Retrieval and Mapping of Soil Organic Carbon Using Sentinel-2A Spectral Images from Bare Cropland in Autumn

Abstract

Soil is the largest carbon reservoir on the terrestrial surface. Soil organic carbon (SOC) not only regulates global climate change, but also indicates soil fertility level in croplands. SOC prediction based on remote sensing images has generated great interest in the research field of digital soil mapping. The short revisiting time and wide spectral bands available from Sentinel-2A (S2A) remote sensing data can provide a useful data resource for soil property prediction. However, dense soil surface coverage reduces the direct relationship between soil properties and S2A spectral reflectance such that it is difficult to achieve a successful SOC prediction model. Observations of bare cropland in autumn provide the possibility to establish accurate SOC retrieval models using the S2A super-spectral reflectance. Therefore, in this study, we collected 225 topsoil samples from bare cropland in autumn and measured the SOC content. We also obtained S2A spectral images of the western Guanzhong Plain, China. We established four SOC prediction models, including random forest (RF), support vector machine (SVM), partial least-squares regression (PLSR), and artificial neural network (ANN) based on 15 variables retrieved from the S2A images, and compared the prediction accuracy using RMSE (root mean square error), R2 (coefficient of determination), and RPD (ratio of performance to deviation). Based on the optimal model, the spatial distribution of SOC was mapped and analyzed. The results indicated that the inversion model with the RF algorithm achieved the highest accuracy, with an R2 of 0.8581, RPD of 2.1313, and RMSE of 1.07. The variables retrieved from the shortwave infrared (SWIR) bands (B11 and B12) usually had higher variable importance, except for the ANN model. SOC content mapped with the RF model gradually decreased with increasing distance from the Wei river, and values were higher in the west than in the east. These results matched the SOC distribution based on measurements at the sample sites. This research provides evidence that soil properties such as SOC can be retrieved and spatially mapped based on S2A images that are obtained from bare cropland in autumn.

Estimation of Soil Organic Carbon Content in the Ebinur Lake Wetland, Xinjiang, China, Based on Multisource Remote Sensing Data and Ensemble Learning Algorithms

“…Therefore, vegetation indices derived from remotely sensed data, soil moisture and brightness indices can represent vegetation biomass and soil characteristics [ 36 , 68 ]. Wang, K. found NDVI, BI, BI2 and SATVI to be the most important variables for predicting farmland SOC in autumn [ 69 ], which is consistent with the results of this paper. Due to the sensitivity of SAVI to soil properties [ 70 ], it is considered the most important variable in agricultural soil organic carbon prediction.…”

Estimation of soil organic carbon in agricultural fields: A remote sensing approach

“…Madugundu et al: Soil organic carbon estimation using satellite images relationship between VIs and SOC. In this study, linear relationships between the SWIR-1 band, VI from the L8 and S2A satellite data were observed against the SOC and were A confirmed with a similar study by Wang et al (2021). The analysis of S2A data showed that the two SWIR broad bands centered at 1610 and 2190 nm are sufficient to satisfactorily predict SOC content, as a result, high values of BSI and low RSR areas were associated with low SOC areas, and vice versa.…”

“…While, vegetation indices (VI), such as the Normalized Difference Vegetation Index (NDVI), NDVIRedEdge, Enhanced Vegetation Index (EVI), Bare Soil Index (BSI), and Reduced Simple Ratio (RSR) were computed and subsequently used for the development of SOC prediction models. ) employed on satellite images such as Landsat (Viscarra-Rossel and Bouma, 2016;Kumar et al, 2016), Sentinel-2 (Wang et al, 2021;Dvorakova et al, 2021), etc. Many studies have been conducted to investigate the characteristics of soil spectral reflection in regions of moderate to high soil fertility levels, but studies in low-fertility soils are still limited in the Arab region, such as Saudi Arabia.…”

Sentinel-2 and Landsat-8 Multi-Temporal Series to Estimate Topsoil Properties on Croplands

“…This is due to the combination of several factors linked with the sensor characteristic (range and spectral resolution) and the distance between the sensor and target surface (atmospheric disturbance, signal quality, spectral and spatial resolution), and the soil surface conditions (moisture, roughness, vegetation residues). Despite this, many recent scientific papers demonstrated the capability of the Copernicus Sentinel-2 Multi-Spectral Instrument (MSI) (hereinafter referred to as S2) and NASA Landsat-8 Operational Land Imager (OLI) (hereinafter referred to as L8) optical data for soil properties prediction and mapping [15][16][17][18][19][20][21], obtaining encouraging results, especially for the SOC content.…”