Variability analysis of soil organic carbon content across land use types and its digital mapping using machine learning and deep learning algorithms

Environmental Monitoring and Assessment, Год журнала: 2025, Номер 197(5)

Опубликована: Апрель 10, 2025

Язык: Английский

---

A dynamic normalized difference index for estimating soil organic matter concentration using visible and near-infrared spectroscopy

Ecological Indicators, Год журнала: 2023, Номер 147, С. 110037 - 110037

Опубликована: Фев. 17, 2023

Accurate, rapid, and non-destructive estimation of soil organic matter (SOM) is crucial for fertility diagnosis precision farming. Due to the complicated unstable spectral characteristics SOM, few SOM indexes have been proposed widely used. In this paper, a new dynamic normalized difference index (DNDI) was constructed estimate using visible near-infrared spectroscopy. A correction factor α used adjust optimal wavelength range obtain more robust features SOM. Different pre-processing methods were applied compared. The support vector machine (SVM) model Partial least square regression (PLSR) calibrated based on DNDI To end, total 111 samples collected in southern coastal plain Laizhou Bay. results showed that by optimization could higher correlation with than two-dimensional (NDI). had maximum 0.88 from first derivative reflectance, NDI correlations most improved standard normal variate transform (SNV), reaching 0.81. For models, exhibited better performance, yielding validation R2, RMSE, RPD 0.78, 0.17 g·kg−1, 2.01, respectively. Our algorithm has strong application potential estimating other properties enhancing ground- satellite-based sensing.

Язык: Английский

---

Soil Data Cube and Artificial Intelligence Techniques for Generating National-Scale Topsoil Thematic Maps: A Case Study in Lithuanian Croplands

Remote Sensing, Год журнала: 2023, Номер 15(22), С. 5304 - 5304

Опубликована: Ноя. 9, 2023

There is a growing realization among policymakers that in order to pave the way for development of evidence-based conservation recommendations policy, it essential improve capacity soil-health monitoring by adopting multidimensional and integrated approaches. However, existing ready-to-use maps are characterized mainly coarse spatial resolution (>200 m) information not up date, making their use insufficient EU’s policy requirements, such as common agricultural policy. This work, utilizing Soil Data Cube, which self-hosted custom tool, provides yearly estimations soil thematic (e.g., exposed soil, organic carbon, clay content) covering all area Lithuania. The pipeline exploits various Earth observation data time series Sentinel-2 satellite imagery (2018–2022), LUCAS (Land Use/Cover Area Frame Statistical Survey) topsoil database, European Integrated Administration Control System (IACS) artificial intelligence (AI) architectures prediction accuracy well (10 m), enabling discrimination at parcel level. Five different models were tested with convolutional neural network (CNN) model achieve best both targeted indicators (SOC clay) related R2 metric (0.51 SOC 0.57 clay). predictions supported uncertainties based on PIR formula (average 0.48 0.61 provide valuable model’s interpretation stability. application final carried out national bare-soil-reflectance composite layers, generated employing pixel-based approach overlaid annual bare-soil using combination vegetation indices NDVI, NBR2, SCL. findings this work new insights generation large scale, leading more efficient sustainable management, supporting agri-food private sector.

Язык: Английский

---

Contribution of the Sentinel-2 spring seedbed spectra to the digital mapping of soil organic carbon concentration

Geoderma, Год журнала: 2024, Номер 449, С. 116984 - 116984

Опубликована: Авг. 1, 2024

Soil organic carbon (SOC) is central to the functioning of terrestrial ecosystems, has climate mitigation potential and provides several benefits for soil health. Understanding spatial distribution SOC can help formulate sustainable management practices. Digital mapping (DSM) uses advanced statistical geostatistical methods estimate properties across large areas. DSM integrates data, topographic features, geology, legacy maps, land remote sensing data. Bare spectra may reflect presence particular components, making satellite derived suitable predictors SOC. from Sentinel-2 were used concentration (SOC%) granulometric fractions in plough layer (0–30 cm) agricultural parcels northern Belgium. Thereafter, estimation performance SOC% was compared three models: one with bare spectra, environmental covariates (topography, granulometry vegetation), a combined model covariates. The sand, silt clay using spring seedbed (R2: 0.53–0.74; RPD: 1.49–2.05; RPIQ: 1.52–2.39) higher than that 0.16; 1.08; 1.32). highest obtained including all 0.28; 1.18; 1.44), but contribution containing small. results provide valuable insights refining property spectral

Язык: Английский

---

Factors controlling peat soil thickness and carbon storage in temperate peatlands based on UAV high-resolution remote sensing

Geoderma, Год журнала: 2024, Номер 449, С. 117009 - 117009

Опубликована: Авг. 23, 2024

Язык: Английский

---

A Framework for Retrieving Soil Organic Matter by Coupling Multi-Temporal Remote Sensing Images and Variable Selection in the Sanjiang Plain, China

Remote Sensing, Год журнала: 2023, Номер 15(12), С. 3191 - 3191

Опубликована: Июнь 20, 2023

Soil organic matter (SOM) is an important soil property for agricultural production. Rising grain demand has increased the intensity of cultivated land development in Sanjiang Plain China, and there a strong SOM monitoring this region. Therefore, Baoqing County Plain, production area, was considered study area. In study, we proposed framework high-accuracy retrieval by coupling multi-temporal remote sensing (RS) images variable selection algorithms. A total 73 surface samples (0–20 cm) were collected 2010, Landsat 5 acquired during bare period (April, May, June) selected from 2008 to 2011. Three algorithms, namely, Genetic Algorithm, Random Frog Competitive Adaptive Reweighted Sampling (CARS), combined with Partial Least Squares Regression (PLSR) build models on spectral bands indices images. The results using single-date image showed that combination algorithms PLSR outperformed alone, CARS best performance (R2 = 0.34, RMSE 15.66 g/kg) among all only applied different year interval groups. To investigate effect acquisition time, divided into various groups, resulting then stacked. accuracy improved as lengthened. optimal result 0.59, 11.81 obtained 2008–2011 group, wherein difference derived 2009, 2011 dominated variables. Moreover, spatial prediction based model consistent distribution SOM. Our suggested couples stacked RS potential retrieval.

Язык: Английский

---

Leveraging legacy data with targeted field sampling for low-cost mapping of soil organic carbon stocks on extensive rangeland properties

Geoderma, Год журнала: 2024, Номер 448, С. 116952 - 116952

Опубликована: Июль 5, 2024

Accurately quantifying high-resolution field-scale soil organic carbon (SOC) stocks is challenging yet crucial for improving site-specific land management and accounting. This challenge even greater when the study units are large heterogenous ranches. utilizes a digital mapping (DSM) approach U.S. legacy dataset, combined with soil, climate, biotic, topographic covariate datasets, to design targeted sampling plan acquiring local samples. The resulting samples were then used in combination data build optimal ranch-scale SOC stock models. We provide an example of this using ranch western as case study. In our we first applied clustering analysis generate spatial clusters. was followed by adopting conditioned Latin hypercube scheme within each cluster, sets strategically selected points. required improved estimates determined have sample size 15 40 cores, respective 13 36 km2 parcels. While modeling results concentrations at relatively homogeneous site eastern Montana showed significant two-fold improvement model fit individually calibration datasets point, opposed selecting dataset whole level, disparity between pixel- ranch-based models inconsequential other two sites Colorado that more spatially diverse terms vegetation cover. Compared concentration (R2 0.3 0.7), performance bulk density (BD) < 0.4) 0.2) poor. Strategies including utilizing subset covariates, incorporating broader-scale national depths did not further improve BD Future work should explore whether addition temporally dynamic environmental covariates can estimates, DSM-supported field strategy be successfully elsewhere.

Язык: Английский

---

Fine-resolution baseline maps of soil nutrients in farmland of Jiangxi Province using digital soil mapping and interpretable machine learning

CATENA, Год журнала: 2024, Номер 249, С. 108635 - 108635

Опубликована: Дек. 9, 2024

Язык: Английский

---

Contributions of Multi‐Method Geophysical Survey to Archaeological Research at the Battlefield of Waterloo

Archaeological Prospection, Год журнала: 2024, Номер 31(3), С. 267 - 287

Опубликована: Июль 1, 2024

ABSTRACT Archaeological prospection is continually expanding into new frontiers, examining increasingly large areas, diverse environmental contexts and varying site types. One area that has received only limited focus historic battlefields. This paper presents results from large‐scale geophysical surveys (> 100 ha) at the Napoleonic battlefield of Waterloo (1815) in Belgium, using fluxgate magnetometry frequency‐domain electromagnetic induction. Despite its international historical significance, professional archaeological research still infancy. We demonstrate how important insights can be gained by methods for identifying features artefacts related to battle developing an understanding various influences acting on present landscape. The largest survey kind undertaken a single site, this approach holds particular potential archaeology, given subtle low‐density nature sought‐after targets extensive site. Such mitigate (though not entirely resolve) challenges resolution scale associated with other investigation. Using representative range examples Waterloo, we consider successes undertaking sites. An integrated incorporates targeted sampling forms ancillary data emphasized more robust interpretation noninvasive sensor data.

Язык: Английский

---

A High-resolution Soil Organic Carbon Map for Great Britain

Sustainable Environment, Год журнала: 2024, Номер 10(1)

Опубликована: Окт. 29, 2024

Soil organic carbon (SOC) is used for soil health, indicating soils' agricultural productivity potential, and correlating with other functions like water capacity biodiversity. SOC stocks are increasingly recognized in climate change mitigation strategies. sequestration represents 25% of all natural solutions to carbon. Current maps Great Britain (GB) limited due their coarse resolution (0.5–1 km). High demand exists fine estimate stock baselines inform field-scale sampling strategies land management. We present concentration at 5 m GB, generated using machine learning accounting physical chemical properties, weather, topography cover (LC). Our model explains 74% variability the evaluation dataset, a RMSE 9.8 (tC ha− 1). pH LC most important predictors. ~ 2704Tg GB soil's top 30 cm: 1403Tg England, 283Tg Wales, 1017Tg Scotland. Neutral grasslands contribute England Wales (37.2% 50.7%). Dwarf heath shrubs, bogs have higher contributions Scotland (22–25%). estimation compares previous studies our map reflects expected distribution different parts under LCs. Its high spatial accuracy enable assessment small scales (a single farm or field) can help develop sustainable management guide by optimizing size cost.

Язык: Английский

---