ARTICLE
Temperature effects on carbon storage are
controlled by soil stabilisation capacities
Iain P. Hartley
1
✉
, Tim C. Hill
1
, Sarah E. Chadburn
& Gustaf Hugelius
3,4
Physical and chemical stabilisation mechanisms are now known to play a critical role in
controlling carbon (C) storage in mineral soils, leading to suggestions that climate warming-
induced C losses may be lower than previously predicted. By analysing > 9,000 soil pro
fi
les,
here we show that, overall, C storage declines strongly with mean annual temperature.
However, the reduction in C storage with temperature was more than three times greater in
coarse-textured soils, with limited capacities for stabilising organic matter, than in
fi
ne-
textured soils with greater stabilisation capacities. This pattern was observed independently
in cool and warm regions, and after accounting for potentially confounding factors (plant
productivity, precipitation, aridity, cation exchange capacity, and pH). The results could not,
however, be represented by an established Earth system model (ESM). We conclude that
warming will promote substantial soil C losses, but ESMs may not be predicting these losses
accurately or which stocks are most vulnerable.
https://doi.org/10.1038/s41467-021-27101-1
OPEN
1
Geography, College of Life and Environmental Sciences, University of Exeter, Exeter EX4 4RJ, UK.
2
Mathematics, College of Mathematics and Physical
Sciences, University of Exeter, Exeter EX4 4RJ, UK.
3
Department of Physical Geography, Stockholm University, SE-106 91 Stockholm, Sweden.
4
Bolin Centre
for Climate Research, Stockholm University, SE-106 91 Stockholm, Sweden.
✉
email:
NATURE COMMUNICATIONS
| (2021) 12:6713 | https://doi.org/10.1038/s41467-021-27101-1 | www.nature.com/naturecommunications
123456789
0():,;
I
n cold and wet regions, low soil temperatures and/or anaerobic
conditions promote the formation of thick organic horizons
and peats, resulting in the storage of ~500 Pg of C
and
helping drive a latitudinal gradient in soil C stocks
. However, for
soils or horizons that are a mixture of organic and mineral
material (referred to as mineral soils), and which contain more
than two-thirds of the global soil C stock
, the role of climatic
variables in controlling C storage has been increasingly
questioned
. It is now recognised that, in the absence of
anaerobic conditions, the long-term persistence of soil organic
matter (SOM) in mineral soils depends more on its accessibility
to decomposers than on its intrinsic decomposability
. Physi-
cochemical stabilisation mechanisms, including the binding of
organic matter with mineral surfaces and physical protection
within aggregates, limit SOM availability to microbes and allow
organic matter to persist. Because these soil properties are so
important in determining C stocks, it has been suggested that the
role of temperature in determining C storage may have been
overestimated, limiting the potential for a substantial release of C
from soils as the climate warms
. However, the extent to which
physicochemical
stabilisation
mechanisms
are
themselves
temperature-sensitive is poorly understood
. For example, it
has both been argued that sorption (stabilisation) reactions
should be more
and less
sensitive than desorption (destabi-
lisation) reactions, with the relative temperature sensitivities
potentially varying with the af
fi
nity of organic matter for mineral
. Evidence has been produced that the decomposition of
older, and more protected SOM, is more sensitive to temperature
than rapidly decomposing SOM, suggesting that destabilisation
reactions may indeed be highly temperature sensitive
–
However, understanding remains very limited. Thus, the extent to
which temperature controls C storage in mineral soils remains
controversial, and it is not known whether the C stored in
mineral soils with large capacities to stabilise C will be more or
less vulnerable to climate warming than the C stored in soils with
more limited stabilisation capacities.
In recent years, important initiatives have started to bring
together soil data from across the globe to provide databases that
can help answer some of the most pressing environmental
questions facing humanity, including the impacts of climate
change
. For example, the World Soil Information (WoSIS)
database published a snapshot of the freely available soil pro
fi
le
data from across six continents
. Critically, more than 9300
pro
fi
les from non-cropland systems included data on both soil
texture and soil C storage by depth (Fig.
a) offering a major
opportunity for improving understanding of how stabilisation
mechanisms control the effects of climate on soil C storage. This
is because soil texture, in particular clay content, has been shown
to be a strong predictor of a soil
’
s C stabilisation capacity
–
directly determining the potential for chemical stabilisation
through the formation of organo-mineral complexes, and indir-
ectly affecting physical protection through its in
fl
uence on
aggregate dynamics
. Thus, it is now possible to empirically
determine whether the effect of temperature on soil C storage is
related to SOM stabilisation capacities.
In this study, by analysing the soil pro
fi
le data
, we investi-
gated whether the effect of temperature on soil C storage varies
with stabilisation capacities, as indicated by soil textural proper-
ties. We hypothesised, because of the evidence that the tem-
perature sensitivity of decomposition is greater for older and
more protected SOM pools
–
, that the effects of temperature
on soil C storage would be strongest in
fi
ne-textured soils with
greater stabilisation capacities. In contrast to our expectations, we
observed a much greater reduction in C storage with increasing
mean annual temperature in coarse-textured soils, suggesting that
unprotected pools are most affected by temperature. The patterns
were observed independently for warm and cool regions, across
different soil sampling depths, and were maintained after
accounting for potentially confounding variables. Our
fi
ndings
emphasise the potential for C to be released from soils in response
to climate warming, help identify the soil C stocks that may be
most vulnerable, and represent a powerful dataset for the eva-
luation of Earth systems models.
Results and discussion
Effects of temperature on C storage in soils with contrasting
stabilisation capacities
. Using a space-for-time approach, we
de
fi
ned the effect of temperature on C storage as the proportional
reduction in C storage for each 10
o
C increase in mean annual
temperature. In this context, a value of 1 indicates no change in C
storage with temperature, values less than 1 indicate C stocks
increase with temperature and values greater than 1 indicate C
stocks decline with temperature, with, for example, a value of 2
indicating that C stocks halve for every 10
o
C increase in tem-
perature. C storage in the top 50 cm of mineral soil declined
strongly with increasing temperature, declining by a factor of ~1.4
per 10
o
C (Fig.
b). Critically, the nature of this relationship was
modi
fi
ed by soil clay content; C storage in
fi
ne-textured soils with
greater stabilisation capacities was affected much less by tem-
perature than C storage in coarse-textured soils (Fig.
a and
Supplementary Fig. 1; factors of up to 1.9 per 10
o
C for coarse-
textured soils versus factors below 1.2 per 10
o
C for
fi
ner-textured
soils). We also demonstrate that the lower effect of temperature
on C storage in
fi
ne-textured soils was retained after accounting
for potentially confounding variation in precipitation, aridity
(actual minus potential evapotranspiration), plant productivity,
soil pH and cation exchange capacity (CEXC) (Fig.
b). While we
focus on the top 50 cm, due to the potential for vertical pro
fi
les of
soil C to be affected by temperature
, very similar results were
observed for the top 20 cm (Supplementary Fig. 2). In addition,
the negative relationship between clay content and the effect of
temperature on C storage was observed independently both above
and below 15
o
C (Fig.
b, c).
The lower effect of temperature on soil C storage in
fi
ne-
textured soils with greater stabilisation capacities was unexpected
given the evidence of the high-temperature sensitivity associated
with the decomposition of more protected SOM pools
–
.
However, the
fi
ndings from our global analysis are in agreement
with a recent Europe-wide synthesis
, which, by compiling data
from soil physical fractionation studies, demonstrated that
mineral-associated C stocks varied less with temperate than freer
particulate pools. Therefore, there is growing evidence that the
effect of temperature on soil C storage is higher in soils
containing a greater proportion of unprotected C.
In the literature, there are apparently contradictory conclusions
in terms of how C storage varies across
fi
ne-scale climate gradients,
in which variation in other factors has been minimised. However,
these may potentially be resolved by considering differences in the
likely extent of SOM stabilisation. For example, on poorly
weathered, relatively coarse-textured, silt loam soils in Alaska,
mineral soil C stocks declined strongly with temperature
. In
contrast, in Hawaiian forests growing on
fi
ne-textured soils with
high concentrations of Al and Fe oxides, very little change in soil C
storage was observed across a gradient of 5
o
C in MAT. This was
despite the fact that, in these Hawaiian forests, C storage in
unprotected pools on the forest
fl
oor was found to decline strongly
with temperature
. We suggest that differences in the extent of
physicochemical protection in the Alaskan versus Hawaiian soils
may explain the contrasting results. Thus, apparently contradictory
fi
ndings may be resolvable within a single framework in which the
relative effect of temperature on C storage in mineral soils declines
ARTICLE
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27101-1
2
NATURE COMMUNICATIONS
| (2021) 12:6713 | https://doi.org/10.1038/s41467-021-27101-1 | www.nature.com/naturecommunications
as the soil
’
s physicochemical stabilisation capacity, and the
proportion of C in protected pools, increase.
Overall, our analysis identi
fi
ed C stored in high-latitude soils
with limited capacities for stabilising organic matter as likely to be
most vulnerable to the impacts of climate change. Such stores,
therefore, may require particular attention given the high rates of
warming taking place in cooler regions. In contrast, the
particularly low effect of temperature on C storage in
fi
ne-
textured soils in warm climates suggests (Fig.
c) that the C stocks
in many tropical soils may be less vulnerable to climate warming.
While a soil warming study in a less weathered tropical soil
identi
fi
ed the potential for high rates of C release
, our results
are consistent with a recent large-scale analysis that concluded
that the temperature sensitivity of soil respiration is generally
lowest in tropical environments
. However, because C storage in
tropical soils has been shown to be potentially vulnerable to
drought
, it should not be concluded that C storage in tropical
soils will be unaffected by climate change. Our results do, though,
suggest that C stocks in coarse-textured soils at high latitudes are
likely to be especially vulnerable to warming (Fig.
b). Finally,
while the dataset contains soil pro
fi
le information for sites across
the full mean annual temperature range investigated (0
–
30
o
C),
and there were data on a minimum of 500 pro
fi
les in every 5
o
C
temperature increment, increasing the amount of data available
for sites with mean annual temperatures below 5
o
C and greater
than 20
o
C would add further con
fi
dence to the
fi
ndings.
Because of their greater stabilisation capacities,
fi
ne-textured soils
store more soil organic matter
. Therefore,
fi
ne and coarse-
textured soils could contain similar absolute quantities of highly
vulnerable C, and the lower effect of temperature in
fi
ne-textured
soils could re
fl
ect the presence of greater quantities of low-
vulnerability organic matter
. Therefore, it is likely to be very
important to quantify stocks of unprotected pools, such as free
particulate C, in soils with contrasting stabilisation capacities, and to
investigate how such stocks vary with climate
. This may make it
possible to identify whether there are still important stocks of
unprotected organic matter that are vulnerable to climate warming
in
fi
ne-textured soils with high stabilisation capacities
Predicting and modelling future rates of C release
. Accurately
predicting the response of soil C storage to global warming
remains a major challenge. While spatial datasets, such as the
ones analysed in this paper, add con
fi
dence to the prediction that
C will be lost overall and help identify the most vulnerable stocks,
they provide limited information on the likely rates or dynamics
Fig. 1 Overall effect of temperature on carbon storage.
The location of the soil pro
fi
les (
a
) and observed overall relationships between C storage in the top
50 cm of mineral soil and mean annual temperature (
b
). The sampling locations (
a
) are colour-coded by mean annual temperature. The soil C stock data
(
b
) have been natural log-transformed and a linear regression
fi
t.
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27101-1
ARTICLE
NATURE COMMUNICATIONS
| (2021) 12:6713 | https://doi.org/10.1038/s41467-021-27101-1 | www.nature.com/naturecommunications
Fig. 2 Texture effects on temperature
–
soil carbon storage relationships.
The effect of texture on the relationships between C storage in the top 50 cm of
mineral soil and mean annual temperature in the raw data (
a
), and after accounting for potential confounding variables (
b
). The y-axes display the
proportional reduction in C storage for each 10
o
C increase in mean annual temperature, with higher values indicating greater reductions in soil C with
temperature. In panel
a
, the slopes of the relationships (solid line), together with their 95% con
fi
dence intervals (dark grey shaded area), are presented for
each of the textural categories, with the slope and 95% con
fi
dence interval for the full dataset (dotted line and light grey shaded areas) also presented
across the graph for comparison. In panel
b
, the relationship between soil C storage and temperature after accounting for variation in annual precipitation
(light blue), gross primary productivity (GPP; dark green), soil pH (purple), aridity (ET/PET; evapotranspiration minus potential evapotranspiration; navy
blue), and cation exchange capacity (CEXC, light green) are shown. The slopes of these relationships (solid lines) together with their 95% con
fi
dence
intervals (shaded area) are presented for each of the textural categories.
ARTICLE
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27101-1
4
NATURE COMMUNICATIONS
| (2021) 12:6713 | https://doi.org/10.1038/s41467-021-27101-1 | www.nature.com/naturecommunications
of C release. In this context, long-term surveys can be extremely
valuable. For example, a recent study in Chinese grasslands was
able to detect warming-induced soil C losses since the 1960s and,
consistent with the global analysis presented here, coarser-
textured soils lost far greater amounts of SOM
. Experimental
soil warming studies also offer opportunities for further deter-
mining the factors controlling soil C storage and predicting rates
of C release, although recent syntheses have produced con
fl
icting
overall
fi
,
. Revisiting the networks of warming studies
and considering the
fi
ndings in the context of soil stabilisation
capacities and changes in pools of protected and unprotected
SOM may allow for a greater understanding of the observed
patterns. For example, C losses from subsoils in response to 5
years of whole pro
fi
le warming were shown to be dominated by
the free particulate C pool
. Therefore, understanding the
responses of different pools to warming may offer the potential to
generate mechanistic understanding, even where changes in total
C storage have been dif
fi
cult to identify. It should though be
recognised that there are major challenges in accurately quanti-
fying relatively short-term changes in soil C stocks, and there are
many other variables that differ between soil warming studies,
including contrasting changes in plant productivity and rates of C
input driven by interactions between C and nutrient cycling
For these reasons, it may not always be possible to determine the
role of soil stabilisation capacities in controlling soil C storage
responses to experimental warming
, and observations collected
across space and time will likely remain important for con-
textualising experimental results.
Soil texture is included as a factor modifying decomposition
rates in the terrestrial C cycle modules of a number of Earth
systems models (ESMs), including the United Kingdom ESM
(UKESM), whose land surface scheme (the Joint UK Land
Environment Simulator (JULES)
) is based around the
Rothamsted C model
. Therefore, we investigated whether
JULES was already able to represent the patterns that we had
observed in the empirical data. In direct contrast to the empirical
data, JULES predicted very little variation in soil C with
temperature in cooler regions (below 15
o
C; Fig.
b), but
predicted a strong effect of temperature on C storage across all
textural classes above 15
o
C (Fig.
c). The pattern across the full
dataset was confounded by the model simulating only a small
number of
fi
ne-textured soils at high latitudes (Fig.
a), and the
fact that the relationship between temperature and soil C storage
differed so strongly above and below 15
o
C. However, crucially,
JULES failed to reproduce the greater effect of temperature on C
storage in coarse-textured soils and overestimated the effect of
temperature on C storage in
fi
ne-textured soils in warmer regions.
These
fi
ndings question whether JULES is identifying accurately
which soil C stocks are most vulnerable to the effects of climate
warming. This is important given the considerable geographical
variation in (1) rates of climate warming and (2) the amounts of
C stored in mineral soil horizons. In recent years, there have been
major efforts made towards developing models that include
physicochemical stabilisation mechanisms and yet can potentially
be run at the global scale
–
. Testing whether such models can
better simulate the observed spatial patterns of C storage in soils
with contrasting stabilisation capacities would increase con
fi
-
dence in projections of future changes in soil C stocks
.
Limitations and future perspectives
. As well as in
fl
uencing rates
of key biological processes, climatic variables also control pedo-
genesis, rates of mineral weathering and therefore in
fl
uence the
reactivity of soil surfaces
. Directly determining the binding
af
fi
nity of mineral surfaces is challenging and would require
detailed information on the type of clay minerals present, as well as
the abundance of key metal oxides
, but there is, currently,
insuf
fi
cient data to assess these more detailed variables at the global
scale
. However, it has been argued that, at broad spatial scales, soil
pH may explain an important proportion of variation in mineral-
binding af
fi
nities
. Furthermore, cation exchange capacity
(CEXC) varies with the type of clay minerals present and the
binding ef
fi
ciencies of the mineral surfaces
. In global analyses,
texture, pH and CEXC tend to be the three edaphic factors that
correlate most strongly with soil C storage
. For these reasons,
we also accounted for variation in both pH and CEXC, and eval-
uated whether the relationship between soil texture and the effect of
temperature on C storage was retained. We found that it was
(Fig.
b). Thus, we conclude, that within this large dataset, clay
content remains a strong predictor of soil stabilisation capacities,
both overall, and after accounting for factors that potentially control
SOM binding af
fi
nities.
While we consider that our analysis of how SOM stabilisation
capacities determine the effects of temperature on soil C storage is
robust, it is also high level. Thus, there is considerable opportunity
to further investigate different vulnerabilities of speci
fi
c pools of
SOM, contrasting the roles of mineral protection versus occlusion
in aggregates
, determining the importance of SOM binding
Fig. 3 Comparison between soil pro
fi
le data and JULES model output.
The effect of texture on the relationships between C storage in the top 50 cm of
mineral soil in the empirical data (solid lines) and JULES output (dashed lines). The slopes of these relationships (solid lines) together with their 95%
con
fi
dence intervals (shaded area) are presented for each of the textural categories. Results for the full mean annual temperate range (
a
), as well as for
subsets of the data for sites with mean annual temperatures below 15
o
C (
b
, blue) and above 15
o
C (
c
, red) are shown.
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27101-1
ARTICLE
NATURE COMMUNICATIONS
| (2021) 12:6713 | https://doi.org/10.1038/s41467-021-27101-1 | www.nature.com/naturecommunications
af
fi
nities
, and linking protection mechanisms with the sources of
the organic matter (e.g. microbial versus plant derived
). A
debate has often revolved around whether climatic versus edaphic
factors are more important in controlling patterns of soil C
storage. Rather, than focusing on which is more important, for
predicting future rates of soil C release, we suggest that a key
priority should be on identifying how key edaphic factors
determine the vulnerability of contrasting soil C stocks to climate
warming. In this context, a recent meta-analysis demonstrated the
importance of soil properties in controlling the temperature
sensitivity of soil respiration, emphasising how responses to global
warming will likely vary substantially between different types of
soils in contrasting geoclimatic zones
.
Using a large global database, we observed declining C storage
with temperature in mineral soils, suggesting that there is the
potential for strong positive feedback to climate warming. Critically,
however, this overall relationship masked differences between soils
with contrasting C stabilisation capacities, as indicated by their
textural properties. The data suggest that there are stabilised pools
of SOM in
fi
ne-textured soils that may be relatively insensitive to
the impacts of climate change, but that unprotected pools may be
substantially more vulnerable to climate warming than currently
predicted. Finally, because at least one major ESM was unable to
re
fl
ect the observed patterns, we argue that ESMs should be
evaluated against their ability to simulate the differences in the
effects of temperature on C storage in soils with contrasting textural
properties in order to reduce uncertainties in projections of the
effect of climate change on future soil C storage.
Methods
Data sources and processing
. Soil data were obtained from the World Soil
Information (WoSIS) database
with the 2016 snapshot
being downloaded and
formatted for analysis in MATLAB (2018a, The MathWorks, Inc., Natick, Mas-
sachusetts, United States). About 14,517 pro
fi
les contained the necessary infor-
mation on soil texture, bulk density and C contents, as well as pH and CEXC. This
subset of pro
fi
les was re
fi
ned by removing potentially disturbed and managed soil
pro
fi
les; locations classi
fi
ed as cropland or urban areas in the MODIS IGBP
landcover MCD12Q1 dataset
were removed from the analysis, leaving a total of
9,326 pro
fi
les. Stocks in the top 20 or 50 cm of mineral soil were quanti
fi
ed by
fi
rst
removing surface organic horizons (de
fi
ned as having gravimetric C contents
greater than 20%
), and then using data in the next 20 or 50 cm. Pro
fi
les with
organic horizons below mineral horizons were discounted due to the increased
possibility of past disturbance. The stocks were calculated for each sampling layer
or horizon based on depth, C content and bulk density. These were then added
together, with the depth of the
fi
nal sampling layer being adjusted to ensure that
stocks represented the top 20 or 50 cm of mineral soil. Average clay contents were
calculated accounting for variation in bulk density with depth. For each sampling
location, long-term average climate information (mean annual temperature and
annual precipitation for 1970
–
2000) was obtained from the WorldClim version 2.0
(
. Soils from sites with a mean annual temperature
below 0
o
C were removed, as previous analyses have identi
fi
ed a stronger effect of
temperature on C storage in cold climates due to freeze-thaw dynamics
. The
AppEEARS (
https://lpdaac.usgs.gov/tools/appeears/
) tool was used to download
MODIS data corresponding to the pro
fi
le locations for the year 2014; annual gross
primary productivity (GPP) from the MOD17A3 product
, 8-day composite
evapotranspiration (ET) and potential evapotranspiration (PET) from the
MOD16A2 product
. Good quality 8-day composite data were used to produce
the mean annual ET/PET fraction. Pro
fi
les with missing MODIS data were
removed from the analysis.
Data analysis
. Data analysis was completed using MATLAB 2018a. Linear
regression was used to determine whether there was an overall relationship
between temperature and C storage in mineral soils, with the C stock data being
natural log-transformed to reduce the extent to which high C stock outliers drove
any relationship and to re
fl
ect the hypothesised roughly exponential relationship
between decomposition rates and temperature. To determine whether the rela-
tionship between soil C storage and the temperature was dependent on soil texture,
the different soil pro
fi
les were allocated to different soil texture bins based on the %
clay content, with each bin representing a 5% increment in percentage clay content
(e.g. 0
–
5, 5
–
10%). All data above a clay content of 55% was combined into one bin
to ensure that the high clay content bins were not characterised by small numbers
of observations. The slopes and 95% con
fi
dence intervals of the slopes for each
relationship were calculated. Differences between texture bins, or between different
texture bins and the overall relationship, were considered statistically signi
fi
cant
when the con
fi
dence intervals did not overlap.
To investigate whether the observed patterns could be caused by confounding
environmental variables, linear relationships between mineral soil C stocks (natural
log-transformed) and plant productivity (GPP), precipitation (annual) and aridity
ET-PET), cation exchange capacity and soil pH (with all calculations using H
+
concentrations) were calculated. The same analysis as outlined above was then run
on the residuals from these relationships. While all the patterns still held, it is
important to note that the aridity calculation already includes temperature in the
calculation of potential evapotranspiration which may explain the reduction in the
gradient of the slopes (Fig.
b and Supplementary Fig. 2b).
UKESM/JULES model runs
. UKESM output data was taken from simulations
provided for the sixth coupled model intercomparison project (CMIP6), which
represents the state of the art in Earth system modelling
. Soil carbon, air tem-
perature (2 m) and clay content data were provided via the Met Of
fi
ce (Andy
Wiltshire
pers. comm
). NPP was downloaded from the public data archive (
data.ceda.ac.uk/badc/cmip6/data/CMIP6/CMIP/MOHC/UKESM1-0-LL
, accessed
12/07/2019), where the other variables are also stored. Soil carbon was simulated
dynamically by JULES via a Rothamsted-C-type scheme, and NPP was simulated by
the dynamic vegetation model in JULES
. Clay content was input to the model from
an ancillary
fi
le. Air temperature was simulated by the coupled atmospheric model in
UKESM, with the UKESM CMIP6 con
fi
guration being used
. Any grid points with
annual mean NPP less than 5 × 10
−
9
kg C m
−
2
s
−
1
were
fi
ltered out as these covered
mostly deserts, which are generally not included or under-sampled in the soil carbon
database, and very low NPP leads to very low soil C in JULES that obscures any soil
texture effect. Model output was then analysed in the same way as the observations.
Data availability
All data used in this manuscript are fully open access and available. The soil data were
obtained from a published snapshot derived from the World Soil Information database
(
https://doi.org/10.17027/isric-wdcsoils.20160003
), the long-term climate data are
available in the WorldClim version 2.0 database (
), while the MODIS
primary productivity, evapotranspiration, and landcover data are available in the
MOD17A3 (
https://doi.org/10.5067/MODIS/MOD17A3.006
), MOD16A2 (
doi.org/10.5067/MODIS/MOD16A2.006
) databases respectively. The UKESM data from the sixth
coupled model intercomparison project (CMIP6) is available in the public data archive
(
https://data.ceda.ac.uk/badc/cmip6/data/CMIP6/CMIP/MOHC/UKESM1-0-LL
Received: 13 July 2021; Accepted: 3 November 2021;
References
1.
Yu, Z. C. Northern peatland carbon stocks and dynamics: a review.
Biogeosciences
9
, 4071
–
4085 (2012).
2.
García-Palacios, P. et al. Evidence for large microbial-mediated losses of soil
carbon under anthropogenic warming.
Nat. Rev. Earth Environ
.
doi.org/10.1038/s43017-021-00178-4
(2021).
3.
Jobbagy, E. G. & Jackson, R. B. The vertical distribution of soil organic carbon
and its relation to climate and vegetation.
Ecol. Appl.
10
, 423
–
436 (2000).
4.
Doetterl, S. et al. Soil carbon storage controlled by interactions between
geochemistry and climate.
Nat. Geosci.
8
, 780
–
783 (2015).
5.
Giardina, C. P., Litton, C. M., Crow, S. E. & Asner, G. P. Warming-related
increases in soil CO2 effux are explained by increased below-ground carbon
fl
ux.
Nat. Clim. Change
4
, 822
–
827 (2014).
6.
Gonzalez-Dominguez, B. et al. Temperature and moisture are minor drivers of
regional-scale soil organic carbon dynamics.
Sci. Rep.
9
, 6422 (2019).
7.
Schmidt, M. W. I. et al. Persistence of soil organic matter as an ecosystem
property.
Nature
478
, 49
–
56 (2011).
8.
Dungait, J. A. J., Hopkins, D. W., Gregory, A. S. & Whitmore, A. P. Soil
organic matter turnover is governed by accessibility not recalcitrance.
Glob.
Change Biol.
18
, 1781
–
1796 (2012).
9.
Smith, P. et al. A comparison of the performance of nine soil organic matter
models using datasets from seven long-term experiments.
Geoderma
81
,
153
–
225 (1997).
10. Wiesmeier, M. et al. Carbon sequestration potential of soils in southeast
Germany derived from stable soil organic carbon saturation.
Glob. Change
Biol.
20
, 653
–
665 (2014).
11. Thornley, J. H. M. & Cannell, M. G. R. Soil carbon storage response to
temperature: an hypothesis.
Ann. Bot.
87
, 591
–
598 (2001).
12. Conant, R. T. et al. Temperature and soil organic matter decomposition rates -
synthesis of current knowledge and a way forward.
Glob. Change Biol.
17
,
3392
–
3404 (2011).
ARTICLE
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27101-1
6
NATURE COMMUNICATIONS
| (2021) 12:6713 | https://doi.org/10.1038/s41467-021-27101-1 | www.nature.com/naturecommunications
13. Jia, Y. et al. Temperature sensitivity of decomposition of soil organic matter
fractions increases with their turnover time.
Land Degrad. Dev.
31
, 632
–
645 (2019).
14. Lefevre, R. et al. Higher temperature sensitivity for stable than for labile soil
organic carbon - Evidence from incubations of long- term bare fallow soils.
Glob. Change Biol.
20
, 633
–
640 (2014).
15. Zhou, X. H., Xu, X., Zhou, G. Y. & Luo, Y. Q. Temperature sensitivity of soil
organic carbon decomposition increased with mean carbon residence time:
Field incubation and data assimilation.
Glob. Change Biol.
24
, 810
–
822 (2018).
16. Ribeiro, E., Batjes, N. H. & van Oostrum, A. J. M.
World Soil Information
Service (WoSIS) - Towards the Standardization and Harminization of World
Soil Data. Procedures Manual 2018, Report 2018/01
(ISRIC - World Soil
Information, 2018).
17. Batjes, N. H. et al. WoSIS: providing standardised soil pro
fi
le data for the
world.
Earth Syst. Sci. Data
9
, 1
–
14 (2017).
18. Jackson, R. B. et al. The ecology of soil carbon: pools, vulnerability and biotic
and abiotic controls.
Annu. Rev. Ecol. Evol. Syst.
48
, 419
–
445 (2017).
19. Hassink, J. The capacity of soils to preserve organic C and N by their
association with clay and silt particles.
Plant Soil
191
, 77
–
87 (1997).
20. Luo, Z. K., Viscarra-Rossel, R. A. & Qian, T. Similar importance of edaphic
and climatic factors for controlling soil organic carbon stocks of the world.
Biogeosciences
18
, 2063
–
2073 (2021).
21. Six, J., Conant, R. T., Paul, E. A. & Paustian, K. Stabilization mechanisms of
soil organic matter: Implications for C-saturation of soils.
Plant Soil
241
,
155
–
176 (2002).
22. Luo, Z., Luo, Y., Wang, G., Xia, J. & Peng, C. Warming-induced global soil
carbon loss attenuated by downward carbon movement.
Glob. Change Biol.
26
, 7242
–
7254 (2020).
23. Lugato, E., Lavallee, J. M., Haddix, M. L., Panagos, P. & Cotrufo, M. F.
Different climate sensitivity of particulate and mineral-associated soil organic
matter.
Nat. Geosci.
14
, 295
–
300 (2021).
24. Kane, E. S. & Vogel, J. G. Patterns of total ecosystem carbon storage with
changes in soil temperature in Boreal Black Spruce Forests.
Ecosystems
12
,
322
–
335 (2009).
25. Nottingham, A. T., Meir, P., Velasquez, E. & Turner, B. L. Soil carbon loss by
experimental warming in a tropical forest.
Nature
584
, 234
–
237 (2020).
26. Haaf, D., Six, J. & Doetterl, S. Global patterns of geo-ecological controls on the
response of soil respiration to warming.
Nat. Clim. Change
11
, 623
–
627
(2021).
27. Kramer, M. G. & Chadwick, O. A. Climate-driven thresholds in reactive
mineral retention of soil carbon at the global scale.
Nat. Clim. Change
8
,
1104
–
1108 (2018).
28. Xin, X. et al. Climate change dominated long-term soil carbon losses of Inner
Mongolian grasslands.
Glob. Biogeochem. Cycles
(2020).
29. Crowther, T. W. et al. Quantifying global soil carbon losses in response to
warming.
Nature
540
, 104
–
108 (2016).
30. van Gestel, N. et al. Predicting soil carbon loss with warming reply.
Nature
554
, E7
–
E8 (2018).
31. Soong, J. L. et al. Five years of whole-soil warming led to loss of subsoil carbon
stocks and increased CO2 ef
fl
ux.
Sci. Adv.
7
, eabd1343 (2021).
32. Zaehle, S. et al. Carbon and nitrogen cycle dynamics in the O-CN land surface
model: 2. Role of the nitrogen cycle in the historical terrestrial carbon balance.
Glob. Biogeochem. Cycles
24
, GB1005 (2010).
33. Clark, D. B. et al. The Joint UK Land Environment Simulator (JULES), model
description - Part 2: Carbon
fl
uxes and vegetation dynamics.
Geosci. Model
Dev.
4
, 701
–
722 (2011).
34. Jenkinson, D. S. The turnover of organic-carbon and nitrogen in soil.
Philos.
Trans. R. Soc. B
329
, 361
–
368 (1990).
35. Robertson, A. D. et al. Unifying soil organic matter formation and persistence
frameworks: the MEMS model.
Biogeosciences
16
, 1225
–
1248 (2019).
36. Abramoff, R. et al. The Millennial model: in search of measurable pools and
transformations for modeling soil carbon in the new century.
Biogeochemistry
137
, 51
–
71 (2018).
37. Wieder, W. R., Grandy, A. S., Kallenbach, C. M. & Bonan, G. B. Integrating
microbial physiology and physio-chemical principles in soils with the
MIcrobial-MIneral Carbon Stabilization (MIMICS) model.
Biogeosciences
11
,
3899
–
3917 (2014).
38. Zhang, H. C. et al. Microbial dynamics and soil physicochemical properties
explain large-scale variations in soil organic carbon.
Glob. Change Biol.
26
,
2668
–
2685 (2020).
39. Doetterl, S. et al. Links among warming, carbon and microbial dynamics
mediated by soil mineral weathering.
Nat. Geosci.
11
, 589
–
593 (2018).
40. Rasmussen, C. et al. Beyond clay: towards an improved set of variables for
predicting soil organic matter content.
Biogeochemistry
137
, 297
–
306 (2018).
41. Mayes, M. A., Heal, K. R., Brandt, C. C., Phillips, J. R. & Jardine, P. M.
Relation between soil order and sorption of dissolved organic carbon in
temperate subsoils.
Soil Sci. Soc. Am. J.
76
, 1027
–
1037 (2012).
42. Quesada, C. A. et al. Variations in soil chemical and physical properties explain
basin-wide Amazon forest soil carbon concentrations.
SOIL
6
, 53
–
88 (2020).
43. Angst, G., Mueller, K. E., Nierop, K. G. J. & Simpson, M. J. Plant- or
microbial-derived? A review on the molecular composition of stabilized soil
organic matter.
Soil Biol. Biochem.
156
, 108189 (2021).
44. Sulla-Menashe, D. & Friedl, M. User Guide to Collection 6 MODIS Land
Cover (MCD12Q1 and MCD12C1) Product. (2018).
45. IUSS Working Group WRB.
World Reference Base for Soil Resources 2014,
Update 2015: International Soil Classi
fi
cation System for Naming Soils and
Creating Legends for Soil Maps
. (FAO, 2015).
46. Fick, S. E. & Hijmans, R. J. WorldClim 2: new 1-km spatial resolution climate
surfaces for global land areas.
Int. J. Climatol.
37
, 4302
–
4315 (2017).
47. Koven, C. D., Hugelius, G., Lawrence, D. M. & Wieder, W. R. Higher
climatological temperature sensitivity of soil carbon in cold than warm
climates.
Nat. Clim. Change
7
, 817
–
822 (2017).
48. Running, S., Mu, Q., Zhao M. MOD17A3 MODIS/Terra Gross Primary
Production Yearly L4 Global 1km SIN Grid V055 [Data set]. (NASA EOSDIS
Land Processes DAAC, 2015).
49. Running, S., Mu, Q., Zhao M. MOD16A2 MODIS/Terra Net
Evapotranspiration 8-Day L4 Global 500m SIN Grid V006. (NASA EOSDIS
Land Processes DAAC, 2017).
50. Eyring, V. et al. Overview of the coupled model intercomparison project phase
6 (CMIP6) experimental design and organization.
Geosci. Model Dev.
9
,
1937
–
1958 (2016).
51. Sellar, A. A. et al. UKESM1: description and evaluation of the UK Earth
system model.
J. Adv. Model. Earth Sy.
11
, 4513
–
4558 (2019).
Acknowledgements
We would like to acknowledge the International Soil Reference and Information Centre
(
) for compiling and providing access to the WoSIS dataset. S.E.C.
acknowledges funding from Natural Environment Research Council (grant no. NE/
R015791/1). G.H. acknowledges funding from the Swedish Research Council (2018-
04516) and the EU H2020 project Nunataryuk (773421).
Author contributions
I.P.H. designed the study in discussion with all coauthors. I.P.H. and T.C.H. completed
the main analysis of the soils data following an initial analysis by G.H. S.E.C. carried out
the analysis of the JULES outputs. The manuscript was drafted by I.P.H. and all authors
contributed to the
fi
nal version.
Competing interests
The authors declare no competing interests.
Additional information
Supplementary information
The online version contains supplementary material
available at
https://doi.org/10.1038/s41467-021-27101-1
.
Correspondence
and requests for materials should be addressed to Iain P. Hartley.
Peer review information
Nature Communications
thanks Yiqi Luo and the other
anonymous reviewers for their contribution to the peer review of this work. Peer reviewer
reports are available.
Reprints and permission information
is available at
http://www.nature.com/reprints
Publisher
’
s note
Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional af
fi
liations.
Open Access
This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative
Commons license, and indicate if changes were made. The images or other third party
material in this article are included in the article
’
s Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not included in the
article
’
s Creative Commons license and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this license, visit
.
© The Author(s) 2021
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-27101-1
ARTICLE
NATURE COMMUNICATIONS
| (2021) 12:6713 | https://doi.org/10.1038/s41467-021-27101-1 | www.nature.com/naturecommunications