1
Auxiliary material for
2
3
Quantifying microbial ecophysiological effects on the carbon
fluxes of forest ecosystems over the conterminous United States
4
5
Guangcun Hao1,2, Qianlai Zhuang2,3*, Qing Zhu2,4 , Yujie He2, Zhenong Jin2 and Weijun
Shen1,
6
7
8
1
9
10
2
11
3
Department of Agronomy, Purdue University, West Lafayette, Indiana, USA
12
4
Earth Science Division, Lawrence Berkeley National Lab, Berkeley, CA, USA
Key Laboratory of Vegetation Restoration and Management of Degraded Ecosystems,
South China Botanical Garden, Chinese Academy of Sciences, Guangzhou 510650,
China
Department of Earth, Atmospheric, and Planetary Sciences, Purdue University, West
Lafayette, Indiana, USA
13
14
*Correspondence to:
15
Qianlai Zhuang
16
qzhuang@purdue.edu
17
18
19
20
21
22
23
24
25
26
27
28
1
29
Appendix.1
30
The Terrestrial Ecosystem Model
31
In TEM, carbon exchanges start from the simulation of gross primary production (GPP).
32
GPP is modeled as a function of multi-scalars that represent a variety of environmental
33
and physical constraints. The formula for calculating daily GPP is:
34
GPP  C max f ( PAR) f ( p) f ( FOLIAGE ) f (T )  f (CA, Gv ) f ( NA) f ( FT )
（1）
35
Where Cmax is the maximum rate of C assimilation by the entire plant canopy under
36
optimal environmental conditions, and f(PAR), f(p), f(FOLIAGE), f(NA) and f(FT)
37
represent the limits of the photosynthetically active radiation (PAR), the leaf phenology,
38
the influence of relative canopy leaf biomass relative to maximum leaf biomass (Zhuang
39
et al. 2002), the limiting effects of plant nitrogen availability and the effects of freeze-
40
thaw dynamics on GPP (Zhuang et al. 2003), respectively. The term CA represents the
41
influence of increasing atmospheric CO2 concentration on GPP, which is modelled by
42
following Michaelis-Menten kinetics (Raich et al. 1991); Gv accounts for changes in leaf
43
conductivity to CO2 resulting from moisture availability, which is based on the estimates
44
of evapotranspiration (ET). Net Ecosystem Production (NEP) is defined as the difference
45
between Net Primary Production (NPP) and Heterotrophic Respiration (RH), and NPP is
46
defined as the difference between Gross Primary Production (GPP) and Autotrophic
47
Respiration (RA).
48
NPP=GPP – RA ;
（3）
49
NEP=NPP - RH ;
（4）
50
Appendix.2
51
Definition of key terminology
52
53
In the paper, soil microbial physiology refers to microbial growth activity and
depolymerization activities. Carbon flux mean GPP, NPP, NEP, RA and RH .
54
55
56
57
Modeling microbial physiology
The changes in microbial biomass are simulated by the subtraction of microbial
death and enzyme production and the CO2 (heterotrophic respiration) emitted through
2
58
microbial respiration from assimilated soluble C, via which O2 is consumed to produce
59
energy for assimilation of dissolved organic C:
dMIC
= ASSIM - CO2 - DEATH - EPROD
dt
60
61
62
(1)
Assimilation is a Michaelis-Menten function scaled to the pool size of microbial
biomass:
[ Sx ]
kM[ Sx ] +[Sx ]
63
ASSIM = V maxuptake  MIC 
64
where V maxuptake is the maximum velocity of the enzymatic reaction when substrate is
65
not limiting. kM [ Sx ] is the corresponding Michaelis constant. The concentration of soluble
66
C substrates at the reactive site of the enzyme ([Sx]) is affected by soil water content, and
67
specifically by diffusion of substrates through soil water films. [Sx] is calculated from
68
3
[Sxsoluble] through [S x ]=[Sxso lub le ]  Dliq  , where  is the volumetric water content of the
69
soil, and Dliq is a diffusion coefficient of the substrate in liquid phase. Diffusion of
70
soluble substrates have been shown to be related to the thickness of the soil water films,
71
which is approximated by the cube of the volumetric water content. It is assumed that the
72
cell surface area available for [Sx] uptake is proportional to the number of cells, and thus
73
the microbial biomass [Davidson et al. 2012]. [Sx] is assumed to be the only substrate for
74
microbial C uptake. Similar to Davidson et al. (2012), the value of Dliq is determined by
75
assuming the boundary condition that all soluble substrate is available at the reaction site
76
for saturated soil (i.e., [Sx ]=[Sxsoluble ]).
77
(2)
CO2 (heterotrophic respiration) is produced as the part of microbial assimilated C
78
not allocated to biomass growth. The production process follows Michaelis-Menten
79
kinetics similar to assimilation but is controlled by the concentration of both [Sx] and O2:
80
CO2  V maxCO2 
[Sx ]
[O2 ]

 MIC
kM[ sx ]  [Sx ] kMO2  [O2 ]
(3)
81
82
The concentration of O2 at the reactive site of the enzyme ([O2]) depends upon
83
diffusion for gases within the soil medium, which is modeled with a simple function of
84
air-filled porosity: [O2 ]=D gas  0.209  a
4/3
. Dgas is a diffusion coefficient for O2 in air,
3
85
0.209 is the volume fraction of O2 in air, and a is the air-filled porosity of the soil. The
86
total porosity is calculated from bulk density (BD) and particle density (PD):
87
a =1-
88
89
BD
-.
PD
V maxuptake , V max CO , and kM [ Sx ] are temperature dependent. Vmaxuptake and
2
V max CO2 follow the Arrhenius equation:
90
Ea uptake


V max uptake =V max uptake0  exp  
 R  (TC +273) 
(4)
91
EaCO2


V maxCO2 =V maxCO20  exp  
 R  (TC +273) 
(5)
92
where V max uptake0 and V max CO20 are the pre-exponential coefficient (i.e., the theoretical
93
decomposition enzymatic reaction rate at Ea = 0), R is the ideal gas constant (8.314 J K-1
94
mol-1), TC is the temperature in Celsius, and Ea uptake and Ea CO2 are the activation energy
95
for [Sx] uptake and CO2 respiration by microbial. High activation energy indicates high
96
temperature sensitivity but reacts slowly. kM [ sx ] is calculated as a linear function of
97
temperature, as adopted in Davidson et al. (2012):
kM [ S x ]  ckM[ S ]  mkM [ S ]  TC
98
x
99
where ckM[ S ] and mkM[ Sx ] are the intercept and slope parameters, respectively. kM O2 is
100
assumed to be constant with respect to temperature for the sake of model parsimony.
101
However, kM O2 could be modeled as a function of temperature when observations are
102
available.
103
104
x
Microbial death is modeled as a first-order process with rate constant rdeath
(Lawrence et al. 2009):
DEATH  rdeath  MIC
105
106
107
108
109
(6)
x
(7)
Enzyme production is modeled as a constant fraction ( rEnz Pr od ) of microbial
biomass (Lawrence et al. 2009):
EPROD  rEnz Pr od  MIC
(8)
The enzyme pool changes with enzyme production and turnover:
4
dEnz
 EPROD  ELOSS
dt
110
111
where the turnover (ELOSS) is modeled as a first-order process with constant rate:
ELOSS  rEnzLoss  Enz
112
113
(10)
The changes in SOC pool varies with external inputs from vegetation litterfall
114
carbon, enzyme turnover, inputs from dead microbial biomass ( MICtoSOC ) and
115
decomposition loss:
116
(9)
dSOC
 inputSOC  DEATH  MICtoSOC  ELOSS  DECAY
dt
(11)
117
where enzymatic decomposition of SOC (DECAY) here is mainly referring to the process
118
through which microbes secrete exoenzymes to convert macromolecules into soluble
119
products (soluble C, denoted as [Sxsoluble]) that can be absorbed and metabolized by
120
microbes. This process follows Michaelis-Menten kinetics with enzyme and substrate
121
(here SOC) constraint:
122
DECAY  V max SOC  Enz 
SOC
kM SOC  SOC
123
where VmaxSOC is the maximum velocity of the enzymatic reaction when substrate is not
124
limiting and is calculated according to Arrhenius function:
125
126
(12)


Ea SOC
V max SOC =V max SOC0  exp  
 R  (temp+273) 
(13)
We assume Michaelis-Menten constant for SOC ( kM SOC ) is invariable with
127
temperature. The soluble C pool ([Sxsoluble]) changes with external inputs, the remaining
128
fraction of dead microbial biomass, and decomposition:
129
dSo lub leC
 DEATH  (1 MICtoSOC)  DECAY  ASSIM
dt
130
This process represents the enzymatic depolymerization of complex molecules to
131
the simpler ones available for microbial uptake. More detailed algorithms are
132
documented in He et al. (2014).
(14)
133
5
134
Appendix.3
135
Data organization and model simulations
136
We organized the observed or estimated GPP, NEP, and meteorological data (e.g.
137
radiation, air temperature, and precipitation) from six representative eddy covariance flux
138
sites for each vegetation type to parameterize the MIC-TEM. Data from other four
139
AmeriFlux forest sites are used to verify the model performance (Table S2). Specifically,
140
we collected all available daily Level 4 Net Ecosystem Exchange (NEE) products
141
(http://public.ornl.gov/ameriflux/) of these sites (Table S2). For each site, if the
142
percentage of remaining missing values for NEE_st or GPP_st (standardized data) is
143
lower than that for NEE_or or GPP_or(original data), we select NEE_or or GPP_or;
144
otherwise, we use NEE_st or GPP_st. The measured NEE is then compared with modeled
145
NEP (Chen and Zhuang et al. 2011).
146
We spun up the model for 120 years to account for the influence of climate inter-
147
annual variability on the initial conditions of the ecosystems. Since historic climate data
148
are not available before 1948, we repeat the data from 1948 to 1987 for 3 times for the
149
spin-up with NECP global datasets at a 0.5 spatial resolution.
150
To quantify the effects of microbial activity on regional carbon dynamics, we
151
applied the both original and revised TEM to the forest ecosystems of the conterminous
152
United States for the period 2006 – 2100 with daily time-step and 0.5° × 0.5° spatial
153
resolution driving data (Kottek et al. 2006) (Fig. S1). The spatially-explicit soil texture,
154
elevation data and vegetation type data are from our previous studies (Melillo et al. 1993;
155
Zhuang et al. 2003). Future climate scenarios from 2006 to 2100 are generated under two
156
RCPs of the Coupled Model Inter-comparison Project phase 5 (CMIP5) with NOAA’s
157
Earth System Models (GFDL-ESM2G). A general description of CMIP5 models and the
6
158
experiment design can be found in Taylor et al. (2012). All climate data sets are
159
resampled to 0.5° × 0.5° spatial resolution with statistical downscaling method from
160
Mitchell et al. (2004).
161
162
163
164
165
166
167
168
169
170
7
171
Table S1 Parameters used in the model. Inversed estimates of specific parameters and parameter
172
ranges used are listed
Process
Assimilation
Decay
Parameter
Unit
Initial
Value
Description
Parameter
range
References
Ea_micup
J mol-1
47000
Soluble and diffused
Sx uptake by microbial
-
Allison et
al. (2010)
Vmax_uptake0
mg Sx cm-3
soil (mg
biomass cm-3
soil)-1 h-1
9.97e6
Maximum microbial
uptake rate
[1.0e4,
1.0e8]
-
c_uptake
mg Sx cm-3
soil
0.1
-
Allison et
al. (2010)
m_uptake
mg Sx cm-3
soil °C-1
0.01
-
Allison et
al. (2010)
Ea_Sx
J mol-1
48092
-
Knorr et al.
(2005)
c_Sx *
mg
assimilated Sx
cm-3 soil
0.1
-
Allison et
al. (2010)
m_Sx *
mg
assimilated Sx
cm-3 soil °C-1
0.01
-
Allison et
al. (2010)
41000
Activation energy of
decomposing SOC to
soluble C
-
Modified
from
Davidson et
al. (2012)
9.17e7
Maximum rate of
converting SOC to
soluble C
[1.0e5,
1.0e8]
-
-
Allison et
al. (2010)
-
Allison et
al. (2010)
-
Davidson et
al. (2012)
mol-1
Ea_SOC
J
Vmax_SOC0
mg
decomposed
SOC cm-3 soil
(mg Enz cm-3
soil)-1 h-1
c_SOC
mg SOC cm-3
soil
400
m_SOC
mg SOC cm-3
soil °C-1
5
kM_O2
cm3O2 cm-3
soil
0.121
Vmax_CO20
mg respired
Sx cm-3 soil h-
c_Sx *
mg
assimilated Sx
cm-3 soil
Temperature regulator
of MM for enzymatic
decay of SOC to
soluble C (kM_SOC)
Temperature regulator
of MM for enzymatic
decay of SOC to
soluble C (kM_SOC)
Michaelis-Menten
constant (MM) for O2
(at mean value of
volumetric soil
moisture)
1.9e7
Maximum microbial
respiration rate
[1.0e6,
1.0e8]
-
0.1
Temperature regulator
of MM for microbial
respiration of
-
Allison et
al. (2010)
1
CO2
production
Temperature regulator
of MM for Sx uptake
by microbes
(kM_uptake)
Temperature regulator
of MM for Sx uptake
by microbes
(kM_uptake)
Activation energy of
microbes assimilating
Sx to CO2
Temperature regulator
of MM for microbial
assimilation of Sx
(kM_Sx)
Temperature regulator
of MM for microbial
assimilation of Sx
(kM_Sx)
8
MIC
turnover
ENZ
turnover
173
m_Sx *
mg
assimilated Sx
cm-3 soil °C-1
0.01
MICtoSOC
%
50
r_death
% h-1
0.02
r_EnzProd
% h-1
5.0e-4
r_EnzLoss
% h-1
0.1
assimilated Sx
(kM_Sx)
Temperature regulator
of MM for microbial
respiration of
assimilated Sx
(kM_Sx)
Partition coefficient for
dead microbial biomass
between the SOC and
Soluble C pool
Microbial death
fraction
Enzyme production
fraction
Enzyme loss fraction
-
Allison et
al. (2010)
-
Allison et
al. (2010)
-
Allison et
al. (2010)
Allison et
al. (2010)
Allison et
al. (2010)
* c_Sx and m_Sx are used in both assimilation and CO2 production calculations
9
174
175
Table S2 Characteristics of AmeriFlux sites used in this study and statistical results for the observed and predicted daily NEP and
GPP at each site for parameterization. The unit of RMSE is g C m-2 day-1
R2
RMSE
References
MIC-TEM NEP
TEM NEP
MIC-TEM GPP
TEM GPP
0.70
0.60
0.94
0.85
1.39
2.45
2.27
5.39
Hollinger et al.
(1999, 2004)
2000-2006
MIC-TEM NEP
TEM NEP
MIC-TEM GPP
TEM GPP
0.72
0.70
0.90
0.87
2.99
4.56
3.80
4.88
Urbanski et al.
(2007)
Evergreen
Forest
2000-2005
MIC-TEM NEP
TEM NEP
MIC-TEM GPP
TEM GPP
0.35
0.10
0.87
0.75
0.76
2.34
2.85
1.59
Monson et al.
(2002)
-121.9519
Evergreen
Forest
2000-2002
MIC-TEM NEP
TEM NEP
MIC-TEM GPP
TEM GPP
0.67
0.30
0.74
0.20
0.85
2.73
2.73
5.55
Falk et al.
(2008)
-86.4131
Deciduous
Forest
2001-2006
MIC-TEM NEP
TEM NEP
MIC-TEM GPP
TEM GPP
0.70
0.54
0.87
0.50
1.85
4.98
5.32
12.94
Schmid et
al.(2000)
Site Name
Latitude
Longitude
Vegetation
type
Years
Howland
Forest West
Tower
(ME, USA)*
45.2091
-68.7470
Evergreen
Forest
2000-2004
Harvard
Forest
(MA, USA)*
42.5378
-72.1715
Deciduous
Forest
Niwot Ridge
(CO ,USA)
40.0329
-105.5464
Wind River
Crane Site
(WA,USA)
45.8205
Morgan
Monroe State
Forest
(IN, USA)
39.3232
10
Willow Creek
(WI,USA)
176
45.8059
-90.0799
Deciduous
Forest
2000-2003
MIC-TEM NEP
TEM NEP
MIC-TEM GPP
TEM GPP
0.69
0.51
0.96
0.71
1.87
2.97
3.49
4.51
Cook et al.
(2004)
*Sites for parameterization
11
177
Fig. S1 Vegetation distribution of forests in the conterminous United States at a resolution of
178
0.5°×0.5°
179
180
12
181
Fig. S2 The mean and standard deviation of the first order sensitivity index (Si) of soil microbial
182
RH with respect to each selected controlling parameters. Six out of ten selected parameters are
183
presented here because the others do not control RH processes. RH means soil microbial
184
respiration at 10cm depth
185
186
13
187
Fig. S3 Sensitivity of RH responding to model input (±10% change) in forest ecosystems. Top
188
panel: evergreen forest at site Howland forest west tower, ME; Low panel: deciduous at site
189
Harvard forest, MA
190
14
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
References
Czimczik CI, Trumbore SE, Carbone MS, Winston GC (2006) Changing sources of soil
respiration with time since fire in a boreal forest. Glob Chang Biol 12:957-971
Chen M, Zhuang QL, Cook QR et al (2011) Quantification of terrestrial ecosystem carbon
dynamics in the conterminous United States combining a process-based biogeochemical
model and MODIS and AmeriFlux data. Biogeosci 8: 2665-2011
Cook BD, Daviss K, Wang W et al (2004) Carbon exchange and venting anomalies in an upland
deciduous forest in northern Wisconsin, USA. Agricultural and Forest Meteorology
126:271-295
Davidson EA, Samanta S, Caramori SS, Savage K (2012) The Dual Arrhenius and Michaelis–
Menten kinetics model for decomposition of soil organic matter at hourly to seasonal
time scales. Glob Chang Biol 18:371-384
Falk M, Wharton S, Schroeder M, Ustin S (2008) Flux partitioning in an old-growth forest:
seasonal and interannual dynamics. Tree Physiology 28: 509-520
He Y, Zhuang QL, Harden J et al (2014) The implications of microbial and substrate limitation
for the fates of carbon in different organic soil horizon types: a mechanistically based
model analysis. Biogeosci 11: 2227-2266
Hollinger D, Goltz S, Davidson E et al (1999) Seasonal patterns and environmental control of
carbon dioxide and water vapour exchange in an ecotonal boreal forest. Glob Chang Biol
5: 891-902
Hollinger D, Aber J, Dail B et al (2004) Spatial and temporal variability in forest–atmosphere
CO2 exchange. Glob Chang Biol 10: 1689-1706
Kottek M, Grieser J, Beck C, Rudolf B, Rubel F (2006) World map of the Köppen-Geiger
climate classification updated. Meteorologische Zeitschrift 15: 259-263
Lawrence CR, Neff JC, Schimel JP (2009) Does adding microbial mechanisms of decomposition
improve soil organic matter models? A comparison of four models using data from a
pulsed rewetting experiment. Soil Biol Biochem 41:1923-1934
Melillo JM, McGuire AD, Kicklighter DW et al (1993) Global climate change and terrestrial net
primary production. Nature 363: 234-240
Monson R, Turnipseed A, Sparks J et al (2002) Carbon sequestration in a high‐elevation,
subalpine forest. Glob Chang Biol 8: 459-478
Schimel D, Melillo J, Tian H, McGuire AD, Kicklighter D (2000) Contribution of increasing
CO2 and climate to carbon storage by ecosystems in the United States. Science 287:
2004-2006
Taylor KE, Stouffer RJ, Meehl GA (2012) An overview of CMIP5 and the experiment design.
Bullet America Meteorol Soci 93: 485-498
Urbanski S, Barford C, Wofsy S, Kucharik C, Pyle E (2007) Factors controlling CO2 exchange
on timescales from hourly to decadal at Harvard Forest. JGR-Biogeosci 112
Yi S, Manies K, Harden J, McGuire AD (2009) Characteristics of organic soil in black spruce
forests: Implications for the application of land surface and ecosystem models in cold
regions. Geophys Rese Lett 36: L05501
Yi S, McGuire AD, Kasischke E, Harden J, Manies K, Mack M, Turetsky M (2010) A dynamic
organic soil biogeochemical model for simulating the effects of wildfire on soil
environmental conditions and carbon dynamics of black spruce forests. JGR 115:G04015
15