Using Local Ensemble Models and Landsat Bare Soil Composites for Large-Scale Soil Organic Carbon Maps DOI

Tom Brög,

Axel Don, Alexander Gocht

et al.

Published: Jan. 1, 2023

National soil organic carbon (SOC) maps are essential to improve greenhouse gas accounting and support climate-smart agriculture. Large-scale SOC models based on wall-to-wall information from remote sensing remain a challenge due the high diversity of natural conditions difficulty for spatial location samples. In this study, we tested if implementation local ensemble (LEM) can be used predictions Landsat-based reflectance composites (SRC) Germany. We divided research area into 30 times km tiles calculated generalized linear (GLM) random, nearby observations. Based GLMs, were predicted aggregated using moving window approach. The variable importance was analyzed identify dependencies in correlation between SRC SOC. For final map, Random Forest (RF) model trained predictions, SRC, full set training samples agricultural inventory. results show that LEM able accuracy (R2 = 0.68; RMSE 5.6 g kg–1), compared single, global 0.52; 6.8 kg–1). spectral bands showed clear patterns throughout area. Differences explained by conditions, influencing properties. Compared widely adopted integration distance covariates such as geographical coordinates, reduce autocorrelation greater extent prediction accuracy, especially underrepresented values. presents new method account increase interpretability DSM models.

Language: Английский

Digital mapping of heavy metals in urban soils: A review and research challenges DOI
Tiezhu Shi, He Li, Ran Wang

et al.

CATENA, Journal Year: 2023, Volume and Issue: 228, P. 107183 - 107183

Published: May 2, 2023

Language: Английский

Citations

33
Using local ensemble models and Landsat bare soil composites for large-scale soil organic carbon maps in cropland DOI Creative Commons
Tom Broeg, Axel Don, Alexander Gocht

et al.

Geoderma, Journal Year: 2024, Volume and Issue: 444, P. 116850 - 116850

Published: March 19, 2024

National soil organic carbon (SOC) maps are essential to improve greenhouse gas accounting and support climate-smart agriculture. Large-scale SOC models based on wall-to-wall information from remote sensing remain a challenge due the high diversity of natural conditions difficulty for spatial location samples. In this study, we tested if implementation local ensemble (LEM) can be used predictions Landsat-based reflectance composites (SRC) Germany. For this, divided research area into 30 times km tiles calculated generalized linear (GLM) random, nearby observations. Based GLMs, were predicted aggregated using moving window approach. The variable importance was analyzed identify dependencies in correlation between SRC SOC. final map, Random Forest (RF) model trained predictions, SRC, full set training samples agricultural inventory. results show that LEM able accuracy (R2 = 0.68; RMSE 5.6 g kg−1), compared single, global 0.52; 6.8 kg−1). spectral bands showed clear patterns throughout area. Differences explained by conditions, influencing properties. Compared widely adopted integration distance covariates such as geographical coordinates, reduce autocorrelation greater extent prediction accuracy, especially underrepresented values. presents new method integrate increase interpretability DSM models.

Language: Английский

Citations

12
Open Remote Sensing Data in Digital Soil Organic Carbon Mapping: A Review DOI Creative Commons
Dorijan Radočaj, Mateo Gašparović, Mladen Jurišić

et al.

Agriculture, Journal Year: 2024, Volume and Issue: 14(7), P. 1005 - 1005

Published: June 26, 2024

This review focuses on digital soil organic carbon (SOC) mapping at regional or national scales in spatial resolutions up to 1 km using open data remote sensing sources, emphasizing its importance achieving United Nations’ Sustainable Development Goals (SDGs) related hunger, climate action, and land conservation. The literature was performed according scientific studies indexed the Web of Science Core Collection database since 2000. analysis reveals a steady rise total 2000, with SOC accounting for over 20% these 2023, among which SDGs 2 (Zero Hunger) 13 (Climate Action) were most represented. Notably, countries like States, China, Germany, Iran lead research. shift towards machine deep learning methods has surged post-2010, necessitating environmental covariates topography, climate, spectral data, are cornerstones prediction methods. It noted that available primarily restrict resolution km, typically requires downscaling harmonize topography (up 30 m) multispectral 10–30 m). Future directions include integration diverse development advanced algorithms leveraging learning, utilization high-resolution more precise mapping.

Language: Английский

Citations

9
Handheld In Situ Methods for Soil Organic Carbon Assessment DOI Open Access
Nancy Loria, Rattan Lal,

Ranveer Chandra

et al.

Sustainability, Journal Year: 2024, Volume and Issue: 16(13), P. 5592 - 5592

Published: June 29, 2024

Soil organic carbon (SOC) assessment is crucial for evaluating soil health and supporting sequestration efforts. Traditional methods like wet digestion dry combustion are time-consuming labor-intensive, necessitating the development of non-destructive, cost-efficient, real-time in situ measurements. This review focuses on handheld methodologies SOC estimation, underscoring their practicality reasonable accuracy. Spectroscopic techniques, visible near-infrared, mid-infrared, laser-induced breakdown spectroscopy, inelastic neutron scattering each offer unique advantages. Preprocessing such as external parameter orthogonalization standard normal variate, employed to eliminate moisture content particle size effects estimation. Calibration methods, partial least squares regression support vector machine, establish relationships between spectral reflectance, properties, SOC. Among 32 studies selected this review, 14 exhibited a coefficient determination (R2) 0.80 or higher, indicating potential accurate estimation using approaches. Each study meticulously adjusted factors range, pretreatment method, calibration model improve accuracy content, highlighting both methodological diversity continuous pursuit precision direct field Continued research validation imperative ensure across diverse environments. Thus, underscores devices with good leveraging that influence its precision. Crucial optimizing farming, these measurements, empowering land managers enhance promote sustainable management agricultural landscapes.

Language: Английский

Citations

8
Assessing the potential of multi-source remote sensing data for cropland soil organic matter mapping in hilly and mountainous areas DOI
Peng Li, Xiaobo Wu,

C Feng

et al.

CATENA, Journal Year: 2024, Volume and Issue: 245, P. 108312 - 108312

Published: Aug. 12, 2024

Language: Английский

Citations

5
A detailed mapping of soil organic matter content in arable land based on the multitemporal soil line coefficients and neural network filtering of big remote sensing data DOI Creative Commons
Д. И. Рухович, П. В. Королева, Alexey Rukhovich

et al.

Geoderma, Journal Year: 2024, Volume and Issue: 447, P. 116941 - 116941

Published: June 12, 2024

Language: Английский

Citations

4
Evaluating Airborne Hyperspectral Scanner (AHS) for the mapping of soil organic matter and clay in a Mediterranean forest ecosystem DOI Creative Commons
Francisco M. Canero, Víctor Rodríguez‐Galiano, Sabine Chabrillat

et al.

CATENA, Journal Year: 2025, Volume and Issue: 252, P. 108889 - 108889

Published: March 4, 2025

Language: Английский

Citations

0
Soil Organic Carbon Prediction and Mapping in Morocco Using PRISMA Hyperspectral Imagery and Meta-Learner Model DOI Creative Commons
Yassine Bouslıhım, Abdelkrim Bouasria, Budiman Minasny

et al.

Remote Sensing, Journal Year: 2025, Volume and Issue: 17(8), P. 1363 - 1363

Published: April 11, 2025

Accurate mapping of soil organic carbon (SOC) supports sustainable land management practices and accounting initiatives for mitigating climate change impacts. This study presents a novel meta-learner framework that combines multiple machine learning algorithms spectra processing to optimize SOC prediction using the PRISMA hyperspectral satellite imagery in Doukkala plain Morocco. The employs two-layer structure models. first layer consists Random Forest (RF), Support Vector Regression (SVR), Partial Least Squares (PLSR). These base models were configured data smoothing, transformation, spectral feature selection techniques, based on 70/30% split. second utilizes ridge regression model as integrate predictions from Results indicated RF SVR performance improved primarily with selection, while PLSR was most influenced by smoothing. approach outperformed individual models, achieving an average relative improvement 48.8% over single R2 0.65, RMSE 0.194%, RPIQ 2.247. contributes development methodologies predicting properties data.

Language: Английский

Citations

0
Application of Machine Learning Algorithms for Digital Mapping of Soil Salinity Levels and Assessing Their Spatial Transferability in Arid Regions DOI Creative Commons
Magboul M. Sulieman, Fuat Kaya, Mohammed Abdalla Elsheikh

et al.

Land, Journal Year: 2023, Volume and Issue: 12(9), P. 1680 - 1680

Published: Aug. 28, 2023

A comprehensive understanding of soil salinity distribution in arid regions is essential for making informed decisions regarding agricultural suitability, water resource management, and land use planning. methodology was developed to identify Sudan by utilizing optical radar-based satellite data as well variables obtained from digital elevation models that are known indicate variations salinity. The includes the transfer areas where similar conditions prevail. geographically coordinated database established, incorporating a variety environmental based on Google Earth Engine (GEE) Electrical Conductivity (EC) measurements saturation extract samples collected at three different depths (0–30, 30–60, 60–90 cm). Thereafter, Multinomial Logistic Regression (MNLR) Gradient Boosting Algorithm (GBM), were utilized spatially classify levels region. To determine applicability model trained reference site target area, Multivariate Environmental Similarity Surface (MESS) analysis conducted. producer’s accuracy, user’s Tau index parameters used evaluate model’s spatial confusion indices computed assess uncertainty. At depths, values area ranged 0.38 0.77, whereas 0.66 0.88, decreasing depth increased. Clay normalized ratio (CLNR), Salinity Index 1, SAR important modeling. It found subsoils middle northwest both had higher level compared topsoil. This study highlighted effectiveness means identifying evaluating management facing significant salinity-related challenges. approach can be instrumental alternative suitable activities regional level.

Language: Английский

Citations

9
Combining Digital Covariates and Machine Learning Models to Predict the Spatial Variation of Soil Cation Exchange Capacity DOI Creative Commons
Fuat Kaya, Gaurav Mishra, Rosa Francaviglia

et al.

Land, Journal Year: 2023, Volume and Issue: 12(4), P. 819 - 819

Published: April 3, 2023

Cation exchange capacity (CEC) is a soil property that significantly determines nutrient availability and effectiveness of fertilizer applied in lands under different managements. CEC’s accurate high-resolution spatial information needed for the sustainability agricultural management on farms Nagaland state (northeast India) which are fragmented intertwined with forest ecosystem. The current study digital mapping (DSM) methodology, based CEC values determined samples obtained from 305 points region, mountainous difficult to access. Firstly, auxiliary data were three open-access sources, including indices generated time series Landsat 8 OLI satellite, topographic variables derived elevation model (DEM), WorldClim dataset. Furthermore, used Lasso regression (LR), stochastic gradient boosting (GBM), support vector (SVR), random (RF), K-nearest neighbors (KNN) machine learning (ML) algorithms systematically compared R-Core Environment Program. Model performance evaluated square root mean error (RMSE), determination coefficient (R2), absolute (MAE) 10-fold cross-validation (CV). lowest RMSE was by RF algorithm 4.12 cmolc kg−1, while others following order: SVR (4.27 kg−1) <KNN (4.45 <LR (4.67 <GBM (5.07 kg−1). In particular, WorldClim-based climate covariates such as annual temperature (BIO-1), precipitation (BIO-12), elevation, solar radiation most important all algorithms. High uncertainty (SD) have been found areas low sampling density this finding be considered future surveys.

Language: Английский

Citations

8