / Harm2 (1990)
46
Cases Cases use the schemes to compute results.
Different results can be computed,using the
same pre-programmed schemes and selecting
different input data At the end of the ELPEN
project (March 2003), the following cases were
stored in the ELPEN system
47
Cases overview

• Allocation of Livestock basis FSS / HARM2
  and Land cover on 1km2
• Allocation of ELPEN Farm type groups basis
  FADN / HARM1, Land cover, LFA, Altitude on 1km2
• Economic Impacts of milk price reduction
  basis economic model, input FADN data / farm
  type group
• Environmental impacts of milk price reduction
• Social who stops and who continues decision
  path for LFA regions, input reference farms

48
Case Allocation of livestock per grid
cell
49
Case allocation of farm type groups
50
Case Economic impact of milk price
reduction Example specialised dairy cattle farms
1) Computation of full costs of milk
input 250 FADN variables, e.g labour
land capital feed
51
Case Economic impact of milk price
reduction Example specialised dairy cattle farms
Small herd size
The results show that there is a correlation
between production costs and herd size
lt 15 spec. dairy farms
High production costs
(or non EU member Switzerland)
52
2) Computation of surplus surplus milk
price - costs In our example costs on long
termCash costs depreciation full
opportunity costs (land,
labour, capital, feed, etc)
Case Economic impact of milk price
reduction
53
Environmental Impact of milk price reduction
This is a political map, which is not very useful
for the impact assessment of milk price
reduction. Instead, we used the following map
54
Environmental Impact of milk price reduction
Fragment of the meta data on Nitrate
vulnerability To determine the nitrate leaching
fraction the Burns model is used (Burns, 1976).
This model, defined in 1975, calculates a nitrate
leaching fraction (f) on nitrate present in the
soil, using soil and climate factors. f (Ed /
(Ed Vm/100))x f nitrate leaching fraction
(-) Ed amount of water drain-ing to the sub
soil(cm/yr) Vm volumetric soil moisture content
at field capacity () x depth factor depending
on the nitrate repartition considered in the soil
(cm)
55
Environmental Impact of milk price reduction
56
Economic impact of milk price reduction on farm
types with environmental consequences
57
Economic impact of milk price reduction on farm
types with environmental consequences
58
Economic impact of milk price reductionon farm
type with environment consequences
EU average forall dairy cattle farms index 100
59
Assumption of previous cases Only dairy
farmers with a surplus income from milk
productionwill continue farming, but... Social
and cultural aspects mayhave a big influence on
decisionsof farmers. An example of rules that
simulate social behaviour is the following
60
(No Transcript)
61
The previous social scheme uses as input stored
reference farm data in the ELPEN system.
Unfortunately the number of reference farms in
the current system is too limited to apply social
schemes like these to predict representative
social behaviour.
We need to interview more farmers in order to
gather representative data to be able to assess
social impacts.
62
Conclusions

• The ELPEN consortium offers
• Support for integrated and targeted policy
  development by
• Making available integrated statistical and
  bio-geographical data for the whole EU in one
  system
• Providing regionally differentiated information
  in maps (using expert knowledge for
  desegregation)
• Differentiating farm responses using a flexible
  typology that can be associated with reference
  farms in the field
• Enabling a better understanding of linkages
  between economic, environmental and social
  impacts of policy measures
• Maintaining and improving the utility of the
  ELPEN system for policy impact assessment

63
Accessing ELPEN
The ELPEN software is available free of charge
athttp//www.objectvision.nl For general
information on the ELPEN project visitthe ELPEN
website http//macaulay.ac.uk/elpen If you are
interested in using our services or the ELPEN
system, please contact us
64
If you are interested in using our services or
the ELPEN system, please contact
us Organisation Country Contact person
Coordinates Macaulay Institute Scotland
Iain Wright Project Coordinator,
Reference Farm Collection Alterra, Green
World Research Netherlands Berien Elbersen
Geographic Data, Regional Indicators,
Desaggregation Procedures Agricultural
Economics Research Institute (LEI)
Netherlands Frans Godeschalk FADN data
provision, preparation of Eurostat data FSS
and Regio Danish Forest and Landscape Research
Institute (FSL) Denmark Erling Andersen
Elpen Farm Typology, Environmental
Indicators Federal Agricultural Research Centre
(FAL) Germany Peter Hinrich Farm Cost
Model, Economic Farm Behaviour Model,
Economic Indic. Agricultural University of
Athens (AUA) Greece Leonidas Louloudis
Socio-cultural indicators and farmers'
behaviour Object Vision Netherlands
Maarten Hilferink DMS and Elpen system software
65
(No Transcript)