Hg and As pollution in the soil-plant system evaluated by combining multispectral UAV-RS, geochemical survey and machine learning

Environmental Pollution, Год журнала: 2023, Номер 333, С. 122066 - 122066

Опубликована: Июнь 19, 2023

The combination of a low-density geochemical survey, multispectral data obtained with Unmanned Aerial Vehicle-Remote Sensing (UAV-RS), and machine learning technique was tested in the search for statistically robust prediction contaminant distribution soil vegetation, zones highly variable pollutant load. To this end, novel methodology devised by means limited study topsoil vegetation combined UAV-RS. verified an area affected Hg As contamination that typifies abandoned mining-metallurgy sites recent decades. A broad selection spectral indices were calculated to evaluate soil-plant system response, four techniques (Multiple Linear Regression, Random Forest, Generalized Boosted Models, Multivariate Adaptive Regression Spline) obtain statistical models. Forest (RF) provided best non-biased models concentration R2 rRMSE (%) ranging from 0.501 0.630 180.72 46.31, respectively, acceptable values RPD RPIQ statistics. mapping content well enough adjusted revealed superior accuracy than Hg, topsoil. results more precise those comparable studies applied satellite or spectrometry data. In conclusion, presented emerges as powerful tool addressing pollution alternative approach classical studies, which are time-consuming expensive.

Язык: Английский

---

Multi-Scale Stereoscopic Hyperspectral Remote Sensing Estimation of Heavy Metal Contamination in Wheat Soil over a Large Area of Farmland

Agronomy, Год журнала: 2023, Номер 13(9), С. 2396 - 2396

Опубликована: Сен. 16, 2023

With the rapid development of China’s industrialization and urbanization, problem heavy metal pollution in soil has become increasingly prominent, seriously threatening safety ecosystem human health. The hyperspectral remote sensing technology provides possibility to achieve non-destructive monitoring contents. This study aimed fully explore potential ground satellite image spectra estimating We chose Xushe Town, Yixing City, Jiangsu Province as research area, collected samples from farmland over two different periods, measured contents metals Cd As laboratory. At same time, under field conditions, we also wheat leaves obtained HuanJing-1A HyperSpectral Imager (HJ-1A HSI) data. first performed various spectral transformation pre-processing techniques on leaf Then, used genetic algorithm (GA) optimized partial least squares regression (PLSR) establish an estimation model contents, while evaluating accuracy model. Finally, best models drew spatial distribution maps area. results showed following: (1) can highlight some hidden information spectra, including mathematical transformations such differentiation; (2) modeling, GA-PLSR higher than PLSR, using a GA for band selection improve model’s stability; (3) provide good ability estimate (relative percent difference (RPD) = 2.72) excellent (RPD 3.25); HJ-1A HSI only distinguishing high low values 1.87, RPD 1.91). Therefore, it is possible indirectly data, identify areas key pollution.

Язык: Английский

---

Shaping the concentration of petroleum hydrocarbon pollution in soil: A machine learning and resistivity-based prediction method

Journal of Environmental Management, Год журнала: 2023, Номер 345, С. 118817 - 118817

Опубликована: Авг. 17, 2023

Язык: Английский

---

Machine Learning in Geosciences: A Review of Complex Environmental Monitoring Applications

Machine Learning and Knowledge Extraction, Год журнала: 2024, Номер 6(2), С. 1263 - 1280

Опубликована: Июнь 5, 2024

This is a systematic literature review of the application machine learning (ML) algorithms in geosciences, with focus on environmental monitoring applications. ML algorithms, their ability to analyze vast quantities data, decipher complex relationships, and predict future events, they offer promising capabilities implement technologies based more precise reliable data processing. considers several vulnerable particularly at-risk themes as landfills, mining activities, protection coastal dunes, illegal discharges into water bodies, pollution degradation soil matrices large industrial complexes. These case studies about provide an opportunity better examine impact human activities environment, specific matrices. The recent underscores increasing importance these contexts, highlighting preference for adapted classic models: random forest (RF) (the most widely used), decision trees (DTs), support vector machines (SVMs), artificial neural networks (ANNs), convolutional (CNNs), principal component analysis (PCA), much more. In field management, following methodologies invaluable insights that can steer strategic planning decision-making accurate image classification, prediction models, object detection recognition, map variable predictions.

Язык: Английский

---

Effects of Chinese “double carbon strategy” on soil polycyclic aromatic hydrocarbons pollution

Environment International, Год журнала: 2024, Номер 188, С. 108741 - 108741

Опубликована: Май 11, 2024

Polycyclic aromatic hydrocarbons (PAHs) and carbon dioxide primarily originate from the combustion of fossil fuels biomass. The implementation Chinese "double strategy" is expected to impact distribution PAH emissions, consequently influencing spatial trend PAHs in surface soil. Therefore, it crucial quantitatively evaluate effectiveness on soil pollution for purpose "the reduction emissions". This study utilized 15,088 individual concentration data 943 samples collected between 2003 2020 China, conjunction with emissions at a 10 km resolution, meta-analysis. calculated this are line global emission inventory (PKU-PAH-2007), relative standard deviation provincial level less than 25 %. Subsequently, novel method was developed using density K

Язык: Английский

---