Rabot et al. (2024). Relevance of the organic carbon to clay ratio as a national soil health indicator

www.elsevier.com/locate/geoderma

Relevance of the organic carbon to clay ratio as a national soil

health indicator

Eva Rabot

Isabelle Cousin

INRAE, Info

&

Sols, 45075 Orl

´

eans, France

b

Laboratoire de G

´

eologie, UMR 8538, Ecole Normale Sup

´

erieure, PSL Research University, CNRS, Paris 75005, France

c

UMR ECOSYS, INRAE, AgroParisTech, Universit

´

e Paris Saclay, 91123 Palaiseau, France

d

Agriculture and Agri-Food Canada, Qu

´

ebec City, Canada

A R T I C L E I N F O

Handling Editor: L. Morgan Cristine

Keywords:

Soil organic carbon

Clay content

Soil structure

A B S T R A C T

The soil organic carbon (SOC) to clay-sized particles ratio (SOC/clay) has recently been selected as an indicator

of the soil organic matter status in managed mineral soils within the framework of the European Soil Monitoring

Law proposal. This indicator was initially developed to predict soil structural quality in a local study in

Switzerland and was subsequently tested at national scales in England and Wales, and in Germany. In this study,

we evaluated if the SOC/clay ratio was relevant to assess the structural quality of soils at the national scale in

mainland France. We additionally evaluated its variant, SOC/(silt

<

20

µ

m

+

clay). We confronted SOC/clay and

SOC/(silt

<

20

µ

m

+

clay) to two indicators of soil structure, the soil bulk density and aggregate stability, and we

tested the effect of land use and soil type using information from the French Soil Quality Monitoring Network

(RMQS). We showed that the SOC/clay and SOC/(silt

<

20

µ

m

+

clay) were poor indicators of the soil bulk

density and aggregate stability. In our analysis, the SOC content was the best indicator of soil structure. Both land

use and soil type had an effect on the SOC/clay value. SOC/clay was found to be strongly affected by soil pH with

acidic soils consistently being classified as healthy according to the threshold of 1/13 and alkaline soils often

being classified as unhealthy. The domain of applicability of SOC/clay excludes soils involving other SOC sta-

bilization mechanisms than associations with the clay fraction and climate is not taken into account. Based on the

RMQS dataset, 63 % of cropland, 81 % of permanent crop and 23 % of grassland soils were below the SOC/clay

threshold of 1/13, which would classify them as unhealthy according to the European Soil Monitoring Law. We

questioned the relevance of the SOC/clay ratio and its proposed threshold of 1/13 as a soil structure indicator,

and more broadly as an indicator of the SOC status of healthy soils for all European pedoclimatic contexts. The

Soil Monitoring Law leaves the possibility of using correction factors for specific soil types or climatic conditions,

which appears necessary for France, because some pedoclimatic contexts will never allow a satisfactory value to

be reached.

1. Introduction

There is an increasing demand for simple indicators to assess soil

health, from the field scale for farmers to evaluate the effect of their

management practices, to the national scale for countries to report the

current status of their soils. In particular, the soil organic carbon (SOC)

to clay-sized particles ratio (SOC/clay) has been selected as an indicator

of SOC status in managed mineral soils (i.e., soils with SOC content

<

20

%) at the European Union level, to be used within the framework of the

European Soil Monitoring Law proposal (COM(2023) 416 final). To be

considered in a healthy condition, mineral soils must present a SOC/clay

greater than 1/13, with the possibility of a correction factor being

applied for specific soil types or climatic conditions. Although it is

referred to as a

“

loss of SOC

”

indicator in the Soil Monitoring Law, the

SOC/clay ratio has originally been developed as an indicator of soil

structural quality. Because the fine fraction (clay or silt

+

clay) was

observed to contribute to SOC protection either directly via organo-

mineral interactions, or indirectly via the formation of aggregates (

von

* Corresponding author.

E-mail address:

eva.rabot@inrae.fr

(E. Rabot).

Contents lists available at

ScienceDirect

Geoderma

journal homepage:

https://doi.org/10.1016/j.geoderma.2024.116829

Received 20 September 2023; Received in revised form 9 February 2024; Accepted 18 February 2024

Geoderma 443 (2024) 116829

2

Lützow et al., 2006

), it seems reasonable to explore whether this ratio

can be used as an indicator of the SOC status, for example like in

Dupla

et al. (2021)

The rationale for using the SOC/clay indicator to assess soil structure

originates from

Dexter et al. (2008)

who developed the concept of

“

complexed organic carbon

”

(COC). COC is assumed to reflect the

fraction of SOC associated with clay-sized particles. This concept implies

that SOC is entirely complexed by clay when the SOC content is lower

than clay/

n

, with

n

close to 10 for the studied soils (here, clay and SOC

contents in mass units). When the SOC content is lower than clay/

n

, a

part of the clay fraction is considered not to be associated with carbon

and contributes to the measured values of dispersible clay (

). Indeed,

and

Getahun et al. (2016)

found a correlation between the calculated non-

complexed clay content and the measured dispersible clay content.

High values of dispersible clay denote a low structural stability in water

(

Czy

˙

z and Dexter, 2015

). Conversely, when SOC content is higher than

clay/

n

, the COC concept implies that a fraction of SOC is not complexed

and would occur in the form of particulate organic matter and that soils

have reached their minimum level of dispersible clay. In other words,

the theory of

Dexter et al. (2008)

postulates that 10 g clay allow com-

plexing 1 g SOC for the studied soils. Therefore, several authors inves-

tigated if a SOC/clay value of 1/10 could be a target value for good soil

structure.

concluded that SOC/clay ratios of 1/

8, 1/10 and 1/13 were appropriate thresholds to distinguish very good,

good, moderate and degraded soil structures as determined by a visual

evaluation of soil structure (CoreVESS method,

Johannes et al., 2017b

).

The same conclusion was reached by

Johannes et al., 2017a; Poeplau and

, using an index

of soil structural quality based on the shape and size of aggregates and

soil texture, although their statistical analysis using boxplots indicated

some overlaps between the different classes of soil structural quality.

De

Jonge et al. (2009)

found that soils with SOC/clay

>

1/10 showed

favorable tilth conditions, contrary to soils with SOC/clay

<

1/10.

Dupla

et al. (2021)

referred to the SOC/clay ratio as the

“

structure vulnera-

bility indicator

”

and calculated the amount of SOC necessary to reach

the ratio of 1/10. In these studies, an effect of land use has been iden-

tified on the proportion of soils classified as degraded, croplands

exhibiting a higher proportion of sites with SOC/clay

<

1/13 as

compared to grasslands and forests (

Don, 2023; Prout et al., 2021

).

The SOC/clay indicator has been used at the national scale in Poland

and in Northern France (

Dexter et al., 2008

), in England and Wales

(

Prout et al., 2022, 2021

), in Germany (

Dupla et al., 2021; Guillaume et al., 2022;

), at the

regional scale in Switzerland (

Johannes et al., 2023, 2017a

) and in England (

Pulley et al., 2023

), or at

the plot scale in Denmark (

de Jonge et al., 2009; Getahun et al., 2016;

) or in Turkey (

). However,

divergent conclusions were drawn about its relevance as an indicator of

soil structural quality. In their study at the national scale,

Prout et al.

(2021)

used the threshold values developped in a local study in

Switzerland by

Rourke et al., 2015; Wiesmeier et al., 2019

validated the

threshold values for England and Wales and further suggested that they

may apply in similar climate zones across Europe. Conversely,

Poeplau

and Don (2023)

considered that the SOC/clay ratio was not a satisfac-

tory indicator in their national study of croplands and grasslands in

Germany, with the example of Chernozems

–

whose structure is known

to be good

–

mostly classified as having a moderate to degraded struc-

ture according to the SOC/clay indicator.

Given all the factors which were identified to control SOC dynamics

at different spatial scales (

O

’

),

it is necessary to further verify if the SOC/clay indicator is relevant to

assess soil structural quality at a national scale in other pedoclimatic

contexts. Since SOC may also be associated with the fine silt fraction

(

von Lützow et al., 2006

), this analysis was extended to its variant, the

SOC/(silt

<

20

µ

m

+

clay) ratio (

Arbor et al., 2023; Le Bissonnais and Arrouays, 1997; Nasta et al.,

nning et al., 2012

). The use of

SOC/clay or SOC/(silt

<

20

µ

m

+

clay), hereafter referred to as SOC

indicators, to assess the soil structural quality imply that they show a

strong relationship with indicators of soil structure. We tested this hy-

pothesis at the national scale in mainland France by using information

from the French Soil Quality Monitoring Network (RMQS). We used two

quantitative indicators of soil structure, bulk density and aggregate

stability, representing two different structural scales (soil samples of a

few hundred cm

3

vs. ten mm

3

) with implications on the range of pro-

cesses involved in soil structure formation and stabilization. Both bulk

density and aggregate stability are affected by SOC and clay contents (e.

g.,

2020; Regelink et al., 2015

). Given the diversity of situations covered by

the RMQS and the expected significant differences in soil structure be-

tween land uses and soil types, we further tested whether there was an

effect of land use and soil type on SOC/clay and SOC/(silt

<

20

µ

m

+

clay). As the SOC/clay ratio is one of the soil health indicators chosen by

the European Commission in the European Soil Monitoring Law, we

quantified the proportion of soils that would be in an unhealthy con-

dition in France, according to the proposed threshold of SOC/clay.

2. Material and methods

2.1. The RMQS database

We used topsoil data from the French Soil Quality Monitoring

Network (RMQS,

https://doi.org/10.15454/QSXKGA)

. This network

represents about 2,200 sites sampled following a 16

×

16 km

2

regular

grid across mainland France. At each site, a first set of measurements

was made using a composite sample based on 25 individual cores,

sampled with a soil auger from 0 to 30 cm, according to an unaligned

random sampling design within a 20

×

20 m

2

area (

Jolivet et al., 2022

Core samples were then bulked and the resulting composite samples

were air-dried and sieved to 2 mm before analysis. Particle-size distri-

bution in five fractions (

<

2, 2

–

20, 20

–

50, 50

–

200, 200

–

2,000

µ

m;

pipette method associated with wet sieving, without removing carbon-

ates, NF X31-107), pH in water (1/5 suspension of soil in water, ISO

10390), calcium carbonate content (CaCO

3

; volumetric method, ISO

10693), SOC content (elemental analysis after dry combustion and

correction for CaCO

3

content, ISO 10694) and cation exchange capacity

(CEC; hexamminecobalt(III) chloride method, NF X31-130) were

analyzed, among others (

Jolivet et al., 2006

). The second set of obser-

vations was collected on an adjacent pit, where soil bulk density was

measured according to the ring method or the excavation method on

three samples distributed vertically between 0 and 30 cm depth (

Jolivet

et al., 2022

). Because the bulk density of the fine earth was not

measured, we only used bulk density values of sites with rock fragment

content

<

20 % in the analysis involving bulk density (1,345 sites). In

the following, the land uses of the RMQS sites were simplified into four

classes: croplands, permanent crops (i.e., orchards, vineyards, olive

groves), grasslands and forests (

Fig. 1

). A few sites fell outside of these

classes and were discarded from the analysis. All the remaining sites

could be considered as mineral soils (SOC content

<

20 %). In the

database, soils of the RMQS are named following the French classifica-

tion system (

Baize and Girard, 2008

). In this study, we grouped them

into ten classes, differing in terms of SOC stabilization mechanisms. The

correspondence in the WRB classification (

IUSS Working Group WRB,

2022

) is as follows: Al-dominated soils [Podzols, Andosols, Umbrisols

except those with qualifier Folic], Ca/Mg-dominated soils [Rendzic

Leptosols, Cambisols with qualifiers Calcaric, Hypereutric, Dolomitic or

Magnesic], organic matter-rich soils [Histosols, Folic Umbrisols], water-

saturated soils [Gleysols, Stagnosols, Planosols], clay-rich soils [Rego-

sols (Clayic), Cambisols (Clayic), Vertisols], sandy or stony soils [Are-

nosols, soils with qualifier Skeletic], tidal water-affected soils [Tidalic

Fluvisols], other Cambisols [Cambisols except those with qualifiers

Calcaric, Hypereutric, Dolomitic, Magnesic or Clayic] and Luvisols. Soils

with strong human influence [Anthrosols and Technosols] and soils

having Solimovic material were excluded from the analysis implying

E. Rabot et al.

Geoderma 443 (2024) 116829

3

grouped soil types.

On a subset of 174 RMQS sites, aggregate stability was measured

using the fast wetting method of

Le Bissonnais (1996)

(ISO 10930) on

additional soil samples collected at the surface of the soil profiles

(

Fig. 1

B). The subset included 102 croplands, 6 permanent crops, 36

grasslands and 30 forests. Following this method, air-dried samples were

manually broken and manually sieved to keep aggregates of 3 to 5 mm in

diameter. Aggregates were dried at 40

◦

C for 24 h just before starting the

aggregate stability test. 5 g of aggregates were immersed in deionized

water for 10 min. The fragment size distribution was then measured first

using a sieve of 50

µ

m immersed in ethanol. The

>

50

µ

m fraction was

oven-dried and manually dry sieved using 2,000

µ

m, 1,000

µ

m, 500

µ

m,

200

µ

m, 100

µ

m and 50

µ

m sieves. Each size fraction was weighted and

the result of the test was expressed as the mean weight diameter (MWD).

MWD was calculated by summing the product of the mean diameter of

each size class by the relative proportion of aggregates in that size class.

The test was replicated three times for each RMQS site and the MWD

value was averaged. The larger the MWD, the higher the aggregate

stability: very unstable for MWD

<

0.4 mm, unstable for MWD between

0.4 and 0.8 mm, medium stability for MWD between 0.8 and 1.3 mm,

stable for MWD between 1.3 and 2.0 mm and very stable for MWD

>

2.0

mm (

).

2.2. Data analyses

Four indicators were tested in relation to soil structural properties:

the SOC content, clay content, SOC/clay ratio and SOC/(silt

<

20

µ

m

+

clay) ratio. Simple linear regressions or non-linear regressions were used

to highlight potential relationships with soil structural properties for

each land use class. We evaluated regression quality using the root-

mean-square error (RMSE, eq.

1

):

RMSE

=

̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅

1

n

∑

n

i

=

1

(

y

i

−

̂

y

i

)

2

√

(1)

where

y

i

and

ˆ

y

i

are measured and predicted values for site

i

and

n

is the

number of observations. For simple linear regressions, we also used the

coefficient of determination (

r

2

). Fits were only shown in the figures

when

r

2

>

0.55 and RMSE

<

1 for linear regressions and when a trend

was sufficiently marked to choose an appropriate function for non-linear

regressions. Outlier detection was used to better understand the domain

of applicability of the SOC indicators and was performed for linear re-

lationships on each land use class. Outlier detection used the Cook

’

s

distance method of the R package

“

performance

”

(

Lüdecke et al., 2021

When detected, outliers were not removed from the analyses. The effect

of the four land uses and soil types on soil properties and SOC indicators

were assessed by using the nonparametric test of Kruskal-Wallis, fol-

lowed by Dunn

’

s multiple comparisons with Bonferroni correction, at

the

p

=

0.001 level. Only grouped soil types with more than 20 sites were

included in the Kruskal-Wallis test. All statistical analyses were per-

formed using the R software (

R Core Team, 2022

3. Results and discussion

3.1. Link between the SOC indicators tested and soil structure indicators

3.1.1. Bulk density

There were no significant relationship between SOC/clay or SOC/

(silt

<

20

µ

m

+

clay) and bulk density for sites with rock fragment

content

<

20 % in croplands and forests (

A and B). For grasslands,

points were distributed in a vertical fashion for SOC/clay

<

1/13 or

SOC/(silt

<

20

µ

m

+

clay)

<

1/20 (

A and B). For higher values of

SOC/clay or SOC/(silt

<

20

µ

m

+

clay), the distribution became more

horizontal and the dispersion increased. A function of the form

y

=

1/(

a

×

x

)

+

b

could be fitted for grasslands, with a RMSE of 0.28 using SOC/

clay as the explanatory variable and a RMSE of 0.25 using SOC/(silt

<

20

µ

m

+

clay). A fair decreasing linear relationship was found with clay

content in croplands, grasslands and forests (

C). Bulk density was

better described by SOC content, where a function of the form

y

= −

a

×

ln(

x

)

+

b

could be fitted for croplands, grasslands and forests (

D).

Using SOC content as the explanatory variable, the RMSE was 0.16 for

forests and 0.12 for croplands and grasslands. It has to be noted that in

this study, we did not attempt to find the best possible pedotransfer

function to predict bulk density, which might depend on other soil

properties and textural class.

The thresholds of 1/8, 1/10 and 1/13 suggested by

and the threshold of 1/20 suggested by

were represented as vertical dotted lines in

A and B. These

thresholds did not appear to be effective in linking SOC/clay nor SOC/

(silt

<

20

µ

m

+

clay) to bulk density. Indeed, the distribution being very

vertical around these thresholds for each land use, the bulk densities

spanned for SOC/clay

=

1/8 were similar to those spanned for SOC/clay

=

1/13.

In their study,

Dexter et al. (2008)

used a subset of the RMQS dataset

corresponding to topsoils of croplands and grasslands in Northern

Fig. 1.

(A) RMQS sites simplified in four land use classes and (B) RMQS sites

with aggregate stability measurements.

E. Rabot et al.

Geoderma 443 (2024) 116829

4

France. They found that the inverse of bulk density was better explained

by SOC content for croplands because of their low levels of SOC. The

inverse of bulk density was better explained by clay content for grass-

lands because of their greater SOC contents. Our results did not support

this finding since bulk density appeared to be explained by both the SOC

and clay contents when considering the whole mainland France (

Johannes et al. (2023, 2017a)

C

and D).

, for soils with a large range of clay

contents, also found that SOC content was a better explanatory variable

of soil bulk density and of some other physical properties (i.e., gravi-

metric water content and air content at

−

100 hPa) than SOC/clay.

3.1.2. Aggregate stability

A linear relationship was found between aggregate stability (MWD)

and SOC/clay and SOC/(silt

<

20

µ

m

+

clay) in croplands and perma-

nent crops and in grasslands (

A and B). The quality of the regres-

sion was better when considering only SOC content or clay content for

croplands and permanent crops (

r

2

=

0.78, RMSE

=

0.41 for both,

C and D). All indicators seemed less efficient for use in grasslands

and forests than in croplands and permanent crops, meaning that the

link with MWD could only be valid for croplands and permanent crops

(

). A higher content of particulate organic matter in forests and

grasslands may explain this observation. According to

Guillaume et al.

(2022)

, the fact that the amount of SOC in particulate organic matter

does not increase proportionally with the clay content may be a limi-

tation of using the SOC/clay ratio as an indicator of soil physical quality

in organic carbon-rich soils. Moreover, other soil properties than SOC

and clay contents would have to be taken into account to better model

MWD. Clay mineralogy, the content, type and concentration of cations,

and sesquioxide and CaCO

3

contents are also known to affect aggregate

stability (

The thresholds of 1/8, 1/10 and 1/13 suggested by

and the threshold of 1/20 suggested by

Fig. 2.

Bulk density as function of (A) soil organic carbon to clay ratio, (B) soil organic carbon to silt

<

20

µ

m

+

clay ratio, (C) clay content and (D) soil organic

carbon content, for topsoils of the RMQS dataset with rock fragment content

<

20 %. Functions plotted in blue correspond to (A and B)

y

=

1/(

a

×

x

)

+

b

, (C)

y

=

a

×

x

+

b

and (D)

y

= −

a

×

ln(

x

)

+

b

. Vertical dotted lines represent the thresholds of 1/8, 1/10 and 1/13 suggested in

and the threshold of 1/20

in

. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

E. Rabot et al.

Geoderma 443 (2024) 116829

5

were represented as vertical dotted lines in

A and B. These

thresholds did not appear satisfying to discriminate different values of

MWD. This observation is in line with the large overlaps found in the

studies of

and

when repre-

senting their qualitative soil structural quality assessments as boxplots.

The same RMQS sites were identified as outliers for SOC/clay and

SOC/(silt

<

20

µ

m

+

clay) (

IUSS Working Group WRB, 2022

). One site under cropland is a Cambisol

(

) with a silty texture and a high SOC

content (SOC

=

47.3 g kg

−

1

, clay

=

167 g kg

−

1

, silt

=

548 g kg

−

1

, pH

=

5.9). Two additional outliers are very sandy forest soils, one being

classified as a Podzol and the other one as an Umbrisols (Hyperdystric)

(

IUSS Working Group WRB, 2022

). They have a high SOC content

relative to their clay content and a pH in water of 4.2. Two of these sites

were also identified as outliers for SOC content with the Cook

’

s distance

method, one in croplands and another one in forests (

).

3.1.3. Soil structure indicators

There is no universally accepted way to characterize soil structure

(

Díaz-Zorita et al., 2002

). In the original study of

Johannes et al., 2017b; Lin et al., 2022

, the SOC/clay ratio was compared with a semi-quantitative

indicator of the soil structural quality, the CoreVESS score (

Johannes

et al., 2017b

). The CoreVESS method is an adaptation for the laboratory

of the Visual Evaluation of Soil Structure (VESS) field method (

). CoreVESS is applied to smaller soil

samples than VESS. Therefore, only the evaluation criteria of the VESS

method that could be observed on soil cylinders of approximately 150

cm

3

were kept in CoreVESS. This method has not been widely used yet

but appeared to be appropriate to detect the degradation of soil struc-

tural porosity (

). In our national

scale study, we used two available quantitative indicators of soil struc-

ture, bulk density and aggregate stability (measured using a fast wetting

method). According to the literature, the CoreVESS method seems to be

related to these two indicators of soil structure. Indeed, an increasing

Fig. 3.

Mean weight diameter (MWD) as function of (A) soil organic carbon to clay ratio, (B) soil organic carbon to silt

<

20

µ

m

+

clay ratio, (C) clay content and (D)

soil organic carbon content, in the topsoil of the RMQS dataset. Solid blue lines represent linear regressions. Outliers are indicated by a circle. Vertical dotted lines

represent the thresholds of 1/8, 1/10 and 1/13 suggested in

and the threshold of 1/20 in

Cornelis et al., 2019; Johannes et al., 2017b; Lin et al.,

. (For interpretation of the

references to colour in this figure legend, the reader is referred to the web version of this article.)

E. Rabot et al.

Geoderma 443 (2024) 116829

6

linear relationship was consistently found between the CoreVESS score

and bulk density (

2022; Mutuku et al., 2021

). A negative relationship was also found with

the MWD, but in pedoclimatic contexts different than France (

Cornelis

et al., 2019; Mutuku et al., 2021

). For this reason, we think that using

bulk density and MWD to assess the relevance of SOC/clay and SOC/

(silt

<

20

µ

m

+

clay) was a reliable approach. In our study, bulk density

and MWD were found to be related to SOC and clay contents (

C and

D and

Le Bissonnais and Arrouays, 1997; Regelink et al.,

C and D), which was consistent with previous studies about

aggregate stability (

2015

) and bulk density (see pedotransfer functions cited in

Arbor et al.,

2023; Nasta et al., 2020

). In particular, the relationship between bulk

density and SOC content was non-linear and the relationship with clay

content was linear (

Johannes et al., 2017b, 2023

C and D). In the literature, a fair positive linear

relationship was found between the CoreVESS score and SOC content

across the CoreVESS range (

Cornelis et al., 2019

) or only for scores

ranging from 1 to 3 (i.e., high structural quality) (

Johannes et al.,

2017b

). A weak negative relationship (

Lin et al., 2022

) or no relation-

ship (

) was found with clay content.

Therefore, dividing the SOC content by the clay content may have little

influence on the results when SOC/clay is confronted to the CoreVESS

score as compared to SOC content alone. On the contrary, this division

strongly affects the results when looking for relationships between SOC/

clay and bulk density, since bulk density is related to SOC content by a

non-linear relationship and to clay content by a linear relationship.

Bulk density varies over time depending on agricultural field oper-

ations and depends on the soil water content during sampling. In the

RMQS dataset, bulk density was measured when the soil water content

was close to field capacity and when the soil has naturally compacted

after cultivation operations (

Jolivet et al., 2022

). In this respect, the

CoreVESS protocol is more rigorous, because it takes into account vol-

ume changes in swell-shrinking soils by first equilibrating free to swell

samples at a given matric potential. Aggregate stability tests also have

their limitations, the main being that the results are highly sensitive to

the protocol used (

Díaz-Zorita et al., 2002

). In the RMQS dataset, this

drawback was taken into account by following an ISO protocol for the

MWD determination. Using a different aggregate stability test could

have given different results in terms of MWD because different mecha-

nisms of aggregate breakdown, such as slaking, differential swelling or

raindrop impact, could be involved depending on the initial soil water

content, rate of wetting and energy applied (

). In the

fast wetting method used in this study, immersion in water of dry ag-

gregates mimics a heavy rain and favors aggregate breakdown by slak-

ing, i.e., aggregate breakdown caused by the compression of entrapped

air during wetting (

We found that SOC/clay and SOC/(silt

<

20

µ

m

+

clay) were not

good indicators of the soil bulk density and aggregate stability, while the

SOC content was a much better predictor. It leads us to question the use

of the SOC/clay ratio as one of the indicators of soil health in managed

ecosystems, as proposed in the Soil Monitoring Law proposal.

3.2. Current situation in France

3.2.1. Effect of land use on the SOC indicators tested and on soil properties

Descriptive statistics are presented in

Table 1

for each land use to

show the range of variation of different soil properties, SOC/clay and

Table 1

Overview of soil properties for four land uses in the topsoil of the RMQS dataset and its aggregate stability subset (SOC: soil organic carbon content, CEC: cation

exchange capacity, MWD: mean weight diameter). Different letters indicate significant differences among the four land uses.

RMQS dataset

Aggregate stability subset

Croplands

Permanent crops

Grasslands

Forests

Croplands

Permanent crops

Grasslands

Forests

Count

878

59

521

581

102

6

36

30

SOC

Mean

17

a

12

b

29

b

34

b

16

a

10

a

25

b

36

b

(g kg

−

1

)

Median

15

9

25

27

14

8

24

26

Min.

3

7

1

5

7

11

Max.

58

39

145

159

47

21

54

85

Clay

Mean

244

a

252

ab

262

a

228

b

257

a

286

a

252

a

278

a

(g kg

−

1

)

Median

217

223

214

189

242

298

213

240

Min.

19

56

62

2

81

125

84

87

Max.

700

591

798

815

613

463

643

646

SOC/clay

Mean

0.079

a

0.054

b

0.126

c

0.231

d

0.067

a

0.036

a

0.112

b

0.150

b

Median

0.067

0.045

0.107

0.139

0.058

0.036

0.102

0.114

Min.

0.012

0.028

0.037

0.029

0.021

0.043

0.042

Max

0.546

0.162

0.721

9.000

0.283

0.052

0.258

0.575

SOC/(silt

+

Mean

0.038

a

0.029

a

0.063

b

0.122

c

0.032

a

0.018

a

0.055

b

0.079

b

clay)

Median

0.032

0.023

0.054

0.070

0.028

0.017

0.049

0.063

Min.

0.008

0.009

0.014

0.022

0.012

0.011

0.027

Max.

0.444

0.087

0.334

6.000

0.125

0.028

0.128

0.310

pH water

Mean

7.0

a

7.6

a

6.3

b

5.6

c

7.1

a

7.7

a

6.2

ab

5.6

b

Median

7.0

8.1

5.9

4.9

7.3

8.3

5.9

4.9

Min.

4.5

4.9

4.5

3.7

5.1

6.0

4.9

4.0

Max.

8.6

8.7

8.6

8.5

8.3

8.6

8.3

8.0

CEC

Mean

14.0

a

14.6

a

14.5

a

13.4

b

15.2

a

14.2

a

12.8

a

15.4

a

(cmol

+

kg

−

1

)

Median

11.8

13.1

10.0

6.3

14.0

14.1

9.4

7.2

Min.

1.0

3.6

0.8

0.2

2.6

3.9

3.5

2.1

Max.

59.6

39.3

63.5

70.1

41.1

24.2

40.3

54.1

CaCO

3

Mean

72

a

111

a

35

b

34

b

58

a

178

a

40

a

36

a

(g kg

−

1

)

Median

1.0

23.5

0.5

1.4

189

0.5

Min.

0

Max.

866

451

706

739

866

416

646

531

Bulk

Mean

1.4

a

1.5

a

1.3

b

1.2

b

1.4

a

1.6

a

1.3

ab

1.1

b

density

Median

1.4

1.5

1.3

1.2

1.4

1.6

1.3

1.1

(g cm

−

3

)

Min.

0.8

1.0

0.4

0.5

0.8

1.4

0.9

0.7

Max.

2.0

1.9

1.8

1.7

1.6

MWD

Mean

0.7

a

0.5

a

1.8

b

2.2

b

(mm)

Median

0.6

0.4

1.9

2.4

Min.

0.2

0.6

0.5

Max.

2.2

1.4

3.2

3.3

E. Rabot et al.

Geoderma 443 (2024) 116829

7

SOC/(silt

<

20

µ

m

+

clay) in the RMQS dataset. We observed significant

differences at

p

<

0.001 between all land uses in the indicators SOC/clay

and SOC/(silt

<

20

µ

m

+

clay), except for SOC/(silt

<

20

µ

m

+

clay) in

croplands and permanent crops.

We used the thresholds of SOC/clay of 1/8, 1/10 and 1/13 suggested

by

to visualize SOC content as a function of clay

content for the four land uses (

). The majority of sites in croplands

and permanent crops had a SOC/clay ratio

<

1/13 (

). A majority of

forest sites had a SOC/clay

>

1/8. The grassland sites were more evenly

distributed between the different classes of SOC/clay. A majority of sites

in croplands and permanent crops presented a SOC/(silt

<

20

µ

m

+

clay)

ratio

<

1/20 (

). Conversely, a majority of forest sites had a SOC/

(silt

<

20

µ

m

+

clay)

>

1/20. The grassland sites were more evenly

distributed.

63 % of cropland, 81 % of permanent crop, 23 % of grassland and 13

% of forest soils showed a SOC/clay

<

1/13, so their soil structure would

be classified as degraded. The proportion of degraded cropland soils was

much higher than in other national-scale studies. In England and Wales,

croplands also exhibited a higher proportion of sites with SOC/clay

<

1/

13, but these sites represented only 38.2 % of the studied sites (

Prout

et al., 2021

). For ley grass, 15 % of the sites had a SOC/clay

<

1/13, 6.6

% for permanent grass and 5.6 % for forests (

Prout et al., 2021

). In

Germany, 37 % of cropland and 14 % of grassland soils were found to

have a SOC/clay

<

1/13 (

). Our data show that

grassland and forest soils have a smaller proportion of sites with SOC/

clay

<

1/13 than croplands. According to

, this trend

supports the use of SOC/clay

=

1/13 as an indicative threshold for soil

degradation, as grassland and forest soils are generally close to semi-

natural systems.

3.2.2. Effect of soil type on the SOC indicators tested

The Kruskal-Wallis test showed a significant difference on SOC/clay

and SOC/(silt

<

20

µ

m

+

clay) depending on our grouped soil types (

p

<

0.001). A summary of the post-hoc Dunn

’

s multiple-comparison test is

given in

Table 2

for SOC/clay. Distinct groups were found for each land

use, from a tendency to high SOC/clay ratios in Al-dominated soils or

Leptosols, to low SOC/clay ratios in Ca/Mg-dominated soils. Results

were different depending on the land use considered (

Table 2

). It shows

that there is an effect of both land use and soil type on the SOC/clay

value.

We represented the SOC content as a function of clay content with

the thresholds of 1/8, 1/10 and 1/13 suggested by

for each of the grouped soil types and land uses in

. The

majority of sites classified as Al-dominated soils and Leptosols had a

SOC/clay ratio

>

1/8 in all land uses (

). Podzols were also found to

have SOC/clay

>

1/8 in the national scale studies of England and Wales

(

Prout et al., 2021

) and Germany (

). Interactions

of SOC with aluminum and iron contributes to SOC stabilization in

Podzols (

von Lützow et al., 2006

), leading to high SOC content for a

given clay content. It is known that, in Andosols, the dominance of

highly reactive short-range-order minerals such as allophane also lead to

store high amounts of SOC (

ogel-Knabner and Amelung, 2021

Because of high agricultural constraints Al-dominated soils and Lep-

tosols were in majority not cultivated. SOC/clay was

<

1/13 in crop-

lands and permanent crops for only 8 % of Al-dominated soils and 13 %

Fig. 4.

Soil organic carbon content as a function of (A) clay content and (B) silt

<

20

µ

m

+

clay content in the topsoil of the RMQS dataset for different land uses. The

thresholds of 1/8, 1/10 and 1/13 are suggested in

and the threshold of 1/20 in

. On the right side of each plot, the

percentage of sites in each soil structural quality class is given, from the best soil structural quality at the top to the worst at the bottom.

Table 2

Summary of the Dunn

’

s multiple-comparison test for Kruskal-Wallis analysis of

variance on ranks. Groups of soil types followed by the same letter are not

significantly different at the

p

=

0.001 level with respect to their soil organic

carbon to clay ratio.

Croplands, permanent crops

Grasslands

Forests

Al-dominated soils

a

ab

a

Leptosols

abc

a

ab

Other Cambisols

ab

bc

Water-saturated soils

bc

abc

bcd

Luvisols

c

bc

cd

Ca/Mg-dominated soils

d

c

d

E. Rabot et al.

Geoderma 443 (2024) 116829

8

of Leptosols.

A high proportion of sites with SOC/clay

>

1/8 was also found in

organic matter-rich soils and sandy or stony soils, but only a few sites

represented these classes in our dataset (

). In the soil types grouped

in organic-matter rich soils (i.e., Histosols and Folic Umbrisols), an

external driver is responsible for the low mineralization and SOC

accumulation (e.g., climate, waterlogging), leading to high SOC/clay

ratios.

The sites were more evenly distributed between the different SOC/

clay classes for Cambisols, water-saturated soils, Luvisols and Ca/Mg-

dominated soils (

). These soil types were more often cultivated

than the previous ones, and croplands and permanent crops tended to

have the lowest SOC/clay ratios (SOC/clay

<

1/13 in croplands and

permanent crops for 48 % of Cambisols, 61 % of water-saturated soils,

65 % of Luvisols and 84 % of Ca/Mg-dominated soils). This observation

highlights a confounding effect between soil type and land use. Cam-

bisols are derived from a variety of parent materials, therefore spanning

a wide variety of soil properties such as pH or composition of clay-sized

minerals (

ogel-Knabner and Amelung, 2021

). It may explain the wide

range of SOC/clay ratios observed in Cambisols. Water-saturated soils

also span a wide variety of soil properties. SOC content has been

observed to increase with the level of waterlogging (

Amendola et al.,

2018; Meersmans et al., 2008; Poeplau et al., 2020

). In the water-

saturated soil class, we could observe that some grassland and forest

sites had some of the highest clay contents of the dataset and had SOC/

clay

<

1/13 (

).

A majority of sites classified as clay-rich soils and tidal water-affected

soils had a SOC/clay

<

1/13, but only a few sites represented these

classes (

). 100 % of clay-rich soils and tidal water-affected soils

under croplands and permanent crops had a SOC/clay

<

1/13. Clay-rich

Vertisols also showed SOC/clay

<

1/13 in Germany (

Poeplau and Don,

2023

Poeplau and Don (2023)

considered that it would take an

immense effort in terms of carbon input to move these clay-rich soils

from a degraded class to a moderate class as defined by the SOC/clay

ratio. Salinity in salt-affected soils is usually responsible for poor plant

growth leading to low carbon inputs into the soil and low SOC contents

(

Wong et al., 2010

). In addition, the presence of sodium ions may cause

clay dispersion, resulting in SOC losses (

Wong et al., 2010

). In the Soil

Monitoring law, the 1/13 threshold was chosen to be a target value to

achieve because it distinguished managed and semi-natural systems in

the study of

. We however showed that the 1/13

threshold does not allow this distinction for every type of soil.

Fig. 5.

Soil organic carbon content as a function of clay content in the topsoil of the RMQS dataset for different soil types and land uses. The thresholds of 1/8, 1/10

and 1/13 are suggested in

. On the right side of each plot, the percentage of sites in each soil structural quality class is given, from the best soil

structural quality at the top to the worst at the bottom.

E. Rabot et al.

Geoderma 443 (2024) 116829

9

One of the main soil properties discriminating different soil types is

soil pH. There is also a known control of pH on SOC stabilization (

Knabner and Amelung, 2021; Rowley et al., 2018

ogel-

), which is partly

responsible for the differences observed by comparing the acidic Al-

dominated soils to the alkaline Ca/Mg-dominated soils in

.

Therefore, we looked in more detail at the variation of SOC/clay with

soil pH (

). The majority of sampling sites had a SOC/clay

>

1/8 for

pH lower than 5, corresponding to a very good soil structure according

to the thresholds proposed by

IUSS Working Group WRB, 2022

. These sites

included Podzols, Cambisols (Hyperdystric) or Umbrisols (Hyper-

dystric), Luvisols and Fluvisols or Fluvic Cambisols developed on sandy

or silty materials (WRB classification,

).

The great majority was under forest land use (

). In the coarse soil

textures of these RMQS sites, when recorded from the soil profile, the

soil structure was single grained. It usually leads to classify soil structure

as poor.

also observed a tendency for high SOC/clay

ratios in acidic soils and a lower effect of soil pH in the range 5.5 to 7.

For pH

>

8 (and

≤

8.7, the maximum pH value found), the range of

SOC/clay values was smaller than in the rest of the pH range (

).

The mean and median clay contents of these sites were larger than in the

whole dataset. SOC/clay was the lowest (

<

1/40) for some RMQS sites

classified as Cambisols (Calcaric) or Renzic Leptosols (

IUSS Working

Group WRB, 2022

) under agricultural land use and a climate charac-

terized by a high annual average temperature in Southwest France. In

addition to clay contents higher than in the whole dataset, they showed

lower mean and median carbon contents than the whole dataset. RMQS

sites with SOC/clay

>

1/8, supposed to have a good structural quality

according to the SOC/clay ratio, were very rare for pH

>

8.2.

3.2.2. Map of healthy and unhealthy soils

Based on the RMQS dataset, 63 % of cropland, 81 % of permanent

crop and 23 % of grassland soils were below the threshold of SOC/clay of

1/13 proposed in the European Soil Monitoring Law (

). The

“

one

out

–

all out

”

principle chosen in this law proposal implies that this

single failing criterion is sufficient to classify these soils as unhealthy.

We sought to observe the geographical trends that would emerge from

the application of a SOC/clay threshold of 1/13 on the French managed

soils (

). If

found that soils with SOC/clay

<

1/13

were not located in a particular region in England and Wales, we found

known geographical patterns in France. The trends observed in our

study, by analyzing the effect of land use, soil type and soil pH on SOC/

clay, are clearly visible spatially at the national scale (

A and B).

Climate also seems to have an additional impact on the results, through

the altitude and latitude.

For example, some soils classified in a healthy condition are located

in the Brittany region and Massif Central (

C). Soils of Brittany are

silty, with low clay contents. They are developed from an eroded

crystalline basement, leading mainly to Cambisols with acid soil pH

(

B). Soils are used for croplands and grasslands (

A). This

region is also known to produce and use the largest quantities of organic

fertilizers in France, due to a developed livestock production (

Loyon,

2017

). The climate is marked by an oceanic influence with medium to

high precipitations (

). Massif Central corresponds to low

to medium elevation mountains, mainly covered by grasslands (

A).

It is also an eroded crystalline basement with some volcanic rocks

responsible for the presence of Andosols (Al-dominated soils,

B).

Soil textures are coarse. Massif Central shows a mountain climate with

low temperatures and high precipitations (

).

On the contrary, we can cite the example of some soils classified as

unhealthy, located in the coast of Languedoc-Roussillon and the Char-

entes regions (

C). In Languedoc-Roussillon, soils are mainly

developed from limestones and quaternary alluvial deposits leading to

alkaline pH (Ca/Mg-dominated soils,

B). They are highly culti-

vated as vineyards (

A). This permanent crop usually returns low

amounts of SOC to the soil. Climate is Mediterranean, with high tem-

peratures and low precipitations (

). The Charentes re-

gion is located in the northern part of the sedimentary basin of

Aquitaine. Soils are cultivated (

A), with a clayey texture and

developed from carbonate materials (Ca/Mg-dominated soils,

B).

The climate is marked by an oceanic influence, with medium to high

precipitations (

).

We have cited here extreme examples, combining several favorable

(Brittany, Massif Central) or unfavorable factors (Languedoc-Roussillon,

Charentes) for the SOC/clay indicator. These factors are not all related to

soil texture or management as taken into account by the SOC/clay in-

dicator. Climate, pH and mineralogy also affect the SOC/clay value. All

these factors also define agricultural constraints which were responsible

for the historical land use and management of soils in France. The fact

that geographical patterns are visible at this national scale supports the

idea of defining threshold values for assessing soil health by pedocli-

matic context, which we could refer to as

“

soil districts

”

in a European

legislative context. The definition of soil districts is, however, still

vague. For their definition, a trade-off will have to be made between the

pedoclimatic homogeneity of soil districts and the number of sampling

points necessary to define robust reference values.

In the national scale study of Germany,

Poeplau and Don (2023)

considered that the SOC/clay indicator was insensitive to changes in

SOC content in both coarse- and fine-textured soils. A problem in clay-

rich soils was also reported by

Çelik et al. (2020)

in Turkey. In a field

study in croplands with clay contents of 50 %, they found a SOC/clay

<

1/13 although the VESS score classified soil structure as good. In our

study, only 2 % of the cropland sites with a clay content

≥

35 % were

classified in a healthy condition, and they were not found further South

than the Charentes region (

C). We did not find healthy soils in

croplands for clay contents

≥

50 %.

Our study also underlined that the normalization of SOC content by

the fine fraction content provided an oversimplified view of the links

between SOC and soil structure. As shown when plotting SOC/clay as

function of soil pH (

), mechanisms such as sorption on aluminum

species cannot be taken into account using the SOC/clay indicator. This

mechanism is however dominant in some soil types (

Amelung, 2021; Rasmussen et al., 2018

ogel-Knabner and

). This is a major problem when

one wants to apply the SOC/clay indicator for a wide range of pedo-

climatic contexts as they exist at a national or European level. In the case

of France, removing soils for which other factors of SOC dynamics than

those related to the crystalline minerals found in the fine fraction are

involved would lead to exclude at least Al-dominated soils, organic

matter-rich soils, tidal water-affected soils and soils with pH

≤

5. These

soils represent only 6 % of the managed mineral soils in the RMQS

dataset. Removing these few sites from our analysis of the link between

SOC indicators and soil structure indicators for managed soils would not

modify our conclusions, i.e., SOC is a better indicator of soil structure

than SOC/clay or SOC/(silt

<

20

µ

m

+

clay).

Fig. 6.

Soil organic carbon to clay ratio (SOC/clay) as function of soil pH, in the

topsoil of the RMQS dataset. Horizontal dotted lines represent the thresholds

suggested in

.

E. Rabot et al.

Geoderma 443 (2024) 116829

10

It is unclear whether the high percentage of highly alkaline soils in an

unhealthy condition is realistic (81 % of the managed RMQS sites with

pH

≥

8,

et al., 2017a; Prout et al., 2021

). These soils are highly represented in France, since 23 %

of the managed RMQS sites have a pH

≥

8. In these soils, some other SOC

stabilization mechanisms are supposed to be involved (

Rowley et al.,

2018

) but in France, many confounding factors also affect their carbon

inputs into the soil and SOC mineralization. These confounding factors

make it difficult to draw a conclusion about the relevance of SOC/clay in

the case of soil pH

≥

8. A large part of managed alkaline soils is under

croplands or permanent crops (85 %), a large part is found under a

climate showing high annual average temperatures in Southern France

(22 %) and a large part is clayey (35 % with a clay content

≥

35 %). In

addition, it cannot be excluded that the use of a particle size analysis

without removing carbonates and leading to overestimates of the fine

fraction content could have influenced the resulting SOC/clay class of

alkaline soils. Additional investigations would be required.

The Soil Monitoring Law leaves the possibility of applying a

correction factor for specific soil types and climatic conditions. It seems

clear that an adaptation of the threshold is required for France, because

it will be nearly impossible to reach the threshold of 1/13 for some soils

under agricultural use, for example under a Mediterranean climate or for

clayey soil textures. As mentioned earlier, the role of clay particles on

SOC content is well known. However, climate, through its effect on

primary productivity and decomposition, is also a strong driver of SOC.

The results of

Chen et al. (2019)

for France and

Pacini et al. (2023)

for

Europe both suggest that, along with soil properties, the maximum

attainable SOC content is strongly driven by climate. These results

suggest that the SOC/clay threshold should take into account climatic

Fig. 7.

Map of the RMQS sites classified as unhealthy because of a SOC/clay value lower than 1/13 in the topsoil, symbolized by (A) land use and (B) grouped soil

type. Four natural or administrative regions mentioned in the text are represented in (C).

E. Rabot et al.

Geoderma 443 (2024) 116829

11

zones.

3.3. Explaining divergent conclusions about the SOC/clay indicator

Previous studies on the SOC/clay indicator have drawn divergent

conclusions about its relevance when confronted with soil structural

indicators, some authors being in favor of the use of SOC/clay (

Johannes

) and others not (

Johannes et al., 2023, 2017b; Lin et al., 2022

).

As presented earlier, CoreVESS is not very sensitive to clay content

(

), contrary to bulk density

and aggregate stability. We also believe that the use of qualitative

physical indicators, although very useful for field applications, to be

compared with SOC/clay was partly responsible for these divergent

conclusions. When qualitative indicators of soil structure were used to

investigate their link with SOC/clay, such as the CoreVESS score or an

index based on the shape and size of aggregates and soil texture, the

observed trends were considered satisfactory by

and

. However, when quantitative indicators of soil

structure were used, such as the inverse of bulk density or the gravi-

metric water content and air content at a given matric potential, the

relationships were unsatisfying (

) or no better

than the relationship with SOC content (

Johannes et al., 2017a

). The use

of quantitative indicators of soil structure in our study allowed using

scatter plot representations instead of the boxplots necessarily used in

and

. If the median values of

SOC/clay presented in the boxplots appeared well discriminated for the

different levels of the qualitative structural measure, boxplots also

showed large overlaps that could indicate some more complex re-

lationships. The continuous variables used in our study allowed visual-

izing these more complex relationships, such as inverse or logarithmic

relationships shown in

, and to highlight the low sensitivity of the

1/8, 1/10 and 1/13 thresholds in the case of France. A second expla-

nation for the inconsistent conclusions about the SOC/clay indicator is

linked to the use of an indicator designed and first tested in a restricted

pedoclimatic environment to be applied throughout Europe. While the

use of locally developed indicators should not be condemned, we

recommend vigilance regarding the use of thresholds defined in other

studies on datasets with wider ranges of variation in soil properties.

implicitly defined a domain of applicability of the

SOC/clay ratio, because they removed peat soils and sites identified as

outliers in the SOC/clay distribution from their study. Their proposal for

a soil degradation threshold of 1/13 was however taken up by the Eu-

ropean Commission. In the same way,

Dupla et al. (2021)

excluded

drained peatlands, sandy soils and heavy clay soils and removed outliers

in the SOC/clay distribution. The thresholds proposed by

were established for a single soil group (Cambi-Luvisols ac-

cording to the WRB classification), by purpose, to limit the variability in

the factors of SOC dynamics. In their study, soil pH ranged from 5 to 8.

4. Conclusion

In this study, we intended (i) to test the relevance of the SOC/clay

and SOC/(silt

<

20

µ

m

+

clay) ratios to assess the structural quality of

soils and (ii) to quantify the proportion of soils that would be considered

as unhealthy in France, using the threshold of SOC/clay chosen in the

European Soil Monitoring Law proposal. We used information from the

French Soil Quality Monitoring Network representing quantitative

measurements of 1,345 bulk densities and 174 aggregate stability tests.

We showed that SOC/clay and SOC/(silt

<

20

µ

m

+

clay) were poor

indicators of the soil bulk density and aggregate stability. The SOC

content was a much better indicator of soil structure. By normalizing the

SOC content by the fine fraction content, these indicators provide an

oversimplified view of the links between SOC and soil structure. Both

land use and soil type had an effect on the SOC/clay value. In particular,

SOC/clay was found to be strongly affected by soil pH, with acidic soils

consistently being classified as healthy according to the threshold of 1/

13 and alkaline soils often being classified as unhealthy. This bias of the

indicator, if used with a fixed threshold of 1/13, also prevents its use as

an indicator of the SOC status for the wide range of pedoclimatic con-

texts of Europe. Additional investigations are however required to

explain the results for alkaline soils. Based on the RMQS dataset, 63 % of

cropland, 81 % of permanent crop and 23 % of grassland soils were

below the threshold of 1/13, which would lead to classify these soils as

unhealthy. The Soil Monitoring Law leaves the possibility of using

correction factors for specific soil types or climatic conditions, which

appears necessary for France, because some pedoclimatic contexts will

never allow a satisfactory value to be reached. A pan-European study of

the correlation of SOC/clay and additional SOC indicators with several

quantitative and visual soil structure variables appears necessary, to

determine their domain of validity and identify, whenever necessary,

relevant threshold values for the diversity of European pedoclimatic

conditions.

CRediT authorship contribution statement

Eva Rabot:

Conceptualization, Formal analysis, Methodology,

Visualization, Writing

–

original draft.

Nicolas P.A. Saby:

Formal

analysis, Writing

–

review

&

editing.

Manuel P. Martin:

Writing

–

re-

view

&

editing.

Pierre Barr

´

e:

Writing

–

review

&

editing.

Claire Chenu:

Writing

–

review

&

editing.

Isabelle Cousin:

Writing

–

review

&

editing.

Dominique Arrouays:

Writing

–

review

&

editing.

Denis Angers:

Writing

–

review

&

editing.

Antonio Bispo:

Writing

–

review

&

editing,

Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial

interests or personal relationships that could have appeared to influence

the work reported in this paper.

Data availability

Data will be made available on request.

Acknowledgments

Thanks are expressed to the reviewers for their constructive and

helpful comments on a previous version of this article. The RMQS pro-

gram is funded by the GIS Sol, a scientific interest group involving the

French ministries in charge of the environment and of agriculture,

INRAE (National Research Institute for Agriculture, Food and the

Environment), ADEME (French Agency for Ecological Transition), IRD

(French National Institute for Sustainable Development), IGN (National

Institute of Geographic and Forest Information), OFB (French Biodi-

versity Agency) and BRGM (French geological survey). The authors

thank the staff of the European Conservatory for Soil Samples (CEES)

who were involved in preparing and archiving soil samples. This work

was supported by a grant from the AgroEcoSystem department from

INRAE.

References

Amendola, D., Mutema, M., Rosolen, V., Chaplot, V., 2018. Soil hydromorphy and soil

carbon: A global data analysis. Geoderma 324, 9

–

17.

https://doi.org/10.1016/j.geoderma.2023.116674

geoderma.2018.03.005

.

Arbor, A., Schmidt, M., Saurette, D., Zhang, J., Bulmer, C., Filatow, D., Kasraei, B.,

Smukler, S., Heung, B., 2023. A framework for recalibrating pedotransfer functions

using nonlinear least squares and estimating uncertainty using quantile regression.

Geoderma 439, 116674.

Baize, D., Girard, M.-C., 2008. R

´

ef

´

erentiel p

´

edologique 2008, Collection Savoir-faire.

´

Editions Quae, Association Française pour l

’

Etude du Sol (Afes), Versailles.

Ball, B.C., Batey, T., Munkholm, L.J., 2007. Field assessment of soil structural quality - a

development of the Peerlkamp test. Soil Use Manag. 23, 329

–

337.

10.1111/j.1475-2743.2007.00102.x

.

E. Rabot et al.

Geoderma 443 (2024) 116829

12

Çelik,

˙

I., Günal, H., Acar, M., Acir, N., Bereket Barut, Z., Budak, M., 2020. Evaluating the

long-term effects of tillage systems on soil structural quality using visual assessment

and classical methods. Soil Use Manag. 36, 223

–

239.

https://doi.org/10.1016/j.scitotenv.2019.02.249

sum.12554

.

Chen, S., Arrouays, D., Angers, D.A., Chenu, C., Barr

´

e, P., Martin, M.P., Saby, N.P.A.,

Walter, C., 2019. National estimation of soil organic carbon storage potential for

arable soils: A data-driven approach coupled with carbon-landscape zones. Sci. Total

Environ. 666, 355

–

367.

Cornelis, W.M., Akodi, D., Komutunga, E., Agaba, C., Ahumuza, E., Oratungye, K., 2019.

Exploring visual soil evaluation and examination methods on highly-weathered

tropical soil. Soil Tillage Res. 195, 104360

https://doi.org/10.1515/intag-2015-0007

still.2019.104360

.

Czy

˙

z, E.A., Dexter, A.R., 2015. Mechanical dispersion of clay from soil into water:

readily-dispersed and spontaneously-dispersed clay. Int. Agrophysics 29, 1

–

7.

https://doi.org/10.5194/hess-13-1485-2009

de Jonge, L.W., Moldrup, P., Schj

ø

nning, P., 2009. Soil Infrastructure, Interfaces

&

Translocation Processes in Inner Space (

“

Soil-it-is

”

): towards a road map for the

constraints and crossroads of soil architecture and biophysical processes. Hydrol.

Earth Syst. Sci. 13, 1485

–

1502.

https://doi.org/10.1016/j.geoderma.2008.01.022

Dexter, A.R., Richard, G., Arrouays, D., Czyz, E.A., Jolivet, C., Duval, O., 2008.

Complexed organic matter controls soil physical properties. Geoderma 144,

620

–

627.

https://doi.org/10.1016/S0167-1987(01)

Díaz-Zorita, M., Perfect, E., Grove, J.H., 2002. Disruptive methods for assessing soil

structure. Soil Tillage Res. 64, 3

–

22.

00254-9

.

Dupla, X., Gondret, K., Sauzet, O., Verrecchia, E., Boivin, P., 2021. Changes in topsoil

organic carbon content in the Swiss leman region cropland from 1993 to present.

Insights from large scale on-farm study. Geoderma 400, 115125.

10.1016/j.geoderma.2021.115125

Getahun, G.T., Munkholm, L.J., Schj

ø

nning, P., 2016. The influence of clay-to-carbon

ratio on soil physical properties in a humid sandy loam soil with contrasting tillage

and residue management. Geoderma 264, 94

–

102.

geoderma.2015.10.002

https://doi.org/10.1016/j.geoderma.2021.115529

Guillaume, T., Makowski, D., Libohova, Z., Bragazza, L., Sallaku, F., Sinaj, S., 2022. Soil

organic carbon saturation in cropland-grassland systems: Storage potential and soil

quality. Geoderma 406, 115529.

.

Guimar

˜

aes, R.M.L., Ball, B.C., Tormena, C.A., 2011. Improvements in the visual

evaluation of soil structure. Soil Use Manag. 27, 395

–

403.

j.1475-2743.2011.00354.x

.

IUSS Working Group WRB, 2022. World reference base for soil resources. International

soil classification system for naming soils and creating legends for soil maps. 4th

edition. International Union of Soil Sciences (IUSS), Vienna, Austria.

Johannes, A., Matter, A., Schulin, R., Weisskopf, P., Baveye, P.C., Boivin, P., 2017a.

Optimal organic carbon values for soil structure quality of arable soils. Does clay

content matter? Geoderma 302, 14

–

21.

geoderma.2017.04.021

https://doi.org/10.1016/j.still.2016.06.001

Johannes, A., Weisskopf, P., Schulin, R., Boivin, P., 2017b. To what extent do physical

measurements match with visual evaluation of soil structure? Soil Tillage Res. 173,

24

–

32.

.

Johannes, A., Sauzet, O., Matter, A., Boivin, P., 2023. Soil organic carbon content and

soil structure quality of clayey cropland soils: A large-scale study in the Swiss Jura

region. Soil Use Manag. 1

–

10

https://doi.org/10.1111/sum.12879

.

Jolivet, C., Arrouays, D., Boulonne, L., Rati

´

e, C., Saby, N., 2006. Le R

e des Sols de France (RMQS). Etat d

’

avancement et premiers r

Gest. Des Sols 13, 149

–

164

https://doi.org/10.4000/cybergeo.23155

Jolivet, C., Almeida Falcon, J.-L., Berch

´

e, P., Boulonne, L., Fontaine, M., Gouny, L.,

Lehmann, S., Maitre, B., Rati

´

e, C., Schellenberger, E., Soler-Dominguez, N., 2022.

French soil quality monitoring network manual RMQS2: second metropolitan

campaign 2016

–

2027, version 3. INRAE, Orl

´

eans, France. https://doi.org/

10.17180/KC64-NY88.

Joly, D., Brossard, T., Cardot, H., Cavailhes, J., Hilal, M., Wavresky, P., 2021. Les types

de climats en France, une construction spatiale. Cybergeo Eur. J. Geogr. Cartograph.

.

K

¨

ogel-Knabner, I., Amelung, W., 2021. Soil organic matter in major pedogenic soil

groups. Geoderma 384, 114785.

https://doi.org/10.1016/j.geoderma.2020.114785

.

Le Bissonnais, Y., 1996. Aggregate stability and assessment of soil crustability and

erodibility : I. Theory and methodology. Eur. J. Soil Sci. 47, 425

–

437.

https://doi.

org/10.1111/j.1365-2389.1996.tb01843.x

https://doi.org/10.1111/j.1365-

Le Bissonnais, Y., Arrouays, D., 1997. Aggregate stability and assessment of soil

crustability and erodibility: II. Application to humic loamy soils with various organic

carbon contents. Eur. J. Soil Sci. 48, 39

–

48.

2389.1997.tb00183.x

.

Lin, L., De Pue, J., Vivanco, A.K.M., Van der Bolt, F., Cornelis, W., 2022. Visual

assessment of soil structural quality across soil textures and compaction levels

–

Part

I: Examination of intact soil cores. Geoderma 426, 116099.

10.1016/j.geoderma.2022.116099

https://doi.org/10.1016/j.wasman.2016.11.040

Loyon, L., 2017. Overview of manure treatment in France. Waste Manag. 61, 516

–

520.

.

Lüdecke, D., Ben-Shachar, M.S., Patil, I., Waggoner, P., Makowski, D., 2021.

performance: An R Package for assessment, comparison and testing of statistical

models. J. Open Source Softw. 6, 3139.

https://doi.org/10.21105/joss.03139

Meersmans, J., De Ridder, F., Canters, F., De Baets, S., Van Molle, M., 2008. A multiple

regression approach to assess the spatial distribution of Soil Organic Carbon (SOC) at

the regional scale (Flanders, Belgium). Geoderma 143, 1

–

13.

10.1016/j.geoderma.2007.08.025

https://doi.org/10.1016/j.still.2021.105135

Mutuku, E.A., Vanlauwe, B., Roobroeck, D., Boeckx, P., Cornelis, W.M., 2021. Visual soil

examination and evaluation in the sub-humid and semi-arid regions of Kenya. Soil

Tillage Res. 213, 105135

org/10.1016/j.geodrs.2020.e00267

Nasta, P., Palladino, M., Sica, B., Pizzolante, A., Trifuoggi, M., Toscanesi, M., Giarra, A.,

D

’

Auria, J., Nicodemo, F., Mazzitelli, C., Lazzaro, U., Di Fiore, P., Romano, N., 2020.

Evaluating pedotransfer functions for predicting soil bulk density using hierarchical

mapping information in Campania. Italy. Geoderma Reg 21, e00267.

https://doi.

.

O

’

Rourke, S.M., Angers, D.A., Holden, N.M., McBratney, A.B., 2015. Soil organic carbon

across scales. Glob. Chang. Biol. 21, 3561

–

3574.

https://doi.org/10.1016/j.scitotenv.2023.165811

gcb.12959

.

Pacini, L., Arbelet, P., Chen, S., Bacq-Labreuil, A., Calvaruso, C., Schneider, F.,

Arrouays, D., Saby, N.P.A., C

´

ecillon, L., Barr

´

e, P., 2023. A new approach to estimate

soil organic carbon content targets in European croplands topsoils. Sci. Total

Environ. 900, 165811

.

Poeplau, C., Don, A., 2023. A simple soil organic carbon level metric beyond the organic

carbon-to-clay ratio. Soil Use Manag. 39, 1057

–

1067.

https://doi.org/10.1002/jpln.202000113

sum.12921

.

Poeplau, C., Jacobs, A., Don, A., Vos, C., Schneider, F., Wittnebel, M., Tiemeyer, B.,

Heidkamp, A., Prietz, R., Flessa, H., 2020. Stocks of organic carbon in German

agricultural soils

—

Key results of the first comprehensive inventory. J. Plant Nutr.

Soil Sci. 183, 665

–

681.

.

Prout, J.M., Shepherd, K.D., McGrath, S.P., Kirk, G.J.D., Haefele, S.M., 2021. What is a

good level of soil organic matter? An index based on organic carbon to clay ratio.

Eur. J. Soil Sci. 72, 2493

–

2503.

https://doi.org/10.1111/ejss.13012

https://doi.org/10.1038/s41598-

Prout, J.M., Shepherd, K.D., McGrath, S.P., Kirk, G.J.D., Hassall, K.L., Haefele, S.M.,

2022. Changes in organic carbon to clay ratios in different soils and land uses in

England and Wales over time. Sci. Rep. 12, 5162.

022-09101-3

https://doi.org/10.1111/sum.12920

Pulley, S., Taylor, H., Prout, J.M., Haefele, S.M., Collins, A.L., 2023. The soil organic

carbon:clay ratio in North Devon, UK: implications for marketing soil carbon as an

asset class. Soil Use Manag. 39, 1068

–

1081.

https://doi.org/10.1007/s10533-018-

R Core Team, 2022. R: A language and environment for statistical computing.

Rasmussen, C., Heckman, K., Wieder, W.R., Keiluweit, M., Lawrence, C.R., Berhe, A.A.,

Blankinship, J.C., Crow, S.E., Druhan, J.L., Hicks Pries, C.E., Marin-Spiotta, E.,

Plante, A.F., Sch

¨

adel, C., Schimel, J.P., Sierra, C.A., Thompson, A., Wagai, R., 2018.

Beyond clay: towards an improved set of variables for predicting soil organic matter

content. Biogeochemistry 137, 297

–

306.

0424-3

.

Regelink, I.C., Stoof, C.R., Rousseva, S., Weng, L., Lair, G.J., Kram, P., Nikolaidis, N.P.,

Kercheva, M., Banwart, S., Comans, R.N.J., 2015. Linkages between aggregate

formation, porosity and soil chemical properties. Geoderma 247

–

248, 24

–

37.

https://doi.org/10.1016/j.geoderma.2015.01.022

.

Rowley, M.C., Grand, S., Verrecchia,

´

E.P., 2018. Calcium-mediated stabilisation of soil

organic carbon. Biogeochemistry 137, 27

–

49.

https://doi.org/10.1007/s10533-017-

0410-1

.

Schj

ø

nning, P., de Jonge, L.W., Munkholm, L.J., Moldrup, P., Christensen, B.T., Olesen, J.

E., 2012. Clay dispersibility and soil friability

—

Testing the soil clay-to-carbon

saturation concept. Vadose Zo. J. 11

https://doi.org/10.2136/vzj2011.0067

https://doi.org/10.1111/j.1365-2389.2006.00809.x

von Lützow, M., K

¨

ogel-Knabner, I., Ekschmitt, K., Matzner, E., Guggenberger, G.,

Marschner, B., Flessa, H., 2006. Stabilization of organic matter in temperate soils:

mechanisms and their relevance under different soil conditions

–

a review. Eur. J.

Soil Sci. 57, 426

–

445.

doi.org/10.1016/j.geoderma.2018.07.026

Wiesmeier, M., Urbanski, L., Hobley, E., Lang, B., von Lützow, M., Marin-Spiotta, E., van

Wesemael, B., Rabot, E., Lie

ß

, M., Garcia-Franco, N., Wollschl

¨

ager, U., Vogel, H.-J.,

K

¨

ogel-Knabner, I., 2019. Soil organic carbon storage as a key function of soils - A

review of drivers and indicators at various scales. Geoderma 333, 149

–

162.

https://

Wong, V.N.L., Greene, R.S.B., Dalal, R.C., Murphy, B.W., 2010. Soil carbon dynamics in

saline and sodic soils: a review. Soil Use Manag. 26, 2

–

11.