Annual grass invasions and wildfire deplete ecosystem carbon storage by >50% to resistant base levels

Abstract

Ecological disturbance can affect carbon storage and stability and is a key consideration for managing lands to preserve or increase ecosystem carbon to ameliorate the global greenhouse gas problem. Dryland soils are massive carbon reservoirs that are increasingly impacted by species invasions and altered fire regimes, including the exotic-grass-fire cycle in the extensive sagebrush steppe of North America. Direct measurement of total carbon in 1174 samples from landscapes of this region that differed in invasion and wildfire history revealed that their impacts depleted soil carbon by 42–49%, primarily in deep horizons, which could amount to 17.1–20.0 Tg carbon lost across the ~400,000 ha affected annually. Disturbance effects on soil carbon stocks were not synergistic, suggesting that soil carbon was lowered to a floor—i.e. a resistant base-level—beneath which further loss was unlikely. Restoration and maintenance of resilient dryland shrublands/rangelands could stabilize soil carbon at magnitudes relevant to the global carbon cycle.

Full text

communications earth & environment Article
https://doi.org/10.1038/s43247-024-01795-9
Annual grass invasions and wildfire
deplete ecosystem carbon storage by
>50% to resistant base levels
Check for updates
Toby M. Maxwell 1,2,HaroldE.Quicke 3, Samuel J. Price 2& Matthew J. Germino 2
Ecological disturbance can affect carbon storage and stability and is a key consideration for managing
lands to preserve or increase ecosystem carbon to ameliorate the global greenhouse gas problem.
Dryland soils are massive carbon reservoirs that are increasingly impacted by species invasions and
altered fire regimes, including the exotic-grass-fire cycle in the extensive sagebrush steppe of North
America. Direct measurement of total carbon in 1174 samples from landscapes of this region that
differed in invasion and wildfire history revealed that their impacts depleted soil carbon by 42–49%,
primarily in deep horizons, which could amount to 17.1–20.0 Tg carbon lost across the ~400,000 ha
affected annually. Disturbance effects on soil carbon stocks were not synergistic, suggesting that soil
carbon was lowered to a floor—i.e. a resistant base-level—beneath which further loss was unlikely.
Restoration and maintenance of resilient dryland shrublands/rangelands could stabilize soil carbon at
magnitudes relevant to the global carbon cycle.
A key strategy for managing earth’s carbon may be increasing sequestration
of carbon into soil, but avoiding losses of ecosystem carbon to the atmo-
sphere may be equally important. Conserving current soil-carbon stocks,
i.e., the amount of carbon stored per unit land area, should contribute
meaningfully to enhancing carbon sequestration where management or
conservation efforts affect vast land areas. One such region is the Inter-
mountain Western U.S., which is vulnerable to carbon loss due to wide-
spread impacts of disturbances such as wildfires and exotic plant invasions,
both of which are exacerbated by climat e change1–4.Further loss of carbon to
the atmosphere could feed back to intensify these disturbances, causing the
region to shift from a net carbon sink to a source by 20505,6.
Cold desert-shrublands, and specifically sagebrush steppe, are thought
to have the greatest relative potential to gain carbon stock among ecoregions
in the Intermountain Western U.S5. Carbon-loss threats to sagebrush steppe
are substantial and result specifically from exotic annual grass (EAG)
invasions and the increased wildfire, which they benefit from and promote
(i.e., the “grass-fire cycle”)1,4,7,8.Thenetimpactofthefeedbackisthecon-
version of diverse, deep-rooted perennial plant communities to shallow-
rooted annual grasslands, which has impacted >50,000,000 ha already and is
occurring at a rate of ~400,000 ha annually9. The exotic annual grasses
include species such as cheatgrass (Bromus tectorum L) that have short and
early growing seasons of weeks to a few months, that otherwise leave fire-
prone mats of senesced, dry, and fine-textured foliar litter1. This contrasts
with the potential for nearly year-round photosynthesis from native
evergreen perennials such as sagebrush, and a correspondingly more
positive annual ratio of ecosystem photosynthesis:respiration10.Impor-
tantly, while the grass-fire cycle is best studied in the arid and semi-arid
Western US, the introduction of exotic grasses to perennial ecosystems and
subsequent conversion to grasslands is globally relevant, impacting fire
behavior in Central and South America, Hawaii, and Australia, among other
places, with unknown consequences to carbon stocks1,7.
Decreases in above- and belowground productivity and biomass
resulting from plant-community shifts from perennial woody and her-
baceous communities to annual grasslands should impact soil carbon
stocks11–13.Wildfire, which is promoted by EAG invasion, is also expected
to locally degrade soil carbon stocks both directly by volatilizing biomass
carbon and indirectly by eliminating deep-rooted woody perennials and
through the loss of carbon-rich topsoil by erosion4,8,14. Detecting changes
in soil carbon in response to disturbance or land management actions is
challenging because there is substantial vertical, horizontal, and temporal
heterogeneity in dry shrubland plant community structure, and soil
water and biogeochemical cycling that must be considered12,15–17. Further,
the effects of disturbance or management on soil carbon are not likely to
be uniform, whether due to heterogeneity in disturbance severity or
because of variation in the capacity of soils to store and stabilize
carbon18–24. Conservation, management, or sequestration of soil
carbon in dry shrublands can be improved with an understanding
of how carbon and its sensitivity to disturbance are distributed across
1Boise State University, Department of Biology, Boise, ID, 83725, USA. 2U.S. Geological Survey, Forest, Rangeland, and Ecosystem Science Center, Boise, ID,
83702, USA. 3Environmental Science U.S. LLC, Cary, NC, 27513, USA. e-mail: mgermino@usgs.gov
Communications Earth & Environment | (2024) 5:669 1
1234567890():,;
1234567890():,;
Content courtesy of Springer Nature, terms of use apply. Rights reserved

the mosaic of plant communities and soil types that dominate these
ecosystems4,10,23,25–27.
A recent review revealed that neither the direct nor the synergistic
effects of wildfire and EAG invasions on soil carbon stocks are well
understood because the two disturbances often co-occur, and their separate
effects are not trivial to ident ify4.Studies on exotic annual grass effects on soil
carbon have not together considered the effects of wildfire, the strong lateral
heterogeneity—i.e. plant-soil microsites—that typifies the dominant plant
community types in sagebrush steppe, the duration of invasion, and the soil
depthswhereimpactsarelikelyevident
4,13. To make more accurate assess-
ments of the distribution of carbon stocks and their disturbance responses,
we designed and implemented a benchmark sampling plan that addressed
(1) the statistical power needed to detect changes in soil carbon across
variable landscapes28, (2) variation in plant-community states and the dif-
ferent soil microsites within them4(Fig. 1), and (3) both the commonly
sampled shallow soils ( < 40 cm) as well horizons up to 1 m depth that are
rarely sampled13,29. Our sampling regime also removed sampling biases that
result from non-systematic errors in the measurement of bulk density and
corresponding up-scaling error30 or incomplete removal of fine roots that
can be misconstrued as shallow-soil carbon under annual grasses.
Specifically, we assessed the impact of EAGs and wildfire on soil carbon
stocks from 0–100 cm depth in three cold-desert ecoregions in the vast
sagebrush-steppe landscape of the Western U.S.: the Northern Great Basin,
the Snake River Plain, and the Idaho Batholith. We sampled sites with all
combinations of burned/unburned and invaded/uninvaded disturbance
history and accounted for spatial heterogeneity within and between plant-
community types by sampling soils and weighting carbon stocks according
to the abundance of key, representative microsites (Fig. 1,methods: carbon
stocks). Our results revealed that wildfire and EAGs threaten soil carbon
stocks, stabilization, and sequestration in the vast sagebrush-steppe range-
lands, which are a relatively well-studied model system for the vast cold-
desert shrub steppes of the world.
Results and discussion
Overview—impact of invasion and burning on carbon stocks
Landscapes that had been invaded or burned had 55% and 93% less
aboveground biomass, respectively, and approximately half the soil carbon
stocks of areas that were not burned and not invaded (Figs. 1and 2). Across
all ecoregions, 66% of the soil carbon occurred in ‘deep’soils—i.e., from
40 cm–1 m, which was the maximum depth sampled—despite greater
carbon concentrations near the surface (Fig. 3), which is common for
semi-arid soils where inorganic mineral carbon can be redistributed to- or
formed in deeper soils31,32.Carbonstocksfrom60–100 cm depth in
unburned or burned landscapes that were invaded were 73% or 43% less
than uninvaded shrublands or grasslands that were unburned and burned,
respectively (Fig. 3). The significant reductions in soil carbon below 60 cm in
invaded/unburned areas was surprising because invaded shrublands had
substantial (18–49%) canopy cover of deep-rooted native perennial shrubs
(Table S2) that should have conferred root and carbon inputs to deep
soils11,12. In addition to revealing the sensitivity of deep soil carbon to
invasion, our findings of reduced deep soil carbon under EAGs in both
burned and unburned areas indicate that EAG invasion impacts likely
involve alteration of both organic and inorganic carbon processes32.Also,
the similarity of carbon stocks across disturbed areas may indicate that there
is a limit to how much carbon can be lost from soils following disturbances
to the biotic regulators of carbon flow, which we propose as a “soil-carbon
floor hypothesis”. Despite this putative soil-carbon floor, the effects of
invasion and wildfire were substantial (Fig. 2). Our results implicate
20.0 Tg/year of carbon loss from the top meter of soil over the ~400,000 ha
of sagebrush shrublands are estimated to be degraded primarily by wildfire
and plant invasions annually, amounting to the loss of ~1% of all carbon
stored in Western U.S. cold desert soils, annually (Fig. 2)5,9. Maintaining
intact sagebrush-steppe by protecting against the annual grass-fire cycle
thus appears to be a highly effective “nature-based”solution to the global
greenhouse gas problem.
Plant community and biomass relationships to invasion, fire, and
soil carbon
Areas that were burned and/or invaded had distinct plant communities
in terms of growth habit, lifespan, species richness, and root architecture
(Fig. 1and Table S2), which we expected, but did not observe, to be pro-
portional to differences in soil carbon storage (Fig. 2and Table 1). This result
implies that interspecific competition and plant community functional
diversity were not a direct factor affecting the observed soil carbon loss, as
might be expected33, but instead that indirect and/or abiotic effects, that were
associated with the plant community shift may explain the lesser soil carbon
Fig. 1 | Photographs of representative sites, the
primary microsites comprising each site, and
generalized predictions for relative differences in
carbon stock above ground and by depth in the
soil for each microsite. Over all soil depths, the
relative ranking of carbon stock by microsite was
predicted to be Shrub>PBG > EAG > BS, where PBG
is perennial bunchgrass, EAG is an exotic annual
grass, and BS is bare soil. Photo credit: Toby
Maxwell.
https://doi.org/10.1038/s43247-024-01795-9 Article
Communications Earth & Environment | (2024) 5:669 2
Content courtesy of Springer Nature, terms of use apply. Rights reserved

in burned or invaded areas. Burning reduced biomass by 93%, but post-fire
invasion increased aboveground biomass of burned areas by 22% (Tukey
HSD, p< 0.05, Fig. 2A). Invaded plots also had greater live herbaceous
biomass (by 2.5- and 1.5-fold) and herbaceous litter (by 1.1- and 2.0-fold) in
unburned and burned plots, respectively, compared to their uninvaded
analogs, which corresponded to greater foliar canopy cover and a lack of
bare soil (Fig. 1and S2; Tables 1 and 2). In unburned areas, differences in
biomass were mainly due to 2.5-fold greater shrub biomass in uninvaded
compared to invaded sites (Fig. 2AandS2).
Mean soil carbon was enriched to 60 cm depth underneath the crowns of
perennials (although the enrichment was not significant, Tukey HSD, p> 0.05,
Fig. S3), similar to previous reports (Fig. 1)15. Somewhat surprisingly, the
greatest microsite-horizon for carbon stock was the deepest soils under bare soil
—i.e., only found in uninvaded sites—and the least carbon storage at
60–100 cm depth was under EAGs (Fig. S3). Possibly the greater amount of
carbon in bare-soil—compared to EAG microsites was due to differences in
acidic root exudates between the microsites that could have impacted inorganic
carbon formation and permanence, which could be evaluated with measure-
ments of root density in relation to organic and inorganic carbon.
Deep soil carbon in drylands should be dominated by inorganic car-
bon, which is redistributed within soil profiles by water31,32.Thus,thewell-
documented greater water-content of bare-soil microsites in dry-
shrublands34 could drive their greater-than-expected deep soil carbon sto-
rage (Fig. S3). Unsurprisingly, the loss of fertile-island microsites created
and maintained by shrubs is a common effect of wildfire, which leads to the
loss of soil-carbon hotspots and thus is a key mechanism by which carbon
storage is reduced with invasion10 (Fig. 1and S3). Where fertile-island
microsites persist after wildfire, they compound their benefittocarbon
Fig. 2 | Aboveground biomass, carbon content, and carbon stock in invaded or
uninvaded and burned or unburned areas. Brown, Yellow, light green and dark
green portions of the bars represent the contribution of bare soil, exotic annual grass
(EAG), perennial bunchgrass (PBG), and shrub biomass to the total aboveground
biomass stock (A). The carbon effects of annual grass invasion and wildfires are
shown in Cfor burning (dotted arrow), invasion (solid arrow), and burning with
invasion (dashed arrow). Bars represent average carbon content (B) or total carbon
stocks (C) for 0–100 cm depth for each landscape condition. Significant differences
(Tukey’s HSD test, p< 0.05) between bars within each panel are represented by
lowercase letters at the top of each bar where a shared letter indicated no significant
difference. Error bars represent standard error. Note the varying scale and units for
each panel.
Fig. 3 | Variation in carbon with soil depth among the plot types. Carbon content
(A,B) and carbon stocks (C,D), aggregated across ecoregions and microsites for
different depth ranges for burned (left) and unburned (right) as well as invaded
(brown, filled points) and uninvaded (green, empty points) plots +/−SE. Significant
differences are indicated where: p< 0.1, *p< 0.05, **p< 0.001, ***p< 0.0001 for
paired t-tests between invaded and uninvaded plots within a single depth fraction
and burned and invaded treatment combination. No symbol is displayed
for p> 0.10.
https://doi.org/10.1038/s43247-024-01795-9 Article
Communications Earth & Environment | (2024) 5:669 3
Content courtesy of Springer Nature, terms of use apply. Rights reserved

retention by facilitating establishment of deep-rooted perennials35,which
confers greater net ecosystem carbon gain compared to areas that are
invaded after fire10 (Fig. S3).
Our finding that invasion and wildfire led to greater differences in
carbon stocks at depth compared to the surface contradicted the prevailing
expectation that near-surface soil carbon is most sensitive to disturbance on
short time scales due to relatively faster cycling near the soil surface29,36,37
(Fig. 3and S3). However, this paradigm stems from a conceptual model of
carbon cycling developed for agriculture and forests29,33,37–39 that may not
apply well to drylands and other low-productivity ecosystems where ratios
of above:below-ground inputs are lower12,40,41. For example, 72% of biomass
is estimated to be underground in drylands, compared to 9% for croplands
or 19% for temperate evergreen forests12.Morethanhalfofthebelowground
biomass in deserts is below 20 cm depth, which helps to explain our finding
that deep soil carbon was sensitive to disturbances that altered the plant
community and reduced aboveground productivity12 (Fig. 3and S3).
Relatively little research has been done to evaluate the impact of dis-
turbances on deep soil carbon cycling. In one notable study in which shallow
and deep soil carbon cycling were assessed after a forest-to-grassland pas-
ture transition, gains in topsoil carbon were balanced by equivalent and
sometimes greater losses of deep soil carbon, indicating an active carbon
pool with ~10 year turnover in deep soils11. While deep soil carbon is
generally considered to be older and more stable than surface soil carbon36,
our data are consistent with the hypothesis that seasonal patterns of root
growth and shedding may lead to meaningful quantities of actively cycled
deep soil carbon in ecosystems where plant productivity is dependent on
deep-soil resources11,42,43. Our work adds to a small, but growing body of
evidence showing that disturbances that lead to shallower rooted plant
communities with shorter growing seasons are expected to affect deep soil
carbononshortertimescalesthanistypicallythought
11,41,42,44.Importantly,
in drylands, slow decomposition rates and the heterogeneous distribution of
rootsinthesoilmakedetectionandattributionofchangesindeepsoil
carbon particularly challenging4.
Evidence for a soil carbon ‘floor’
Burning and invasion led to somewhat similarreductionsinbothsoilcar-
bon stocks and content (Fig. 2B, C), despite their unique impacts on vege-
tation, litter, and other aspects of the ecosystem that affect carbon flow and
storage10. It is possible that most of the readily mineralizable, labile soil
carbon was lost from either burning or invasion, thereby decreasing soil
carbon stocks to a minimal and base “floor”level, at which remaining
carbon was strongly bound or chemically recalcitrant. By analogy, the
remaining carbon could be referred to as the soil-carbon “basement”.In
support of the soil-carbon floor hypothesis, a soil-carbon “ceiling”has been
recognized that helps explain why soil carbon does not always increase
following organic inputs19,36,45–47. Awareness of soil-carbon floors and ceil-
ings is important in evaluating or predicting stability of soil carbon, speci-
fically where or when it may be unresponsive to perturbation.
All soil resources exist in a gradient of forms or matrices that affect
reactivity and availability to transformation or transport, and thus a similar
trendshouldbeidentifiable for soil carbon. For example, soil water can be
freely available or tightly bound at the same gravimetric water content,
depending on the texture and structure of a soil48. Similarly, phosphorus
content in soil can be weakly related to its bioavailability because it is often
most abundant as phosphate or as organic phosphorous, both of which tend
to strongly sorb to soil minerals, rendering them relatively unavailable for
plant uptake49. Soil nitrogen can also be soluble, mobile, and bioavailable
when oxidized to nitrate, or plant-available yet relatively immobile when
reduced to ammonium, or relatively unavailable as organic nitrogen50.
The form of basement soil carbon could vary with the processes that
formed it51–53. Inorganic carbon tends to be 10–17-foldgreaterthanorganic
carbon stocks in dryland soils and, moreover, concentrated at ~30 cm depth
Table 1 | Mixed effects model results. Model structure and parameter results for carbon content (g/kg soil), carbon stocks, and
aboveground biomass for two different models of each parameter, varying in their fixed and random effects. Estimates are
reported relative to the intercept case for each model. Nesting of random effects is represented by parentheses, independent
sets of random effects are separated by a comma, and interactions are represented by a colon
Response Random Intercept (s) Parameter Estimate (SE) p≥|z |
Soil carbon content (g/kg)aSite (transect position (depth class)), microsite Intercept (burned, invaded) 2.02 (0.20) <0.0001
Uninvaded 0.22 (0.085) 0.0104
Unburned 0.22 (0.060) 0.0002
Burning:invasion interaction −0.21 (0.075) 0.0048
Soil carbon stocks (Mg/ha)aSite (transect position) Intercept (burned, invaded) 4.82 (0.26) <0.0001
Uninvaded −0.085 (0.10) 0.39
Unburned 0.76 (0.26) 0.0032
Aboveground biomass carbon −0.089 (0.028) 0.0017
Species richness −0.14 (0.037) 0.0002
Burning:invasion interaction 1.64 (0.37) <0.0001
Soil carbon stocks (Mg/ha)aSite (transect position), Burning: Invasion interaction Intercept 4.07 (0.23) <0.0001
Aboveground biomass carbon 0.021 (0.0056) 0.0002
Species richness −0.048 (0.034) 0.15
Aboveground biomass (Mg/ha)aSite (replicate ID), microsite Intercept (burned, invaded) 6.27 (0.86) <0.0001
Unburned −1.21 (0.29) <0.0001
Uninvaded 1.36 (0.48) 0.0042
Burning:invasion interaction 1.38 (0.39) 0.0004
Aboveground biomass (Mg/ha)aSite (replicate ID), Burning:invasion interaction Intercept (bare soil) 4.51 (0.60) <0.0001
EAG microsite 3.42 (0.41) <0.0001
Perennial bunchgrass microsite 1.43 (0.26) <0.0001
Shrub microsite 3.92 (0.26) <0.0001
aModel coefficients estimated using a Gaussian distribution of the log response.
https://doi.org/10.1038/s43247-024-01795-9 Article
Communications Earth & Environment | (2024) 5:669 4
Content courtesy of Springer Nature, terms of use apply. Rights reserved

or deeper31,54. Soil inorganic carbon tends to accumulate in dryland soils
because hydrologic events that leach and remove minerals out of the soil
profile are rare, leading to an accumulation of eluviated or precipitated
carbonate minerals at the depth of the wetting front31,55. Mineral carbon is
fully oxidized and does not function as an electron acceptor for microbial
respiration, or as a nutrient, in the way that mineral nitrogen can, for
example. Thus, while much of soil inorganic carbon is derived from pho-
tosynthesis, its fate and transport are mostly abiotically and particularly
hydraulically determined31,56. In drylands soil inorganic carbon may provide
a substantial and stable basement compared to ecosystems where soil car-
bon is predominantly organic and/or meteoric water is more abundant, and
soil properties conducive to redistribution of the carbon outside the eco-
system by leaching or respiration. However, while soil inorganic carbon is
often disregarded as recalcitrant, altered soil pH, hydrology or the cycling of
nitrogen that occurs with EAG invasions could conceivably and rapidly
affect soil inorganic carbon stocks that formed over long timeframes (1 K to
>1 M years)10,16,57,58.
Disturbance effects on soil carbon stability
Carbon tends to be more abundant and stable in finer-textured soils, and
therefore, more reliably stored, thus changes in soil texture due to invasion
or burning could impact carbon stability and storage59–62. Model estimates
from our data revealed the counter-intuitive result that for each percentage-
point reduction in fine particles (silt+clay content), soil carbon stocks would
be expected to be 1.0 Mg/ha greater (generalized linear mixed effects model;
Table S3). Thus, burning or invasion could have secondary effects that
impacted the observed carbon-texture relationship. EAGs can reduce soil
erosion, and because they can germinate in the first fall rains that follow a
fire, sites that were invaded could have avoided aeolian or alluvial erosion,
particularly in burned sites through the first growing season8,63,64.Alter-
natively, soil texture may simply be a confounding factor because EAG
invasion is facilitated by finer soil textures in semi-arid landscapes57,65,66.
Were it not for the finer soil texture in invaded, lower carbon, sites, we would
expect to have seen stronger negative effects of invasion on soil carbon
stocks because EAG selection effects should have a benefit of greater storage
and stabilization properties of silt and clay particles that cannot be conferred
by native vegetation (Fig. S5).
As our data revealed, biogeochemical measures of soil carbon stability
—i.e., physical protection by association with fine soil particles, or chemical
recalcitrance47,62—mayonlypartiallyexplainsoil carbon stability, particu-
larly where disturbances are frequent, overlapping, or widespread. For
example, biogeochemical carbon-stability metrics have complex responses
to wildfire that are poorly described in dry shrublands23 and an unknown
relationship to EAG invasion10. While the effect of disturbances on micro-
scale soil carbon stability is a topic that deserves more study, our data suggest
that two additional forms of carbon stabilization at the ecosystem scale that
should be considered. First, ecogeomorphic stability, which can be thought
of as the effect of plant-community composition and soil surface structure
on erosion and soil hydrology8,56,67. Aeolian erosion can amount to cen-
timeters of topsoil loss annually after fire, and could selectively affect soil
particles that store biogeochemically stable carbon—i.e. silt, clay, and
microaggregates8,62,68,69. Plant communities that rapidly regrow after wildfire
can resist this erosion, preventing substantial carbon losses from these
ecosystems8,63. Second, ecological stability refers to the resilience, or the
tendency of a plant community to return to its previous state following
disturbance70, and therefore the likelihood that plant-soil interactions that
affect soil carbon storage are stable over long periods of time (Fig. 4). For
example, sites that are not burned or invaded that effectively turn into low
diversity shrublands due to overgrazing and the depletion of native
resprouting perennials, as has occurred across much of the western U.S71,
may maintain a substantial carbon stock, but are also the most vulnerable to
carbon loss, because, when they burn, natural recovery of native perennials
is unlikely (Fig. 4)72. Lack of perennial recovery portends further loss in
carbon gain due to a positive shift in annual net ecosystem carbon fluxes (i.e.
to a source from a sink5), and the greater wind erosion losses expected from
shrub microsites after wildfire69. The effect of plantcommunitycomposition
on carbon stabilization can be direct—through altered magnitude, chem-
istry, and vertical distribution of carbon inputs11,53, or indirect by affecting
soil hydrology or nitrogen cycling in ways that could impact soil resource
availability and water infiltration56,57,73 and thus primary productivity.
Therefore, restoration and maintenance of native-perennial shrublands that
are resistant to EAG invasions and resilient to wildfire should confer stability
to soil carbon stocks and be more responsive to restoration to the greater-
carbon-storing shrubland state (Figs. 2and 4)35,74.
Improvements on inference
Here, to the best of our knowledge, we provide the first comprehensive,
robust sampling to evaluate the relationship of soil carbon to EAG invasion
and wildfire—extensive disturbances that are closely linked to millions of
Fig. 4 | Ball-and-cup diagram conceptualizing results on changes in soil carbon
resulting from plant community state transitions associated with exotic annual
grass invasion and wildfire. Relative soil carbon stocks (left axis) and ecological
stability of the carbon (right axis) are represented by the four “ground levels’of soil
on the y-axis for each ecosystem state (x-axis) in this study. The plant community for
each ecosystem state is represented within each cup (also see Fig. 1), as is the relative
aboveground biomass (intensity of blue coloring) and soil carbon with depth
(intensity of brown shading), where darker shading represents more carbon. Arrows
show perturbations that cause state transitions among plant communities to a new
state where wildfire is represented by shaded orange/red, invasion in black, and state
transitions to states that have more or less carbon in green (natural recovery likely)
or red (natural recovery unlikely), respectively. Red arrows represent the ecosystem
states where intervention/restoration would be most appropriate. The height of the
hill between states is proportional to the severity of a disturbance (or restoration
intervention) that would be required to shift the ecosystem from one state to another.
For example, a more intense restoration/disturbance would be required to stimulate
recovery to the greatest carbon state for the burned/invaded compared to the
unburned/invaded state. This original image was made for exclusive use by the
authors by Mason Otis.
https://doi.org/10.1038/s43247-024-01795-9 Article
Communications Earth & Environment | (2024) 5:669 5
Content courtesy of Springer Nature, terms of use apply. Rights reserved

acres of land management in dry shrublands3,4. Few studies have measured
soil carbon beyond 30 cm depth and even fewer accounted for heterogeneity
within and between plant-communities by sampling the relevant plant-soil
microsites that are common to dryland ecosystems4,13. Studies have also
lacked a statistically defensible sampling design that paired invaded and
uninvaded sites with adequate assurance about the longevity of the exotics
with and without their covarying burn effects4. Our results corroborated the
hypothesized negative overall effects of EAGs on soil carbon that were
previously developed in studies that compared invaded and uninvaded sites
between studies without consideration of the effect of soil type, climate, and
disturbances on soil carbon stocks4,13. In contrast, we did not observe greater
surface-soil carbon in invaded areas. The available related studies typically
did not describe in adequate detail whether their soil processing methods did
the onerous sifting that would have removed the fine roots of EAGs, and
thus may have had included coarse organics as soil carbon, making accurate
meta-analysis or inter-study comparison of carbon stocks challenging and
sometimes impossible4.
Soil carbon is generally considered to be greater in areas that receive
more moisture, especially along precipitation gradients in drylands where
primary productivity is moisture limited75. However, we found that wildfire
and EAG invasions disrupted the established positive climate-biomass-soil
carbon relationship (Fig. 2and Table 1)33,76. For example, we observed that
the Idaho Batholith had 40–49% greater precipitation and, accordingly, 28%
greater biomass in intact native perennial communities that had not been
burned or invaded compared to undisturbed areas of the Northern Basin
and Range or Snake River Plain. However, in disturbed areas of the Idaho
Batholith, biomass ranged from 16% to 42% less, and soil carbon was 62%
less than in areas that were similar in terms of burning and invasion in the
other drier ecoregions (Fig. 2; S2 and S4). Across all sites we found that
differences in species richness were not intuitive (i.e. negatively related to soil
carbon) and aboveground biomass was not proportional to changes in soil
carbon in response to disturbance. These observations were contrarian to
previous studies, and were due to the disruptive effects of EAG invasion and
wildfire (Table 1)33,39,76.
Modeled estimates of carbon stocks are often restricted to shallow
soils37,38,77,78. Our data confirmed observations from other studies that dis-
turbances can affect deep soils just as strongly as shallow soils (Figs. 2–4and
Table 1)11,36.Thus, while sampling deeper soils is more challenging, our data
show that models that do not incorporate deep soils may underestimate the
carbon losses due to disturbances, especially perennial-to-annual, or forest-
to-pasture conversions that may have a lesser net effect on the total carbon in
surface soils (Figs. 3Aand4)11,41. There are still significant barriers to
modeling soil carbon at high resolution for single time-points, even in the
top 30–40 cm of soil78,79, and little data is available to validate or para-
meterize deep soil carbon responses to disturbances13,36. Therefore, more
research describing how wildfire and plant community composition
interact to affect carbon cycling throughout the soil column would be
beneficial.
This research provides robust evidence for significant and substantial
effects of wildfire and plant invasions on soil carbon and shows that those
effects were contrary to theoretical expectations based on differences in
aboveground biomass, species richness, and soil texture between plots.
However, critical questions remain about the mechanisms driving differ-
ences in soil carbon. Specifically, more study is needed to improve under-
standing of how the different forms of carbon are affected by plant invasions
due to any differences in organic or inorganic C, rather than total carbon or,
further, due to any differences in particulate or mineral associated organic
carbon pools which have important implications for the long-term stability
of carbon47. Similarly, differences in fire frequency that occur in grasslands
could lead to growth of the pyrogenic carbon-pool which could take millenia
to accumulate80, but could eventually offset some of the short-term carbon
losses. The primary land use in this ecosystem is grazing, and thus it will also
be important to test whether there are any interactions between grazing,
wildfire, and plant invasions81. Additionally, to understand the empirical
mechanisms that govern changes in carbon is an important part of
improving the capacity of earth system models to represent our findings.
Specifically, it is not well understood if and whether there are differences in
soil hydrology, evapotranspiration, root architecture, or plant functional
type due to invasion or wildfire that may be impacting the flow and chemical
and biological processing of both organic and inorganic carbon. Finally,
state-of-the art ecosystem carbon models77 account for microbial commu-
nity structure and function, and among many microbial processes that
represent the processing of carbon in the soil, it could significantly improve
both empirical understanding and also modeling capabilities to know
whether the carbon-use efficiency of the microbial community is altered by
these disturbances.
Conclusions
By experimentally isolating the effects of wildfire and EAG invasion, we
showed that soil carbon stocks and stability in semi-arid landscapes are
threatened by EAG invasions and associated wildfire. Surprisingly, there
was no synergistic or even additive effect of invasion and wildfire on the loss
of soil carbon stocks, which indicated that soil carbon may have been
lowered to a ‘floor’, beneath which further carbon loss would be unlikely.
Biogeochemical and soil physical processes are often cited as the main
factors affecting soil carbon stability, but here we show that exotic plant
invasion and wildfire are ecological forces that strongly and rapidly desta-
bilize carbon accumulated over times scales ranging from decades to mil-
lenia, or more. While there are increasing calls to sequester carbon in soils as
a‘natural climate solution’, and it may be possible to increase carbon stocks
in dryland/shrub steppe soils through careful management, a related and
more achievable goal may be to invest in the ecological stabilization of
current soil carbon stocks by restoration and management of intact native
perennial communities that are resilient to wildfire and resistant to invasion
that are known to be net carbon sinks, and thus could at least partially offset
the potential loss of 20 Tg carbon annually in western North America alone.
Methods
Approach
We measured plant community composition and soil carbon stocks in
dominant microsite types within sites that were: (1) burned and invaded—
Annual grass dominated near monocultures of cheatgrass (Bromus tec-
torum L.), and in one case, medusahead (Taeniatherum caput-medusae), (2)
burned and uninvaded—perennial bunchgrass stands of Bluebunch
wheatgrass (Pseudoroegneria spicata), Siberian Wheatgrass (Agropyron
fragile),squirreltail(Elymus elymoides and Elymus multisetus), and/or Idaho
Fescue (Festuca idahoensis), (3) unburned and invaded—cheatgrass and big
sagebrush (Artemesia tridentata) and (4) unburned and uninvaded—intact
shrub-steppe dominated by big sagebrush often with perennial bunch-
grasses (Fig. 1and Table S2), in each of three cold-desert ecoregions: The
Snake River Plain, Northern Basin and Range, and the Idaho Batholith82.All
sites have been grazed to some extent over the last 100+years and are best
described as wildlands disturbed by pastoral grazing that created an inroad
for exotic annual grass invasion and associated wildfire (as described for the
region by Mack 1986).
Site classification as (un)burned or (un)invaded
Sites were defined as burned if there had been a wildfire 20–50 years prior to
sampling according to the historical wildfire maps from the U.S. National
Interagency Fire Center83.Asitewasdefined as unburned if it had not
burned for at least 50 years, which is the minimum recovery time for
Wyoming big sagebrush communities which have historical fire return
intervals of 171–342 years, while perennial grasslands are thought to burn
every 35–100 years84,85. Invasion history was determined by a combination
of on-the-ground knowledge from local Bureau of Land Management
agency staff and by using Rangeland Analysis Platform to assess the extent
and duration of EAG invasions. Specifically, maps of herbaceous annual
cover were obtained for the areas of interest for years 1998–2001 and
2018–2021 (rangeland analysis platform—RAP86). A plot was considered
invaded if it had >50% herbace ous annual cover for the 20 years prior to our
https://doi.org/10.1038/s43247-024-01795-9 Article
Communications Earth & Environment | (2024) 5:669 6
Content courtesy of Springer Nature, terms of use apply. Rights reserved

sampling and uninvaded if it had <20% herbaceous annual cover for 20
years. This was determined by a combination of communications with local
land managers and by quantifying maximum cover values of herbaceous
annuals for each plot from RAP data (see above) for years 1998–2001 (i.e., a
4-year window 20 years prior to sampling) and 2018–2021 (i.e., for the 4
years prior to sampling). At the time of sampling all burned or unburned
invaded plots had 100% or 51-82% canopy cover of EAGs, respectively
(Table S2). The accuracy of RAP products is variable at the ~1 km2scales of
our plots87,88. While our application of RAP was at a smaller scale than the
most accurate applications of RAP (12 km2), our interest was not to
quantitatively assess EAG cover at a specific time, but rather to classify
annual grass cover as either extremely high or low over multiple years, and
thus we believe our application of the data to be appropriate87,88.
Ecoregion, site, and plot descriptions
The Snake River Plain ecoregion is xeric with lacustrine and alluvial
parent materials deposited by Pleistocene Lake Idaho and the Snake
River. The natural potential vegetation is sagebrush steppe, with
greasewood-saltbrush in more arid subregions89. Unburned plots were
located at 870 meters above sea level at approximately 43.328312 N,
116.382486 W within the Birds of Prey National Conservation Area and
have no recorded burn history. The burned sites were located approxi-
mately 14 km NW of the unburned sites on Kuna Butte, a small volcanic
butte located on the Snake River Plain at approximately 43.448739 N,
116.462508 W. Burned plots were at anelevation of 935 m above sealevel
within the boundary of the 1957, 1983 and 1987 ‘Kuna Butte’fires and
the 1996 Coyote Butte fire. Burned-uninvaded plots were located within a
1983 drill seed that established the current stand of bluebunch wheat-
grass. All plots were flat, soils lacked restrictive horizons and had silt
loam texture. Mean annual temperature and precipitation were 11.0 °C
and 253 mm, respectively (Table S1). Long-term mean annual tem-
perature and precipitation data were extracted from 800 m gridded cli-
mate data provided by the PRISM climate group (Table S1)90.
The Northern Basin and Range ecoregion is characterized by dissected
lava plains, rolling hills, alluvial fans, valleys, and scattered mountains. It is
generallyhigherinelevationandcoolerthantheSnakeRiverPlain.The
natural potential vegetation is sagebrush grasslands or saltbush greasewood,
with cool season grasses. Higher elevations support mountain sagebrush
and juniper woodlands89. All plots were in the Owyhee Mountains South-
west of Boise, Idaho, USA and were all co-located at approximately
43.492878 N, 116.992677 W at 1115 meters above sea level. Unburned plots
were preserved ‘islands’of sagebrush shrubs that, while inside documented
fire perimeters, did not burn. Burned plots were burned in the 2001 Trimbly
creek fire and the 2015 Soda fire while burned uninvaded plots were located
within the bounds of a successful drill seed of Siberian Wheatgrass that
occurred immediately after the fire. All plots were flat and had a highly
developed restrictive horizon at approximately 25–30 cm depth and gravelly
loam soil texture. Mean annual temperature and precipitation were 10 °C
and 269 mm, respectively (Table S1)90.
The Idaho Batholith ecoregion is mountainous, deeply dissected,
underlain by granitic rocks, and characterized by limited fertility and water-
holding capacity. Much of the Idaho Batholith is montane, but lower ele-
vations support sagebrush shrublands and perennial grasslands89.Our
burned plots were established at approximately 43.55986 N, 116.07119 W at
1114 meters above sea level within the bounds of the 1957 Rocky Canyon
fire and the 2003 High Crow fire, and our unburned plots at 43.57403,
−116.06201 at 1278 meters above sea level, all within the Boise River
Wildlife Management Area. The last recorded burn at the ‘unburned’sites
was the 1959 Oberbillig fire, and all sites burned in the 1957 Rocky Canyon
fire. EAG establishment in the Boise Foothills subregion of the Idaho
Batholith is aspect dependent and therefore our invaded plots were located
on NW aspects and uninvaded plots on S/SE aspects. Plots ranged from
18–24° slopes and were loamy sand to sandy loam in texture. Long-term
average mean annual temperature and precipitation were 10.5 °C and
377 mm, respectively (Table S1)90.
Vegetation monitoring
Canopy cover was measured at each plot by line point intercept. A 50 m
transect was randomly established within the bounds of the appropriate
treatment combinations (of burning and invasion). At every 50 cm along the
transect, plant species that intercepted a pin flag were recorded at all levels of
the canopy for a total of 100 points. In a second pass of the same transect,
canopy gaps between identified microsite types that were present at the site
were determined and their length represented along the transect recorded
(Fig. 1and Table S2). Short statured perennial grasses such as Poa secunda
that occasionally occupied interspaces was not considered as a perennial for
the purposes of the canopy gap assessment. Microsite composition was
determined by continuous lengths of canopy cover for each microsite type
along the transect, and considered continuous when breaks in that microsite
type were <20 cm length. If an EAG or perennial bunchgrass was present
below a sagebrush canopy, that microsite was listed as sagebrush canopy.
Bunchgrass microsites were considered to extend beyond their basal area,
i.e., any location underneath the canopy of a bunchgrass was recorded as a
perennial bunchgrass microsite.
Soil sampling
At 5, 15, 25, 35, and 45 m locations along the transect, soils were sampled
from the closest example of each type of microsite found in the canopy gap
measurement described above. Soils were sampled using a slide hammer soil
core sampler (AMS 3”x6”steel soil core sampler, American Falls, ID) for
the top 40 cm of soil (2 cores) which were then separated into depth frac-
tions by hand. When possible, a gas-powered drill was then used to drill
deeper, and soils sampled from 5 cm depth increments within the 40–60 and
60–100 depth ranges. Soils could not always be sampled at the exact same
depth even within a transect due to the variable depth and presence of
impenetrable calcic hardpans, and thus analyses were done on ‘binned’
groups of soil samples within depth ranges that reflected the horizonation of
each soil (see below). To assess the bulk density of the soils, soil pits were dug
at 5, 25, and 45 m along the transect and soil cores extracted using thin
walled, 5 cm diameter steel cylinders to carefully re move a known volume of
soil from 0–5cm,5–10 cm, 10–15 cm, 20–40 cm, 40–60 cm and 60–100 cm
depth ranges. For soils down to 15 cm, 2 cylinders of approximately 60 cm3
volume were removed and combined as an aggregate sample. For samples
below 20 cm depth, 3 samples were taken evenly across the depth range and
aggregated. The number of samples taken was determined by applying an a
priori test of statistical power based on soil carbon concentrations and
variation from a study in sagebrush steppe at the Idaho National Lab on the
eastern Snake River Plain that had measured soil carbon concentrations
under perennial canopies and in bare soil microsite locations91.Thesam-
pling campaign resulted in 1174 soil samples. The number of replicates per
treatment combination can be found in the data release for this manuscript
published on the dryad digital repository.
Soil processing and analysis
Soilswereovendriedat60°Cfor48h,oruntiltheweightofthesampleshad
stabilized. Surface soils that had been stored in plastic sleeves used in the
slide hammer-soil corer were separated into 0–5, 5–10, and 10–15 cm
depths by careful removal from the cores with a spoon. Depth ranges were
adjusted for compaction when the depth of sampling did not match the core
height. Soils were sieved to 2 mm, coarse and fine roots removed, ground to
a powder, and analyzed for total carbon with a Costech ECS 4010 elemental
analyzer (Costech Analytical Technologies, Inc., Valencia, CA). A duplicate
sample was included every 10 samples, and an entire sample run of
~40 samples re-run if any duplicate weas >10% different, including for
internally validated soil standards.
Texture of mineral soil
Soil texture was measured in a representative subset of 793 samples using the
micro pipette method92. Approximately 5 g of soil was added to 40 mL 0.5%
(w/v) sodium hexametaphosphate and shaken overnight to separate
aggregate particles. After shaking, a 5 mL micropipette was used to extract
https://doi.org/10.1038/s43247-024-01795-9 Article
Communications Earth & Environment | (2024) 5:669 7
Content courtesy of Springer Nature, terms of use apply. Rights reserved

2.5 mL soil solution from 2.5 cm depth in the tubes at approximately 11 s
(sand particles had settled) and 1 h 46 min (sand and silt particles had
settled) after shaking was stopped. Settling times were calculated from
Stokes law assuming spherical soil particles of density 2.60 g/cm3settling at
24 °C (the measured lab temperature) and that sand particles were >50 µm
and clay <2 µm diameter.
Bulk density
Bulk density was measured using the pit method93—i.e., by digging soil pits
with flat vertical faces and extracting intact cores by inserting thin-walled
5 cm diamet er steel pipe that was cut to 5 cm lengths, excavating the cylinder
from the soil, and bringing the soil to the lab for measurement. Soil from
each core was dried at 105 °C for 48 h, or until the weight stabilized. Soil was
sieved to 2 mm and fine roots separated. Rock volume was measured by
displacement of water in graduated cylinders. Bulk density was calculated
according to the fine earth method30 where coarse fragment volume was
subtracted from core volume to calculate bulk density of the mineral soil
(<2 mm) fraction only.
Carbon stocks
Carbon stocks for each sample were calculated as a function of bulk density,
sample depth range, and carbon content on a per mass basis. (Eq. 1)
Carbon stock ¼soil carbon content gcarbon
gsoil

×bulk density gsoil
cm3soil

×sample depth cm
ðÞ 1Mg
1;000;000g
×100;000;000cm2
ha
ð1Þ
Carbon stocks for each plot type (i.e. each combination of burning/
invasion within the site) were reported as microsite-weighted carbon stocks
by weighting the carbon stock for each microsite by the relative canopy
cover for each microsite that was prese nt at a plot (see vegetation monitoring
above, and Fig. 1and Table S2) and then summing the average weighted
values for each microsite at an individual transect location (Eq. 2). The sum
of relative cover for all microsites at a site was always 100 and standard
carbon stocks are calculated as in Eq. 1.
Microsite weighted carbon stocks ¼
Standard carbon stockmicrosite A ×relative covermicrosite A þ
Standard carbon stockmicrosite B ×relative covermicrosite B þ
Standard carbon stockmicrosite C ×relative covermicrosite C
ð2Þ
Reported total carbon stocks for all site types within an ecoregion were
the sum of all measurable carbon stocks from 0–100 cm depth. Soil samples
were binned into depth categories of 0–5cm, 5–10 cm, 10–15 cm,
15–40 cm, 40–60 cm, and 60–100 cm (Figs. 2and 3). Carbon content
measurements of each sample within a depth range for an individual soil pit
were averaged and converted into stocks representing the entire depth
range. For some sites, soil was <100 cm deep and our pits ended at either
bedrock or consolidated bedrock which was not always possible to sample,
and also contained no mineral soil. The result was that the same depth
ranges are not represented at all plots, however, our carbon stock mea-
surements represent the total measurable carbon stocks in mineral soil from
0–100 cm in each site type . To estimate the effect of burning and invasion on
carbon stocks across the region, carbon stocks were averaged across ecor-
egions (Figs. 2and 3). Similarly, to estimate the microsite and ecoregion
effect on carbon stocks, carbon stocks were averaged across burning,
invasion and ecoregion or microsite effects (Figs. 3and S4) To estimateEAG
invasion impacts on soil carbon stocks across the 526,100 ha of sagebrush
steppe that is degraded annually (as reported in the abstract), differences in
total carbon stocks between burned/invaded and unburned/uninvaded sites
(Fig. 2) were multiplied by that area and scaled by 69%, which is the portion
of that degradation that was attributed to annual grass invasion9. Specifi-
cally, we report the observed range of loss of 47.1–55.2 Mg/ha carbon at our
sites which amounts to 42–49% compared to 113 Mg/ha in unburned
uninvaded communities (Fig. 2). This loss was multiplied by 526,100ha, or
the area of sagebrush that are estimated to be degraded each year and then
again multiplied by 69%, which is the portion of the degradation that has
been attributed to annual grass invasion, generating the range of 17.1–20 Tg
carbon that could be lost across the area of intact sagebrush steppe that is
disappearing annually.
Biomass
Herbaceous biomass and litter were collected from a within 1 m2frame
placed at randomly determined distances from the transect (but less than
5 m) at 5, 25, and 45m transect locations. Within the frame, each
microsite type that was identified on a transect was independently
sampled, and from within each microsite type, the following biomass
classes were collected separately: herbaceous, EAG, woody litter, and
herbaceous litter. Shrub biomass in Mg/ha was estimated by combining
shrub density measurements made at our plots with allometric equations
developed specifically for sagebrush shrubs94. Shrub density was esti-
mated using a frequency-density approach, by counting shrubs within a
5, 9, or 13 m radius until at least 5 mature shrubs were included in the
area. Three counts were made per transect within a 5 m radius at 5,
25, 45 m transect locations. Shrub biomass was estimated using Eq. 3
below, where diameter 1 was the largest diameter of the shrub canopy,
diameter 2 was perpendicular to diameter 1, and the shrub volume was
calculated as the volume of an elliptical cylinder based on diameter 1
and 2, and the shrub height (Pyke et al.94). All measurements were in cm
or cm3.
Biomass kg

¼0:0462 0:00294 Diameter1 þ0:00422 Diameter2
þ0:00000112 Shrub volume
ð3Þ
Statistics
Significant differences in carbon content or stocks between invaded and
uninvaded plots for all depth ranges of the soil were determined by paired t-
tests (Fig. 3,p< 0.05). All data from all ecoregions and microsites were
averaged and tests were conducted individually for each set of invaded and
uninvaded sites that were either burned or unburned. The error of carbon
stock measurements (Figs. 2and 3; S4) was the standard error of the mean
for all replicates of carbon stocks across all plots that were used to calculate a
reported value. The net effect of invasion and burning on soil carbon content
(% carbon by mass), and carbon stocks was assessed with generalized linear
mixedeffectsmodelsusingthepackageglmmTMBinR
95,96. Model struc-
tures are detailed in Table 1. Model residuals were normally distributed
(Shapiro–wilks test, p> 0.05), and there was no correlation between model
predicted values and residuals. Post-hoc Tukey HSD tests were conducted
on soil carbon stock data using HSD.test function from the agricolae package
(Fig. 2;S3andS4)
97.
Reporting summary
Further information on research design is available in the Nature Portfolio
Reporting Summary linked to this article.
Data availability
Data are available in the Dryad digital data repository https://doi.org/10.
5061/dryad.d2547d88k.
Received: 6 February 2024; Accepted: 27 September 2024;
References
1. Brooks, M. L. et al. Effects of invasive alien plants on fire regimes.
BioScience 54, 677 (2004).
https://doi.org/10.1038/s43247-024-01795-9 Article
Communications Earth & Environment | (2024) 5:669 8
Content courtesy of Springer Nature, terms of use apply. Rights reserved

2. Bradley, B. A. Regional analysis of the impacts of climate change on
cheatgrass invasion shows potential risk and opportunity. Glob.
Change Biol. 15, 196–208 (2009).
3. Pilliod, D. S., Welty, J. L. & Toevs, G. R. Seventy-five years of
vegetation treatments on public. Rangelands in the Great Basin of
North America. Rangelands 39,1–9 (2017).
4. Maxwell, T. M. & Germino, M. J. The effects of cheatgrass invasion on
US Great Basin carbon storage depend on interactions between plant
community composition, precipitation seasonality, and soil climate
regime. J. Appl. Ecol. 59, 2863–2873 (2022).
5. Zhu, Z. & Reed, B. C. Baseline and Projected Future Carbon Storage
and Greenhouse-Gas Fluxes in Ecosystems of the Western United
States. Professional Paper 1797. (United States Geological
Survey, 2012).
6. Bradford, M. A. et al. Managing uncertainty in soil carbon feedbacks to
climate change. Nat. Clim. Change 6, 751–758 (2016).
7. D’Antonio, C. M. & Vitousek, P. M. Biological Invasions by Exotic
Grasses, the Grass/fire Cycle, and Global Change. Annu. Rev. Ecol.
Syst. 23,63–87 (1992).
8. Hasselquist, N. J., Germino, M. J., Sankey, J. B., Ingram, L. J. & Glenn,
N. F. Aeolian nutrient fluxes following wildfire in sagebrush steppe:
implications for soil carbon storage. Biogeosciences 8, 3649–3659
(2011).
9. Doherty, K. et al. A Sagebrush Conservation Design to Proactively
Restore America’s Sagebrush Biome. Open-File Report 2022-1081.
(United States Geological Survey, 2022).
10. Germino, M. J., Belnap, J., Stark, J. M., Allen, E. B. & Rau, B. M.
Ecosystem Impacts of Exotic Annual Invaders in the Genus Bromus.
In: Exotic Brome-Grasses in Arid and Semiarid Ecosystems of the
Western US (eds. Germino, M. J., Chambers, J. C. & Brown, C. S.)
61–95 (Springer International Publishing, Cham, 2016).
11. Trumbore, S. E., Davidson, E. A., Barbosa De Camargo, P., Nepstad,
D. C. & Martinelli, L. A. Belowground cycling of carbon in forests and
pastures of eastern Amazonia. Glob. Biogeochem. Cycles 9,
515–528 (1995).
12. Jobbágy, 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).
13. Nagy, R. C. et al. A synthesis of the effects of cheatgrass invasion on
US Great Basin carbon storage. J. Appl. Ecol. 58, 327–337 (2021).
14. Abney, R. B. & Berhe, A. A. Pyrogenic carbon erosion: implications for
stock and persistence of pyrogenic carbon in soil. Front. Earth Sci. 6,
26 (2018).
15. Bechtold, H. A. & Inouye, R. S. Distribution of carbon and nitrogen in
sagebrush steppe after six years of nitrogen addition and shrub
removal. J. Arid Environ. 71, 122–132 (2007).
16. Austin, A. T. et al. Water pulses and biogeochemical cycles in arid and
semiarid ecosystems. Oecologia 141, 221–235 (2004).
17. De Graaff, M.-A., Throop, H. L., Verburg, P. S. J., Arnone, J. A. &
Campos, X. A synthesis of climate and vegetation cover effects on
biogeochemical cycling in shrub-dominated drylands. Ecosystems
17, 931–945 (2014).
18. Schlesinger, W. Evidence from chronosequence studies for a low
carbon-storage potential of soils. Nature 348, 232–234 (1990).
19. Stewart, C. E., Paustian, K., Conant, R. T., Plante, A. F. & Six, J. Soil
carbon saturation: concept, evidence and evaluation.
Biogeochemistry 86,19–31 (2007).
20. Boyd, C. S., Davies, K. W. & Hulet, A. Predicting fire-based perennial
bunchgrass mortality in big sagebrush plant communities. Int. J.
Wildland Fire 24, 527 (2015).
21. Bossio, D. A. et al. The role of soil carbon in natural climate solutions.
Nat. Sustain. 3, 391–398 (2020).
22. Bai, Y. & Cotrufo, M. F. Grassland soil carbon sequestration: current
understanding, challenges, and solutions. Science 377, 603–608
(2022).
23. Pellegrini, A. F. A. et al. Fire effects on the persistence of soil organic
matter and long-term carbon storage. Nat. Geosci. 15,5–13 (2022).
24. Matthews, H. D., Zickfeld, K., Koch, A. & Luers, A. Accounting for the
climate benefit of temporary carbon storage in nature. Nat. Commun.
14, 5485 (2023).
25. Bradley, B. A., Houghton, R. A., Mustard, J. F., & Hamburg, S.P.
Invasive grass reduces aboveground carbon stocks in shrublands of
the Western US. Glob. Change Biol. 12, 1815–1822 (2006).
26. Campos, X., Germino, M. & De Graaff, M.-A. Enhanced precipitation
promotes decomposition and soil C stabilization in semiarid
ecosystems, but seasonal timing of wetting matters. Plant Soil 416,
427–436 (2017).
27. McAbee, K., Reinhardt, K., Germino, M. J. & Bosworth, A. Response of
aboveground carbon balance to long-term, experimental enhancements
in precipitation seasonality is contingent on plant community type in
cold-desert rangelands. Oecologia 183,861–874 (2017).
28. Stanley, P., Spertus, J., Chiartas, J., Stark, P. B. & Bowles, T. Valid
inferences about soil carbon in heterogeneous landscapes.
Geoderma 430, 116323 (2023).
29. Harrison, R. B., Footen, P. W. & Strahm, B. D. Deep Soil Horizons:
Contribution and Importance to Soil Carbon Pools and in Assessing
Whole-Ecosystem Response to Management and Global Change.
For. Sci.57,67–76 (2011).
30. Throop, H. L., Archer, S. R., Monger, H. C. & Waltman, S. When bulk
density methods matter: Implications for estimating soil organic
carbon pools in rocky soils. J. Arid Environ. 77,66–71 (2012).
31. Zamanian, K., Pustovoytov, K.& Kuzyakov, Y. Pedogenic carbonates:
Forms and formation processes. Earth-Sci. Rev. 157,1–17 (2016).
32. Plaza, C. et al. Soil resources and element stocks in drylands to face
global issues. Sci. Rep. 8, 13788 (2018).
33. Chen, S. et al. Plant diversity enhances productivity and soil carbon
storage. Proc. Natl. Acad. Sci. USA 115, 4027–4032 (2018).
34. Hoover, A. N. & Germino, M. J. A common-garden study of resource-
island effects on a native and an exotic, annual grass after fire. Rangel.
Ecol. Manag. 65, 160–170 (2012).
35. Germino, M. J. et al. Thresholds and hotspots for shrub restoration
following a heterogeneous megafire. Landsc. Ecol. 33, 1177–1194
(2018).
36. Schmidt, M. W. I. et al. Persistence of soil organic matter as an
ecosystem property. Nature 478,49–56 (2011).
37. Crowther, T. W. et al. Quantifying global soil carbon losses in response
to warming. Nature 540, 104–108 (2016).
38. Arrouays, D., Deslais, W. & Badeau, V. The carbon content of topsoil
and its geographical distribution in France. Soil Use Manag. 17,
7–11 (2001).
39. Sothe, C., Gonsamo, A., Arabian, J. & Snider, J. Large scale mapping
of soil organic carbon concentration with 3D machine learning and
satellite observations. Geoderma 405, 115402 (2022).
40. Raich, J. W. & Nadelhoffer, K. J. Belowground carbon allocation in
forest ecosystems: global trends. Ecology 70, 1346–1354 (1989).
41. Brown, S. & Lugo, A. E. Effects of forest clearing and succession on
the carbon and nitrogen content of soils in Puerto Rico and US Virgin
Islands. Plant Soil 124,53–64 (1990).
42. Moreland, K. et al. Deep in the Sierra Nevada critical zone: saprock
represents a large terrestrial organic carbon stock. Environ. Res. Lett.
16, 124059 (2021).
43. Kulmatiski, A., Adler, P. B., Stark, J. M. & Tredennick, A. T. Water and
nitrogen uptake are better associated with resource availability than
root biomass. Ecosphere 8, e01738 (2017).
44. Barger, N. N. et al. Woody plant proliferation in North American
drylands: a synthesis of impacts on ecosystem carbon balance. J.
Geophys. Res. 116, G00K07 (2011).
45. Six, J. & Paustian, K. Aggregate-associated soil organic matter as an
ecosystem property and a measurement tool. Soil Biol. Biochem. 68,
A4–A9 (2014).
https://doi.org/10.1038/s43247-024-01795-9 Article
Communications Earth & Environment | (2024) 5:669 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved

46. Rasmussen, C. et al. Beyond clay: towards an improved set of
variables for predicting soil organic matter content. Biogeochemistry
137, 297–306 (2018).
47. Lavallee, J. M., Soong, J. L. & Cotrufo, M. F. Conceptualizing soil
organic matter into particulate and mineral‐associated forms to
address global change in the 21st century. Glob. Change Biol. 26,
261–273 (2020).
48. Vereecken, H. et al. Using pedotransfer functions to estimate the van
Genuchten–Mualem soil hydraulic properties: a review. Vadose Zone
J9, 795–820 (2010).
49. Cross, A. F. & Schlesinger, W. H. Biological and Geochemical
Controls on Phosphorus Fractions in Semiarid Soils.
Biogeochemistry.52, 155–172 (2001)
50. Chapin, F. S., Matson, P. A. & Vitousek, P. M. in Principles of Terrestrial
Ecosystem Ecology. 259–296 (Springer, 2012).
51. Throckmorton, H. M. et al. The soil matrix increases microbial C
stabilization in temperate and tropical forest soils. Biogeochemistry
122,35–45 (2015).
52. Sokol, N. W. & Bradford, M. A. Microbial formation of stable soil
carbon is more efficient from belowground than aboveground input.
Nat. Geosci. 12,46–53 (2019).
53. Kelly‐Slatten, M. J. et al. Root traits of perennial C
4
grasses contribute
to cultivar variations in soil chemistry and species patterns in
particulate and mineral‐associated carbon pool formation. GCB
Bioenergy 15, 613–629 (2023).
54. Stanbery, C. et al. Controls on the presence and storage of soil
inorganic carbon in a semi-arid watershed. CATENA 225,
106980 (2023).
55. Schlesinger, W. H. The formation of caliche in soils of the Mojave
Desert, California. Geochim. Cosmochim. Acta 49,57–66 (1985).
56. Stavi, I., Lavee, H., Ungar, E. D. & Sarah, P. Ecogeomorphic feedbacks
in semiarid rangelands: a review. Pedosphere 19, 217–229 (2009).
57. Boxell, J. & Drohan, P. J. Surface soil physical and hydrological
characteristics in Bromus tectorum L. (cheatgrass) versus Artemisia
tridentata Nutt. (big sagebrush) habitat. Geoderma 149, 305–311
(2009).
58. Zamanian, K., Zhou, J. & Kuzyakov, Y. Soil carbonates: the
unaccounted, irrecoverable carbon source. Geoderma 384, 114817
(2021).
59. Schimel, D. S. et al. Climatic, edaphic, and biotic controls over storage
and turnover of carbon in soils. Glob. Biogeochem. Cycles 8,
279–293 (1994).
60. Oades, J. M. The retention of organic matter in soils. Biogeochemistry
5,35–70 (1988).
61. Kirschbaum, M. U. F. & Mueller, R. Net Ecosystem Exchange:
Workshop Proceedings, CRC for Greenhouse Accounting, April 2001.
(CRC for Greenhouse Accounting, Canberra, 2001).
62. Poeplau, C. et al. Isolating organic carbon fractions with varying
turnover rates in temperate agricultural soils—A comprehensive
method comparison. Soil Biol. Biochem. 125,10–26 (2018).
63. Craddock, G. & Pearse, C. K. Surface Run-off and Erosion on Granitic
Mountain Soils of Idaho as Influenced by Range Cover, Soil
Disturbance, Slope and Precipitation Intensity. (1938).
64. Roundy, B. A., Hardegree, S. P., Chambers, J. C. & Whittaker, A.
Prediction of cheatgrass field germination potential using wet thermal
accumulation. Rangel. Ecol. Manag. 60, 613–623 (2007).
65. Bishop, T. B. B. et al. Spatiotemporal patterns of cheatgrass
invasion in Colorado Plateau National Parks. Landsc. Ecol. 34,
925–941 (2019).
66. Price, S. J. & Germino, M. J. Variability in weather and site properties
affect fuel and fire behavior following fuel treatments in semiarid
sagebrush-steppe. J. Environ. Manage. 353, 120154 (2024).
67. Williams, A. J., Buck, B. J. & Beyene, M. A. Biological soil crusts in the
Mojave Desert, USA: micromorphology and pedogenesis. Soil Sci.
Soc. Am. J. 76, 1685–1695 (2012).
68. Li, J., Okin, G. S. & Epstein, H. E. Effects of enhanced wind erosion on
surface soil texture and characteristics of windblown sediments. J.
Geophys. Res. Biogeosci. 114, 2008JG000903 (2009).
69. Sankey, J. B., Germino, M. J., Sankey, T. T. & Hoover, A. N. Fire effects
on the spatial patterning of soil properties in sagebrush steppe, USA: a
meta-analysis. Int. J. Wildland Fire 21, 545 (2012).
70. Holling, C. S. Resilience and stability of ecological systems. Annu.
Rev. Ecol. Syst. 4,1–23 (1973).
71. Mack, R. N. Alien Plant Invasion into the Intermountain West: A Case
History. In: Ecology of Biological Invasions of North America and
Hawaii. Ecological Studies, Vol. 58 191–213 (Springer, New
York, 1986).
72. Chambers, J. C. et al. Resilience to stress and disturbance, and
resistance to bromus tectorum L. invasion in cold desert shrublands of
Western North America. Ecosystems 17, 360–375 (2014).
73. Morris, K. A., Saetre, P., Norton, U. & Stark, J. M. Plant community
effects on soil moisture and nitrogen cycling in a semi-arid ecosystem.
Biogeochemistry 159, 215–232 (2022).
74. Chambers, J. C. et al. Operationalizing resilience and resistance
concepts to address invasive grass-fire cycles. Front. Ecol. Evol. 7,
185 (2019).
75. Post, W. M., Emmanuel, William, R., Zinke, P. J. & Stangenberger, A.
G. Soil carbon pools and world life zones. Nature 298, 156–159 (1982).
76. Fornara, D. A. & Tilman, D. Plant functional composition influences
rates of soil carbon and nitrogen accumulation. J. Ecol. 96, 314–322
(2008).
77. Wieder,W.R.,Grandy,A.S.,Kallenbach,C.M.,Taylor,P.G.&Bonan,
G. B. Representing life in the Earth system with soil microbial functional
traits in the MIMICS model. Geosci. Model Dev. 8, 1789–1808 (2015).
78. Pierson, D. et al. Optimizing process-based models to predict current
and future soil organic carbon stocks at high-resolution. Sci. Rep. 12,
10824 (2022).
79. Hirmas, D. R., Amrhein, C. & Graham, R.C. Spatial and process-based
modeling of soil inorganic carbon storage in an arid piedmont.
Geoderma 154, 486–494 (2010).
80. Bird, M. I., Wynn, J. G., Saiz, G., Wurster, C. M. & McBeath, A. The
pyrogenic carbon cycle. Annu. Rev. Earth Planet. Sci. 43, 273–298
(2015).
81. Stanley, P. L., Wilson, C., Patterson, E., Machmuller, M. B. & Cotrufo,
M. F. Ruminating on soil carbon: applying current understanding
to inform grazing management. Glob. Change Biol. 30,e17223
(2024).
82. U.S. Environmental Protection Agency. Primary Distinguishing
Characteristics of Level III Ecoregions of the Continental United States
(U.S. Environmental Protection Agency, 2013).
83. National Interagency Fire Center. Historical Wildland Fires (National
Interagency Fire Center, 2018).
84. Baker, W. L. Fire and restoration of sagebrush ecosystems. Wildl.
Soc. Bull. 34, 177–185 (2006).
85. Bukowski, B. E. & Baker, W. L. Historical Fire regimes, reconstructed
from land-survey data, led to complexity and fluctuation in sagebrush
landscapes. Ecol. Appl. 23, 546–564 (2013).
86. Allred, B. W. et al. Improving Landsat predictions of rangeland
fractional cover with multitask learning and uncertainty. Methods
Ecol. Evol. 12, 841–849 (2021).
87. Applestein, C. & Germino, M. J. How do accuracy and model
agreement vary with versioning, scale, and landscape heterogeneity
for satellite-derived vegetation maps in sagebrush steppe? Ecol.
Indic. 139, 108935 (2022).
88. Applestein, C. & Germino, M. J. Satellite-derived plant cover maps
vary in performance depending on version and product. Ecol. Indic.
155, 110950 (2023).
89. US EPA, Ecoregions of North America.https://www.epa.gov/eco-
research/ecoregions-north-america (US Environmental Protection
Agency, 2015).
https://doi.org/10.1038/s43247-024-01795-9 Article
Communications Earth & Environment | (2024) 5:669 10
Content courtesy of Springer Nature, terms of use apply. Rights reserved

90. PRISM Climate Group, Oregon State University, https://prism.
oregonstate.edu (PRISM Climate Group, Oregon State University,
2023) Data created Feb 2014, accessed September 2023.
91. Huber, D. P. et al. Vegetation and precipitation shifts interact to alter
organic and inorganic carbon storage in cold desert soils. Ecosphere
10, e02655 (2019).
92. Miller, W. P. & Miller, D. M. A micro-pipette method for soil mechanical
analysis. Commun. Soil Sci. Plant Anal. 18,1–15 (1987).
93. Walter, K., Don, A., Tiemeyer, B. & Freibauer, A. Determining soil bulk
density for carbon stock calculations: a systematic method
comparison. Soil Sci. Soc. Am. J. 80, 579–591 (2016).
94. Pyke, D. A. et al. Region-wide ecological responses of arid Wyoming
big sagebrush communities to fuel treatments. Rangel. Ecol. Manag.
67, 455–467 (2014).
95. Brooks, M. E. et al. glmmTMB balances speed and flexibility among
packages for zero-inflated generalized linear mixed modeling. RJ9,
378 (2017).
96. R CoreTeam. R: A Language and Environment for StatisticalComputing.
R Foundation for Statistical Computing (R Core Team, 2022).
97. Lenth, R. V. Estimated marginal means, aka least-squares means [R
Package Emmeans Version 1.6. 0]. (2021).
Acknowledgements
This project would not have been possible without the hard work and
dedication of Evan Blodgett,Krystal Busby,Katie Bush, Bill Davidson,Austin
Davis, Noah Johnson, Gavin Kerr, Chad Kluender, Andrew Lague, Brynne
Lazarus, Darius Liles, Molly Long, Sophie Steppe, Jayna Thompson, and
Chloe Watt. Sampling was done with permission from the local Bureau of
Land Management offices for the Northern Basin and Range and Snake
River Plain ecoregions, and from the Idaho Department of Fish and Game
Boise River Wildlife Management Area management staff for the Idaho
Batholithecoregion. Funding wasprovided for this project by Environmental
Science U.S.LLC, Cary NC, 27513. Any use of trade,firm, or product names
is for descriptive purposesonly and does not imply endorsement by theU.S.
Government.
Author contributions
M.J.G and H.E.Q. conceived the idea. T.M.M., S.J.P., and M.J.G. designed
the experiment. T.M.M. and S.J.P. led the field sampling campaign. S.J.P.
performedthe geospatial analysis. T.M.M.wrote the first draftand co-led the
writing of the manuscript with M.J.G., analyzed the data, and performed the
statistical analysis. All authors contributed to the interpretation of resultsand
edited and approved the final manuscript.
Competing interests
H.E.Q. is employed by Environmental Science US LLC, a solutions provider
for rangeland protection.
Additional information
Supplementary information The online version contains
supplementary material available at
https://doi.org/10.1038/s43247-024-01795-9.
Correspondence and requests for materials should be addressed to
Matthew J. Germino.
Peer review information Communications Earth & Environment thanks
Jingyi Ding, Yong Zhou and the other, anonymous, reviewer(s) for their
contribution to the peer review of this work. Primary handling editors: Erica
Buscardo and Martina Grecequet. A peer review file is available.
Reprints and permissions 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 affiliations.
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 licence, and indicate if changes
were made. The images or other third party material in this article are
included in the article’s Creative Commons licence, unless indicated
otherwise in a credit line to the material. If material is not included in the
article’s Creative Commons licence 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
licence, visit http://creativecommons.org/licenses/by/4.0/.
This is a U.S. Government work and not under copyright protection in the US;
foreign copyright protection may apply 2024
https://doi.org/10.1038/s43247-024-01795-9 Article
Communications Earth & Environment | (2024) 5:669 11
Content courtesy of Springer Nature, terms of use apply. Rights reserved

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com