Auxiliary Material For

Auxiliary material for

Quantifying microbial ecophysiological effects on the carbon fluxes of forest ecosystems over the conterminous United States

Guangcun Hao1,2, Qianlai Zhuang2,3*, Qing Zhu2,4 , Yujie He2, Zhenong Jin2 and Weijun Shen1,

1Key Laboratory of Vegetation Restoration and Management of Degraded Ecosystems, South China Botanical Garden, Chinese Academy of Sciences, Guangzhou 510650, China

2Department of Earth, Atmospheric, and Planetary Sciences, Purdue University, West Lafayette, Indiana, USA

3Department of Agronomy, Purdue University, West Lafayette, Indiana, USA

4Earth Science Division, Lawrence Berkeley National Lab, Berkeley, CA, USA

*Correspondence to:

Qianlai Zhuang

Appendix.1

The Terrestrial Ecosystem Model

In TEM, carbon exchanges start from the simulation of gross primary production (GPP). GPP is modeled as a function of multi-scalars that represent a variety of environmental and physical constraints. The formula for calculating daily GPP is:

（1）

Where Cmax is the maximum rate of C assimilation by the entire plant canopy under optimal environmental conditions, and f(PAR), f(p), f(FOLIAGE), f(NA) and f(FT) represent the limits of the photosynthetically active radiation (PAR), the leaf phenology, the influence of relative canopy leaf biomass relative to maximum leaf biomass (Zhuang et al. 2002), the limiting effects of plant nitrogen availability and the effects of freeze-thaw dynamics on GPP (Zhuang et al. 2003), respectively. The term CA represents the influence of increasing atmospheric CO2 concentration on GPP, which is modelled by following Michaelis-Menten kinetics (Raich et al. 1991); Gv accounts for changes in leaf conductivity to CO2 resulting from moisture availability, which is based on the estimates of evapotranspiration (ET). Net Ecosystem Production (NEP) is defined as the difference between Net Primary Production (NPP) and Heterotrophic Respiration (RH), and NPP is defined as the difference between Gross Primary Production (GPP) and Autotrophic Respiration (RA).

NPP=GPP – RA ; （3）

NEP=NPP - RH ; （4）

Appendix.2

Definition of key terminology

In the paper, soil microbial physiology refers to microbial growth activity and depolymerization activities. Carbon flux mean GPP, NPP, NEP, RA and RH .

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 microbial respiration from assimilated soluble C, via which O2 is consumed to produce energy for assimilation of dissolved organic C:

(1)

Assimilation is a Michaelis-Menten function scaled to the pool size of microbial biomass:

(2)

where is the maximum velocity of the enzymatic reaction when substrate is not limiting. is the corresponding Michaelis constant. The concentration of soluble C substrates at the reactive site of the enzyme ([Sx]) is affected by soil water content, and specifically by diffusion of substrates through soil water films. [Sx] is calculated from [Sxsoluble] through , whereis the volumetric water content of the soil, and Dliq is a diffusion coefficient of the substrate in liquid phase. Diffusion of soluble substrates have been shown to be related to the thickness of the soil water films, which is approximated by the cube of the volumetric water content. It is assumed that the cell surface area available for [Sx] uptake is proportional to the number of cells, and thus the microbial biomass [Davidson et al. 2012]. [Sx] is assumed to be the only substrate for microbial C uptake. Similar to Davidson et al. (2012), the value of Dliq is determined by assuming the boundary condition that all soluble substrate is available at the reaction site for saturated soil (i.e., ).

CO2 (heterotrophic respiration) is produced as the part of microbial assimilated C not allocated to biomass growth. The production process follows Michaelis-Menten kinetics similar to assimilation but is controlled by the concentration of both [Sx] and O2:

(3)

The concentration of O2 at the reactive site of the enzyme ([O2]) depends upon diffusion for gases within the soil medium, which is modeled with a simple function of air-filled porosity:. Dgas is a diffusion coefficient for O2 in air, 0.209 is the volume fraction of O2 in air, and is the air-filled porosity of the soil. The total porosity is calculated from bulk density (BD) and particle density (PD):

.

,, and are temperature dependent. and follow the Arrhenius equation:

(4)

(5)

where and are the pre-exponential coefficient (i.e., the theoretical decomposition enzymatic reaction rate at Ea = 0), R is the ideal gas constant (8.314 J K-1 mol-1), is the temperature in Celsius, and and are the activation energy for [Sx] uptake and CO2 respiration by microbial. High activation energy indicates high temperature sensitivity but reacts slowly. is calculated as a linear function of temperature, as adopted in Davidson et al. (2012):

(6)

where and are the intercept and slope parameters, respectively. is assumed to be constant with respect to temperature for the sake of model parsimony. However, could be modeled as a function of temperature when observations are available.

Microbial death is modeled as a first-order process with rate constant(Lawrence et al. 2009):

(7)

Enzyme production is modeled as a constant fraction () of microbial biomass (Lawrence et al. 2009):

(8)

The enzyme pool changes with enzyme production and turnover:

(9)

where the turnover (ELOSS) is modeled as a first-order process with constant rate:

(10)

The changes in SOC pool varies with external inputs from vegetation litterfall carbon, enzyme turnover, inputs from dead microbial biomass () and decomposition loss:

(11)

where enzymatic decomposition of SOC (DECAY) here is mainly referring to the process through which microbes secrete exoenzymes to convert macromolecules into soluble products (soluble C, denoted as [Sxsoluble]) that can be absorbed and metabolized by microbes. This process follows Michaelis-Menten kinetics with enzyme and substrate (here SOC) constraint:

(12)

where is the maximum velocity of the enzymatic reaction when substrate is not limiting and is calculated according to Arrhenius function:

(13)

We assume Michaelis-Menten constant for () is invariable with temperature. The soluble C pool ([Sxsoluble]) changes with external inputs, the remaining fraction of dead microbial biomass, and decomposition:

(14)

This process represents the enzymatic depolymerization of complex molecules to the simpler ones available for microbial uptake. More detailed algorithms are documented in He et al. (2014).

Appendix.3

Data organization and model simulations

We organized the observed or estimated GPP, NEP, and meteorological data (e.g. radiation, air temperature, and precipitation) from six representative eddy covariance flux sites for each vegetation type to parameterize the MIC-TEM. Data from other four AmeriFlux forest sites are used to verify the model performance (Table S2). Specifically, we collected all available daily Level 4 Net Ecosystem Exchange (NEE) products (http://public.ornl.gov/ameriflux/) of these sites (Table S2). For each site, if the percentage of remaining missing values for NEE_st or GPP_st (standardized data) is lower than that for NEE_or or GPP_or(original data), we select NEE_or or GPP_or; otherwise, we use NEE_st or GPP_st. The measured NEE is then compared with modeled NEP (Chen and Zhuang et al. 2011).

We spun up the model for 120 years to account for the influence of climate inter-annual variability on the initial conditions of the ecosystems. Since historic climate data are not available before 1948, we repeat the data from 1948 to 1987 for 3 times for the spin-up with NECP global datasets at a 0.5 spatial resolution.

To quantify the effects of microbial activity on regional carbon dynamics, we applied the both original and revised TEM to the forest ecosystems of the conterminous United States for the period 2006 – 2100 with daily time-step and 0.5° × 0.5° spatial resolution driving data (Kottek et al. 2006) (Fig. S1). The spatially-explicit soil texture, elevation data and vegetation type data are from our previous studies (Melillo et al. 1993; Zhuang et al. 2003). Future climate scenarios from 2006 to 2100 are generated under two RCPs of the Coupled Model Inter-comparison Project phase 5 (CMIP5) with NOAA’s Earth System Models (GFDL-ESM2G). A general description of CMIP5 models and the experiment design can be found in Taylor et al. (2012). All climate data sets are resampled to 0.5° × 0.5° spatial resolution with statistical downscaling method from Mitchell et al. (2004).

13

Table S1 Parameters used in the model. Inversed estimates of specific parameters and parameter ranges used are listed

Process / Parameter / Unit / Initial Value / Description / Parameter range / References
Assimilation / 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 / Temperature regulator of MM for Sx uptake by microbes (kM_uptake) / - / Allison et al. (2010)
m_uptake / mg Sx cm-3 soil °C-1 / 0.01 / Temperature regulator of MM for Sx uptake by microbes (kM_uptake) / - / Allison et al. (2010)
Ea_Sx / J mol-1 / 48092 / Activation energy of microbes assimilating Sx to CO2 / - / Knorr et al. (2005)
c_Sx * / mg assimilated Sx cm-3 soil / 0.1 / Temperature regulator of MM for microbial assimilation of Sx (kM_Sx) / - / Allison et al. (2010)
m_Sx * / mg assimilated Sx cm-3 soil °C-1 / 0.01 / Temperature regulator of MM for microbial assimilation of Sx (kM_Sx) / - / Allison et al. (2010)
Decay / Ea_SOC / J mol-1 / 41000 / Activation energy of decomposing SOC to soluble C / - / Modified from Davidson et al. (2012)
Vmax_SOC0 / mg decomposed SOC cm-3 soil (mg Enz cm-3 soil)-1 h-1 / 9.17e7 / Maximum rate of converting SOC to soluble C / [1.0e5, 1.0e8] / -
c_SOC / mg SOC cm-3 soil / 400 / Temperature regulator of MM for enzymatic decay of SOC to soluble C (kM_SOC) / - / Allison et al. (2010)
m_SOC / mg SOC cm-3 soil °C-1 / 5 / Temperature regulator of MM for enzymatic decay of SOC to soluble C (kM_SOC) / - / Allison et al. (2010)
kM_O2 / cm3O2 cm-3 soil / 0.121 / Michaelis-Menten constant (MM) for O2 (at mean value of volumetric soil moisture) / - / Davidson et al. (2012)
CO2
production / Vmax_CO20 / mg respired Sx cm-3 soil h-1 / 1.9e7 / Maximum microbial respiration rate / [1.0e6, 1.0e8] / -
c_Sx * / mg assimilated Sx cm-3 soil / 0.1 / Temperature regulator of MM for microbial respiration of assimilated Sx (kM_Sx) / - / Allison et al. (2010)
m_Sx * / mg assimilated Sx cm-3 soil °C-1 / 0.01 / Temperature regulator of MM for microbial respiration of assimilated Sx (kM_Sx) / - / Allison et al. (2010)
MIC turnover / MICtoSOC / % / 50 / Partition coefficient for dead microbial biomass between the SOC and Soluble C pool / - / Allison et al. (2010)
r_death / % h-1 / 0.02 / Microbial death fraction / - / Allison et al. (2010)
ENZ turnover / r_EnzProd / % h-1 / 5.0e-4 / Enzyme production fraction / - / Allison et al. (2010)
r_EnzLoss / % h-1 / 0.1 / Enzyme loss fraction / - / Allison et al. (2010)

* c_Sx and m_Sx are used in both assimilation and CO2 production calculations

13

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

Site Name / Latitude / Longitude / Vegetation type / Years / R2 / RMSE / References
Howland Forest West Tower
(ME, USA)* / 45.2091 / -68.7470 / Evergreen Forest / 2000-2004 / 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)
Harvard Forest
(MA, USA)* / 42.5378 / -72.1715 / Deciduous Forest / 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)
Niwot Ridge (CO ,USA) / 40.0329 / -105.5464 / 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)
Wind River Crane Site (WA,USA) / 45.8205 / -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)
Morgan Monroe State Forest
(IN, USA) / 39.3232 / -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)
Willow Creek (WI,USA) / 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

13

Fig. S1 Vegetation distribution of forests in the conterminous United States at a resolution of 0.5°×0.5°

Fig. S2 The mean and standard deviation of the first order sensitivity index (Si) of soil microbial RH with respect to each selected controlling parameters. Six out of ten selected parameters are presented here because the others do not control RH processes. RH means soil microbial respiration at 10cm depth

Fig. S3 Sensitivity of RH responding to model input (±10% change) in forest ecosystems. Top panel: evergreen forest at site Howland forest west tower, ME; Low panel: deciduous at site Harvard forest, MA

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