Modelling C and N dynamics with
MAGIC model
from annual to seasonal/monthly
time step
Filip Oulehle, Jack Cosby, Chris Evans,
Jakub Hruška, Jiří Kopáček, Filip
Moldan, Dick Wright
Introduction
• Effective immobilization of deposited nitrogen
is a common feature of most acid sensitive
catchments
• Saturation hypothesis suggests that N
immobilization should decline
• Nitrate leaching may become important
acidifying component
• Prediction of future N immobilization is
probably the biggest uncertainty in
acidification/eutrophication modelling
Modelling nitrogen with MAGIC
MAGIC (Model for Acidification of Groundwater In Catchments)
-Developed to predict the long-term effects of acidic deposition on surface water
chemistry
-Model simulates soil and surface water chemistry in response to changes in drivers such
as deposition of S and N, land use practices, climate…
- As sulphate concentrations have decreased, in response to the decreased S deposition,
nitrate (NO3) has become increasingly important. In acid soils much of the NO3 leached
from soil is accompanied by the acid cations H+ and inorganic aluminium (Ali)
-In the early versions of MAGIC (version 1-5) retention of N was calculated empirically as
a fraction of N deposited from input-output budgets
-Later on fraction N retained was described as a function of the N richness of the
ecosystem (soil C/N ratio in this case) – version 7
Soil C/N and N leaching - empirical evidence
Soil C/N seems to be a good predictor
of N leaching on a spatial scale
Lovett et al., Ecosystems (2002) 5: 712-718
Oulehle et al. Ecosystems (2008) 11: 410–425
Some limitations
- Soil C/N is vegetation specific
- C/N ratio does not necessarily reflect the N-richness of the actively cycling
component of the organic matter
- C/N ratio does not appear to be useful in understanding relatively short-term
changes in N dynamics
- „hard“ to detect changes in soil C/N under field conditions
Modelling nitrogen with MAGIC
% of N retention
100
80
In version 7 of MAGIC model – C/N soil ratio is the
fundamental control on N leaching
60
40
Two
shortcomings:
20
1) Over the short-term large changes in N leaching cannot be accounted for
0
0
10 in the
20 C/N
30 ratio
40 since
50 the C/N ration of soil organic matter changes only
by changes
soil C/N
slowly.
Limitation2)ofThe
thisC/N
approach
example
from
Čertovo
ratio of– bulk
organic
matter
is inLake
reality a consequence rather
than the driver of the long-term retention and loss of N from the soil pool.
MAGIC v5:
Čertovo
Čertovo
Čertovo
LakeLake
-Lake
Bohemian
- -Bohemian
Bohemian
Forest
Forest
Forest
(Czech
(Czech
(Czech
Republic)
Republic)
Republic)
N retention modelled as a first140
140
observed data
observed
data
120
120
order function of N deposition.
NO
NO33- (µeq/l)
(µeq/l)
100
100
80
80
MAGIC v5
MAGIC v7
MAGIC v7:
N retention modelled as a
function of N richness of the
ecosystem
60
60
40
40
20
20
00
1940
1940
1960
1960
1980
1980
2000
2000
2020
2020
Oulehle et al. Environmental Pollution (2012) 165: 158–166
Modelling nitrogen with MAGIC
Alternative formulation of N retention in new version of MAGIC (MAGIC v7ext) is based
directly on the microbial processes which determine the balance of N mineralization and
immobilization.
Conceptually developed by Jack Cosby
- Inorganic N enters the model as
deposition (wet and dry)
- Time series of plant litter and N fixation
(litter C and N) are external inputs to
SOM. At each time step, decomposers
process some of the C and N content of
SOM (FC1 and FN1). A portion of this C
and N turnover returns to the SOM as
decomposer biomass (FC2 and FN2), while
the remainder is lost from SOM as CO2
and NH4 (FC3 and FN3) or as DOC and
DON (FC4 and FN4).
Modelling nitrogen with MAGIC
Čertovo Lake - Bohemian Forest (Czech Republic)
120
observed data
100
MAGIC v7ext
80
60
40
20
1960
1980
2000
30
100
25
80
20
60
15
40
10
5
20
0
0
1940
1960
1980
2000
2020
The simulation can be further
improved by including the negative
effect of acidification on turnover of
SOM during the period of peak S
deposition – rationale well explained
in
Kopáček
et
al.
(2013)
Biogeochemistry 115: 33-51
2020
35
140
C/N pool- (mol/mol)
NO3 (µeq/l)
Carbon fraction processed (%)
0
1940
Constant carbon turnover (FC1)
Ali conc. (umol+/L)
NO3- (µeq/l)
140
120
30
100
25 80
Čertovo Lake - Bohemian Forest (Czech Republic)
observed data
MAGIC v7ext - C
turnover adjusted
60
20 40
20
15 0
1850
1940
1900
1960
1950
1980
2000
2000
2050
2020
Oulehle et al. Environmental Pollution (2012) 165: 158–166
Modelling nitrogen with MAGIC
Summary:
• The new formulation of C and N processes in the soil gives a more satisfactorily
simulation of the observed trends in NO3 in water compared to previous versions of
the MAGIC model.
• The new formulation simulates both rapid (and amplified) ecosystem responses to
changes in deposition of N, as well as the long-term changes in soil C/N resulting from
chronic N deposition and accumulation in SOM.
Limitations:
• Balanced C cycle, i.e. a constant soil C pool
• DOC and DON adjusted to fit the measured data
Modelling nitrogen with MAGIC
• Preliminary testing of MAGIC performance in monthly time step
• Soil organic matter decomposition and N uptake driven by changes in soil
temperature – Q10 fce (calculated externaly)
• potential application in climate change scenario assessment
Čertovo lake
Soil temp
Air temp
6000
20
C mmol m-2 month-1
15
°C
10
5
0
-5
5000
140
C decomp (Q10=3)
120
N uptake (Q10=4)
100
4000
80
3000
60
2000
40
1000
20
-10
-15
2003
0
1997
2004
2005
2006
0
2000
2003
2006
2008
2011
N mmol m-2 month-1
25
Modelling nitrogen with MAGIC
Seasonal MAGIC applied on four sites:
Čertovo lake (CZ) – seasonal data available 1998-2010
Gwy (Cymru) – 1980-2010
Storgama (Norway) – 1975-2010
Gårdsjön NITREX (Sweden) – 1990-2010
DIN (meq m-2 year-1)
500
400
300
Čertovo Lake
Gwy
Storgama
Gårdsjön
Gårdsjön fertilizer N input
200
100
0
1950
1970
1990
DIN annual deposition
2010
Modelling nitrogen with MAGIC
150
40
Čertovo Lake inlet
Čertovo Lake output
Gwy
Storgama
Gårdsjön
Soil C/N (mol mol-1)
N-NO3 (meq m-2 year-1)
200
100
50
0
1950
30
20
Čertovo Lake inlet
Gwy
Storgama
Gårdsjön
10
1970
1990
2010
The 4 sites are at various stages in N saturation
1950
1970
1990
2010
Modelling nitrogen with MAGIC
Čertovo Lake inlet
Inputs
N depositon*
Outputs
N-NO3 leaching
Gwy
Storgama
Gårdsjön
mmol m-2 % of input mmol m-2 % of input mmol m-2 % of input mmol m-2 % of input
139
122
69
338
96
N-NO3 leaching observed
100
DON leaching
Denitrification
Soil accumulation
27
8
11
69
37
30
25
19
6
8
24
7
49
9
13
8
20
6
40
17
7
35
* In respect of Gårdsjön = deposition + fertilizer input
N saturation
Čertovo Lake >> Gwy > Storgama > Gårdsjön
22
6
19
25
10
51
22
7
312
7
2
92
Modelling nitrogen with MAGIC
N-NO3 (mmol m-3month-1)
150
100
50
0
1997
150
N-NO3 (mmol m-3month-1)
100
Čertovo Lake inlet
1999
2001
2003
2005
2007
2009
50
0
1975
50
350
Storgama
100
1985
1995
2005
Gwy
0
1980
N-NO3 (mmol m-3month-1)
N-NO3 (mmol m-3month-1)
200
1985
1990
1995
2000
2005
2010
Gårdsjön
300
250
200
150
100
50
0
1990
1995
2000
2005
2010
Modelling nitrogen with MAGIC
N-NO3 (mmol m-2 month-1)
30
15
0
1997
15
N-NO3 (mmol m-2 month-1)
30
Čertovo Lake inlet
1999
2001
2003
2005
2007
2009
5
0
1975
15
45
Storgama
10
Gwy
0
1980
N-NO3 (mmol m-2 month-1)
N-NO3 (mmol m-2 month-1)
45
1985
1990
1995
2000
2005
2010
Gårdsjön
30
15
0
1985
1995
2005
1990
1995
2000
2005
2010
Modelling nitrogen with MAGIC
20
Gwy
10
y = 1.06x - 0.82
R² = 0.88
0
0
15
N-NO3 MODELLED (mmol m -2)
N-NO3 MODELLED (mmol m-2)
20
Čertovo Lake inlet
y = 1.10x + 0.74
R² = 0.54
0
20
40
N-NO3 OBSERVED (mmol m-2)
0
15
Storgama
N-NO3 MODELLED (mmol m-2)
N-NO3 MODELLED (mmol m-2)
40
10
20
N-NO3 OBSERVED (mmol m -2)
Gårdsjön
10
10
5
y = 0.91x + 0.15
R² = 0.43
0
0
5
10
15
N-NO3 OBSERVED (mmol m-2)
y = 1.05x + 0.24
R² = 0.30
5
0
0
5
10
15
-2
N-NO3 OBSERVED (mmol m )
Modelling nitrogen with MAGIC
120
Čertovo Lake inlet (1998-2009)
NO3 (mmol m-3)
80
60
40
MODELLED
Gwy (1980-2010)
OBSERVED
100
NO3 (mmol m-3)
40
MODELLED
OBSERVED
30
20
10
20
0
0
1
3
4
5
6
7
Month
8
9
10
11
1
12
100
MODELLED
Storgama (1975-2010)
2
3
4
5
6
7
Month
8
9
10
20
10
11
12
MODELLED
Gårdsjön (1990-2009)
OBSERVED
OBSERVED
NO3 (mmol m-3)
NO3 (mmol m-3)
30
2
80
60
40
20
0
0
1
2
3
4
5
6
7
Month
8
9
10
11
12
1
2
3
4
5
6
7
Month
8
9
10
11
12
Summary
• Despite reasonable model fit of cumulative N leaching across sites, only
Čertovo calibration has shown satisfactory fit between modelled and
observed NO3 dynamic.
•
•
•
This might be a result of uniform Q10 fce used across sites – need to try site specific Q10 for
decomposition and N uptake
Generally overestimation of N leaching during the winter months in Gwy - lack of proper
winter? Decomposition and N uptake more tightly coupled at this heathland site?
Presented examples pointed out that in Čertovo and Storgama catchments N dynamics behave
quite similar, despite different level of N saturation (in other words Strogama might be fairly
sensitive to N deposition).
• In respect of Gårdsjön, lack of seasonality in NO3 leaching might be a
consequence of artificial fertilizing.
•
Are N addition experiments able to mimic altered N soil transformations caused by gradual enrichment
through N deposition?
• Current MAGIC version is able to reproduce N seasonality as a result of
coupled C and N dynamic
•
Further development should focus on:
soil C dynamic (C sequestration – more C soil pools?)
feedback between soil acidity and C decomposition
Linkages between DOC availability and soil heterotrophic respiration
Three mechanisms could lead to lower amount of bioavailable dissolved organic C (DOC) for
the microbial community (Kopáček et al., 2013)
(1) Increased abundance of N for plant uptake, causing lower C allocation to plant roots
(2) Chemical suppression of DOC solubility by soil acidification
(3) Enhanced mineralisation of DOC due to increased abundance of electron acceptors in the
form of sulphate and nitrate - in anoxic soil micro-sites.
CO2 measurements
Treatment addition
Week 1
Treatment addition
Treatment addition
Week 2
Week 3
Leachate analysis
Leachate analysis
Week 4
Leachate analysis
Soil analysis
Soil analysis
800 ueq L-1
Control
H2SO4 HCl
NaOH
NaCl
Linkages between DOC availability and soil heterotrophic respiration
Linkages between DOC availability and soil heterotrophic respiration
200
soilwater DOC
180
160
1st treat
2nd treat
3rd treat
140
mg/L
120
100
80
60
40
20
0
Ctrl
H2SO4
HCl
NaOH
NaCl
Solution applications had immediate effect on DOC concentration in soil water.
Linkages between DOC availability and soil heterotrophic respiration
Standardized soil respiration (CO2 flux)
1.3
1.2
NaOH
1.1
1.0
NaCl
0.9
H2SO4
0.8
1st treatment
0.7
1
49
97
145
2nd treatment
193
Hour
241
3rd treatment
289
337
Solution applications had immediate effect on DOC concentration in soil water and on soil
respiration. In the end of the experiment, alkaline solution enhanced soil respiration by 20%
compared to control, whereas acid treatment suppressed soil respiration by 15% compared
to control. Neutral treatment has only short-term effect (suppression) on soil respiration.