Next Article in Journal
Gully Head-Cuts Inventory and Semi-Automatic Gully Extraction Using LiDAR and Topographic Openness—Case Study: Covurlui Plateau, Eastern Romania
Previous Article in Journal
A Multi-Attribute Approach for Low-Carbon and Intensive Land Use of Jinan, China
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Land
Volume 12
Issue 6
10.3390/land12061198
land-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Mapping Soil Organic Carbon in Floodplain Farmland: Implications of Effective Range of Environmental Variables

by
Zihao Wu
1,2,3,
Yiyun Chen
4,
Yuanli Zhu
1,2,3,
Xiangyang Feng
2,
Jianxiong Ou
2,
Guie Li
1,2,
Zhaomin Tong
4 and
Qingwu Yan
1,2,*
1
Research Center for Transformation Development and Rural Revitalization of Resource-Based Cities in China, China University of Mining and Technology, Xuzhou 221116, China
2
School of Public Policy and Management, China University of Mining and Technology, Xuzhou 221116, China
3
Research Center for Land Use and Ecological Security Governance in Mining Area, China University of Mining and Technology, Xuzhou 221116, China
4
School of Resource and Environmental Science, Wuhan University, Wuhan 430079, China
*
Author to whom correspondence should be addressed.
Land 2023, 12(6), 1198; https://doi.org/10.3390/land12061198
Submission received: 14 May 2023 / Revised: 1 June 2023 / Accepted: 5 June 2023 / Published: 8 June 2023
Browse Figures
Versions Notes

Abstract

:
Accurately mapping soil organic carbon (SOC) is conducive to evaluating carbon storage and soil quality. However, the high spatial heterogeneity of SOC caused by river-related factors and agricultural management brings challenges to digital soil mapping in floodplain farmland. Moreover, current studies focus on the non-linear relationship between SOC and covariates, but ignore the effective range of environmental variables on SOC, which prevents the revelation of the SOC differentiation mechanism. Using the 375 samples collected from the Jiangchang Town near Han River, we aim to determine the main controlling factors of SOC, reveal the effective range of environmental variables, and obtain the spatial map of SOC by using the gradient boosting decision tree (GBDT) model and partial dependence plots. Linear regression was used as a reference. Results showed that GBDT outperformed linear regression. GBDT results show that the distance from the river was the most important SOC factor, confirming the importance of the Han River to the SOC pattern. The partial dependence plots indicate that all environmental variables have their effective ranges, and when their values are extremely high or low, they do not respond to changes in SOC. Specifically, the influential ranges of rivers, irrigation canals, and rural settlements on SOC were within 4000, 200, and 50 m, respectively. The peak SOC was obtained with high clay (≥31%), total nitrogen (≥1.18 g/kg), and total potassium contents (≥11.1 g/kg), but it remained steady when these covariates further increased. These results highlight the importance of revealing the effective range of environmental variables, which provides data support for understanding the spatial pattern of SOC in floodplain farmland, achieving carbon sequestration in farmland and precision agriculture. The GBDT with the partial dependence plot was effective in SOC fitting and mapping.

Graphical Abstract

1. Introduction

Floodplain farmland is one of the most productive agricultural ecosystems due to its fertile alluvial soil [1,2,3]. Seasonal flooding causes fertile sediment to settle in low-lying areas and sufficient irrigation water sources promote crop growth, allowing many crop residues to enter the soil [4,5,6]. Consequently, floodplain farmland has great potential for soil carbon sequestration. The soil organic carbon (SOC) content is a key indicator of soil fertility [7], and accurately mapping the SOC in floodplain farmland is conducive to evaluating the carbon storage and the soil quality.
The current digital soil mapping method often uses easily-obtained Scorpan factors to predict the SOC [8,9,10]. This method fits the relationship between SOC and environmental variables and then applies this relationship to predict the SOC at the unsampled region. Soil types, climate, land use type, vegetation index, terrain, parent materials, and distance factors have been widely used in SOC mapping [11,12,13,14,15]. However, many soil properties, such as soil texture, nutrient content, and pH, are associated with the SOC content [16,17,18], but they are rarely used as covariates in SOC mapping due to the lack of spatial distribution maps. With the launch of the GlobalSoilMap.net project [19], many national or global soil maps have been estimated and shared, such as the SoilGrids dataset (250 m resolution) [20] and the China High Resolution National Soil Information Grid Basic Attribute Dataset (90 m resolution) [21]. These global or national high-resolution soil property datasets may help improve the accuracy of regional SOC mapping.
In addition to the above-mentioned soil forming factors, rivers also control the process and spatial pattern of the SOC in the floodplain. River flooding redistributes the soil properties through the process of erosion and sedimentation [4,5]. The effect of rivers on soil properties weakens as the distance increases. Several studies found a significant correlation between the distance from the river (Dis_River) and the soil texture [22], the pH [23], the salinity [24], and the sulfur [25]. However, the relationship between SOC and Dis_River remains controversial; some studies found a positive correlation between them [1,26,27,28], while some found a negative correlation [29,30]. Therefore, clarifying the relationship between SOC and Dis_River will help elucidate the spatial pattern of SOC in floodplain.
Many studies have found that environmental variables only have an impact on SOC within a certain range [31,32,33,34]. For example, Chen, et al. [34] found that the amount of soil loss rapidly decreases when the vegetation coverage increases from 0 to 80%, but the amount of soil loss remains basically unchanged when the vegetation coverage increases from 80% to 100%. Wang, et al. [33] found that potential CO2 production in soil is stimulated at intermediate salinities (5‰ to 7.5‰) but inhibited by salinities ≥15‰. These studies indicate that identifying the effective range of environmental variables is crucial for optimizing agricultural management to achieve carbon sequestration in soil. However, such a complex non-linear relationship is difficult to fit well with a simple linear, logarithmic or power function.
The gradient boosting decision tree (GBDT) model with a partial dependence plot is an effective tool for determining the effective range of covariates. Specifically, the GBDT model is a boosting ensemble learning model based on the CART algorithm with high prediction accuracy [35,36]. In some cases, GBDT outperforms the support vector machine, random forest, Cubist, and extreme gradient boosting models [37,38,39,40]. Moreover, the non-linear relationship explored by the GBDT model can be visualized by using a partial dependence plot [41,42,43], which is conducive to determining the effective range of environmental variables, understanding the spatial pattern of SOC, and optimizing agricultural management.
Jiangchang Town is a typical alluvial plain with hundreds of years of farming history. The region is adjacent to the Han River, experiencing occasionally flooding during summer [44]. The joint effect of natural environment and agricultural practice leads to the strong spatial heterogeneity of the SOC in the region. Therefore, this study aims to determine the main controlling factors of SOC, reveal the effective range of environmental variables, and obtain the spatial map of SOC in the region by using the GBDT model and a partial dependence plot.

2. Materials and Methods

2.1. Study Area

Jiangchang Town (30°31′–30°39′ N and 112°52′–113°01′ E) is located in the central part of Jianghan Plain, China (Figure 1). The region covers approximately 79.80 km2, and its elevation ranges from −14–112 m. The land use pattern contains large areas of irrigated land, a few paddy fields, scattered rural settlements, ponds, greenlands (i.e., woodland and grassland), and developed irrigation canals. The study area has a typical subtropical monsoon climate. The mean annual temperature (MAT) is approximately 16.2 °C, and the mean annual precipitation (MAP) is roughly 1100 mm. The region is dominated by Fluvisols and contains a few Gleysols. The study area is a typical alluvial plain and is adjacent to the Han River. The Han River provides sufficient water for agricultural irrigation. However, the terrain of the Han River is relatively high, resulting in frequent floods in summer.
  • RS: rural settlement.

2.2. SOC Content of Sampling Points

A total of 375 topsoil farmland samples (0–30 cm) were collected in 2015 (Figure 1), including 36 paddy fields and 339 irrigated land points. The average sampling density is 6.65 samples per km2. The samples were air-dried in a laboratory and then gently crushed to pass through a 20-mesh nylon sieve. The SOC content was measured by using the combustion oxidation-titration method according to Chinese National Environmental Protection Standards (HJ 658-2013) [45].
The histograms of the SOC content are shown in Figure 2. The SOC content ranges from 4.8 g/kg to 23.0 g/kg. The mean and the median are close, namely, 12.05 and 11.80, respectively. The SOC content is normally distributed, and its skewness coefficient is 0.387. The coefficient of variation (CV) is 27.42%, indicting a moderate level of variation [46].

2.3. Environmental Variables

On the basis of the Scorpan model, we selected 15 environmental variables of SOC from the aspects of soil properties, relief, organism, and distance factor (Table 1). The soil type, MAT, and MAP variation in the study area is very small, so they are excluded from the environmental variable set.
The land use map (30 m) in 2015 was obtained based on the supervised classification of Landsat 8 images (28 July 2015, http://www.gscloud.cn, accessed on 1 March 2023) (Figure 1). Four distance factors, including the distance from the nearest river, rural settlement, irrigated canal, and pond, were calculated based on the location of the sampling points and the land use map. The elevation, the slope, and the topographic wetness index (TWI) [47] were obtained and calculated based on the ASTER DEM data (30 m, http://www.gscloud.cn, accessed on 1 March 2023) and are shown in Figure 3a–c, respectively. The time-series normalized difference vegetation index (NDVI) images were calculated by using all Landsat 5/8 images in 2015, and the maximum NDVI value for each pixel was obtained to generate the maximum NDVI map (30 m). The maximum NDVI map in 2015 reflected the crop growth and was acquired from the Chinese Resource and Environment Science and Data Center (https://www.resdc.cn/, accessed on 1 March 2023).
The total nitrogen (TN), total phosphorus (TP), and total potassium (TK) contents reflect the soil nutrients and the fertilization situation, while the pH reflects the soil acidity and alkalinity. The data were obtained from the Soil SubCenter, National Earth System Science Data Center, National Science and Technology Infrastructure of China (2010–2018) (http://soil.geodata.cn, accessed on 1 May 2022) (Figure 3g–j). The soil texture, including the clay, silt, and sand contents, was obtained from the SoilGrids dataset provided by the World Soil Information Service (Figure 3d–f) [48].
The mapping of the environmental variables was completed using the ArcGIS software platform.

2.4. GBDT Model

GBDT is a typical boosting ensemble learning model based on the CART algorithm [35,36]. The GBDT model is an additive model that serially trains a set of CART regression trees (Equation (1)). The fitting result of each tree should be as near as possible to the residual between the current prediction result and the real value (Equation (2)), that is, the negative gradient direction of the current loss function (Equation (3)). The fitting process ends when the prediction deviation is lower than the threshold. The final prediction result is equal to the sum of the prediction results of all the decision trees.
F T x = ∑ t = 1 T f t x ,
f t x = Y − F T − 1 x ,
L Y , F T x = [ Y − F T x ] 2 ,
where F T x is the prediction result that contains T decision trees, f t x is the tth decision tree, Y is the real value, and L Y , F T x is the loss function.
In this study, a 10-fold cross-validation procedure was used to obtain the best interaction depth and number of decision trees, and the two parameters were used in training the GBDT model. The partial dependence plots were then used to show the non-linear relationship and effective ranges of the environmental variables. The modeling and visualization of the GBDT model were completed in the R studio software by using the gbm and pdp packages. Univariate linear regression (ULR) was used as a reference.

3. Results

3.1. Impact of Rivers on Soil Properties

Figure 3 indicates that various soil properties changed with the distance from the rivers. Specifically, the clay content, the TN, the TP, and the TK increased with the distance from the Han River, while the silt content, the sand content, and the pH decreased with the distance from the Han River. Figure 4 presents a scatter atlas of the soil properties and Dis_River at each sampling point, where the blue trend lines fit their collaborative change trend. The clay content, the TN, the TP, and the TK show an overall positive relation with the Dis_River, and the pH is negatively related with Dis_River. The silt and sand contents fluctuated greatly with the increasing distance, and their change trend is the opposite. These results indicate that the spatial distribution of the soil properties was affected by rivers, and the relationship between the soil properties and Dis_River is non-linear.

3.2. Model Evaluation of ULR and GBDT

Table 2 shows the R2 of each environmental variable and SOC fitted by ULR and GBDT, respectively. The R2 of each environmental variable fitted by the GBDT model is higher than that of the ULR model, and the R2 of half the environment variables in the GBDT model is more than five times that in the ULR model. Particularly, many factors, such as Dis_RS, Dis_pond, Dis_IC, the TWI, the sand content, and slope, have an R2 of only 0.001 in the ULR model, but the GBDT model can increase the R2 to over 0.04. The relative importance ranking of each variable in the ULR and GBDT models is basically consistent, and Dis_River is the most important factor.
The total 16 covariates separately explained 16.9% and 37.7% variation of SOC via ULR and GBDT models. These results indicate that the relationship between SOC and each environmental variable was more non-linear rather than linear, and the GBDT model captures more accurate relationships. Therefore, only the results of the GBDT model are shown in the subsequent sections.

3.3. Non-Linear Response of Environmental Variables to SOC

Figure 5 shows the relative importance of the explanatory variables. Dis_River is the most influential factor of SOC, with a relative importance of 26.3%, which is much higher than that of the other variables. The clay content follows Dis_River, with an importance of 16.2%. The importance values of Dis_RS, TK, Dis_IC, pH, and TN range from 5% to 10%, and those of the other variables are less than 5%.
  • Dis_River, Dis_RS, Dis_pond, and Dis_IC: the distance from the nearest river, rural settlements, pond, and irrigated canals, respectively; TWI: topographic wetness index; LU: land use.
Figure 6 contains the partial dependence plots of the eight most important environmental variables, exhibiting the specific non-linear change and effective range of these variables. Dis_River, the clay content, the TK, and the TN are positively related with the SOC, while Dis_RS, pH, Dis_IC, and elevation are negatively associated with the SOC.
The environmental variables did not affect the SOC within all their value ranges but have certain effective ranges. For example, the SOC increased rapidly with the increase in Dis_River between 0 and 4000 m and was stable when Dis_River exceeded 4000 m. The change trends of the clay content, the TK, and the TN curves are similar: maintaining stable, rapid growth, and returning steady. The effective ranges of the clay content, the TK, and the TN on the SOC were 29% to 31%, 10.7 g/kg to 11.1 g/kg, and 1.06 g/kg to 1.18 g/kg, respectively. The SOC rapidly decreased with the distances from the rural settlements and the irrigation canals and remained low when the distances exceeded 50 and 400 m, respectively. The change trends of the pH and the elevation are similar: stable, rapid decline, and returning steady. The effective ranges of the pH and the elevation were 7.1 to 7.3 and 25 m to 32 m, respectively.
Land 12 01198 g006 550
Figure 6. Partial dependence plots of the most important 8 environmental variables (a–h).
Figure 6. Partial dependence plots of the most important 8 environmental variables (a–h).
Land 12 01198 g006
  • Dis_River, Dis_RS, and Dis_IC: the distance from the nearest river, rural settlements, and irrigated canals, respectively.
The two-factor interaction effects of the top eight important variables on the SOC is revealed via GBDT model and exhibited in Figure 7. The value of a certain position is the R2 of the two corresponding explanatory variables for SOC. By comparing the R2 (X1∩X2) and the sum of f R2 (X1) and R2 (X2) in Figure 7, we find that any two variables are not completely independent but have interactive effects. Most of the interaction effects are negative, meaning that R2 (X1∩X2) is less than the sum of R2 (X1) and R2 (X2). The joint effects of pH and TN and pH and elevation are positive. The two covariates (i.e., Dis_River and clay content) explain 18.5% of the variation of SOC, which is equivalent to the total explanatory power of the other 14 variables.
  • Dis_River, Dis_RS, and Dis_IC: the distance from the nearest river, rural settlements, and irrigated canals, respectively.

3.4. Spatial Distribution of SOC

The spatial map of the SOC estimated by the GBDT model is shown in Figure 8. The SOC was high in the areas far from the Han River, such as the eastern and northern parts of the study area, and low in the areas near the Han River, such as the southern and western parts of the study area. The comparison of the spatial maps of the SOC and the environmental variables revealed that the high SOC value area coincided with the high clay content and TK value areas and the low pH value area. In addition, the shorter the distance to rural settlements and irrigation channels is, the higher the SOC content is in the farmland. In summary, the spatial distribution of SOC was influenced by the location of the Han River, the rural settlements, and the irrigation canals and the distribution of various soil properties.

4. Discussion

4.1. Impact of Environmental Variables on SOC Variation

Rivers are the primary influencing factor on the SOC in the study area. The Han River directly and indirectly affects the spatial pattern of SOC in the study area. Through the topographic map and field investigations, we found that the Han River is located at a high-altitude region, causing frequent flooding during summer. Consequently, large amounts of river sediment and soil along the river are carried away by floods to low-relief areas far from the Han River, directly causing the deposition of soil organic matter [49]. In addition, sand and silt particles are mostly deposited near the Han River (Figure 4), while clay particles are mostly deposited far from the Han River. The Han River affects the spatial distribution of the SOC content by affecting the soil texture. This phenomenon explains why the Dis_River is positively related with the SOC. In this study, Dis_River is the most important factor on the SOC, with a relative importance of 26.3%, which is much higher than that of the other variables. Several studies also found significant positive relationships between the SOC and Dis_River, and most of the study areas are alluvial plain, like our study area [1,26,27,28,50].
Soil nutrients and texture are important factors on the regional SOC. A high soil nutrient content is beneficial for crop growth, allowing more residues to return to the soil [51,52]. Therefore, soil nutrients are often positively related with the SOC [53,54]. Compared with a high sand content, a high clay content is beneficial for maintaining soil moisture and nutrients, promoting vegetation growth and organic matter accumulation [55,56]. Consequently, the clay and silt contents often show a positive relation with the SOC, while the sand content often shows a negative relation with the SOC [17,57,58]. In this study, the clay content is the second important factor of SOC, and the relative importance of the TN and the TK exceeds 5%. However, the effects of the TP and the silt and sand contents on SOC are relatively small.
Farmland near rural settlements often experiences long-term cultivation and fertilization, resulting in a thick plow layer and a high organic matter content. In addition, the closer the dry land is to the irrigation canal, the better the soil water and heat conditions are, which is conducive to the accumulation of SOC. Thus, the negative relation of Dis_RS and Dis_IC to SOC is proven [59,60,61], which is consistent with the results of this article.

4.2. Implications of the Effective Range of Environmental Variables

This study revealed the effective range of each influencing factor of SOC, providing valuable information for agricultural management. For example, acidic substances can be added to alkaline soil to keep the pH below 7.1. Such a neutral or slightly acidic soil environment is conducive to the growth of most crops and the accumulation of SOC in farmland [62]. Moreover, clay-rich soil can be added to soil with a low clay content for the clay content to reach 31%, which is beneficial for water and fertilizer conservation, vegetation growth, and SOC accumulation [63]. However, an excessive clay content leads to poor soil drainage and permeability, which is not conducive to the growth of vegetation roots and the accumulation of SOC [64,65]. Figure 4d,g suggest that the SOC remained at a high level when TN ≥ 1.18 g/kg and TK ≥ 11.2 g/kg, providing data support for precise fertilization. Excessive fertilization has a weak impact on the accumulation of SOC in farmland and may lead to agricultural non-point source pollution [66,67]. These results indicated that maintaining the soil nutrients and texture within an appropriate range contributed to the high SOC content.
The influencing ranges of the Han River are approximately 4 km, which may be due to the high terrain there, making the region less affected by floods. In addition, particular attention should be paid to farmland soils within 1.5 km of the Han River, where the SOC content is low. The relationship between SOC and Dis_IC becomes steady when the Dis_IC is larger than 400 m. This may be due to the limited irrigation range of the canal, and farmland far from the canal can only be irrigated by rainwater. Therefore, it is necessary to build irrigation and drainage systems to improve soil fertility.

4.3. Limitations

This study explored the non-linear relationship and the effective ranges of environmental variables. However, the accuracy of the research results is limited by the quality of environmental variables. For example, the spatial map of the TN, the TP, and the TK was estimated via the ensemble learning model and the soil sampling data collected in China from 2010 to 2018 [21]. The values of these variables in this study area may deviate from the true values.
The relationship between the soil properties and Dis_River is also influenced by the size, type, and stage (e.g., upstream, midstream, and downstream) of the river and the seasons [24,68]. In addition, the relationship between the pH, the soil texture, and the SOC may be different for different crop types [69,70]. However, this study did not consider these factors due to data limitations.

5. Conclusions

This study determined the main controlling factors of SOC, revealed the effective range of environmental variables, and obtained the spatial map of SOC by using the GBDT model and partial dependence plots. The results show that GBDT outperforms ULR. The GBDT results show that Dis_River is the most important factor of SOC, confirming the importance of the Han River to the SOC pattern. The partial dependence plots indicate that all the environmental variables have effective ranges, and when their values are extremely high or low, they do not respond to changes in SOC. Moreover, any two variables are not completely independent but have interactive effects. These results highlight the importance of revealing the effective ranges of environmental variables, which provide data support for understanding the spatial pattern of SOC in floodplain farmland, achieving carbon sequestration in farmland, and achieving precision agriculture. The GBDT with partial dependence plots is an effective method for the fitting and mapping of SOC.

Author Contributions

Conceptualization, Z.W. and Q.Y.; Methodology, Z.W. and Y.C.; Software, Z.T.; Validation, J.O. and G.L.; Formal analysis, Y.C.; Investigation, Z.W.; Resources, Y.Z. and X.F.; Data curation, Y.Z. and X.F.; Writing–original draft, Z.W.; Writing–review & editing, Y.C. and Q.Y.; Visualization, Y.Z. and J.O.; Supervision, Q.Y.; Funding acquisition, Z.W. and Q.Y. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the National Natural Science Foundation of China (Grant No. 42201447), the Fundamental Research Funds for the Central Universities (Grant No. 2022-11278), and the Third Comprehensive Scientific Expedition to Xinjiang in China—Geological Hazards and Ecological Environment Investigation of National Major Energy Channel on the North Slope of Tianshan Mountains (Grant No. 2022xjkk1004).

Data Availability Statement

No applicable.

Acknowledgments

This research was supported by We are grateful for the data support from the Hubei Geological Survey and Soil SubCenter, National Earth System Science Data Center, National Science & Technology Infrastructure of China.

Conflicts of Interest

The authors declare no conflict of interest.

Abbreviations

GBDTgradient boosting decision tree
ULRUnivariate linear regression
SOCsoil organic carbon
TNtotal nitrogen
TPtotal phosphorus
TKtotal potassium
LUland use
NDVInormalized difference vegetation index
TWItopographic wetness index
Dis_Riverthe distance from the nearest river
Dis_RSthe distance from the nearest rural settlements
Dis_pondthe distance from the nearest pond
Dis_ICthe distance from the nearest irrigated canals

References

  1. Chimweta, M.; Nyakudya, I.W.; Jimu, L. Fertility status of cultivated floodplain soils in the Zambezi Valley, northern Zimbabwe. Phys. Chem. Earth 2018, 105, 147–153. [Google Scholar] [CrossRef]
  2. Yu, Y.; Shi, Y.; Li, M.; Wang, C.; Zhang, L.; Sun, Z.; Lei, B.; Miao, Y.; Wang, W.; Liu, B.; et al. Land-use type strongly affects soil microbial community assembly process and inter-kingdom co-occurrence pattern in a floodplain ecosystem. Appl. Soil Ecol. 2022, 179, 104574. [Google Scholar] [CrossRef]
  3. Salamanca-Carreno, A.; Velez-Terranova, M.; Vargas-Corzo, O.M.; Perez-Lopez, O.; Castillo-Perez, A.F.; Pares-Casanova, P.M. Relationship of Physiographic Position to Physicochemical Characteristics of Soils of the Flooded-Savannah Agroecosystem, Colombia. Agriculture 2023, 13, 220. [Google Scholar] [CrossRef]
  4. Yu, Y.; Liu, H.; Zhang, L.; Sun, Z.; Lei, B.; Miao, Y.; Chu, H.; Han, S.; Shi, Y.; Zheng, J. Distinct response patterns of plants and soil microorganisms to agronomic practices and seasonal variation in a floodplain ecosystem. Front. Microbiol. 2023, 14, 95. [Google Scholar] [CrossRef]
  5. Qin, Y.; Xin, Z.; Wang, D. Comparison of topsoil organic carbon and total nitrogen in different flood-risk riparian zones in a Chinese Karst area. Environ. Earth Sci. 2016, 75, 1038. [Google Scholar] [CrossRef]
  6. Beyene, A.A.; Verhoest, N.E.C.; Tilahun, S.; Alamirew, T.; Adgo, E.; Nyssen, J. Irrigation efficiency and shallow groundwater in anisotropic floodplain soils near Lake Tana, Ethiopia. Irrig. Drain. 2019, 68, 365–378. [Google Scholar] [CrossRef]
  7. Liu, M.; Han, G.; Zhang, Q. Effects of agricultural abandonment on soil aggregation, soil organic carbon storage and stabilization: Results from observation in a small karst catchment, Southwest China. Agric. Ecosyst. Environ. 2020, 288, 106719. [Google Scholar] [CrossRef]
  8. McBratney, A.B.; Santos, M.L.M.; Minasny, B. On digital soil mapping. Geoderma 2003, 117, 3–52. [Google Scholar] [CrossRef]
  9. Minasny, B.; McBratney, A.B. Digital soil mapping: A brief history and some lessons. Geoderma 2016, 264, 301–311. [Google Scholar] [CrossRef]
  10. Wang, L.; Li, Z.; Wang, D.; Liao, S.; Nie, X.; Liu, Y. Factors controlling soil organic carbon with depth at the basin scale. Catena 2022, 217, 106478. [Google Scholar] [CrossRef]
  11. Lamichhane, S.; Kumar, L.; Wilson, B. Digital soil mapping algorithms and covariates for soil organic carbon mapping and their implications: A review. Geoderma 2019, 352, 395–413. [Google Scholar] [CrossRef]
  12. Meng, X.; Bao, Y.; Liu, J.; Liu, H.; Zhang, X.; Zhang, Y.; Wang, P.; Tang, H.; Kong, F. Regional soil organic carbon prediction model based on a discrete wavelet analysis of hyperspectral satellite data. Int. J. Appl. Earth Obs. Geoinf. 2020, 89, 102111. [Google Scholar] [CrossRef]
  13. Lamichhane, S.; Adhikari, K.; Kumar, L. National soil organic carbon map of agricultural lands in Nepal. Geoderma Reg. 2022, 30, e00568. [Google Scholar] [CrossRef]
  14. Liu, S.; Xie, X.; Wang, X.; Feng, X.; Hou, X.; Wang, S.; Lin, K.; Huang, M.; Jia, S.; Hou, Y.; et al. Distribution characteristics and prediction model of farmland soil organic carbon in eastern China. Environ. Res. Commun. 2022, 4, 055012. [Google Scholar] [CrossRef]
  15. Yang, J.; Fan, J.; Lan, Z.; Mu, X.; Wu, Y.; Xin, Z.; Miping, P.; Zhao, G. Improved Surface Soil Organic Carbon Mapping of SoilGrids250m Using Sentinel-2 Spectral Images in the Qinghai-Tibetan Plateau. Remote Sens. 2023, 15, 114. [Google Scholar] [CrossRef]
  16. Dai, Y.; Zhou, P.; Guo, X.; Luo, P.; Chen, X.; Wu, J. Role of environmental factors on concentrations and ratios of subsoil C-N-P in subtropical paddy fields. J. Soils Sediments 2023, 23, 1999–2010. [Google Scholar] [CrossRef]
  17. Lu, J.; Feng, S.; Wang, S.; Zhang, B.; Ning, Z.; Wang, R.; Chen, X.; Yu, L.; Zhao, H.; Lan, D.; et al. Patterns and driving mechanism of soil organic carbon, nitrogen, and phosphorus stoichiometry across northern China’s desert-grassland transition zone. Catena 2023, 220, 106695. [Google Scholar] [CrossRef]
  18. Lv, X.; Jia, G.; Yu, X.; Niu, L. Vegetation and Topographic Factors Affecting SOM, SOC, and N Contents in a Mountainous Watershed in North China. Forests 2022, 13, 742. [Google Scholar] [CrossRef]
  19. Macmillan, R.A.; McBratney, A. GlobalSoilMap.net—From planning, development and proof of concept to full-scale production mapping. In Proceedings of the 19th World Congress of Soil Science, Brisbane, Australia, 1–6 August 2010. [Google Scholar]
  20. Poggio, L.; de Sousa, L.M.; Batjes, N.H.; Heuvelink, G.B.M.; Kempen, B.; Ribeiro, E.; Rossiter, D. SoilGrids 2.0: Producing soil information for the globe with quantified spatial uncertainty. Soil 2021, 7, 217–240. [Google Scholar] [CrossRef]
  21. Liu, F.; Wu, H.; Zhao, Y.; Li, D.; Yang, J.-L.; Song, X.; Shi, Z.; Zhu, A.X.; Zhang, G.-L. Mapping high resolution National Soil Information Grids of China. Sci. Bull. 2022, 67, 328–340. [Google Scholar] [CrossRef]
  22. da Silva, R.B.; Iori, P.; Moraes Tavares, R.L.; de Souza, Z.M.; de Lima, C.C.; de Melo Silva, F.A.; Bento, M.d.S. Do water dynamics and land use in riparian areas change the spatial pattern of physical-mechanical properties of a Cambisol? Precis. Agric. 2022, 23, 984–1007. [Google Scholar] [CrossRef]
  23. Rad, M.R.P. Spatial variability of soil texture fractions and pH in a flood plain (case study from eastern Iran). Catena 2018, 160, 275–281. [Google Scholar]
  24. Wang, X.; Zhang, H.; Zhang, Z.; Zhang, C.; Zhang, K.; Pang, H.; Bell, S.M.; Li, Y.; Chen, J. Reinforced soil salinization with distance along the river: A case study of the Yellow River Basin. Agric. Water Manag. 2023, 279, 108184. [Google Scholar] [CrossRef]
  25. Lu, Q.; Bai, J.; Fang, H.; Wang, J.; Zhao, Q.; Jia, J. Spatial and seasonal distributions of soil sulfur in two marsh wetlands with different flooding frequencies of the Yellow River Delta, China. Ecol. Eng. 2016, 96, 63–71. [Google Scholar] [CrossRef] [Green Version]
  26. Xu, H.; Wu, S.; Diehl, J.A. The Influence of Harbin Forest-River Ecological Corridor Construction on the Restoration of Mollisols in Cold Regions of China. Forests 2022, 13, 652. [Google Scholar] [CrossRef]
  27. Lamichhane, S.; Kumar, L.; Adhikari, K. Digital mapping of topsoil organic carbon content in an alluvial plain area of the Terai region of Nepal. Catena 2021, 202, 105299. [Google Scholar] [CrossRef]
  28. Pahlavan-Rad, M.R.; Dahmardeh, K.; Brungard, C. Predicting soil organic carbon concentrations in a low relief landscape, eastern Iran. Geoderma Reg. 2018, 15, e00195. [Google Scholar] [CrossRef]
  29. Chen, L.; Yin, Z.; Tang, M.; Li, T.; Xu, D. Distribution and Genesis of Organic Carbon Storage on the Northern Shelf of the South China Sea. Int. J. Environ. Res. Public Health 2022, 19, 11367. [Google Scholar] [CrossRef] [PubMed]
  30. Sutfin, N.A.; Wohl, E. Substantial soil organic carbon retention along floodplains of mountain streams. J. Geophys. Res.-Earth Surf. 2017, 122, 1325–1338. [Google Scholar] [CrossRef]
  31. Matteau, J.-P.; Celicourt, P.; Letourneau, G.; Gumiere, T.; Walter, C.; Gumiere, S.J. Association between irrigation thresholds and promotion of soil organic carbon decomposition in sandy soil. Sci. Rep. 2021, 11, 6733. [Google Scholar] [CrossRef] [PubMed]
  32. Bezak, N.; Auflic, M.J.; Mikos, M. Reanalysis of Soil Moisture Used for Rainfall Thresholds for Rainfall-Induced Landslides: The Italian Case Study. Water 2021, 13, 1977. [Google Scholar] [CrossRef]
  33. Wang, C.; Tong, C.; Chambers, L.G.; Liu, X. Identifying the Salinity Thresholds that Impact Greenhouse Gas Production in Subtropical Tidal Freshwater Marsh Soils. Wetlands 2017, 37, 559–571. [Google Scholar] [CrossRef]
  34. Chen, J.; Xiao, H.; Li, Z.; Liu, C.; Wang, D.; Wang, L.; Tang, C. Threshold effects of vegetation coverage on soil erosion control in small watersheds of the red soil hilly region in China. Ecol. Eng. 2019, 132, 109–114. [Google Scholar] [CrossRef]
  35. Friedman, J.H. Greedy function approximation: A gradient boosting machine. Ann. Stat. 2001, 29, 1189–1232. [Google Scholar] [CrossRef]
  36. Ke, G.; Meng, Q.; Finley, T.; Wang, T.; Chen, W.; Ma, W.; Ye, Q.; Liu, T.-Y. LightGBM: A Highly Efficient Gradient Boosting Decision Tree. In Proceedings of the 31st Annual Conference on Neural Information Processing Systems (NIPS), Long Beach, CA, USA, 4–9 December 2017. [Google Scholar]
  37. Du, Z.; Yang, L.; Zhang, D.; Cui, T.; He, X.; Xiao, T.; Xie, C.; Li, H. Corn variable-rate seeding decision based on gradient boosting decision tree model. Comput. Electron. Agric. 2022, 198, 107025. [Google Scholar] [CrossRef]
  38. Liu, Y.; Ren, W.; Liu, C.; Cai, S.; Xu, W. Displacement-Based Back-Analysis Frameworks for Soil Parameters of a Slope: Using Frequentist Inference and Bayesian Inference. Int. J. Geomech. 2022, 22, 04022026. [Google Scholar] [CrossRef]
  39. Zhu, T.; Tao, C. Prediction models with multiple machine learning algorithms for POPs: The calculation of PDMS-air partition coefficient from molecular descriptor. J. Hazard. Mater. 2022, 423, 127037. [Google Scholar] [CrossRef] [PubMed]
  40. Huan, Y.; Song, L.; Khan, U.; Zhang, B. Stacking ensemble of machine learning methods for landslide susceptibility mapping in Zhangjiajie City, Hunan Province, China. Environ. Earth Sci. 2023, 82, 35. [Google Scholar] [CrossRef]
  41. Wu, Z.; Chen, Y.; Yang, Z.; Liu, Y.; Zhu, Y.; Tong, Z.; An, R. Spatial distribution of lead concentration in peri-urban soil: Threshold and interaction effects of environmental variables. Geoderma 2023, 429, 116193. [Google Scholar] [CrossRef]
  42. Lin, P.; Weng, J.; Brands, D.K.; Qian, H.; Yin, B. Analysing the relationship between weather, built environment, and public transport ridership. IET Intell. Transp. Syst. 2020, 14, 1946–1954. [Google Scholar] [CrossRef]
  43. Shao, Q.; Zhang, W.; Cao, X.; Yang, J. Nonlinear and interaction effects of land use and motorcycles/E-bikes on car ownership. Transp. Res. Part D Transp. Environ. 2022, 102, 103115. [Google Scholar] [CrossRef]
  44. Hao, W.; Hao, Z.; Yuan, F.; Ju, Q.; Hao, J. Regional Frequency Analysis of Precipitation Extremes and Its Spatio-Temporal Patterns in the Hanjiang River Basin, China. Atmosphere 2019, 10, 130. [Google Scholar] [CrossRef] [Green Version]
  45. Li, N.; Li, Y.; Lou, R.; Xu, H.; Saeed, L. Effects of Fe(II) and organic carbon on nitrate reduction in surficial sediments of a large shallow freshwater lake. J. Environ. Manag. 2023, 336, 117623. [Google Scholar] [CrossRef] [PubMed]
  46. Wilding, L.P. Spatial variability: Its documentation, accommodation and implication to soil survey. In Proceedings of the Spatial Variations, Las Vegas, NV, USA, 30 November–1 December 1984. [Google Scholar]
  47. Cao, S.; Lu, A.; Wang, J.; Huo, L. Modeling and mapping of cadmium in soils based on qualitative and quantitative auxiliary variables in a cadmium contaminated area. Sci. Total Environ. 2017, 580, 430–439. [Google Scholar] [CrossRef]
  48. Batjes, N.H.; Ribeiro, E.; van Oostrum, A. Standardised soil profile data to support global mapping and modelling (WoSIS snapshot 2019). Earth Syst. Sci. Data 2020, 12, 299–320. [Google Scholar] [CrossRef] [Green Version]
  49. Yuan, J.; Zhong, X. Sediment pollution and nitrogen release at the sediment-water interface in Changjiang River and its tributary, the lower Han River Basin. Water Environ. J. 2020, 34, 672–682. [Google Scholar] [CrossRef]
  50. Amosse, J.; Le Bayon, R.-C.; Gobat, J.-M. Are urban soils similar to natural soils of river valleys? J. Soils Sediments 2015, 15, 1716–1724. [Google Scholar] [CrossRef] [Green Version]
  51. Roussis, I.; Kakabouki, I.; Bilalis, D. Comparison of growth indices of Nigella sativa I. under different plant densities and fertilization. Emir. J. Food Agric. 2019, 31, 231–247. [Google Scholar] [CrossRef] [Green Version]
  52. Chen, M.; Wang, X.; Ding, X.; Liu, L.; Wu, L.; Zhang, S. Effects of organic fertilization on phosphorus availability and crop growth: Evidence from a 7-year fertilization experiment. Arch. Agron. Soil Sci. 2022, 69, 2092–2103. [Google Scholar] [CrossRef]
  53. Hu, Y.; Zhang, Y.; Liu, J.; Chen, X.; Zhang, J.; Yao, Y. Nitrogen-rich animal and plant wastes as fertilizer improve the soil carbon/nitrogen ratio and plant branching and thickening of young walnut trees under deficit irrigation conditions. Arch. Agron. Soil Sci. 2023, 1–16. [Google Scholar] [CrossRef]
  54. Shen, X.; Ma, J.; Li, Y.; Li, Y.; Xia, X. The Effects of Multiple Global Change Factors on Soil Nutrients across China: A Meta-Analysis. Int. J. Environ. Res. Public Health 2022, 19, 15230. [Google Scholar] [CrossRef]
  55. Yang, S.; Zhao, W.; Pereira, P. Determinations of environmental factors on interactive soil properties across different land-use types on the Loess Plateau, China. Sci. Total Environ. 2020, 738, 140270. [Google Scholar] [CrossRef] [PubMed]
  56. Huang, W.; Mirabito, A.J.J.; Tenesaca, C.G.G.; Mejia-Garcia, W.F.F.; Lawrence, N.C.C.; VanLoocke, A.L.; Kaleita, A.L.; Hall, S.J.J. Controls on organic and inorganic soil carbon in poorly drained agricultural soils with subsurface drainage. Biogeochemistry 2023, 163, 121–137. [Google Scholar] [CrossRef]
  57. Ding, S.J.; Zhang, X.F.; Yang, W.L.; Xin, X.L.; Zhu, A.N.; Huang, S.M. Soil Nutrients and Aggregate Composition of Four Soils with Contrasting Textures in a Long-Term Experiment. Eurasian Soil Sci. 2021, 54, 1746–1755. [Google Scholar] [CrossRef]
  58. Su, Y.-G.; Huang, G.; Lin, S.-N.; Huang, Z.-Y.; Wu, G.-P.; Cheng, H. Patterns of organic carbon and nitrogen stocks in soil particle-size fractions along an aridity gradient in Northern China?s deserts. Catena 2023, 221, 106785. [Google Scholar] [CrossRef]
  59. Liu, B.; Qian, J.; Zhao, R.; Yang, Q.; Wu, K.; Zhao, H.; Feng, Z.; Dong, J. Spatio-Temporal Variation and Its Driving Forces of Soil Organic Carbon along an Urban-Rural Gradient: A Case Study of Beijing. Int. J. Environ. Res. Public Health 2022, 19, 15201. [Google Scholar] [CrossRef]
  60. Caulfield, M.E.; Fonte, S.J.; Tittonell, P.; Vanek, S.J.; Sherwood, S.; Oyarzun, P.; Borja, R.M.; Dumble, S.; Groot, J.C.J. Inter-community and on-farm asymmetric organic matter allocation patterns drive soil fertility gradients in a rural Andean landscape. Land Degrad. Dev. 2020, 31, 2973–2985. [Google Scholar] [CrossRef]
  61. Wu, Z.; Chen, Y.; Yang, Z.; Zhu, Y.; Han, Y. Mapping Soil Organic Carbon in Low-Relief Farmlands Based on Stratified Heterogeneous Relationship. Remote Sens. 2022, 14, 3575. [Google Scholar] [CrossRef]
  62. Sun, M.; Chen, S.; Kurle, J.E. Interactive Effects of Soybean Cyst Nematode, Arbuscular-Mycorrhizal Fungi, and Soil pH on Chlorophyll Content and Plant Growth of Soybean. Phytobiomes J. 2022, 6, 95–105. [Google Scholar] [CrossRef]
  63. Blomquist, J.; Englund, J.-E.; Berglund, K. Soil characteristics and tillage can predict the effect of ‘structure lime’ on soil aggregate stability. Soil Res. 2022, 60, 373–384. [Google Scholar] [CrossRef]
  64. Suzuki, L.E.A.S.; Reinert, D.J.; Alves, M.C.; Reichert, J.M. Critical Limits for Soybean and Black Bean Root Growth, Based on Macroporosity and Penetrability, for Soils with Distinct Texture and Management Systems. Sustainability 2022, 14, 2958. [Google Scholar] [CrossRef]
  65. Chen, S.; Mao, X.; Shang, S. Response and contribution of shallow groundwater to soil water/salt budget and crop growth in layered soils. Agric. Water Manag. 2022, 266, 107574. [Google Scholar] [CrossRef]
  66. Li, X.; Shang, J. Spatial interaction effects on the relationship between agricultural economic and planting non-point source pollution in China. Environ. Sci. Pollut. Res. 2023, 30, 51607–51623. [Google Scholar] [CrossRef]
  67. Lu, H.; Wang, R.; Dun, C.; Wang, D.; Zhang, H.; Cui, P.; Dai, Q.; Zhang, H. Effects of Controlled-release Bulk Blending Fertilizer on Wheat Yield of Different Varieties under Various Soil Basic Fertility. Environ. Pollut. Bioavailab. 2022, 34, 273–283. [Google Scholar] [CrossRef]
  68. Guan, Q.; Feng, L.; Tang, J.; Park, E.; Ali, T.A.; Zheng, Y. Trends in River Total Suspended Sediments Driven by Dams and Soil Erosion: A Comparison Between the Yangtze and Mekong Rivers. Water Resour. Res. 2022, 58, e2022WR031979. [Google Scholar] [CrossRef]
  69. Wang, S.; Hu, K.; Feng, P.; Qin, W.; Leghari, S.J. Determining the effects of organic manure substitution on soil pH in Chinese vegetable fields: A meta-analysis. J. Soils Sediments 2023, 23, 118–130. [Google Scholar] [CrossRef]
  70. Wang, D.; Wang, Z.; Zhang, J.; Zhou, B.; Lv, T.; Li, W. Effects of Soil Texture on Soil Leaching and Cotton (Gossypium hirsutum L.) Growth under Combined Irrigation and Drainage. Water 2021, 13, 3614. [Google Scholar] [CrossRef]
Land 12 01198 g001 550
Figure 1. Location of the study area and sampling points.
Figure 1. Location of the study area and sampling points.
Land 12 01198 g001
Land 12 01198 g002 550
Figure 2. Histograms of SOC content. CV: coefficient of variation.
Figure 2. Histograms of SOC content. CV: coefficient of variation.
Land 12 01198 g002
Land 12 01198 g003 550
Figure 3. Spatial distribution maps of Elevation (a), slope (b), TWI (c), clay content (d), silt content (e), sand content (f), TN (g), TP (h), TK (i), pH (j), and NDVI (k).
Figure 3. Spatial distribution maps of Elevation (a), slope (b), TWI (c), clay content (d), silt content (e), sand content (f), TN (g), TP (h), TK (i), pH (j), and NDVI (k).
Land 12 01198 g003
Land 12 01198 g004 550
Figure 4. Relationship between the distance from the rivers and SOC (a), TN (b), TP (c), TK (d), pH (e), clay content (f), silt content (g), and sand content (h).
Figure 4. Relationship between the distance from the rivers and SOC (a), TN (b), TP (c), TK (d), pH (e), clay content (f), silt content (g), and sand content (h).
Land 12 01198 g004
Land 12 01198 g005 550
Figure 5. Importance ranking of environmental factors on SOC.
Figure 5. Importance ranking of environmental factors on SOC.
Land 12 01198 g005
Land 12 01198 g007 550
Figure 7. Two-factor interaction effects on the SOC content.
Figure 7. Two-factor interaction effects on the SOC content.
Land 12 01198 g007
Land 12 01198 g008 550
Figure 8. Spatial distribution of SOC content.
Figure 8. Spatial distribution of SOC content.
Land 12 01198 g008
Table
Table 1. Environmental variables of SOC.
Table 1. Environmental variables of SOC.
AspectsVariablesData Source and Link
soil propertiesTNChinese National Earth System Science Data Center (90 m resolution), http://soil.geodata.cn/index.html, accessed on 1 May 2022
TP
TK
pH
clay contentSoilGrids v2.0 (250 m resolution), https://soilgrids.org/, accessed on 1 March 2023
silt content
sand content
reliefelevationASTER DEM (30 m resolution), http://www.gscloud.cn, accessed on 1 March 2023
slope
TWI
organismNDVIChinese Resource and Environment Science and Data center (30 m resolution), https://www.resdc.cn/, accessed on 1 March 2023
land useSupervised classification of Landsat 8 images (30 m resolution), http://www.gscloud.cn, accessed on 1 March 2023
distance factorsdistance from the riverBased on land use map
distance from the rural settlement
distance from the irrigated canal
distance from the pond
Table
Table 2. Model evaluation of ULR and GBDT.
Table 2. Model evaluation of ULR and GBDT.
VariablesR2 of ULRR2 of GBDTVariablesR2 of ULRR2 of GBDT
Dis_River0.105 ***0.147 ***Dis_RS0.0010.057 ***
pH0.034 ***0.088 ***Dis_pond0.0010.047 ***
Clay0.028 **0.086 ***TN0.023 **0.047 ***
TK0.023 **0.080 ***Dis_IC0.0010.045 ***
Silt0.032 ***0.076 ***TWI0.0010.046 ***
NDVI0.0090.075 ***Sand0.0010.043 ***
TP0.065 ***0.072 ***Slope0.0010.040 ***
Elevation0.010 *0.058 ***LU0.0010.004
Dis_River, Dis_RS, Dis_pond, and Dis_IC: the distance from the nearest river, rural settlements, pond, and irrigated canals, respectively; TWI: topographic wetness index; LU: land use. *, **, *** denotes p < 0.05, 0.01, 0.001, respectively (two-tailed).
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Wu, Z.; Chen, Y.; Zhu, Y.; Feng, X.; Ou, J.; Li, G.; Tong, Z.; Yan, Q. Mapping Soil Organic Carbon in Floodplain Farmland: Implications of Effective Range of Environmental Variables. Land 2023, 12, 1198. https://doi.org/10.3390/land12061198

AMA Style

Wu Z, Chen Y, Zhu Y, Feng X, Ou J, Li G, Tong Z, Yan Q. Mapping Soil Organic Carbon in Floodplain Farmland: Implications of Effective Range of Environmental Variables. Land. 2023; 12(6):1198. https://doi.org/10.3390/land12061198

Chicago/Turabian Style

Wu, Zihao, Yiyun Chen, Yuanli Zhu, Xiangyang Feng, Jianxiong Ou, Guie Li, Zhaomin Tong, and Qingwu Yan. 2023. "Mapping Soil Organic Carbon in Floodplain Farmland: Implications of Effective Range of Environmental Variables" Land 12, no. 6: 1198. https://doi.org/10.3390/land12061198

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop