Mapping Soil Organic Carbon Research Articles

Estimation of soil organic carbon by combining hyperspectral and radar remote sensing to reduce coupling effects of soil surface moisture and roughness

Soil organic carbon (SOC) is important in the global carbon cycle. Accurate estimation of SOC content in cultivated land is a prerequisite for evaluating the carbon sequestration potential and quality of soils. However, existing SOC prediction studies based on hyperspectral remote sensing neglect the spectral response of the physical properties of surface soil, leading to inadequate model generalization. With the exponential growth of remote sensing data, the development of pixel-level soil spectral correction methods based on multi-source remote sensing data has become an interesting and challenging topic. This method aims to minimize the effect of soil physical properties on spectra, thus addressing the poor spatiotemporal transferability of SOC prediction models due to uncertain variations in surface soil physical properties. In this study, a soil spectral correction strategy is constructed using satellite hyperspectral image (HSI) and synthetic aperture radar (SAR) images through multi-order polynomial regression and convolutional neural networks. This strategy considers soil physical variables such as soil moisture (SM) content and root mean square height (RMSH) of soil surface roughness. The soil spectral correction model and SOC content prediction model were established using 80 soil samples collected from Site 1. Afterward, the performance and transferability of both models were verified using the remaining 25 samples from Site 1 and 50 samples from Site 2. The results showed that: 1) The effect of SM and RMSH on the soil pixel spectrum can be significantly reduced after correcting HSI using soil spectral correction strategy. The correlation coefficients between the corrected pixel spectrum and the ground-based spectrum increase by over 60 % compared with those between the original spectrum and the ground-based spectrum. 2) Soil spectral correction improves the prediction accuracy and mapping capability of HSI for SOC content, with the highest RP2 of 0.743 and RMSEP of 3.455 g/kg at Site 1. 3) Compared with the original HSI-based SOC prediction model, the soil spectral correction strategy based on multi-order polynomial and convolutional neural network reduced the RMSEP of SOC prediction results at Site 2 by 5.082 g/kg and 5.454 g/kg, and the RP2 increased by 0.390 and 0.409, respectively. 4) When predicting SOC content from raw HIS, SM and RMSH contribute to more than 60 % of the bias, with SM having a larger bias than RMSH. The findings of this study emphasize the influence of soil physical properties on SOC prediction and contribute to the existing research on SOC mapping using HSI and SAR data.

Using local ensemble models and Landsat bare soil composites for large-scale soil organic carbon maps in cropland

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

Mapping Soil Organic Carbon Stock Using Hyperspectral Remote Sensing: A Case Study in the Sele River Plain in Southern Italy

Mapping soil organic carbon (SOC) stock can serve as a resilience indicator for climate change. As part of the carbon dioxide (CO2) sink, soil has recently become an integral part of the global carbon agenda to mitigate climate change. We used hyperspectral remote sensing to model the SOC stock in the Sele River plain located in the Campania region in southern Italy. To this end, a soil spectral library (SSL) for the Campania region was combined with an aerial hyperspectral image acquired with the AVIRIS–NG sensor mounted on a Twin Otter aircraft at an altitude of 1433 m. The products of this study were four raster layers with a high spatial resolution (1 m), representing the SOC stocks and three other related soil attributes: SOC content, clay content, and bulk density (BD). We found that the clay minerals’ spectral absorption at 2200 nm has a significant impact on predicting the examined soil attributes. The predictions were performed by using AVIRIS–NG sensor data over a selected plot and generating a quantitative map which was validated with in situ observations showing high accuracies in the ground-truth stage (OC stocks [RPIQ = 2.19, R2 = 0.72, RMSE = 0.07]; OC content [RPIQ = 2.27, R2 = 0.80, RMSE = 1.78]; clay content [RPIQ = 1.6 R2 = 0.89, RMSE = 25.42]; bulk density [RPIQ = 1.97, R2 = 0.84, RMSE = 0.08]). The results demonstrated the potential of combining SSLs with remote sensing data of high spectral/spatial resolution to estimate soil attributes, including SOC stocks.

Digital mapping of soil organic carbon in volcanic soils after prolonged eruption Mt. Sinabung, Karo Regency, North Sumatra, Indonesia

Mount Sinabung (at Karo Regency, North Sumatra) has been considered as an active volcano since 2010, and after three years of inactivity, the volcano’s activity resumed in 2013. The deposition of volcanic ash buried everything on the soil surface with various thicknesses. Volcanic ash is a valuable inorganic material and consists mostly of primary minerals. But with time, these materials can initiate carbon storage through the revegetation process. This study investigates carbon storage and sequestration in volcanic soils affected by the intermittent eruptions of Mt. Sinabung. A total of 34 soil samples were collected from areas of 3 to 7 km from the eruptive vent. The samples were analyzed to determine labile-C, very labile-C, total-C, Organic-C, non-crystalline-C, and metal complex-C fractions. Regression kriging (RK) was applied to spatially predict the carbon distribution. The results showed that the highest labile-C was 1.65% found in the Southeast slope and the lowest, 1.20%, in the Southern slope. The highest value of very labile-C was in the Southeast (1.20%), and the lowest was in the Northern (0.46%). The highest (11.66%) and lowest values (8.84%) of total C were in the Northeast and South slopes, respectively. The highest Organic-C value was detected 7.42% in the Northeast volcanic soils, and the lowest, 5.74%, is in the South slopes soil. The lowest non-crystalline- C was found in the soils of Northeast slope (0.60%), and the highest was in the Southeast (0.82%) soil. While the highest metal complex-C value was 0.88% in the Southeast, and the lowest is 0.36% in the soil of South slope. These data show that the highest carbon storage is parallel with the direction of the ash distribution, which tends to the Southeast side of Mt. Sinabung. The results of this study demonstrated that volcanic regions have a strong resilience capacity to bounce back after devastating natural hazards. Volcanic deposits interact with atmospheric water to initiate vegetation regrowth and create a carbon pool within their particles.

Effect of training sample size, sampling design and prediction model on soil mapping with proximal sensing data for precision liming

Site-specific estimation of lime requirement requires high-resolution maps of soil organic carbon (SOC), clay and pH. These maps can be generated with digital soil mapping models fitted on covariates observed by proximal soil sensors. However, the quality of the derived maps depends on the applied methodology. We assessed the effects of (i) training sample size (5–100); (ii) sampling design (simple random sampling (SRS), conditioned Latin hypercube sampling (cLHS) and k-means sampling (KM)); and (iii) prediction model (multiple linear regression (MLR) and random forest (RF)) on the prediction performance for the above mentioned three soil properties. The case study is based on conditional geostatistical simulations using 250 soil samples from a 51 ha field in Eastern Germany. Lin’s concordance correlation coefficient (CCC) and root-mean-square error (RMSE) were used to evaluate model performances. Results show that with increasing training sample sizes, relative improvements of RMSE and CCC decreased exponentially. We found the lowest median RMSE values with 100 training observations i.e., 1.73%, 0.21% and 0.3 for clay, SOC and pH, respectively. However, already with a sample size of 10, models of moderate quality (CCC > 0.65) were obtained for all three soil properties. cLHS and KM performed significantly better than SRS. MLR showed lower median RMSE values than RF for SOC and pH for smaller sample sizes, but RF outperformed MLR if at least 25–30 or 75–100 soil samples were used for SOC or pH, respectively. For clay, the median RMSE was lower with RF, regardless of sample size.

A Deep Learning Approach to Estimate Soil Organic Carbon from Remote Sensing

Monitoring soil organic carbon (SOC) typically assumes conducting a labor-intensive soil sampling campaign, followed by laboratory testing, which is both expensive and impractical for generating useful, spatially continuous data products. The present study leverages the power of machine learning (ML) and, in particular, deep neural networks (DNNs) for segmentation, as well as satellite imagery, to estimate the SOC remotely. We propose a new two-stage pipeline for remote SOC estimation, which relies on using a DNN trained to classify land cover to perform feature extraction, while the SOC estimation is performed by a different ML model. The first stage is an image segmentation DNN with the U-Net architecture, which is trained to estimate the land cover for an observed geographical region, based on multi-spectral images taken by the Sentinel-2 satellite constellation. This estimator is subsequently used to extract the latent feature vector for each of the output pixels, by rolling back from the output (dense) layer of the U-Net and accessing the last available convolutional layer of the same dimension as our desired output. The second stage is trained on a set of feature vectors extracted at the coordinates for which manual SOC measurements exist. We tested a variety of ML models and report on their performance. Using the best extremely randomized trees model, we generated a spatially continuous map of SOC estimations for the region of Tuscany, in Italy, with a resolution of 10 m, to share with the researchers as a means of validating the results and to demonstrate the efficiency of the proposed approach, which can can easily be scaled to create a global continuous SOC map.

Soil Loss Estimation by Water Erosion in Agricultural Areas Introducing Artificial Intelligence Geospatial Layers into the RUSLE Model

The existing digital soil maps are mainly characterized by coarse spatial resolution and are not up to date; thus, they are unable to support the physical process-based models for improved predictions. The overarching objective of this work is oriented toward a data-driven approach and datacube-based tools (Soil Data Cube), leveraging Sentinel-2 imagery data, open access databases, ground truth soil data and Artificial Intelligence (AI) architectures to provide enhanced geospatial layers into the Revised Universal Soil Loss Equation (RUSLE) model, improving both the reliability and the spatial resolution of the final map. The proposed methodology was implemented in the agricultural area of the Imathia Regional Unit (northern Greece), which consists of both mountainous areas and lowlands. Enhanced soil maps of Soil Organic Carbon (SOC) and soil texture were generated at 10 m resolution through a time-series analysis of satellite data and an XGBoost (eXtrene Gradinent Boosting) model. The model was trained by 84 ground truth soil samples (collected from agricultural fields) taking into account also additional environmental covariates (including the digital elevation model and climatic data) and following a Digital Soil Mapping (DSM) approach. The enhanced layers were introduced into the RUSLE’s soil erodibility factor (K-factor), producing a soil erosion layer with high spatial resolution. Notable prediction accuracy was achieved by the AI model with R2 0.61 for SOC and 0.73, 0.67 and 0.63 for clay, sand, and silt, respectively. The average annual soil loss of the unit was found to be 1.76 ton/ha/yr with 6% of the total agricultural area suffering from severe erosion (>11 ton/ha/yr), which was mainly found in the mountainous border regions, showing the strong influence of the mountains in the agricultural fields. The overall methodology could strongly support regional decision making and planning and environmental policies such as the European Common Agricultural Policy (CAP) and the Sustainable Development Goals (SDGs).

Upscaling Soil Organic Carbon Measurements at the Continental Scale Using Multivariate Clustering Analysis and Machine Learning

AbstractEstimates of soil organic carbon (SOC) stocks are essential for many environmental applications. However, significant inconsistencies exist in SOC stock estimates for the U.S. across current SOC maps. We propose a framework that combines unsupervised multivariate geographic clustering (MGC) and supervised Random Forests regression, improving SOC maps by capturing heterogeneous relationships with SOC drivers. We first used MGC to divide the U.S. into 20 SOC regions based on the similarity of covariates (soil biogeochemical, bioclimatic, biological, and physiographic variables). Subsequently, separate Random Forests models were trained for each SOC region, utilizing environmental covariates and SOC observations. Our estimated SOC stocks for the U.S. (52.6 ± 3.2 Pg for 0–30 cm and 108.3 ± 8.2 Pg for 0–100 cm depth) were within the range estimated by existing products like Harmonized World Soil Database, HWSD (46.7 Pg for 0–30 cm and 90.7 Pg for 0–100 cm depth) and SoilGrids 2.0 (45.7 Pg for 0–30 cm and 133.0 Pg for 0–100 cm depth). However, independent validation with soil profile data from the National Ecological Observatory Network showed that our approach (R2 = 0.51) outperformed the estimates obtained from Harmonized World Soil Database (R2 = 0.23) and SoilGrids 2.0 (R2 = 0.39) for the topsoil (0–30 cm). Uncertainty analysis (e.g., low representativeness and high coefficients of variation) identified regions requiring more measurements, such as Alaska and the deserts of the U.S. Southwest. Our approach effectively captures the heterogeneous relationships between widely available predictors and the current SOC baseline across regions, offering reliable SOC estimates at 1 km resolution for benchmarking Earth system models.

Satellite-based soil organic carbon mapping on European soils using available datasets and support sampling

Soil organic carbon (SOC) plays a major role in the global carbon cycle and is an important factor for soil health and fertility. Accurate mapping of SOC and other influencing parameters are crucial to guide the optimization of agricultural land management to maintain and restore soil health, to increase soil fertility, and thus to quantify its potential for sequestering CO2. Remote sensing and machine learning techniques offer promising approaches for predicting SOC distribution. In this study, we used remote sensing data and machine learning algorithms to map SOC at regional to large scale, which we then combined with temporospatial and spectral signature-based soil sampling to integrate local ground measurements. A rigorous validation approach was performed where several independent unseen datasets with a high number of samples were used, which additionally involved densely sampled fields. We found that our approach could predict SOC with an average percentage error of less than 10 % with an R2 of 0.91 using support sampling on croplands located on mineral soils, demonstrating the potential of remote sensing, machine learning, and specific ground measurements for mapping SOC. Our results suggest that this approach could make small carbon differences measurable and inform carbon sequestration efforts and improve our understanding of the impacts of land use and field management practices on soil carbon cycling.

Revealing the hidden carbon in forested wetland soils

Inland wetlands are critical carbon reservoirs storing 30% of global soil organic carbon (SOC) within 6% of the land surface. However, forested regions contain SOC-rich wetlands that are not included in current maps, which we refer to as ‘cryptic carbon’. Here, to demonstrate the magnitude and distribution of cryptic carbon, we measure and map SOC stocks as a function of a continuous, upland-to-wetland gradient across the Hoh River Watershed (HRW) in the Pacific Northwest of the U.S., comprising 68,145 ha. Total catchment SOC at 30 cm depth (5.0 TgC) is between estimates from global SOC maps (GSOC: 3.9 TgC; SoilGrids: 7.8 TgC). For wetland SOC, our 1 m stock estimates are substantially higher (Mean: 259 MgC ha−1; Total: 1.7 TgC) compared to current wetland-specific SOC maps derived from a combination of U.S. national datasets (Mean: 184 MgC ha−1; Total: 0.3 TgC). We show that total unmapped or cryptic carbon is 1.5 TgC and when added to current estimates, increases the estimated wetland SOC stock to 1.8 TgC or by 482%, which highlights the vast stores of SOC that are not mapped and contained in unprotected and vulnerable wetlands.

Collaborative Utilization of Sentinel-1/2 and DEM Data for Mapping the Soil Organic Carbon in Forested Areas Based on the Random Forest

Optical remote sensing data are widely used for constructing soil organic carbon (SOC) mapping models. However, it is challenging to map SOC in forested areas because atmospheric water vapor affects the results derived from optical remote sensing data. To address this issue, we utilized Sentinel-1, Sentinel-2, and digital elevation model (DEM) data to obtain a comprehensive feature set (including S1-based textural indices, S2-based spectral indices, and DEM-derived indices) to map the SOC content in forested areas. The features set were the predictor variables, and the measured SOC content was the dependent variable. The random forest algorithm was used to establish the SOC model. The ratio of performance to inter-quartile range (RPIQ) was 2.92 when the S2-based spectral indices were used as predictor variables. When the comprehensive feature set was utilized as the model input, the model achieved an RPIQ of 4.13 (R2 = 0.91, root mean square error (RMSE) = 9.18), representing a 41.44% improvement in model accuracy. The average SOC content in the Greater Khingan Mountains was 43.75 g kg−1. The northern and southwestern parts had higher SOC contents (>54.93 g kg−1), while the southeastern and northwestern parts had lower contents (<39.83 g kg−1). This discrepancy was primarily attributed to agricultural activities. The results indicate that using a comprehensive feature set and the random forest algorithm is a reliable approach for estimating the spatial distribution of the SOC content in forested areas and is suitable for forest ecology and carbon management studies.

Comparing Laboratory and Satellite Hyperspectral Predictions of Soil Organic Carbon in Farmland

Mapping soil organic carbon (SOC) accurately is essential for sustainable soil resource management. Hyperspectral data, a vital tool for SOC mapping, is obtained through both laboratory and satellite-based sources. While laboratory data is limited to sample point monitoring, satellite hyperspectral imagery covers entire regions, albeit susceptible to external environmental interference. This study, conducted in the Yuncheng Basin of the Yellow River Basin, compared the predictive accuracy of laboratory hyperspectral data (ASD FieldSpec4) and GF-5 satellite hyperspectral imagery for SOC mapping. Leveraging fractional order derivatives (FODs), various denoising methods, feature band selection, and the Random Forest model, the research revealed that laboratory hyperspectral data outperform satellite data in predicting SOC. FOD processing enhanced spectral information, and discrete wavelet transform (DWT) proved effective for GF-5 satellite imagery denoising. Stability competitive adaptive re-weighted sampling (sCARS) emerged as the optimal feature band selection algorithm. The 0.6FOD-sCARS RF model was identified as the optimal laboratory hyperspectral prediction model for SOC, while the 0.8FOD-DWT-sCARS RF model was deemed optimal for satellite hyperspectral prediction. This research, offering insights into farmland soil quality monitoring and strategies for sustainable soil use, holds significance for enhancing agricultural production efficiency.

A high-resolution map of soil organic carbon in cropland of Southern China

An accurate and fine map of soil organic carbon (SOC) plays a vital role in understanding the global carbon cycle and achieving soil carbon sequestration potential. Although global and national maps of SOC are already available with various spatial resolutions, limited sample size, and relative coarse resolution hinder its accuracy and application in local regions. Here, we collected 13,424 soil samples and information of 45 environmental covariates from cropland in Jiangxi Province, Southern China during 2012 and 2013. Then, the optimal covariates were selected by a recursive feature elimination algorithm for mapping SOC. After that, we deployed the random forest (RF) to produce a fine map (30 m) of SOC in the cropland of Jiangxi Province and 100 times bootstrapping was performed to calculate the prediction uncertainty. Finally, we determined the impacts of various covariates on SOC variability using RF and partial least squares structural equation modeling. Our results showed that compared with the predictive model without soil management information, introducing soil management information improved the predictability of SOC with an increase in R2 by 7.35 % (0.73 vs 0.68) and decrease in RMSE by 7.03 % (2.91 vs 3.13 g kg−1). Our results well estimated the uncertainty of the predicted result with a PICP of 0.91 for a 90 % prediction interval. Soil properties and soil management activities make the largest contribution for modelling SOC. Specifically, the total nitrogen content, straw return amount, total potassium content, and multi-resolution valley bottom flatness were found as the most important factors for mapping SOC in the cropland of our study area. Overall, this study deepened our knowledge of the variation of SOC and also emphasizes that incorporating soil management information could help us to achieve more accurate predictions of SOC.

Mapping Soil Organic Carbon Stock and Uncertainties in an Alpine Valley (Northern Italy) Using Machine Learning Models

In this study, we conducted a comprehensive analysis of the spatial distribution of soil organic carbon stock (SOC stock) and the associated uncertainties in two soil layers (0–10 cm and 0–30 cm; SOC stock 10 and SOC stock 30, respectively), in Valchiavenna, an alpine valley located in northern Italy (450 km2). We employed the digital soil mapping (DSM) approach within different machine learning models, including multivariate adaptive regression splines (MARS), random forest (RF), support vector regression (SVR), and elastic net (ENET). Our dataset comprised soil data from 110 profiles, with SOC stock calculations for all sampling points based on bulk density (BD), whether measured or estimated, considering the presence of rock fragments. As environmental covariates for our research, we utilized environmental variables, in particular, geomorphometric parameters derived from a digital elevation model (with a 20 m pixel resolution), land cover data, and climatic maps. To evaluate the effectiveness of our models, we evaluated their capacity to predict SOC stock 10 and SOC stock 30 using the coefficient of determination (R2). The results for the SOC stock 10 were as follows: MARS 0.39, ENET 0.41, RF 0.69, and SVR 0.50. For the SOC stock 30, the corresponding R2 values were: MARS 0.45, ENET 0.48, RF 0.65, and SVR 0.62. Additionally, we calculated the root-mean-squared error (RMSE), mean absolute error (MAE), the bias, and Lin’s concordance correlation coefficient (LCCC) for further assessment. To map the spatial distribution of SOC stock and address uncertainties in both soil layers, we chose the RF model, due to its better performance, as indicated by the highest R2 and the lowest RMSE and MAE. The resulting SOC stock maps using the RF model demonstrated an accuracy of RMSE = 1.35 kg m−2 for the SOC stock 10 and RMSE = 3.36 kg m−2 for the SOC stock 30. To further evaluate and illustrate the precision of our soil maps, we conducted an uncertainty assessment and mapping by analyzing the standard deviation (SD) from 50 iterations of the best-performing RF model. This analysis effectively highlighted the high accuracy achieved in our soil maps. The maps of uncertainty demonstrated that the RF model better predicts the SOC stock 10 compared to the SOC stock 30. Predicting the correct ranges of SOC stocks was identified as the main limitation of the methodology.

Topographic correction of visible near‐infrared reflectance spectra for horizon‐scale soil organic carbon mapping

AbstractUnderstanding soil organic carbon (SOC) response to global change has been hindered by an inability to map SOC at horizon scales relevant to coupled hydrologic and biogeochemical processes. Standard SOC measurements rely on homogenized samples taken from distinct depth intervals. Such sampling prevents an examination of fine‐scale SOC distribution within a soil horizon. Visible near‐infrared hyperspectral imaging (HSI) has been applied to intact monoliths and split cores surfaces to overcome this limitation. However, the roughness of these surfaces can influence HSI spectra by scattering reflected light in different directions posing challenges to fine‐scale SOC mapping. Here, we examine the influence of prescribed surface orientation on reflected spectra, develop a method for correcting topographic effects, and calibrate a partial least squares regression (PLSR) model for SOC prediction. Two empirical models that account for surface slope, aspect, and wavelength and two theoretical models that account for the geometry of the spectrometer were compared using 681 homogenized soil samples from across the United States that were packed into sample wells and presented to the spectrometer at 91 orientations. The empirical approach outperformed the more complex geometric models in correcting spectra taken at non‐flat configurations. Topographically corrected spectra reduced bias and error in SOC predicted by PLSR, particularly at slope angles greater than 30°. Our approach clears the way for investigating the spatial distributions of multiple soil properties on rough intact soil samples.

Mapping soil organic carbon in northern France using adaptive zoning regression kriging based on LUCAS dataset

To improve the accuracy of soil organic carbon (SOC) mapping, an adaptive zoning regression kriging (AZ_RK) framework was proposed, which simultaneously considers spatial heterogeneity and the influence of multiple environmental variables. In AZ_RK, the study area was first adaptively divided into several spatially homogeneous zones. Then, the SOC distribution was mapped in each zone separately considering the relationships between SOC and environmental variables. The results indicated that the influence of environmental variables on SOC in different zones differed. The proposed zoning strategy was proven to be valid, which can account for 36.6% of the spatial variance in SOC. Furthermore, AZ_RK outperformed other methods with the lowest RMSE and the highest R 2 ( R cv 2 = 0.495 and R p 2 = 0.481). We concluded that AZ_RK can effectively cluster spatially homogeneous zones and adaptively map SOC with high estimation accuracy, which can serve as a valuable model reference for precise fertilization in regions characterized by spatial heterogeneity.

Predicting and Mapping Soil Organic Carbon Using Remote Sensing and Machine Learning Algorithms

Predicting and Mapping Soil Organic Carbon Using Remote Sensing and Machine Learning Algorithms

Sentinel-1 Imagery Used for Estimation of Soil Organic Carbon by Dual-Polarization SAR Vegetation Indices

Despite optical remote sensing (and the spectral vegetation indices) contributions to digital soil-mapping studies of soil organic carbon (SOC), few studies have used active radar remote sensing mission data like that from synthetic aperture radar (SAR) sensors to predict SOC. Bearing in mind the importance of SOC mapping for agricultural, ecological, and climate interests and also the recently developed methods for vegetation monitoring using Sentinel-1 SAR data, in this work, we aimed to take advantage of the high operationality of Sentinel-1 imaging to test the accuracy of SOC prediction at different soil depths using machine learning systems. Using linear, nonlinear, and tree regression-based methods, it was possible to predict the SOC content of soils from western Bahia, Brazil, a region with predominantly sandy soils, using as explanatory variables the SAR vegetation indices. The models fed with SAR sensor polarizations and vegetation indices produced more accurate results for the topsoil layers (0–5 cm and 5–10 cm in depth). In these superficial layers, the models achieved an RMSE in the order of 5.0 g kg−1 and an R2 ranging from 0.16 to 0.24, therefore explaining about 20% of SOC variability using only Sentinel-1 predictors.

China's conservation and restoration of coastal wetlands offset much of the reclamation-induced blue carbon losses.

China's coastal wetlands have experienced large losses and gains with rapid coastal reclamation and restoration since the end of the 20th century. However, owing to the difficulties in mapping soil organic carbon (SOC) in blue carbon stocks of coastal wetlands on a national scale, little is known about the spatial pattern of SOC stock in China's coastal wetlands and the loss and gain of SOC stock following coastal reclamation, conservation, and restoration over the past decades. Here, we developed a SOC stock map in China's coastal wetlands at 30 m spatial resolution, analyzed the spatial variability and driving factors of SOC stocks, and finally estimated SOC losses and gains due to coastal reclamation and wetland management from 1990 to 2020. We found that the total SOC stocks in China's coastal wetlands were 77.8 Tg C by 2020 with 3.6 Tg C in mangroves, 8.8 Tg C in salt marshes, and 65.4 Tg C in mudflats. Temperature, rainfall, and seawater salinity exerted the highest relative contributions to SOC spatial variability. The spatial trend of SOC density gradually decreased from south to north except for Liaoning province, with the lowest density in Shandong province. About 24.9% (19.4 Tg C) of SOC stocks in China's coastal wetlands were lost due to high-intensity reclamation, but SOC stock gained from conservation and restoration offset the reclamation-induced losses by 58.2% (11.3 Tg C) over the past three decades. These findings demonstrated the great potential of conservation and restoration of coastal wetlands in reversing the loss trend of blue carbon and contributing to the mitigation of climate change toward carbon neutrality. Our study provides significant spatial insights into the stocks, sequestration, and recovery capacity of blue carbon following rapid urbanization and management actions, which benefit the progress of global blue carbon management.

Effects of land conversion to cropland on soil organic carbon in montane soils of Northeast China from 1985 to 2020

Soil organic carbon (SOC) plays important roles in soil ecological function, soil conservation, and the global carbon cycle. With the increase of population and food demand, more and more croplands were reclaimed in mountainous areas. However, land conversion to cropland can lead to evident changes of SOC. The innovation of this study was to quantify the changes of SOC for different land covers in montane soils. Eight hundred twenty-nine topsoil samples were collected from dryland in Northeast China, and each soil sample was collected at least 200 m off the road at every 5–10 km interval, and five sites were mixed as one soil sample. Landsat 8 images were acquired from the bare soil period, and land cover data for 1985–2020 were obtained from Yang and Huang (2021). The surface reflectance of the 829 topsoil samples (0–20 cm) was extracted from Landsat 8 images. Then, a SOC prediction hybrid model was built after k-means clustering based on measured SOC using the potassium dichromate heating method, and a probability hybrid model was developed for SOC mapping. Land conversion to cropland was assigned to eight periods (i.e., five-year intervals from 1985 to 2020) based on land cover data, and the conversion percentages of different land cover types to cropland were calculated. Seven mountainous sub-regions were selected in Northeast China based on distribution of cropland and mountains to analyze the effects of land conversion to cropland on SOC. Our study showed that (1) the hybrid model had a higher accuracy than the global model, and the hybrid model led to the R2 of 0.77 and RMSE of 4.66 g kg−1 for validation samples. (2) Land conversion to cropland was mainly from forest in the 1980 s (proportion of 67.13%) and from grassland after 2015 (proportion of 52.24%). (3) SOC decreases with an average rate of 0.30 g kg -1 per five years from 1985 to 2020, and the slope of land conversion to cropland increases with an average rate of 0.35° per five years from 1985 to 2020; (4) SOC of land conversion to cropland from forests decreased by 2.04 g kg−1, and wetland has an evident decrease trend (decrease by 1.84 g kg−1) and an increasing trend from barren (increase by 4.07 g kg−1). Our results suggested that land conversion to cropland has a negative effect on SOC in montane soils and should control land conversion to cropland to protect natural resources.