PERSPECTIVES
doi:10.1038/nature10386
Persistence of soil organic matter as an
ecosystem property
Michael W. I. Schmidt
1
*, Margaret S. Torn
2,3
*, Samuel Abiven
1
, Thorsten Dittmar
4,5
, Georg Guggenberger
6
, Ivan A. Janssens
7
,
Markus Kleber
8
, Ingrid Ko
¨
gel-Knabner
9
, Johannes Lehmann
10
, David A. C. Manning
11
, Paolo Nannipieri
12
, Daniel P. Rasse
13
,
Steve Weiner
14
& Susan E. Trumbore
15
Globally, soil organic matter (SOM) contains more than three times as much carbon as either the atmosphere or
terrestrial vegetation. Yet it remains largely unknown why some SOM persists for millennia whereas other SOM
decomposes readily—and this limits our ability to predict how soils will respond to climate change. Recent analytical
and experimental advances have demonstrated that molecular structure alone does not control SOM stability: in fact,
environmental and biological controls predominate. Here we propose ways to include this understanding in a new
generation of experiments and soil carbon models, thereby improving predictions of the SOM response to global
warming.
Understanding soil biogeochemistry is essential to the stewardship of
ecosystem services provided by soils, such as soil fertility (for food,
fibre and fuel production), water quality, resistance to erosion and
climate mitigation through reduced feedbacks to climate change. Soils
store at least three times as much carbon (in SOM) as is found in either
the atmosphere or in living plants
1
. This major pool of organic carbon
is sensitive to changes in climate or local environment, but how and
on what timescale will it respond to such changes? The feedbacks
between soil organic carbon and climate are not fully understood,
so we are not fully able to answer these questions
2–7
, but we can
explore them using numerical models of soil-organic-carbon cycling.
We can not only simulate feedbacks between climate change and
ecosystems, but also evaluate management options and analyse carbon
sequestration and biofuel strategies. These models, however, rest on
some assumptions that have been challenged and even disproved by
recent research arising fromnew isotopic, spectroscopic and molecular-
marker techniques and long-term field experiments.
Here we describe how recent evidence has led to a framework for
understanding SOM cycling, and we highlight new approaches that
could lead us to a new generation of soil carbon models, which could
better reflect observations and inform predictions and policies.
The conundrum of SOM
About a decade ago, a fundamental conundrum was articulated
8
: why,
when organic matter is thermodynamically unstable, does it persist in
soils, sometimes for thousands of years? Recent advances in physics,
material sciences, genomics and computation have enabled a new
generation of research on this topic. This in turn has led to a new
view of soil-organic-carbon dynamics—that organic matter persists
not because of the intrinsic properties of the organic matter itself, but
because of physicochemical and biological influences from the sur-
rounding environment that reduce the probability (and therefore
rate) of decomposition, thereby allowing the organic matter to persist.
In other words, the persistence of soil organic carbon is primarily not
a molecular property, but an ecosystem property.
This emerging view has not been fully implemented in global models
or research design, for a variety of reasons. First, the knowledgegathered
in the past decade has often been published in outlets of traditionally
separated disciplines. As a result, confusion has arisen because these
different disciplines can use the same vocabulary to mean different
things, or vice versa. For example, ‘decomposition rates’ may mean
the rate of mass loss of fresh litter, the production rate of CO
2
in a
laboratory incubation, or the rate inferred from input and loss of an
isotopic tracer present in plant inputs to soil
9,10
. Second, the complexity
of the soil system is difficult to incorporate into one conceptual model or
to translate into a tractable yet accurate numerical model. Soil is a realm
in which solid, liquid, gas and biology all interact, and the scale of spatial
structures spans many orders of magnitude (from nanometre minerals
to football-sized soil clods). Indeed, the spatial heterogeneity of biota,
environmental conditions and organic matter may have a dominant
influence on carbon turnover and trace gas production in soils. Last,
the new knowledge remains more qualitative than quantitative. In many
cases, it tells us what is important and suggests new model structures,
but not how to parameterize them.
Recent insights into carbon cycling
Since pioneering work in the 1980s
11
, new insights gathered across
disciplines (ranging from soil science to marine science, micro-
biology, material science and archaeology) have challenged several
foundational principles of soil biogeochemistry and ecosystem models;
in particular, the perceived importance of the ‘recalcitrance’ of the
input biomass (the idea that molecular structure alone can create stable
organic matter) and of humic substances (biotic or abiotic condensa-
tion products). New observations show these to be only marginally
*These authors contributed equally to this work.
1
Department of Geography, University of Zurich, 8050 Zu
¨
rich, Switzerland.
2
Earth Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720, USA.
3
Energy and Resources
Group, University of California, Berkeley, California 94720, USA.
4
Max Planck Research Group for Marine Geochemistry, University of Oldenburg, Institute for Chemistry and Biology of the Marine
Environment, 26129 Oldenburg,Germany.
5
Max Planck Research Group for Marine Geochemistry, Max Planck Institute for Marine Microbiology, 28359 Bremen, Germany.
6
Institute of Soil Science, Leibniz
Universita
¨
t Hannover, 30419 Hannover, Germany.
7
Department of Biology, University of Antwerp, 2610 Wilrijk, Belgium.
8
Department of Crop and Soil Science, Oregon State University, Corvallis, Oregon
97331, USA.
9
Lehrstuhl fu
¨
r Bodenkunde, Technische Universita
¨
tMu
¨
nchen, 85354 Freising, Germany.
10
Department of Crop and Soil Sciences, Cornell Center for a Sustainable Future, Cornell University,
Ithaca, New York 14853, USA.
11
School of Civil Engineering and Geosciences, Institute for Research on Environment and Sustainability, Newcastle University, Newcastle NE1 7RU, UK.
12
Department of
Plant, Soil and Environmental Sciences, University of Firenze, 50144 Firenze, Italy.
13
Norwegian Institute for Agricultural and Environmental Research, 1432 A
˚
s, Norway.
14
Structural Biology, Weizmann
Institute, 76100 Rehovot, Israel.
15
Max Planck Institute for Biogeochemistry, 07745 Jena, Germany.
00 MONTH 2011 | VOL 000 | NATURE | 1
1
important for organic matter cycling
12,13
. Furthermore, loose use of the
term ‘recalcitrance’ has significantly confused the discussion in the
past.
We need to ensure that the conceptual framework that supports
our understanding of soil carbon cycling is consistent with observa-
tions and has a mechanistic basis, as only then can we start to make the
necessary advances in terrestrial ecology and improve our ability to
predict soil responses to changes in climate, vegetation or manage-
ment. Here we articulate key insights into soil carbon cycling synthe-
sized from research of the past decade, and describe the research
challenges they pose for the coming decade.
Molecular structure and decomposition
The initial decomposition rate of plant residues correlates broadly
with indices of their bulk chemical composition, such as the nitrogen
content or the fraction of plant residue that cannot be solubilized by
strong acid treatments (often operationally defined as ‘lignin’)
14
.
Accordingly, the molecular structure of biomass and organic material
has long been thought to determine long-term decomposition rates in
the mineral soil. However, using compound-specific isotopic analysis,
molecules predicted to persist in soils (such as lignins or plant lipids)
have been shown to turn over more rapidly than the bulk of the
organic matter (Fig. 1)
12,15–17
. Furthermore, other potentially labile
compounds, such as sugars, can persist not for weeks but for decades.
We therefore cannot extrapolate the initial stages of litter decomposi-
tion to explain the persistence of organic compounds in soils for
centuries to millennia—other mechanisms protect against decom-
position. Perhaps certain compounds require co-metabolism with
another (missing) compound, or microenvironmental conditions
restrict the access (or activity) of decomposer enzymes (for example,
hydrophobicity, soil acidity, or sorption to surfaces
18
).
Soil humic substances
The prevalence of humic substances in soil has been assumed for
decades
19
. Previous generations of soil chemists relied on alkali
and acid extraction methods
20
and observations of the extracted (or
residual) functional-group chemistry to describe the presence of
operationally defined ‘humic and fulvic acids’ and ‘humin’. Humin
was thought to comprise large, complex macromolecules that were
the largest and most stable SOM fraction. However, we now under-
stand that these components represent only a small fraction of total
organic matter
13,21–23
: direct, in situ observations, rather than verifying
the existence of these large, complex molecules, in fact find smaller,
simpler molecular structures, as visualized in Fig. 2 (refs 13, 22, 23).
Some of what is extracted as humic acids may be fire-derived
24,25
,
although these compounds are rare in soil without substantial fire-
derived organic matter. In any case, there is not enough evidence to
support the hypothesis that the de novo formation of humic polymers
is quantitatively relevant for humus formation in soils.
Fire-derived organic matter
Fire-derived organic matter (also called char, black carbon or pyro-
lysed carbon) is found in many soils, sediments and water bodies, and
can comprise up to 40% of total SOM in grasslands and boreal for-
ests
26
. It is not inert, but its decomposition pathways remain a mys-
tery. Fire-derived carbon was suspected to be more stable in soil than
other organic matter because of its fused aromatic ring structures and
the old radiocarbon ages of fire residues isolated from soil
27
. However,
Bulk SOM
Chemical compound class
Plant-derived Alkanoic acids
n-Alkanes
Lignin
Proteins
Hexoses
Pentoses
Total saccharides
Glucosamine
PLFA Gram-negative
PLFA Gram-positive
Bacterial hexosamines
Microbial
origin
Different
biological
sources
Fire-derived organic matter
Mean residence time (years)
Mean residence time (years)
0 50 100 200 300
0 50 100 200
??
300
Figure 1
|
Molecular structure does not control long-term decomposition of
soil organic matter (SOM). Certain plant-derived molecules (classically, long-
chain alkanoic acids, n-alkanes, lignin and other structural tissues) often persist
longer than others while leaf or needle biomass is decaying. In mineral soil,
however, these relatively persistent components appear to turn over faster than
the bulk soil (top row), except for fire-derived organic matter (bottom row).
Even components that appear chemically labile, including proteins and
saccharides of plant and microbial origin (‘Different biological sources’),
instead seem to turn over (on average) at rates similar to those of bulk SOM,
that is, on the order of years or even decades. Thus, over time, the importance of
initial quality fades and the initially fast-cycling compounds are just as likely to
persist as the slow
12,15
. This figure compiles data from surface horizons of 20
long-term field experiments (up to 23 years) in temperate climate, using
13
C
labelling to trace the residence time of bulk SOM and of individual molecular
compounds. The variation in turnover time is also seen in the compounds of
microbial origin analysed for
13
C content, phospholipid fatty acids (PLFA)
produced by Gram-negative and Gram-positive bateria and amino sugars
(hexosamines). Redrawn from ref. 15 (with permission); for clarity, we have
excluded outliers, and we have added the tentative data on fire-derived organic
matter. Data points: thin horizontal lines, 10th and 90th percentiles; box, 25th
and 75th percentiles; central vertical line, median.
RESEARCH PERSPECTIVES
2 | NATURE | VOL 000 | 00 MONTH 2011
2
fire-derived carbon does undergo oxidation and transport, as we now
know from archaeological settings
28
, soils
29,30
, and from breakdown
products in river
31
and ocean water
32,33
. In a field experiment, fire-
derived residues were even observed to decompose faster than the
remaining bulk organic matter, with 25% lost over 100 years (ref. 29).
Spectroscopic characterization shows that combustion temperature
affects the degree of aromaticity and the size of aromatic sheets, which
in turn determine short-term mineralization rates
22,34–36
.Toreconcile
the observations of decomposability with the old radiocarbon ages of
fire-derived carbon deposits
37,38
, it has been suggested that physical
protection and interactions with soil minerals play a significant part
in black-carbon stability over long periods of time
39
.
Influence of roots
Root-derived carbon is retained in soils much more efficiently than are
above-ground inputs of leaves and needles
40–42
. Isotopic analyses and
comparisons of root and shoot biomarkers confirm the dominance of
root-derived molecular structures in soil
43
and of root-derived carbon
in soil microorganisms
44
. Preferential retention of root-derived carbon
has been observed in temperate forests
45,46
, for example, where below-
ground inputs, including fungal mycelia, make up a bigger fraction of
new carbon in SOM than do leaf litter inputs
44,47
. In addition to many
above-ground inputs being mineralized in the litter layer, root and
mycorrhizal inputs have more opportunity for physico-chemical inter-
actions with soil particles
40
. At the same time, fresh root inputs may
‘prime’ microbial activity, leading to faster decomposition of older
organic matter
48,49
as well as changing community composition
50
.
Carbon allocation by plants thus plays an important part in soil carbon
dynamics, but it is not known how future changes in plant allocation
will affect soil carbon stocks
51
.
Physical disconnection
The soil volume occupied by micro-organisms is considerably less
than 1%: this occupied volume is distributed heterogeneously in
small-scale habitats, connected by water-saturated or unsaturated
pore space
18
. The availability of spatially and temporally diverse habi-
tats probably gives rise to the biodiversity that we see in soil, but this
fragmentation of habitat may restrict carbon turnover. At present, we
are far from being able to quantify the complex processes of soil
structure development and fragmentation, which have different space
scales and timescales depending on soil type, texture and manage-
ment
10
. The physical disconnection between decomposer and organic
matter is likely to be one reason for persistence of deep SOM. The
specific pedological processes operating in a given soil type that influ-
ence the distribution of organisms and substrates, such as bioturba-
tion and formation of preferential flow paths, need to be taken into
account to understand and quantify subsoil carbon dynamics, and
thus its vulnerability to decomposition
52
.
Deep soil carbon
There is a lot more deep soil carbon than we once thought, and the
underlying processes inhibiting its turnover are still largely unknown.
Despite their low carbon concentrations, subsoil horizons contribute
to more than half of the global soil carbon stocks
53
. In fact, the res-
ponse of deep soils to land-use change can equal that from the top
30 cm of soil, even though typically only the shallow depths are expli-
citly represented in models
54
. Inputs of carbon to the subsoil include
dissolved organic matter, root products, and transported particulates
from the surface
55
, but the relative importance of different sources is
not known
56
. Based on depth trends of elemental composition
(decreasing C/N ratio), isotopic composition (increasing d
13
C values)
and individual organic compounds, microbial products make up
more organic matter in subsoil horizons than do plant compounds
57
.
Organic matter in subsoil horizons is characterized by very long
turnover times that increase with depth—radiocarbon ages of 1,000 to
.10,000 years are common—but the reasons for this are not clear.
Microbial activi ty may be reduced by suboptimal environmental con-
ditions, nutrient limitation or energy scarcity, and organic matter may
be less accessible because of its sparse density or association with
reactive mineral surfaces. Microbial biomass decreases with soil
depth
58
, and community composition changes to reflect an increase
in substrate specialization
59
. Recent studies suggest that energy limi-
tation, or the converse—‘priming’ (see below) by root exudates or
dissolved organic carbon—is an important factor in the subsur-
face
48,49
. Most studies concerning these factors, however, have been
conducted in the laboratory, and their relevance in situ needs evalu-
ation. If we do not understand these mechanisms of stabilization, we
cannot predict the vulnerability of deep SOM to change.
Thawing permafrost
Permafrost soils store as much carbon (up to 1,672 3 10
15
g; ref. 60) as
was believed a decade ago to exist in all soils worldwide. During
permafrost thaw, which is expected to become widespread owing to
climate change, much of this SOM may be vulnerable to rapid miner-
alization
61
if it is primarily stabilized by freezing temperatures
62
. There
is evidence that old carbon is mobilized following permafrost
thaw
61,63
, which indicates that organic matter previously locked in
the permafrost is highly vulnerable. Moreover, the accelerated decom-
position may increase nitrogen availability, which would amplify the
direct effects of warming on microbial activity. Alleviation of nitrogen
limitation in tundra experiments led to large and rapid carbon losses,
including older carbon
64,65
. Over the very long term, however, forma-
tion of pedogenic reactive minerals in former permafrost soils may act
to stabilize SOM
66,67
, and development of soil structure may lead to
physical disconnection between organic matter and decomposers.
Despite some important recent research, surprisingly little is known
about permafrost biogeochemistry and how the landscape would
evolve with warming. Key questions surround the extent to which
permafrost carbon is additionally stabilized by other processes beyond
freezing, and the extent to which the active layer becomes saturated and
anaerobic. The extent, rates and spatial variability of these processes are
still largely unknown for permafrost soils, forming one of the major
uncertainties in predicting climate–carbon feedbacks.
Soil
Direct observation
by in situ imaging
and spectroscopy
Interpretation
Soil
Wet chemical
extraction and
characterization
a Historical view
b Emerging understanding
Humic
macromolecules
exist in soil
Simple
biomolecules
exist in soil
Interpretation
Observed
Observed
Figure 2
|
In soil, the existence of humic substances has not been verified by
direct measurements. a, Based on chemical analysis of the extracted materials
(Observed), the de novo formation of humic polymers (Interpretation) was
postulated to be an important source of recalcitrant SOM. b, Direct high-
resolution in situ observations with non-destructive techniques (Observations)
have been able to explain the functional group chemistry of the extracted humic
substances as relatively simple biomolecules (Interpretation), without the need
to invoke the presence of unexplainable macromolecules
100
. Moreover, the
chemical mixture of SOM is spatially distinct on a nanometre scale, and the
aromatic/carboxylate-rich compounds characteristic of the bulk extracted
humic substances have not been found in situ even when looking at the
submicrometre scale (using near-edge X-ray fine structure spectroscopy
combined with scanning transmission X-ray microscopy)
22
.
PERSPECTIVES RESEARCH
00 MONTH 2011 | VOL 000 | NATURE | 3
3
Soil micro-organisms
Soil microbial diversity and activity can be characterized at molecular
resolution, but the quantitative linkages to ecosystem function are
uncertain
68,69
. Soil micro-organisms influence SOM cycling not only
via decomposition but also because microbial products are themselves
important components of SOM
70
. As a result, environmental change
can influence soil carbon cycling through changes in both metabolic
activity and community structure. For example, microbial community
shifts following nitrogen additions can have large effects on decom-
position rates
50,51
. New genetic and protein-based tools enable the
quantification of soil microbiological abundance and functioning
(for example, enzymatic gene expression), and can describe the micro-
bial community composition with very high taxonomic resolution
71
.
Nevertheless, the challenge remains of synthesizing this immense
amount of detailed information
72
and linking it to the rates and routes
of SOM processing. To quantitatively relate microbial genomics to
ecosystem function, we need a better understanding of microbial func-
tional redundancy.
Implications of these insights
Taken together, these eight insights paint a broad picture of carbon
cycling in soil that has implications for fundamental research, land
management, and climate change prediction and mitigation (Fig. 3).
They suggest that the molecular structure of plant inputs and organic
matter has a secondary role in determining carbon residence times
over decades to millennia, and that carbon stability instead mainly
depends on its biotic and abiotic environment (it is an ecosystem
property). Most soil carbon derives from below-ground inputs and
is transformed, through oxidation by microorganisms, into the sub-
stances found in the soil. By moving on from the concept of recal-
citrance and making better use of the breadth of relevant research, the
emerging conceptual model of soil organic carbon cycling will help to
unravel the mysteries surrounding the fate of plant- and fire-derived
inputs and how their dynamics vary between sites and soil depths,
and to understand feedbacks to climate change. We argue that the
persistence of organic matter in soil is largely due to complex inter-
actions between organic matter and its environment, such as the
interdependence of compound chemistry, reactive mineral surfaces,
climate, water availability, soil acidity, soil redox state and the pres-
ence of potential degraders in the immediate microenvironment. This
does not mean that compound chemistry is not important for decom-
position rates, just that its influence depends on environmental factors.
Rather than describing organic matter by decay rate, pool, stability or
level of ‘recalcitrance’—as if these were properties of the compounds
themselves—organic matter should be described by quantifiable
environmental characteristics governing stabilization,such as solubility,
molecular size and functionalization
73
.
Soil response to global environmental change
We now consider how these insights affect our use of numerical models.
Such models are powerful tools for quantifying the complex interac-
tions and feedbacks that will underpin soil responses to globalchange.A
variety of models that include SOM dynamics have informed our res-
ponse to environmental issues, including agricultural management,
bioremediation and environmental water research
74
. Most model test-
ing, however, has been at local-to-regional spatial scales, spanning sea-
sons to decades (although the century-long Rothamsted experiments
are a noteworthy exception). In the long term or at a global scale,
mechanisms of SOM stabilization and destabilization that are not cur-
rently embedded in models have the potential to dominate soil carbon
dynamics, making it vital that models are correct for the right reasons.
Recent model intercomparisons reveal large differences among predic-
tions of soil carbon stocks and fluxes in the next century, for example
2
,
demonstrating how sensitive global carbon cycling is to assumptions
about SOM decomposition dynamics.
Recent advances in our mechanistic understanding of soils, such as
those described above, have not yet been incorporated into the widely
used models of SOM cycling, which are all structured around the idea
that a type, or pool, of organic material will have an intrinsic decay
rate
75–78
. These models rely on simple proxies—such as soil texture as
Condensation
reactions
Creation of
new stable
compounds
Condensation
reactions
Creation of
new stable
compounds
Fresh plant litter (leaves, stems, roots and rhizosphere); re residues
Fresh plant litter (leaves)
Rhizosphere
Inputs
Molecular structure
determines timescale
of persistence
Physical
disconnection
(e.g. from
enzymes,
decomposers,
e-acceptors)
Freezing /
thawing
Physical
disconnection
(e.g. from
enzymes,
decomposers,
e-acceptors)
Freezing /
thawing
Deep soil carbon: age of carbon reects
timescale of process. Rapid destabilization
possible with change in environmental conditions
7
8
5
6
4
2
3
1
a Historical view b Emerging understanding
Sorption/
desorption
(organo-mineral
associations)
Sorption/
desorption
(organo-mineral
associations)
Microbial products
Rhizosphere
inputs
Figure 3
|
A synopsis of all eight insights, contrasting historical and
emerging views of soil carbon cycling. The historical view (a) has emphasized
above-ground plant carbon inputs and organic matter in the top 30 cm of soil.
Stable organic matter is seen to comprise mainly selectively preserved plant
inputs and de novo synthesis products like humic substances, whose chemical
complexity and composition render them nearly inert relative to microbial
degradation. The emerging understanding (b) is that the molecular structure of
organic material does not necessarily determine its stability in soil (1; molecular
structure). Rather, SOM cycling is governed by multiple processes (5) shaped
by environmental conditions (such as physical heterogeneity). Plant roots and
rhizosphere inputs (4; roots) make a large contribution to SOM, which is
mainly partial degradation and microbial products and fire residues (3) rather
than humic substances (2). The vulnerability of deep soil carbon (6; deep
carbon) to microbial degradation (8; soil micro-organisms) in a changing
environment, such as thawing permafrost (7; thawing permafrost) remains a
key uncertainty.
RESEARCH PERSPECTIVES
4 | NATURE | VOL 000 | 00 MONTH 2011
4
a surrogate for sorption and other organo-mineral interactions, and
litter quality (such as lignin:N ratios or structural carbon groupings)
as a means of partitioning plant inputs into pools of different turnover
times—but in general these parameters are not consistent with the
observations that are starting to emerge. Global models largely ignore
deep mineral soils and are only now beginning to address the accu-
mulation and loss of carbon in peatlands and permafrost
79
. Even more
importantly, parameterizations based on litter chemistry may correl-
ate well to initial rates of litter decomposition, but they have little
relationship to the rates of decomposition for microbial residues or
to organic matter sorbed to mineral surfaces or isolated in aggregates.
Moreover, most models that make any allowances for microbial bio-
mass treat it as a pool of carbon, rather than as an agent that affects the
decomposition rate of SOM. The large disagreement among predic-
tions of soil carbon fluxes in a warmer world highlights both the
complexity of the many potential feedbacks to climate and the uncer-
tainty that arises as a result
2
.
How does the perspective that SOM persistence is an ecosystem
property inform our understanding of the response of decomposition
to warming? The conventional assumption that older SOM is recalcit-
rant implies that this large carbon pool is highly temperature sensitive,
because Arrhenius kinetics tells us that reactions with higher activation
energies are more temperature-sensitive than those with low activation
energies
80–82
. Our ecosystem perspective suggests that the mechanisms
governing the timing and magnitude of a response to a change in
temperature are far more complex than this, as further physical, chem-
ical and biological mechanisms controlling decomposition and stabil-
ization would also be affected
81,83,84
. A recent incubation study of soils
from a wide range of sites found that lower initial decomposition rates
were associated with higher temperature sensitivity but not with any
change in SOM quality indices
85
, suggesting that multiple stabilization
mechanisms are temperature sensitive. Nevertheless, it is not yet pos-
sible to predict the integrated response of decomposition to changes in
climate. In fact, we could use the ability to accurately predict temper-
ature response as a guide to the degree of mechanistic representation
that we need in our next generation of soil carbon models.
Phyto-engineering
Phyto-engineering to produce plant tissues high in chemical com-
pounds resistant to rapid mineralization, such as plant lipids and
lignin, has been suggested as a means to increase carbon sequest-
ration
86
. This strategy is called into question, however, if the molecu-
lar structure of plant compounds does not determine stability on the
timescales necessary for significant carbon sequestration
12,87
. More
generally, sequestration strategies based on adding recalcitrant mater-
ial to soils, whether through plant selection for recalcitrant tissues or
through biochar amendments, must be re-evaluated. Enhancing root
carbon input to soils might be a more promising avenue, but it is not
known what root properties influence rhizodeposition rates or
stability
43
, or the extent to which root inputs will stimulate (prime)
decomposition of other SOM.
Biochar
Biochar (intentionally pyrolysed biomass) has gained much attention
in recent years as a means to increase soil fertility and store carbon in
soil for decades to centuries
88
. However, certain types of biochar can
degrade relatively rapidly in some soils, probably depending on the
conditions under which they were produced, which suggests that
pyrolysis could be optimized to generate a more stable biochar. But
as with natural fire residues, persistence over the long term may also
be affected by interactions with minerals and by soil conditions (for
microorganisms capable of char oxidation and for abiotic oxidation).
Whether interactions of fire-derived carbon with soil minerals may be
manipulated to enhance stability, and what the trade-offs might be
with fertility benefits, are not known. Biochar is likely to be a useful
part of sequestration-mitigation strategies, but more understanding
of the variation in its decay rates is needed before we can develop
simple (that is, policy-relevant) quantitative relationships between
biochar additions and expected sequestration.
Vulnerability of soil to degradation
The vulnerability of SOM to degradation will depend on the nature of
the disturbance as well as the stabilization and destabilization
mechanisms at play in a given ecosystem. Hence, as with carbon
stability, the vulnerability of soil stocks should not be assessed accord-
ing to the classes of organic matter present, but rather according to the
mechanisms through which organic matter is stabilized or made
assimilable in that soil, and how these interacting physical, chemical
and biological factors respond to change
5,6
. Improved understanding
of SOM destabilization is needed to enhance efforts to avoid soil
degradation and accelerate recovery of degraded soils.
The way forward
Soils are now in the ‘front line’ of global environmental change—we
need to be able to predict how they will respond to changing climate,
vegetation, erosion and pollution so that we can better understand
their role in the Earth system and ensure that they continue to provide
for humanity
89
and the natural world. The conceptual framework of
soil carbon cycling presented here, that residence time is a property of
the interactions between organic matter and the surrounding soil
ecosystem, will help us get nearer to these goals
7
. This will require
developed and entirely new lines of research and modelling, includ-
ing: (1) applying a new generation of field experiments and analytical
tools to study the processes driving SOM stabilization and destabiliza-
tion; (2) developing a new generation of soil biogeochemistry models
that represent the mechanisms driving soil response to global change;
and (3) joining forces and connecting the disparate research com-
munities that are studying, managing and predicting SOM cycling
and terrestrial ecology.
The next generation of experiments
Although not a novel recommendation, we cannot overstate the need
for long-term, manipulative experiments designed to test soil-based
hypotheses. In some countries, long-term ecological observational net-
works already exist, but most were designed with vegetation or hydro-
logy goals. Many are in danger of being discontinued. Although
preserving these experiments is crucial, they may not be sufficient to
untangle individual soil processes. In the near term, new disciplines and
techniques could be applied to ongoing experiments, allowing the
investigation of changes that occurred after decades of manipulation
90
.
Focus is needed on long-term, controlled manipulations of entire soil
profiles (that is, to a metre or more depth) to investigate distinct
mechanisms in situ, and on observatories allowing quantification of
budgets, such as large-scale lysimeters. In addition, research approaches
are needed that combine manipulations with spatial gradients—and
thus timescales—for variables and processes of interest. These new
experiments should be designed to help determine the key soil func-
tional traits for understanding and modelling thresholds in SOM stor-
age and loss. Such traits, including soil depth, mineral charge density
and pH, vary spatially, but we suggest that their spatial distributions are
ultimately predictable according to geologic setting, disturbance and
management history, climate and ecosystem plant characteristics—in
other words, the six state factors: climate, organisms, relief, parent
material, time and human activity
91
. One of the major weaknesses of
current models is the lack of representation of edaphic characteristics
(that is, those physical and chemical features that are intrinsic to the
soil)—and the fact that the major stabilization mechanisms will vary
spatially with soil type and topographic positions.
Tracing pathways, fluxes and biology
When combined with manipulative experiments, new analytical tech-
niques and instrumentation to study elements, isotopes and molecules
PERSPECTIVES RESEARCH
00 MONTH 2011 | VOL 000 | NATURE | 5
5