Professional Documents
Culture Documents
Mapping Soil Organic Matter Content Using Sentinel-2 Synthetic Images at Different Time Intervals in Northeast China (1)
Mapping Soil Organic Matter Content Using Sentinel-2 Synthetic Images at Different Time Intervals in Northeast China (1)
To cite this article: Chong Luo, Wenqi Zhang, Xinle Zhang & Huanjun Liu (2023)
Mapping soil organic matter content using Sentinel-2 synthetic images at different time
intervals in Northeast China, International Journal of Digital Earth, 16:1, 1094-1107, DOI:
10.1080/17538947.2023.2192005
1. Introduction
Soil organic matter (SOM) can promote the formation of soil aggregates, improve the conditions of
physical, chemical and biological soil processes, and improve the absorption and buffering of soil
(Page, Dang, and Dalal 2020). In addition, the carbon stored in SOM is an important part of the
global carbon cycle and is thus critical for the global carbon balance (Coppola et al. 2022). The accu-
racies of soil quality and regional carbon cycle assessments are also affected by the lack of high-pre-
cision spatial distribution maps of SOM content (Meng et al. 2020; Poggio et al. 2021). Accurate and
rapid mapping of regional SOM content is critical for high-quality management of cultivated land
and sustainable agricultural development.
The SOM content is one of the main factors affecting soil spectral characteristics (Angelopoulou
et al. 2019; Moura-Bueno et al. 2019). Laboratory hyperspectral studies have demonstrated that the
CONTACT Huanjun Liu huanjunliu@yeah.net State Key Laboratory of Black Soils Conservation and Utilization, Northeast
Institute of Geography and Agroecology, Chinese Academy of Sciences, 130102, Changchun, People’s Republic of China
© 2023 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/),
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The terms on which this
article has been published allow the posting of the Accepted Manuscript in a repository by the author(s) or with their consent.
INTERNATIONAL JOURNAL OF DIGITAL EARTH 1095
SOM content is related to the visible shortwave infrared (350-2500 nm) and mid-infrared (2.5-
25 nm μm) wavelengths (Wang et al. 2022a). This is mainly due to the absorption of polar covalent
bonds (OH, CH and NH functional groups) to the infrared frequency (Bahureksa et al. 2021). Many
researchers have developed SOM estimation algorithms based on this finding. Although these
methods can accurately predict the SOM content for a single point, they cannot be used to obtain
SOM contents on a regional scale (Bao et al. 2020; Conforti et al. 2015; Liu et al. 2019). Most com-
monly-used satellites include visible shortwave infrared (350–2500 nm) bands, making it possible
to use remote sensing images to map regional SOM contents.
Based on the relationship between SOM and the soil spectrum, many researchers have begun to
use common satellites to map regional SOM content (Biney et al. 2022; John et al. 2022). In recent
years, the prediction methods used to map SOM have included linear regression and nonlinear
regression (support vector machines, random forest regression, and Cubist models) (Gomes
et al. 2019; Lamichhane, Kumar, and Wilson 2019; Luo et al. 2022a; Mahmoudzadeh et al. 2020;
Pouladi et al. 2019; Tajik, Ayoubi, and Zeraatpisheh 2020). In recent years, some researchers
have used ANN deep-learning algorithms to map SOM contents, showing good prediction per-
formances and model portability (Meng et al. 2022b). In addition to improving SOM prediction
methods, image selection is also an important part of SOM mapping (Demattê et al. 2018; Silvero
et al. 2021). Multitemporal synthetic images have been used to obtain pure bare-soil pixels to avoid
the impact of soil environmental factors on SOM mapping and have their advantages over single
images have been demonstrated (Luo et al. 2022b; Luo et al. 2022c; Wang et al. 2022b). Finally,
the addition of environmental covariates has proven effective (Wang et al. 2022b).
Some recent signs of progress have enabled us to predict the spatial distributions of large-scale
SOM in the bare-soil period. On the one hand, a number of free and high-performance satellites
have been successfully launched (Landsat-9, Sentinel-2) (Claverie et al. 2018; Masek et al. 2020),
allowing us to obtain near real-time, high-spatial-resolution satellite images around the world. On
the other hand, the development of the GEE has enabled many researchers to gain vast computing
power (Dong et al. 2016; Gorelick et al. 2017). Compared to traditional remote sensing image proces-
sing methods, the biggest advantage of the GEE platform is that it can easily process multitemporal
images (Huang et al. 2017; Xiong et al. 2017). GEE has been used to obtain multitemporal synthetic
images for crop classifications, wetland extractions, land use classifications and other research (Jin
et al. 2019; Mahdianpari et al. 2019; Teluguntla et al. 2018; You and Dong 2020). With the continuous
improvements in data availability and computing platforms, more researchers have attempted large-
scale mapping of the spatial distribution of SOM (Luo et al. 2022a; Luo et al. 2022b). The key to large-
scale SOM mapping is selecting an appropriate time window corresponding to bare-soil conditions.
The temporal and spatial resolutions of remote sensing images are mutually restricted (Emilien,
Thomas, and Thomas 2021). Building cloud-free and gap-filled images with proper and regular
time intervals is essential for using high-spatial-resolution images to reconstruct bare-soil time-series
image sets (Griffiths, Nendel, and Hostert 2019). In crop-mapping research, crop types differ from
year to year, so remote sensing image data collected in the same year are usually used (Luo et al.
2022b; Tomppo, Antropov, and Praks 2019). In cultivated soils, the relative stability of the organic
matter content can be ensured without large-scale land remediation (Dou et al. 2019). Therefore,
when constructing regular time interval cloud-free and gap-filled bare soil images, we can use the
same time period synthesis method for many years (Demattê et al. 2018; Silvero et al. 2021). The
key to constructing a bare-soil synthetic image set is selecting the appropriate time interval (Belgiu
and Csillik 2018; You and Dong 2020). The shorter the time interval is, theoretically, the smaller
the soil variation among different years are. However, performing image synthesis with too narrow
an interval may lead to more gaps (due to cloud pollution) (Dong et al. 2016; Griffiths, Nendel,
and Hostert 2019). Therefore, determining the appropriate time interval for performing image syn-
thesis to reconstruct the time series of cloudless images in the bare-soil period is very important.
In this study, farmland in Northeast China was the research object and we used all available Sentinel-
2 images obtained from 2019 to 2022, soil sampling points, different image synthesis strategies, random
1096 C. LUO ET AL.
forest (RF) regression and GEE platforms to determine the best Sentinel-2 synthetic image (the best
time interval and the best time period). The specific objectives are as follows: (1) to determine the
image synthesis time interval that yields the highest SOM spatial distribution map accuracy among
the following choices: 10, 15, 20 d or 30 d; (2) to identify the time window distributions of the synthetic
images at different time intervals; (3) to determine whether adding environmental covariate data
improves the soil mapping accuracy compared to using Sentinel-2 alone. The resulting moderate-
spatial-resolution (10-m) SOM content spatial distribution map of Northeast China can provide the
latest soil chemical quality information and aid in assessments of terrestrial carbon cycle simulations.
Figure 1. Distribution map of sampling points: (a) false-color composite; (b) soil type map; and (c) elevation map.
be used for splitting or regression (Svetnik et al. 2003). The method mainly involves the selection of
decision tree types and the selection of s categories of decision trees according to specific tasks (Cutler
et al. 2007). The random forest regression algorithm is used in scenarios in which the data dimension is
relatively low (tens of dimensions) and the accuracy is high (Strobl, Malley, and Tutz 2009).
The RF regression model in GEE was used and the ‘ee. Classifier. smileRandomForest()’ program
was used to predict SOM. The number of trees was set to 400, mainly to balance the accuracy and
operation efficiency; for VariablesPerSplit, we used the default value of ‘null’; for minLeafPopula-
tion, we use dthe default value of 1; for bagFraction, we used the default value of 0.5; for maxNodes,
we used the default value of ‘null’; and for seed, we used the default value of 0.
3. Results
3.1. SOM mapping using synthetic images at different time intervals
The R2 and RMSE changes in the SOM prediction accuracies for four different time-interval scen-
arios (10, 15, 20 d and 30 d) are shown in Tables 2–5. In all four scenarios, the SOM prediction
accuracy reached the highest value around May. In the S1 scenario, DOY 130 to 140 had the highest
accuracy, with an R2 value of 0.445 and RMSE of 1.430%. In the S2 scenario, DOY 120 to 135 had
the highest accuracy, with an R2 of 0.467 and RMSE of 1.397%. In the S3 scenario, DOY 120 to 140
had the highest accuracy, with an R2 of 0.511 and RMSE of 1.341%. In the S4 scenario, DOY 120 to
150 had the highest accuracy, with an R2 of 0.478 and RMSE of 1.382%. In conclusion, the SOM
prediction accuracies in different time intervals first increased and then decreased, and the highest
SOM was obtained when using the 20-d time interval.
Table 2. SOM prediction accuracies at different time intervals with 10-d increments.
DOY Label R2 RMSE
60–70 20192022060070 0.305 1.603
70–80 20192022070080 0.306 1.601
80–90 20192022080090 0.329 1.567
90–100 20192022090100 0.317 1.575
100–110 20192022100110 0.311 1.589
110–120 20192022110120 0.288 1.646
120–130 20192022120130 0.406 1.481
130–140 20192022130140 0.445 1.430
140–150 20192022140150 0.381 1.494
150–160 20192022150160 0.399 1.490
160–170 20192022160170 0.364 1.474
170–180 20192022170180 0.250 1.706
different time intervals were mainly high values, and the prediction errors of the synthetic image at the
10-d time interval were relatively large in the high-value area (SOM > 6%). After adding environmental
covariates, the absolute SOM prediction error decreased greatly for different SOM content intervals.
4. Discussion
4.1. Importance of selecting appropriate synthetic images with appropriate time intervals
for SOM mapping
Different studies have used synthetic images obtained in different time intervals, including synthetic
images spanning more than one year or single months in many years; it was found that the best option
is to use synthetic images spanning multiple years or single months for mapping SOM (Luo et al.
2022b; Wang et al. 2022b). However, while monthly synthesis data are commonly used in related
research, this time interval is not necessarily the optimal time period for synthesis (Chong et al.
2021; Griffiths, Nendel, and Hostert 2019; You et al. 2021). In studies related to crop mapping, images
from the same year are usually used for synthesis, and the shorter the time interval is, the better the
mapping effect. Soil is more stable than crops, and many studies have used synthetic images taken
over many years to obtain more useful pixels (Mendes et al. 2021; Silvero et al. 2021). The shorter
the time period of the synthetic image is, the closer the collection time of the images in the interval
Table 3. SOM prediction accuracies at different time intervals with 15-d increments.
DOY Label R2 RMSE
60–75 20192022060075 0.282 1.628
75–90 20192022075090 0.325 1.564
90–105 20192022090105 0.366 1.526
105–120 20192022105120 0.375 1.520
120–135 20192022120135 0.467 1.396
135–150 20192022135150 0.457 1.406
150–165 20192022150165 0.392 1.497
165–180 20192022165180 0.275 1.655
1100 C. LUO ET AL.
Table 4. SOM prediction accuracies at different time intervals with 20-d increments.
DOY Label R2 RMSE
60–80 20192022060080 0.274 1.636
80–100 20192022080100 0.279 1.617
100–120 20192022100120 0.375 1.510
120–140 20192022120140 0.511 1.341
140–160 20192022140160 0.406 1.471
160–180 20192022160180 0.357 1.544
Table 5. SOM prediction accuracies at different time intervals with 30-d increments.
DOY Label R2 RMSE
60–90 20192022060090 0.358 1.527
90–120 20192022090120 0.369 1.518
120–150 20192022120150 0.478 1.382
150–180 20192022150180 0.338 1.558
is, but the large-scale phenology, snowmelt time, and plowing time differ, leading synthetic images
with different states in different regions, for example, in mid-April in Northeast China (Yang et al.
2022). At this time, the Sanjiang Plain in the northern area of the region has just completed the snow-
melt period, and the soil moisture content is very high, while the soil on the Liaohe Plain in the south
of Northeast China is relatively dry (Qi et al. 2021). For relatively long synthesis time intervals, the
median value of the image set was used to obtain an image with the same relative state in different
regions of Northeast China, but longer time intervals may introduce more redundant information,
thus resulting in deviations in the optimal state of the synthetic image. We found that when different
time intervals were used, the 20-day interval synthetic image had the best SOM prediction accuracy,
possibly because the 20-day interval balances the above two problems.
4.3. Effects of environmental covariate data and remote sensing data on SOM predictions
Figure 5 shows that climate covariates and terrain covariates are of high importance in SOM pre-
dictions. Pedogenesis postulates that natural soil-forming factors and man-made soil-forming fac-
tors work together to affect soil physical and chemical properties, including climate, biology, parent
INTERNATIONAL JOURNAL OF DIGITAL EARTH 1101
Figure 2. Box plots of SOM prediction accuracies for different scenarios: (a) 10d_20192022130140; (b) 15d_20192022120135; (c)
20d_20192022120140; (d) 30d_20192022120150; and (e) 20d_20192022120140 after adding environmental covariates.
material, topography, hydrology, time, cultivation, fertilization, and irrigation factors (Lin 2011;
Minasny, McBratney, and Salvador-Blanes 2008). Environmental covariates can optimally rep-
resent natural soil formation factors, and our research also showed that environmental covariates
were very important for predicting SOM (Table 6 and Figure 5). However, it is difficult to use
environmental covariates to monitor anthropogenic soil-formation factors (Hengl et al. 2014).
Remote sensing image data can be used to directly observe the surface of bare soil, especially in
Northeast China, where the bare-soil period is long (Dou et al. 2019; Luo et al. 2022c). The spatial
resolution of the Sentinel-2 remote sensing images was better than those of the environmental cov-
ariates, and these two data types can be combined to obtain a SOM spatial distribution map with a
relatively high spatial resolution (Drusch et al. 2012a). Therefore, remote sensing data and environ-
mental covariate data have their own advantages in SOM predictions and combining environmental
covariate and remote sensing data is a future direction of SOM prediction research.
1102 C. LUO ET AL.
Figure 4. Changes in the SOM prediction accuracies of different synthetic image intervals: (a) 10-d interval; (b) 15-d interval; (c)
20-d interval; and (d) 30-d interval.
INTERNATIONAL JOURNAL OF DIGITAL EARTH 1103
5. Conclusion
The results presented herein indicated that the 20-d synthetic image interval is a suitable time inter-
val for mapping SOM, and the accuracy of the results obtained with an excessively long or short
time interval was decreased. In addition, the best time period for predicting SOM was always within
May. The consideration of environmental covariates can effectively improve the SOM mapping
accuracy, and information regarding environmental covariates should be fully incorporated in
1104 C. LUO ET AL.
future research. Based on the 10-m spatial resolution SOM map obtained in this study, researchers
can improve the quality of agricultural ecosystem modeling, and government managers can rely on
these high spatial resolution SOM spatial data. The distribution map obtained herein can be applied
to optimize farmland protection and agricultural development strategies in different regions.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Funding
This work was supported by the Strategic Priority Research Program of the Chinese Academy of Sciences
(XDA28100000), the K. C. Wong Education Foundation, Jilin Provincial Development and Reform Commission
Innovation Capacity Building Project (grant number 2021C044-10) and the Special fund project for high-tech indus-
trialization of science and technology cooperation between Jilin Province and the Chinese Academy of Sciences
(2021SYHZ0013).
References
Angelopoulou, T., N. Tziolas, A. Balafoutis, G. Zalidis, and D. Bochtis. 2019. “Remote Sensing Techniques for Soil
Organic Carbon Estimation: A Review.” Remote Sensing 11: 676. doi:10.3390/rs11060676
Bahureksa, W., M. M. Tfaily, R. M. Boiteau, R. B. Young, M. N. Logan, A. M. McKenna, and T. Borch. 2021. “Soil
Organic Matter Characterization by Fourier Transform ion Cyclotron Resonance Mass Spectrometry (FTICR
MS): A Critical Review of Sample Preparation, Analysis, and Data Interpretation.” Environmental Science &
Technology 55: 9637–9656. doi:10.1021/acs.est.1c01135
Bao, Y., X. Meng, S. Ustin, X. Wang, X. Zhang, H. Liu, and H. Tang. 2020. “Vis-SWIR Spectral Prediction Model for
Soil Organic Matter with Different Grouping Strategies.” CATENA 195: 104703. doi:10.1016/j.catena.2020.104703
Belgiu, M., and O. Csillik. 2018. “Sentinel-2 Cropland Mapping Using Pixel-Based and Object-Based Time-Weighted
Dynamic Time Warping Analysis.” Remote Sensing of Environment 204: 509–523. doi:10.1016/j.rse.2017.10.005
Biney, J. K. M., R. Vašát, S. M. Bell, N. M. Kebonye, A. Klement, K. John, and L. Borůvka. 2022. “Prediction of Topsoil
Organic Carbon Content with Sentinel-2 Imagery and Spectroscopic Measurements Under Different Conditions
Using an Ensemble Model Approach with Multiple pre-Treatment Combinations.” Soil and Tillage Research 220:
105379. doi:10.1016/j.still.2022.105379
Breiman, L. 2001. “Random Forests.” Machine Learning 45: 5–32. doi:10.1023/A:1010933404324
Chong, L., H.-J. Liu, L.-P. Lu, Z.-R. Liu, F.-C. Kong, and X.-L. Zhang. 2021. “Monthly Composites from Sentinel-1
and Sentinel-2 Images for Regional Major Crop Mapping with Google Earth Engine.” Journal of Integrative
Agriculture 20: 1944–1957. doi:10.1016/S2095-3119(20)63329-9
Claverie, M., J. Ju, J. G. Masek, J. L. Dungan, E. F. Vermote, J.-C. Roger, S. V. Skakun, and C. Justice. 2018. “The
Harmonized Landsat and Sentinel-2 Surface Reflectance Data set.” Remote Sensing of Environment 219:
145–161. doi:10.1016/j.rse.2018.09.002
Conforti, M., A. Castrignanò, G. Robustelli, F. Scarciglia, M. Stelluti, and G. Buttafuoco. 2015. “Laboratory-based
Vis–NIR Spectroscopy and Partial Least Square Regression with Spatially Correlated Errors for Predicting
Spatial Variation of Soil Organic Matter Content.” Catena 124: 60–67. doi:10.1016/j.catena.2014.09.004
Coppola, A. I., S. Wagner, S. T. Lennartz, M. Seidel, N. D. Ward, T. Dittmar, C. Santín, and M. W. Jones. 2022. “The
Black Carbon Cycle and its Role in the Earth System.” Nature Reviews Earth & Environment 3: 516–532. doi:10.
1038/s43017-022-00316-6
Cucchi, M., G. P. Weedon, A. Amici, N. Bellouin, S. Lange, H. Müller Schmied, H. Hersbach, and C. Buontempo.
2020. “WFDE5: Bias-Adjusted ERA5 Reanalysis Data for Impact Studies.” Earth System Science Data 12:
2097–2120. doi:10.5194/essd-12-2097-2020
INTERNATIONAL JOURNAL OF DIGITAL EARTH 1105
Cutler, D. R., T. C. Edwards, Jr., K. H. Beard, A. Cutler, and K. T. Hess. 2007. “Random Forests for Classification in
Ecology.” Ecology 88: 2783–2792. doi:10.1890/07-0539.1
Demattê, J. A. M., C. T. Fongaro, R. Rizzo, and J. L. Safanelli. 2018. “Geospatial Soil Sensing System (GEOS3):
A Powerful Data Mining Procedure to Retrieve Soil Spectral Reflectance from Satellite Images.” Remote Sensing
of Environment 212: 161–175. doi:10.1016/j.rse.2018.04.047
Dong, J., X. Xiao, M. A. Menarguez, G. Zhang, Y. Qin, D. Thau, C. Biradar, and B. Moore. 2016. “Mapping Paddy
Rice Planting Area in Northeastern Asia with Landsat 8 Images, Phenology-Based Algorithm and Google Earth
Engine.” Remote Sensing of Environment 185: 142–154. doi:10.1016/j.rse.2016.02.016
Dou, X., X. Wang, H. Liu, X. Zhang, L. Meng, Y. Pan, Z. Yu, and Y. Cui. 2019. “Prediction of Soil Organic Matter
Using Multi-Temporal Satellite Images in the Songnen Plain, China.” Geoderma 356: 113896. doi:10.1016/j.
geoderma.2019.113896
Drusch, M., U. Del Bello, S. Carlier, O. Colin, V. Fernandez, F. Gascon, B. Hoersch, et al. 2012b. “Sentinel-2: ESA’s
Optical High-Resolution Mission for GMES Operational Services.” Remote Sensing of Environment 120: 25–36.
doi:10.1016/j.rse.2011.11.026
Drusch, M., U. Del Bello, S. Carlier, O. Colin, V. Fernandez, F. Gascon, B. Hoersch, C. Isola, P. Laberinti, and P.
Martimort. 2012a. “Sentinel-2: ESA’s Optical High-Resolution Mission for GMES Operational Services.”
Remote Sensing of Environment 120: 25–36. doi:10.1016/j.rse.2011.11.026
Emilien, A.-V., C. Thomas, and H. Thomas. 2021. “UAV & Satellite Synergies for Optical Remote Sensing
Applications: A Literature Review.” Science of Remote Sensing 3: 100019. doi:10.1016/j.srs.2021.100019
Gomes, L. C., R. M. Faria, E. de Souza, G. V. Veloso, C. E. G. Schaefer, and E. I. Fernandes Filho. 2019.
“Modelling and Mapping Soil Organic Carbon Stocks in Brazil.” Geoderma 340: 337–350. doi:10.1016/j.
geoderma.2019.01.007
Gorelick, N., M. Hancher, M. Dixon, S. Ilyushchenko, D. Thau, and R. Moore. 2017. “Google Earth Engine:
Planetary-Scale Geospatial Analysis for Everyone.” Remote Sensing of Environment 202: 18–27. doi:10.1016/j.
rse.2017.06.031
Griffiths, P., C. Nendel, and P. Hostert. 2019. “Intra-annual Reflectance Composites from Sentinel-2 and Landsat for
National-Scale Crop and Land Cover Mapping.” Remote Sensing of Environment 220: 135–151. doi:10.1016/j.rse.
2018.10.031
Hengl, T., J. M. de Jesus, R. A. MacMillan, N. H. Batjes, G. B. Heuvelink, E. Ribeiro, A. Samuel-Rosa, B. Kempen, J. G.
Leenaars, and M. G. Walsh. 2014. “SoilGrids1km—Global Soil Information Based on Automated Mapping.” PloS
one 9: e105992. doi:10.1371/journal.pone.0105992
Huang, H., Y. Chen, N. Clinton, J. Wang, X. Wang, C. Liu, P. Gong, et al. 2017. “Mapping Major Land Cover
Dynamics in Beijing Using all Landsat Images in Google Earth Engine.” Remote Sensing of Environment 202:
166–176. doi:10.1016/j.rse.2017.02.021
Jin, Z., G. Azzari, C. You, S. Di Tommaso, S. Aston, M. Burke, and D. B. Lobell. 2019. “Smallholder Maize Area and
Yield Mapping at National Scales with Google Earth Engine.” Remote Sensing of Environment 228: 115–128.
doi:10.1016/j.rse.2019.04.016
John, K., Y. Bouslihim, K. I. Ofem, L. Hssaini, R. Razouk, P. B. Okon, I. A. Isong, P. C. Agyeman, N. M. Kebonye, and
C. Qin. 2022. “Do Model Choice and Sample Ratios Separately or Simultaneously Influence Soil Organic Matter
Prediction?” International Soil and Water Conservation Research 10: 470–486. doi:10.1016/j.iswcr.2021.11.003
Kulu, E. 2021. “Satellite Constellations-2021 Industry Survey and Trends.” 35th Annual Small Satellite Conference.
Lamichhane, S., L. Kumar, and B. Wilson. 2019. “Digital Soil Mapping Algorithms and Covariates for Soil Organic
Carbon Mapping and Their Implications: A Review.” Geoderma 352: 395–413. doi:10.1016/j.geoderma.2019.05.
031
Lin, H. 2011. “Three Principles of Soil Change and Pedogenesis in Time and Space.” Soil Science Society of America
Journal 75: 2049–2070. doi:10.2136/sssaj2011.0130
Liu, S., H. Shen, S. Chen, X. Zhao, A. Biswas, X. Jia, Z. Shi, and J. Fang. 2019. “Estimating Forest Soil Organic Carbon
Content Using vis-NIR Spectroscopy: Implications for Large-Scale Soil Carbon Spectroscopic Assessment.”
Geoderma 348: 37–44. doi:10.1016/j.geoderma.2019.04.003
Luo, C., Y. Wang, X. Zhang, W. Zhang, and H. Liu. 2022a. “Spatial Prediction of Soil Organic Matter Content Using
Multiyear Synthetic Images and Partitioning Algorithms.” CATENA 211: 106023. doi:10.1016/j.catena.2022.106023
Luo, C., X. Zhang, X. Meng, H. Zhu, C. Ni, M. Chen, and H. Liu. 2022b. “Regional Mapping of Soil Organic Matter
Content Using Multitemporal Synthetic Landsat 8 Images in Google Earth Engine.” CATENA 209: 105842. doi:10.
1016/j.catena.2021.105842
Luo, C., X. Zhang, Y. Wang, Z. Men, and H. Liu. 2022c. “Regional Soil Organic Matter Mapping Models Based on the
Optimal Time Window, Feature Selection Algorithm and Google Earth Engine.” Soil and Tillage Research 219:
105325. doi:10.1016/j.still.2022.105325
Mahdianpari, M., B. Salehi, F. Mohammadimanesh, S. Homayouni, and E. Gill. 2019. “The First Wetland Inventory
map of Newfoundland at a Spatial Resolution of 10 m Using Sentinel-1 and Sentinel-2 Data on the Google Earth
Engine Cloud Computing Platform.” Remote Sensing 11: 43. doi:10.3390/rs11010043
1106 C. LUO ET AL.
Mahmoudzadeh, H., H. R. Matinfar, R. Taghizadeh-Mehrjardi, and R. Kerry. 2020. “Spatial Prediction of Soil
Organic Carbon Using Machine Learning Techniques in Western Iran.” Geoderma Regional 21: e00260. doi:10.
1016/j.geodrs.2020.e00260
Masek, J. G., M. A. Wulder, B. Markham, J. McCorkel, C. J. Crawford, J. Storey, and D. T. Jenstrom. 2020. “Landsat 9:
Empowering Open Science and Applications Through Continuity.” Remote Sensing of Environment 248: 111968.
doi:10.1016/j.rse.2020.111968
Mendes, W.d.S., J. A. M. Demattê, N. E. Q. Silvero, and L. Rabelo Campos. 2021. “Integration of Multispectral and
Hyperspectral Data to map Magnetic Susceptibility and Soil Attributes at Depth: A Novel Framework.” Geoderma
385: 114885. doi:10.1016/j.geoderma.2020.114885
Meng, X., Y. Bao, J. Liu, H. Liu, X. Zhang, Y. Zhang, P. Wang, H. Tang, and F. Kong. 2020. “Regional Soil Organic
Carbon Prediction Model Based on a Discrete Wavelet Analysis of Hyperspectral Satellite Data.” International
Journal of Applied Earth Observation and Geoinformation 89: 102111. doi:10.1016/j.jag.2020.102111
Meng, X., Y. Bao, H. Liu, X. Zhang, and X. Wang. 2022a. “A new Digital Soil Mapping Method with Temporal-
Spatial-Spectral Information Derived from Multi-Source Satellite Images.” Geoderma 425: 116065. doi:10.1016/
j.geoderma.2022.116065
Meng, X., Y. Bao, Y. Wang, X. Zhang, and H. Liu. 2022b. “An Advanced Soil Organic Carbon Content Prediction
Model via Fused Temporal-Spatial-Spectral (TSS) Information Based on Machine Learning and Deep Learning
Algorithms.” Remote Sensing of Environment 280: 113166. doi:10.1016/j.rse.2022.113166
Minasny, B., A. B. McBratney, and S. Salvador-Blanes. 2008. “Quantitative Models for Pedogenesis—A Review.”
Geoderma 144: 140–157. doi:10.1016/j.geoderma.2007.12.013
Moura-Bueno, J. M., R. S. D. Dalmolin, A. ten Caten, A. C. Dotto, and J. A. Demattê. 2019. “Stratification of a Local
VIS-NIR-SWIR Spectral Library by Homogeneity Criteria Yields More Accurate Soil Organic Carbon
Predictions.” Geoderma 337: 565–581. doi:10.1016/j.geoderma.2018.10.015
NASA, J. 2020. NASADEM Merged DEM Global 1 arc Second V001 [Data Set]. NASA EOSDIS Land Processes
DAAC. Accessed July 08, 2020. doi:10.5067/MEaSUREs/NASADEM.
Nelson, D. W., and L. Sommers. 2013. “A Rapid and Accurate Procedure for Estimation of Organic Carbon in Soils.”
Proceedings of the Indiana Academy of Science 84: 456–462.
O’Kelly, B. C. 2004. “Accurate Determination of Moisture Content of Organic Soils Using the Oven Drying Method.”
Drying Technology 22: 1767–1776. doi:10.1081/DRT-200025642
Ou, Y., A. N. Rousseau, L. Wang, and B. Yan. 2017. “Spatio-temporal Patterns of Soil Organic Carbon and pH in
Relation to Environmental Factors—A Case Study of the Black Soil Region of Northeastern China.”
Agriculture, Ecosystems & Environment 245: 22–31. doi:10.1016/j.agee.2017.05.003
Page, K. L., Y. P. Dang, and R. C. Dalal. 2020. “The Ability of Conservation Agriculture to Conserve Soil Organic
Carbon and the Subsequent Impact on Soil Physical, Chemical, and Biological Properties and Yield.” Frontiers
in Sustainable Food Systems 4: 31. doi:10.3389/fsufs.2020.00031
Poggio, L., L. M. De Sousa, N. H. Batjes, G. Heuvelink, B. Kempen, E. Ribeiro, and D. Rossiter. 2021. “SoilGrids 2.0:
Producing Soil Information for the Globe with Quantified Spatial Uncertainty.” Soil 7: 217–240. doi:10.5194/soil-
7-217-2021
Pouladi, N., A. B. Møller, S. Tabatabai, and M. H. Greve. 2019. “Mapping Soil Organic Matter Contents at Field Level
with Cubist, Random Forest and Kriging.” Geoderma 342: 85–92. doi:10.1016/j.geoderma.2019.02.019
Qi, W., L. Feng, H. Yang, and J. Liu. 2021. “Spring and Summer Potential Flood Risk in Northeast China.” Journal of
Hydrology: Regional Studies 38: 100951. doi:10.1016/j.ejrh.2021.100951
Qian, S.-E. 2021. “Hyperspectral Satellites, Evolution, and Development History.” IEEE Journal of Selected Topics in
Applied Earth Observations and Remote Sensing 14: 7032–7056. doi:10.1109/JSTARS.2021.3090256
Saptoro, A., M. O. Tadé, and H. Vuthaluru. 2012. “A Modified Kennard-Stone Algorithm for Optimal Division of
Data for Developing Artificial Neural Network Models.” Chemical Product and Process Modeling 7. doi:10.
1515/1934-2659.1645
Shafizadeh-Moghadam, H., F. Minaei, H. Talebi-khiyavi, T. Xu, and M. Homaee. 2022. “Synergetic use of Multi-
Temporal Sentinel-1, Sentinel-2, NDVI, and Topographic Factors for Estimating Soil Organic Carbon.”
CATENA 212: 106077. doi:10.1016/j.catena.2022.106077
Silvero, N. E. Q., J. A. M. Demattê, M. T. A. Amorim, N.V.d. Santos, R. Rizzo, J. L. Safanelli, R. R. Poppiel, W. d. S.
Mendes, and B. R. Bonfatti. 2021. “Soil Variability and Quantification Based on Sentinel-2 and Landsat-8 Bare Soil
Images: A Comparison.” Remote Sensing of Environment 252: 112117. doi:10.1016/j.rse.2020.112117
Stevens, A., and L. Ramirez-Lopez. 2014. “An Introduction to the Prospectr Package.” R Package Vignette, Report
No.: R Package Version 0.1 3.
Strobl, C., J. Malley, and G. Tutz. 2009. “An Introduction to Recursive Partitioning: Rationale, Application, and
Characteristics of Classification and Regression Trees, Bagging, and Random Forests.” Psychological Methods
14: 323–348. doi:10.1037/a0016973
Svetnik, V., A. Liaw, C. Tong, J. C. Culberson, R. P. Sheridan, and B. P. Feuston. 2003. “Random Forest: A
Classification and Regression Tool for Compound Classification and QSAR Modeling.” Journal of Chemical
Information and Computer Sciences 43: 1947–1958. doi:10.1021/ci034160g
INTERNATIONAL JOURNAL OF DIGITAL EARTH 1107
Tajik, S., S. Ayoubi, and M. Zeraatpisheh. 2020. “Digital Mapping of Soil Organic Carbon Using Ensemble Learning
Model in Mollisols of Hyrcanian Forests, Northern Iran.” Geoderma Regional 20: e00256. doi:10.1016/j.geodrs.
2020.e00256
Teluguntla, P., P. S. Thenkabail, A. Oliphant, J. Xiong, M. K. Gumma, R. G. Congalton, K. Yadav, and A. Huete. 2018.
“A 30-m Landsat-Derived Cropland Extent Product of Australia and China Using Random Forest Machine
Learning Algorithm on Google Earth Engine Cloud Computing Platform.” Isprs Journal of Photogrammetry
and Remote Sensing 144: 325–340. doi:10.1016/j.isprsjprs.2018.07.017
Tomppo, E., O. Antropov, and J. Praks. 2019. “Cropland Classification Using Sentinel-1 Time Series: Methodological
Performance and Prediction Uncertainty Assessment.” Remote Sensing 11: 2480. doi:10.3390/rs11212480
Wang, S., K. Guan, C. Zhang, D. Lee, A. J. Margenot, Y. Ge, J. Peng, W. Zhou, Q. Zhou, and Y. Huang. 2022a. “Using
Soil Library Hyperspectral Reflectance and Machine Learning to Predict Soil Organic Carbon: Assessing Potential
of Airborne and Spaceborne Optical Soil Sensing.” Remote Sensing of Environment 271: 112914. doi:10.1016/j.rse.
2022.112914
Wang, X., L. Wang, S. Li, Z. Wang, M. Zheng, and K. Song. 2022b. “Remote Estimates of Soil Organic Carbon Using
Multi-Temporal Synthetic Images and the Probability Hybrid Model.” Geoderma 425: 116066. doi:10.1016/j.
geoderma.2022.116066
Wang, L., X. Wang, D. Wang, B. Qi, S. Zheng, H. Liu, C. Luo, H. Li, L. Meng, and X. Meng. 2021. “Spatiotemporal
Changes and Driving Factors of Cultivated Soil Organic Carbon in Northern China’s Typical Agro-Pastoral
Ecotone in the Last 30 Years.” Remote Sensing 13: 3607. doi:10.3390/rs13183607
Wu, X., R. Zhu, Y. Long, and W. Zhang. 2022. “Spatial Trend and Impact of Snowmelt Rate in Spring Across China’s
Three Main Stable Snow Cover Regions Over the Past 40 Years Based on Remote Sensing.” Remote Sensing 14:
4176. doi:10.3390/rs14174176
Xiong, J., P. S. Thenkabail, M. K. Gumma, P. Teluguntla, J. Poehnelt, R. G. Congalton, K. Yadav, and D. Thau. 2017.
“Automated Cropland Mapping of Continental Africa Using Google Earth Engine Cloud Computing.” Isprs
Journal of Photogrammetry and Remote Sensing 126: 225–244. doi:10.1016/j.isprsjprs.2017.01.019
Yang, Y., R. Chen, G. Liu, Z. Liu, and X. Wang. 2022. “Trends and Variability in Snowmelt in China Under Climate
Change.” Hydrology and Earth System Sciences 26: 305–329. doi:10.5194/hess-26-305-2022
You, N., and J. Dong. 2020. “Examining Earliest Identifiable Timing of Crops Using all Available Sentinel 1/2 Imagery
and Google Earth Engine.” ISPRS Journal of Photogrammetry and Remote Sensing 161: 109–123. doi:10.1016/j.
isprsjprs.2020.01.001
You, N., J. Dong, J. Huang, G. Du, G. Zhang, Y. He, T. Yang, Y. Di, and X. Xiao. 2021. “The 10-m Crop Type Maps in
Northeast China During 2017–2019.” Scientific Data 8: 41. doi:10.1038/s41597-021-00827-9
Zhou, T., Y. Geng, J. Chen, J. Pan, D. Haase, and A. Lausch. 2020. “High-resolution Digital Mapping of Soil Organic
Carbon and Soil Total Nitrogen Using DEM Derivatives, Sentinel-1 and Sentinel-2 Data Based on Machine
Learning Algorithms.” Science of The Total Environment 729: 138244. doi:10.1016/j.scitotenv.2020.138244