Miljø- og Fødevareudvalget 2017-18
MOF Alm.del Bilag 172
Offentligt
1839586_0001.png
December 21, 2017
Summary of the analysis regarding prerequisites for the Food and Agriculture Package
The Danish analysis (Petersen, 2017) is entitled “Analyse af forudsætninger for landbrugspakken”.
The main theme of this analysis is the large uncertainty regarding the most appropriate value of
the marginal leaching, and the lacking communication of this large uncertainty. The main focus of
the analysis is the claims and calculations in Børgesen et al. (2015) by Aarhus University.
Regarding the scientific calculations underlying the Danish Agricultural Package (2015), Aarhus
University relied entirely on the NLES4 leaching model (Kristensen et al., 2008). This model
calculates an average marginal leaching of 18 %, meaning that 18 % of additional nitrogen fertiliser
is presumed to be leached from the root zone, as a national average for Denmark. During the
period 2003 – 2011 the average marginal leaching was assumed to be approx. 30 - 33 %, which is
in accordance with the marginal leaching of the previously utilised NLES3 model, which is
described by Kristensen et al. (2003).
This change from 30 – 33 % to 18 % caused a substantial easing in the additional measures needed
to compensate for the extra added nitrogen fertiliser.
The aim of the analysis was to investigate the scientific evidence that supported this considerable
change in presumed marginal leaching. Aarhus University claimed that the NLES4 model was
superior to the previous NLES3 model mainly due to:
1. NLES4 is based on a larger number of observations (N = 1467) than NLES3 (N = 1299).
2. The NLES4-calculated marginal leaching is in better accordance with national and international
studies.
Regarding claim no. 1, the analysis points out that an increased number of observations is not as
such a quality criterion in terms of realistic calculations of marginal leaching.
Regarding claim no. 2, the acclaimed better accordance with national and international studies is
based on field experiments from 5 scientific articles.
Regarding these 5 articles the analysis points out:
Three of these articles (Engström et al., 2010; Delin and Stenberg, 2014; Manevski et al., 2015)
treated one-year field trials with varying nitrogen applications. Aarhus University compared
the marginal leaching from these three articles directly with the average long-term marginal
leaching. As the long-term marginal leaching invariably will be higher than the one-year
leaching, this is an obvious statistical error, leading to a bias.
When making a more detailed numerical analysis of the marginal leaching from the trials, and
comparing values with the NLES4 one-year average, all 10 field trials had considerable larger
marginal leaching than NLES4.
 MOF, Alm.del - 2017-18 - Bilag 172: Orientering om brev til EU-kommissionen fra Danmarks Naturfredningsforening vedr. analyse af forudsætning for landbrugspakken
1839586_0002.png
The forth article (Wachendorf et al., 2004) assembles a number of field trials, and generates a
formula for the leaching. From this formula the marginal leaching can be calculated, and the
NLES4 long-term marginal leaching is compared with this calculation. But this approach
generates a bias, as some of the trials have a shorter time-span than the 6 years it takes for
NLES4 to reach the long-term marginal leaching level. Further, correcting for this bias is not
possible without a comprehensive effort, as not all trial durations are directly accessible. Due
to this bias which further cannot be corrected directly, this study cannot be used for proper,
unbiased comparisons.
The fifths and last article (Pandey et al., 2015) compares 3 different cropping systems, one
conventional and two organic systems. The one organic system has a separate crop rotation,
whereas the other crop rotation involved is used both for the conventional and the other
organic system. Aarhus University claims, that a simple two-dimensional linear regression
(leaching versus total nitrogen input) for the three rather different cropping systems can
represent some sort of common marginal leaching. This claim is incorrect, as the analysis
explains in detail. There are many other factors (dimensions) involved, including the
interaction between applied fertiliser and nitrogen fixation, the different dynamics of animal
manure and mineral fertiliser, different crop rotations and different nitrogen soil pool
developments. The two dimensions considered will not be able to span all the underlying
factors/dimensions, and hence cannot constitute a true and accurate system comparison. Due
to the profound methodologic problems of this oversimplification, this study cannot be used
for scientifically correct comparisons with marginal leaching.
Claim no. 2 regards NLES4 being in better accordance with national and international studies. The
analysis points out that the marginal leaching calculated by NLES4 is not supported by the 5
articles. The calculations that could be carried out without bias or other methodological problems
showed a much higher marginal leaching than the modelled national average (see table 1,
Appendix). This does indicate problems with the calculated average of 18 %, but the data set is
comparatively small and further not representative for Denmark regarding crop composition.
Therefore, the analysis does not claim that the assumed 18 % with certainty can be considered too
low a value. What the analysis does claim, is that this value is very uncertain, and based on the
current knowledge, the previous range of approx. 30 – 33 % is just as realistic as the present 18 %.
./. The Danish Ministry of Environment and Food referred to the acclaimed better accordance with
national and international studies as justification for changing the calculation of the leaching in the
answer no. 568 (MOF alm. del) given 31
st
March 2016 to Member of the Danish Parliament Ida
Auken (attached).
The curve function of the leaching in NLES4 is quite unrealistic. Even when going from approx. 16
% below economic optimum for nitrogen fertilisation to 50 % above economic optimum an almost
linear response is obtained, according to the figure below. It is common agronomical knowledge
(e.g. Lord and Mitchell, 1998; Goulding et al., 2000; Delin and Stenberg, 2014) that a distinct rise in
the leaching level will typically begin near economical optimum. The lack of realism regarding the
response is also evident from figure 1 – 6 in the analysis (see Appendix).
 MOF, Alm.del - 2017-18 - Bilag 172: Orientering om brev til EU-kommissionen fra Danmarks Naturfredningsforening vedr. analyse af forudsætning for landbrugspakken
1839586_0003.png
A model that is unable to mimic this well-known effect appropriately, has hence been used for
assisting in comprehensive decision making regarding environmental effects.
A
90
80
Leaching (kg N/ha)
70
60
50
40
30
20
10
0
0
20
40
60
80
100
Economic
optimum
Extra nitrogen fertiliser (kg N/ha)
Figure.
NLES4-calculated average for the Danish average nitrogen leaching. The starting point is
nitrogen fertiliser application levels as in 2011. The vertical dotted line shows the economical
optimum. The present figure is a part of figure 7 in the analysis.
The last part of the analysis is largely independent of the above calculations. This part considers
the large and unexplained difference in marginal leaching between the two model versions NLES3
and NLES4. It is explained how this implies significant methodological problems with the
calculations of the marginal leaching. The following key questions were never addressed during
the preparation of the Food and Agriculture Package:
To what extent is the difference in marginal leaching caused by differences in model
structure and to what extent is it caused by differences in underlying data?
What specific conditions in NLES3 and NLES4 (structure and data basis) constitute the
reason for the difference in marginal leaching?
Which model has the most accurate calculation of marginal leaching and what criteria can
be used to determine this?
The analysis hence points out questions that need to be answered. The above three questions
need to be clarified before it is possible to undertake any kind of fact-based discussion of whether
adjustments should be made to the marginal leaching relative to the previous assumption of a
value of 30 - 33%.
 MOF, Alm.del - 2017-18 - Bilag 172: Orientering om brev til EU-kommissionen fra Danmarks Naturfredningsforening vedr. analyse af forudsætning for landbrugspakken
Conclusions of the analysis
The development of the NLES4 marginal leaching, relative to increasing nitrogen application,
differs from the development calculated from experimental data (Figures 1B - 6B in the analysis,
see also Appendix). In addition, the general level of marginal leaching calculated with NLES4 is
significantly lower (table 1 in the analysis, see also Appendix) than the marginal leaching that can
be calculated from the experiments.
Based on data selected by Aarhus University to support the general level of the nationwide NLES4-
calculations with experimental results, it is concluded that the model used for calculation of
marginal leaching underlying The Danish Agricultural Package, cannot be used for real-life
estimations of nitrogen emissions resulting from the easing of the Danish fertilization rules.
Furthermore, the choice of using NLES4 rather than NLES3 is arbitrary, as this choice was not
based on systematic quality criteria.
The analysis thus demonstrates that there is no scientific support for the Food and Agriculture
Package's lowering of the estimated marginal leaching at national level from 30 – 33 % to 18 %.
 MOF, Alm.del - 2017-18 - Bilag 172: Orientering om brev til EU-kommissionen fra Danmarks Naturfredningsforening vedr. analyse af forudsætning for landbrugspakken
1839586_0005.png
Appendix
This section addresses table 1 and figure 1 – 7 from the analysis.
Table 1.
Results for calculation of the marginal leaching, based on data from Delin and Stenberg
(2014), Engström et al. (2010), and Manevski et al. (2015). From Petersen (2017).
Article
Delin, Stenberg
Engström et al.
Manevski et al.
Manevski et al.
Manevski et al.
Manevski et al.
Manevski et al.
Manevski et al.
Crops
Winter oilseed rape
Oat
Maize after maize
Maize after maize
Maize after maize, cover crops
Location
Götala, SV
Götala, SV
Foulum, DK
Jyndevad, DK
Foulum, DK
Jyndevad, DK
Foulum, DK
Jyndevad, DK
Foulum, DK
Jyndevad, DK
Marginal leaching at Factor relative to
N-norm (%)
NLES4-average (%)
14.0
175
23.8
298
98.9*
1236*
28.5
356
22.1
276
20.2
42.6
81.2
43.6
64.2
253
533
1015
545
803
Maize after maize, cover crops
Maize after clovergrass
Maize after clovergrass
Maize after cl.grass, cover
Manevski et al.
crops
Maize after cl.grass, cover
Manevski et al.
crops
*This value can be considered an extreme outlier.
These one-year trial results were compared with the average one-year model-derived result. All
marginal leaching values were much higher than the average one-year model-derived marginal
leaching at national level, which is 8 %. Ideally, all results should be compared to a specific NLES4
mimic of each experiment. Such specific comparisons are not attempted by Aarhus University and
not attempted in the analysis.
Aarhus University selected five articles which were used for comparison with the NLES4 model
results at a national level. The crop and cropping system composition in these 5 articles is far from
representative for Danish agriculture. Conventional grown cereals are underrepresented, and
maize and organic farming is overrepresented, relative to the present agricultural practice.
Two of these articles (Wachendorf et al. 2004; Pandey et al. 2015) were unfit for scientifically
stringent comparisons with model results, as elaborated in the analysis. Wachendorf et al. (2004)
regarded grass and clover grass, and Pandey et al. (2015) primarily regarded organic farming (6
organic trials and 2 conventional trials, carried out in 3 locations).
The below figures are shown with their original Danish text.
Figures 1-6, A, X-axis is applied nitrogen fertiliser (kg N/hectare).
Figures 1-6, A, Y-axis is applied nitrogen leaching (kg N/hectare).
Figures 1-6, B, X-axis is applied nitrogen fertiliser relative to the recommended amount.
Figures 1-6, B, Y-axis is applied marginal nitrogen leaching (%).
The yellow curve in figures 1-6, B, shows the national average marginal leaching at the given
relative nitrogen application level.
 MOF, Alm.del - 2017-18 - Bilag 172: Orientering om brev til EU-kommissionen fra Danmarks Naturfredningsforening vedr. analyse af forudsætning for landbrugspakken
1839586_0006.png
Figure 7A, X-axis is applied nitrogen fertiliser
above
the level in 2011 (kg N/hectare) as national
average.
Figure 7A, Y-axis is nitrogen leaching (kg N/hectare) as national average.
Figures 7B, X-axis is applied nitrogen fertiliser
above
the level in 2011 (kg N/hectare) as national
average.
Figures 7B, Y-axis is marginal nitrogen leaching (%). This is shown at the 6-year and above level
(black curve) and at the 1-year level (yellow curve).
The vertical, stipulated lines in Figure 7 show the economically optimal application.
A
90
80
70
2007
2008
2009
Gennemsnit
B
100
90
80
NLES4
Forsøg
Udvaskning (kg N/ha)
Marginaludvaskning (%)
60
50
40
30
20
10
0
0
20
40
60
80
100
120
140
70
60
50
40
30
20
10
0
0
0,2
0,4
0,6
0,8
1
1,2
1,4
Tilført mineralsk N (kg N/ha)
N-tilførsel i forhold til norm
Figur 1.
Udvaskning (A) og marginaludvaskning (B) fra havre tilført forskellige mængder
kvælstofgødning, på basis af data fra Delin og Stenberg (2014).
A
100
90
80
Mild vinter
Kold vinter
Gennemsnit
B
100
90
80
NLES4
Forsøg
Marginaludvaskning (%)
Udvaskning (kg N/ha)
70
60
50
40
30
20
10
0
0
20
40
60
80
100
120
140
160
180
200
70
60
50
40
30
20
10
0
0
0,2
0,4
0,6
0,8
1
1,2
1,4
Tilført mineralsk N (kg N/ha)
N-tilførsel i forhold til norm
 MOF, Alm.del - 2017-18 - Bilag 172: Orientering om brev til EU-kommissionen fra Danmarks Naturfredningsforening vedr. analyse af forudsætning for landbrugspakken
1839586_0007.png
Figur 2.
Udvaskning (A) og marginaludvaskning (B) fra vinterraps tilført forskellige mængder
kvælstofgødning, på basis af data fra Engström et al., 2010.
A
220
200
180
160
140
120
100
80
60
40
20
0
0
Foulum
Jyndevad
B
100
90
NLES4
Forsøg Foulum
Forsøg Jyndevad
Marginaludvaskning (%)
50
100
150
200
250
80
70
60
50
40
30
20
10
0
0
0,2
0,4
0,6
0,8
1
1,2
1,4
1,6
Udvaskning (kg N/ha)
Tilført mineralsk N (kg N/ha)
N-tilførsel i forhold til norm
Figur 3.
Udvaskning (A) og marginaludvaskning (B) fra majs der efterfølger majs, tilført forskellige
mængder kvælstofgødning, på basis af data fra Manevski et al. (2015).
A
80
70
Foulum
Jyndevad
B
100
90
80
NLES4
Forsøg Foulum
Forsøg Jyndevad
Marginaludvaskning (%)
Udvaskning (kg N/ha)
60
50
40
30
20
10
0
0
50
100
150
200
70
60
50
40
30
20
10
0
0
0,2
0,4
0,6
0,8
1
1,2
1,4
Tilført mineralsk N (kg N/ha)
N-tilførsel i forhold til norm
Figur 4.
Udvaskning (A) og marginaludvaskning (B) fra majs med efterafgrøder der efterfølger
majs, tilført forskellige mængder kvælstofgødning, på basis af data fra Manevski et al. (2015).
 MOF, Alm.del - 2017-18 - Bilag 172: Orientering om brev til EU-kommissionen fra Danmarks Naturfredningsforening vedr. analyse af forudsætning for landbrugspakken
1839586_0008.png
A
220
200
180
Foulum
Jyndevad
B
100
90
80
NLES4
Forsøg Foulum
Forsøg Jyndevad
Udvaskning (kg N/ha)
160
140
Marginaludvaskning (%)
70
60
50
40
30
20
10
0
120
100
80
60
40
20
0
0
20
40
60
80
100
120
140
0
0,2
0,4
0,6
0,8
1
1,2
1,4
1,6
1,8
Tilført mineralsk N (kg N/ha)
N-tilførsel i forhold til norm
Figur 5.
Udvaskning (A) og marginaludvaskning (B) fra majs der efterfølger kløvergræs, tilført
forskellige mængder kvælstofgødning, på basis af data fra Manevski et al. (2015).
A
180
160
140
Foulum
Jyndevad
B
100
90
80
NLES4
Forsøg Foulum
Forsøg Jyndevad
Marginaludvaskning (%)
0
20
40
60
80
100
120
140
Udvaskning (kg N/ha)
120
100
80
60
40
20
0
70
60
50
40
30
20
10
0
0
0,2
0,4
0,6
0,8
1
1,2
1,4
1,6
1,8
Tilført mineralsk N (kg N/ha)
N-tilførsel i forhold til norm
Figur 6.
Udvaskning (A) og marginaludvaskning (B) fra majs med efterafgrøder der efterfølger
kløvergræs, tilført forskellige mængder kvælstofgødning, på basis af data fra Manevski et al.
(2015).
 MOF, Alm.del - 2017-18 - Bilag 172: Orientering om brev til EU-kommissionen fra Danmarks Naturfredningsforening vedr. analyse af forudsætning for landbrugspakken
1839586_0009.png
A
90
80
70
60
50
40
30
B
25
20
Marginaludvaskning (%)
Udvaskning (kg N/ha)
15
6-årig
marginal-
udvaskning
10
20
10
0
0
20
Økonomisk
optimum
5
Økonomisk
optimum
1-årig
marginal-
udvaskning
0
40
60
80
100
0
20
40
60
80
100
Ekstra mineralsk N (kg N/ha)
Ekstra mineralsk N (kg N/ha)
Figur 7.
NLES4-beregnede landsgennemsnit for udvaskning (A) og marginaludvaskning (B).
Udgangspunktet er kvælstofgødskning som i 2011. De lodrette stiplede linjer viser økonomisk
optimum. Se afsnit 3 i Appendiks for detaljerne i beregningen af figurens kurver.
References
Børgesen, C.D., Thomsen, I.K., Hansen, E.M., Kristensen, I.T., Blicher-Mathiesen, G., Rolighed, J.,
Jensen, P.N., Olesen, J.E., Eriksen, J. 2015. Notat om tilbagerulning af tre generelle krav,
Normreduktion, Obligatoriske efterafgrøder og Forbud mod jordbearbejdning i efteråret. Aarhus
Universitet. Notat til NaturErhvervstyrelsen den 11. november 2015.
http://pure.au.dk/portal/files/95991713/Notat_om_tilbagerulning_af_tre_generelle_krav_Normr
eduktion_Obligatoriske_efterafgr_der_og_Forbud_mod_jordbearbejdning_i_efter_ret_111115.pd
f
Delin, S., Stenberg, M. 2014. Effect of nitrogen fertilization on nitrate leaching in relation to grain
yield response on loamy sand in Sweden. European Journal of Agronomy 52, 291-296.
Engström, L., Stenberg, M., Aronsson, H., Lindén, B. 2010. Reducing nitrate leaching after winter
oilseed rape and peas in mild and cold winters. Agronomy for Sustainable Development 31, 337–
347.
Goulding, K.W.T., Poulton, P.R., Webster, C.P., Howe, M.T. 2000. Nitrate leaching from the
Broadbalk Wheat Experiment, Rothamsted, UK, as influenced by fertilizer and manure inputs and
the weather. Soil Use and Management 16, 244-250.
Kristensen, K., Jørgensen, U. og Grant, R. (2003) Genberegning af modellen N -LES. 12 pp.
Baggrundsnotat til Grant, R. og Waagepetersen, J. (2003)
http://www2.dmu.dk/1_viden/2_publikationer/3_ovrige/rapporter/vmpii/genberegning_af_mode
llen_nles.pdf
 MOF, Alm.del - 2017-18 - Bilag 172: Orientering om brev til EU-kommissionen fra Danmarks Naturfredningsforening vedr. analyse af forudsætning for landbrugspakken
1839586_0010.png
Kristensen, K., Waagepetersen, J., Børgesen, C.D., Vinther, F.P., Grant, R., Blicher-Mathiesen,
G. 2008. Reestimation and further development in the model N-LES - N-LES3 to N-LES4. Det
Jordbrugsvidenskabelige Fakultet, Aarhus Universitet og Danmarks Miljøundersøgelser, Aarhus
Universitet. DJF rapport 139.
http://pure.au.dk/portal/files/2423945/DJF_rapport_Markbrug_139
Lord, E.I., Mitchell, R.D.J. 1998. Effect of nitrogen inputs to cereals on nitrate leaching from sandy
soils. Soil Use and Management, 14, 78–83.
Manevski, K., Børgesen, C.D., Andersen, M.N., Kristensen, I.S. 2015. Reduced nitrogen leaching by
intercropping maize red fescue on sandy soils in Northern Europe: a combined field and modeling
study. Plant and Soil 388, 67-85.
Pandey, A., Li, F., Askegaard, M., Rasmussen, I.A., Olesen, J.E. 2015. Nitrogen balances in organic
and conventional arable crop rotations and their relations to nitrogen yield and leaching losses.
Agriculture, Ecosystems and Environment. Publication status unknown.
Petersen, B.M. 2017. Analyse af forudsætninger for landbrugspakken. The Danish Society for
Nature Conservation, November 2017.
http://www.dn.dk/media/24635/analyse-af-forudsaetninger-for-landbrugspakken.pdf
Wachendorf, M., Büchter, M., Trott, H., Taube, F. 2004. Performance and environmental effects of
forage production on sandy soils. II. Impact of defoliation system and nitrogen input on nitrate
leaching losses. Grass and Forage Science, 59, 56-68.