Page 1
Page 1/20
Direct yield bene ts of soil carbon increases in low-
carbon soils: A global meta-analysis of cover
cropping co-bene ts
Isaac Vendig (
isaacvendig@berkeley.edu )
University of California, Berkeley
https://orcid.org/0000-0002-1651-4490
Aidee Guzman
University of California Berkeley
Gisel De Lacerda
University of California, Berkeley
Kenzo Esquivel
University of California, Berkeley
Lauren Ponisio
University of Oregon
Timothy Bowles
University of California, Berkeley
https://orcid.org/0000-0002-4840-3787
Analysis
Keywords:
Posted Date:
July 27th, 2022
DOI:
https://doi.org/10.21203/rs.3.rs-1848483/v1
License:
This work is licensed under a Creative Commons Attribution 4.0 International License.
Read Full License
Version of Record:
A version of this preprint was published at Nature Sustainability on May 29th, 2023.
See the published version at
https://doi.org/10.1038/s41893-023-01131-7
.
Page 2
Page 2/20
Abstract
Cropland management practices that restore soil organic carbon (SOC) are increasingly presented as
climate solutions that provide the additional bene t of yield enhancement. But how often these bene ts
align at the farm level — the scale of farmers’ decision-making — remains uncertain. We examined
concurrent SOC and yield responses to cover cropping, including their direct connection, with a global
meta-analysis. We showed that cover cropping simultaneously increased yields and SOC in 59.7% of 434
paired observations. Direct yield bene ts from SOC enhancement were evident in soils with initial SOC
concentrations below 9.2 g kg
-1
. These yield bene ts did not vary with nitrogen inputs or cover crop type,
suggesting they are not substitutable with fertilization. In all types of soils, the largest yield increases (up
to +20.2%) coincided with the largest SOC increases (up to +16.8%) when legume cover crops were
integrated into systems with simpli ed rotations or with nitrogen inputs < 136 kg N ha
-1
season
-1
, thus
providing substantial bene ts for farmers and society regardless of direct effects of SOC on yield.
Background
Soil organic carbon (SOC) is considered a critical component of soil health. In agroecosystems, soil
health is a metaphor that describes the degree to which soils support multiple functions beyond just crop
productivity
1,2
. SOC in uences multiple soil-based ecosystem services, such as nutrient cycling and
retention, soil aeration and structural integrity
3
, climate regulation
4
, and possibly crop productivity
5
. The
concentration of SOC has thus become one of the most common metrics for assessing the state of a
soil’s health
6
.
Despite the various bene ts that SOC is thought to provide
7
, agricultural expansion and conventional
agricultural intensi cation have dramatically depleted SOC across the world
8
. Practices that regenerate
SOC are garnering increasing attention for their potential to restore soil functionality while simultaneously
drawing down atmospheric carbon
9,10
. Cover cropping is one such cropland practice. Grown on fallow
soils otherwise left bare, cover crops increase organic matter inputs to the soil in the form of crop detritus
and root exudates. This in turn supports soil microbial biomass which turns over and forms microbial
necromass, a key component of stable SOC
11,12
. A recent meta-analysis showed that cover cropping
increases SOC by 0.32 Mg C ha
-1
yr
-1
on average, highlighting its potential to restore some portion of the
133 Pg of global SOC that has been lost from croplands
8,13
. But the extent to which farmers will
voluntarily adopt C sequestering practices hinges on more than just their potential to mitigate climate
change or restore soil health
14,15
.
How a practice in uences crop productivity and farm pro tability is central to farmers’ management
decisions. Cover cropping can affect both input costs (e.g., via reducing fertilizer input costs following a
Page 3
Page 3/20
legume cover crop) and crop yields. Recent meta-analyses show that cover cropping typically increases
crop yields
16,17
. Yield increase estimates range from 6% to 33% depending on cash crop type, cover crop
type, fertilizer additions and other factors, although some studies show crop yield decreases as well
16,17
.
However, since syntheses of how cover cropping affects SOC and yields have been conducted separately,
it is not known how often cover cropping simultaneously increases SOC and yields (co-bene ts) at the
same location, increases or decreases one but not the other (trade-offs), or even decreases both SOC and
yields (co-costs). Perhaps more importantly, it is also not known if there are management, edaphic, or
environmental conditions in which the largest yield increases are most likely to align with the largest SOC
increases. Understanding the potential for co-bene ts will help inform decision-making at the farm level
and will help identify areas of overlap between farm level bene ts and bene ts for society that might
occur at regional or global scales.
When yield increases do result from cover cropping, a critical knowledge gap is the relative role of
changes in SOC in driving these increases, versus other cover cropping effects, such as nutrient
scavenging
18
. Understanding the role that SOC plays in yield changes under cover cropping would
contribute to recent calls to better quantify the relationship between SOC and yields generally
5,19
.
The widespread expectation that increasing SOC will increase crop productivity exists
8,11,20,21
because, as
part of soil organic matter, SOC is related to many soil properties and functions that are important for
plant productivity like nutrient and water provisioning. However, actual evidence of a relationship between
SOC and yield remains contradictory and inconclusive
5,22–24
. One reason for the lack of consensus is the
challenge of manipulating SOC in such a way that the effect of nutrient additions can be separated from
other possible bene ts of SOC
25
. Attempts to circumvent this challenge use observational data, but the
lack of controls and covariation between SOC and other environmental and management variables create
complex interactions that can be di cult to tease apart even using multivariate approaches
5,22
. Using
similar meta-analytic techniques, recent studies have reported positive effects of SOC on yield
5,24
, little to
no effects
22
, and negative effects
23
. In addition, observational studies examining SOC-to-yield
relationships span very wide ranges of SOC
5,24
. Regional or global SOC-to-yield relationships are
generally not applicable to an individual farmer making management changes since they encompass
very wide ranges of SOC rather than the often modest SOC increases that occur on-farm following
changes to management.
Meta-analysis of studies on agricultural practices expected to shift SOC, such as cover cropping, provides
an alternative approach to quantifying the SOC-to-yield relationship
5
. By pairing treatments with relevant
control values, relationships between changes in SOC and changes in yield can be quanti ed in such a
Page 4
Page 4/20
way that eliminates the confounding effects that result from observational data (e.g., between climate or
edaphic factors that in uence both SOC and yields). While other effects can also confound or obscure
the SOC-to-yield relationship in this approach (e.g., increases in both nitrogen availability and SOC from
legume cover crops), building a broad yield model that includes such interactions can increase
con dence in the causality of SOC effects.
We use a global meta-analysis to determine how cover cropping affects SOC and crop yields
simultaneously, and the extent to which changes in crop yield (
Δ
Yield
) are related to changes in SOC
(
Δ
SOC
). We thus build on previous meta-analyses that assess how cover cropping affects SOC or yields
individually by linking these responses together in a paired treatment-control meta-dataset. We asked 3
questions: 1) are co-bene ts, i.e., simultaneous increases in crop yields and SOC, the most common
response to cover cropping? 2) Do changes in SOC link directly to changes in yield and, if so, is this
association nutrient related? 3) Regardless of direct links between SOC and yield, are there edaphic,
environmental, or management conditions where co-bene ts of increased SOC and yield from cover
cropping are more likely to be maximized? We compiled an exhaustive database of paired yield and SOC
responses to cover cropping and constructed models with important factors mediating their individual
and joint responses. By building comprehensive models to identify and quantify important predictors of
yield and SOC changes from cover cropping, our study not only helps with farm-level decisions regarding
cover cropping, but also informs policymakers seeking to quantify the impact of cropland carbon
sequestration on global food production capacity.
Results
Joint impacts of cover cropping on crop yields and SOC
Cover cropping had a strong positive effect on both SOC and yield. The modeled change across all
management types and all sites was + 10.9% [95% CI: 7.5–14.5] for yield (Fig. 3A) and + 9.8% [95% CI:
7.0–12.6] for SOC (Fig. 4A). In 59.7% of the 434 paired observations in our dataset, cover cropping
increased both SOC and yields (Fig. 1; top right quadrant). Trade-offs, in which either SOC or yield
increased while the other decreased, accounted for about one-third of observations. In 20.7% of paired
comparisons, cover crops increased SOC but decreased yield; in 12.9% of cases, cover crops increased
yields but decreased SOC (Fig. 1; top left and bottom right quadrants). Co-costs, in which cover cropping
negatively affected both yields and SOC, accounted for 6.7% of paired observations (Fig. 1; bottom left
quadrant). Funnel plots showed no indication of publication bias (Figure S1).
Explaining variability in crop yield responses to cover
cropping
Page 5
Page 5/20
Out of 29 management and environmental variables considered as possible moderators of the effect of
cover crops on cash crop yield, our yield change (
Δ
Yield
) model included an interaction between SOC
change (
Δ
SOC
) and initial SOC as well as rotational complexity and N fertilizer, each interacting with cover
crop type (legume vs. non-legume) (Table 1). Rotational complexity is the number of different cash crop
species rotated in a given plot throughout the length of an experiment and N fertilizer is the N (kg N ha
− 1
)
added to both treatment and control plots in each cash crop season. Marginal
of our
Δ
Yield
model was
0.27 and conditional
was 0.89. Addition of edaphic variables such as soil texture and sampling depth
did not improve model t (see Table S1). The average initial SOC concentration of our dataset was 15.5 ±
9.2 g kg
− 1
(standard deviation) at an average sampling depth of 0-18.4 cm ± 7.3 cm (standard deviation).
All adjusted generalized variance in ation factor (GVIF) values were below 3, indicating the effect of
multicollinearity was negligible.
Δ
Yield
Model Results
Variable
F-value
df
p-value
Initial SOC
0.30
1,92
0.59
Δ
SOC
2.69
1,67
0.11
Cover Crop Type
0.25
1,27
0.62
Rotational Complexity
9.90
2,71
<0.001
N Fertilizer
31.26
1,27
<0.001
Absolute Latitude
1.42
1,87
0.24
Δ
SOC
Initial SOC
4.90
1,85
<0.05
Rotational Complexity Cover Crop Type
20.47
2,25
<0.001
N Fertilizer Cover Crop Type
29.83
1,24
<0.001
Δ
SOC
Cover Crop Type
2.17
1,24
0.15
Δ
SOC
N Fertilization
0.01
1,63
0.91
Table 1. Type III ANOVA results from our
Δ
Yield
model (n = 417). df is numerator and denominator degrees
of freedom, respectively, with Kenward-Roger approximation for denominator degrees of freedom.
Δ
Yield
and
Δ
SOC
are the log cash crop yield and SOC response ratios, respectively. Initial SOC is SOC (g kg
-1
)
prior to cover cropping. Cover crop type is binary categorical; legume vs non-legume coded 0 and 1,
R
2
R
2
Page 6
Page 6/20
respectively. N fertilization is in-season cash crop N fertilization (kg N ha
-1
season
-1
). Rotational
complexity is the number of different cash crop species in rotation throughout the experiment. p-values in
italics are considered signi cant at
We found that SOC changes from cover cropping (
Δ
SOC
) were directly associated with yield changes
(
Δ
Yield
), but only in soils with initial SOC values of 9.2 g kg
− 1
or less (Fig. 2). In soils with initial SOC
values of 5 g kg
− 1
, for instance, a 10% increase in SOC (i.e., a change from 5 g kg
− 1
C to 5.5 g kg
− 1
C)
was associated with a 2.8% yield increase. In soils with initial SOC values greater than 9.2 g kg
− 1
,
Δ
SOC
was not signi cantly associated with
Δ
Yield
. The
Δ
SOC
-to-
Δ
Yield
relationship did not differ between cover
crop types (legume vs. non-legume) and did not vary across differing levels of N fertilization.
The effect of rotational complexity on
Δ
Yield
differed between legume cover crops and non-legume cover
crops (Fig. 3B, Fig. 3C). Holding all else at its average,
Δ
Yield
in legume cover crop treatments was
signi cantly greater in continuous cash crop monocultures (+ 21.0%, 95% CI: 15.0–27.3) versus rotations
with two (+ 9.3%, 95% CI: 1.7–17.5) cash crop species (Fig. 3B). For rotations with 3 or more cash crops,
Δ
Yield
from legume cover crops was not statistically different from zero. For non-legume cover crops, the
magnitude of
Δ
Yield
across rotational complexity groups varied but not signi cantly so. At the average
level of N fertilization in our dataset, non-legume cover crops signi cantly increased yield only in plots
with 3 or more cash crops in rotation (+ 19.2%, 95% CI: 6.9–33.0) (Fig. 3C).
Δ
Yield
from non-legume cover
crops in continuous monoculture and two-crop rotations was positive but overlapped zero (+ 5.1%, 95%
CI: -1.0–11.5; +4.5%, 95% CI: -3.3–12.9, respectively).
We found that increased N fertilization reduced
Δ
Yield
in legume cover crop treatments but did not have a
signi cant effect on
Δ
Yield
from non-legume cover crops (Fig. 3D, Fig. 3E). Legume cover crops in low N
systems (12.9 kg N ha
− 1
season
− 1
, one standard deviation below the mean N fertilization of our dataset)
increased yield by + 17.8% (95% CI: 11.3–24.7) and in average N systems (85.9 kg N ha
− 1
season
− 1
)
increased yield by + 10.6% (95% CI: 4.8–16.8) (Fig. 3D). In systems receiving more than 136 kg N ha
− 1
season
− 1
, we found no statistically signi cant effect of legume cover crops. Non-legume cover crops
provided yield increases in low (+ 7.5%, 95% CI: 0.9–14.4), average (+ 9.4%, 95% CI: 3.0–16.2), and high
(+ 11.4%, 95% CI: 3.7–19.6) N systems (Fig. 3E).
SOC responses to cover cropping
Our
Δ
SOC
model included absolute latitude, cover crop type (legume vs non-legume), and initial SOC as
variables which moderated the effect of cover crops on SOC (Table 2). Marginal
was 0.11 and
conditional
was 0.92. Addition of other environmental and management variables such as aridity,
R
2
R
2
Page 7
Page 7/20
mean annual precipitation, N and phosphorus fertilization, and tillage did not improve model t (see
Table S2 for full list).
Δ
SOC
Model Results
Variable
F value
df
p-value
Cover Crop Type
16.52
1,25
<0.001
Absolute Latitude
7.22
1,90
<0.05
Initial SOC
4.52
1,90
<0.01
Table 2. Type III ANOVA results from our
Δ
SOC
model (n = 434). df is numerator and denominator degrees
of freedom, respectively, with Kenward-Roger approximation for denominator degrees of freedom.
Δ
SOC
is
the log SOC response ratio. Cover Crop Type is binary categorical; non-legume vs legume. Absolute
Latitude is the absolute value of study latitude. Initial SOC is SOC (g kg
-1
) prior to cover cropping. p-
values in italics are considered signi cant at
While both cover crop types signi cantly increased SOC, we found that legume cover crops increased
SOC (+ 15.4%, 95% CI: 11.5–19.5) more than non-legume cover crops (+ 11.2%, 95% CI: 7.2– 15.3) (Fig.
4B). Cover crops became less effective at increasing SOC as absolute latitude increased (Fig. 4C). In
addition, we found that the largest SOC increases occurred in sites with lower initial SOC values. In sites
with initial SOC values of 6.3 g kg
− 1
, for instance, cover cropping resulted in a + 15.5% SOC increase (95%
CI: 11.3–19.9) while at the average initial SOC value in our dataset of 15.5 g kg
− 1
, cover cropping
increased SOC by 13.3% (95% CI: 9.4–17.2) (Fig. 4D).
Discussion
In our meta-analysis of 82 studies spanning 5 continents, we found that cover crops increased crop yields
concurrently with SOC in 59.7% of 434 paired observations, thus providing a win-win outcome for farmers
and society a majority of the time.
Δ
SOC
was directly associated with
Δ
Yield
only in soils with relatively low
SOC prior to cover cropping. The yield bene t of increased SOC did not diminish in systems with higher N
inputs and did not differ between cover crop types (legume vs non-legume), indicating that N inputs
cannot substitute for changes in SOC that link to higher yields. The largest SOC increases occurred in
legume cover crop treatments (+ 15.4%) and the largest yield increases also occurred from legume cover
crops in systems with low to average N inputs and in 1–2 crop rotations (up to + 21.0%).
Page 8
Page 8/20
Direct relationships between changes in SOC and yield
Our experimentally based approach identi ed a
Δ
SOC
-to-
Δ
Yield
response that does not vary based on N
inputs or with legume vs. non-legume cover crops. A negative
Δ
SOC
by N fertilization interaction would
have indicated that the yield bene t from SOC was substitutable for N inputs and therefore N related.
Likewise, if the
Δ
SOC
-to-
Δ
Yield
relationship differed between legume and non-legume cover crops, then
some portion of the SOC bene t likely would have been a re ection of yield bene ts from N xation. In
the absence of nutrient interactions with
Δ
SOC
, the link we found between
Δ
SOC
and
Δ
Yield
is likely better
explained by bene ts of increased SOC like reduced compaction and increased aeration
3
. Our results thus
help to identify and quantify the yield bene ts of soil improvement provided by SOC for which fertilization
cannot substitute.
Our analysis clari es contrasting results of observational meta-analyses regarding whether the yield
bene t of additional SOC declines in higher SOC soils
5,24
. It also likely explains the nding of no
relationship between yield and SOC reported in an observational meta-analysis of Danish farms
22
in
which there were very few observations with SOC concentrations below 10 g kg
− 1
. The yield bene t of
increased SOC that we identi ed is larger than that reported in a previous observational data meta-
analysis. For a hypothetical increase from 5 g kg
− 1
to 8 g kg
− 1
, our model predicted a + 16.2% yield
increase, compared to the + 10% yield increase previously reported
5
. Although farms with average SOC
levels are unlikely to see direct yield bene ts of soil improvement from increased SOC, our ndings
demonstrate that there is a direct material incentive for farms on degraded soils to increase SOC levels at
least until 9.2 g kg
− 1
.
Aligning carbon sequestration goals with maximum yield
bene ts
Regardless of direct links between
Δ
SOC
and
Δ
Yield
, we found that a substantial opportunity to sequester
carbon as SOC while also increasing crop yield lies in the incorporation of legume cover crops into
systems with few (one to two) cash crops in rotation. Legume cover crops provided increases of + 15.4%
SOC and + 21.0% yield in continuous monocrop cultures. In two crop rotations, legume cover crops
increased yield by + 9.3% while the + 15.4% SOC increase remained unchanged (i.e.,
Δ
SOC
did not vary
across rotation types). We did not nd evidence that the presence of legume cash crops like soy in more
complex rotations accounted for the diminishing yield bene t of legume cover crops in more complex
rotations (see Table S1). Yield bene ts of crop rotational diversi cation are well-known
26,27
and based on
our results here, appear to be redundant with legume cover crops in more complex rotational systems. In
contrast, non-legume cover crops did not signi cantly increase yield in monocultures or simple two-crop
rotations but did increase yield with three or more cash crops in rotation, although there were relatively
few data points in this grouping (n = 22). This suggests a need for further research on how to optimize
cover crops in more complex cash crop rotations.
Page 9
Page 9/20
We identi ed low to average N input systems as other key farm types where carbon sequestration goals
can align with maximum yield increases through the use of legume cover crops. While
Δ
SOC
did not vary
across N fertilization levels, we found that in N limited systems, the yield bene t of leguminous N xation
is relatively greater. In low N systems (de ned here as 12.9 kg N ha
− 1
season
− 1
), legume cover crops
increased yield by + 17.8% and increased SOC by + 15.4%. While still providing a notable yield increase,
the yield bene t of legume cover crops was + 10.6% at the average level of N fertilization in our dataset
(85.9 kg N ha
− 1
season
− 1
) and remained signi cant and positive through 136 kg N ha
− 1
season
− 1
at
which point
Δ
Yield
from legumes was not distinguishable from the null (no effect). Legumes could thus
allow for reducing synthetic N fertilizer inputs while maintaining yields, which also comes with
environmental bene ts (e.g., from reductions in greenhouse gasses from fertilizer production).
The larger SOC response from legumes compared to non-legume cover crops (+ 15.4% vs. +11.2%,
respectively) contrasts with no effect of cover crop type found in prior meta-analyses
13,28
, possibly due to
their more limited datasets. With relatively more labile plant inputs that microbes e ciently use, legumes
may be particularly effective at building soil organic matter pools, including mineral associated organic
matter, that are both stable and supply N
12,29−31
. The negative relationship between latitude and
Δ
SOC
is
likely related to reduced biomass production for cover crops at higher absolute latitudes
28
. Cover crops in
these latitudes have less time to grow in colder conditions with less sunlight. This nding is in line with a
recent meta-analysis which found the length between planting and termination to be an important
predictor of SOC response
28
. Our nding that
Δ
SOC
was greatest in soils with lower initial SOC is
reasonable in that a similar absolute SOC increase will translate to a larger relative change in soils with
lower initial SOC levels.
Our global meta-analysis demonstrates that the goal of building soil carbon through cover cropping
aligns with the goal of increasing or maintaining crop yields. Importantly, since these goals align at the
site level ~ 60% of the time, bene ts of higher yields for farmers are achievable concurrently with the
societal bene t of carbon sequestration
10
. Non-nutrient yield bene ts related to SOC (e.g., improved soil
structure) were only evident in soils with below average SOC. This nding suggests that direct yield
bene ts from SOC increases could help motivate farmers’ adoption of SOC enhancing practices in soils
with low SOC, but other incentives will be needed for farmers with SOC levels greater than ~ 10 g kg
− 1
.
We therefore suggest that determining the conditions for which changes to agricultural management
provide co-bene ts for crop yields and SOC — rather than establishing universal relationships between
SOC and yield — will be more useful for spurring agricultural transitions that produce food while also
mitigating climate change. To achieve carbon sequestration goals while supporting crop yields,
diversifying simpli ed rotations with legumes is a promising strategy given that legumes often provided
the largest bene t to both SOC and yields. Likewise, in low to average N input systems, the greatest yield
bene ts can be aligned with the greatest SOC bene ts through the use of legume cover crops. For
systems with complex rotations or high N inputs, non-legume cover crops are a better choice to support
yield goals while increasing SOC. Identifying when and where agricultural management practices deliver
Page 10
Page 10/20
direct bene ts to farmers and contribute to climate change mitigation will help with the urgent need to
increase the carbon sink of agricultural lands.
Methods
Study Selection
We selected cover cropping studies according to the following criteria: 1) the experimental design
includes one or more replicated cover cropping treatments, de ned as a non-harvested crop grown
between productive seasons; 2) the study includes a clear control as either bare fallow or spontaneous
off-season regrowth (e.g., “winter weeds”); 3) data are available for both SOC and cash crop yield, each
measured no more than one year apart; 4) cash crop yield is measured as fruit or grain; 5) yield and SOC
are available as yearly or monthly values rather than averages across multiple years (for maximum
accuracy in matching SOC values with associated yields); and 6) annual fertilizer inputs are equal across
control and treatment or are administered based on pre-season soil tests. Potted plant experiments were
not included in our dataset.
We began our literature search with the study lists of two recent cover cropping meta-analyses
13,17
and
subsequently searched ISI Web of Science for additional studies that matched our criteria using the
search string TS=((cover crop* OR catch crop OR fallow OR green manure) AND carbon AND yield). In
October 2020, the date of our nal search, our search string returned 2,451 studies. If an article reported
only SOC data or yield data, we used key terms related to the experiment to search Google Scholar for
articles reporting on the same experiment in order to ll in the missing data. In 11 instances, gray
literature sources such as master’s theses, dissertations and conference proceedings were used to
supplement data from peer-reviewed publications. In addition to Google Scholar searches, 36 authors
were contacted for additional data or methodological clari cations, out of which 8 responded and 3
provided additional data and/or information.
Our nal dataset spanned 5 continents and contained data from 82 distinct experiments gathered from
120 sources (107 peer reviewed journal articles, 6 master’s theses, 2 dissertations, 3 publicly available
datasets, and 2 conference proceedings). A list of data sources used in the study along with extraction
notes is provided in the supplementary material.
Data Compilation and Extraction
We quanti ed the effect of cover crops on yield and SOC using the log response ratio, calculated as the
natural log of the cover crop treatment value divided by that of the respective bare fallow control. Within a
given study, a treatment value was matched to a control value in the response ratio (RR) only if both
groups differed in no other respect than the use of cover cropping (e.g., same tillage regime, same N
application, etc.) and if the treatments were sampled at the same time. This aspect of our study design
allowed us to control for confounding effects that would otherwise be introduced in a direct comparison
of raw values between studies such as environmental conditions, management decisions, or edaphic
Page 11
Page 11/20
factors. In the case of the yield response to cover cropping, our use of the RR allowed us to make
comparisons across crops with different morphological characteristics (e.g., tomatoes vs. cotton)
because weight units are normalized by the ratio. Site-level initial SOC values were not available for some
of the studies in our dataset. To approximate missing site-level values, we used the earliest SOC sample
available for the non-cover crop control, assuming that the eld had likely been under a no-cover crop
planting regime prior to the initiation of the cover cropping experiment. Although differing sampling
depths across studies have the potential to obscure trends when comparing raw SOC values, we did not
nd that sampling depth was a signi cant predictor of initial SOC values in our dataset. We therefore
opted to test initial SOC effects using raw SOC values.
Data Analysis
We collected sampling variances when available to assign weights to data points. However, only 30% of
studies reported some form of variance. Following previous work, we chose instead to weight our
observations using sample size of the treatment and control groups which gave high weight to larger,
well-replicated studies
32–34
. Our weighting formula (Eq. 1) includes the common weighting ratio based on
treatment group sample size ( ) and control group sample size ( ) as well as a correction term dividing
by the total number of observations contributed by a given study (N). This additional step is meant to
ensure that no study contributes a disproportionate amount to the nal model simply because it
contained more extractable data points than another
35
.
(Eq. 1)
We modeled study site as a random effect to account for the non-independence of these data points, and
nested sampling year within study site to account for temporal non-independence. To build models for
both
Δ
SOC
and
Δ
Yield
, we implemented a model selection process which utilized Akaike Information
Criterion (AIC)
36
scores to select nal predictors which we had hypothesized may be mechanistically
related to
Δ
SOC
or
Δ
Yield
. Variable relevance was determined by comparing weighted mixed effect models
of each variable as a solitary predictor of each response variable against the corresponding model
containing only the intercept. Because of incomplete data for certain predictor variables, model
comparisons between the solitary predictor and the intercept-only model were done using complete data
subsets for the solitary predictor. If the regression containing the solitary predictor variable resulted in an
AIC score more than two units below that of the intercept-only regression (i.e.,
Δ
AIC < 2), the variable was
included in our nal multiple regression model. We did not perform any further model selection because
complex model selection decisions are often subjective and can change results considerably
37
. For our
Δ
Yield
model, we tested interaction terms between
Δ
SOC
and soil texture metrics, as well as an interaction
between
Δ
SOC
and initial SOC (SOC concentration prior to cover cropping), as per previous ndings
5
.
Lastly, we tested interaction terms between cover crop type (legume vs non-legume) and yield predictor
variables whose effects we hypothesized may be in uenced by N xation such as N fertilization,
rotational complexity and
Δ
SOC
.
n
t
n
c
W =
×
n
t
×
n
c
n
t
+
n
c
1
N
Page 12
Page 12/20
In both models, we checked for collinearity among variables using generalized variance in ation factors
(GVIF) with the following adjustment to allow for comparability across variables with differing degrees of
freedom
38
(df):
. We considered adjusted GVIF values of 3 and higher
to indicate potential collinearity
39
. Variables with foreseeable problems of collinearity such as mean
annual temperature and precipitation, absolute latitude, and aridity index were assessed separately in
multiple regressions on the basis of AIC.
All analyses were performed using R Statistical Software v4.2.0
40
. We built mixed effect regressions
using the package ‘lme4’
41
and determined xed effect F-values using a type III ANOVA in the ‘stats’
package
40
. We used the package ‘emmeans’ to quantify interaction effects
42
. We used pairwise
comparison in the package ‘emmeans’ to determine signi cant differences among levels of categorical
variables using a = 0.05 with a Bonferroni adjustment for multiple comparisons
43
. To determine the
signi cance of different levels of our moderating factors, we checked to see whether their 95% con dence
intervals (95% CI) overlapped zero, with no overlap indicating a rejection of the null (zero effect) at a =
0.05. We used the Kenward-Roger approximation for denominator degrees of freedom in all p-value
calculations
44
.
Declarations
Author contributions
I.V., T.M.B., and L.P. conceived the ideas and designed methodology; I.V. and G.D.L.C. collected the data;
I.V., A.G., and K.E. analyzed the data; I.V. and T.M.B. led the writing of the manuscript. All authors
contributed critically to the drafts and gave nal approval for publication.
Acknowledgments
We thank the many researchers whose studies supported this meta-analysis.
Funding
We thank the Organic Center for providing funding to T.B. and I.V. for this project. T.B. also acknowledges
USDA NIFA Agriculture and Food Research initiative (Grant 2017- 67013-26254).
AdjustedGV IF
= (
GV IF
)
1
2
df
Page 13
Page 13/20
Con ict of Interest
The authors have no con icts of interests related to this work.
Data Availability
Data used in this meta-analysis are available from the Dryad Digital Repository, accession.
References
1. Kibblewhite, M. G., Ritz, K. & Swift, M. J. Soil health in agricultural systems. Philosophical
Transactions of the Royal Society B: Biological Sciences
363
, 685–701 (2008).
2. Janzen, H. H., Janzen, D. W. & Gregorich, E. G. The ‘soil health’ metaphor: Illuminating or illusory? Soil
Biology and Biochemistry
159
, (2021).
3. Bronick, C. J. & Lal, R. Soil structure and management: a review. Geoderma
124
, 3–22 (2005).
4. Robertson, P.
et al. Farming for Ecosystem Services: An Ecological Approach to Production
Agriculture. BioScience
64
, 404–415 (2014).
5. Old eld, E. E., Bradford, M. A. & Wood, S. A. Global meta-analysis of the relationship between soil
organic matter and crop yields. SOIL
5
, 15–32 (2019).
. Manlay, R. J., Feller, C. & Swift, M. J. Historical evolution of soil organic matter concepts and their
relationships with the fertility and sustainability of cropping systems. Agriculture, Ecosystems &
Environment
119
, 217–233 (2007).
7. Lal, R. Enhancing crop yields in the developing countries through restoration of the soil organic
carbon pool in agricultural lands. Land Degradation & Development
17
, 197–209 (2006).
. Sanderman, J., Hengl, T. & Fiske, G. J. Soil carbon debt of 12,000 years of human land use.
Proceedings of the National Academy of Sciences
114
, 9575–9580 (2017).
9. Sherwood, S. & Uphoff, N. Soil health: research, practice and policy for a more regenerative
agriculture. Applied Soil Ecology
15
, 85–97 (2000).
10. Zomer, R. J., Bossio, D. A., Sommer, R. & Verchot, L. v. Global Sequestration Potential of Increased
Organic Carbon in Cropland Soils. Scienti c Reports
7
, 15554 (2017).
11. McDaniel, M. D., Tiemann, L. K. & Grandy, A. S. Does agricultural crop diversity enhance soil microbial
biomass and organic matter dynamics? A meta-analysis. Ecological Applications
24
, 560–570
(2014).
12. Kallenbach, C. M., Grandy, A. S., Frey, S. D. & Diefendorf, A. F. Microbial physiology and necromass
regulate agricultural soil carbon accumulation. Soil Biology and Biochemistry
91
, 279–290 (2015).
13. Poeplau, C. & Don, A. Carbon sequestration in agricultural soils via cultivation of cover crops – A
meta-analysis. Agriculture, Ecosystems & Environment
200
, 33–41 (2015).
Page 14
Page 14/20
14. Bradford, M. A.
et al. Soil carbon science for policy and practice. Nature Sustainability
2
, 1070–1072
(2019).
15. Esquivel, K. E.
et al. The “Sweet Spot” in the Middle: Why Do Mid-Scale Farms Adopt Diversi cation
Practices at Higher Rates? Frontiers in Sustainable Food Systems
5
, (2021).
1 . Marcillo, G. S. & Miguez, F. E. Corn yield response to winter cover crops: An updated meta-analysis.
Journal of Soil and Water Conservation
72
, 226–239 (2017).
17. Daryanto, S., Fu, B., Wang, L., Jacinthe, P.-A. & Zhao, W. Quantitative synthesis on the ecosystem
services of cover crops. Earth-Science Reviews
185
, 357–373 (2018).
1 . Fageria, N. K., Baligar, V. C. & Bailey, B. A. Role of Cover Crops in Improving Soil and Row Crop
Productivity. Communications in Soil Science and Plant Analysis
36
, 2733–2757 (2005).
19. Chabbi, A.
et al. Aligning agriculture and climate policy. Nature Climate Change
7
, 307–309 (2017).
20. Lal, R. Beyond Copenhagen: mitigating climate change and achieving food security through soil
carbon sequestration. Food Security
2
, 169–177 (2010).
21. Lehmann, J. & Kleber, M. The contentious nature of soil organic matter. Nature
528
, 60–68 (2015).
22. Oelofse, M.
et al. Do soil organic carbon levels affect potential yields and nitrogen use e ciency? An
analysis of winter wheat and spring barley eld trials. European Journal of Agronomy
66
, 62–73
(2015).
23. Schjønning, P.
et al. The Role of Soil Organic Matter for Maintaining Crop Yields: Evidence for a
Renewed Conceptual Basis. in vol. 150 35–79 (Advances in Agronomy, 2018).
24. Pan, G., Smith, P. & Pan, W. The role of soil organic matter in maintaining the productivity and yield
stability of cereals in China. Agriculture, Ecosystems & Environment
129
, 344–348 (2009).
25. Celestina, C., Hunt, J. R., Sale, P. W. G. & Franks, A. E. Attribution of crop yield responses to
application of organic amendments: A critical review. Soil and Tillage Research
186
, 135–145
(2019).
2 . Bowles, T. M.
et al. Long-Term Evidence Shows that Crop-Rotation Diversi cation Increases
Agricultural Resilience to Adverse Growing Conditions in North America. One Earth
2
, 284–293
(2020).
27. Ponisio, L. C.
et al. Diversi cation practices reduce organic to conventional yield gap. Proceedings of
the Royal Society B: Biological Sciences
282
, 20141396 (2015).
2 . McClelland, S. C., Paustian, K. & Schipanski, M. E. Management of cover crops in temperate climates
in uences soil organic carbon stocks: a meta-analysis. Ecological Applications
31
, e02278 (2021).
29. Cotrufo, M. F., Wallenstein, M. D., Boot, C. M., Denef, K. & Paul, E. The Microbial E ciency-Matrix
Stabilization (MEMS) framework integrates plant litter decomposition with soil organic matter
stabilization: do labile plant inputs form stable soil organic matter? Global Change Biology
19
, 988–
995 (2013).
30. Jilling, A.
et al. Minerals in the rhizosphere: overlooked mediators of soil nitrogen availability to
plants and microbes. Biogeochemistry
139
, 103–122 (2018).
Page 15
Page 15/20
31. Daly, A. B.
et al. A holistic framework integrating plant-microbe-mineral regulation of soil bioavailable
nitrogen. Biogeochemistry
154
, 211–229 (2021).
32. Pittelkow, C. M.
et al. Productivity limits and potentials of the principles of conservation agriculture.
Nature
517
, 365–368 (2015).
33. Zhang, T., Chen, H. Y. H. & Ruan, H. Global negative effects of nitrogen deposition on soil microbes.
The ISME Journal
12
, 1817–1825 (2018).
34. Chen, C., Chen, H. Y. H., Chen, X. & Huang, Z. Meta-analysis shows positive effects of plant diversity
on microbial biomass and respiration. Nature Communications
10
, 1332 (2019).
35. Ma, Z. & Chen, H. Y. H. Effects of species diversity on ne root productivity in diverse ecosystems: a
global meta-analysis. Global Ecology and Biogeography
25
, 1387–1396 (2016).
3 . Akaike, H. Maximum likelihood identi cation of Gaussian autoregressive moving average models.
Biometrika
60
, 255–265 (1973).
37. Hobbs, N. T. & Hilborn, R. Alternatives To Statistical Hypothesis Testing In Ecology: A Guide To Self
Teaching. Ecological Applications
16
, 5–19 (2006).
3 . Fox, J. & Monette, G. Generalized Collinearity Diagnostics. J Am Stat Assoc
87
, 178–183 (1992).
39. Zuur, A. F., Ieno, E. N. & Elphick, C. S. A protocol for data exploration to avoid common statistical
problems. Methods in Ecology and Evolution
1
, 3–14 (2010).
40. R Core Team. R: A Language and Environment for Statistical Computing. (2022).
41. Bates, D., Mächler, M., Bolker, B. & Walker, S. Fitting Linear Mixed-Effects Models Using lme4. Journal
of Statistical Software
67
, (2015).
42. Lenth, R. emmeans: Estimated Marginal Means, aka Least-Squares Means. (2022).
43. Dunn, O. J. Multiple Comparisons among Means. J Am Stat Assoc
56
, 52–64 (1961).
44. Kenward, M. G. & Roger, J. H. Small Sample Inference for Fixed Effects from Restricted Maximum
Likelihood. Biometrics
53
, 983–997 (1997).
Figures
Page 16
Page 16/20
Figure 1
Scatterplot of co-occurring SOC and yield changes from cover crops (n = 434). In 59.7% of observations,
SOC increases occurred alongside yield increases from cover crops (i.e., provided co-bene ts). In 20.7% of
observations, cover crops increased yield, but decreased SOC, and in 12.9% cover crops decreased SOC
but increased yield (trade-offs). Cover crops decreased both SOC and yield in 6.7% of data points (co-
costs). 2 outlier points have been cut off from the limits of this graph to aid in visualization of the data.
Page 17
Page 17/20
Figure 2
Yield change associated with a 10% increase in SOC (e.g., from 5 g kg
-1
to 5.5 g kg
-1
) at differing levels of
initial SOC. Initial SOC is SOC (g kg
-1
) prior to cover cropping (0-18.4 cm depth on average). Shaded
bands are 95% CIs. Increased SOC is positively associated with yield (red) only in sites with below
average initial SOC. 90% of observations fell within the initial SOC range shown.
Page 18
Page 18/20
Figure 3
Cash crop yield change from cover cropping at different levels of rotational complexity (“rotations”) and
N fertilizer (kg N ha
-1
season
-1
) in our yield model. Selected N fertilizer levels are dataset mean ± sd with
low, average, and high N corresponding to 12.9, 85.9, 158.9 kg N ha
-1
season
-1
, respectively. Rotational
complexity (“Rotations”) is a count of the number of different cash crop species rotated on a given plot
across the length of the experiment. Yield change estimates are shown for both legume and non-legume
cover crops. Letters are pairwise comparison results with different letters indicating signi cantly different
effect sizes at
. Numbers in parentheses are observations in each grouping (not included for N fertilizer because
displayed estimates correspond to selected values along a continuous axis rather than groupings). Error
bars are 95% CIs.
Page 19
Page 19/20
Figure 4
SOC change from both cover crop types (legume vs non-legume) as well as SOC change estimates at
selected values of absolute latitude in our yield model. Absolute Latitude is the absolute value of study
latitude. Initial SOC is SOC prior to cover cropping (selected values are dataset mean ± sd with low,
average, high corresponding to 6.3 g kg
-1
, 15.5 g kg
-1
, 24.7 g kg
-1
, respectively). Cover Crop Type is binary
categorical; non-legume vs legume. Letters are pairwise comparison results with different letters
indicating signi cantly different effect sizes at
. Numbers in parentheses are observations in each grouping (not included for absolute latitude and initial
SOC because displayed estimates correspond to selected values along a continuous axis rather than
groupings). Error bars are 95% CIs.
Page 20
Page 20/20
Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.
AppendixForSubmitting.docx
ArticleInfoForSubmitting.xlsx
ExtractionNotesForSubmitting.xlsx