Towards Optimal Variable Selection Methods for Soil Property Prediction Using a Regional Soil Vis-NIR Spectral Library DOI Creative Commons
Xianglin Zhang, Jie Xue, Yi Xiao

et al.

Remote Sensing, Journal Year: 2023, Volume and Issue: 15(2), P. 465 - 465

Published: Jan. 12, 2023

Soil visible and near-infrared (Vis-NIR, 350–2500 nm) spectroscopy has been proven as an alternative to conventional laboratory analysis due its advantages being rapid, cost-effective, non-destructive environmentally friendly. Different variable selection methods have used deal with the high redundancy, heavy computation, model complexity of using full spectra in spectral modelling. However, most previous studies a linear algorithm selection, application non-linear remains poorly explored. To address current knowledge gap, based on regional soil Vis-NIR library (1430 samples), we evaluated seven algorithms together three predictive predicting properties. Our results showed that Cubist overperformed partial least squares regression (PLSR) random forests (RF) properties (R2 > 0.75 for organic matter, total nitrogen pH) when spectra. Most can greatly reduce number bands therefore simplified models without losing accuracy. The also there was no silver bullet optimal among different algorithms: (1) competitive adaptive reweighted sampling (CARS) always performed best PLSR algorithm, followed by forward recursive feature (FRFS); (2) elimination (RFE) genetic (GA) generally had better accuracy than others algorithm; (3) FRFS performance RF algorithm. In addition, matched outcome this study provides valuable reference information spectroscopic techniques algorithms.

Language: Английский

Remote sensing of soil degradation: Progress and perspective DOI Creative Commons
Jingzhe Wang, Jianing Zhen, Weifang Hu

et al.

International Soil and Water Conservation Research, Journal Year: 2023, Volume and Issue: 11(3), P. 429 - 454

Published: March 15, 2023

Soils constitute one of the most critical natural resources and maintaining their health is vital for agricultural development ecological sustainability, providing many essential ecosystem services. Driven by climatic variations anthropogenic activities, soil degradation has become a global issue that seriously threatens environment food security. Remote sensing (RS) technologies have been widely used to investigate as it highly efficient, time-saving, broad-scope. This review encompasses recent advances state-of-the-art ground, proximal, novel RS techniques in degradation-related studies. We reviewed RS-related indicators could be monitoring properties. The direct (mineral composition, organic matter, surface roughness, moisture content soil) indirect proxies (vegetation condition land use/land cover change) evaluating were comprehensively summarized. results suggest these above are effective degradation, however, no system established date. also discussed RS's mechanisms, data, methods identifying specific phenomena (e.g., erosion, salinization, desertification, contamination). investigated potential relations between Sustainable Development Goals (SDGs) challenges prospective use assessing degradation. To further advance optimize technology, analysis retrieval methods, we identify future research needs directions: (1) multi-scale degradation; (2) availability data; (3) process modelling prediction; (4) shared dataset; (5) decision support systems; (6) rehabilitation degraded resource contribution technology. Because difficult monitor or measure all properties large scale, remotely sensed characterization related particularly important. Although not silver bullet, provides unique benefits studies from regional scales.

Language: Английский

Citations

128
Soil inorganic carbon, the other and equally important soil carbon pool: Distribution, controlling factors, and the impact of climate change DOI
Amin Sharififar, Budiman Minasny, Dominique Arrouays

et al.

Advances in agronomy, Journal Year: 2023, Volume and Issue: unknown, P. 165 - 231

Published: Jan. 1, 2023

Language: Английский

Citations

58
A mobile-based system for maize plant leaf disease detection and classification using deep learning DOI Creative Commons
Faiza Khan, Noureen Zafar, Muhammad Naveed Tahir

et al.

Frontiers in Plant Science, Journal Year: 2023, Volume and Issue: 14

Published: May 15, 2023

Artificial Intelligence has been used for many applications such as medical, communication, object detection, and tracking. Maize crop, which is the major crop in world, affected by several types of diseases lower its yield affect quality. This paper focuses on this issue provides an application detection classification maize using deep learning models. In addition to this, developed also returns segmented images leaves thus enables us track disease spots each leaf. For purpose, a dataset three named Blight, Sugarcane Mosaic virus, Leaf Spot collected from University Research Farm Koont, PMAS-AAUR at different growth stages contrasting weather conditions. data was training prediction models including YOLOv3-tiny, YOLOv4, YOLOv5s, YOLOv7s, YOLOv8n reported accuracy 69.40%, 97.50%, 88.23%, 93.30%, 99.04% respectively. Results demonstrate that model higher than other applied shown excellent results while localizing area leaf accurately with confidence score. latest compared approaches available literature. Also, worked sugarcane mosaic virus first time. Further, high have embedded mobile provide real-time facility end users within few seconds.

Language: Английский

Citations

45
Improving model parsimony and accuracy by modified greedy feature selection in digital soil mapping DOI Creative Commons
Xianglin Zhang, Songchao Chen, Jie Xue

et al.

Geoderma, Journal Year: 2023, Volume and Issue: 432, P. 116383 - 116383

Published: Feb. 24, 2023

In the context of increasing soil degradation worldwide, spatially explicit information is urgently needed to support decision-making for sustaining limited resources. Digital mapping (DSM) has been proven as an efficient way deliver from local global scales. The number environmental covariates used DSM rapidly increased due growing volume remote sensing data, therefore variable selection necessary deal with multicollinearity and improve model parsimony. Compared Boruta, recursive feature elimination (RFE), variance inflation factor (VIF) analysis, we proposed use modified greedy (MGFS), regression. For this purpose, using quantile regression forest, 402 samples 392 were map spatial distribution organic carbon density (SOCD) in Northeast North China. result showed that MGFS selected most parsimonious only 9 (e.g., brightness index, mean annual temperature), much lower than RFE (22 covariates), VIF (30 Boruta (76 covariates). repeated validation (50 times) derived performed better (R2 0.60, LCCC 0.74, RMSE 13.80 t ha−1) these full covariates, 0.48–0.57, 0.64–0.72, 14.24–15.79 ha−1). Despite similar performance uncertainty estimate (PICP), had lowest (0.86) indicated by index. addition, best computation efficiency when considering steps prediction. Given advantages over VIF, a high potential fine-resolution practices, especially studies at broad scale involving heavy on millions or billions pixels.

Language: Английский

Citations

44
Validation of uncertainty predictions in digital soil mapping DOI Creative Commons

Jonas Schmidinger,

G.B.M. Heuvelink

Geoderma, Journal Year: 2023, Volume and Issue: 437, P. 116585 - 116585

Published: July 11, 2023

It is quite common in digital soil mapping (DSM) to quantify the uncertainty of issued predictions, that make probabilistic predictions. Yet, little attention has been paid its validation. Probabilistic predictions are only value for end users if they reliable and ideally also sharp. Reliability refers consistency between predicted conditional probabilities observed frequencies independent test data. Sharpness concentration a probability distribution function, i.e. narrowness. The prediction interval coverage (PICP) currently used DSM validate reliability intervals but it ignorant potential one-sided bias boundaries. Therefore, we propose extend current validation procedure with metrics broader literature. These not evaluate format quantiles or full distributions. We suggest quantile (QCP) integral transform (PIT) histogram as alternatives PICP proper scoring rules relative comparisons competing models. As rules, present score (IS) continuous ranked (CRPS), which can be decomposed into part (RELI). illustrated use these case study using pH organic carbon from LUCAS-soil database. Thereby, five different models were compared: reference null model (NM), regression forest (QRF), post-processing random (QRPP RF), kriging external drift (KED) neural network (QRNN). For KED QRNN, was found. This apparent shown by PIT QCP. RELI summarized trends found QCP, histograms one numerical value. CRPS IS especially harsh outliers low sharpness. According IS, best obtained QRF QRPP RF worst NM.

Language: Английский

Citations

42
A high-resolution map of soil organic carbon in cropland of Southern China DOI
Bifeng Hu, Modian Xie, Yue Zhou

et al.

CATENA, Journal Year: 2024, Volume and Issue: 237, P. 107813 - 107813

Published: Jan. 12, 2024

Language: Английский

Citations

22
SHAP values accurately explain the difference in modeling accuracy of convolution neural network between soil full-spectrum and feature-spectrum DOI
Liang Zhong, Guo Xi, Meng Ding

et al.

Computers and Electronics in Agriculture, Journal Year: 2024, Volume and Issue: 217, P. 108627 - 108627

Published: Jan. 13, 2024

Language: Английский

Citations

21
The effect of covariates on Soil Organic Matter and pH variability: a digital soil mapping approach using random forest model DOI Creative Commons
Yassine Bouslıhım, Kingsley John, Abdelhalim Miftah

et al.

Annals of GIS, Journal Year: 2024, Volume and Issue: 30(2), P. 215 - 232

Published: Jan. 29, 2024

This research focuses on understanding the spatial variation of Soil Organic Matter (SOM) and pH levels in North Morocco. The study employs a comprehensive approach to enhance predictive modelling, incorporating Boruta algorithm for effective environmental covariates selection optimizing model parameters through hyperparameter optimization. Utilizing Random Forest (RF) with remote sensing indices topographic features, predicts SOM identify key contributors their variability. prediction saw significant success, notable correlation such as RVI, NDVI, TNDVI. These indices, indicative vegetation health productivity, emerged primary influencers SOM. In comparison, influence features like elevation, slope, aspect was found be less significant. Conversely, predicting challenging due minimal variability within dataset. Addressing this limitation could involve dataset expansion or alternative models low-correlated data handling. Despite RF model's limited efficacy prediction, an observable between identified, consistent prior research. Areas higher exhibited lower values, indicating relative soil acidification from organic matter decomposition. study's demonstrated potential using but enhancing is essential. Future may explore expansion, diverse sampling, testing better performance datasets. offers valuable insights advanced development enriches management practices.

Language: Английский

Citations

17
The prevalent life cycle of agricultural flash droughts DOI Creative Commons
Miguel A. Lovino, M. Josefina Pierrestegui, Omar V. Müller

et al.

npj Climate and Atmospheric Science, Journal Year: 2024, Volume and Issue: 7(1)

Published: March 19, 2024

Abstract This work examines the characteristics and prevalent life cycle of agricultural flash droughts globally. Using ERA5 data, study introduces a drought indicator based on soil water availability. approach integrates root-zone moisture hydraulic properties, such as field capacity wilting point, to couple rapid depletion plant stress. Our findings reveal that present their higher frequency predominantly during critical growth periods crops. Notably, these exhibit similar regardless location or climatic regime. The primary cause is precipitation deficit, but evapotranspiration also plays significant role. In an energy-limited environment, rapidly increases before onset decreases intensification period system becomes water-limited. Upon concluding period, most crops experience stress, diminishing yields.

Language: Английский

Citations

16
Creating soil districts for Australia based on pedogenon mapping DOI Creative Commons

Quentin Styc,

Julio Pachόn,

Wartini Ng

et al.

Geoderma, Journal Year: 2025, Volume and Issue: 454, P. 117164 - 117164

Published: Jan. 10, 2025

Language: Английский

Citations

2