Evaluation of soil fertility using combination of Landsat 8 and Sentinel‑2 data in agricultural lands

Abstract

Today, remote sensing is widely used to estimate soil properties. Because it is an easy and accessible way to estimate soil properties that are difficult to estimate in the field. Based on this, to evaluate the soil fertility (SF), soil sampling was performed irregularly from the surface depth of 0–30 cm in 216 points, 11 soil properties were measured, and the soil fertility index (SFI) was calculated by soil properties. Simultaneously, we combined satellite images of Landsat 8 and Sentinel-2 using the Gram-Schmidt algorithm. Finally, multiple linear regression SFI was calculated using satellite data, as well as the spatial distribution of SFI was obtained in very low, low, moderate, high, and very high classes. Our findings showed that the combination of Landsat 8 and Sentinel-2 data using the Gram-Schmidt algorithm has a higher correlation with SFI than when these data are individually. Therefore, combined Landsat 8 and Sentinel 2 data were used for SFI modeling. Using model selection procedure indices (including Cp, AIC, and ρc criteria), the visible range bands, notably blue (r = 0.65), green (r = 0.63), and red (r = 0.61), provide the best model for estimating SFI (R² = 0.43, Cp = 3.34, AIC = -277.4, and ρc = 0.44). Therefore, these bands were used to estimate the SFI index. Also, the spatial distribution of the SIF index showed that the most significant area was related to the low class, and the lowest area belonged to the high and very high fertility classes. According to these results, it can be concluded that using the combination of Landsat 8 and Sentinel 2 bands to estimate soil fertility index in agricultural lands can increase the accuracy of soil fertility estimation.

Full text

Vol.: (0123456789)
1 3
Environ Monit Assess (2024) 196:131
https://doi.org/10.1007/s10661-024-12301-1
RESEARCH
Evaluation ofsoil fertility using combination ofLandsat 8
andSentinel‑2 data inagricultural lands
MingZhang· MohammadKhosraviAqdam· HassanAbbasFadel·
LeiWang· KhloodWaheeb· AnghamKadhim· JamalHekmati
Received: 7 September 2023 / Accepted: 2 January 2024
© The Author(s), under exclusive licence to Springer Nature Switzerland AG 2024
Abstract Today, remote sensing is widely used to esti-
mate soil properties. Because it is an easy and accessible
way to estimate soil properties that are difficult to esti-
mate in the field. Based on this, to evaluate the soil fer-
tility (SF), soil sampling was performed irregularly from
the surface depth of 0–30cm in 216 points, 11 soil prop-
erties were measured, and the soil fertility index (SFI)
was calculated by soil properties. Simultaneously, we
combined satellite images of Landsat 8 and Sentinel-2
using the Gram-Schmidt algorithm. Finally, multiple
linear regression SFI was calculated using satellite data,
as well as the spatial distribution of SFI was obtained
in very low, low, moderate, high, and very high classes.
Our findings showed that the combination of Landsat
8 and Sentinel-2 data using the Gram-Schmidt algo-
rithm has a higher correlation with SFI than when these
data are individually. Therefore, combined Landsat 8
and Sentinel 2 data were used for SFI modeling. Using
model selection procedure indices (including Cp, AIC,
and ρc criteria), the visible range bands, notably blue
(r = 0.65), green (r = 0.63), and red (r = 0.61), provide
the best model for estimating SFI (R2 = 0.43, Cp = 3.34,
AIC = -277.4, and ρc = 0.44). Therefore, these bands
were used to estimate the SFI index. Also, the spatial
distribution of the SIF index showed that the most sig-
nificant area was related to the low class, and the lowest
area belonged to the high and very high fertility classes.
According to these results, it can be concluded that
using the combination of Landsat 8 and Sentinel 2 bands
to estimate soil fertility index in agricultural lands can
increase the accuracy of soil fertility estimation.
Keywords Gram-Schmidt algorithm· Remote
sensing· Soil fertility index· Soil fertility
M.Zhang(*)
Department ofResources andEnvironment, Anhui
Vocational andTechnological College ofForestry,
Hefei, Anhui230031, China
e-mail: ahlyxy2023@163.com
M.KhosraviAqdam
Administration ofEducation, KurdistanProvince, Saqqez, Iran
H.AbbasFadel
National University ofScience andTechnology, DhiQar, Iraq
L.Wang
Ministry ofEcology andEnvironment, Nanjing Institute
ofEnvironmental Sciences, Nanjing210042, China
e-mail: wlofrcc@126.com
K.Waheeb
Medical Technical College, Al-Farahidi University,
Baghdad, Iraq
A.Kadhim
Department ofOptical Techniques, Al-Zahrawi University
College, Karbala, Iraq
J.Hekmati
Department ofHorticultural Sciencess, University Campus
2, University ofGuilan, Rasht, Iran

Environ Monit Assess (2024) 196:131
1 3
131 Page 2 of 13
Vol:. (1234567890)
Introduction
Soil is a heterogeneous environment, and understand-
ing the spatial variability of soil fertility (SF) and their
properties in topsoil layers vital for precise manage-
ment to achieve optimum yield (Khosravi Aqdam etal.,
2023a, 2023b; Miran etal., 2021). SF evaluation is to
estimate the amount of native and added nutrients that
can be made available to the plant. Optimum SF condi-
tions increase crop yield, better absorption of nutrients
during plant growth, and create sustainable agricultural
conditions (Jin etal., 2021; Yageta etal., 2019).
Nutrients, as the main indicators of SF, play an
essential role in agricultural productivity, food secu-
rity, and sustainable agricultural development (Kes-
havarzi etal., 2022; Yang etal., 2023). It is clear that
in cultivated soils, plant growth-limiting challenges
are associated with nutrient deficiencies or toxicities
(Rezapour et al., 2022). The primary tool for assess-
ing SF is soil analysis (Nariyanti etal., 2022). Based
on this, investigating the nutrient status in agricultural
lands depends on field sampling and laboratory analysis
(Song etal., 2018). This method is accurate but cannot
quickly determine nutrient content (Zhang etal., 2022a,
2022b; Y. Wang et al., 2021). Estimating nutrients as
one of the main factors of SF requires alternative meth-
ods that can increase the speed and accuracy of these
nutrients. Advanced technologies, such as remote sens-
ing and digital soil mapping (DSM) (Keshavarzi etal.,
2023; Rahbar Alam Shirazi etal., 2023; Zeraatpisheh
etal., 2020), will lead to the development of estimating
soil properties (Arthur Endsley etal., 2020; Keshavarzi
etal., 2023). Obtaining accurate maps using the men-
tioned methods allows planning for large area (Kes-
havarzi etal., 2022). Because these maps provide useful
information for researchers to identify areas that need
management and protection (Kaya etal., 2022).
Remote sensing is a new method for studying spa-
tial variability of soil properties that can solve prob-
lems related to soil properties (Rahbar Alam Shirazi
et al., 2023; Yageta et al., 2019). In recent decades,
remote sensing data has often been used to estimate
soil properties (Kaur etal., 2020; Zeraatpisheh et al.,
2020). In general, a wide range of satellite images with
different spatial resolutions are available to the public
and are constantly updated. Therefore, they provide an
opportunity to estimate soil and plant characteristics
using spectral indices analysis (Khosravi Aqdam etal.,
2023a, 2023b; Zhang etal., 2020). Different research,
such as Arthur Endsley etal. (2020) and Miran etal.
(2021) was conducted in this field. The results of these
researches have led to useful results in estimating soil
and plant characteristics using satellite images.
The Sentinel-2 satellite supplies a collection of
images with appropriate frequency and medium spec-
tral resolution, with multispectral bands widely used
in agriculture because it provides free images with
a 5-day review time (Wang etal., 2022; Zhao et al.,
2023a, 2023b). Landsat 8 is another multispectral sat-
ellite with a spatial resolution of 30m, and its images
are free and have been used by different researchers
to estimate other soil properties (Miran et al., 2021;
Zhang et al., 2022a, 2022b). In addition, SF can be
calculated by physical, chemical, and biological soil
properties and converted into a single value called the
soil fertility index (SFI) (Panwar etal., 2011; K. Zhang
etal., 2022a, 2022b). Therefore, it is crucial to inves-
tigate the use of combining satellite images to esti-
mate SF in agricultural lands. In this regard, careful
consideration was given for selecting a study area that
encompassed a diverse range of landscape patterns,
vegetation types, and topographic properties because
we assumed that combining the images of Sentinel-2
and Landsat 8 could provide a reasonable estimate of
SF in these regions. Based on this, the primary objec-
tives of this study were: (1) investigate the intricate
relationship between SF and the combination of Land-
sat 8 and Sentinel-2 data within agricultural lands of
northwestern Iran, and (2) The spatial estimation of
SFI using the data obtained from the Gram-Schmidt
algorithm. By addressing these objectives, this study
aims to provide valuable insights into sustainable land
management and agricultural practices in the region.
Materials andmethods
Study sites and stages
The study area extends across 3375 km2 of agricul-
tural lands in northwest Iran (between 45o30′00′′
to 47o00′00′′ longitude and 35o50′00′′ to 36o30′00′′
latitude) (Fig. 1a). The annual precipitation average
is ~ 350 mm and the annual average air temperature
is ~ 9.2C◦. Land uses include agricultural Lands, forests,
grasslands, urban, water, and rock outcrops (Fig. 1b).
The conquering soil orders are Entisols and Inceptisols.
The major agricultural commodities are winter wheat,

Environ Monit Assess (2024) 196:131
1 3
Page 3 of 13 131
Vol.: (0123456789)
barley, and peas. Furrow irrigation has been practiced
in the region for decades, even though the shrinking
water resources compeled growers towards more effi-
cient irrigation systems such as central pivot systems.
Soil sampling and laboratory analysis
In the summer of 2022, 216 soil samples were col-
lected (Fig. 1c). The method of soil sampling was
stratified random from 0 to 30 cm soil depth in agri-
cultural lands (Fig.1a). This sampling method does
not need previous information about spatial varia-
tions of soil properties (C. Cao etal., 2022). In this
method, every population unit has the same chance of
being chosen, and The selection of sampling units is
independent. Soil samples were packed in zip bags.
First, soil samples were air-dried and passed through
a 2-mm sieve. Then, physicochemical soil properties
were determined (Table1).
Satellite data
Sentinel‑2 data andimage analysis
In this study, sentinel-2 satellite images were prepared
to the nearest time of soil sampling from the https://
scihub. coper nicus. eu website for each region. Senti-
nel-2 has 13 bands with magnifications ranging from
10 to 60 m (Sadeghi et al., 2017; Yue et al., 2023).
The bands 2, 3, 4, and 8, with a spatial resolution of
10m, has located in the spectral range of visible (Blue,
Green, Red), Near-infrared (NIR), and bands 11 and
12, with a spatial resolution of 20m has located in the
spectral range of short-wave infrared-1 (SWIR1) and
short-wave infrared-2 (SWIR2) were used. In the next
step, all the corrected bands were transferred to the
spatial resolution of 20m using the nearest neighbor
method. Finally, a sentinel-2 data set, including visible
and the NIR to the SWIR with a spatial resolution of
Fig. 1 Location of sampling points in Iran (a), land use map (b), and sampling points (c)

Environ Monit Assess (2024) 196:131
1 3
131 Page 4 of 13
Vol:. (1234567890)
20m, was prepared (Table2) (Chen etal., 2020). In
this stage, ENVI version 5.3 software was used.
Landsat 8 data andimage analysis
In this research, Landsat 8 satellite images from https://
earth explorer. usgs. gov/ website, with a spatial reso-
lution of 30 m, was prepared for the sampling time of
each region. Then, geometric and radiometric correc-
tions were done (Padró etal., 2017). In the next step,
bands 2, 3, 4, 5, 6, and 7, located in the Blue, Green,
Red, NIR, SWIR1, and SWIR2 spectral ranges, were
used for making visible and NIR to the SWIR data,
respectively. In this case, a data set similar to Senti-
nel-2 was created for Landsat 8 data with a spatial res-
olution of 30 m (Table2). This step was also carried
out in ENVI version 5.3 software.
Combination ofLandsat 8 andSentinel‑2 data
To combine Landsat 8 and Sentinel-2 data,
the Gram-Schmidt (G-S) algorithm was used
(Schowengerdt, 2012). This algorithm is one of
the most important algorithms for combining satel-
lite images that can maintain spectral nature when
merging multi-gauge data and transferring low spa-
tial power metric pixels to high spatial resolution.
Another advantage of this method is that spectral
reflection changes appear better, leading to better
identification of complications (Klonus & Ehlers,
2009). To prevent errors, similar bands of Landsat
8 with Sentinel-2 were combined by the G-S algo-
rithm. In this case, a data set was created according
to Table2. Similar to the previous step, ENVI ver-
sion 5.3 software was used.
Table 1 List and
description of the soil
properties used in this study
EC: electrical conductivity,
OC: organic carbon,
CCE: calcium carbonate
equivalent, TN: Total N,
AP: Available phosphor,
AK: Available potassium,
Zn: Zinc, Cu: Copper, and
Fe: Iron
Property* Laboratory method Reference
Soil texture Hydrometer (Bouyoucos, 1962)
pH Saturated soil-paste (Thomas, 1996)
EC (dS m−1) Saturated soil-paste (Rhoades, 1996)
OC (%) Dichromate oxidation (Nelson & Sommmers, 1983)
CCE (%) Reaction with HCl (Nelson, 1982)
TN (%) Kjeldahl (Bremner, 1996)
AP (mg kg−1) Olsen method (Olsen etal., 1982)
AK (meq l−1) Flame photometer (USDA-NRCS, 1996)
Zn (mg kg−1)Extracted by the DTPA (Lindsay & Norvell, 1978)
Cu (mg kg−1)
Fe (mg kg−1)
Table 2 Properties of Landsat 8 and Sentinel-2 bands before and after the Gram-Schmidt (G-S) algorithm
NIR: Near-infrared
SWIR: Short-wave infrared
Primary bands By Final bands
Landsat 8 bands (µm) Resolution
(m)
Sentinel-2 bands (µm) Resolution
(m)
Gram-Schmidt
algorithm
Resolution
(m)
Band name
Band 2—Blue (0.45–0.51) 10 Band 2—Blue (0.458–0.523) 30 20 Blue
Band 3—Green (0.53–0.59) 10 Band 3—Green (0.543–0.578) 30 20 Green
Band 4—Red (0.64–0.67) 10 Band 4—Red (0.650–0.680) 30 20 Red
Band 5—NIR (0.85–0.88) 10 Band 8—NIR (0.785–0.900) 30 20 NIR
Band 6—SWIR1 (1.57–1.65) 20 Band 11—SWIR1 (1.565–
1.655)
30 20 SWIR1
Band 7—SWIR2 (2.11–2.25) 20 Band 12—SWIR2 (2.100–2.280) 30 20 SWIR2

Environ Monit Assess (2024) 196:131
1 3
Page 5 of 13 131
Vol.: (0123456789)
Data analysis
In this study, SFI was used to evaluate the SF status
at sampling points (Tunçay etal., 2021). This model
uses a parametric method to determine the SF using
a quantitative grading assigned to each feature. If
a property is desirable, a degree of a maximum of
100 is given to it. If the same property is limited,
it is assigned a lower degree (Table3). It should be
noted that the range of some soil nutrients accord-
ing to the status of these nutrients in agricultural
soils extracted from the database of SF and fertil-
ity assessment of agricultural lands of Iran (https://
www. swri. ir) was extracted.
After ranking each nutrient and soil properties
according to the ranges presented in Table3, the SFI was
calculated according to the following formula (Eq.1).
where R max is the maximum ratio (A + B + ….
…… + K)/11, and A, B …: rating value for each diag-
nostic factor.
Variable Selection and spatial distribution of SFI
At this stage of the research, Pearson correlation
was obtained between Landsat 8 bands with SFI,
Sentinel-2 with SFI, and the data obtained from
(1)
SFI
=Rmax ×
√
A
100 ×
B
100
×
Gram-Schmidt algorithm with SFI. Then the data-
set that had the highest correlation were used for
modeling. To find the best model for estimating
SFI, model selection pressure indices including
akaike information criterion (AIC), coefficient of
determination (R2), mallows’s Cp (Cp), and con-
cordance (ρc) (Eqs. 2–5) were used (Cao et al.,
2023; Kaiser & Hogan, 2007). To find the best
regression model for SFI estimation, SAS JMP
software was used (Tang etal., 2023; Zhang etal.,
2022a, 2022b; Zhao etal., 2023a, 2023b). Using
this method, the best model for SFI estimation was
obtained. Then, the spatial distribution map of SFI
was obtained in Arc GIS 10.8 software and was
classified using Table4 of SF classes of agricul-
tural lands in this area.
(2)
AIC =In(n)





n
i=1
(Obs −prd)2
Obs −Obs2





+2
Np
(3)
R2=1−
n
i=1(Obs −prd)
2

n
i=1

Obs −Obs

2
(4)
Cp
=
RSSp
𝜎
2+2p−
n
Table 3 Ranking of nutrients affecting SF
CL: clay loam, SCL: sandy clay loam, vfSL: very fine sandy loam, L: loam, C: clay, SL: sandy loam, fSL: fine sandy loam, S: sand,
LS: loamy sand, SiCL: silty clay loam, SiL: silty loam, Si: silty, SC: sandy clay, SiC: silty clay
Soil properties The grade assigned to each property
10 20 50 80 100 10
A-TN (%) > 0.32 < 0.045 0.045–0.09 0.09–0.17 0.17–0.32 < 0.045
B-AP (mg kg−1) > 20 < 5 5–10 10–15 15–20 < 5
C-AK (meq l−1) 110–288 < 51 > 975 51–110 288–975 < 51
D-Zn (mg/kg) > 1 < 0.25 0.25–0.5 0.5–0.75 0.75–1 < 0.25
F- Fe (mg/kg) > 7.5 - - 2.5–5 5–7.5 -
E-Cu (mg/kg) > 1 - - < 0.25 0.25–1 -
G- CCE (%) < 15 > 60 40–60 25–40 15–25 > 60
H- EC (dS m-1) 0–2 > 8 6–8 4–6 2–4 > 8
I- pH 6.5–7.5 > 8.5 and < 4.5 4.5–5.5 5.5–6.5 7.5–8.5 > 8.5 and < 4.5
J- OM (%) > 3 0-.5 0.5–1 1–2 2–3 0-.5
K- soil texture CL,SCL, SiCL S, LS SL, fSL C (> 50%), SC, SiC L, vfSL, Si, SiL,
C (< 50%)
S, LS

Environ Monit Assess (2024) 196:131
1 3
131 Page 6 of 13
Vol:. (1234567890)
where Obs, Prd,
Obs
, and
Prd
are the measured,
predicted, average measured, average predicted values
of the model, respectively. ρ is Pearson’s correlation
coefficient between the predictions and observations,
n is the number of sampling points, σ2Obs and σ2Pred
are the corresponding variances. NP is the number of
parameters that need to be predicted. Also, σ2 is from
the model with all predictors, and RSSp indicates the
RSS from a model with p parameters (Wang et al.,
2023; Yue etal., 2023; Zhao etal., 2023a, 2023b).
Results
Statistical analysis
Descriptive statistics of measurement properties are
presented in Table5. As can be seen, there is a sig-
nificant coefficient of change for sand (50.63%), silt
(21.54%), and clay (34.55%), resulting in eight soil
texture classes. The soil texture of this area includes
clay, silty clay, silty clay loam, clay loam, silty loam,
loam, sandy clay loam, sandy loam. Such diversity
can cause variation in soil physicochemical proper-
ties and diversity in SF. The minimum, maximum,
mean, and coefficient of variation (CV) values for
EC are 0.45, 1.71, 1.02 (dS.m−1), and 27.36 (%),
respectively. The minimum values, maximum, mean,
and CV for CCE are 1.4, 41.6, 12.4, and 92.62 (%).
These values show that it has the highest CV among
soil properties. These values indicate that the soils in
the study area fall into the low to high-alkaline group
(Vinhal-Freitas etal., 2017).
Soil organic carbon (SOC) is one of the most
important soil properties that plays a vital role in SF
(5)
𝜌
c=
2𝜌𝜎
Obs
𝜎
Prd
𝜎2
Obs
+𝜎2
Prd
+(Prd +Obs)
2
and is today considered as one of the most impor-
tant soil properties (Nariyanti etal., 2022; K. Zhang
et al., 2022a, 2022b). The minimum values, maxi-
mum, mean, and CV for SOC are 0.65, 1.56, 1.24,
and 18.91. Jat etal. (2018) classified SOC as less than
1% in the very low group and higher than 3% for SOC
in the high group. Therefore, according to the mean
of SOC in the agricultural lands of this region, it can
be said that these lands are in the middle group (Cao
etal., 2022).
The values of CV in macronutrients, includ-
ing TN, AP, and AK, were 18.91, 87.67, and 15.70
(%), respectively. These values show that the CV of
AP is the highest value. Considering that all lands in
the studied area include agricultural lands, it can be
expected that phosphorus fixation in some lands in
this region can cause high CV of AP in this region
(Tang et al., 2023; Y. Zhao et al., 2023a, 2023b).
Also, the values of soil micronutrients, including
Zn, Fe, and Cu, indicate that the values of these ele-
ments in the agricultural lands of this region are low
(Table5).
Scoring of the determining factors of SFI
In this study, using macronutrients (TN, AP, and AK),
micronutrients (Fe, Zn, and Cu), chemical properties
(CCE, OC, EC, pH), and soil texture, SFI was pre-
dicted. Each grade was scored using the Jenks (1967)
method. The descriptive statistics of these scores are
described in Table6. As the results show, the highest
score is given to the values of AP, CCE, EC, pH, and
soil texture. The lowest scores are given to AP, Zn, and
OC. Also, according to the results, it is observed that
the coefficient of variations in the scores of AK and EC
is equal to zero. This can be due to a sufficient amount
of AK due to the mineralogy of the area and no prob-
lem of salinity in the agricultural lands of this region.
Correlations between remote sensing data and SFI
and multiple linear regression between SFI and
remote sensing data
Pearson correlation coefficient between bands obtained
from Landsat 8 with SFI (Fig.2a), Sentinel-2 with SFI
(Fig. 2b), and combing Landsat 8 and Sentinel-2 sat-
ellite images showed that there was a significant posi-
tive correlation between bands and SFI (Fig.2c). The
Table 4 SFI classification (Tunçay etal., 2021)
Remarks: SFI (soil fertility index)
SFI value SFI class
0–25 Very low
25–50 Low
50–75 Moderate
75–90 High
90–100 Very high

Environ Monit Assess (2024) 196:131
1 3
Page 7 of 13 131
Vol.: (0123456789)
results of Pearson correlation between SFI and the
bands obtained from the combination of Landsat 8 and
Sentinel-2 satellites showed that there was the highest
correlation of this index with the blue band (r = 0.65),
and the lowest correlation with NIR band (r = 0.53)
(Fig.2c). Based on this, the data combined with Gram-
Schmidt algorithm was used for SF modeling. The best
multiple linear regression (MLR) model for SFI estima-
tion was obtained using R2, Cp, AIC, and ρc indices.
According to the results, it was observed that the best
MLR model for SFI estimation is given in Eq.6. The
values of R2, Cp, AIC, and ρc are 0.42, 3.34, -277.4,
286.5, and 0.44, respectively. As it is clear from the
equation results, the best bands for estimating the SFI of
the visible range bands include Blue, Green, and Red.
Spatial variability and distribution maps of SFI
After obtaining the best equation using the model
selection procedure, the spatial distribution of the SFI
index was determined using the best regression equa-
tion. Figure3 shows the spatial distribution of SFI for
agricultural lands of the study area. According to this
figure, the highest area (70.266%) is in the low class,
(6)
SIF
=−6.10 +(1.10 ×Blue)−(0.19 ×Red)+(0.13 ×Green)
R
2
=0.424
Cp =3.34
AIC =−
277.4
𝜌c=
0.44
Table 5 Statistical
properties of studied soil in
all districts
Soil properties Minimum Maximum Mean Standard
deviation
Coefficient of
variation (%)
A-TN (%) 0.07 0.16 0.12 0.02 18.91
B-AP (mg kg−1) 0.02 40.06 9.57 8.39 87.67
C-AK (meq l−1) 390.25 804.72 571.98 89.81 15.70
D-Zn (mg/kg) 0.37 0.62 0.52 0.07 12.78
F- Fe (mg/kg) 4.25 5.75 5.23 0.38 7.26
E-Cu (mg/kg) 0.05 0.75 0.45 0.18 40.55
G- CCE (%) 1.40 41.60 12.42 11.51 92.62
H- EC (ds m-1) 0.45 1.71 1.02 0.28 27.36
I- pH 6.87 8.11 7.50 0.28 3.79
J- OM (%) 0.65 1.56 1.24 0.23 18.91
K-Sand (%) 8.00 61.00 28.48 14.42 50.63
K-Silt (%) 22.00 56.00 38.58 8.31 21.54
K-Clay (%) 11.00 56.00 32.90 11.37 34.55
Table 6 Descriptive
statistics of the scores given
to each soil characteristic to
calculate the SFI index
Statistic Minimum Maximum Mean Standard
deviation
Coefficient of
variation (%)
A-TN (%) 20 50 45.48 10.81 23.76
B-AP (mg kg−1) 10 100 38.77 30.59 78.92
C-AK (meq l−1) 80 80 80.00 0.00 0.00
D-Zn (mg/kg) 20 50 38.90 14.58 37.49
F- Fe (mg/kg) 50 80 72.60 13.02 17.93
E-Cu (mg/kg) 50 80 75.48 10.81 14.32
G- CCE (%) 20 100 87.40 20.89 23.90
H- EC (ds m-1) 100 100 100.00 0.00 0.00
I- pH 80 100 89.59 10.06 11.23
J- OM (%) 20 50 45.89 10.39 22.63
K- soil texture 50 100 76.03 24.31 31.97

Environ Monit Assess (2024) 196:131
1 3
131 Page 8 of 13
Vol:. (1234567890)
and the lowest area (0.002%) is in the very high class
(Table7).
Discussion
This study aimed to obtain SF based on the SFI in a large
part of the agricultural lands in northwest Iran (Fig.3).
Accordingly, a wide range of soil properties was deter-
mined in 216 locations in a large part of the agricultural
lands of northwestern Iran. Knowledge of SF in agricul-
tural lands can lead to agricultural development and sus-
tainable development (Nariyanti etal., 2022).
In this study, three categories of data were used to
estimate SF, which the best of them were the com-
bined data of Landsat 8 and Sentinel 2 using the
Gram-Schmidt algorithm (Fig. 2). So, we used a
combination of data from Landsat 8 and Sentinel-2
to estimate SF. The results showed that visible range
bands can be used to estimate SF in agricultural lands
in northwestern Iran with almost appropriate accu-
racy. In this study, we tried to increase the estima-
tion accuracy using the combination of Landsat 8 and
Sentiel-2 satellite image bands by the Gram-Schmidt
algorithm. In addition, we tried to use model selec-
tion prediction accuracy indices (including Cp, ρc,
and AIC) to select a model that had the highest accu-
racy for estimating SF, which was much better than
the results of the study they did (Eq.6).
In a study T. Wang etal. (2023), which was done
to predict SOC, 33 environmental permeates extracted
from Sentinel-2 and DEM were used. The results of this
study showed that the use of environmental indicators
for estimating SOC is dependent on sampling method.
They concluded that modifying the sampling method
could increase the prediction accuracy of soil properties
estimation models. Therefore, it can be said that using
sampling methods such as the Latin hypercube sam-
pling method (cLHS), the accuracy of the SFI estima-
tion can be greatly increased using the combination of
Landsat 8 and Sentinel-2 bands (Zhang etal., 2020).
The spatial estimation of SFI in the study area showed
that most studied areas were in low and very low fertil-
ity groups (about 80%) (Fig.3 and Table7). SOC, pH,
Fig. 2 Pearson correlation between satellite data and SFI
[Landsat with SFI (a), Sentinel-2 with SFI (b), and Gram-
Gram-Schmidt algorithm with SFI (c)]
▸

Environ Monit Assess (2024) 196:131
1 3
Page 9 of 13 131
Vol.: (0123456789)
soil texture, water, and nutrients affect SF (Kumar etal.,
2018). According to the field observation, it was found
that the pattern of fertilizer application in the study area
is such that most of the nutrients, including N and P, are
added to the soil through fertilizer.
In a study conducted by Miran et al. (2021) in rela-
tion to nutrients in agricultural lands in northwestern
Iran، it was found that farmers mainly consume two
Urea and Diammonium phosphate fertilizers in the
agricultural lands of this region, and in very few cases
micronutrients are used. Therefore, there is a lack of
micronutrients in most of the lands of this region.
This action by farmers to increase SF in the long run
will reduce SF on these lands and have irreparable
consequences in the region (Putra et al., 2020).
Based on the results of this study, it was found that
soils in the region have high pH and high CCE content
(Table5). The high pH of soil solution affects the solu-
bility of nutrients. Micronutrients, including Fe, Cu, Zn,
and Mn, have very low solubility in calcareous soils
(Tamfuh etal., 2018). Therefore, the low availability
of these nutrients in these soils reduces SF. In a study
conducted by Khosravi Aqdam etal., (2023a, 2023b) in
Fig. 3 Spatial distribution of SFI in the study area
Table 7 Area and percentage of each of the SFI classes of the
area studied
SFI class Area (ha) Area (%)
Very low 31,906.16 9.53
Low 235,204.94 70.28
Moderate 67,408.41 20.15
High 210.02 0.06
Very high 5.57 0.002
Total 334,735.10 100.00

Environ Monit Assess (2024) 196:131
1 3
131 Page 10 of 13
Vol:. (1234567890)
the agricultural lands of northwestern Iran, it was found
that macro and micronutrients of soil in all the eleva-
tions of this region are in a low group. This research
determined that the study of SF in this area is essential
and requires advanced methods to evaluate SF. There-
fore, combining Landsat 8 and Sentinel-2 satellite data
can improve SF estimation in agricultural lands.
Another reason for low fertility in studied area
could be related to low organic matter of soil.
Therefore, to maintain SF and compensate for the lack
of SOC, it is necessary to carry out continuous soil
conservation programs to maintain SF by restoring
SOC. Because intensive agricultural activities and
lack of attention to soil conservation programs, apart
from reducing SOC in these lands, is one of the main
factors of nutrient loss N, P, and K which ultimately
leads to a decrease in SF in the agricultural lands of
this region (Cambou et al., 2016). Although, different
levels of SOC is affected by many issues such as soil
properties and environmental factors. But new studies
show that the potential effect of land use changes will
contribute more to SOC levels. Therefore, the most
effective mechanism for reducing SOC losses is to
prevent land use changes in arid and semi-arid regions (T.
Wang etal., 2023). Also, non-observance of crop rotation
or monoculture, reduces soil potential and fertility. In
addition, this type of cultivation has negative effects such
as reducing biodiversity, increasing the growth of weeds,
and increasing diseases and fungi (Zhang etal., 2020).
Other factor that reduces SF in this area is agricul-
tural activities in such a way that these activities
decrease SOC. Among these activities are insufficient
use of animal and plant manures and burning plant
residues, all of which reduce SF. In general, it seems
that low SOC, high pH (pH > 7.5), low humidity due
to semi-arid climatic conditions (precipita-
tion = 250–300), low availability of nutrients, and low
amount of SOC in the studied areas cause low fertility
status. This means that adding organic fertilizers,
including animal manure and plant residues, using
chemical fertilizers according to soil and plant degra-
dation, and using micronutrients fertilizers can
increase soil potential for crop production in these
areas. In general, according to the results of this study,
it can be concluded that the combination of data from
Landsat 8 and Sentinel-2 can be estimated with high
accuracy of SF because the results of this study are in
line with the results of the research conducted in this
region (Tunçay et al., 2021). In addition، another
advantage of the results over similar studies is that
increasing the accuracy of spatial images using the
algorithms used in this study can improve the quality
of the research and more accurate knowledge of the
spatial distribution of SF in the agricultural lands of
this region. Finally, according to the results of this
study, it can be seen that the use of the Gram-Schmidt
algorithm to combine the data of different satellites
can increase the accuracy of soil properties estimation
models. Also, the use of advanced methods such as
machine learning methods can greatly increase the
accuracy of prediction models.
Conclusion
Investigating SFI and obtaining the spatial distribu-
tion of SFI by combining Landsat 8 and Sentinel-2
data showed that this data can provide acceptable esti-
mation accuracy of SF in agricultural lands. Also, the
results showed that most of the land in this area is in
the low SF group. The limitations of SF in this area are
high pH and low SOC of soil. In addition, macro and
micronutrients deficiency in the agricultural lands is
another reason for decreasing SF in this region. In gen-
eral, using the combination of Landsat 8 and Sentiel-2
data can be one of the ways to know the exact spatial
distribution for SF estimating, and policymakers and
managers of agriculture and SF can use this method to
understand the purposeful understanding of SF in this
area, which can finally identify factors limiting SF.
Acknowledgements The authors are grateful for the financial
support provided for this study by Anhui University.
Author contribution Ming Zhang; formal analysis: Ming
Zhang; data curation: Mohammad Khosravi Aqdam and Jamal
Hekmati; writing—original draft preparation: Ming Zhang;
review and editing: Mohammad Khosravi Aqdam, Hassan abbas
fadel, Lei Wang, Khlood waheeb, Angham kadhim, and Jamal
Hekmati; funding acquisition: Ming Zhang and Lei Wang.
All authors have read and agreed to the published version of
the manuscript.
Funding This work was supported by Anhui University
Natural Science Key Research Program “Study on forest car-
bon stocks and their dynamics in different vegetation types”
(2022AH052730); Anhui Province 2021 Excellent Talent
Support Program for Colleges and Universities Key Pro-
jects “Characterization and quality evaluation of soil envi-
ronment in Dangshan Pear Park, Anhui Province, China”
(gxyqZD2021166).

Environ Monit Assess (2024) 196:131
1 3
Page 11 of 13 131
Vol.: (0123456789)
Data availability The datasets generated during and/or
analyzed during the current study are available from the cor-
responding author on reasonable request. The data of digital
elevation model are available in the earth explorer website
(https:// earth explo rer. usgs. gov/).
Declarations
Ethics approval All authors have read, understood, and have
complied as applicable with the statement on “Ethical responsi-
bilities of Authors” as found in the Instructions for Authors and
are aware that with minor exceptions, no changes can be made
to authorship once the paper is submitted.
Competing interests The authors declare no competing interests.
References
Arthur Endsley, K., Kimball, J. S., Reichle, R. H., & Watts,
J. D. (2020). Satellite monitoring of global surface soil
organic carbon dynamics using the SMAP level 4 car-
bon product. Journal of Geophysical Research: Biogeo‑
sciences, 125(12), e2020JG006100. https:// doi. org/ 10.
1029/ 2020J G0061 00
Bouyoucos, G. J. (1962). Hydrometer method improved for
making particle size analyses of soils 1. Agronomy Jour‑
nal, 54(5), 464–465. https:// doi. org/ 10. 2134/ agron j1962.
00021 96200 54000 50028x
Bremner, J. M. (1996) Nitrogen total. In D. L. Sparks (Ed.),
Methods of Soil Analysis Part 3: Chemical Methods, SSSA
Book Series 5, Soil Science Society of America, Madison,
Wisconsin (pp. 1085–1122). John Wiley. https:// doi. org/
10. 2136/ sssab ookse r5.3. c37
Cambou, A., Cardinael, R., Kouakoua, E., Villeneuve, M.,
Durand, C., & Barthès, B. G. (2016). Prediction of soil
organic carbon stock using visible and near infrared reflec-
tance spectroscopy (VNIRS) in the field. Geoderma, 261,
151–159. https:// doi. org/ 10. 1016/j. geode rma. 2015. 07. 007
Cao, C., Wang, J., Kwok, D., Cui, F., Zhang, Z., Zhao, D., Li,
M. J., & Zou, Q. (2022). webTWAS: A resource for disease
candidate susceptibility genes identified by transcriptome-
wide association study. Nucleic Acids Research, 50(D1),
D1123–D1130. https:// doi. org/ 10. 1093/ nar/ gkab9 57
Cao, Y., Xu, N., Wang, H., Zhao, X., & Ahmad, A. M. (2023).
Neural networks-based adaptive tracking control for full-
state constrained switched nonlinear systems with peri-
odic disturbances and actuator saturation. International
Journal of Systems Science, 54(14), 2689–2704. https://
doi. org/ 10. 1080/ 00207 721. 2023. 22419 59
Chen, Q., Miao, F., Wang, H., Xu, Z. X., Tang, Z., Yang, L., &
Qi, S. (2020). Downscaling of satellite remote sensing soil
moisture products over the Tibetan Plateau based on the
random forest algorithm: Preliminary results. Earth and
Space Science, 7(6), e2020EA001265. https:// doi. org/ 10.
1029/ 2020E A0012 65
Jat, H., Datta, A., Sharma, P., Kumar, V., Yadav, A., Choud-
hary, M., Choudhary, V., Gathala, M., Sharma, D., & Jat,
M. (2018). Assessing soil properties and nutrient availabil-
ity under conservation agriculture practices in a reclaimed
sodic soil in cereal-based systems of North-West India.
Archives of Agronomy and Soil Science, 64(4), 531–545.
https:// doi. org/ 10. 1080/ 03650 340. 2017. 13594 15
Jenks, G. F. (1967). The data model concept in statistical map-
ping. International Yearbook of Cartography, 7, 186–190.
Jin, J., Wang, L., Müller, K., Wu, J., Wang, H., Zhao, K., Ber-
ninger, F., & Fu, W. (2021). A 10-year monitoring of soil
properties dynamics and soil fertility evaluation in Chi-
nese hickory plantation regions of southeastern China.
Scientific Reports, 11(1), 23531. https:// doi. org/ 10. 1038/
s41598- 021- 02947-z
Kaiser, R. B., & Hogan, R. (2007). Chapter10 the dark side of
discretion: Leader personality and organizational decline
(Being there even when you are not) (pp. 173–193). Emer-
ald Group Publishing Limited. http:// ndl. ether net. edu. et/
bitst ream/ 12345 6789/ 47884/1/ dx. doi. org/ 10. 1016/ S1479
3571(07) 04009-6. 3d
Kaur, G., Das, K., & Hazra, J. (2020). Soil nutrients prediction
using remote sensing data in western india: An evaluation
of machine learning models. In IGARSS 2020‑2020 IEEE
International Geoscience and Remote Sensing Sympo‑
sium. Waikoloa, HI: IEEE. https:// doi. org/ 10. 1109/ IGARS
S39084. 2020. 93242 01
Kaya, F., Keshavarzi, A., Francaviglia, R., Kaplan, G.,
Başayiğit, L., & Dedeoğlu, M. (2022). Assessing machine
learning-based prediction under different agricultural
practices for digital mapping of soil organic carbon and
available phosphorus. Agriculture, 12(7), 1062. https://
doi. org/ 10. 3390/ agric ultur e1207 1062
Keshavarzi, A., del Árbol, M. Á. S., Kaya, F., Gyasi-Agyei,
Y., & Rodrigo-Comino, J. (2022). Digital mapping of
soil texture classes for efficient land management in the
Piedmont plain of Iran. Soil Use and Management, 38(4),
1705–1735. https:// doi. org/ 10. 1111/ sum. 12833
Keshavarzi, A., Kaya, F., Başayiğit, L., Gyasi-Agyei, Y., Rod-
rigo-Comino, J., & Caballero-Calvo, A. (2023). Spatial
prediction of soil micronutrients using machine learning
algorithms integrated with multiple digital covariates.
Nutrient Cycling in Agroecosystems, 127(1), 137–153.
https:// doi. org/ 10. 1007/ s10705- 023- 10303-y
Khosravi Aqdam, K., Asadzadeh, F., Rezapour, S., & Nouri,
A. (2023a). Comparative assessment of soil fertility
across varying elevations. Environmental Monitoring
and Assessment, 195(8), 1007. https:// doi. org/ 10. 1007/
s10661- 023- 11610-1
Khosravi Aqdam, K., Rezapour, S., Asadzadeh, F., & Nouri, A.
(2023b). An integrated approach for estimating soil health:
Incorporating digital elevation models and remote sensing
of vegetation. Computers and Electronics in Agriculture,
210, 107922. https:// doi. org/ 10. 1016/j. compag. 2023. 107922
Klonus, S., & Ehlers, M. (2009). Performance of evaluation
methods in image fusion. In 2009 12th International Con‑
ference on Information Fusion. Seattle, WA: IEEE.

Environ Monit Assess (2024) 196:131
1 3
131 Page 12 of 13
Vol:. (1234567890)
Kumar, U., Nayak, A. K., Shahid, M., Gupta, V. V., Panneersel-
vam, P., Mohanty, S., Kaviraj, M., Kumar, A., Chatterjee,
D., & Lal, B. (2018). Continuous application of inorganic
and organic fertilizers over 47 years in paddy soil alters
the bacterial community structure and its influence on rice
production. Agriculture, Ecosystems & Environment, 262,
65–75. https:// doi. org/ 10. 1016/j. agee. 2018. 04. 016
Lindsay, W. L., & Norvell, W. (1978). Development of a DTPA
soil test for zinc, iron, manganese, and copper. Soil Sci‑
ence Society of America Journal, 42(3), 421–428.
Miran, N., Rasouli Sadaghiani, M. H., Feiziasl, V., Sepehr, E.,
Rahmati, M., & Mirzaee, S. (2021). Predicting soil nutri-
ent contents using Landsat OLI satellite images in rain-
fed agricultural lands, northwest of Iran. Environmental
Monitoring and Assessment, 193(9), 607. https:// doi. org/
10. 1007/ s10661- 021- 09397-0
Nariyanti, S., Herawati, A., Herdiansyah, G., Irianto, H., Riptanti,
E. W., & Qonita, A. (2022). Soil fertility index based on
altitude: A comprehensive assessment for the cassava devel-
opment area in Indonesia. Annals of Agricultural Sciences,
67(2), 158–165. https:// doi. org/ 10. 1016/j. aoas. 2022. 10. 001
Nelson, D. A., & Sommers, L. E. (1983). Total carbon, organic
carbon, and organic matter. Methods of soil analysis: Part
chemical and microbiological properties, 9, 539–579.
Nelson, R. E. (1982). Carbonate and gypsum. In A. L. Page
(Ed.), Methods of soil analysis. Part 2. Agron. Monogr.
(Vol. 9, 2nd ed., pp. 181–197). Madison, WI: ASA and
SSSA.
Olsen, S., Sommers, L., & Page, A. (1982). Methods of Soil
Analysis. Part, 2(1982), 403–430.
Padró, J.-C., Pons, X., Aragonés, D., Díaz-Delgado, R., García,
D., Bustamante, J., Pesquer, L., Domingo-Marimon, C.,
González-Guerrero, Ò., & Cristóbal, J. (2017). Radiometric
correction of simultaneously acquired Landsat-7/Landsat-8
and Sentinel-2A imagery using pseudoinvariant areas (PIA):
Contributing to the Landsat time series legacy. Remote Sens‑
ing, 9(12), 1319. https:// doi. org/ 10. 3390/ rs912 1319
Panwar, P., Pal, S., Reza, S., & Sharma, B. (2011). Soil fertility
index, soil evaluation factor, and microbial indices under
different land uses in acidic soil of humid subtropical
India. Communications in Soil Science and Plant Analy‑
sis, 42(22), 2724–2737. https:// doi. org/ 10. 1080/ 00103
624. 2011. 622820
Putra, S. S., Putra, E. T. S., & Widada, J. (2020). The effects
of types of manure and mycorrhizal applications on sandy
soils on the growth and yield of curly red chili (Capsicum
annum L.). Journal of Sustainable Agriculture, 35(2),
258–267. https:// doi. org/ 10. 20961/ carak atani. v35i2. 34971
Rahbar Alam Shirazi, F., Shahbazi, F., Rezaei, H., & Biswas,
A. (2023). Digital assessments of soil organic carbon stor-
age using digital maps provided by static and dynamic
environmental covariates. Soil Use and Management,
39(2), 948–974. https:// doi. org/ 10. 1111/ sum. 12900
Rezapour, S., Siavash Moghaddam, S., Nouri, A., & Khosravi
Aqdam, K. (2022). Urbanization influences the distribu-
tion, enrichment, and ecological health risk of heavy met-
als in croplands. Scientific Reports, 12(1), 3868. https://
doi. org/ 10. 1038/ s41598- 022- 07789-x
Rhoades, J. (1996). Salinity: Electrical conductivity and total dis-
solved solids. Methods of soil analysis: part 3 chemical meth‑
ods, 5, 417–435. https:// doi. org/ 10. 2136/ sssab ookse r5.3. c14
Sadeghi, M., Babaeian, E., Tuller, M., & Jones, S. B. (2017).
The optical trapezoid model: A novel approach to remote
sensing of soil moisture applied to Sentinel-2 and Land-
sat-8 observations. Remote Sensing of Environment, 198,
52–68. https:// doi. org/ 10. 1016/j. rse. 2017. 05. 041
Schowengerdt, R. A. (2012). Techniques for image processing
and classifications in remote sensing. Academic Press.
https:// doi. org/ 10. 1016/j. rse. 2017. 05. 041
Tamfuh, P. A., Temgoua, E., Wotchoko, P., Boukong, A., &
Bitom, D. (2018). Soil properties and land capability eval-
uation in a mountainous ecosystem of North-West Came-
roon. Journal of Geoscience and Environment Protection,
6(7), 15–33. https:// doi. org/ 10. 4236/ gep. 2018. 67002
Tang, F., Wang, H., Zhang, L., Xu, N., & Ahmad, A. M.
(2023). Adaptive optimized consensus control for a class
of nonlinear multi-agent systems with asymmetric input
saturation constraints and hybrid faults. Communications
in Nonlinear Science and Numerical Simulation, 126,
107446. https:// doi. org/ 10. 1016/j. cnsns. 2023. 107446
Thomas, G. W. (1996). Soil pH and soil acidity. Methods of
soil analysis: part 3 chemical methods, 5, 475–490.
Tunçay, T., Kılıç, Ş, Dedeoğlu, M., Dengiz, O., Başkan, O., & Bay-
ramin, I. (2021). Assessing soil fertility index based on remote
sensing and gis techniques with field validation in a semiarid
agricultural ecosystem. Journal of Arid Environments, 190,
104525. https:// doi. org/ 10. 1016/j. jarid env. 2021. 104525
USDA-NRCS, N. (1996). Soil survey laboratory methods man-
ual. In Soil Survey Investigations Report No. 42. Version
30 USDA-NRCS.
Vinhal-Freitas, I. C., Corrêa, G. F., Wendling, B., Bobuľská,
L., & Ferreira, A. S. (2017). Soil textural class plays a
major role in evaluating the effects of land use on soil
quality indicators. Ecological Indicators, 74, 182–190.
Wang, Y., Huang, H., & Chen, X. (2021). Predicting organic
matter content, total nitrogen and ph value of lime concre-
tion black soil based on visible and near infrared spectros-
copy. Eurasian Soil Science, 54, 1681–1688. https:// doi.
org/ 10. 1134/ S1064 22932 11101 44
Wang, Q., Putri, N. A., Gan, Y., & Song, G. (2022). Combining
both spectral and textural indices for alleviating saturation
problem in forest LAI estimation using Sentinel-2 data.
Geocarto International, 37(25), 10511–10531. https:// doi.
org/ 10. 1016/j. ecoli nd. 2016. 11. 020
Wang, T., Zhang, L., Xu, N., & Alharbi, K. H. (2023). Adaptive
critic learning for approximate optimal event-triggered
tracking control of nonlinear systems with prescribed per-
formances. International Journal of Control, 1–15 https://
doi. org/ 10. 1080/ 00207 179. 2023. 22508 80
Yageta, Y., Osbahr, H., Morimoto, Y., & Clark, J. (2019).
Comparing farmers’ qualitative evaluation of soil fertility
with quantitative soil fertility indicators in Kitui County,
Kenya. Geoderma, 344, 153–163. https:// doi. org/ 10.
1016/j. geode rma. 2019. 01. 019
Yang, R., Song, S., Chen, S., Du, Z., & Kong, J. (2023). Adap-
tive evaluation of green manure rotation for a low fertil-
ity farmland system: Impacts on crop yield, soil nutrients,
and soil microbial community. CATENA, 222, 106873.
https:// doi. org/ 10. 1016/j. catena. 2022. 106873
Yue, S., Niu, B., Wang, H., Zhang, L., & Ahmad, A. M. (2023).
Hierarchical sliding mode-based adaptive fuzzy control
for uncertain switched under-actuated nonlinear systems

Environ Monit Assess (2024) 196:131
1 3
Page 13 of 13 131
Vol.: (0123456789)
with input saturation and dead-zone. Robotic Intelligence
and Automation, 43(5), 523–536.
Zeraatpisheh, M., Jafari, A., Bodaghabadi, M. B., Ayoubi, S.,
Taghizadeh-Mehrjardi, R., Toomanian, N., Kerry, R., &
Xu, M. (2020). Conventional and digital soil mapping in
Iran: Past, present, and future. CATENA, 188, 104424.
https:// doi. org/ 10. 1016/j. catena. 2019. 104424
Zhang, K., Jia, C., Song, Y., Jiang, S., Jiang, Z., Wen, M.,
Huang, Y., Liu, X., Jiang, T., & Peng, J. (2020). Analy-
sis of Lower Cambrian shale gas composition, source and
accumulation pattern in different tectonic backgrounds: A
case study of Weiyuan Block in the Upper Yangtze region
and Xiuwu Basin in the Lower Yangtze region. Fuel, 263,
115978. https:// doi. org/ 10. 1016/j. fuel. 2019. 115978
Zhang, H., Zou, Q., Ju, Y., Song, C., & Chen, D. (2022a). Distance-
based support vector machine to predict DNA N6-methylad-
enine modification. Current Bioinformatics, 17(5), 473–482.
https:// doi. org/ 10. 2174/ 15748 93617 66622 04041 45517
Zhang, K., Jiang, S., Zhao, R., Wang, P., Jia, C., & Song, Y.
(2022b). Connectivity of organic matter pores in the Lower
Silurian Longmaxi Formation shale, Sichuan Basin, South-
ern China: Analyses from helium ion microscope and
focused ion beam scanning electron microscope. Geological
Journal, 57(5), 1912–1924. https:// doi. org/ 10. 1002/ gj. 4387
Zhao, H., Wang, H., Xu, N., Zhao, X., & Sharaf, S. (2023a).
Fuzzy approximation-based optimal consensus control for
nonlinear multiagent systems via adaptive dynamic pro-
gramming. Neurocomputing, 553, 126529. https:// doi. org/
10. 1016/j. neucom. 2023. 126529
Zhao, Y., Niu, B., Zong, G., Zhao, X., & Alharbi, K. H.
(2023b). Neural network-based adaptive optimal contain-
ment control for non-affine nonlinear multi-agent systems
within an identifier-actor-critic framework. Journal of the
Franklin Institute, 360(12), 8118–8143. https:// doi. org/ 10.
1016/j. jfran klin. 2023. 06. 014
Publisher’s Note Springer Nature remains neutral with regard
to jurisdictional claims in published maps and institutional
affiliations.
Springer Nature or its licensor (e.g. a society or other partner)
holds exclusive rights to this article under a publishing
agreement with the author(s) or other rightsholder(s); author
self-archiving of the accepted manuscript version of this article
is solely governed by the terms of such publishing agreement
and applicable law.