The CLM5 Nitrogen cycle
Rosie Fisher, Dave Lawrence, Will Wieder, Gordon Bonan, Keith Oleson, Peter Lawrence, Sean Swenson,
Danica Lombardozzi, Ahmed Tawfik, Justin Perket, Erik Kluzek, Ben Andre, Bill Sacks, Mariana Vertenstein
Charlie Koven, Bill Riley, Bardan Ghmire (LBNL) Anthony Walker (ORNL), Chonggang Xu, Ashehad Ali (LANL)
Mingjie Shi & Josh Fisher (NASA-JPL) Eddie Brzostek (WVU), Quinn Thomas (VT),
NCAR-CLM4(CN)
Impact of Nitrogen on simulated carbon
fertilization
Issues raised with the CLM4.0(CN)
“Plants get Nitrogen for free”
The FUN* Model A marketplace for Nitrogen Uptake*Fixation and Uptake of Nitrogen
Hypothesis: Plants will take up N from the
cheapest sources
Solution to FUN model
Cgrowth = Cnpp - Cnuptake
Ngrowth = Nuptake
Nuptake = Cnuptake / CNuptake_cost Ngrowth = Cgrowth / CNtarget
Solve for maximum growth
GPP - Mresp
CNtarget/CNuptake_cost +1Cnuptake =
The FUN* Model A marketplace for Nitrogen Uptake*Fixation and Uptake of Nitrogen
FUNMod.F90
“Leaf Nitrogen content is static”
The FlexCN Model Variable carbon:nitrogen ratios
Increase in productivity due to change C:N ratio
Hypothesis: Plants will vary their tissue Carbon:Nitrogen ratio as N availability varies in
space and time
Zaehle et al. 2014
‘FlexCN’ allows for tissue-level variation in C:N ratio relative to target parameter.
Standalone FlexCN model tested by Ghmire et al. (BGC)
Increase in productivity due to increased NUE (fertilization)Increase in productivity due to increased leaf allocation
N Uptake
“N pool”
C Allocation to -leaves
-fine roots -stems
-coarse roots
Fraction of Allocation to -leaves -fine roots -stems -coarse roots
Actual N Allocation to
-leaves -fine roots
-stems -coarse roots
x
If N uptake is too low, C:N ratios will increase
GPP - MResp - GRresp - cost of N uptake
Fixation
Active Uptake
RetranslocationFlexCN N allocation scheme
NutrientCompetitionFlexibleCNMod.F90
“Photosynthetic capacity does not respond to the environment”
The LUNA* Model How best to use the Nitrogen you have?
Predicted optimal photosynthetic capacity
Hypothesis: Leaf Nitrogen is distributed so that light capture, carboxylation and respiration are co-limiting
*Leaf Use of Nitrogen for Assimilation
LUNA performance vs. observations
0 50 100 150
050
150
Predicted Vc,max25
Observed Vc,max25 (a) r2=0.59
ME=0.46
0 20 40 60 80 120
050
100
Predicted V c,max25
Observed Vc,max25 (b) r2=0.63
ME=0.52
0 20 40 60 80 100
050
150
Predicted V c,max25
Observed Vc,max25 (c) r2=0.46
ME=0.46
0 100 200 300 400
0200
400
Predicted J max25
Observed Jmax25 (d) r2=0.5
ME=0.33
0 50 100 200
0100
300
Predicted Jmax25
Observed Jmax25 (e) r2=0.85
ME=0.75
0 50 100 150 200
0100
200
300
Predicted J max 25Observed Jmax25 (f ) r2=0.45
ME=0.45
Herbs Shrubs Trees
Ali et al. 2015
LunaMod.F90
FUN flex-CN reconciliation
FUN-FlexCN coupling• The FUN model targets a fixed C/N ratio
• This intrinsically does not allow flexible CN ratio.
• We thus need to change Cnuptake to allow for this
Cgrowth = Cnpp - Cnuptake
Ngrowth = Nuptake
Nuptake = Cnuptake / CNuptake_cost Ngrowth = Cgrowth / CNtarget
Solve for maximum growth
GPP - Mresp
CNtarget/CNuptake_cost +1Cnuptake =
C allocation to uptake responds to CNuptake-cost and CNactual
Cnuptake = Cadj x (GPP-MR)
(CNtarget / CNuptake-cost) + 1.0
Cadj = 1.0 - (CNuptake-cost-Pa) / Pb
Cadj = Cadj + (1.0 - Cadj) x (CNactual- CNtarget)/ Pc
FUN equationAdjustment factor
Reduce C allocation with cost
Increase C allocation with high C:N
Offline FlexCN-FUN feedback behavior
Years of simulation Years of simulation
Pa = 5 : Pb= 200 : Pc= 80
Ncost=100 C/N
Ncost=10 C/N
Ncost=100 C/N
Ncost=10 C/N
Ncost=100 C/N
Ncost=10 C/N
Ncost=100 C/N
Ncost=10 C/N
Retranslocation
Cost of retrans-location gC/gN
C:N ratio of remaining leaf tissue (decreasing N content)
Cost of below ground extraction (fixation or active)
‘Free’ retranslocation
threshold ‘Paid-for’ retranslocation
limit
Iteration for each litter fall timestep
Remaining N goes to litter
Schematic of Retranslocation Algorithm
N limitation in CLM5Nitrogen is not abundant, for some reason:
slow decomposition? high leaching or denitrification? low productivity & fixation rates? lower deposition?
N uptake becomes more expensive
A higher fraction of NPP is spent on uptake. Tissue C:N ratios increase
N available for photosynthesis declinesNPP for growth decreases
Metrics of N limitationA higher fraction of NPP is spent on uptake. Tissue C:N ratios increase
N available for photosynthesis declines
NPP for growth decreases
Things I need Plot of CN ratios Plot of N spent
NutrientCompetitionFlexibleCNMod.F90
LeafC:N ratio
Fraction of NPP spent on N uptake
New parameters of N model• Nitrogen cost factors
• Fixation
• Active uptake
• Retranslocation
• Target leafCN ratio
• Flexible leafCN parameters
• LUNA parameters (only one is tunable)
Sensitivity Analysis of N cycle parameters
ITERATION (1-6 = 0.2 0.5 1.0 1.5 2.0)
Sensitivity Analysis of non-N parameters
ITERATION (1-6 = 0.2 0.5 1.0 1.5 2.0)
Conclusions• The new CLM5 nitrogen cycle model is substantially different to
the CLM4.5 and CLM4.0.
• We are making progress on understanding the behavior and interactions in the new model
• Much remains to be tested and understood (see parameterization talk)
• The model allows comparisons with many new data streams (N fixation, CN ratio, Vcmax variation)
• …and also fixes numerous theoretical problems with the existing CLM N cycle model