Mapping surface soil organic carbon density of cultivated land using machine learning in Zhengzhou DOI Creative Commons
Hengliang Guo,

Jinyang Wang,

Dujuan Zhang

et al.

Environmental Geochemistry and Health, Journal Year: 2024, Volume and Issue: 47(1)

Published: Nov. 28, 2024

Research on soil organic carbon (SOC) is crucial for improving sinks and achieving the "double-carbon" goal. This study introduces ten auxiliary variables based data from a 2021 land quality survey in Zhengzhou multi-objective regional geochemical survey. It uses geostatistical ordinary kriging (OK) interpolation, as well classical machine learning (ML) models, including random forest (RF) support vector (SVM), to map density (SOCD) topsoil layer (0 − 20 cm) of cultivated land. partitions sampling assess generalization capability with Zhongmu County designated an independent test set (dataset2) remaining training (dataset1). The three models are trained using dataset1, directly applied dataset2 evaluate compare their performance. distribution SOCD SOCS soils various types textures analyzed optimal interpolation method. results indicated that: (1) average SOC densities predicted by OK RF, SVM 3.70, 3.74, 3.63 kg/m2, precisions (R2) 0.34, 0.60, 0.81, respectively. (2) ML achieves significantly higher predictive precision than traditional interpolation. RF model's 0.21 model more precise estimating stock. (3) When dataset2, exhibited superior capabilities (R2 = 0.52, MSE 0.32) over 0.32, 0.45). (4) spatial surface area exhibits decreasing gradient west east south north. total stock estimated at approximately 10.76 × 106t. (5) integration attribute variables, climatic remote sensing data, techniques holds significant promise high-precision high-quality mapping agricultural soils.

Language: Английский

Mapping surface soil organic carbon density of cultivated land using machine learning in Zhengzhou DOI Creative Commons
Hengliang Guo,

Jinyang Wang,

Dujuan Zhang

et al.

Environmental Geochemistry and Health, Journal Year: 2024, Volume and Issue: 47(1)

Published: Nov. 28, 2024

Research on soil organic carbon (SOC) is crucial for improving sinks and achieving the "double-carbon" goal. This study introduces ten auxiliary variables based data from a 2021 land quality survey in Zhengzhou multi-objective regional geochemical survey. It uses geostatistical ordinary kriging (OK) interpolation, as well classical machine learning (ML) models, including random forest (RF) support vector (SVM), to map density (SOCD) topsoil layer (0 − 20 cm) of cultivated land. partitions sampling assess generalization capability with Zhongmu County designated an independent test set (dataset2) remaining training (dataset1). The three models are trained using dataset1, directly applied dataset2 evaluate compare their performance. distribution SOCD SOCS soils various types textures analyzed optimal interpolation method. results indicated that: (1) average SOC densities predicted by OK RF, SVM 3.70, 3.74, 3.63 kg/m2, precisions (R2) 0.34, 0.60, 0.81, respectively. (2) ML achieves significantly higher predictive precision than traditional interpolation. RF model's 0.21 model more precise estimating stock. (3) When dataset2, exhibited superior capabilities (R2 = 0.52, MSE 0.32) over 0.32, 0.45). (4) spatial surface area exhibits decreasing gradient west east south north. total stock estimated at approximately 10.76 × 106t. (5) integration attribute variables, climatic remote sensing data, techniques holds significant promise high-precision high-quality mapping agricultural soils.

Language: Английский

Citations

0