Soil organic matter evolution during long-term bare fallow and implications for sensitivity to global change P.
a,* Barré ,
b,* Cécillon ,
c,* Plante ,
d Chenu ,
e Christensen ,
c Fernandez ,
L. A.F. C. B.T. J.M. C.M. Gherardia,c,d, S. Houotf, T. Kättererg, F. van Oorth, C. Peltrec & P.R. Poultoni
aGeology
laboratory, UMR CNRS-ENS, Ecole normale supérieure, 75005 Paris, France; bUnité Ecosystèmes Montagnards, IRSTEA, Grenoble, France; cEarth and Environmental Science, University of Pennsylvania, Philadelphia, USA; dBIOEMCO laboratory, UMR ParisVI-ParisXII-AgroParisTech-CNRS-IRD-ENS, AgroParisTech, Campus Grignon, 78850 Thiverval-Grignon, France; eAarhus University, Department of Agroecology, AU Foulum, DK-8830 Tjele, Denmark; fEGC laboratory, UMR INRA-AgroParisTech, AgroParisTech, Campus Grignon, 78850 Thiverval-Grignon, France; gSwedish University of Agricultural Sciences, Department of Soil & Environment, SE-75007 Uppsala, Sweden; hPESSAC laboratory, INRA, F-78026 Versailles, France; iRothamsted Research, Department of Sustainable Soils and Grassland Systems, Harpenden, UK * Equal contribution to the work
INTRODUCTION
SAMPLES & METHODS
•Better characterization of soil organic carbon (SOC) biogeochemical stability is necessary to more accurately predict SOC sensitivity to global change and disturbance. •Long-term bare fallow (LTBF) experiments, in which soil organic matter (SOM) is gradually enriched in stable C as labile components decompose, provide a unique opportunity to study stable SOC without the inherent artefacts induced by extraction procedures. •We investigated bulk soil samples from five LTBF experiments using thermal (thermogravimetry and differential scanning calorimetry; TG-DSC) and molecular (diffuse reflectance infrared Fourier transform spectroscopy; DRIFTS) techniques to answer: 1. Is SOM decomposition energy constrained (i.e., by its energetic cost/benefit ratio for decomposers), as suggested by Fontaine et al. (2007)? 2. Does SOM converge to a specific chemistry during long-term decomposition?
Rothamsted
400
Energy density (J/mg OM)
TG-T50 (°C)
Versailles (79 years)
380 360
C chemistry • Samples were analyzed by DRIFT spectroscopy. Spectra were baseline corrected and normalized to the CH2 symmetric stretching band (2851cm-1). • Carbonated Grignon site was discarded • DRIFT spectra were subjected to a principal component analysis (PCA) in the mineral free CH3/CH2 stretching absorption bands region (2800-3000 cm-1)
C chemistry
10
20
30
40
50
60
70
5 0
80
0
10
20
30
40
50
60
70
80
0 0
10
20
30
40
50
60
70
80
10 20
30
40
50
60
70
80
10 20
30
40 50
60
70
80
15
Ultuna (53 years)
Energy density (J/mg OM)
TG-T50 (°C)
420 400
10
380 360
5
340 0
10
20
30
40
50
60
70
80
420 Rothamsted (49 years) 400
15 Energy density (J/mg OM)
TG-T50 (°C)
• Samples were heated from 20°C to 750 °C and analyzed by TG & DSC • Two indicators were computed: TG-T50, temperature at which half of SOM was combusted, and energy density, total exothermic energy content divided by TG mass loss
10
0
10
380 360
5
340 0
10
20
30
40
50
60
70
0 0
80
420 Grignon (48 years) 400
15 Energy density (J/mg OM)
TG-T50 (°C)
C energy
15
340
10
380 360 340
5 0
0 420
10
20
30
40
50
60
70
80
0 15
Askov (27 years)
Energy density (J/mg OM)
TG-T50 (°C)
Askov
Grignon & Versailles
C energy 420
Ultuna
• Samples came from 5 European LTBF experiments, 27 to 79 years old. Samples from control plots (plots with regular C inputs) were available at 4 sites.
400
The first two axes of the PCA of DRIFT spectra in the alkyl-C region accounted for 91% of total variance
10
380 360 340
5 0
0
10
20
30
40
50
60
70
Experiment duration (years)
80
0
10 20
30
40 50
60 70
80
Experiment duration (years)
All BF plots (black diamonds) followed consistent energetic trends, with increased burning temperature (TG-T50) and decreased energy density with increasing bare fallow duration. Control plots (green squares) showed no trends in time
High PC1 scores are characterized by an increase in CH2 relative to CH3 (i.e. increase in hydrocarbon chain length). Conversely, higher PC2 scores indicate a decrease in hydrocarbon chain length Clear and opposite chemical pathways to stable SOC (see black arrows) were observed for Versailles and Rothamsted BF plots regarding alkyl-C chemistry, while no chemical evolution was observed for Ultuna or Askov bare fallows compared to control plots
CONCLUSIONS •In spite of differing characteristics among the five LTBF sites (e.g. texture, climate, past land-use), SOM followed consistent energetic trends during SOC stabilization. The rate of this consistent energetic pathway to stable SOC is ecosystem dependent. •The decrease in energy density and the increase in combustion temperature suggest that priming may be the only means available for microbial decomposition of the remaining C, and that SOC stability may be a function of the high energetic cost/benefit ratio for microbial decomposition. •Conversely, no consistent chemical trends were observed. The absence of chemical homogeneization during SOM evolution, evidenced recently for litter decomposition (2 years litter bag experiment; Wickings et al., 2012) is also observed for SOM stabilization in soils (30 to 80 years LTBF experiments). •This shows that energy is a major constraint on SOC evolution during stabilization but that this mandatory energetic evolution could be reached through various chemical evolutions. In other words, stable SOC persists because of its intrinsic energetic characteristics, but its chemistry and the rate of its energetic evolution are ecosystem properties. References : - Fontaine et al. (2007) Stability of organic carbon in deep soil layers controlled by fresh carbon supply, Nature, 450, 277-281 - Wickings et al. (2012) The origin of litter chemical complexity during decomposition, Ecology Letters, 15,1180-1188
MOLTER and l’ADEME contributed to the funding of this work
