Global Soil Salinity Prediction by Open Soil Vis-NIR Spectral Library DOI Creative Commons
Yin Zhou, Songchao Chen, Bifeng Hu

et al.

Remote Sensing, Journal Year: 2022, Volume and Issue: 14(21), P. 5627 - 5627

Published: Nov. 7, 2022

Soil salinization is one of the major degradation processes threatening food security and sustainable development. Detailed soil salinity information increasingly needed to tackle this global challenge for improving management. Soil-visible near-infrared (Vis-NIR) spectroscopy has been proven be a potential solution estimating soil-salinity-related (i.e., electrical conductivity, EC) rapidly cost-effectively. However, previous studies were mainly conducted at field, regional, or national scale, so application Vis-NIR scale needs further investigation. Based on an extensive open spectral library (61,486 samples with both EC spectra), we compared four predictive models (PLSR, Cubist, Random Forests, XGBoost) in EC. Our results indicated that XGBoost had best model performance (R2 0.59, RMSE 1.96 dS m−1) predicting whereas PLSR relatively limited ability 0.39, 2.41 m−1). The also showed auxiliary environmental covariates coordinates, elevation, climatic variables) could greatly improve prediction accuracy by models, performed 0.71, 1.65 outcomes study provide valuable reference broad-scale coupling spectroscopic technique easily obtainable covariates.

Language: Английский

High-resolution agriculture soil property maps from digital soil mapping methods, Czech Republic DOI
Daniel Žížala, Robert Minařík, Jan Skála

et al.

CATENA, Journal Year: 2022, Volume and Issue: 212, P. 106024 - 106024

Published: Jan. 19, 2022

Language: Английский

Citations

53
Improving soil organic carbon predictions from a Sentinel–2 soil composite by assessing surface conditions and uncertainties DOI Creative Commons
Klara Dvorakova, Uta Heiden, Karin Pepers

et al.

Geoderma, Journal Year: 2022, Volume and Issue: 429, P. 116128 - 116128

Published: Nov. 10, 2022

Soil organic carbon (SOC) prediction from remote sensing is often hindered by disturbing factors at the soil surface, such as photosynthetic active and non–photosynthetic vegetation, variation in moisture or surface roughness. With increasing amount of freely available satellite data, recent studies have focused on stabilizing reflectance building composites using time series images. Although composite imagery has demonstrated its potential SOC prediction, it still not well established if resulting spectra mirror fingerprint optimal conditions to predict topsoil properties (i.e. a smooth, dry bare soil). We collected 303 photos surfaces Belgian loam belt where five main classes were distinguished: smooth seeded soils, crusts, partial cover growing crop, moist soils crop residue cover. Reflectance then extracted Sentinel–2 images coinciding with date photos. After was removed an NDVI < 0.25, Normalized Burn Ratio (NBR2) calculated characterize threshold NBR2 0.05 found be able separate unfavorable i.e. wet covered residues. Additionally, we that normalizing dividing each band mean all spectral bands) allows for cancelling albedo shift between crusts seed–bed conditions. built exposed southern Belgium part Noord-Holland Flevoland Netherlands (covering spring periods 2016–2021). used per pixel content means Partial Least Squares Regression Model (PLSR) 10–fold cross–validation. The uncertainty models assessed via interval ratio (PIR). cross validation model gave satisfactory results (mean 100 bootstraps: efficiency coefficient (MEC) = 0.48 ± 0.07, RMSE 3.5 0.3 g C kg–1, RPD 1.4 0.1 RPIQ 1.9 0.3). maps show decreases when number scenes increases, reaches minimum least six are PIR pixels 12.4 while predicted 14.1 kg–1). against independent data set showed median difference 0.5 kg–1 2.8 measured (average 13.5 kg–1) contents field scale. Overall, this compositing method shows both realistic within regional patterns.

Language: Английский

Citations

52
Mapping Brazilian soil mineralogy using proximal and remote sensing data DOI Creative Commons
Nícolas Augusto Rosin, José Alexandre Melo Demattê, Raúl Roberto Poppiel

et al.

Geoderma, Journal Year: 2023, Volume and Issue: 432, P. 116413 - 116413

Published: March 8, 2023

Minerals control many soil functions and play a crucial role in addressing global existential issues. Measuring the abundance of minerals is laborious, costly, time-consuming task; however, spectroscopy can be useful tool to overcome this issue. This work aimed map major mineralogical components soils Brazil from surface 1 m deep at spatial resolution 30 m. Spectral data Brazilian Soil Library with Vis-NIR-SWIR was used estimate haematite, goethite, kaolinite, gibbsite. These were spatialized using digital mapping techniques. We also developed novel framework obtain bare reflectance for areas without natural or anthropic exposure (continuous image) it as covariate. their abundances successfully estimated by reflectance. Haematite predictions presented most accurate results Random Forest models, followed gibbsite, goethite. The validation reference found R2 0.64 (haematite), 0.40 (goethite), 0.20 (kaolinite/Kt), 0.29 (gibbsite/Gbs), (Kt/Kt + Gbs). resulting maps accordance geology, pedology, climate, relief revealed distribution mineral finer than existing geological pedological maps, reaching farm level detail.

Language: Английский

Citations

29
Transferability of Covariates to Predict Soil Organic Carbon in Cropland Soils DOI Creative Commons
Tom Broeg, Michael Blaschek, Steffen Seitz

et al.

Remote Sensing, Journal Year: 2023, Volume and Issue: 15(4), P. 876 - 876

Published: Feb. 4, 2023

Precise knowledge about the soil organic carbon (SOC) content in cropland soils is one requirement to design and execute effective climate food policies. In digital mapping (DSM), machine learning algorithms are used predict properties from covariates derived traditional mapping, elevation models, land use, Earth observation (EO). However, such DSM models trained for a specific dataset region have so far only allowed limited general statements be made that would enable transferred different regions. this study, we test transferability of SOC using five covariate groups: multispectral reflectance composites (satellite), legacy data (soil), model derivatives (terrain), parameters (climate), combined (combined). The was analyzed two federal states southern Germany: Bavaria Baden-Wuerttemberg. First, baseline were each state with performing best both cases (R2 = 0.68/0.48). Next, tested samples other whose not during calibration. Only satellite transferable, but accuracy declined cases. final step, (mixed-data models) applied separately. This process significantly improved accuracies satellite, terrain, while it showed no effect on decreased based covariates. experiment underlines importance EO transfer extrapolation models.

Language: Английский

Citations

24
Soil health in Latin America and the Caribbean DOI Creative Commons
Raúl Roberto Poppiel, Maurício Roberto Cherubin, Jean Jesus Novais

et al.

Communications Earth & Environment, Journal Year: 2025, Volume and Issue: 6(1)

Published: Feb. 24, 2025

Language: Английский

Citations

1
High resolution middle eastern soil attributes mapping via open data and cloud computing DOI
Raúl Roberto Poppiel, José Alexandre Melo Demattê, Nícolas Augusto Rosin

et al.

Geoderma, Journal Year: 2020, Volume and Issue: 385, P. 114890 - 114890

Published: Dec. 29, 2020

Language: Английский

Citations

51
Soil property maps with satellite images at multiple scales and its impact on management and classification DOI
Nélida Elizabet Quiñonez Silvero, José Alexandre Melo Demattê, Júlia de Souza Vieira

et al.

Geoderma, Journal Year: 2021, Volume and Issue: 397, P. 115089 - 115089

Published: March 30, 2021

Language: Английский

Citations

50
An Analysis of Bare Soil Occurrence in Arable Croplands for Remote Sensing Topsoil Applications DOI Creative Commons
Nada Mzid, Stefano Pignatti, Wenjiang Huang

et al.

Remote Sensing, Journal Year: 2021, Volume and Issue: 13(3), P. 474 - 474

Published: Jan. 29, 2021

A better comprehension of soil properties and processes permits a progress in agricultural management effectiveness, together with diminution environmental damage more beneficial use resources. This research investigated the usage multispectral (Sentinel-2 MSI) satellite data at farm/regional level, for identification agronomic bare presence, utilizing bands spectral range from visible to shortwave infrared. The purpose was assess frequency cloud-free time-series images available during year typical areas, needed development digital mapping (DSM) approaches applications, using hyperspectral missions such as current PRISMA planned EnMAP or CHIME. exploited Google Earth Engine platform, by processing all Sentinel-2 throughout time span four years. Two main results were obtained: (i) frequency, indicating where when pixel (or an field) detected surface three representative areas Italy, (ii) temporal sensitivity analysis, providing acquisition useful applicable retrieval variables interest. It shown that, order provide effective monitoring capability, revisit five seven days is required, which less than specifications e.g., CHIME missions, but could be addressed combining two sensors.

Language: Английский

Citations

47
Sentinel-2 and Landsat-8 Multi-Temporal Series to Estimate Topsoil Properties on Croplands DOI Creative Commons
Fabio Castaldi

Remote Sensing, Journal Year: 2021, Volume and Issue: 13(17), P. 3345 - 3345

Published: Aug. 24, 2021

The spatial and temporal monitoring of soil organic carbon (SOC), other properties related to erosion, is extremely important, both from the environmental economic perspectives. Sentinel-2 (S2) Landsat-8 (L8) time series increase probability observe bare fields in croplands, thus, monitor over large regions. In this regard, work suggests an automated pixel-based approach select only pure pixels S2 L8 series, make a synthetic image (SBSI). SBSIs measured framework European LUCAS survey were used calibrate SOC, clay, CaCO3 prediction models. results highlight high correlation between laboratory spectra median spectra, especially for SBSI obtained by three-year collection, which provides satisfactory terms SOC accuracy (RPD: 1.74). comparison demonstrated higher capability sensor accuracy, mainly due greater resolution bands visible region. Whereas, neither nor could accurately predict clay content. This because low spectral their SWIR that prevent exploitation narrow features these two attributes. study prove can estimate croplands using selects retrieves reliable spectra.

Language: Английский

Citations

46
Spatial prediction of soil surface properties in an arid region using synthetic soil image and machine learning DOI
Salman Naimi, Shamsollah Ayoubi, José Alexandre Melo Demattê

et al.

Geocarto International, Journal Year: 2021, Volume and Issue: 37(25), P. 8230 - 8253

Published: Oct. 23, 2021

Evaluation of spatial variability and mapping soil properties is critical for sustainable agricultural production in arid lands. The main objectives the present study were to spatialize organic carbon (SOC), particle size distribution(clay, sand, silt contents), calcium carbonate equivalent (CCE) by integrating multisource environmental covariates, including digital elevation model (DEM) remote sensing data machine learning (Cubist, Cu random forest, RF) an region Iran. Additionally, Synthetic Soil Images (SySI) achieved from multi-temporal images bare pixels based on Landsats 4, 5, 7, 8, a DEM. Three hundred topsoil samples (0–30 cm depth) collected conditioned Latin hypercube sampling (cLHS) approach Afzar district, Fars province, southern models calibrated validated 10-fold cross-validation approach, performance was evaluated using root mean square error (RMSE), ratio interquartile distance (RPIQ), coefficient determination (R2). Also, prediction accuracy assessed relative RMSE (rRMSE). best RPIQ index showed that predicting clay (1.89) had good prediction, sand (1.64), SOC (1.55), CCE (1.59) fair while (1.13) performed poorly. We found RF highest lowest accuracies (rRMSE = 14.31%) 43.93%), respectively. discovered combination high-quality RS variables derived DEM reasonably able predict properties. revealed strong promise enhance mapping, especially regions with limited data. Moreover, application can reduce cost and, accordingly, mapping.

Language: Английский

Citations

46